Citation
A novel electrospray ion trap mass spectrometer for photodissociation of biological molecules

Material Information

Title:
A novel electrospray ion trap mass spectrometer for photodissociation of biological molecules
Creator:
Stephenson, James Lee, Jr., 1961-
Publication Date:
Language:
English
Physical Description:
ix, 428 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Analyzers ( jstor )
Electric fields ( jstor )
Electric potential ( jstor )
Electrodes ( jstor )
Ion traps ( jstor )
Ions ( jstor )
Lasers ( jstor )
Mass spectroscopy ( jstor )
Photolysis ( jstor )
Quadrupoles ( jstor )
Chemistry thesis, Ph. D
Dissertations, Academic -- Chemistry -- UF
Mass spectrometry ( lcsh )
Photodissociation ( lcsh )
City of Gainesville ( local )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1995.
Bibliography:
Includes bibliographical references (leaves 405-426).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by James Lee Stephenson, Jr.

Record Information

Source Institution:
University of Florida
Rights Management:
Copyright James Lee Jr. Stephenson. 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:
022247319 ( ALEPH )
34351142 ( OCLC )

Downloads

This item has the following downloads:


Full Text









A NOVEL ELECTROSPRAY ION TRAP MASS SPECTROMETER FOR
PHOTODISSOCIATION OF BIOLOGICAL MOLECULES


Bil

JAMES LEE STEPHENSON, JR.


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 1995



UNIVERSITY OF FLORTDA L1,R ART


1780

J)95

























"I have little patience with scientists who take a board of wood, look for its thinnest part, and drill a great number of holes where drilling is easy."

Albert Einstein in "Einstein's Philosphy of Science" Reviews of Modern Physics 1949, 21, no. 3.



"Just work hard and be honest, and everything else will take care of itself..."

Orson Calvin "O.C." Pearson ....my Grandfather













ACKNOWLEDGMENTS


I would like to begin by expressing my unfeigned gratitude to Dr. Richard A. Yost for allowing me to pursue my own research ideas (no matter how bad) over the last five years. It has been a most satisfying experience personally and professionally. I also would like to thank the following members of my committee: Dr. John Gander for his insightful comments on carbohydrate chemistry; Dr. Jim Winefordner for the endless jokes; Dr. Bob Kennedy for the career discussions on academic life; Dr. John Eyler for convincing me that a modified ring electrode would actually work; and Dr. Dave Powell for those grueling 10 mile Sunday runs and brutal Tuesday night track workouts. My survival here was also due in no small part to Jeanne Karably, Susan Ciccarone, and Donna Balkcom, all of whom pointed me in the right direction, provided me with the right paperwork, and told me where to be in order to graduate.

Financial support for this research came from a variety of sources including the University of Florida Division of Sponsored Research, the Office of Naval Research, the ACS Analytical Division Fellowship (sponsored by The Procter & Gamble Company), and a Dissertation Fellowship from the College of Liberal Arts and Sciences at the University of Florida.


iii








A great deal of credit for the success of this project goes to Matt Booth who has been my collaborator, co-worker, and good friend; from San Jose to Gainesville, Matt always had an idea, a beer, or a word of encouragement that kept me going. This dissertation also would not have been possible without the Herculean design and machining effort put forth by Joe Shalosky; he is truly an artist. In addition, I would like to acknowledge Scott Quarmby for assistance with electronic design of the rf circuitry and Stephen Boud for the set-up/operation of the pulsed CO2 laser.

My "Yost Group" experience was made possible by a whole cast of characters including Uli Bemier, Tim Griffin, Tracie Williams, Shannan Carlson, and John Laycock, just to name a few. To the "Night Shift of the Yost Lab," Nathan Yates, Don Eades, Jodie Johnson, and Brad Coopersmith, I am truly grateful for the late-night beers at the Salty Dog and the "Hut" experience, which made writing this dissertation a much less painful process.

I wish to thank my parents, James L and Vivian P. Stephenson, for instilling in me the value of education, hard work, and humility. Their unconditional support during the last five years has been invaluable.

Lastly, I wish to thank my wife Tracy (a.k.a Chica) for her love, support, and understanding. She organized the weekend getaways, the La Chua trail walks, B.E. breaks, and parties which added spice to our lives and memories to hang on to.


iv













TABLE OF CONTENTS



ACKNOW LEDGMENTS ..................................... iii

ABSTRACT .............................................. viii

CHAPTERS

1 INTRODUCTION ...................................... 1

The Quadrupole Ion Trap Mass Spectrometer ................. 3
H istory .. . .. . .. .. .. . . . . .. . . . . .. . 3
Theoretical Aspects of Ion Motion ..................... 7
Ion Activation .............. ................... 17
General Operation .............................. 21
Electrospray Ionization ................................. 28
Basic Principles of ion Formation .................... 30
Molecular Weight Determination ..................... .35
Photodissociation .................................... 36
Infrared Multiple Photon Dissociation (IRMPD) ........... .38
The Photon Absorption Process ..................... .41

2 FUNDAMENTAL INVESTIGATIONS OF IRMPD IN THE
QUADRUPOLE ION TRAP ......................... .47

Instrum entation ...................................... 47
Experimental Design ............................. 48
Ion Trap Operation without Helium Buffer Gas ........... .57
The Multi-Pass Ring Electrode ...................... 63
Effects of Ion Motion on Photodissociation
Efficiency .................................... 70
Dipolar Excitation ................................ 72
Quadrupolar Excitation ............................ 82
The Photon Absorption Process .......................... 93
IRM PD Kinetics ...................................... 97
Consecutive Reactions ................................ 100
Buffer Gas Effects .................................. 116


v








Wavelength Dependence/nfrared Spectroscopy of
Gas-Phase Ions ................................ 120

3 THE RF-ONLY OCTOPOLE ION TRANSMISSION GUIDE ....... .129

Ion Behavior in Electromagnetic Fields .................... 131
Octopole Electrode Arrangement ........................ 138
Octopole Field Potentials .............................. 141
Electric Field Strength Calculations .................. 145
Electrode Geometry ............................ 149
Equations of Motion ............................ 150
Octopole Design Considerations ........................ 155
Effective Trapping Potential ....................... 157
Assembly and Construction ....................... 161
Factors Affecting Ion Transmission ....................... .169
Initial Entry Angle Conditions ...................... 172
RF Am plitude ................................ 180
RF Frequency ................................ 183
Kinetic Energy ................................ 189
Collisional Focusing ............................ 194

4 ELECTROSPRAY/lON TRAP INSTRUMENTATION: DESIGN
AND OPERATION .............................. 196

General Overview .................................. 196
Instrument Design .................................. 196
Vacuum Manifold and Pumping System .............. 197
Electrospray Ion Source .......................... 204
RF-Only Octopole .............................. 210
Analyzer Assembly .............................. 226
Detector Assembly .............................. 243
Photodissociation Set-Up ......................... 246
System Interconnections .............................. 254
Instrument Characterization ............................ 258
Ion Trap High Mass Theory/Operation ................ 263
Basic ESI Operation ............................ 270
Octopole RF Level .............................. 282
Octopole Offset ................................ 287
Ion Gate Lens ................................ 291
RF Level/Ion Injection ............................ 294
Ion Isolation .................................. 303
Collision-induced Dissociation (CID) ................. 309
Negative Ion Mode ............................. 322


vi








5 PHOTODISSOCIATION OF BIOLOGICALLY IMPORTANT
MOLECULES: PROTEINS, CARBOHYDRATES,
AND OLIGONUCLEOTIDES ....................... 328

General Overview of Structural Elucidation ................. 328
Peptides and Proteins ................................ 338
Human Angiotensin I ............................ 338
Gram icidin D .................................. 342
Carbohydrates and Oligosaccharides ..................... 343
Monosaccharide Cleavage ........................ 348
Raffinose ..................................... 363
Stachyose .................................... 374
O ligonucleotides .................................... 377
RNA Dim ers .................................. 382
Carbohydrate Antibiotics .............................. 389
Macrolide Antibiotics Erythromycin ................. 390

6 CONCLUSIONS AND FUTURE WORK .................... .399

REFERENCE LIST ........................................ 405

BIOGRAPHICAL SKETCH .................................. 427


vii













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

A NOVEL ELECTROSPRAY ION TRAP MASS SPECTROMETER FOR
PHOTODISSOCIATION OF BIOLOGICAL MOLECULES By

James L Stephenson, Jr.

December, 1995

Chairperson: Dr. Richard A. Yost
Major Department: Chemistry

The combined techniques of photodissociation and mass spectrometry have been used extensively to study the fundamental aspects of gas-phase ion chemistry. Within the last several years, a great deal of interest has been shown in employing photodissociation as an analytical tool for the structural elucidation of biological molecules, due in part to the limitations of traditional tandem mass spectrometric techniques (e.g. collision-induced dissociation) in providing relevant structural information. This dissertation presents the design and characterization of an electrospray ion trap mass spectrometer, capable of performing photodissociation experiments on a wide range of biological molecules.

Previous photodissociation studies have been limited to fundamental investigations due mainly to the limited photoabsorption cross-sections observed for organic ions. In order to increase the photodissociation efficiency, three


viii








spherically symmetric mirrors were placed in the radial plane of the ion trap ring electrode to increase the photoabsorption pathlength of the incident radiation. The performance of this multipass ring electrode was characterized by conducting experiments on ion motion, kinetics, consecutive reactions, spectroscopy and buffer gas pressure to examine their effects on photodissociation efficiency of a series of model compounds.

Next, a novel ion injection system was designed and built in order to transfer biological ions from the electrospray source to the ion trap mass spectrometer in the most efficient manner possible. The instrumentation consisted of a modified electrospray interface coupled to an rf-only octopole, which transmitted ions from a high pressure region near the ion source to the lower pressure region of the ion trap analyzer. The theory, design, and characterization of the rf-only octopole is discussed, along with an analysis of the ion transmission properties of the device.

Finally, a detailed investigation into the photodissociation of ions from carbohydrates, peptides/proteins and oligonucleotides is presented. Sensitivity considerations, fragmentation patterns, and comparison studies to collisional activation are discussed. An analysis of the type of structural information obtained from photodissociation showed its applicability in solving biological problems.


ix













CHAPTER 1
INTRODUCTION


Over the last decade, perhaps the most important advancement in quadrupole ion trap mass spectrometry has been that of tandem mass spectrometry or MS" for structural elucidation of organic ions. The most frequently used method for the activation of these ions has been collisional activation, commonly known as collision-induced dissociation (CID).' The major factors that contribute to the success of CID experiments in the quadrupole ion trap mass spectrometer (QITMS) include the ability to perform tandem-in-time as opposed to tandem-in-space MS/MS experiments, the efficient conversion of parent ions to product ions (typically 10-50%), and, most importantly, the high collision cross sectional area observed for CID (on the order of 10 to 200 A2). 1These advantages arise in part because in the quadrupole ion trap, uniquely amongst tandem mass spectrometers, kinetic energy is imparted to the parent ions only between collisions.

To date, the majority of the published applications using the QITMS for structural elucidation has employed collisional-activation as the method of choice for ion activation.4' More recently, the attention of many researchers has focused on the fundamental understanding of the collisional-activation process in the QITMS.' Even though collisional-activation cannot provide all the answers in


1







2

tandem (MS/MS") mass spectrometry, there has only been a limited effort to investigate alternative techniques for ion activation in trapping instruments (including ion cyclotron resonance and quadrupole ion trap mass spectrometers).

This dissertation presents an alternative method for the activation of polyatomic ions in the gas phase, that of photon absorption or what is frequently called photo-induced dissociation (PID). Photodissociation has been used extensively by physical chemists to study fundamental properties of gas-phase ions. The combination of mass spectrometry (employing both ion trap and ion cyclotron resonance instruments) and photodissociation has been used successfully to investigate the chemical kinetics, reactivity, and spectroscopy of various ionic species. The long storage times and instrumental configuration of trapping instruments are ideally suited for photodissociation experiments. Some advantages of trapping instruments include the measurement of photon-induced ion decay as a function of laser irradiance time, the use of the multiphoton absorption processes to study fragmentation, and the use of the photodissociation spectrum as a fingerprint for determination of isomeric ion structures.

This introductory chapter begins with the relevant history of the quadrupole ion trap mass spectrometer, followed by a brief discussion of ion motion, the principles of ion activation, and general QITMS operation procedures. Since the technique of electrospray ionization (ESI) is used to generate gas-phase ions for the biological studies presented in this dissertation, an introduction to the basic







3

principles of ESI also is included. The chapter concludes with a discussion of photodissociation that covers the basics of the photon absorption mechanism and provides an introduction to the infrared multiple photon dissociation (IRMPD) process.


The Quadrupole Ion Trap Mass Spectrometer


History


The origin of the quadrupole ion trap mass spectrometer dates back to the original patent of Paul and Steinwedel, which disclosed the operation of the quadrupole mass filter and the quadrupole ion trap.' The first studies by Paul and coworkers centered on the use of the device for ion storage over long periods of time.8 Early detection procedures measured the power absorption of stored ions, which utilized an rf voltage applied to both endcaps. In 1959, the original mass selective detection studies were performed by Fischer, who demonstrated the ability of the device to obtain unit mass resolution for a series of krypton isotopes.9 In the early to mid sixties, spectroscopic studies of ions by Dehmelt and Majors showed that high resolution studies were possible for all ground state, metastable, atomic, and molecular ions. 102 The first mass discrimination experiments were performed in 1968 by Dawson and Whetten.'3 This led to the first use of the quadrupole ion trap as a true mass spectrometer. These experiments utilized an external detector (electron multiplier) for the detection of ions ejected through holes in the endcap electrodes. In 1971, it was discovered







4

that a combination of both rf and dc voltages applied to the ring electrode produced trapping conditions that were favorable for mass-selective storage.'" The storage conditions used in these experiments were defined by the upper and lower apices of the Mathieu stability diagram.4

The combination of a quadrupole ion trap (used as an ion source) and a quadrupole mass filter was used extensively by Todd, Lawson, and Bonner for the analysis of ejected ions from the ion trap, the general characterization/behavior of trapped ions, chemical ionization, and ion molecule kinetics."''7 This hybrid instrument, called the QUISTOR (or QUadrupole Ion STORe), was operated with the endcaps held at ground potential and a combination of rf and dc voltages applied to the ring electrode. An electron gate was used to pulse electrons into the center volume of the ion trap where ionization of the neutral sample gas(es) occurred. The detection event consisted of a dc pulse applied to one or both endcap electrodes, which extracted ions out towards the quadrupole mass filter. The late 1970s also saw the development of mass-selective isolation by Fulford and March.18 Here, the ion of interest was moved to the apex point in the stability diagram where all ions of smaller m/z were unstable in the axial z-direction and all ions of higher m/z were unstable in the radial r-direction. In 1979, Fulford et al. performed the first experiments that involved resonance excitation at a given ion population's unique frequency of motion. This caused either ion ejection from the trap or collision-induced dissociation of the ions of interest."







5

In the early 1980s, a landmark series of papers by Hughes, March, and Young first demonstrated the use of IR photodissociation in a quadrupole ion trap.2" The technique of infrared multiple photon dissociation was used for studying the gas-phase ion chemistry of the proton-bound dimer of 2-propanol. Other early IRMPD experiments focused on wavelength dependence, collisional cooling, dissociation efficiency, and analyte pressure dependence of various gasphase systems.2 A more detailed discussion of the instrumentation, experimental parameters, and theoretical aspects of IRMPD in the quadrupole ion trap appears in the photodissociation section of this chapter.

Perhaps the most important development in quadrupole ion trap mass spectrometry was the mass selective instability scan developed by Stafford et al.24.25 This mode of operation was radically different from previous modes in that the rf amplitude applied to the ring electrode was ramped linearly with respect to time. This enabled ions of increasingly higher mass-to-charge ratio (m/z) values to be sequentially ejected from the ion trap. A mass spectrum could then be recorded as a function of the time it takes ions of various m/z to be ejected to the detector. Furthermore, by pressurizing the ion trap analyzer with approximately 10- torr of a light buffer gas (helium or hydrogen), drastic improvements in mass resolution, sensitivity, and dynamic range were obtained for ion trap operation.2 The mechanism of action for these improvements involved the collision of ions with the relatively slow and much less massive background gas atoms or







6

molecules. These collisions caused viscous damping of ionic motion, thereby focusing the ion cloud to the center of the trap.

These improvements led to the first commercially available ion trap detector (lTD 700) developed by Finnigan MAT. This product was designed as a low cost benchtop GC/MS detector, which gained popularity for trace level environmental and clinical analysis. Further improvements in dynamic range were obtained with automatic gain control (AGC), which regulated the number of ions stored in the trap as a function of sample concentration.Y This limited space charge effects and ion-molecule reactions typically seen with constant ionization time experiments.

The first fully functional research-grade ion trap mass spectrometer was developed by Kelley et al. in 1985.2 The Finnigan MAT Ion Trap Mass Spectrometer (ITMS) possessed a wide range of experimental capabilities including electron ionization (El), chemical ionization (CI), mass isolation (e.g., apex isolation), tandem MS", and user-defined software. With the advent of resonance ejection (axial modulation) in 1988, further improvements in resolution and dynamic range were achieved.3* The technique of resonance ejection was also responsible for the extension of the mass range of the quadrupole ion trap to well beyond m/z 50,000.

Rapid growth in the field of ion trap mass spectrometry over the last six to seven years has been driven by the coupling of external ion sources to the device. Some of the external ion sources coupled to the ITMS include fast atom







7

bombardment (FAB)", glow discharge (GD)3'", electron and chemical ionization (El/Cl)", and electrospray ionization (ESI).3 The discovery of high resolution ion trap mass spectrometry by Schwartz et al., combined with advances in high mass analysis published previously, has enabled ion trap mass spectrometry to take a lead role in the analysis of biomolecules.3' The advances discussed in this dissertation address many of the issues currently limiting ion trap mass spectrometry in the analysis of biomolecules. Other recent advances in the field include the use of broadband waveforms (stored waveform inverse Fourier transforms, SWIFT) for MS/MS analysis and for mass isolation. In addition, alternative techniques to scanning and detection of ions from those traditionally used over the last 10 years have been developed.'

Several reviews published over the last few years cover many of the aforementioned developments in greater detail."4 Several books encompassing ion trap mass spectrometry as well as biological mass spectrometry, cover a range of topics from basic instrumental principles to applications development.4'* Theoretical Aspects of Ion Motion


A fundamental understanding of ion motion in the quadrupole ion trap is important in evaluating various ion activation techniques (collisional activation or surface-induced dissociation, frequently called SID). Some of these evaluation criteria include dissociation efficiencies, understanding ion-neutral collisions, and collisions of ions with surfaces. Therefore, a fundamental comprehension of ion







8

motion is important for describing the conversion of an ion's kinetic energy (for CID and SID) of motion into the vibrational energy needed to accomplish fragmentation in an MS/MS experiment. For efficient photodissociation, the amount of time required for sufficient absorption of a photon(s) is directly related to the overlap of the ion trajectory with the incident radiation. Consequently, it is important to grasp the basic concepts of ion motion and how they might be applied to the interaction between ions and light.

The motion of an ion in a quadrupole field can be described mathematically by the Mathieu equation, a second-order linear differential equation originally used to characterize the vibrational motion of a stretched skin or membrane." The derivation begins by assuming the presence of an ideal quadrupolar field (no space charge effects due to other ions), defined by 0, the potential at any point (x,y,z) in that field:


4= ()x +y2 +yz2) (1_1)
2
ro

where oo is the applied electric potential, A, o, and y are a weighting constants in the x, y and z directions respectively, and ro is the inscribed radius of the ring electrode. Any oscillatory and dc potentials applied to the ring electrode can be represented mathematically in the form:







9


*O=U-VO_, cos 0 t (1-2)

where a is the angular frequency of the rf trapping field applied to the ring electrode (in rad s-, which is equal to 2nf where f is the frequency in Hz), U is the applied dc voltage, VO. is the zero-to-peak amplitude of the rf voltage, and t is the time variable.2"

For any quadrupolar device, the field is uncoupled in the three coordinate directions, so that the forces acting on an ion are independent of one another. For this condition to hold, equation 1-1 must satisfy the LaPlace condition where the field strength at the center of the ion trap must equal zero. Therefore, when the potential 0 is applied to the ring electrode (xy plane) and the two hyperbolic endcaps are held at ground potential, the potential at any point 0 within the device is represented by equation 1 1.23

Because the geometry of the quadrupole ion trap has cylindrical symmetry, the x and y components are combined to give a single radial component defined by x2+y2=r2. The orientation in space of the ring electrode and two hyperbolic endcaps needed to define a quadrupolar trapping field is r.2=2z2, where ro is the radius of the ring electrode and z is the center-to-endcap distance.-2 The hyperbolic shape of the electrodes in this geometrical configuration can be seen in figure 1-1. The equations defining the hyperbolic shape of the ring electrode and endcap electrodes are given as:






















Figure 1-1:


The geometric configuration of the quadrupole ion trap. Shown are the three hyperbolic electrodes which comprise the analyzer. The center-to-endcap and the center-to-ring distances are indicated by zo and r, respectively.

















Filament Endcap






Ring Electrode






Exit Endcap


-A


10





zo

ro





AV








12


1 (r2 -2z 2)= 1 (1-3)
r0


1 (r2-2z2)-1 (1-4)
2z:

The equations of motion for an ion in both the r- and z-directions can be derived from the forces exerted on the ion independently in each direction. By applying Newton's second law of motion, substituting equation 1-2 for 00 in equation 1-1, applying the condition x2+y2=r2, and differentiating, the equations of motion in the r- and z-direction are obtained:


d t2 2
_+_2+ (U-Vcos~t)r=O (1-5)
dt2 2mro



d + 4e (U-Vcos~t)z=O (1-6)
dt2 2mr2


where m is the mass of the ion of interest and t is the time variable. These equations are examples of the Mathieu equations, whose solutions have been studied extensively.23 The general form of the Mathieu equation can be expressed as:


+(au 2qcos2 )u=0 (1-7)
d t2

where u represents r or z, and "(=Qt/2. By performing a series of operations and substitutions, the parameters au and qu can be evaluated for the quadrupole ion







13

trap. These parameters determine whether or not ion motion is stable or unstable. For motion in both the r and z direction the values of ar and q,, become:


a =-2ar = (1-8)
m(r +2z 1)Q2


qz=2qr= 4eV
m(r2+2z_)02
The Mathieu stability parameters au and qu are directly proportional to the applied dc and rf voltages on the ring electrode. Since au and q, are inversely proportional to m/z, the motion of all ions can be expressed in terms of these Mathieu stability parameters.

In equations 1-8 and 1-9, the relationship the between axial (zo) and radial

(r.) dimensions of the ion trap are defined theoretically by z0=r0/2"2. However, all commercially manufactured ion traps to date posses a stretched geometry in the z-direction. Therefore, instead of a theoretically defined z4 distance of 0.707 cm, the ion trap is stretched axially 10.7% with a new zo value of 0.783 cm.3 In a previous report by Johnson et al.', the general form of equations 1-8 and 1-9 (also known as the Knight equations) were found to be a good approximation for experimental measurements to determine q,. and a, values made with the stretched ion trap geometry.

The stability diagram, which specifies stable ion trajectories, can then be defined as the intersection of the solutions to the Mathieu equation in the z-







14

direction with those in the r-direction. In this overlap region, ions are stable in both the z and r directions and can be stored in an ion trap. The overlap region which is closest to the a, q origin is of the most practical significance, since the voltages are less and thus the electronics needed to apply the appropriate rf and dc voltages to the ring electrode can be readily designed. Figure 1-2 shows the stability diagram for the three-dimensional quadrupole ion trap." The ordinate and abscissa are expressed in terms of the dimensionless quantities a and q respectively.

Ions within the stability region exist in a pseudopotential well, with a depth of about 1 eV near the origin and 10 eV near the right hand (z) stability edge. The lines drawn across the stability diagram are called iso-6, lines which describe detailed trajectories of ions at that particular point. Iso-6 lines essentially determine the frequency of ion motion; they are found in the general solution to the Mathieu equation:


u(Z)=A Cn cos(2n+6)Z+B EC2nsin(2n+6) (1-10)

where the Cn coefficients are the amplitude components of oscillation and the (2n+6) terms represent the frequency components of ion motion. These quantities can be calculated for the corresponding values of a and q using recurrence relations/continued fractions." The relationship between ion frequency (wj and 6 can be defined as follows:




























Figure 1-2:


Mathieu stability diagram for the quadrupole ion trap. The dimensionless quantity q. is directly proportional to the applied rf voltage, while the dimensionless quantity az is directly proportional to the applied dc voltage. The 6. lines are a direct relation to the secular frequency of motion for a given ion. Adapted from reference 23.








16





0.2 1.0 0
0.8 0.1
00. 0.6 0.7
0.1 -- 6. 0.2



-0..4
0.2 0.3 0.3
0
0.4

-0.1
0.5

-0.2 --6r
-.6

-0.3
00.7

-0.40.0
0.8

-0.5 --.9


-0.6


-0.7
0.2 0.4 0.6 0.8 1.0 1.2 1.4


qz







17


ne = (2n + 6U) 12(1-11) where 0<56_<51 and n=0, 1, 2, 3..... Consequently, the main secular frequency (n=0) is calculated as 6UD/2.23

As mentioned earlier, the fundamental driving force behind the derivation of the field equations for the quadrupole ion trap is the fact the motions in the rand z-directions are independent of one another. This concept enabled Paul and others to describe the motion of ions with a simplified mathematical treatise.5 As a better understanding (both theoretical and experimental) of ion motion in the quadrupole ion trap was gained, it was realized that coupled motion between the r- and z-directions was indeed possible. In 1989, March et al. published a theoretical derivation of this coupled motion (discussed later in chapter 2) under resonance excitation conditions.Y This dissertation presents a novel way to experimentally confirm the presence of coupled motion in the quadrupole ion trap, employing the technique of photodissociation. Ion Activation


The most common result of ion activation of any polyatomic species in the gas phase is unimolecular dissociation. Unimolecular dissociation in trapping instruments typically occurs from a stable ion which has been "activated" and made unstable. The fragments observed from this dissociation depend on the structure of the parent ion and can thus provide structural information on the







18

parent species itself. The ion activation techniques which drive the unimolecular dissociation process are usually evaluated by a series of four criteria: (1) energy deposition in the ion of interest; (2) energy distribution or bandwidth of the deposition process; (3) variability of the deposited energy; and (4) reaction crosssection for the ion activation technique.'

Several different types of reactions in tandem mass spectrometry (MS/MS) can be used to bring about the ion activation process. Collisional activation is the most widely used method for ion activation, operating at translational energies ranging from approximately 10 eV to about the 10 KeV range. Fragmentation of polyatomic species via collisional activation is called collision-induced dissociation (CID).*" Another increasingly popular method of ion activation is surfaceinduced dissociation (SID). Energy transfer in the SID process can be quite large (on the order of about 8 eV), where the amount of energy transferred is controlled by the translational energy of the incident ions.6'63 A third method for the activation of polyatomic ions involves the interaction of the parent ion with a beam of electrons, a process called electron excitation. Product ion spectra generated by electron excitation are characterized by broad energy distributions with an upper limit approaching the energy of the electrons used for ion activation." The final method typically employed for ion activation involves the absorption of photons from a light source (e.g., laser irradiation), with the resulting fragmentation process referred to as photodissociation. 69 Photodissociation holds several advantages over the other ion activation techniques, including the







19

ability to impart a wide range of well defined energies to the parent ion of interest. The ability to finely control the energy deposition process with photon absorption, and thus the fragmentation process, is the central focus of this dissertation.

It is beyond the scope of this dissertation to discuss in detail the fundamental aspects of the ion activation techniques other than photodissociation (e.g., collisional activation, surface-induced dissociation, and electron excitation). Instead, a general overview of the unimolecular dissociation process is presented in order to give the reader a better understanding of the results achieved with ion activation (e.g., fragmentation process).

Unimolecular reactions in the gas phase have been studied extensively and can provide a large amount of information concerning rates of dissociation, energy partitioning, and development of models to describe a particular gasphase reaction system. Mass spectrometry provides a unique environment for evaluating gas-phase systems since reactions can take place in collision-free conditions. In addition, both parent and product ions can be isolated and identified with high specificity. Unimolecular dissociations have been used successfully to explain electron ionization of simple molecules, and the theory is now currently being adapted to help predict the products from simple collisional activation experiments.70 This theory, termed the quasiequilibrium theory (QET), has associated with it four basic assumptions important not only for the statistical theory of electron ionization, but for general considerations of unimolecular reactions in the gas phase. The assumptions of the basic theory are as follows:







20

(1) the dissociation event is long when compared to the time needed for ionization or excitation; (2) the rate of the dissociation event is much slower than the corresponding rate for the redistribution of the energy deposited during ionization/activation; (3) fragmentation is the result of a series of competing and consecutive reactions; and (4) an internal energy equilibrium is achieved where the energy is randomized over all internal states with equal probability.'

After the initial energy deposition, the weaker bonds are preferentially broken, thus revealing mass fragments indicative of the parent ion. This would mean that a mass spectrum or MS/MS spectrum depends only on the amount of energy deposited into the ion and not on how that energy was deposited. For a typical MS/MS experiment, energy can be deposited during the ionization event and/or in a subsequent ion activation reaction.'

The product ions produced in an MS/MS experiment are determined by several factors in addition to the amount of internal energy deposited into the system. These include both the time frame for the reaction to occur and the individual microscopic rate constants for each dissociation pathway. Since there exists an inverse relationship between ion lifetimes and the rate of dissociation of an ion, examining the plot of mass spectra as a function of ion lifetime can yield information similar to that obtained with traditional breakdown curves. Early studies by Morgan et al. have reported that rate constants in excess of 2 x1 0-7 s were required for daughter ions to be dissociated in the collision cell of a reverse geometry (magnetic sector followed by electrostatic sector, BE) double-focusing







21

instrument." By applying the basic ideas of transition state theory to individual microscopic rate constants, the expression for the intemal-energy-dependent rate constant becomes:


k(E)1 W*(E-EO) 1 W*(E-Eo) (1-12)
h dW(E)/dE h p(E)
where W(E) is the number of energy states of the ion with energy less than or equal to E, p is the energy level density (dW/dE), W* is the same as W except that the transition state is assumed for the ion, E. is the ion activation energy and therefore (E-E0) can be defined as the internal energy of the ion. A schematic representation of the relevant energetics2 is shown in figure 1-3. General Operation


The quadrupole ion trap instrumentation used and developed for this dissertation was based on an early design by Kelley et al. of Finnigan MAT (San Jose, CA).27'28 A schematic of the Finnigan MAT ITMST" is shown in figure 1-4. In the normal mode of operation, electrons are gated into the central volume of the analyzer (through a 1/16" hole) by pulsing a gate electrode to + 180 V for a specified ionization time period. The ions formed by electron ionization are trapped by an rf voltage applied to the ring electrode of the analyzer (while the endcap electrodes are held at ground potential). The detection event is accomplished by ramping the amplitude of the rf voltage applied to the ring

















Figure 1-3:


Schematic representation of the energy terms relating to the ionization (or dissociation) of a polyatomic molecule: I., Ionization energy of the molecule P; Eth, thermal excitation of a molecule P prior to ionization; Etd, energy transferred to P by the incident electron, photon, or collision; E, resulting internal energy in the ion; q,, reaction coordinate for P+ -- A+ + B; E.,, the activation energy for P+ A+ + B; e',t, activation energy for the reverse reaction A+ + B P*+; AH *, AH08*, standard enthalpy changes for P -. A+ + B + e at 0 K and 298 K, respectively; Do=AH0'*, standard enthalpy change for the dissociation P -- A + B; 1-(A), ionization energy of the fragment A. Only two of the many vibrational degrees of freedom of P, P+, P', and A' are represented; the potential surfaces of A and B are omitted. The spacing of vibrational levels is greatly exaggerated relative to D. and e,,: 1, is typically 2-4 times Do. The level (a) represents the energy of the dissociated neutral system, A + B, with the species in their ground states. Most of these energy terms refer to the differences in energy between specific levels; the exceptions are AH2S* and AH*, which are the usual thermodynamic quantities. Adapted from reference 72.















~P


eact E


E -Ect r
act act


q, 0


L.


H


Etfd


A H (A)
.98 0i-


I I I-.


4


I


I. I


Eth


pI


H298- Ho


t


A+B


0)


AH~


DoI


z


i






















Figure 1-4:


Schematic diagram of the Finnigan MAT Ion Trap Mass Spectrometerm items) The drive frequency applied to the ring electrode is 1.1 MHz. The supplementary rf generator can be used to apply both dipolar (as illustrated in the figure) and quadrupolar excitation signals to the endcap electrodes.












Filament

>11 End Cap


Ring Electrode


V


Electron Multiplier


A-


Detector To
(I


End Cap


Preamplifier on Signal)


zzt


F


Amplifier and RF Generator, Fundamental RF Voltage


Scan Acquisition
Processor (Computer)


Amplifier and RF Generator, Supplementary RF Voltage


01







26

electrode, thus causing ions of successively higher m/z ratios to be ejected from the ion trap to the detector (mass selective instability).24.

Increased resolution and sensitivity can be obtained when axial modulation is used in conjunction with the mass selective instability scan. Axial modulation is accomplished by applying an auxiliary ac voltage to the endcap electrodes 1800 out of phase.' This dipolar resonance excitation frequency is set to approximately 525 kHz just below the 6,= 1 boundary (at 550 kHz) of the stability diagram. The improved peak shape and resolution obtained is due to presence of a uniform electric field in the z-direction, which reduces ion shielding and space charge effects normally seen with the mass selective instability scan where the field strength at the center of the trap is zero.23,3 This technique has also been used successfully to extend the mass range of the quadrupole ion trap."3 A discussion of mass range extension will follow in chapter four of this dissertation. Other applications for the technique of resonance excitation include notchfiltering27, high resolution'1674, ion isolation74, and collision-induced dissociation" studies.

Ion isolation can be accomplished using a variety of methods. A combination of rf and dc voltages applied to the ring electrode can be used for either apex or two-step isolation.7"'79 These two methods utilize the edges of the stability diagram to preferentially isolate the ion of interest. Resonance excitation using forward and reverse scans is also used for ion isolation.4 By applying a high frequency dipolar excitation signal to the endcap electrodes and increasing







27

the rf amplitude applied to the ring electrode, all m/z values below a specific ion of interest are ejected from the trap. Next, by lowering the dipolar frequency and decreasing the rf amplitude, masses of higher m/z values are ejected, thus storing the desired m/z ion or range of ions.

Broadband techniques have also been employed for ion isolation studies. The various methods available include stored waveform inverse Fourier transform (SWIFT)', the use of multiple single discrete frequencies"',", and random noise.2 Perhaps the most effective means for ejecting a large range of ions is that of SWIFT or random noise. These techniques have proven to be very successful, since a large number of signals (of various frequencies) can be applied simultaneously to the endcap electrodes. In essence, sensitivity is increased because the ion trap is selectively filled with the ion(s) of interest and not with unwanted matrix or background ions.

The ability to implement user-defined scanning strategies through the Ion Catcher Mass Spectrometer (ICMS) software developed in this laboratory was critical for the success of the experiments in this dissertation.3 Specifically, the development of ICMS software allowed for the computer control (TTL signal) of external devices such as a pulsed C02 laser, a continuous wave (cw) C02 laser, and a pulsed-valve controller.' The FORTH programming option permitted the design of intricate experiments involving resonant excitation frequency, laser control, pulsed-valve control, and ion cool time (vibrational relaxation). In







28

addition, scan table times could be adjusted up to 1 s, to facilitate various photodissociation experiments.


Electrospray Ionization


Perhaps the most successful liquid chromatography/mass spectrometry (LC/MS) interface to date is that of the electrospray (ESI) ion source. Over the last seven to eight years, the volumes of research in the field have produced an abundance of methodologies and applications for the identification and quantitation of biological species. Some of the reasons for the rapid growth rate of ESI include the speed with which commercial instrumentation was developed, the ability to couple ESI with microscale capillary separation techniques, and the ability to perform tandem mass spectrometry on multiply charged species."

The first use of electrospray as an ionization technique for biological species was reported over 25 years ago by Dole and coworkers who defined many of the operational parameters used today."'" Dole's original detection scheme involved ion mobility and ion retardation methods, since an appropriate mass spectrometer was not available. Dole was also the first to report the multiple charging effect seen with large biological species." The first reports of ESI combined with mass spectrometry were by Fenn and Aleksandrov et a]. in 1984.m,8 Aleksandrov's group was also the first to interface an LC to an electrospray ionization source connected to a magnetic sector mass spectrometer.9*







29

The most important aspect of ESI is that of multiple charging. In a landmark paper published in 1988, Fenn and coworkers reported as many as 45 charges attached to proteins of molecular weight 40,000 da.' This work was quickly duplicated and extended by both Smith and Covey. The multiple charging effect permitted the analysis of high molecular weight biological species using existing quadrupole instrumentation, due to the lower mass-to-charge ratios that can be obtained with multiple charging. Another advantage of the multiple charging process is that a more accurate molecular weight determination can be obtained from a distribution of multiply charged peaks.47 In addition, tandem mass spectrometry has been shown to be useful for the dissociation of multiply charged species. The most frequently used instrument for structural elucidation studies of electrosprayed ions is the triple quadrupole mass spectrometer. One of the most notable applications of tandem mass spectrometry and ESI ionization has been the work of Don Hunt and colleagues." The Hunt group successfully used only femtomoles of material to identify histocompatibility complex-bound peptides. This work ultimately led to the identification of a peptide with high binding affinity for cytotoxic killer T cells.94

The use of tandem mass spectrometry (triple quadrupole) in conjunction with ESI is not the only active area of research in the ESI field. Other areas of great interest include mechanistic investigations"'*, surface-induced dissociation'1'02, ion-molecule reactions', non-covalent interactions'*",







30

collisional activation""', and photodissociation."1112 Several recent reviews on ESI coupled with a variety of applications can be found in the literature.4.1s115

The purpose of this section is to provide the reader with a general understanding of the electrospray process and examine some of the more recent advances in the field. A brief discussion of charged droplet formation is followed by a section on the chemistry of multiply charged ions to give the reader knowledge of the basic physical principles of ESI.


Basic Principles of Ion Formation


The production of ions in electrospray mass spectrometry is comprised of two steps: the production of highly charged droplets with their dispersal at atmospheric pressure and the evaporation of these droplets to produce multiply charged ions.115 The production of highly charged droplets begins with a small flow of liquid through a simple metal capillary (stainless steel needle) which operates at an elevated electric potential relative to a counter electrode. The potential of this electric field on the capillary is typically between 3 and 6 kV relative to the counter electrode placed about 0.3 to 2 cm away. The counter electrode has an orifice where charged clusters, ions, or droplets are passed into the mass spectrometer. Charge accumulations occur at the liquid surface due to the application of the electric field. Therefore, flow rate, solution resistivity, and surface tension are important variables in droplet production. The bias of the capillary needle relative to the counter electrode can be selected to produce







31

either positively or negatively charged droplets. The electric potential essentially disrupts the flow of liquid from the capillary tip resulting in production of the charged droplets. To aid in the desolvation process, the droplets are typically entrained in a nebulization gas such as nitrogen, oxygen, or even SF.. Oxygen and SF act as electron scavengers for negative ion electrospray or when spraying pure water solutions."

Solution resistivities on the order of < 10-5 0 are required for stable spray conditions at room temperature. This corresponds to a solution conductivity of aqueous electrolytes of approximately 1 O-4normal (N), where normality is defined as one gram molecular weight of the dissolved substance divided by the hydrogen equivalent of the substance per liter of solution. The higher the surface tension (the higher the aqueous fraction in the solution), the higher the threshold voltage needed for the onset of the electrospray process. The relationships between the electrospray onset voltage V., the surface tension T,, the needle or capillary radius r, and the needle to counter electrode distance h can be shown to approximate:


Voo-(Tsr)12In 4h (1-13)
r

Because the droplet size decreases with increasing solution conductivity, lower flow rates are required for solutions which are highly conductive. The dependence of solution conductivity (a) on ESI ion current (1) is relatively weak (Icc OP2) 8







32

The presence of a dry nebulization gas at approximately 800 C and the use of a heated capillary (counter electrode) can aid significantly in the desolvation process of the charged droplet. The formation of multiply charged ions from the charged droplet is an area of great debate. The droplets eventually reach a point where the repulsive coulombic forces approach those of the cohesive forces (surface tension) that hold the droplet together. At this point the droplet may form an ion by one of two proposed mechanisms: droplet fission at the Raleigh limit or direct field evaporation of the droplet."'"' A discussion of the relevant thermodynamics of these two processes is beyond the scope of this dissertation. However, it is the belief of this author that the direct evaporation theory as set forth by Iribarne and Thomson applies to most situations."'19

Once the charged droplets/ions pass through the counter electrode region, they proceed through a differentially pumped region containing one or two skimmer cones. A schematic diagram of the typical electrospray source developed by Fenn and coworkers can be seen in figure 1-5, showing the electrospray needle, counter electrode, differentially pumped region, skimmer cones, and dc lens injection system." Details of the various ion injection systems used are discussed in chapters 3 and 4, along with details of a new ion injection system (for quadrupole ion traps) for electrosprayed ions using an rf-only octopole beam guide.






















Figure 1-5:


Typical ESI ion source used in mass spectrometry. The glass capillary can be replaced by a heated stainless steel capillary. The skimmer and ion source lens assembly design will vary depending upon the instrumental configuration.











Cylindrical Electrode



4F


Glass Capillary
I


Tube Lens

Liquid Sample






I Rotary
Pumps


Electrospray
Needle Assembly


Drying Gas


IF


Skimme


N%1*


Ion Source r Lenses
Baffle















Turbo Pump I


Analyzer


Turbo Pump







35


Molecular Weight Determination


Molecular weight data from ESI spectra can easily be obtained due to the charge state distribution associated with the ionization process. Typically, the width of the charge state distribution is approximately half that of the highest charge state, although the effects of the various factors (e.g., pH, applied potential) are not yet well understood.12""2 The adjacent peaks (of the multiply charged ion distribution) in the spectrum of positively charged biopolymers usually vary by one charge. Therefore, in order to determine molecular weight

(M,) from an ESI spectrum (where the charge varies on adjacent peaks by the addition or subtraction of one proton), the following expressions are used:


PIZi = Mr + Mazi = Mr + 1.0079z, (1-14)

where p, is the m/z of interest and z, is the charge on p,, and Ma is assumed to be the charge carrying species (proton). By examining another peak in the charge distribution spectrum, another equation can be generated for a m/z higher than the previous example (p2>P1) that is j peaks away from p,:


P2(Z -j)= Mr+1.0079(z, -j) (1-15)

Equations (1-14) and (1-15) can then be solved for p, yielding:


j(p2-1.0079) (1-16)
(P2-P1)
The value of the molecular weight is then calculated by evaluating z, to the nearest whole number.12 Improved precision can be obtained by performing







36

the same calculation for the entire series of multiply charged ions. The accuracies which have been reported to date for proteins and other biopolymers over 100 kDa (molecular weight) are better than 0.005%." One of the best mass accuracy measurements to date was recorded for myoglobin, with an observed error of less than 1 ppm obtain using a FTICR mass spectrometer.124 A new method of molecular weight determination developed by Hagen and Monning uses a multiplicative correlation algorithm for processing charge distribution data.12s The ability of the technique to accurately determine molecular weight increases as the (M+H)* signal is spread out over larger and larger charge state distributions.


Photodissociation


Photo-induced dissociation is the next most frequently used method for activation of polyatomic ions after collisional activation. The range of internal energies present after the photon absorption process is much narrower than that obtained with collisional energy transfer. Therefore, the usefulness of PID for the study of ion structures is greatly enhanced. However, the reduced absorption cross-sections observed with photodissociation (10-2 A2) compared to those of collision-induced dissociation (10 to 200 A2) can limit this technique for analytical applications. The recent availability of higher powered light sources over a wider range of wavelengths should provide greater flexibility for photodissociation as a routine analytical technique.'-







37

The process of photodissociation for a positive ion can be described by the following equation:


nhv
A *2 A-** P + N (1-17)
relaxation dissociation

where A+ is the ion of interest, n is the number of photons absorbed, hv is the photon energy, A+* is the excited state, and P+ represents the product ion (with loss of neutral N). For photodissociation to occur several prerequisites must be met. The most important criteria include the absorption of photons with energy hv, the existence of excited states above the dissociation threshold, a slow relaxation rate compared to light absorption (multiphoton processes), and dissociation rates which are fast on the time scale of the type of mass spectrometer employed.1'6

The information obtained from a photodissociation experiment can address a variety of gas-phase chemistry issues. One of the most important issues is the difference observed in fragmentation spectra between PID and CID. The narrow well defined energy transfer distribution step in PID typically leads to the dissociation process via the fragmentation pathway with the lowest activation energy (especially for visible and infrared wavelengths).""

In addition, wavelength-dependent spectra can be obtained as long as the internal energy of the ion population is above the dissociation threshold for the wavelength of interest. The photodissociation spectrum can then be compared with the typical absorption spectrum of the neutral molecules, provided that light







38

absorption can occur for the given structure of the ion. The information obtained from these experiments can be used as a fingerprint in the determination of ion structures.126,27

In trapping instruments, photo-induced ion decay can be measured as a function of irradiance time. These data can be used effectively to distinguish isomeric ion structures in the gas-phase. Kinetic energy release data can also be used (below 100 mV) to add important ion fragmentation information.2' Infrared Multiple Photon Dissociation (IRMPD)


Infrared multiple photon dissociation is particularly well suited for use with trapping instruments. The ability to trap ions for extended periods of time at low pressures in both the ion cyclotron resonance (ICR) cell and the quadrupole ion trap mass spectrometer allows for the sequential absorption of infrared photons via low intensity infrared radiation. The IRMPD process was first characterized by Beauchamp and co-workers for positive ions using an ICR mass spectrometer.-1"' These and other studies focused primarily on the chemical kinetics, reactivity, and spectroscopy of various ionic species. Brauman and coworkers were the first to study the vibrational relaxation of gas-phase ions and the accompanying physics of pulsed megawatt infrared multiphoton dissociation using a C02 laser."' For the case of trapped negative ions, irradiation can produce both a photodissociation and an electron photodetachment spectrum. 2
134







39

As with the ICR technique, the quadrupole ion trap is capable of storing ions for long periods of time. The storage capability makes the quadrupole ion trap mass spectrometer very compatible with a wide range of experiments using light. One of the first successful uses of the ion trap in conjunction with photodissociation involved the study of the proton-bound dimer of 2-propanol utilizing a cw infrared laser.20 2-Propanol was chosen for study since its gasphase ion chemistry is well known and the formation of the protonated dimer is easily accomplished. The instrumental configuration consisted of an ion trap connected directly to the ion source of a quadrupole mass filter (QUISTOR mode of operation). A detailed description of the optimization of ion trapping characteristics for studies of ion photodissociation is a QUISTOR can be found in March and Hughes.'

These earlier studies of the IRMPD process in the ion trap were performed with a single-pass ring electrode design, with a 3 mm diameter hole on the center axis of the ring electrode as the entrance aperture for the low power cw C02 laser beam. Upon reaching the other side of the ring electrode, a portion of the beam passed through a 0.8 mm diameter hole and through a NaCl window, where the laser power was monitored externally. The remainder of the laser beam was reflected by the ring electrode throughout the QUISTOR. The pressure of 2propanol was adjusted to 5 mPa so that photodissociation of the proton-bound dimer, (2M+H)', at m/z 121 could occur at an appreciable rate. At pressures optimum for the formation of the proton-bound dimer (13 mPa), no laser-induced







40

dissociation was observed due to collisional deactivation of the vibrationally excited proton-bound dimer.20.2 At the low pressures used in these experiments, the dissociation rate constant kD was related to the phenomenologically defined cross section a- and the photon flux e by the following equation: kD= D4) (1-18)

The highest absorption cross section for 2-propanol was found to be at a wavenumber of 944 cm-', with the corresponding absorption of 10 photons. The dissociation rate constant kD was determined to be 2.2 s-', assuming first order dependence on photon flux.

Photodissociation experiments for the proton-bound dimer of 2-propanol (m/z 121) showed three different photoreaction channels open for the IRMPD process. March and Hughes give a detailed description for verification of the various reaction pathways, the photodissociation of the various isotopic analogues of 2-propanol, and the ion relaxation processes involved.22"

The same experimental apparatus has also been used to investigate the gas-phase ion chemistry of ethanethiol, 1- and 2-propanethiol, and 1hydroxyethanethiol.'-"' The collisionally-cooled, proton-bound dimers of ethanethiol, 1 -propanethiol, and 2-propanethiol were unaffected by laser irradiation at 944 cm1. However, the proton-bound dimer of 2-hydroxyethanethiol was thought to contain a S-Hf-S linkage which was shown to be transparent at the same wavelength. Isomer differentiation by multiphoton dissociation of the proton-bound dimer of propanone (m/z 117) and protonated diacetone alcohol







41

(m/z 117) has also been demonstrated."" The authors were able to differentiate the two isomeric m/z 177 ions by an aldol condensation reaction observed with the fragmentation of protonated diacetone alcohol. This reaction was not observed with the propanone dimer in the gas phase.

Investigation into the wavelength dependence of IRMPD photodissociation efficiency of the proton-bound dimer of ethanol using the QUISTOR demonstrated the observable frequency shifts of the C-0 stretch in the IR region.'," Application of the quistor technique to the study of photodissociation rates by varying relaxation time, buffer gas pressure, and analyte pressure has yielded data for the proton-bound dimers of isopropanol, 2-d,-2-propanol, and ethanol. The authors also studied the effect of collision rate on the defined photodissociation cross section. The results obtained for the fully relaxed protonbound dimer population showed access to the lowest Ea pathway, thus demonstrating the only variable observed was that of the collisional deactivation process (corresponding to higher collision rates 5 ms4).23


The Photon Absorption Process



The interaction of gas-phase molecules with infrared light is currently one of the most rapidly developing fields in chemical physics. Since the first discovery of infrared photon absorption by ions in the gas phase, the field has attracted the attention of researchers from a variety of disciplines."141 This section is intended to give the reader a general overview of the mechanism of the







42

IRMPD process and to outline the applications of this process to large biological ions in the gas phase.

The majority of investigations into the mechanism of the photon absorption process have utilized pulsed or continuous wave (cw) lasers operating in the wavelength range 9.5 to 10.5 pm. The fundamental model of photon absorption and subsequent dissociation is based on the extensive studies of the SF, system at 10.6 pm.'" The basic concepts of this model also apply to the IRMPD process of other gas-phase systems.

In the model proposed by Black et al., there are three successive regions for photon absorption as the energy deposited in the ion increases.'4" A schematic energy diagram in figure 1-6 shows the three individual regions (1,11, and 111) corresponding to coherent multiphoton interaction, incoherent single photon interaction, and dissociation threshold/channels. In the coherent multiphoton interaction region (1), vibrational states are essentially discrete and the corresponding absorption of a single IR photon is a straightforward process. However, successive transitions to higher vibrational levels are not possible since the vibrational spacing is now out of resonance with the monochromatic IR photons. Therefore, in order to further excite the ion of interest other mechanisms which compensate for these anharmonicities must be in operation."

For any polyatomic ion the density of the vibrational levels increases with internal energy. The rate of this increase depends upon the magnitude of the frequencies of the individual modes and the molecular complexity of the ion.



























Figure 1-6:


Schematic energy level diagram as originally applied to the IRMPD of SF.. In region 1, coherent multiphoton interaction describes one of the mechanisms (intensity-dependent powerbroadening effects) whereby non-resonant absorption can take place in a sparse region of vibrational levels. In region 11, the quasicontinuum, resonant absorption steps are always possible, and can be treated as stepwise (incoherent) excitation processes. Region III lies above the dissociation threshold. Adapted from reference 145.










Dissociation channels


Dissociation threshold

Incoherent
single-photon
interaction


Quasi -continuum


Coherent multiphoton interaction


ladder
9
8-
7
6

45 3 _4 2 3 I 2


44


Ill


A
A


6


I







45

Eventually, these vibrational levels merge to form what is termed the quasicontinuum or the incoherent single-photon interaction region (11). Ions in region 11 are characterized by their ability to undergo the resonant absorption process for a given monochromatic laser frequency (although there may be some structure within the quasicontinuum itself). If an ion can be excited through the discrete vibrational levels in region I, then there always exists a path for the absorption process to occur through the quasicontinuum. Polyatomic ions can be excited into the quasicontinuum by a variety of mechanisms including thermal excitation, collisional or particle excitation, exothermic chemical reactions, or electronic excitation followed by internal conversion or forced multiphoton excitation. For the biological ions used in this dissertation, no excitation methods are needed to push ions into the quasicontinuum, since these species are sufficiently large that they already exist in the quasicontinuum as a result of internal thermal energy content."

As the ion continues to absorb energy, it eventually reaches a point where it obtains enough energy to dissociate (region 111). If randomization of this excess internal energy above the dissociation threshold occurs at a much faster rate than the dissociation process itself, then the traditional statistical theories associated with unimolecular dissociation (the QET theory discussed previously), and that of Rice-Ramsperger-Kassel-Marcus (RRKM theory, i.e. the application of transitionstate theory to unimolecular reactions) can be applied to the photodissociation process."5 Applications of RRKM theory include relating the excess energy







46

absorbed above the dissociation threshold to the lifetime of the excited species, and estimation of the energy distributions within the product ions from the IRMPD process.













CHAPTER 2
FUNDAMENTAL INVESTIGATIONS OF IRMPD IN THE QUADRUPOLE ION TRAP


The second chapter of this dissertation focuses on the fundamental aspects of the photodissociation (IRMPD) process in the quadrupole ion trap mass spectrometer. A large part of the discussion centers around the design of a novel multipass optical arrangement for use with IRMPD. This design circumvents previous problems of limited IR laser power, small IR absorption cross sections for many polyatomic organic species, and the limited ion statistics of trapping and detection of ions for IRMPD in the quadrupole ion trap. Fundamental investigations (using model compounds formed by El/Cl) into consecutive reactions, ion motion effects, the photon absorption process, kinetics, buffer gas effects, and infrared spectroscopy employing a cw CO2 laser are presented.6'11 This work is the basis for future applied studies addressed in chapters 4 and 5 of this dissertation.


Instrumentation


The instrumentation addressed in this section covers the experimental design (Finnigan MAT ITMS parameters), ion trap operation without He buffer gas, and a detailed description of the multipass optical arrangement necessary for all


47







48

subsequent experiments performed in this and all of the following dissertation chapters. The basis for the experimental design is derived from previous studies by Watson et al. in the ICR mass spectrometer."'4 A detailed description of the instrumentation employed has been published.47"48 Experimental Design


All experiments were performed on a Finnigan MAT (San Jose, CA) ion trap mass spectrometer items) Except for the pulsed-valve experiments, the ion trap was operated with no He buffer gas in the vacuum chamber. The base pressure of the instrument was 3.5x1 0- torr (uncorrected), as indicated on a Bayard-Alpert ion gauge (Granville-Phillips, Boulder, CO) mounted on the vacuum manifold. The Teflon ring electrode spacers were found to absorb strongly at 944 cm' during the laser irradiation period, which desorbed both neutral and ionic species that affected both ion storage efficiency and detection; they were therefore removed. The vacuum manifold was equipped with a modified flange containing a ZnSe window to pass IR radiation. The modified software used (ICMS) provided two TTL pulses for computer control of both the cw C02 laser and the pulsedvalve apparatus.' The experimental arrangement can be seen in figure 2-1.

The cw C02 laser employed was an Apollo Model 570 that is line tunable over a wavelength range of 1099-924 cm" and has a beam diameter of approximately 1 cm. The maximum laser power obtainable was 50 W at 944 cm'. The laser power supply was modified with electronics that transformed an























Figure 2-1:


Instrumental configuration for IRMPD in the ion trap.






4$ To Computer He line
Pulsed Valv


Pulsed Valve Controller o

I Pu1sd -..


Sample


Inlet Valves


CO2 Laser


Beam


I


Turning


To Laser


Power


Mirror


Supply


From cw Laser c


C)


valve Ion Trap


/


mmml


ZnSe Window Flange


U I


ITMS
Electronics


TTL 1 TTL 2







51

incoming TTL signal to the CMOS logic used by the laser. A spectrum analyzer (Optical Engineering, Model 16-A, Santa Rosa, CA) was placed directly in-line with the cw C02 laser beam for wavelength measurements. All laser energy measurements (Coherent Radiation Model 410 power meter, Boulder, CO) were taken inside the ITMS vacuum manifold to correct for beam loss at the turning mirror surface and ZnSe window. Because of space considerations, the ion trap analyzer was removed from the ITMS vacuum manifold for the energy measurements. The power meter was then positioned at the exact location where the laser beam would enter the ion trap analyzer. The laser was placed parallel to the ITMS manifold with the beam reflected at a 90" angle by a gold coated mirror as seen in figure 2-1. Laser beam alignment was accomplished with a HeNe laser placed in-line with the cw C02 laser. The analyzer of the ITMS was fitted with a modified mounting bracket to allow for rotation of the ion trap analyzer from its original fixed position to facilitate alignment (figure 2-2).

The pulsed-valve used in the experiments (Series 9, General Valve Corporation, Fairfield, NJ) was mounted on the opposite flange from the ZnSe window (see figure 2-1) and was used to pulse He into the ion trap to increase trapping efficiency. The pulsed valve was placed 0.5 cm from the outer diameter of the ring electrode. Its horizontal position was between the modified ring electrode and the entrance endcap. The pulsed-valve controller (built at the University of Florida) was controlled by an external TTL pulse generated by the ITMS electronics. The timing diagram shown in figure 2-3 displays both the laser























Figure 2-2:


Adjustable mounting bracket for the ion trap analyzer assembly. The three slotted holes allow for rotation of the analyzer to facilitate laser alignment with the entrance aperture on the ring electrode.












0,745"


Center mark 2,750' diameter


Center mark for
2.532" slots




02,523"

9"'02.750"
1.250"


/


-I-


---0


0.152 2.000"


I


/


135*


cn wi


@-14w"






















Figure 2-3:


ITMS scan function and timing diagram for a typical IRMPD experiment (figure not to scale). 1, pre-ionization/pulsed-valve on; 2, ionization-chemical self-ionization reaction; 3, two-step mass isolation; 4, vibrational relaxation; 5, laser on; 6, laser decay; 7, acquisition.












































I


55


00


...........




c


...........










in






...........


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







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


...........


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







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

















. . . . . . . .


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


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


- Cd







56

control TTL pulse and the pulsed-valve TTL control pulse for a given scan function.

Bis(2-methoxyethyl)ether (diglyme; ACS reagent grade purchased from Fisher Scientific, Fairlawn, NJ), 3-bromo-1-propene (allyl bromide; ACS reagent grade purchased from Aldrich, Milwaukee, WI), 12-crown-4 ether, and 15-crown-5 ether (ACS reagent grade purchased from Sigma, St. Louis, MO) were introduced via a Granville-Phillips (Boulder, CO) fine metering valve system directly into the ion trap manifold. Depending on the experiment, the sample pressure ranged from 1.1 to 3.6x10- torr. Formation of the protonated diglyme, 12-crown-4 ether, or 15-crown-5 ether was accomplished by self chemical ionization, predominantly due to the reaction of the low mass even-electron fragment ions from electron ionization (El) with the neutral molecules for approximately 400 ms. The resulting [M+H]* ion of interest was then mass isolated by a two-step rf/dc isolation routine.15 Next a 400 ms delay was included to allow for removal of any excess internal energy by radiative and/or collisional cooling. The [M+H]* ion was then irradiated for a specified time period with the cw CO2 laser. After the laser irradiation period, a 10 ms time period was incorporated for decay of the laser output when the high voltage was turned off. For all photodissociation efficiency measurements, the first sequential product ion from the IRMPD of the [M+H]* was resonantly ejected (q = 0.3, frequency = 118.1 kHz, amplitude = 100 mV, values approximate) during the laser irradiation period (periods 5 and 6 in figure 2-3). This prevented the possible occurrence of any ion molecule reactions as the







57

result of the reaction of the sequential absorption product(s) with neutral sample molecules. The remaining ions were then mass analyzed via resonance ejection with qzea = 0.89 and an amplitude of 1.5 V (zero to peak)."'e Ion Trap Operation without Helium Buffer Gas


The presence of a light buffer gas (He or H2) at a relatively high pressure (1 mtorr) has been shown to enhance resolution, sensitivity, and improve detection limits associated with operation of the quadrupole ion trap.26 Unfortunately, the presence of a buffer gas at 1 mtorr can significantly decrease photodissociation efficiencies in the infrared region. Since the time between absorption of consecutive photons for the IRMPD process is on the millisecond time scale, ions which have acquired some fixed amount of internal energy from the photon absorption process (but not enough to reach the dissociation threshold) can undergo collisional damping, thus reducing the photodissociation efficiency of IRMPD.151"52 Therefore, in order to properly evaluate the performance of the newly designed multipass-ring electrode, the ITMS was operated without the presence of He buffer gas so as not to interfere with the photodissociation process. This section presents a practical dialogue on ion trap operation without buffer gas and discusses the relevant instrument parameters (without He buffer gas) which affect trapping efficiency, resolution, and mass range.

One of the most important aspects of no He ion trap operation is trapping efficiency. The two instrument parameters which affect trapping efficiency the







58

most are the ionization and cool times. Since there is no appreciable concentration of buffer gas (only sample gas pressure in the 107 torr range), longer ionization times are needed to generate large numbers of "reagent" ions for the self-chemical ionization process or for the production of molecular ions [M+-]. Typical ionization times for experiments with no He buffer gas are on the order of 60 to 100 ms. Under normal He buffer gas conditions, 60-100 ms ionization times would severely space charge the ion trap for sample pressures in the 1 0-7torr range. However, since trapping efficiency is severely reduced with no buffer gas, a larger number of ions must be made (necessitating a longer ionization time) to trap enough ions so as to limit the large statistical variations associated with low ion densities.

Perhaps even more important than generating large numbers of ions is the cool time associated with the experiment. Wu and Brodbelt showed that for low pressure operation of the ion trap (pressures < 9x1 0-5 torr), increased cool times led to very efficient storage conditions similar to those with 1 mtorr of He buffer gas." Over any given time period, it was shown that about 50-100 collisions were needed to cool ions to the center of the ion trap. The authors reported that for shorter storage times and lower pressures, the majority of the ions in the expanded ion cloud are likely to be accelerated into an endcap rather than through the endcap exit holes during the mass-selective instability scan. Also, it has been postulated that ions on the outer fringes of the ion cloud could receive







59

too much kinetic energy from the rf field for coherent ejection during the analytical scan.I3'154

For many of the experiments discussed in this chapter, the cool time and the reaction time for the formation of the [M+ H]+ ion for diglyme and the crown ethers was combined (since during any reaction time, collisional cooling can also occur). In many instances, the cool/reaction times were on the order of 400 ms. At these extended cool times, ion stability (ion-molecule reactions) can become problematic. For the case of an even-electron species, such as protonated diglyme, [M+H]+, at m/z 135, storage times of over one minute showed no appreciable dissociation/reaction of the parent ion species. However, for an oddelectron species like ionized allyl bromide ([M+] m/z 120 and m/z 122), long storage/cool times can produce unwanted ion-molecule reaction products. As seen in figure 2-4, an appreciable rate of reaction (2.5 s-) of m/z 120 (the M* of allyl bromide) with the neutral allyl bromide present in the trap produced a decrease in the parent ion signal intensity over a 100 ms time period. The reaction of the parent species with the neutral allyl bromide produced the allyl carbocation at m/z 41. The mechanism of this reaction was thought to occur through a collision between the [M+-] ions at m/z 120 and 122 with the neutral species; this energetic collision (which has a substantially higher average kinetic energy than that observed when He buffer gas is present) initiates charge-site migration to the allyl portion of the molecule and loss of the bromine radical as shown in equation 2-1.






















Figure 2-4:


Kinetic data from the selective mass storage of the 79Br isotope from the molecular ion of allyl bromide, m/z 120. A plot of In (l/1,) versus time yields a straight line, with the slope equivalent to the rate coefficient of the ion-molecule reaction of m/z 120 with neutral allyl bromide.








0.20





0.00 0

00 0



-0.20-0.400 20 40 60 80 100
Storage Time in ms


0)







62


[C3H.Brf* + C3H5Br [C3H5]* + Br* + C3H5Br (2-1)

After the desired cool/reaction time for a given ion population, ion isolation was accomplished by either apex or two-step isolation."7" The tuning of the necessary rf/dc voltage combinations to store only ions of a single m/z required fine adjustments of 0.1 to 0.5 V. This meticulous procedure was required because under conditions of no He buffer gas, the stability edges of the Mathieu diagram become very steep. These steep edges were found to be extremely sensitive to ion population; without a fine rf/dc voltage control, a large percentage of the parent ion population of interest could be ejected during the isolation event.

During the mass selective instability scan (no helium buffer gas), the optimum resonance ejection parameters were found at a q..,=0.89 and an amplitude of 1.5 V(a). These values gave peak widths slightly smaller than 0.5 mass units at full width half maximum (FWHM). This observation was consistent with that of the high resolution mode of operation, where reduced peak widths were observed only using resonance ejection techniques.' Operation of the ion trap without helium buffer gas (no resonance ejection scan), produced larger peak widths at FWHM (>0.5 mass units) and reduced signal-to-noise (s/n) ratios for the peaks of interest. However, ion statistics and signal reproducibility were found to be somewhat more reliable without using the resonance ejection scan.

For verification of the various reaction pathways, notch filter ejection experiments were performed during the laser irradiation period. Frequency probes for protonated diglyme (m/z 135) without He buffer gas required







63

significantly less resonant excitation amplitude (55 mV for no He) compared with that of the same experiment with He buffer gas (245 mV) (see figure 2-5). In addition, large shifts in the maximum ejection frequency were observed. These shifts were attributed to a lack of helium buffer gas, which was needed for a well defined frequency distribution. Also observed in figure 2-5, frequency bandwidths were much greater (approximately 0.4 kHz with He) for the case where no He buffer gas was present (2 kHz). All frequency data in the experiment were taken for a constant number of ions stored in the trap (as measured by the electron multiplier and detector circuits) so as to minimize bandwidth and frequency shift phenomena typically observed with different ion populations. Although the absolute number of counts for protonated diglyme (m/z 135) was matched in the profile scan mode for the case of He buffer gas versus no He buffer gas, there could be substantially more ions present which were not detected for the aforementioned reasons in this section.


The Multi-Pass Ring Electrode


The 1.0-cm-diameter IR laser beam (Gaussian profile) was attenuated by a 0.3 cm (1/8") entrance aperture on the ring electrode. The laser was aligned such that the center portion of the Gaussian beam profile passed through the entrance aperture. The ring electrode was modified by incorporation of three polished stainless steel spherical concave mirrors (radius of curvature = 2.0 cm) mounted on the inner surface of the ring, as shown in figure 2-6. The




















Figure 2-5:


Frequency optimization curves for protonated diglyme (with and without He buffer gas) using dipolar excitation. The frequency was incremented every 0.1 kHz. For the case where He buffer gas was present, a well defined optimum frequency was obtained due to the well defined trajectories of the ions confined to the center of the ion trap.








400


m/z 135 parent (He present)


300

m/z 135 parent (He not present)



200
Sum of m/z 103 &
m/z 59 Product Ions
(He present)



100- 11



Sum of m/z 103 & m/z 59 Product Ions (He not present)

0 --r

114.0 115.0 116.0 117.0 118.0


Frequency in kHz


0)





















Figure 2-6:


Modified ring electrode for multipass IRMPD experiments. Mirror positions and the eight laser passes across the radial plane of the ring electrode, along with approximate photon density int eh radial plane of the ring electrode, are shown. The positions of mirrors A and B determine the number of laser transversals across the radial plane of the ring electrode.










Ring Electrode CW Co2
Laser Beam


Spherical





a Spherical Mirirrrr
1 a sr = 2 .0 c m..........................................





2 Passes I3 Passes E4 Passes


0)







68

approximate photon density (assuming constant intensity across the attenuated beam width) observed in the radial plane of the ring electrode can be seen in figure 2-6. The most critical adjustment of the mirror system was the separation of the centers of curvature of the mirrors labeled A and B. This separation distance determines the number of beam transversals across the ring electrode: 4,8,12, or any other multiple of 4. The mirrors were mounted on the ring electrode such that the centers of curvature of Mirrors A and B were on the front surface of mirror C, and the center of curvature of mirror C was halfway between mirrors A and B.'1 This method of mirror alignment establishes a system of conjugate foci on the reflecting surfaces of mirrors A,B, and C. Consequently, light leaving the surface of mirror A is focused by mirror C on the surface of mirror B, and the light leaving mirror B is then focused back to the original point on mirror A. Similarly, any light leaving mirror C and going to either mirror A or B is focused back to mirror C at some point offset from the original one1" (see figure 2-6).

This technique for extending the optical pathlength in restricted volumes has many advantages over previous designs which incorporate flat mirror systems or a spherical mirror and a truncated prism scheme.'" One advantage is the ease in making adjustments since all tolerances but the horizontal angles of mirrors A and B are usually quite large. Other inaccuracies which are introduced are small and not cumulative. Another advantage is that light losses on mirrors surfaces are kept to a minimum. Since there are only two reflections (at normal







69

incidence), spots, dust or pinholes on mirrors A and B have a much less serious effect since the light from any point on the mirror surface always goes back to the same point; therefore, if there is a spot on the mirror surface, the light falling on the spot is lost but on the second reflection from that mirror no more light is lost. Yet another advantage is that there is only one transmission of light through an entrance aperture; no light travels through any glass or other optical material where losses due to reflection can occur.15

Each mirror (and its mounting bracket) was constructed from a single piece of stainless steel. At one end of each piece, the radius of curvature was cut (r = 2.0 cm) and the surface was highly polished. The three mirrors were mounted into precision-drilled holes in the ring electrode, positioned such that the alignment was automatic; no realignment has been needed since the original assembly of the multipass ring electrode (36 months). A small machine screw was used to hold each mirror-mirror mount assembly in place on the ring electrode. Laser alignment was set such that the center portion of the 1-cm Gaussian beam profile was transmitted through the entrance aperture, thus yielding high photodissociation efficiency. With efficiencies already greater than those published for previous designs on a QUISTOR trap or ICR cell, condensing the beam down to 0.3 cm was considered less important than examining the numerous ways in which gas-phase ion chemistry can be studied via IRMPD with this unique design. Furthermore, with the beam focused to 0.3 cm, damage could possibly occur on the surface of the ion trap mirror when the laser is tuned







70

to a strong IR emission line (e.g., 10.59 pm). With continued use over a 36-month period, no degradation of the ion trap mirror surfaces has been observed with the unfocused laser beam.147


Effects of Ion Motion on Photodissociation Efficiency


The effect of resonance excitation on ion motion in the quadrupole ion trap is a rapidly evolving area of research, encompassing both simulation and experimental investigations.'" The original reports of resonant excitation of ions at their secular frequency of motion (where w, = the fundamental axial frequency and w, = the fundamental radial frequency) due to an externally applied ac field on the endcap electrodes was reported by Paul and Fischer."' It was originally thought that radial excitation (w,) would increase ion trajectories only in the r direction and that axial excitation (wi) would increase ion trajectories in the z direction. However, with the nonideal (stretched) quadrupole ion trap, ion motion in the r- and z-direction are coupled, leading to a complex series of interactions between ion motion in both the r- and z-directions.'6" Today, resonance excitation of stored ions in the quadrupole ion trap is accomplished by applying an auxiliary ac signal to the endcap electrodes. The frequency of the ac signal applied corresponds to a particular frequency component of ion motion; the greater the contribution of the component frequency to ion motion, the greater the ability of the ion to absorb power from the applied field. The larger the power







71

absorption from the applied field, the greater the increase of an ion's kinetic energy (trajectory) for a particular frequency component.

The two most common types of resonance excitation are the dipolar and quadrupolar techniques. Dipolar excitation (standard on all existing commercial instrumentation) is performed by applying two auxiliary ac signals 1800 out of phase to the endcaps. Whereas quadrupolar excitation is performed by application of the same ac signal to both endcaps. Experimentally, both dipolar and quadrupolar resonance excitation have been used for CID studies, SID studies, and resonance ejection (axial modulation) in the quadrupole ion trap. More recently, there has been a concerted effort to understand the nuances of ion motion during the excitation period.5" 6

In this dissertation, a unique method using IRMPD for the evaluation of the coupled motion phenomena observed with resonance excitation is reported. The basic theory behind ion motion has been reported in chapter one of this work and elsewhere. The two subsections entitled Dipolar Excitation and Quadrupolar Excitation briefly discuss the theoretical aspects of resonance excitation, as well as present relevant photodissociation studies concerning coupled ion motion. This technique takes advantage of the high photon density in the radial plane of the ring electrode, which can be used to determine the presence of coupled axial excitation (w) while performing radial (w,) excitation experiments.

All coupled motion data presented in this study are plotted as photodissociation efficiency versus excitation voltage (dipolar or quadrupolar







72

excitation). The photodissociation efficiency (PD) for a given experiment is defined as the fraction of the original ion population photodissociated over a given exposure time for a specified laser irradiance:


PDi -0 (2-2)

where I is the signal intensity of the dissociating ion at the end of the exposure period and 1. is the signal intensity after the same period without irradiation. This definition of 10 corrects for any unimolecular or collision-induced dissociation that may occur."


Dipolar Excitation


Dipolar excitation is the most common form of resonance excitation employed today. The techniques of collision-induced dissociation, notch filtering, and axial modulation all use commercially available dipolar excitation circuits and have been well characterized experimentally. Part of the reason for the success of the aforementioned techniques is due to the asymmetric nature of the dipolar field. Since the signals applied to the endcap electrodes are 1800 out of phase, axial motion is favored over radial motion during the excitation period.' This can be understood by examining the relationship between ion motion and the dipolar mode of excitation. As an ion approaches each endcap, the potential on that endcap needs to be of the appropriate polarity to obtain maximum power absorption. With the application of a 1800 out-of-phase signal, the polarity on







73

each endcap alternates at the same frequency as the ion's secular frequency in the axial direction.16

The trapping conditions observed with the application of a dipolar excitation signal can be significantly different than those seen by direct application of a quadrupolar trapping field to the ring electrode. Therefore, the resulting ion trajectories no longer follow traditional Mathieu parameters, since ion motion is now controlled by both the quadrupolar trapping field and the dipolar excitation field. Previously reported simulation(s) for the application of dipolar fields for resonance excitation typically employ numerical methods and simplified models to describe ion motion.63 Exact solutions cannot be obtained because of electrode surface geometry considerations (machining of hyperbolic surfaces, truncation effects, and endcap holes) and the presence of higher-order fields which result in a perturbed quadrupolar trapping field, therefore affecting ion motion. The aforementioned reasoning results in coupled motion in the r- and zdirections, thus making exact solutions difficult to solve mathematically.

A preliminary investigation to determine the influence of storage conditions (under the influence of dipolar excitation) on photodissociation efficiency was undertaken with protonated diglyme. To minimize the effects of collisions on photodissociation efficiency, the ion trap was again operated with no He buffer gas. A detailed description on the instrumentation employed and the procedures for data acquisition, can be found in the Experimental Design section of this chapter.







74

To determine the appropriate dipolar excitation frequencies for the ion motion studies, the (M+H)+ ion of protonated diglyme was stored for a period of 10 ms at a q, value of 0.3. During the storage period, a dipolar excitation signal was applied to the endcaps with an amplitude of 6 V,_,. The frequency range probed was from 25 kHz to 500 kHz, with the frequency incremented at intervals of 10 Hz. A plot of the intensity of m/z 135 of protonated diglyme versus the excitation frequency yielded a series of absorption bands with the center frequency assigned as the minimum intensity value of m/z 135 for a given absorption band. For absorption bands which were offscale for a 6 V,., dipolar excitation amplitude, the center frequency was taken as the center point of the FWHM for the band. To assign the ion's frequency from an absorption band as accurately as possible, several instrumental parameters must be strictly controlled. These include varying only one instrumental parameter per scan table and allowing appropriate stabilization times for changes in the various instrumental parameters (rf voltages, dc voltages, dipolar frequency, etc.). A detailed description of all the factors involved has been published by Eades et al.12"1"6

Typically, the minimum value of frequency optimization curve is taken as a close approximation of an ion's component frequency of motion. As previously mentioned, shifts in the observed frequency due to the instrumental variations mentioned above can cause inaccuracies in proper frequency assignment. However, other experimental factors can also contribute to inaccuracies in frequency assignment for a given component of ion motion. These factors







75

include the amplitude of the applied excitation signal (frequency shifts to higher values with increasing amplitude), He buffer gas pressure as described by the physics of the damped harmonic oscillator model (frequency shifts to lower values with increasing gas pressure), and space charge or ion-ion interaction considerations (frequency shifts to lower values with increasing ion population). In addition, arbitrary shifts in an ion's frequency can also occur due to the uncertainty of the exact RF voltage applied to the ring electrode of the ion trap. Consequently, the experimentally determined ion component frequencies are only qualitative estimates of the true ion frequencies.

The power absorption spectrum of protonated diglyme (no He buffer gas present) for q,=0.3 and az=0 is shown in figure 2-7. Three main frequency bands were observed at wz (118.3 kHz), 2w, (235.9 kHz), and 3w, (349.8 kHz). The broad frequency band at 55.1 kHz could correspond to the W/2 band, however the broad bandwidth precludes direct frequency assignment at this time. An unknown frequency band at 446.0 kHz is also present. The largest absorption band (the ion's secular frequency of motion in the z-direction) at w, was very broad due to the large excitation amplitude used. Observation of the smaller bands at 2wZ and 3w2 can be attributed partly to higher-order field effects (hexapolar) on ion motion and have been successfully predicted using forced Mathieu equations as developed by Williams et al." To account for power absorption at these frequencies, several different explanations (or combinations thereof) are plausible, including: 1) the presence of higher order fields






















Figure 2-7:


The power absorption spectrum of protonated diglyme for an applied dipolar excitation signal (6 V,.,) at q,=0.3 and a,=O (no He buffer gas present). The applied dipolar signal was started at a frequency of 25 kHz and incremented at 0.1 kHz intervals to 500 kHz. The symbol wZ refers to the fundamental frequency of motion for an ion in the z-direction.









1200 1000 800 600



400 200



0


II'


30) z
f=349.8 kHz
2 w f=446.0 kHz
f=235.9 kHz 4j


(o z
f=118.3 kHz
4


.3 U UI


100


200


300


400


500


Dipolar Excitation Frequency in kHz


5
f=55.1 kHz







78

(hexapolar); 2) harmonic considerations of the applied dipole field; or 3) direct contributions to ion motion. A detailed investigation into the origin of these absorption bands has been performed by Eades12 and Vedel.16 These two authors were independently able to verify the existence of the 2w, frequency band as a component of ion motion and not just part of the harmonics from the excitation field. Other absorption bands observed by Eades and Vedel which were not seen in these experiments include w/2 and w,+w,. Several plausible explanations for this observation include the q. value used, the large w, band width, the absence of He buffer gas, and the sensitivity of the chemical probe (protonated diglyme).

To examine the effect of dipolar excitation on ion (axial) motion, the protonated diglyme ions at m/z.1 35 were irradiated with a cw C02 laser during the dipolar excitation (10 ms) period. Since the time between consecutive photon absorption for the IR laser in this experiment is on the ms time scale, instantaneous ion trajectories or velocities cannot be determined. However, C02 lasers can be used to evaluate the time-averaged behavior of the ion (cloud) population. The multipass optical design used in these experiments produces a high photon density in the radial plane of the ring electrode as shown by figure 2-6. Any ions whose axial excursions exceed the beam width of the laser (3 mm) should show a marked decrease in photodissociation efficiency."

To test this theory, protonated diglyme ions were excited at a dipolar frequency of 118 kHz (wa), which was determined from the power absorption







79

spectrum. The effect on the photodissociation efficiency of exciting the [M+H]' ions of diglyme (no He present) using dipolar resonant excitation (during period 5 in figure 2-3) is seen in figure 2-8. With increasing dipolar resonant excitation voltage applied to the endcaps, the axial excursions of the ions increase, decreasing the fraction of time the ions spend in the radial plane of the ring electrode which, in turn, decreases the extent of interaction between the stored ions and photons. The decrease in photodissociation efficiency can be further rationalized by considering the inset on figure 2-8. As long as the ion's maximum excursion (trajectory) in the axial or z-direction does not exceed the laser beam width (dashed line figure 2-8), the photodissociation efficiency remains relatively constant. However, if the ions gain enough kinetic energy from the applied dipolar field so that ion trajectories exceed the laser beam width in the axial direction, a significant decrease in photodissociation efficiency is observed. This decrease in efficiency can be explained by considering instantaneous ion velocity arguments. As an ion moves away from the center of any rf-only device (quadrupole or ion trap), the magnitude of the restoring forces become larger, as for any harmonic oscillator. Therefore, as an ion reaches its maximum displacement from the center of the ring electrode, its instantaneous velocity is very slow compared to that of the same ion as it accelerates through the center of the device. This means that when the peak-to-peak excursions of the individual ion trajectories are larger than the laser beam width in the axial direction, the ions spend most of their time outside the radial plane of the ring electrode where the





















Figure 2-8:


Effect of resonant excitation (w,= 118.3 kHz, qz=0.3, a,=0) voltage on photodissociation efficiency. When the axial excursions of the ions exceed the width of the CO2 laser beam (as indicated in the figure inset by the solid line), photodissociation efficiency drops off dramatically, since the ions spend a significant amount of time at their maximum excursions outside the beam width. Error bars are defined as the standard deviation of the mean.









0.8


0.7-1 Laser Beam
Width

0.6
z

0.5- L>


0.4
o 1

0.3- Ion Trajectory


0.2


0.1


0.0- i I I I I I i I I
0 10 20 30 40 50 60 70
Dipolar Resonant Excitation Voltage in mV







82

photon density is highest (see solid line figure 2-8). This displacement results in the observed sharp decrease in photodissociation efficiency for a dipolar excitation amplitude of 15 mV.16

The unique advantage of the multi-pass ring electrode system (as shown above) is its sensitivity in detecting the presence of axial excitation. By exciting ions in the radial or r-direction (using quadrupolar excitation), the multi-pass ring electrode system should be able to detect the presence of coupled motion in the z-direction. However, this system would not be useful in detecting coupled ion motion in the r-direction since there is a large photon density in the radial plane of the ring electrode, which would not result in a drop-off in photodissociation efficiency if the ions were excited radially. In the next section, a series of experiments is described in which quadrupolar excitation/photodissociation experiments are used to detect axial excitation (z-direction) when a given ion population is excited only in the radial or r-direction."


Quadrupolar Excitation


Over the last several years, the use of quadrupolar excitation as an experimental tool has drawn increasing attention. '''6 The basic premise of quadrupolar excitation is the application of the same auxiliary ac potential (inphase) to both endcap electrodes. The application of this in-phase ac signal creates a symmetric field similar to the quadrupolar rf trapping field on the ring electrode. Therefore, quadrupolar excitation does not favor either radial (r-







83

direction) or axial (z-direction) excitation of a given ion population with the application of the excitation signal. This phenomenon can be understood by examining the relationship between ion motion and the quadrupolar mode of excitation. As an ion approaches each endcap, the potential on that endcap needs to be of the appropriate polarity to obtain maximum power absorption. With the application of an in-phase auxiliary ac potential, the excitation signal must be applied at twice the ion's axial or radial frequency. Figure 2-9 demonstrates the sequence of events needed for maximum power absorption using quadrupolar excitation compared with that of dipolar excitation. Previously published theoretical simulations have predicted strong power absorption at the 2wU frequencies (where u is either for the r- or z-directions) for quadrupolar excitation. '" -'716,6

The trapping conditions observed with the application of a quadrupolar excitation signal can be significantly different than those seen by direct application of a quadrupolar trapping field to the ring electrode. Therefore, the resulting ion trajectories no longer follow traditional Mathieu parameters, since ion motion is now controlled by both the quadrupolar trapping field and the excitation field. Due to the quadrupolar nature of the excitation field, exact solutions to the equations of motion can be calculated (March et al.)'". In this report, March and coworkers were able to predict the presence of the 2wU bands, with the 2wz band having a somewhat greater power absorption than the 2w, band.5'"'


















Figure 2-9:


Effect of the applied excitation signal (e.g dipolar or quadrupolar) on ion motion in the quadrupole ion trap. The symbol w, refers to the fundamental frequency of motion of an ion in the z-direction. For maximum power absorption under dipolar excitation conditions, the ion's frequency should match that of the applied supplementary ac signal as seen in (a). The negative cycles of the supplementary ac coincide with the ion frequency in the z-direction, thus leading to an increase in amplitude for the ion trajectory and corresponding increase in ion kinetic energy. For quadrupolar excitation where the applied supplementary ac signal is in phase with respect to both endcaps, the applied frequency must be two times that of the fundamental ion frequency to obtain maximum power absorption. An applied supplementary ac at the identical frequency of motion of an ion in the z-direction leads to a significantly reduced power absorption spectrum (as seen experimentally in figure 2-10).























rf=O




v-vv


Quadrupolar Excitation (Exit Endcap)O


Quadrupolar Excitation (Entrance Endcap)


Dipolar Excitation (entrance endcap)


Dipolar Excitation (Exit Endcap)







86

An investigation to determine the influence of storage conditions (under the influence of quadrupolar excitation and thus determine the presence of coupled ion motion in the axial or z-direction while exciting the ions radially) was undertaken using protonated 12-crown-4 ether. As mentioned previously, once the ion trajectories exceed the 3 mm laser beam width in the axial or z-direction, a marked drop off in photodissociation efficiency will occur, indicating the presence of excitation in the axial direction. To minimize the effects of collisions on photodissociation efficiency, the ion trap was again operated with no He buffer gas. A detailed description of the instrumentation employed, limitations of frequency measurements, and the procedures for data acquisition can be found in the Experimental Design and Dipolar Excitation sections of this chapter.

To determine the appropriate quadrupolar excitation frequencies for the ion motion studies, the [M+H)f ion of protonated 12-crown-4 ether was stored and analyzed as described previously in the Dipolar Excitation section of this chapter. The frequency range probed was from 25 kHz to 500 kHz, with the frequency incremented at intervals of 10 Hz. A plot of the intensity of m/z 177 of protonated 12-crown-4 ether versus the excitation frequency yielded a series of absorption bands with the center frequency assigned as the minimum intensity value of m/z 177 for a given absorption band. For absorption bands which were offscale for a 6 V,_, quadrupolar excitation amplitude, the center frequency was taken as the center point of the FWHM for the band.16







87

The power absorption spectrum of protonated 12-crown-4 ether (no He buffer gas present) for q,=0.3 and a,=0 is shown in figure 2-10. Three main frequency bands were observed at w, (59.2 kHz), 2w, (126.0 kHz), and 2w, (239.8 kHz). Two of these absorption bands, at 2w, and 2w,, were offscale, while the W, band showed limited absorption. The magnitude of the 2w, and 2w. bands was approximately equal, which corresponds well with results reported by Eades et al.'2" Also, the magnitudes of all absorption bands observed agree well with the theoretical values calculated by March.515,11216 Note that the absorption bands at 2wZ and 2w are not directly related to frequencies of ion motion, but instead represent only the absorption of power. This can be explained by the frequency doubling of the quadrupolar excitation signal applied to the endcaps, such that the charge on an endcap is of the appropriate polarity to obtain maximum power absorption (figure 2-9). Therefore, this absorption of power at 2wr and 2wZ represents actual ion motion at w, and wz, respectively.

To examine the effect of coupled ion motion using quadrupolar excitation, the appropriate excitation signal was applied at the three frequencies observed (w,, 2w, and 2w) with concurrent laser irradiation, as mentioned previously in this chapter. A plot of the photodissociation efficiency versus the quadrupolar excitation voltage applied to the endcaps yields the three curves shown in figure 2-11. As expected for the 2wz band (representing wz frequency component), there is observed a steep drop-off in photodissociation efficiency as the axial excursions of the ions exceed the laser beam width (for the reasons mentioned






















Figure 2-10:


The power absorption spectrum of protonated 12-crown-4 ether for an applied quadrupolar excitation signal (6 V,_,) at q,=0.3 and a,=0 (no He buffer gas present). The applied quadrupolar signal was started at a frequency of 25 kHz and incremented at 0.1 kHz intervals to 500 kHz. The symbols w, and w, refer to the fundamental frequencies of motion for an ion in the z- and r-direction respectively.















KA\vx


0)r
f=59.2 kHz




20 r kl f=126.0 kH z


800 600



400 200



0


200


2 oz
f=239.8 kHz
j


300


400


500


Quadrupolar Excitation Frequency in kHz


1200


1000


I
100


-~


I


CO
(0





















Figure 2-11:


Application of quadrupolar excitation signals at 59.2 kHz (w,), 126.0 kHz (2w,), and 239.8 kHz (2w,) to determine the presence of coupled motion observed in the z-direction with quadrupolar excitation. The observed decrease in photodissociation efficiency for the 2W, band cannot be explained by radial excitation phenomena due to the presence of high photon density across radial plane of the ring electrode.









1.00




0.80




0.60- 2 o




0.40 2o z




0.20




0.00

0 500 1000 1500 2000 2500 3000 3500 4000

Quadrupolar Resonant Excitation Voltage in mV




Full Text
Figure 4-5: The modified Analytica electrospray ion source showing the elimination of the second skimmer
cone, lens L2 and lens L3. Also shown is the stainless steel adaptor ring used to extend the first
skimmer cone region 0.250" forward for coupling to the rf-only octopole.


38
absorption can occur for the given structure of the ion. The information obtained
from these experiments can be used as a fingerprint in the determination of ion
structures.126,127
In trapping instruments, photo-induced ion decay can be measured as a
function of irradiance time. These data can be used effectively to distinguish
isomeric ion structures in the gas-phase. Kinetic energy release data can also be
used (below 100 mV) to add important ion fragmentation information.128
Infrared Multiple Photon Dissociation (IRMPD)
Infrared multiple photon dissociation is particularly well suited for use with
trapping instruments. The ability to trap ions for extended periods of time at low
pressures in both the ion cyclotron resonance (ICR) cell and the quadrupole ion
trap mass spectrometer allows for the sequential absorption of infrared photons
via low intensity infrared radiation. The IRMPD process was first characterized by
Beauchamp and co-workers for positive ions using an ICR mass
spectrometer.129,130 These and other studies focused primarily on the chemical
kinetics, reactivity, and spectroscopy of various ionic species. Brauman and co
workers were the first to study the vibrational relaxation of gas-phase ions and the
accompanying physics of pulsed megawatt infrared multiphoton dissociation
using a C02 laser.131 For the case of trapped negative ions, irradiation can
produce both a photodissociation and an electron photodetachment spectrum.132'
134


236
endcap, which facilitates efficient transfer of ions from the octopole to the ion
trap).
The octopole was positioned by an aluminum mounting piece with an
internal diameter designed to brace the Delrin support at the end of the octopole
rod assembly. The aluminum mounting piece was designed with three large cut
outs to aid pumping around the analyzer region. The piece was mounted on top
of the ion trap analyzer assembly using three alumina ceramic supports. A
schematic diagram of the octopole support can be seen in figure 4-13.
The next part of the analyzer assembly was an exit tube lens made from
stainless steel to transfer the ions from the mass analyzer to the detector. This
exit lens consisted of a single piece of stainless steel with a reduced orifice tube
lens used to reduce field penetration effects from the 20 kV dynode. The back
side of this assembly was threaded to accept a variety of tube lens extensions.
The tube lens extension was needed to help transfer ions from the analyzer to the
detector because the physical constraints of the manifold would not allow
placement of the detector assembly within 2" of the exit endcap. The exit tube
lens was mounted to the bottom of the analyzer assembly again using alumina
ceramic supports. The exit tube lens assembly and corresponding extension
piece are shown in figure 4-14.
The analyzer stack was assembled in the following manner: (1) aluminum
mounting plate (bottom); (2) exit tube lens; (3) exit endcap; (4) multipass ring
electrode, (5) exit endcap (e.g., ion entrance), (6) octopole mounting assembly;


346
carbohydrate or oligosaccharide can also be linked to a variety of other
biochemically important compounds to form glycoproteins, peptidoglycans,
lipopolysaccharides, and glycoshpingolipids. To determine the structure of the
oligosaccharide, or oligosaccharide portion of a complex biomolecule, several
steps are necessary to establish structure: (1) determination of the reducing and
nonreducing ends of branched or straight chain oligosaccharides, (2)
determination of monosaccharide ratios and ring sizes, (3) determination of the
internal sequence of individual monosaccharides, (4) determination of branching
points (if any), (5) anomeric data relating to linkage type and bonding, (6)
presence of modified sugars possibly containing acyl groups, phosphates,
sulfates, pyruvates, cyclic acetals, or taurine moieties. To obtain the
aforementioned information, a myriad of techniques are employed, including wet
chemical, enzymatic, antibody or lectin affinity, chromatography (thin layer, gas,
liquid, paper), nuclear magnetic resonance, circular dichroism, and mass
spectrometry.47,294'297
The major contribution of mass spectrometry to carbohydrate analysis has
been that of high sensitivity (nanomole to femtomole) studies which have included
selective detection of glycopeptides in protein digests, multi-residue confirmation
of aminoglycoside antibiotics, sequencing of cationized carbohydrate antibiotics,
identification and linkage position determination of reducing ends in
oligosaccharides, and the structural analysis of steroidal oligoglycosides.275,298"301
Collision-induced dissociation studies needed to help determine oligosaccharide


415
Molecules Kompa, K.L.; Smith, S.D., eds.; Springer-Vertag: New York,
1979, 134-137.
152. Quigley, G.P. in Chemical Physics 3, Advances in Laser Chemistry Zewail,
A.H., ed.; Springer-Verlag: New York, 1978, 374-383.
153. Wu, H.F.; Brodbelt, J.S. Int J. Mass Spectrom. Ion Processes 1992, 115,
67-81.
154. Reiser, H.-P, Kaiser, R.E.; Savickas, P.J.; Cooks, R.G. Int. J. Mass
Spectrom Ion Processes 1991, 106, 237.
155. White, J.U. J. Opt. Soc. Am. 1942, 32, 285-288.
156. March, R.E.; McMahon, A.W. Allinson, E.T.; Londry, F.A.; Alfred, R.L; Todd,
J.F.J.; Vedel, F. Int. J. Mass Spectrom. Ion Processes 1990, 99, 109-124.
157. March, R.E.; Londry, F.A.; Alfred, R.L.; Todd, J.F.J.; Penman, A.D.; Vedel,
F.; Vedel, M. Int. J. Mass Spectrom. Ion Processes 1991, 110, 159-178.
158. Alfred, R.L.; Londry, F.A.; March, R.E. Int. J. Mass Spectrom. Ion Processes
1993, 125, 171-185.
159. Hemberger, P.H.; Nogar, N.S.; Williams, J.D.; Cooks, R.G.; Syka, J.E.P.
Chem. Phys. Lett. 1992, 191, 405-410.
160. Williams, J.D.; Cooks, R.G.; Syka, J.E.P.; Hemberger, P.H.; Nogar, N.S. J.
Am. Soc. Mass Spectrom. 1993, 4, 792-292.
161. Wang, Y. Fundamental Study of Magntica! and Electrical Ion Traps Ph.D.
Dissertation, Bremen, Germany, 1992.
162. Eades, D.M. Higher Order Field Effects on Stored Ions in a Quadrupole Ion
Trap, Ph.D. Dissertation, University of Florida, Gainesville, FL, 1994.
163. Reiser, H-P; Julian Jr., R.K.; Cooks, R.G. Int. J. Mass spectrom. Ion
Processes 1992, 121, 49-63.
164. Eades, D.M.; Johnson, J.V.; Yost, R.A. in Proceedings of the 41st ASMS
Conference on Mass Spectrometry and Allied Topics San Francisco, CA,
1993, 693a-693b.
165. Williams, J.D.; Reiser, H-P.; Kaiser, R.E.; Cooks, R.G. Int. J. Mass
Spectrom. Ion Processes 1991, 108, 199-219.


300
the dashed lines in figure 4-29). The representative spectra are shown in figure
4-30. For an rf level of m/z 25, the doubly-charged peak at m/z 262 predominates
with only a small contribution from the singly-charged species at m/z 524. At a
higher rf level of m/z 38, the singly-charged [M+H]+ ion is approximately 33% of
the m/z 262 (+2 charge state) base peak. It is clear from this simple low mass
range example that a series of ion injection steps at various rf levels would be
appropriate in order to obtain a "truer" representative ESI mass spectrum. For the
horse muscle apomyoglobin and bovine heart cytochrome c samples, discussed
in the Basic ESI Operation section, the rf level was incremented during the ion
injection period (see table 4-4) to obtain the spectra seen in figures 4-23 and 4-
24.
Another interesting aspect of figure 4-30 is the presence of the m/z 393
peak (or the y3 ion), indicating the loss of methionine from either the singly-
charged ion at m/z 524 or the doubly charged ion at m/z 263. This peak
essentially arises from the ion trap offset value of -5 V (translational energy),
although preliminary evidence indicates that injection at somewhat higher q2
values can contribute to fragmentation. It should also be noted that some ions
may be eliminated entirely from the spectrum if CID occurs before or during the
ion injection into the analyzer. Therefore, to determine the importance of the CID
process during ion injection becomes increasingly difficult, especially for multiply-
charged systems where coulombic repulsion forces are strong and the
corresponding energy deposition for CID is small.


Figure 4-30: Spectra of the peptide MRFA taken at an rf level (low-mass cut
off) of m/z 25 and m/z 38. This experiment demonstrates the
need for either scanning or incrementing the rf level during ion
injection in order to obtain a representative ESI spectrum.


Figure 4-37: MS3 spectrum of the b9+2 ion from angiotensin I. The predominant ions observed are the higher
m/z ions b6 and b8. The unknown peaks may correspond to cyclization for the peptide backbone
structure; however, without proper labeling experiments this cannot be confirmed.


93
predominates, but there exists a significant amount of axial excitation associated
with the application of the 2u, frequency in the quadrupolar excitation
mode.57156'157
The Photon Absorption Process
The minimum number of photons (n) needed to reach the thermodynamic
threshold for dissociation via a given reaction channel was calculated by dividing
the enthalpy change by the photon energy:
A H = nhv Na (2-3)
where AH is the enthalpy change, n is the number of photons absorbed, h is
Planks constant, v is the frequency of the radiation used, and NA is Avogadros
number.
The consecutive reaction sequence for protonated diglyme as it absorbs
IR photons is seen in equations 2-4 and 2-5:
[C6H1503r nhv [C5H02r CHjOH (2-4)
[C5H02]* + nhv [CgHyO]- C2H40 (2-5)
The AHf values for neutral methanol, neutral acetaldehyde, and protonated
diglyme were obtained directly from the literature.170 The AHf values for the
photofragments m/z 103 and 59 were estimated by using the quantum
mechanical AM1 (Austin Model 1) calculations based on the neglect of diatomic
differential overlap approximation.171'172 The accuracy of the AM1 calculations was


194
beam image exiting from the octopole as small as possible. This reduces the exit
angles of the ions as they leave the device, thus leading to improved ion
transmission properties and a well defined translational ion energy.253
Collisional Focusing
Collisional focusing of ions in an rf-only device was first reported by
Douglas and French.257 In their experiments, an rf-only quadrupole was used to
inject ions from both atmospheric and electrospray ionization sources into a mass
analyzing (rf/dc) quadrupole. For ion injection energies on the order of 1 to 30
eV, ion transmission efficiencies were found to increase through the rf-only
quadrupole at pressures up to 8x1 O'3 torr. The increased ion transmission
observed was found in direct contrast with those predicted by traditional
scattering models. The collisional focusing capabilities of the rf-only quadrupole
were found to be quite similar to those observed in the three dimensional
quadrupole ion trap.257
Douglas and French reported that collisional focusing improved with
increasing ion mass and not mass-to-charge ratio. The improved ion transmission
was attributed mainly to a significant loss in the ions axial kinetic energy. In
addition, higher mass ions (at the same q value of lower mass ions) were
transmitted with higher rf voltages applied to the rods, thus leading to a greater
well depth and better confinement along the central axis of the rf-only quadrupole.
Monte Carlo simulations of this phenomenon provided good agreement with


Figure 2-11: Application of quadrupolar excitation signals at 59.2 kHz (a/r), 126.0 kHz (2wr), and 239.8 kHz (2coz)
to determine the presence of coupled motion observed in the z-direction with quadrupolar
excitation. The observed decrease in photodissociation efficiency for the 2a/r band cannot be
explained by radial excitation phenomena due to the presence of high photon density across radial
plane of the ring electrode.


Analyzer Mounti
01.850'
Top View
Plate (Bottom)
/- 03.750'
Dimensions for
Inner and Outer
I.D., and Cut-outs
for Set Screws
2-56 tap, 0.0B9" drill
Number: 1
Material: Aluminum
Modifications: Anodize
All Units are in Inches
Page 1 of 5


Top V¡ew(s)
Tube Lens Assembly
Side View(s)
0.326
L
0.531
| j 0,389'
ifl iu
-ve=>
C
r540*
T tf=3
0.560'
Thread to 0,215
appropriate size
7T
All Units are in Inches
240


Figure 4-1:
Mounting plate design for attachment of the vacuum manifold to the instrument support table.
Circular portions indicate 8" outline of 500 L/s turbomolecular pumps.


Raffnose
Bia
m/z 164 (-NH3)
m/z 343 Z2a
m/z 326 (-NH3)
D-Glalactose
D-Glucose
m/z 343
m/z 326 (-NH3)
D-Fructose
m/z 164 (-NH3)
Zia
373


126
spectrum with that of the positive ion IRMPD spectrum is feasible. To obtain the
gas-phase neutral spectrum of allyl bromide (3-bromopropene), a Nicolet 7199
FTIR spectrometer equipped with a quartz cell and KBr windows (10 cm
pathlength) was used. The cell was filed with allyl bromide to a pressure of 0.45
torr. In figure 2-20 is shown the gas-phase neutral spectrum (850-1350 cm'1) and
the IRMPD spectrum of allyl bromide. The widths of the doublet peaks at 944 cm'
1 and 976 cm'1 are significantly smaller than those of the corresponding neutral.
The allyl bromide ions stored in the ion trap undergo at most one to two collisions
during the laser irradiance period, meaning the spectrum observed will not be
collisionally broadened. For the case of the gas-phase neutral, at a pressure of
0.45 torr collisional broadening is observed as evidenced by the approximate 90
cm'1 bandwidth of the doublet peak at 920 cm'1. As mentioned previously, the
widths of the peaks obtained with the ion trap were again somewhat larger than
those observed with the ICR cell (which follows the same reasoning stated
above). The individual absorption peaks observed in the IRMPD spectrum are
due to the overlap of sharp C02 laser lines with the sharp molecular absorption
bands. For larger more complex molecules/ions, where the absorption of the first
photon may be in the vibrational quasicontinuum, a more broad featureless
spectrum may be expected.148,189


Filament
End Cap
Electron Multiplier Detector
To Preamplifier
(Ion Signal)
Amplifier and
RF Generator,
Fundamental
RF Voltage
£
Scan Acquisition
Processor
(Computer)
£
- ¡
Ring Electrode ^
End Cap
V
JT
Amplifier and
RF Generator,
Supplementary
RF Voltage


151
(3-42)
which has the general solution:
Z = Z0 + v2(t t0)
(3-43)
with z defined as the initial position in the z-direction and vz the velocity in the
z-direction.
The force acting on a charged particle in the xy plane of an octopolar field
is obtained by combining Newtons law of motion with the Lorentz force law
(magnetic field strength of zero). The total force F is defined as:
F = m = ma = eE
dt
(3-44)
where v is the velocity in the xy plane, e is the charge on the ion, and E is the
electric field strength vector as defined in equation 3-31. Using the component
vectors of the electric field strength, equation 3-44 can be rewritten in rectangular
coordinates as:
,¡ = iF(¡iE
dt2 dt2 m x m
(3-45)
substituting for E, (equation 3-34) and Ey (equation 3-37), the equations of motion
in the x- and y-direction in rectangular coordinates are written as:


Intensity
3000 i
2500
2000
1500
1000
500
(M+3H)
+3
HPF or PFH
Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-His-Leii
200
300
400
500
600
700
m/z
800
900
1000 1100 1200
341


148
X


Displacement (mm) |Z|
176


249
The vacuum-air interface for both the pulsed and cw lasers consisted of a
2.75" flange modified to accept a 1.5" ZnSe window (Melles Griot, Irvine, CA). The
vacuum seal was made by two teflon rings placed on either side of the ZnSe
window.
The photodissociation set-up for the cw laser can be seen in figure 4-17.
A 13" x 20" optical table was constructed and used for mounting the various ion
optics. The 1 cm diameter unfocused beam was passed through a beam selector
(designed to pass IR radiation) and reflected at a 90 angle off a gold plated
mirror. The beam then entered the mass spectrometer through the ZnSe window.
A helium-neon (632.8 nm) laser (model 05 LLR 851, 5 mW power output, from
Melles Griot, Irvine, CA) was used to expedite the alignment of the cw C02 laser.
The helium-neon laser was placed on top of the instrument table and aimed
directly at the 1" beam selector (Melles Griot, Irvine, CA) which reflected the 632.8
nm light and passed IR (9.0 to 11.0 //m) radiation. This greatly simplified the
alignment process since the helium-neon laser did not need to be placed in line
with the cw C02 laser for beam alignment.
In the case of the pulsed laser (as seen in figure 4-18), the beam selector
was not used since the beam shape from the laser (approximately a 1" x 1.25"
square beam) exceeded the dimensions of the beam selector. As mentioned
previously, the beam was first reflected at a 90 angle off the surface of a gold
plated mirror. To focus the beam down to the appropriate size to pass through
the entrance aperture of the multipass ring electrode, a convex focusing mirror


Photodissociation Efficiency 1-(I/I0)
Laser Beam
Width
$
Ion Trajectory
| i i i i | i i i i | i i r~r | i i i i |
10 20 30 40 50 60
Dipolar Resonant Excitation Voltage in mV
00


Electron
245


348
Monosaccharide Cleavage
The first step in obtaining relevant carbohydrate sequence data employing
photodissociation is understanding the ring fragmentations of simple
monosaccharides subunits. The goal was to determine the number of ring
substituents on an individual monosaccharide. In this section the ring
dissociations of 2-deoxy-D-glucoseand 1 -O-methyl-D-glucopyranoside (structures
shown in figure 5-5) are discussed, which were used as a feasibility study for
more advanced applications (e.g., glycosidic bond cleavages). The samples were
introduced via a solids probe into a Finnigan MAT ITMS mass spectrometer set
up for photodissociation experiments as described (see figure 2-1) in chapter 2.
Sample ionization was accomplished by ammonia Cl with reaction times on the
order of 65 to 90 ms and ammonia pressures of 5.5x1 O'6 torr. For CID
experiments, the helium buffer gas pressure was 1.0x10^ torr (uncorrected). To
maximize photodissociation efficiency, the ion trap was operated without helium
buffer gas.
The CID mass spectrum of the ammonium adduct of 2-deoxy-D-glucose
(m/z 182) is shown in figure 5-6. The observed dissociation efficiency of the
ammonium adduct ion was 90%. The major peak produced at m/z 164 involves
the loss of neutral H20 from the 2-deoxy-D-glucose ring. The peak at m/z 147
also showed the loss of neutral H20 with an additional loss of ammonia. Low
intensity ions at m/z 146 [M+NH4+-2H20] and m/z 129 [M-2H20-NH3]+ indicated
the loss of a second neutral H20 from the monosaccharide ring. To determine


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 E512T48FH_FPG1JV INGEST_TIME 2015-08-13T22:11:25Z PACKAGE AA00029830_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


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
A NOVEL ELECTROSPRAY ION TRAP MASS SPECTROMETER FOR
PHOTODISSOCIATION OF BIOLOGICAL MOLECULES
By
James L. Stephenson, Jr.
December, 1995
Chairperson: Dr. Richard A. Yost
Major Department: Chemistry
The combined techniques of photodissociation and mass spectrometry
have been used extensively to study the fundamental aspects of gas-phase ion
chemistry. Within the last several years, a great deal of interest has been shown
in employing photodissociation as an analytical tool for the structural elucidation
of biological molecules, due in part to the limitations of traditional tandem mass
spectrometrictechniques (e.g. collision-induced dissociation) in providing relevant
structural information. This dissertation presents the design and characterization
of an electrospray ion trap mass spectrometer, capable of performing
photodissociation experiments on a wide range of biological molecules.
Previous photodissociation studies have been limited to fundamental
investigations due mainly to the limited photoabsorption cross-sections observed
for organic ions. In order to increase the photodissociation efficiency, three
Vlll


Figure 1-5:
Typical ESI ion source used in mass spectrometry. The glass capillary can be replaced by a
heated stainless steel capillary. The skimmer and ion source lens assembly design will vary
depending upon the instrumental configuration.


Figure 3-15: (a) Secular motion of an ion in an rf-only octopole, m/z 1000, V0.P=400 V, ion entry 0.5 mm, angle
of 3, w=1.2 MHz. (b) Transverse kinetic energy in an rf-only octopole, conditions as in (a).
Adapted from reference 253.


130
only to focus the ion beam and not to recapture" scattered ions for ion injection,
an alternative method which could recapture scattered ions and focus them into
appropriate trajectories for ion injection would be desirable.
One technique which fulfills this requirement is the rf-only multipole. RF-
only multipoles have been used extensively in both analytical and physical mass
spectrometry217,218 The ability of these devices to focus ions in a high pressure
environment can be understood by examining the forces exerted on a given ion
population as it moves through an rf-only device. As an ion is displaced from the
center axis of the device, the restoring force acting upon that ion (in an rf-only
multipole with 2n electrodes) is proportional to the (n-l)* power of that
displacement, where n=2,3, and 4 for a quadrupole, hexapole, and octopole,
respectively219 Therefore, as ions are displaced from the center of the rf-only
device due to collisions with neutral atoms or molecules, the restoring forces
present recapture the displaced ions and successfully transfer them from a high
pressure region (e.g., electrospray ion source) to a lower pressure region (e.g.,
ion trap analyzer) with minimal scattering losses. Based on the above discussion,
the octopole would be the logical choice for an ion injection device due to the
greater restoring forces present for n=4 case.
Previously, the rf-only octopole has been used as an ion-molecule reaction
cell220,221, as a collision cell in tandem quadrupole instruments222, as an ion
injection device for triple quadrupole instruments223, and as a device for
determining ion-molecule reaction cross sections and energetics (i.e., translational


58
most are the ionization and cool times. Since there is no appreciable
concentration of buffer gas (only sample gas pressure in the 10'7 torr range),
longer ionization times are needed to generate large numbers of "reagent" ions
for the self-chemical ionization process or for the production of molecular ions
[M+ ]. Typical ionization times for experiments with no He buffer gas are on the
order of 60 to 100 ms. Under normal He buffer gas conditions, 60-100 ms
ionization times would severely space charge the ion trap for sample pressures
in the 10'7 torr range. However, since trapping efficiency is severely reduced with
no buffer gas, a larger number of ions must be made (necessitating a longer
ionization time) to trap enough ions so as to limit the large statistical variations
associated with low ion densities.
Perhaps even more important than generating large numbers of ions is the
cool time associated with the experiment. Wu and Brodbelt showed that for low
pressure operation of the ion trap (pressures < 9x105 torr), increased cool times
led to very efficient storage conditions similar to those with 1 mtorr of He buffer
gas.153 Over any given time period, it was shown that about 50-100 collisions
were needed to cool ions to the center of the ion trap. The authors reported that
for shorter storage times and lower pressures, the majority of the ions in the
expanded ion cloud are likely to be accelerated into an endcap rather than
through the endcap exit holes during the mass-selective instability scan. Also, it
has been postulated that ions on the outer fringes of the ion cloud could receive


69
incidence), spots, dust or pinholes on mirrors A and B have a much less serious
effect since the light from any point on the mirror surface always goes back to the
same point; therefore, if there is a spot on the mirror surface, the light falling on
the spot is lost but on the second reflection from that mirror no more light is lost.
Yet another advantage is that there is only one transmission of light through an
entrance aperture; no light travels through any glass or other optical material
where losses due to reflection can occur.155
Each mirror (and its mounting bracket) was constructed from a single piece
of stainless steel. At one end of each piece, the radius of curvature was cut (r =
2.0 cm) and the surface was highly polished. The three mirrors were mounted
into precision-drilled holes in the ring electrode, positioned such that the
alignment was automatic; no realignment has been needed since the original
assembly of the multipass ring electrode (36 months). A small machine screw
was used to hold each mirror-mirror mount assembly in place on the ring
electrode. Laser alignment was set such that the center portion of the 1-cm
Gaussian beam profile was transmitted through the entrance aperture, thus
yielding high photodissociation efficiency. With efficiencies already greater than
those published for previous designs on a QUISTOR trap or ICR cell, condensing
the beam down to 0.3 cm was considered less important than examining the
numerous ways in which gas-phase ion chemistry can be studied via IRMPD with
this unique design. Furthermore, with the beam focused to 0.3 cm, damage
could possibly occur on the surface of the ion trap mirror when the laser is tuned


150
(x4 6x2y2 + y4) = 1 (3-39)
ro
coordinates are defined in dimensionless units, the three variables x, y, and r
become:
X = ; Y = -2-; R = (3-40)
r r0 r0
Therefore, the equation of the electrode contour line (from equation 3-39)
becomes:
X4 6X2Y2 + Y4 = 1 (3_41)
Accordingly, if a charged particle were to hit an electrode surface then the values
of X, Y, and R must be > 1219
Equations of Motion
The derivation for the equations of motion in an octopolar field need only
be considered in the xy plane. The force acting on an ion in the z-direction, or
along what is frequently termed the beam axis, is always zero. If the pole pieces
for the octopole are perfectly parallel and fringing fields are neglected, then the
ion will move along the axis of the octopole with a constant velocity determined
by its kinetic energy (offset potential) and entrance angle when it enters the
octopole. From Newtons law of motion, the z term is defined as:


CHAPTER 3
THE RF-ONLY OCTOPOLE ION TRANSMISSION GUIDE
Over the last eight years, ion injection into the quadrupole ion trap has
become one of the most popular areas of research in mass spectrometry. Since
the original report of ion injection using an external El ion source by Louris et
al.192, a myriad of different ion sources have been interfaced with the ion trap.
These include but are not limited to El/Cl192'194, fast atom bombardment (FAB)195,
particle beam196, thermospray197, electrospray35,198 202, glow-discharge203205,
atmospheric pressure206,207, inductively coupled plasma (ICP)208, laser
desorption209-213, super critical fluid214, and resonance enhanced multiphoton
ionization (REMPI)215"216 sources. In all the aforementioned literature, every
external ionization source has utilized some form of dc lens system for ion
injection into the ion trap. Although dc lens systems may be the simplest to
design and construct for an ion injection system, they are not necessarily the
most efficient way to transfer ions from an external source to the ion trap analyzer.
A stack of dc lenses has a poor conductance for letting neutrals introduced along
with the ions to be pumped away. This is especially true for high pressure
ionization sources such as electrospray or atmospheric pressure ionization where
a large number of collisions between ionized species and gas-phase neutrals can
occur, effectively scattering a large portion of the ions. Since dc lenses serve
129


142
variable, r, 0, and t, respectively.219
If the potential impressed between the octopole electrodes and the
potential of the oscillatory electric field generated between the electrodes (by the
individual octopole electrode potentials) have only one sinusoidal component
(e.g., assuming the electrode shape and alignment are perfect), then the general
solution to the Laplace equation may be written as:
(r,e,t) = (Akrk + Bkr"k)(Cksink0 + Dkcosk6)[U Vcos where Bk, Ck, and Dk are constants and (r,0) are defined as in figure 3-2.
Since 0(r,0,t) must be an even function (Ck=0) and must satisfy the following
boundary conditions:
= -^4>o = -^[u Vcoso)(t t0)] if r = r0 and 6 = 0 (3_25)
= 0 if 6 = ; 3; 5;...
8 8 8
Then the potential of an octopolar field with eight poles becomes:
cos 46
(3-26)
withp0 (the electrode potential) defined as in equation 3-20. To convert equation
3-26 from polar to rectangular coordinates, Moivres formula and the binomial
expansion can be used with:


Three-part figure showing: (1) the tube lens assembly with beveled ends to limit field penetration
from the 20 kV dynode, and to aid in ion transfer to the detector, (2) variable tube lens extension
piece complete with threaded end for extension to the 20 kV dynode detector, and (3) the tube
lens electrical connection.
Figure 4-14:


Length (mm)
150
200
187


Intensity
388


83
direction) or axial (z-direction) excitation of a given ion population with the
application of the excitation signal. This phenomenon can be understood by
examining the relationship between ion motion and the quadrupoiar mode of
excitation. As an ion approaches each endcap, the potential on that endcap
needs to be of the appropriate polarity to obtain maximum power absorption.
With the application of an in-phase auxiliary ac potential, the excitation signal
must be applied at twice the ions axial or radial frequency. Figure 2-9
demonstrates the sequence of events needed for maximum power absorption
using quadrupoiar excitation compared with that of dipolar excitation. Previously
published theoretical simulations have predicted strong power absorption at the
2uu frequencies (where u is either for the r- or z-directions) for quadrupoiar
excitation.57,156,157,162,166
The trapping conditions observed with the application of a quadrupoiar
excitation signal can be significantly different than those seen by direct application
of a quadrupoiar trapping field to the ring electrode. Therefore, the resulting ion
trajectories no longer follow traditional Mathieu parameters, since ion motion is
now controlled by both the quadrupoiar trapping field and the excitation field.
Due to the quadrupoiar nature of the excitation field, exact solutions to the
equations of motion can be calculated (March et al.)158. In this report, March and
coworkers were able to predict the presence of the 2uu bands, with the 2wz band
having a somewhat greater power absorption than the 2wr band.57,156-158


209
controlling the Finnigan MAT ITMS manifold temperature. The 220 V output was
stepped down via a transformer to 24 V (< 2 amp output current). Typical
operating temperatures were in the 170 to 215 C range. For this temperature
range, little if any thermal degradation or induced fragmentation of the [M+nH]n+
(where n=1, 2, 3...) ion(s) was observed.
The electrospray source was pumped by two 500 L/min rotary pumps
(model UNO 016B, Balzers Inc., Hudson, NH). A Harvard Apparatus model 22
syringe pump (South Natick, MA) was used for the direct infusion of samples into
the electrospray source.
To facilitate coupling of the rf-only octopole to the ESI source, several
design changes were made. First, to simplify the ion optics and obtain maximum
ion transfer efficiencies, the second skimmer cone, lens L2 and lens L3 were
removed from the electrospray source (see figure 4-4). An adapter ring was then
constructed from stainless steel to extend the remaining skimmer cone 0.250
forward. The adapter ring was mounted between the base plate of the
electrospray head and the skimmer cone, employing o-ring seals between the
skimmer cone and base plate assemblies. The alignment tool used to determine
the exact distance from the heated capillary exit to the skimmer cone was then
modified to compensate for the stainless steel adapter ring to maintain a constant
distance between the heated capillary and the skimmer cone at 3.5 to 4.0 mm.
The electrospray needle was also moved forward to compensate for the adjusted
heated capillary position. In figure 4-5 is shown the modified electrospray source


Figure 1-2: Mathieu stability diagram for the quadrupole ion trap. The
dimensionless quantity qz is directly proportional to the applied
rf voltage, while the dimensionless quantity is directly
proportional to the applied dc voltage. The 62 lines are a direct
relation to the secular frequency of motion for a given ion.
Adapted from reference 23.


Figure 4-15:
Complete drawing of the analyzer assembly including the alignment ring, octopole mount,
aluminum mounting plate, ion trap, tube lens, and detector assembly.


3
principles of ESI also is included. The chapter concludes with a discussion of
photodissociation that covers the basics of the photon absorption mechanism and
provides an introduction to the infrared multiple photon dissociation (IRMPD)
process.
The Quadrupole Ion Trap Mass Spectrometer
History
The origin of the quadrupole ion trap mass spectrometer dates back to the
original patent of Paul and Steinwedel, which disclosed the operation of the
quadrupole mass filter and the quadrupole ion trap.7 The first studies by Paul and
coworkers centered on the use of the device for ion storage over long periods of
time.8 Early detection procedures measured the power absorption of stored ions,
which utilized an rf voltage applied to both endcaps. In 1959, the original mass
selective detection studies were performed by Fischer, who demonstrated the
ability of the device to obtain unit mass resolution for a series of krypton
isotopes.9 In the early to mid sixties, spectroscopic studies of ions by Dehmelt
and Majors showed that high resolution studies were possible for all ground state,
metastable, atomic, and molecular ions.10-12 The first mass discrimination
experiments were performed in 1968 by Dawson and Whetten.13 This led to the
first use of the quadrupole ion trap as a true mass spectrometer. These
experiments utilized an external detector (electron multiplier) for the detection of
ions ejected through holes in the endcap electrodes. In 1971, it was discovered


305
Mass of
Resonance
Interest Point
QOO#

OOO*
000

Stability
Limit
OQO*
OOO*
0
q
0.908


132
number of electrodes (pole pieces). The derivation in this section represents a
summary of the work of Friedman et al.232 While not intended to cover in detail
all the mathematical nuances of the derivation, this qualitative discussion should
provide the reader with the necessary background to understand the basic
physics behind ion motion.
From Newtons law of motion the differential equation which describes the
motion of a charged particle through any field is defined as:
F=m (3-D
dt
where F is the force on the ion, m is the mass of the ion, and r is the radius
vector. For an electromagnetic field with components E for the electric field and
B for the magnetic field, the force on an ion is given by the Lorentz force law
where:
F = e(E + v x B) i3"2)
where e is the charge on the ion and v is the ion velocity. Since the rf-only field
applied to the multipole device is a time-varying field, the general form of the
Maxwell equations can be used to compute the E and B fields:
V D = p v x E =
at
(3-3)
V B=0 vxH=J+-^
at
here D is the electric displacement vector, p is the charge density in the medium,


111
Because the rate of formation of [C3H70]+ (m/z 59) is proportional to the ion
counts for [C5H1102]+ (m/z 103) the rate is zero at time ^ and has a maximum
value when [C5H1102]+ (m/z 103) reaches a maximum. Atthis maximum point, the
value of k, can be obtained:
-[C5;2r = 0 k,[C6H,503]* k2[C5H02]* (2-19)
at
^ ^[^6^15^3] (2-20)
2 [CsH.Ar
For protonated diglyme, k, was found to be 35 s'1. The derivation procedure does
not take into account the small back reaction (k<0.5 s'1) of m/z 59 with neutral
diglyme at a pressure of 3.6x1 O'7 torr to form [M+H]+ at m/z 135. In this case the
rate of formation of protonated diglyme from the back reaction only effects the
photodissociation of [M+H]+ from diglyme (formed via ion-molecule reactions) at
longer irradiance times (t>40 ms) where the reaction approaches completion.
The consecutive reaction sequence for 15-crown-5 ether (figure 2-16) was
also examined. To form a significant number of [M+H]+ ions at m/z 221, neutral
15-crown-5 ether was allowed to react with low molecular weight even-electron
species generated from El of the neutral. Reaction times for the protonated 15-
crown-5 ether were on the order of 400 ms. Following mass isolation of m/z 221,
the laser irradiance time was varied from 0 to 100 ms to track the series of
consecutive reactions. All reaction channels were verified using MS" experiments


Filament Endcap
Ring Electrode
Exit Endcap


5
In the early 1980s, a landmark series of papers by Hughes, March, and
Young first demonstrated the use of IR photodissociation in a quadrupole ion
trap.20"22 The technique of infrared multiple photon dissociation was used for
studying the gas-phase ion chemistry of the proton-bound dimer of 2-propanol.
Other early IRMPD experiments focused on wavelength dependence, collisional
cooling, dissociation efficiency, and analyte pressure dependence of various gas-
phase systems.23 A more detailed discussion of the instrumentation, experimental
parameters, and theoretical aspects of IRMPD in the quadrupole ion trap appears
in the photodissociation section of this chapter.
Perhaps the most important development in quadrupole ion trap mass
spectrometry was the mass selective instability scan developed by Stafford et
al.24,25 This mode of operation was radically different from previous modes in that
the rf amplitude applied to the ring electrode was ramped linearly with respect to
time. This enabled ions of increasingly higher mass-to-charge ratio (m/z) values
to be sequentially ejected from the ion trap. A mass spectrum could then be
recorded as a function of the time it takes ions of various m/z to be ejected to the
detector. Furthermore, by pressurizing the ion trap analyzer with approximately
10"3 torr of a light buffer gas (helium or hydrogen), drastic improvements in mass
resolution, sensitivity, and dynamic range were obtained for ion trap operation.26
The mechanism of action for these improvements involved the collision of ions
with the relatively slow and much less massive background gas atoms or


422
259. Kaiser, R.E.; Cooks, R.G.; Moss, J.; Hemberger, P.H. RapidCommun. Mass
Spectrom. 1989, 3, 50-53.
260. Kaiser, R.E.; Cooks, R.G.; Stafford, G.C.; Syka, J.E.P.; Hemberger, P.H. Int.
J. Mass Spectrom. Ion Processes 1991, 106, 79-115.
261. Booth, M. M.; Stephenson, Jr. J. L; Johnson, J. V.; Yost, R. A.
Fundamental and Practical Considerations of Mass Range Extension in a
Quadrupole Ion Trap Mass Spectrometer Presented to the Florida Section
of the American Chemical Society, April 3 May 2, 1992, Coco Beach, FL.
262. Chowdhury, S.K.; Katta, V.; Chait, B.T. J.Am. Chem. Soc. 1990, 112, 9012-
9013.
263. Loo, J.A.; Ogorazlek Loo, R.R.; Udseth, H.R.; Edmonds, C.G.; Smith, R.D.
Rapid Commun. Mass Spectrom. 1991, 5, 101-105.
264. McDonald, C.C.; Phillips, W.D.; Glickson, J.D. J. Am. Chem. Soc. 1971,93,
235-246.
265. Creighton, T.E. in Proteins, Structures and Molecular Principles W.H.
Freeman: New York, 1984.
266. Tanford, C. in Advances in Protein Chemistry Anfinsen, C.G.; Anson, M.L.;
Edsall, J.T.; Richards, R.M., eds.; Academic Press: New York, 1968, vol.
23, p.121.
267. Timasheff, S.N. Acc. Chem. Res. 1970, 3, 62.
268. Franks, R.; England, D. CRC Critical Rev. Biochem. 1975, 3, 165.
269. Herskovits, T.T.; Gadgefbeku, B.; Jaillet, H.T. J. Biol. Chem. 1970, 245,
2599.
270. Wilkinson, K.D.; Mayer, A.N. Arch. Biochem. Biophys. 1986, 250, 390-399.
271. Kishore, M.N.; Ghosh, P.K. Int. J. Mass spectrom. Ion Phys. 1979, 29, 345-
350.
272. Rajagopal, C.; Ghosh, P.K. Int. J. Mass Spectrom. Ion Phys. 1981, 41, 1-6.
273. Kaiser, R.E.; Cooks, R.G.; Syka, J.E.P.; Stafford, G.C. Rapid Commun.
Mass Spectrom. 1990, 4, 30.


Figure 4-22:
Mathieu stability diagram indicating the resonance ejection points for various degrees of mass
range extension (from reference 261).


56
control TTL pulse and the pulsed-valve TTL control pulse for a given scan
function.
Bis(2-methoxyethyl)ether (diglyme; ACS reagent grade purchased from
Fisher Scientific, Fairlawn, NJ), 3-bromo-1-propene (allyl bromide; ACS reagent
grade purchased from Aldrich, Milwaukee, Wl), 12-crown-4 ether, and 15-crown-5
ether (ACS reagent grade purchased from Sigma, St. Louis, MO) were introduced
via a Granville-Phillips (Boulder, CO) fine metering valve system directly into the
ion trap manifold. Depending on the experiment, the sample pressure ranged
from 1.1 to 3.6x1 O'7 torr. Formation of the protonated diglyme, 12-crown-4 ether,
or 15-crown-5 ether was accomplished by self chemical ionization, predominantly
due to the reaction of the low mass even-electron fragment ions from electron
ionization (El) with the neutral molecules for approximately 400 ms. The resulting
[M+H]+ ion of interest was then mass isolated by a two-step rf/dc isolation
routine.150 Next a 400 ms delay was included to allow for removal of any excess
internal energy by radiative and/or collisional cooling. The [M+H]+ ion was then
irradiated for a specified time period with the cw C02 laser. After the laser
irradiation period, a 10 ms time period was incorporated for decay of the laser
output when the high voltage was turned off. For all photodissociation efficiency
measurements, the first sequential product ion from the IRMPD of the [M+H]+
was resonantly ejected (q2 = 0.3, frequency = 118.1 kHz, amplitude = 100 mV,
values approximate) during the laser irradiation period (periods 5 and 6 in figure
2-3). This prevented the possible occurrence of any ion molecule reactions as the


Figure 5-16: Photodissociation spectra of the ammonium adduct of stachyose
for one, two, and three laser pulses. After two laser pulses The
entire sequence for stachyose was obtained. With the addition
of a third laser pulse, ring fragmentations (e.g. loss of H20
indicating the presence of the hydroxyl group) were observed for
the terminal monosaccharide residue (m/z 128 and m/z 146).


Figure 4-18:
Photodissociation set-up for the pulsed-laser. The beam selector was removed since the beam
width of the pulse laser exceeded that of the beam selector. Alignment was accomplished by
using a Helium-Neon laser and an additional gold plated turning mirror. A focusing lens was
added to make the beam diameter 0.2 cm as it enters the multipass ring electrode.


CHAPTER 6
CONCLUSIONS AND FUTURE WORK
As illustrated by the overall success of this project to date, the future for the
ESI/ion trap is very promising. Alternative ion activation techniques for structural
elucidation such as photodissociation show a great deal of promise in tackling the
difficult problems associated with analytical biochemistry. The major
accomplishment of this work is reflected not just in low detection limits, but also
in the application of innovative new ideas in order to push ion trap mass
spectrometry to its limits.
In order to accomplish the major goal of the project; photodissociation of
biological molecules, a very methodical course was taken. It began with
increasing the photoabsorption pathlength of the photodissociation process, to
overcome the problems (e.g. low absorption cross-sections for organic species,
and limited power and tunability of the C02 laser output) traditionally associated
with IRMPD. This increase in pathlength was accomplished by mounting three
spherically asymmetric concave mirrors in the radial plane of the ring electrode,
thereby producing a multipass cell. It was demonstrated that the rate coefficients
for various photodissociation processes were increased by a factor of 8 to 10 over
the traditional single-pass experiments. Other fundamental investigations included
the photon absorption process, IRMPD kinetics, the study of consecutive
399


100
80
60
40
20
0
la
m/z 158
,5A
la
m/z 100
/
0JA
m/z 116
\
la
m/z 174
(M+H)+
m/z 734
Zp or Y0p- h2o
m/z 558
Consecutive Water Losses
m/z 540 and m/z 522
Aglycone Ring Cleavages
Y
*op
m/z 576
(m-h2o>
m/z 716
+
too
200 300
400 500
m/z
600
700
800
395


51
incoming TTL signal to the CMOS logic used by the laser. A spectrum analyzer
(Optical Engineering, Model 16-A, Santa Rosa, CA) was placed directly in-line with
the cw C02 laser beam for wavelength measurements. All laser energy
measurements (Coherent Radiation Model 410 power meter, Boulder, CO) were
taken inside the ITMS vacuum manifold to correct for beam loss at the turning
mirror surface and ZnSe window. Because of space considerations, the ion trap
analyzer was removed from the ITMS vacuum manifold for the energy
measurements. The power meter was then positioned at the exact location where
the laser beam would enter the ion trap analyzer. The laser was placed parallel
to the ITMS manifold with the beam reflected at a 90 angle by a gold coated
mirror as seen in figure 2-1. Laser beam alignment was accomplished with a He-
Ne laser placed in-line with the cw C02 laser. The analyzer of the ITMS was fitted
with a modified mounting bracket to allow for rotation of the ion trap analyzer from
its original fixed position to facilitate alignment (figure 2-2).
The pulsed-valve used in the experiments (Series 9, General Valve
Corporation, Fairfield, NJ) was mounted on the opposite flange from the ZnSe
window (see figure 2-1) and was used to pulse He into the ion trap to increase
trapping efficiency. The pulsed valve was placed 0.5 cm from the outer diameter
of the ring electrode. Its horizontal position was between the modified ring
electrode and the entrance endcap. The pulsed-valve controller (built at the
University of Florida) was controlled by an external TTL pulse generated by the
ITMS electronics. The timing diagram shown in figure 2-3 displays both the laser


41
(m/z 117) has also been demonstrated.137'139 The authors were able to
differentiate the two isomeric m/z 177 ions by an aldol condensation reaction
observed with the fragmentation of protonated diacetone alcohol. This reaction
was not observed with the propanone dimer in the gas phase.
Investigation into the wavelength dependence of IRMPD photodissociation
efficiency of the proton-bound dimer of ethanol using the QUISTOR demonstrated
the observable frequency shifts of the C-0 stretch in the IR region.23140
Application of the quistor technique to the study of photodissociation rates by
varying relaxation time, buffer gas pressure, and analyte pressure has yielded
data for the proton-bound dimers of isopropanol, 2-dT-2-propanol, and ethanol.
The authors also studied the effect of collision rate on the defined
photodissociation cross section. The results obtained for the fully relaxed proton-
bound dimer population showed access to the lowest Ea pathway, thus
demonstrating the only variable observed was that of the collisional deactivation
process (corresponding to higher collision rates > 5 ms'1).23
The Photon Absorption Process
The interaction of gas-phase molecules with infrared light is currently one
of the most rapidly developing fields in chemical physics. Since the first
discovery of infrared photon absorption by ions in the gas phase, the field has
attracted the attention of researchers from a variety of disciplines.141'143 This
section is intended to give the reader a general overview of the mechanism of the


406
Gauge" British Patent 1,225,272, 1971.
15. Todd, J.F.J; Bonner, R.F.; Lawson, G. J. Chem. Soc. Chem. Commun.
1972, 1179-1180.
16. Lawson, G.; Bonner, R.F.; Mather, R.E.; Todd, J.F.J; March, R.E. J. Chem.
Soc., Faraday Trans. 1 1976, 73, 545-557.
17. Bonner, R.F.; Lawson, G.; Todd, J.F.J.; March, R.E. Adv. Mass Spectrom.
1974, 6, 377-384.
18. Fulford, J.E.; March, R.E. Int. J. Mass Spectrom. Ion Phys. 1978, 26, 155-
162.
19. Fulford, J.E.; Hoa, D.N.; Hughes, R.J.; March, R.E.; Bonner, R.F.; Wong,
G.J. J. Vac. Sci. Technol. 1980, 17, 829-835.
20. Hughes, R.J.; March, R.E.; Young, A.B. Int. J. Mass Spectrom. Ion Phys.
1982, 42, 255-263.
21. Hughes, R.J.; March, R.E.; Young, A.B. Can J. Chem. 1983, 61, 824-833.
22. Hughes, R.J.; March, R.E.; Young, A.B. Can. J. Chem. 1983, 61, 834-845.
23. March, R.E.; Hughes, R.J. in Quadrupole Storage Mass Spectrometry John
Wiley & Sons: New York, 1989.
24. Stafford, G.C.; Kelley, P.E.; Reynolds, W.E.; Syka, J.E.P. in Proceedings of
the 31st ASMS Conference on Mass Spectrometry and Allied Topics
Boston, MA, 1983, 48-49.
25. Stafford, G.C.; Kelley, P.E.; Bradford, D.C. Am. Lab. 1983, 51-57.
26. Stafford, G.C.; Kelley, P.E.; Syka, J.E.P.; Reynolds, W.E.; Todd, J.F.J. Int.
J. Mass Spectrom. Ion Processes 1984, 60, 85-98.
27. Kelley, P.E.; Stafford, G.C.; Syka, J.E.P.; Reynolds, W.E. Louris, J.N.; Amy,
J.W.; Todd, J.F.J. in Proceedings of the 33rd ASMS conference on Mass
Spectrometry and Allied Topics San Diego, CA, 1985, 707-708.
Kelley, P.E.; Syka, J.E.P.; Ceja, P.C.; Stafford, G.C.; Louris, J.N.;
Grutzmacher, H.F.; Kuck, D.; Todd, J.F.J. in Proceedings of the 34th ASMS
conference on Mass Spectrometry and Allied Topics Cincinnati, OH, 1986,
963-964.
28.


2400
1800
1200
600
0
m/z 221
m/z 45
100
Irradiance Time in ms
w


389
Compared with the spectra generated from the CID of oligonucleotides, the
types of ions observed in photodissociation spectra differed markedly.308 For CID
spectra, preferential cleavage occurs at the CO of the sugar (charge retention
of the 3 end) to form w-type fragment ions. Complementary cleavages to form
ions with a (a-B(A)) type fragmentation were also observed for oligonucleotides
containing adenine base(s). The photodissociation spectra presented in this
section show preferential cleavage at the PO, which has a fairly high
photoabsorption cross-section; these data correspond well with those obtained
using IRMPD in the FTICR.111
Carbohydrate Antibiotics
Carbohydrate antibiotics are some of the most important compounds in all
of biochemistry. In the next decade, and increasing emphasis will be placed on
discovery of new carbohydrate antibiotics, as microorganisms continue to develop
resistance to current antibiotic treatments. Over the years, one of the most
important classes of carbohydrate antibiotics has been those of macrolide
antibiotics.310 Macrolide antibiotics all contain a large lactone ring (aglycone of
12 to 22 atoms) with no nitrogen atoms and few double bonds. Linked to the
aglycone ring are one or more sugars which can be nitrogen-containing. These
sugar linkages to the aglycone ring are critical for biological activity. The
structural determination of the various functional groups comprising these
compounds and other carbohydrate antibiotics has long been of


188
octopole seems to follow the same relationship(s) as those of the rf-only
quadrupole.253,254 In this case, the fundamental wavelength of motion (4^) is
defined as:
lab
sec
f
(3-57)
sec
where vlab is the longitudinal velocity and is the secular frequency of motion.
The secular frequency of motion for a quadrupole is related to the fundamental
applied rf frequency by:
= f
'sec
q2
2y/2
(3-58)
with f defined as the applied rf frequency and q2 defined for a quadrupole as:
=
4eV
_ 22
mo) r0
(3-59)
Combining equations 3-57, 3-58, and 3-59 yields for the fundamental secular
wavelength:
'sec
= /mE
lab
4u2r02f
eV
(3-60)
here is derived from the standard relationship of kinetic energy and velocity
with v= v,ab.222'253'255 Since the square of the ions secular frequency is inversely
proportional to the value of q (equation 3-59), a significant reduction in rf drive
frequency can lead to unstable ion trajectories as seen in figure 3-14.253


Figure 4-13: Octopole mount assembly for interfacing the octopole with the ion trap analyzer. The cut-out slots
are for increased pumping in the analyzer region.


174


m/z
o
o
K)
O
4-
O
o\
O
00
o
Intensity
'M
O
O
O
o
o
'Jl
o
o
O
o
o
J I L
J 1L
I I I
J I L
S-S
s"
II
z
* j
o
li
2
+
o
i!
00
Irradian ce Time=120 ms
(M+NH4-3H,o-NHj)+ Energy=2.0 J
m/z 111
(M+NH4-2HjO-NH,)
m/z 129
Intensity
i
h-
K
K>
u>
o
Ul
O
o
o
O
o
O
o
o
o
o
o
o
> I I I I > 1 *I
+
355


4001
£
c
4-
s
114.0 115.0 116.0 117.0 118.0
Frequency in kHz
O)
Ol


86
An investigation to determine the influence of storage conditions (under the
influence of quadrupolar excitation and thus determine the presence of coupled
ion motion in the axial or z-direction while exciting the ions radially) was
undertaken using protonated 12-crown-4 ether. As mentioned previously, once
the ion trajectories exceed the 3 mm laser beam width in the axial or z-direction,
a marked drop off in photodissociation efficiency will occur, indicating the
presence of excitation in the axial direction. To minimize the effects of collisions
on photodissociation efficiency, the ion trap was again operated with no He buffer
gas. A detailed description of the instrumentation employed, limitations of
frequency measurements, and the procedures for data acquisition can be found
in the Experimental Design and Dipolar Excitation sections of this chapter.
To determine the appropriate quadrupolar excitation frequencies for the ion
motion studies, the [M+H]+ ion of protonated 12-crown-4 ether was stored and
analyzed as described previously in the Dipolar Excitation section of this chapter.
The frequency range probed was from 25 kHz to 500 kHz, with the frequency
incremented at intervals of 10 Hz. A plot of the intensity of m/z 177 of protonated
12-crown-4 ether versus the excitation frequency yielded a series of absorption
bands with the center frequency assigned as the minimum intensity value of m/z
177 for a given absorption band. For absorption bands which were offscale for
a 6 Vp_p quadrupolar excitation amplitude, the center frequency was taken as the
center point of the FWHM for the band.167


Table 2-1. Thermochemical data for the photodissociation of protonated diglyme.
Reaction
^^f(reactlon)
kJ/mol
^^((products)
kJ/mol
^^f(reactants)
kJ/mol
Photons
C6H1503+ + nhv -* C5H1102+ + CH3OH
37
C5H1102+ = 352*
CH3OH = -201b
C0H15O3+ = 114b
n ;> 3
C5H1102+ + nhv -*> C3H70+ + C2H40
68
C3H70+ = 586*
C2H40 = -166b
C6H1102+ = 352*
n 2r 6
a AHf values obtained by AM1 calculations references 167, 168.
b AHf values obtained from reference 166.
CO
OI


136
theorem, states that for any vector field around a closed path which gives a value
of zero (as in equation 3-12), the vector field may be represented as the
divergence of a scalar field with:
E = -V(x,y,z) (3'13)
The proof for equation 3-13 can be found in Friedman et al.232 Substituting
equation 3-13 into the first Maxwell equation of 3-9 gives:
V V<|>(x,y,z) =0 (3_14)
the V*V term is called the LaPlacian as is defined as V2. This transforms
equation 3-14 to:
V2(x,y,z) = 0 i315)
Putting Laplaces equation 3-15 into rectangular coordinates gives:
V2 = V V
8 ? a ,* 3 '
i +1 + k
k dX dy dz
(3-16)
- JL + Ji- + JL
ax2 + ay2 + az2
Since equation 3-15 is a result of the Maxwell relationships from equation 3-9, any
function of 0(x,y,z) which is a solution of equation 3-15 (with an E field defined
by equation 3-13) satisfies the requirements of Maxwells relationships in equation
3-9 232
The differential equation for ion motion for any configuration (e.g., any


Figure 3-7:
Schematic diagram of the teflon plug spacer, brass alignment jig, and brass alignment ring used
for proper alignment of the eight octopole rods during the spot welding process.


313
Table 4-8. Instrumental parameters for the MS/MS of the triply-charged m/z 433
ion of human angiotensin I.
Parameter
value
Qz
0.3
MS/MS Frequency
118 kHz
Resonance Excitation Amplitude
1.5 Vp.p
Resonance Excitation Time
20 ms
He Buffer Gas Pressure
1.0x1 O'4 torr (uncorr.)


Figure 1-3: Schematic representation of the energy terms relating to the ionization (or dissociation) of a
polyatomic molecule: l2, Ionization energy of the molecule P; Elh, thermal excitation of a molecule
P prior to ionization; E,fd, energy transferred to P by the incident electron, photon, or collision; E,
resulting internal energy in the ion; q1t reaction coordinate for P+ -* A+ + B; >c(, the activation
energy for P+ A+ + B; e^,, activation energy for the reverse reaction A+ + B -* P+; AH0. AH298,
standard enthalpy changes for P A+ + B + e at 0 K and 298 K, respectively; D0=AH0, standard
enthalpy change for the dissociation P -* A + B; lz(A), ionization energy of the fragment A. Only
two of the many vibrational degrees of freedom of P, P+, P+*, and A+ are represented; the potential
surfaces of A and B are omitted. The spacing of vibrational levels is greatly exaggerated relative
to D0 and eac{: \z is typically 2-4 times D0. The level (a) represents the energy of the dissociated
neutral system, A + B, with the species in their ground states. Most of these energy terms refer
to the differences in energy between specific levels; the exceptions are AH298 and AH0, which are
the usual thermodynamic quantities. Adapted from reference 72.


ZnSe Window Blank Flange
206


Figure 4-9:
(a) Scope trace from the frequency scan of the function generator in the SWR bridge test set-up.
The display range shown is from 1.520 MHz to 1.720 MHz. (b) Expanded view of the scope trace
from part (a). Channel 2 represents the marker frequency (TTL). The marker frequency indicates
the frequency of the function generator when the bottom of the sweep peak is aligned with the
marker TTL signal.


261
parameters for the spectrum are shown in table 4-2. The acquisition was for a
low-mass cut-off m/z 30 and a He bath gas pressure of 1.0x10^ torr. The
spectrum represents the average of 5 microscans. The PFTBA spectrum
obtained is typical of ion trap spectra in general, which show enhanced sensitivity
for the high mass end of the spectrum.
In the case of El operation, the ion gate (quadrupole entrance lens L3) is
at a negative voltage for pulsing ions into the octopole and at a high positive
voltage to prevent ions from entering the octopole. This is the opposite case for
electrospray operation (discussed below) where a positive voltage is used to gate
positive ions into the octopole and a high negative voltage is used to stop the
ions from entering the octopole. For the case of electrospray, ions are
undergoing a rapid jet expansion going from the liquid phase to the gas phase.
This means the dispersion of the ions is much greater. To effectively inject ions
as close as possible to the center axis of the octopole, the voltage must be
positive in order to focus the ion beam into the octopole. Since the El ionization
process occurs at sample pressures well below 1x1 O'6 torr, the ion beam is more
effectively extracted and focused from the ion source with a negative voltage as
opposed to a positive voltage where ions would strike the wall of the ion gate
tube lens.


Figure 2-14: IRMPD ion growth curves for protonated diglyme as a function of laser irradiance time. Two
reaction channels were observed: one for the formation of the m/z 103 fragment from the (M+H)+
ion and the other for the formation of the m/z 59 fragment from m/z 103. Some back reaction of
m/z 59 with neutral diglyme to reform the (m/z 135) (M+H)+ Ion was observed at longer irradiance
times. Error bars are defined as the standard deviation of the mean.


5 PHOTODISSOCIATION OF BIOLOGICALLY IMPORTANT
MOLECULES: PROTEINS, CARBOHYDRATES,
AND OLIGONUCLEOTIDES 328
General Overview of Structural Elucidation 328
Peptides and Proteins 338
Human Angiotensin I 338
Gramicidin D 342
Carbohydrates and Oligosaccharides 343
Monosaccharide Cleavage 348
Raffinose 363
Stachyose 374
Oligonucleotides 377
RNA Dimers 382
Carbohydrate Antibiotics 389
Macrolide Antibiotics Erythromycin 390
6 CONCLUSIONS AND FUTURE WORK 399
REFERENCE LIST 405
BIOGRAPHICAL SKETCH 427
vii


CHAPTER 1
INTRODUCTION
Over the last decade, perhaps the most important advancement in
quadrupole ion trap mass spectrometry has been that of tandem mass
spectrometry or MS" for structural elucidation of organic ions. The most
frequently used method for the activation of these ions has been collisional
activation, commonly known as collision-induced dissociation (CID).1 The major
factors that contribute to the success of CID experiments in the quadrupole ion
trap mass spectrometer (QITMS) include the ability to perform tandem-in-time as
opposed to tandem-in-space MS/MS experiments, the efficient conversion of
parent ions to product ions (typically 10-50%), and, most importantly, the high
collision cross sectional area observed for CID (on the order of 10 to 200 A2).1-3
These advantages arise in part because in the quadrupole ion trap, uniquely
amongst tandem mass spectrometers, kinetic energy is imparted to the parent
ions only between collisions.
To date, the majority of the published applications using the QITMS for
structural elucidation has employed collisional-activation as the method of choice
for ion activation.4,5 More recently, the attention of many researchers has focused
on the fundamental understanding of the collisional-activation process in the
QITMS.6 Even though collisional-activation cannot provide all the answers in
1


BIOGRAPHICAL SKETCH
James L. Stephenson, Jr. was bom in Richmond, Virginia, on October 18,
1961, to James L. and Vivian P. Stephenson. He spent his childhood building
tree forts, playing football, and riding/wrecking a wide variety of bicycles. In 1980,
he graduated from Hermitage High School and went on to attend college at East
Carolina University in Greenville, North Carolina. While attending East Carolina
University, he was active in Phi Sigma Pi Fraternity and spent an excess amount
of time taking math courses (as evidenced by chapter 3 of this dissertation). He
was awarded the Claude Pennock Todd Fellowship his senior year and graduated
cum laude in 1984 with a B.S. in biochemistry.
His first real job (i.e., not a lifeguard) was as an analytical chemist for
California Analytical Laboratories in Richmond, Virginia. It was here that his
interest in mass spectrometry was kindled. In the fall of 1985, he met his future
wife, Tracy A. Reitz of Alexandria, Virginia; the two were married in October 1987.
As his interest in mass spectrometry grew, he went to work as an Field Engineer
for Finnigan MAT in Washington, D.C. During this period, he was fortunate
enough to work with Dr. Henry Fates of the National Institutes of Health, where he
learned a great deal about mass spectrometry and life in general.
Jim and Tracy moved to San Jose, California, in 1988 due to Jims
427


221
correct tune for the rf circuit. Once a coarse frequency reading of 1.6 MHz was
obtained (see figure 4-9a), the function generator was reset to scan from 1.6 to
1.7 MHz. The marker frequency was then adjusted to match the lowest point on
the SWR trace which indicated an optimum operating frequency of 1.659 MHz
(figure 4-9b).
To run the rf-only octopole at the highest frequency possible (for reasons
discussed in the previous chapter), the taps of the rf-coils were moved in 18 turns
on the 32 turn coils. This significantly increased the optimum rf frequency from
1.2 MHz to the 1.659 MHz value obtained above. The frequency increase can be
explained by the inverse relationship between the applied rf and the inductance
of the coil (e.g., a reduction of the coil inductance leads to an increase in the
applied rf frequency). Although this modification significantly reduces the gain of
the rf, the voltage requirements needed for operation in the rf-only mode are
significantly reduced compared to mass filter operation. A plot of the detected
rf versus the measured rf output (figure 4-10) shows a maximum rf amplitude of
approximately 2000 V^p. The rf output was measured with a standard scope
probe which has a significant amount of capacitance that will affect the tuning of
the rf circuit; therefore the 2000 V^p output is only an approximation of V^.
Another more complex solution to increase the frequency is to either adjust the
capacitance of the octopole device, or the matching capacitors or the rf circuit,
since there exists an inverse relationship between the applied rf frequency and
capacitance.


Dissociation Efficiency
Ionization Time in ms
334


332
fragmentation. In addition, several tuning parameters including helium buffer gas
pressure, ion frequency, and ion population, which further complicate the single
frequency CID experiment even for an experienced user are not concerns in PID
experiments. Even when broadband excitation techniques are employed to
minimize ion frequency and ion population effects, dissociation efficiencies
observed are somewhat lower than in the corresponding single-frequency
experiments, and can be drastically lower than in photodissociation experiments
when a sufficient photoabsorption cross-section exists for the ions of interest.
The effect of ion population on the dissociation efficiency of protonated 12-
crown-4 ether produced by Cl is shown in figure 5-1. Ionization time (shown on
the x-axis) refers directly to stored ion population in the ion trap. For the case of
the single-frequency CID experiment (where the tuning parameters were optimized
for a low number of ions stored in the ion trap), a significant decrease in
photodissociation efficiency was observed as the ionization time and hence the
ion population increased. Typically, an increase in ion population results in a shift
of the fundamental frequency of ion motion to lower frequency, thus resulting in
a decrease in dissociation efficiency for the given CID tune parameters (see figure
5-1 ).293 In the case of broadband excitation, where a whole range of frequencies
are excited over a given time period, the dissociation efficiency is independent of
ion population. However, a decrease is observed in dissociation efficiency
compared to that of the optimum single frequency results.


32
The presence of a dry nebulization gas at approximately 80 C and the use
of a heated capillary (counter electrode) can aid significantly in the desolvation
process of the charged droplet. The formation of multiply charged ions from the
charged droplet is an area of great debate. The droplets eventually reach a point
where the repulsive coulombic forces approach those of the cohesive forces
(surface tension) that hold the droplet together. At this point the droplet may form
an ion by one of two proposed mechanisms: droplet fission at the Raleigh limit
or direct field evaporation of the droplet.116,117 A discussion of the relevant
thermodynamics of these two processes is beyond the scope of this dissertation.
However, it is the belief of this author that the direct evaporation theory as set
forth by Iribarne and Thomson applies to most situations.118,119
Once the charged droplets/ions pass through the counter electrode region,
they proceed through a differentially pumped region containing one or two
skimmer cones. A schematic diagram of the typical electrospray source
developed by Fenn and coworkers can be seen in figure 1-5, showing the
electrospray needle, counter electrode, differentially pumped region, skimmer
cones, and dc lens injection system.120 Details of the various ion injection
systems used are discussed in chapters 3 and 4, along with details of a new ion
injection system (for quadrupole ion traps) for electrosprayed ions using an rf-only
octopole beam guide.


374
instrument parameters is needed to induce fragmentation (only an initial time
investment at the beginning of the days analysis).
Stachvose
The next compound studied was the tetrasaccharide stachyose (see figure
5-10 for structure). Stachyose (a-D-galactopyranosyl-[1 -6]-or-D-galactopyranosyl-
[1 -6]-cr-D-glucopyranosyl-[1 -2]-/?-D-fructofuranoside) differs from raffinose by the
one additional D-galactose on the nonreducing end and a [1-2] /? linkage from
glucose to fructofuranoside. Electrospray and instrumental conditions were the
same as for raffinose in the previous section. The sample concentration was 20
pmol///L infused directly through the ESI source.
The single-frequency CID spectrum of the ammonium adduct of stachyose
(m/z 684) is shown in figure 5-14. The major fragment peaks in the spectrum
were observed at m/z 505 and m/z 488 (B^ or ZgJ, indicating the loss of the
nonreducing galactosyl terminal or the fructosyl residue which has its reducing
end bonded /?-[1-2] to the glucopyranosyl moiety. The peak at m/z 488 shows
the same cleavage with loss of ammonia. The peaks observed at m/z 343 and
m/z 326 indicate the loss of an internal sugar residue, either D-galactose or D-
glucose (B^ or ZJ. The small peak at approximately m/z 589 may indicate a
ring-opening reaction which cleaves in the 3 and 5 position of D-galactose to form
the 3,5X3a ion or cleaves in the 2 and 4 position of D-fructose to produce the 2,4A3a
ion. Isotopic labeling experiments would be required for direct mass assignment
of this ion. The peak at m/z 680 was an occasional artifact peak observed, and


411
88. Yamashita, M.; Fenn, J.B. J. Phys. Chem. 1984, 88, 4451-4459.
89. Aleksandrov, M.L; Gall, L.N.; Krasnov, V.N.; Nikolaev, V.I.; Pavlenko, V.A.;
Shkurov, V.A. Dokl. Akad. Nauk. SSSR 1984, 277, 379-383.
90. Aleksandrov, M.L; Gall, L.N.; Krasnov, V.N.; Nikolaev, V.I.; Pavlenko, V.A;
Shkurov, V.A.; Baram, G.I.; Gracher, M.A.; Knorre, V.D.; Kusner, Y.S.
Bioorg. Kim. 1985, 11, 705-708.
91. Meng, C.K.; Mann, M.; Fenn, J.B. Phys. D. 1988, 10, 361-368.
92. Loo, J.A.; Udseth, H.R.; Smith, R.D. Biomed. Environ. Mass Spectrom.
1988, 17, 411-414.
93. Covey, T.R.; Bonner, R.F.; Shushan, B.I.; Henion, J.D. Rapid Comm. Mass
Spectrom. 1988, 2, 249-256.
94. Cox, A.L; Skipper, J.; Chen, Y.; Henderson, R.A.; Darrow, T.L;
Shabanowitz, J.; Englehard, V.H.; Hunt, D.F.; Slingluff, C.L Science 1994,
264, 716-719.
95. Hunt, D.F.; Michel, H.; Dickerson, T.A.; Shabanowitz, J.; Cox, A.L;
Sakaguchi, K.; Appella, E.; Grey, H.M.; Sette, A. Science 1992, 256, 1817-
1820.
96. Hunt, D.F.; Henderson, R.A.; Shabanowitz, J.; Sakaguchi, K.; Michel, H.;
Sevilir, N.; Cox, A.L; Appella, E.; Englehard, V.H. Science 1992,255,1261-
1263.
97. Henderson, R.A.; Michel, H.; Sakaguchi, K.; Shabanowitz, J.; Appella, E.;
Hunt, D.F.; Englehard, V.H. Science 1992, 255, 1264-1266.
98. Ikonomou, M.G.; Blades, A.T. Kebarle, P.AnaJ. Chem. 1991,63,1989-1998.
99. Kebarle, P.; Tang, L. Anal. Chem. 1993, 65, 972A-986A.
100. Fenn, J.B. J. Am. Soc. Mass Spectrom. 1993, 4, 524-535.
101. Chorush, P.A.; Little, D.D.; Beu, S.C.; Wood, T.D.; McLafferty, F.W. Anal.
Chem. 1995, 67, 1042-1046.
102. Jones, J.C.; Dongue, A.R.; Sonogyi, A.; Wysocki, V.H. J. Am. Chem. Soc.
1994, 116, 8368-8369.


Figure 3-13: (a) RF voltage dependence (V=100 and 150 V) in a quadrupole cell, q2 values are 0.43 and 0.65
respectively, m/z 100, ion entry 1 mm/3, 1.2 MHz. (b) RF voltage dependence (V=100 and 150
V) in an octopole cell, q4 values are 3.46 and 5.20 respectively. Remaining conditions as in (a).
Adapted from reference 253.


Figure 3-11: (a) Ion displacement dependence (at 1, 2, and 3 mm) in an rf-only quadrupole for m/z 1000 with
Vop=400 V and an ion entry angle of 0. (b) Ion displacement dependence (at 1, 2, and 3 mm)
in an rf-only octopole, conditions same as in (a). Adapted from reference 253.


97
f0r the CBr stretch (as compared to the COC stretch) observed in the gas-
phase IR spectroscopy.
IRMPD Kinetics
As mentioned previously, the photodissociation efficiency (PD) for a given
experiment is defined as the fraction of the original ion population
photodissociated over a given exposure time for a specified laser irradiance:
PD=1-
' n
*0,
(2-2)
where I is the signal intensity of the dissociating ion at the end of the exposure
L f'
period and l0 is the signal intensity after the same period without irradiation. This
definition of l0 corrects for any unimolecular or collision-induced dissociation that
may occur. The photodissociation yield observed (PD) is generally dependent on
the wavelength of the laser, while the observation of a particular reaction route is
independent of laser wavelength. From equation 2-2, a first order relationship for
IRMPD kinetics (equation 2-8) was obtained:
In (l/l0) = -kDt (2-8)
where kD is defined as the photodissociation rate coefficient.130 By plotting ln(l/l0)
versus t (figure 2-12) the rate coefficient obtained for the diglyme [M+H]+ ion was
kD = 97.2 1.9 s1.146,147 This value was significantly higher than the 2-30 s'1
values obtained at higher laser irradiances reported in the literature.23,129,130
Clearly the multipass ring electrode enhances photodissociation efficiency to a


35
Molecular Weight Determination
Molecular weight data from ESI spectra can easily be obtained due to the
charge state distribution associated with the ionization process. Typically, the
width of the charge state distribution is approximately half that of the highest
charge state, although the effects of the various factors (e.g., pH, applied
potential) are not yet well understood.121,122 The adjacent peaks (of the multiply
charged ion distribution) in the spectrum of positively charged biopolymers
usually vary by one charge. Therefore, in order to determine molecular weight
(Mr) from an ESI spectrum (where the charge varies on adjacent peaks by the
addition or subtraction of one proton), the following expressions are used:
p^-M, + Maz1 = Mr + 1.0079z1 (1-14)
where p, is the m/z of interest and z, is the charge on p1t and Ma is assumed to
be the charge carrying species (proton). By examining another peak in the
charge distribution spectrum, another equation can be generated for a m/z higher
than the previous example (p2>p1) that is j peaks away from pr
p2(z1-j) = Mr+1.0079(z1-j)
Equations (1-14) and (1-15) can then be solved for p1 yielding:
_ j(p2-1.0079)
1 (P2-P1)
(1-15)
(1-16)
The value of the molecular weight is then calculated by evaluating zn to the
nearest whole number.91,123 Improved precision can be obtained by performing


Figure 2-19:
Photodissociation efficiency as a function of laser wavelength for protonated diglyme at an energy
of 0.252 J (1.1x107 torr diglyme). Error bars are defined as the standard deviation of the mean.


197
transmission properties of this unique arrangement. Next, a short review of the
pertinent design considerations for rf-only octopoles (from chapter 3) is presented
along with the exact specifications for operation of the rf-only octopole used in
this instrument. The final parts of this section cover the design and construction
of the analyzer and detector assemblies.
Vacuum Manifold and Pumping System
A Finnigan MAT ITMS frame was used as a base for construction of the
vacuum manifold for the ESI/ion trap. The standard Finnigan MAT manifold and
wooden table top were removed from the frame assembly to allow for placement
of the new vacuum manifold. Three new table top sections and one side panel
were machined from 0.5" thick aluminum panels to construct the new frame
assembly. A series of 1" square aluminum rods were then used to elevate and
allow level placement of the vacuum manifold on the instrument frame. This level
placement (parallel to the floor within 0.02") was necessary to help eliminate
any laser alignment problems associated with future photodissociation
experiments. The table top panel used for mounting the vacuum manifold was
constructed with three precision-machined holes/slots to accommodate two
turbomolecular pumps and allow for placement of a Conflat adapter flange
needed for the electrospray interface. A schematic diagram of the table top
design is shown in figure 4-1.


137
number of pole pieces) can now be found. Thus, for any ion moving in a field
with n conductors and constant voltage 0¡(¡=1,2,3....n), the equation for the
surface of the ith conductor is z=f¡(x,y). The equation of motion 0(x,y,z) is then
found by satisfying LaPlaces equation 3-15 and meeting the boundary conditions:
(x,y,f (x,y)) = ¡ (i = 1,2,3, n) 3-17
Therefore, the potential applied to any of the electrodes (pole pieces) at their
surfaces must equal the defined potential function.
Once the potential equation is known, Newtons law of motion can be used
to find the differential equation of motion by:
F = ^ = -e V<}>(x,y,z) (3-18)
dr2
the method is approximately valid for time varying potentials giving:
4>¡ = 4*DC ~ (3-19)
which is subject to the limiting conditions imposed from equation 3-10, where the
applied potential does not change too rapidly for the length of the multipole
pieces. From equation 3-19, rods and 0AC refers to the amplitude of the time-varying potential of frequency ()
applied to the same multipole rods. Equation 3-19 can also be rewritten into the
more familiar form where for the applied potential 0O:
4>0 = U V cos where U is the applied DC voltage and V is the amplitude of the time-varying


Figure 3-1: The electrode structure of the ideal octopole (n=4). The contour lines of the hyperbola-like cross-
sections of the electrodes are given by a polynomial of the 4th order. The distribution of the
electrode potentials for the octopole electrode system is also shown. Adapted from reference 219.


229
signal, electron multiplier voltage, exit tube lens, applied (e.g., quadrupolar or
dipolar) excitation signal (2) and pulsed-valve control (2). Four 0.5" diameter 4"
long brass rods were attached to the non-vacuum side of the 8" Conflat flange to
allow for proper clearance of the aforementioned feedthroughs when the analyzer
assembly was set upright on a lab bench for maintenance.
Two aluminum mounting plates were used to contain the ion trap analyzer,
exit tube lens and octopole mounting assembly. The mounting plates were drilled
out to accommodate the optical rail assembly and to pass the electrical
connections to the endcap electrodes and exit tube lens. The three holes drilled
for the ion trap analyzer were offset 13.3 from center, so that alignment of the
entrance aperture of the multipass ring electrode was parallel to the incoming
laser irradiation. This condition was achieved when the 8" Conflat was turned
such that (from a top view) opposing pairs of rods were in line with the plane of
the table top. The position of the mounting plates, and therefore the analyzer
assembly, was fixed by two set screws on each plate. A schematic diagram of
the aluminum mounting plate can be seen in figure 4-12.
The ion trap analyzer was from a standard Finnigan MAT ITMS, with the
entrance endcap replaced with an exit endcap. The two endcaps were positioned
such that the seven holes on both exit endcaps were symmetrically aligned. In
principle, this alignment will reduce the hexapolar non-linear field contributions to
the quadrupolar trapping field. In addition, the use of an exit endcap for an
entrance endcap allows placement of the rf-only octopole to within 0.050" of the


C \
Scale
vertical 50 mV/div
horizontal 20 kHz/div
v /
1.520
units of MHz
1.720
(b)
/ \
Scale
vertical 50 mV/div
horizontal 5 kHz/div
\ J
1.640
units of MHz
1.690
223


298
a2 c2 8W
Or + 6, =
r 2 mQ2
(4-8)
where r and z represent the radial and axial displacements, m is the mass of the
ion, Q is the angular frequency of the rf drive, and q, a, and 6 are trapping
parameters. The maximum kinetic energy any ion can have is defined by the
physical constraints of the device where r0 and are the maximum amplitudes
of the radial and axial displacements, respectively. Assuming that ion motion
during the injection process is limited to the z-direction, with 6z=(qz/21/2), then qz
can be defined as:
q2
4 W
z0Q \ m
(4-9)
As the ions undergo collisions, energy is redistributed in the r-direction; when
sufficient energy is removed, equation 4-9 is satisfied. Thus, for low qz values, the
minimum rf level is inversely proportional to the square root of the mass.11,192 In
table 4-7 the minimum rf level, and threshold qz value (both obtained
experimentally from figure 4-29), and the theoretically calculated qz value
(obtained by using an arbitrary constant) are shown. As seen from table 4-7, the
experimental qz values agreed well with the theoretically calculated values.
To obtain a representative ESI spectrum (based on the aforementioned
arguments) of the compound of interest, it may be necessary to increment the
minimum rf level during the ion injection period. To fully understand this concept,
ESI spectra were taken at rf cut-off levels of m/z 25 and m/z 38 respectively (see


Figure 5-6: The MS2 and MS3 IRMPD spectra of the ammonium adduct of 2-
deoxy-D-glucose. In the first experiment the [M+NH4+] ion at
m/z 182 was mass isolated and fragmented at a q of 0.3. For
the MS3 experiment the first sequential product ion from the MS2
process at m/z 164 was mass isolated and fragmented at a q of
0.3. A significant reduction in CID efficiency marked the MS3
spectrum.


Figure 2-17:
Effect of collisions on the photodissociation rate of protonated diglyme using N, Ar, He, and neutral
diglyme as target gases. All pressure measurements are corrected values. Error bars are defined
as the standard deviation of the mean.


161
case of the square well. This means ions of differing m/z ratios can be
transmitted to the ion trap at a constant translational energy (due to the large flat
bottom" portion of the well, where ion kinetic energies have only small
perturbations), which yields higher transmission and trapping efficiencies. For the
quadrupole and hexapole, the radial potential is more triangular in shape and,
therefore, not as many ions can be transferred at constant kinetic energy thus
leading to a decrease in ion injection efficiency. This phenomena can be
explained by the dependence of the radial potential on the normalized ion
displacement (r/r0)2"'2. The effective radial potential in an octopole is proportional
to r6, which provides a large trapping volume for the ions of interest.221-227 In the
case of the other extreme where n=2 for the quadrupole, the radial potential is
proportional to r2 with a maximum trapping energy one-fourth that of the octopole.
Other advantages of the rf-only octopole include the ability to shield low energy
ions from stray electric fields (or contact potentials) and reduced effects from
space charging.221-227
Assembly and Construction
A unique method for the assembly and construction of the rf-only octopole
was developed by Joe Shalosky and the author at the University of Florida. The
four solid support pieces for the octopole (see figure 3-5) were made from
Delrin. Delrin was used because the material can be threaded and tapped (e.g.,
machinable) and the surface friction generated by Delrin as it rubs against


2000
1500
1000
500
0
Asp-Arg-Val-Tyr-IIe-His-Pro-Phe-His (b*2 fragment)
I I 1 I I I l l I | I I I l| 1 I I I | l l I I | I I I I | I l M |
500 600 700 800 900 1000 1100
m/z
324


243
and (7) aluminum mounting plate (top). The assembly was held together by three
stainless steel rods which were bolted through the two aluminum mounting plates.
Aluminum was used in the design of the mounting plates and octopole mounts
in order to reduce the strain (e.g., from weight) on the optical rail system. All
aluminum pieces were anodized with a 0.002" natural hardcoat finish. The
aluminum was anodized for several reasons including: (1) to ensure that all
support pieces would become non-conductors; (2) to reduce pump-down time
associated with trapped water and oxygen found on porous pure aluminum
solids; and (3) to increase the ease of movement of the analyzer assembly
against the stainless steel optical rail. A complete diagram of the analyzer
assembly is shown in figure 4-15.
The final piece of the analyzer assembly was an anodized aluminum
centering ring which mates directly with the centering ring in the vacuum
manifold. This allows for proper alignment of the ESI-octopole-analyzer assembly
of the instrument and provides additional mechanical support for the optical rail
system (see figure 4-15).
Detector Assembly
The detector used in the instrument consisted of a 20 kV off-axis dynode
and a continuous electron multiplier. This detector assembly was purchased
directly from Finnigan MAT (San Jose, CA) and modified to fit the ESI/ion trap
instrument. A stainless steel mounting bracket was used to hold the rear base


114
as described previously. The results obtained gave a series of irreversible
consecutive reactions of the form:
[C10H21O5r [C6H1303]+ [C4H902]+ ^ [C2H50]*
m/z 221 m/z 133 m/z 89 m/z 45 (2_21)
The sequential reaction equations are derived as before, applying the boundary
conditions at t=0 mentioned above. The rate equations for the sequential four-
stage reaction are seen below:179'181
-dlC^O-ij.'. = k,(C10H2tO5l* (2-22)
dJCeH^OjT = ki(CioH2io5]* k2[CH,303]* (2-23)
dt
- d [C4H302]* = k2[CeH,303]* k3[C4H9o2]* (2-24)
d|Cy k3[C4H902]- (2-25)
By multiplying through using the appropriate integrating factor(s) (for equation 2-
23, ux=ek(2)t, and for equation 2-24, ux=ek(3)t) to form exact differentials,
subsequent integration/substitution (as described previously) gives:
[Ci0H2iO5]- = [CmH^Ojlie"1''1
(2-26)


Unit vectors and components of the electric field strength in rectangular and polar coordinates.
UK1 Uy, Ex, and Ey are unit vectors and transposed vectors in the rectangular coordinate system
respectively. U* Url E*, and Er are the angle of rotation and displacement from center vectors in
the polar coordinate system respectively. Adapted from reference 219.
Figure 3-3:


Photodissociation Efficiency l-(I/Io)
Laser Delay Time in s


119
gases.185-189 Traditionally, collisional deactivation efficiency increases with mass
and polarizability of the neutral molecule. For collisions involving He and Ar, the
loss of vibrational energy from the protonated diglyme can occur only by a
vibrational to translational energy transfer. For the N2 collision partner, vibrational
energy transfer is possible but not realistic due to the large difference in
vibrational frequencies. In all probability, the greater collisional quenching
efficiency of N2 is due to its greater polarizability as compared to that of Ar. The
collisional deactivation process was most efficient when the protonated diglyme
molecules collided with the diglyme neutral. In this case, proton transfer (i.e.
symmetric charge transfer) occurs, and the close match between the vibrational
frequencies of the colliding ion and molecule facilitates intermolecular vibrational
energy transfer.130
Pulsed valve experiments were conducted to determine if the increase in
trapping efficiency during ionization associated with the addition of He buffer gas
to the ion trap analyzer would interfere with the collision-free requirements
subsequently needed for IRMPD. A 1.6 ms pulse of He gas (10 psi He back
pressure) was found to trap the maximum number of diglyme (M+H)+ ions. The
signal intensity of protonated diglyme (m/z 135) by using a 1.6 ms pulse of He
gas during the pre-ionization period was higher by a factor of seven than when
He was not used in conjunction with the ionization event. The ionization time for
both experiments was 1.0 ms.146,147


153
corresponding stability diagram is now dependent on the initial conditions of ion
motion.
Equations 3-48 and 3-49 can be converted into their dimensionless form
by substituting equation 3-40 for the x, y, and r values and subsequently
transforming the time variable to a dimensionless quantity. The conversion of the
equations of motion to this form means that all numerical calculations relating to
the octopole device are in dimensionless units. The advantages of the
dimensionless equation of motions are twofold: (1) the constants and variables
of the octopole will lie in a somewhat more narrow numerical range compared to
their true values; and (2) keeping track of exact units and numerical values of the
system parameters (which can be quite complex especially for the case involving
coupled ion motion observed with the octopole) is simplified.219
In order to understand the restoring force generated by the octopole
electrode arrangement, the equation of motion in the complex form must be
generated. Each of the two dimensional vectors involved in ion motion (position
vector(s), electric field vectors, velocity vector, and acceleration vector) are
functions of the real time variable t in the complex xy plane. Taking this into
account, the equation of motion in the complex form is:
Iff* Jt[U-Veo..(1-t0)](Z)-,.0 (3-52)
dt mr0
where Z is the complex conjugate of Z which has the following definition:


74
To determine the appropriate dipolar excitation frequencies for the ion
motion studies, the (M+H)+ ion of protonated diglyme was stored for a period of
10 ms at a qz value of 0.3. During the storage period, a dipolar excitation signal
was applied to the endcaps with an amplitude of 6 Vp_p. The frequency range
probed was from 25 kHz to 500 kHz, with the frequency incremented at intervals
of 10 Hz. A plot of the intensity of m/z 135 of protonated diglyme versus the
excitation frequency yielded a series of absorption bands with the center
frequency assigned as the minimum intensity value of m/z 135 for a given
absorption band. For absorption bands which were offscale for a 6 VP.P dipolar
excitation amplitude, the center frequency was taken as the center point of the
FWHM for the band. To assign the ions frequency from an absorption band as
accurately as possible, several instrumental parameters must be strictly controlled.
These include varying only one instrumental parameter per scan table and
allowing appropriate stabilization times for changes in the various instrumental
parameters (rf voltages, dc voltages, dipolar frequency, etc.). A detailed
description of all the factors involved has been published by Eades et al.162-164
Typically, the minimum value of frequency optimization curve is taken as
a close approximation of an ions component frequency of motion. As previously
mentioned, shifts in the observed frequency due to the instrumental variations
mentioned above can cause inaccuracies in proper frequency assignment.
However, other experimental factors can also contribute to inaccuracies in
frequency assignment for a given component of ion motion. These factors


Photodissociation Efficiency (1-I/Iq)
Quadrupolar Resonant Excitation Voltage in mV
CD


Figure 4-19:System interconnect diagram for the ESI/ion trap system. All electrical connections are shown.


Figure 5-4: IRMPD of the singly charged [M+H]+ ion of gramicidin D. The [M+2H]+ ion is present for mass
calibration purposes. The two peaks just above the [M+H]+ ion represent the sodium and
potassium adducts of gramicidin D respectively. The peak at [M-61]+ represents cleavage of the
modified carboxy terminus with subsequent hydrogen ion migration.


94
checked by comparing the value obtained in the literature (114 kJ/mol) for
protonated diglyme (m/z 135) against that of the AM1 routine (100 kJ/mol). A
summary of the AHf values for the two consecutive reactions of diglyme is given
in table 2-1. The errors observed in the tabulated values (e.g., AHf C5H1102+ and
AHf C3H70+) were on the order of 15 kJ/mol as determined by Katritzky et al. for
14 different cationic species by using the AM1 calculation.1'3 The results
observed for the reaction channels were n > 3 for the lower energy process and
n > 6 for the higher energy process as shown in table 2-1. The total number of
photons absorbed by the C-0 stretches in protonated diglyme was n > 9 for
formation of the m/z 59 fragment. The calculated photon energy at 944 cm"1 was
0.117 eV.
For the case of allyl bromide, only a one-step reaction for the photon
absorption process is involved (at 944 cm"1), as shown in equation 2-6:
[C3H5Br]+ + nhv -> [C3H5]+ + Br (2-6)
The AHf values for neutral bromine, protonated allyl bromide, and the allyl cation
were obtained directly from the literature.174 The result observed for the single
reaction channel was n > 5 for the formation of the allyl cation, as shown in table
2-2. The larger number of photons needed to reach the dissociation threshold for
the single allyl bromide reaction channel compared to that of the lower energy
channel for protonated diglyme was confirmed experimentally by the longer
induction periods associated with the allyl bromide reaction channel. These
results also agreed well with the smaller photoabsorption cross-sections expected


413
120. Fenn, J.B. Anal. Chem. 1985, 57, 675-679.
121. Loo, J.A.; Udseth, H.R.; Smith, R.D. Anal. Biochem. 1989, 179, 404-412.
122. Guevremont, R.; Siu, K.W.M.; LeBlanc, J.C.Y.; Berman, S.S. J. Am. Soc.
Mass Spectrom. 1990, 1,
123. Mann, M.; Meng, C.K.; Fenn, J.B. Anal. Chem. 1989, 61, 1702-1708.
124. Beu, S.C.; Senko, M.W.; Quinn, J.P.; Wampler, F.M. Ill; McLafferty, F.W. J.
Am. Soc. Mass Spectrom. 1993, 4, 557-565.
125. Hagen, J.J; Monning, C.A. Anal. Chem. 1994, 66, 1877-1883.
126. van der Hart, W.J. Mass spectrom. Reviews 1989, 8, 237-268.
127. van der Hart, W.J. Int. J. Mass Spectrom. Ion Processes 1991, 118/119,
617-633.
128. Dunbar, R.C.; Weddle, G.H. J. Phys. Chem. 1988, 92, 5706.
129. Bomse, D.S.; Woodin, R.L.; Beauchamp, J.L. J. Am. Chem. Soc. 1979,101,
5503-5512.
130. Thorne, L.R.; Beauchamp, J.L. in Gas Phase Ion Chemistry, Vol. 3: Ions
and Light Bowers, M.T., ed.; Academic Press: London, 1984, pp 42-95.
131. Rosenfeld, R.N.; Jasinski, J.M.; Brauman, J.l.J.Am. Chem. Soc. 1979,101,
5503-5512.
132. Turnas, W.; Foster, R.F.; Brauman, J.l. Isr. J. Chem. 1984, 24, 223-231.
133. Drzaic, P.S.; Marks, J.; Brauman, J.l. in Gas Phase Ion Chemistry, Vol. 3:
Ions and Light Bowers, M.T., ed.; Academic Press: London, 1984, pp 168-
207.
134. Mead, R.D.; Lineberger, W.C.; in Gas Phase ion Chemistry, Vol. 3: Ions
and Light: Bowers, M.T., ed.; Academic Press: London, 1984, pp 214-246.
135. March, R.E.; Kamar, A.; Young, A.B. in Advances in Mass Spectrometry
1985: Proc. 10th Int. Mass Spectrom. Confer. Wales, 1986, 949-950.
136. Kamar, A. Application of the Quadrupole Ion Storage Trap (QUISTOR) to the
study of Gas Phase lon/Molecule Reactions Ph.D. Dissertation, Queens


Figure 3-14: RF frequency dependence (w=0.6, 1.2, and 2.4 MHz) in an rf-only octopole; q4 values are 13.9,
3.47, and 0.87 respectively. Ion entry conditions are for m/z 100 at 1 mm off axis and a 1 angle,
rf voltage was 100 V0P. Adapted from reference 253.


Figure 4-35:
Mechanism for formation of internal y type ions from peptides which have proline within the peptide
backbone (e.g. not located on a terminus).


383
d for charge retention on the 5 end and w, x, y, and z for charge retention on the
3 end of the oligonucleotide. Subscripts indicate the number of bases contained
in the fragment. Fragmentation of the nucleoside bases is represented by an
upper case B, where B is the individual base A, G, C, T, or U. A subscript is
assigned to the B symbol to represent the position of the base from the 5 end
of the molecule. Bases are represented parenthetically to avoid confusion with
the normal sequence terms (e.g., B^A)).
The merits of photodissociation for DNA/RNA sequence analysis are
discussed in this section. Two sets of RNA dimers were studied in order to
determine the feasibility of the technique (pulsed-C02 laser photodissociation) for
future sequencing experiments.
RNA Dimers
Two RNA dimers, adenyl adenosine (ApA) and adenyl cytidine (ApC), were
used for the photodissociation experiments. The samples were obtained from the
Core Biotechnology Facility at the University of Florida, and were purified to
remove sodium salts. Sample concentrations were 2 pmol///L in a 70:30
methanol:water solution. Electrospray conditions and instrumental parameters
were set as described in chapter 4 of this dissertation. The photodissociation set
up employed the pulsed laser described in chapter 4.
The lone parent species produced for the negative ion electrospray for the
two dimers was a singly-charged negative ion (M-H)' at m/z 595 for ApA and m/z


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.
-f
Richard A. Yosl/Chairman
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.
John R. Eyler j
Pressor 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.
(/ Professor of Microbiology and Cell
Science
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.
Robert T. Kennedy
Assistant 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.
David H. Powell
Associate Scientist of Chemistry


Figure 3-8:
Schematic diagram of the complete octopole assembly with Delrin supports, stainless steel
mounting assemblies, and octopole rods. The inscribed radius (r0) is 2.94 mm.


271
set-up for high mass operation as described in the previous section. In figures
4-23 and 4-24 are seen the ESI/ion trap mass spectra of (500 fmol///L) horse
muscle apomyoglobin and (350 fmol///L) bovine heart cytochrome c, respectively.
A summary of the operating parameters for the electrospray source can be found
in table 4-3.
The spectrum of myoglobin (figure 4-23) shows a charge state distribution
ranging from +11 (m/z 1542) to +27 (m/z 628). Beginning at charge state +16
(m/z 1060), a small shoulder begins to appear on the right-hand side of the peaks
ranging up to charge state +11 (m/z 1542). This can be attributed to either
resonance ejection voltage optimization effects or ESI adduct ion formation (or a
combination of both). In terms of resonance ejection voltage, larger amplitudes
are needed to efficiently eject higher m/z ions from the ion trap. Many times in
high mass operation, the chosen resonance ejection voltage is a compromise
over the mass range of interest (e.g., not optimum for lower or higher mass ions,
typically observed for m/z ranges over 500). A possible solution to this problem
might involve the ramping of the resonance ejection voltage over the mass range
scanned.
For the other scenario (where adduct ions with the solvent could be
forming), this could be explained by the low capillary-skimmer bias of +15 V.
Low bias values typically result in minimization of collisional desolvation in the
atmosphere-vacuum interface. Characteristically, collisional desolvation is
greatest for higher charge states, thus giving lower residual adduction (as
observed in figure 4-23). For capillary-skimmer biases greater than +25 V,


325
To evaluate the potential of the instrument for negative ion electrospray
analysis, the peptide bradykinin (APPGFSPPA) was used. Bradykinin is
vasodilator belonging to a group of hypotensive peptides called plasma kinins284
To form negative ions, the sample (20 pmol//^l) was dissolved in a 3 mm NH4OH
solution which removes a proton from the carboxy terminus. The polarity of the
dynode, ion trap offset, octopole offset, ion gate lens, and electrospray needle
voltages were reversed for negative ion operation. Oxygen was used for the
counter-current gas flow as an electron scavenger. In figure 4-38 is shown the
negative ion spectrum of Bradykinin. The only ion present is the [M-H]' from the
loss of a proton on the carboxy terminus. The ion injection time was 3.0 ms with
^injectset at 0-2. The sensitivity in the negative ion mode compared to that of the
positive ion mode was a factor of 10 to 100 less. Several possible reasons for
this include source design, peptide stability in highly basic solutions, and
ionization efficiency.


Analyzer
Vlounting Plate
(Bottom)
Dimensions of Ceramic Mounting Holes
Number: 1
Material: Aluminum
Modifications: Anodize
All Units are in Inches
Page 2 of 5
232


338
Peptides and Proteins
This section discusses the photodissociation of the triply-charged ion of
human angiotensin I and the singly-charged ion of the peptide antibiotic
gramicidin D. All experiments in this section were performed with a continuous
wave C02 laser as described previously. The fragmentation nomenclature follows
that of Roepstorff discussed in the CID section of chapter 4.279 Unfortunately, after
preliminary data was obtained for these compounds, a series of malfunctions in
the cw lasers power supply halted all future studies with this laser on peptides,
oligosaccharides, and oligonucleotides. Further experiments were performed with
a pulsed C02 laser.
Human Angiotensin I
One of the first peptides investigated with the new instrument was human
angiotensin I. A continuous wave photodissociation set-up (as described in
chapter 4) was employed for the initial peptide experiments. The electrospray
conditions used were the same as those indicated in chapter 4 for the CID
studies. The m/z 433 ion (triply-charged) of angiotensin I was mass-selected in
the ion trap as described previously. The helium buffer gas pressure was initially
1.0x10^ torr (uncorrected). For this buffer gas pressure, no fragmentation was
observed with the continuous wave C02 laser (for irradiation times up to 400 ms).
This indicated that the large number of collisions of the triply-charged parent ion
with helium buffer gas deactivated (via collisional cooling) the parent ion


CHAPTER 2
FUNDAMENTAL INVESTIGATIONS OF IRMPD IN THE
QUADRUPOLE ION TRAP
The second chapter of this dissertation focuses on the fundamental aspects
of the photodissociation (IRMPD) process in the quadrupole ion trap mass
spectrometer. A large part of the discussion centers around the design of a novel
multipass optical arrangement for use with IRMPD. This design circumvents
previous problems of limited IR laser power, small IR absorption cross sections
for many polyatomic organic species, and the limited ion statistics of trapping and
detection of ions for IRMPD in the quadrupole ion trap. Fundamental
investigations (using model compounds formed by El/Cl) into consecutive
reactions, ion motion effects, the photon absorption process, kinetics, buffer gas
effects, and infrared spectroscopy employing a cw C02 laser are presented.14"148
This work is the basis for future applied studies addressed in chapters 4 and 5
of this dissertation.
Instrumentation
The instrumentation addressed in this section covers the experimental
design (Finnigan MAT ITMS parameters), ion trap operation without He buffer gas,
and a detailed description of the multipass optical arrangement necessary for all
47


360
the newly formed product ion absorbs enough photons to quickly dissociate to
the next sequential product ion.
For the other monosaccharide investigated in this study (1-0-methyl-/?-D-
glucopyranoside), the pyranose ring contained a methoxy substitution at position
one. The reason for examining the fragmentation of this compound was to
investigate the ability of IRMPD to cleave substituted groups on the pyranose ring
structure (e.g., naturally occurring as well as permethylated derivatives). The
methoxy substituent is the simplest case and serves quite well for evaluation
purposes. The conditions for instrument operation were the same as described
above for 2-deoxy-D-glucose.
The IRMPD spectrum of 1 -0-methyl-/?-D-glucopyranoside is shown in figure
5-9. For an irradiance time of 100 ms, structurally diagnostic ions at m/z 194
[M+NH4+-H20] indicating the loss of one hydroxyl group, m/z 180 [M+NH4+-
CH3OH] indicating the loss of the substituted methoxy group, m/z 162 [M+NH4+-
CH30H-H20] showing the loss of methoxy and one hydroxyl, m/z 144 [M+NH4+-
CH30H-2H20] showing the loss of methoxy and two hydroxyl groups, and m/z
126 [M+NH4+-CH30H-3H20] marking the methoxy loss plus 3 hydroxyl groups
(i.e., removed of all ring substituents) are observed.
Although photodissociation experiments have been shown to remove all
substituents from the pyranose ring structure to aid in ring identification, it does
not indicate the substitution position for each individual loss. Perhaps the most
important information provided by photodissociation of the ammonium adducts
of monosaccharides will be the identification of "post translationally" modified


31
either positively or negatively charged droplets. The electric potential essentially
disrupts the flow of liquid from the capillary tip resulting in production of the
charged droplets. To aid in the desolvation process, the droplets are typically
entrained in a nebulization gas such as nitrogen, oxygen, or even SF6. Oxygen
and SF6 act as electron scavengers for negative ion electrospray or when
spraying pure water solutions.84
Solution resistivities on the order of < 10'5 Q are required for stable spray
conditions at room temperature. This corresponds to a solution conductivity of
aqueous electrolytes of approximately 1CT* normal (N), where normality is defined
as one gram molecular weight of the dissolved substance divided by the
hydrogen equivalent of the substance per liter of solution. The higher the surface
tension (the higher the aqueous fraction in the solution), the higher the threshold
voltage needed for the onset of the electrospray process. The relationships
between the electrospray onset voltage V0, the surface tension T the needle or
capillary radius r, and the needle to counter electrode distance h can be shown
to approximate:
V0(Tsr)1sln (1-13)
r
Because the droplet size decreases with increasing solution conductivity, lower
flow rates are required for solutions which are highly conductive. The
dependence of solution conductivity (a) on ESI ion current (I) is relatively weak
(l

254
with a focal length of 4" was placed in line with the beam just in front of the ZnSe
window (see figure 4-18). This lens focused the beam down to a 0.5 cm spot
size. A coarse alignment was obtained by placing the helium-neon laser in line
with the laser beam. Fine adjustment of the alignment was performed by placing
burn-paper at the entrance aperture of the multipass ring electrode to determine
spot size and position. The major disadvantage of using the pulsed laser is in the
alignment procedure, since the instrument must be vented to ensure proper beam
position.
System Interconnections
The system interconnect diagram for the ESI/ion trap instrument is shown
in figure 4-19. The ion trap offset voltage was controlled by the selective mass
storage unit of the ITMS electronics. The balun circuit was floated and modified
to output 18 V^p. The mass set, octopole offset, endcap offset, exit tube lens, ion
gate, and ESI capillary offset voltages were controlled by a Finnigan MAT TSQ 46
voltage power supply. The ion gate voltage was controlled by a bipolar supply,
which was variable for focusing either negative or positive ions (depending on the
switch settings for the ion gate control board built and attached to the Power
Control PCB of the ITMS electronics) into the ion trap. The voltage ranges and
calibration data (using a Fluke 75 DVM, Everett, WA) for the power supply are
shown in table 4-1.


125
7 cm'1 compared to 12 cm'1 for the ion trap.147,189 This difference may arise from
the difference between ion temperatures in the QITMS and the ICR. Ion
translational temperatures (measured with argon-nitrogen kinetic thermometer
reactions) in the QITMS were found to lie in the range 1700-3300 K, whereas
those in the ICR were found to be 500-880 K.190191 The higher kinetic energy of
the ions in the QITMS may correlate with a higher internal energy content, thus
leading to a larger absorption bandwidth because of a higher percentage of
molecules in the vibrational quasicontinuum.
A direct comparison between the infrared spectrum (wavelength
dependence) of protonated diglyme and the corresponding neutral spectrum may
not be feasible due to the presence of the extra OH bond observed with the
protonated species. This extra OH bond could result in a change of the
bonding orbitals and thus affect the COC stretch. However, any direct
comparison between IRMPD and neutral IR spectra must be done with care due
to resolution considerations and the limited spectral coverage of the C02 laser.189
Although direct assignment of various P,Q, or R branches may be difficult, a
general pattern or profile for the gas-phase absorption spectrum is possible.
A direct comparison between gas-phase neutral and positive ion IR
spectrum can be obtained for allyl bromide.148189 Since during the ionization
event a nonbonding electron is removed from the bromine substituent of the
molecule, there is a negligible change in the bonding orbitals which control the
CBr force constant. Therefore, a comparison of the neutral IR gas-phase


287
In almost all applications, the rf voltage applied to the octopole can remain
constant while not significantly affecting the ion transmission properties of a wide
range of ions. When compared to an rf-only quadrupole system (with the same
system parameters as for the octopole and r0 adjusted to produce a quadrupolar
field), all the calculated q2 values exceed the stability limit of 0.908 (see table 4-6).
The value of the Mathieu parameter q2 for a quadrupolar field can be calculated
from the following relationship:13,241
2eV0_P
_ 2,2
(Du r0
(4-7)
Therefore, in order to observe a spectrum over a wide mass range, the rf-only
quadrupole might need an rf voltage ramp (during the ion injection period) in
order to observe all the ions of interest in a given full scan spectrum.
Octopole Offset
The octopole offset can be defined as the potential applied to the octopole
rods in order to extract ions from the skimmer cone region of the ion source. For
the positive ion mode the value of the offset voltage is negative, ranging from a
few hundred millivolts (negative) to approximately -15 volts. The dc offset voltage
is applied equally to all eight pole pieces of the octopole via the rf voltage circuit.
The octopole offset is typically set one to two volts higher than the corresponding
ion trap offset voltage (positive ion mode). This allows for sufficient deceleration
of the ions as they are injected into the mass analyzer region. In figure 4-27 is


Intensity
Dipolar Excitation Frequency in kHz


Figure 2-1:
Instrumental configuration for IRMPD in the ion trap.


Peptide Fragmentation Scheme
P
-Cleavage and Charge Migration
Hydrogen Migration
(a- Carbon)
Ion Type b2
Ion Type y3


316
from the amino terminus). The complementary reaction (which involves charge
retention on the amide nitrogen) for the formation of y type ions is characterized
by a hydrogen ion migration, typically from the a-carbon. The y ion in figure 4-34
is termed y3 because it contains three amino acid residues from the original
peptide structure (counting from the carboxy terminus).280'282
The interpretation of the MS/MS spectrum in figure 4-33 can be explained
by the charge localizations associated with the triply-charged parent ion at m/z
433. The fragmentation indicates that protons are located on Arg2 and His9, with
the third proton free to move along the peptide backbone between these two
residues. This leads to the large amount of (acylium ion type) mid-range
fragmentation observed (b4, b5, b6, and b8). The corresponding series of ions (a4,
ag, ag, and aj arise from the loss of neutral carbon monoxide from the b series
of ions. Some doubly-charged ions are also observed (b6+2 and b9+2), associated
with basic residues His6 and His9 respectively. A complementary series of y ions
also appear (y2, y3, y4) in the spectrum with an unusually high intensity observed
for the y3 ion (proline effect). Attempts of MS3 experiments on the y3 ion to verify
its structure produced no observable fragmentation due to the stability of the ion
in the gas phase. In addition, no a7-b7 pair of ions is observed presumably due
to the proline effect.280'282
The peaks labeled RV, HPF/PFH, and PFHL in figure 4-33 are internal
sequence y type ions, with the latter two possibly driven by the presence of the
pro7 residue. These peaks arise due to the high basicity of proline in the gas-


329
different results that can be seen with CID of multiply-charged species.108 On the
other hand, highly basic peptides such as the neuropeptides of the dynorphin
series exhibit very little, if any, fragmentation (typically immonium ions and small
N-terminal fragments). Even for highly-charged basic fragments (with protons
anchored into position) where coulombic repulsions are strong, fragmentation is
still limited.110
In addition, the broad energy distribution associated with the CID process
can further complicate the spectral interpretation process. Also, for singly-
charged ions of m/z > 2500, kinetic and collisional effects limit the amount of
structural information available.289,290 Therefore, it is desirable to explore
alternative methods for the activation of biological species in the gas-phase.
Photon absorption is one possible alternative that overcomes many of the
aforementioned short-comings of collisional activation. A well-defined energy
deposition process (independent of converting an ions translational energy into
the vibrational energy needed for fragmentation) and the long ion-storage times
associated with trapping instruments make photodissociation an ideal choice for
structural elucidation of biological molecules.
Early studies using photodissociation for structural studies of biological
species in the ICR employed a wide range of wavelengths from infrared (IR) to
ultraviolet (UV).1111291,292 Williams et al. reported photodissociation efficiencies on
the order of 100% for the peptide alamethicin using 193 nm light when the ions
were confined to the beam path.291 Corresponding CID experiments involving


226
In figure 4-11 is shown a schematic diagram of the complete rf circuitry
used to control the rf-only octopole. The octopole offset (extraction voltage) and
mass set (a zero to ten volt set point for producing the rf amplitude) were
obtained from a Finnigan MAT TSQ 46 lens voltage controller. A Stanford
Research System model DS345 (used above) arbitrary waveform generator was
used to produce the 1.659 MHz frequency for the rf amplifier circuit.
Analyzer Assembly
The analyzer assembly was designed to fit on an 8" rotatable Conflat
flange. The rotatable flange was used to facilitate the alignment of the multipass
ring electrode with the incoming laser (pulsed or continuous wave) radiation.
Welded to the rotatable flange were four stainless steel rods which were used as
an optical rail system to facilitate movement of the analyzer assembly relative to
the laser window and detector arrangement. A 1 /4 Swagelock union was welded
into the 8" Conflat for direct coupling of a He buffer gas line or connection to a
pulsed-valve as described in chapter two of this dissertation. A Negretti valve
(Negretti LTD, Southampton, UK) was used to control the back pressure of He for
constant pressure or pulsed valve experiments. The 8" Conflat was also equipped
with three straight high voltage feedthroughs (rated to 20 kV) and 7 single-ended
MHV feedthroughs rated to 5 kV (Insulator Seal Incorporated, Hayward, CA). One
of the high voltage feedthroughs was used for the 20 kV applied to the off-axis
conversion dynode, while the seven MHV feedthroughs were used for the detector


Figure 4-3:
A simplified schematic diagram (top view) of the ESI/ion trap system including the electrospray
source, octopole assembly, ion trap analyzer and detector assembly.


410
75. Louris, J.N.; Cooks, R.G.; Syka, J.E.P.; Kelley, P.E.; Stafford, G.C.; Todd,
J.F.J. Anal. Chem. 1987, 59, 1677-1685.
76. Todd, J.F.J.; Bexon, J.J.; Smith, R.D.; Weber-Grabau, M.; Kelley, P.E.;
Syka, J.E.P.; Stafford, G.C.; Bradshaw, S.C. in Proceedings of the 16th
Meeting of the British Mass Spectrometry Society York, U.K., 1987, 206-
209.
77. Tucker, D.B.; Hamiester, C.H.; Bradshaw, S.C.; Hokeman, D.J.; Weber-
Grabau, M. in Proceedings of the 36th ASMS Conference on Mass
Spectrometry and Allied Topics San Francisco, CA, 1988, 662-663.
78. Gronowska, I.W.; Paradisi, C.; Traldi, P.; Vettori, U. Rapid Comm. Mass
Spectrom., 1990, 4, 306-313.
79. Yates, N.A.; Yost, R.A.; Bradshaw, S.C.; Tucker, D.B. in Proceedings of the
39th ASMS Conference on Mass Spectrometry and Allied Topics Nashville,
TN, 1991, 132-133.
80. Eades, D.M.; Yates, N.A.; Yost, R.A. in Proceedings of the 39th ASMS
Conference on Mass Spectrometry and Allied Topics Nashville, TN, 1991,
1491-1492.
81. Kelley, P.E.; Hoekman, D.J.; Bradshaw, S.C. in Proceedings of the 41st
ASMS Conference on Mass Spectrometry and Allied Topics San Francisco,
CA, 1993, 453a-453b.
82. McLuckey, S.A.; Goeringer, D.E.; Glish, G.L. Anal. Chem. 1992, 64, 1455-
1460.
83. Yates, N.A. ICMS Software, Department of Chemistry, University of Florida,
1993.
84. Smith, R.D.; Loo, J.A.; Ogorzalek, L; Busman, M.; Udseth, H.R. Mass
Spectrom. Reviews 1991, 10, 359-451.
85. Dole, M.; Mack, L.L; Hines, R.L.; Mobley, R.C.; Ferguson, L.D.; Alice, M.B.
J. Chem. Phys. 1968, 49, 2240-2249.
86. Mack, L.L; Kralik, P.; Rheude, A.; Dole, M. J. Chem. Phys. 1970, 52, 4977-
4986.
87. Gieniec, J.; Mack, L.L; Nakamae, K.; Gupta, C.; Kumar, V.; Dole, M.
Biomed. Mass Spectrom. 1984, 11, 259-268.


Octopole Assembly


Displacement (mm) | Z |
185
(a)
(b)
4-0
2 0
50
100
200
Length (mm)
150


183
various multipoles. The force on the ion in the quadrupole has a linear
dependence on displacement from the center of the device. The force exerted
on the ions in the octopole, however, is cubed with respect to the distance off the
center axis of the device. At large distances off the center axis, trajectories can
become quite large and ion motion can become unstable as shown in figure 3-
13.253
In this case, the authors report that the best possible operating parameters
for the rf-only octopole would be for low ion entry angles and small off-axis
entrance distances.253 The instrument designed in our laboratory has the
octopole placed as close as possible (0.015" away from the skimmer cone) to the
electrospray interface. By placing the rf-only octopole this close to the skinner
cone, a larger cross section of the beam image is near the center axis of the
device and the divergence angle of the entering ions is minimized. This, coupled
with the incoherent oscillations observed for octopole operation at various m/z
ratios, make it an excellent alternative to traditional dc lens or rf-only quadrupole
ion injection systems.
RF Frequency
As the rf frequency applied to the octopole is increased, the wavelength of
the secular motion also increases as seen in figure 3-14. This agrees well with
the fundamental relationship observed between the secular wavelength of motion
and the applied rf frequency. The fundamental wavelength of motion for an


343
irradiated. In figure 5-4 is shown the corresponding photodissociation spectrum.
The major fragment ion observed was a rearrangement loss of the modified
carboxy terminus of the peptide. The [M-61]+ ion at m/z 1818 is formed by
hydrogen migration from somewhere along the peptide backbone (preferentially
on the or carbon relative to the tryptophan residue in position 15), and subsequent
cleavage of the amide bond between the modified carboxy terminus and the
tryptophan residue in position 15. The two peaks to the immediate right of the
[M+H]+ ion (m/z 1879) are the sodium and potassium adducts of gramicidin D.
Photodissociation of these species yielded no loss of the modified carboxy
terminus. The appearance of the [M-61]+ ion as the major fragment peak in the
spectrum suggests that protonation occurs preferentially on the modified carboxy
terminus of the peptide (e.g., the most basic site) and charge site driven
fragmentation results. The only other structurally relevant peak observed was the
b14 ion at m/z 1636. As seen with the previous IRMPD experiment, a sequence-
specific b ion is produced which could be used to obtain ladder sequence
information to verify peptide sequences.
Carbohydrates and Oligosaccharides
One of the biggest challenges in biochemistry is the identification and
sequence analysis of complex carbohydrates and oligosaccharides. Perhaps the
most difficult problem in carbohydrate chemistry is the investigation of biologically
active binding-recognition systems. The biologically active portion of a


Figure 4-23:
ESI spectrum of horse muscle apomyoglobin showing charge states ranging from +27 to +11.
The sample concentration was 700 fmol///L at a flow rate of 3 /^L/min.


157
Effective Trapping Potential
To understand the effect of trapping potential on multipole device design,
a discussion of the forces acting on an ion and thus displacing it from the center
axis of the device is necessary. The general equation which describes this
displacement (for the effective radial potential energy; xy plane) for any multipole
device is defined as:
(3-56)
where V#ff is the effective trapping potential in the xy plane, n is the number of
poles, e is the charge on the ion, V0 is the amplitude of the rf voltage (zero-to-
peak), m is the mass of the ion, u is the rf frequency, r0 is the inscribed radius,
and r is the displacement distance of the ion from the center axis of the
device.221227
A plot of effective trapping potential (Vctf) versus ion displacement from the
central axis (r) of three different rf-only devices (quadrupole n=2, hexapole n=3,
and octopole n=4) is shown in figure 3-4. The octopole parameters used are
seen in table 3-1. These parameters are the actual design specifications of the
rf-only octopole developed in our laboratory. All parameters for the three
multipole systems in table 3-1 are constant except for the value of n. The plot
shown in figure 3-4 is for the +3 charge state of bovine insulin. The radial
potential of the octopole has steep repulsive walls which approximate the ideal


Figure 5-8: Consecutive reaction curves for the IRMPD of the ammonium
adduct of 2-deoxy-D-g!ucose. The top half of the figure shows
the sum of the major product ion intensities, plus the presence
of the direct dissociation of the complex to showing the
ammonium species at m/z 18 (dashed line). The bottom portion
of the figure represents the individual curves comprising the
major product ion intensity curve in the top half of the figure.


382
methods.306,307 Recent successes in the field of electrospray have enabled
ionization of oligonucleotides up to 76 base pairs.115 Typical electrospray spectra
include sodium attachment to the multiply-charged anions of the oligonucleotides;
this can complicate molecular weight analysis and has proved to be the single
most limiting factor in electrospray mass spectrometry of this compound class.115
In addition, the weaker signals generated by the oligonucleotide anions can limit
sensitivity of the electrospray technique (compared to the more abundant positive
ion signal for peptides and proteins).
Perhaps the most difficult challenge of negative ion electrospray for the
analysis of oligonucleotides is to obtain structural information from MS/MS
experiments. To date, only a handful of articles has appeared on the MS/MS of
negatively-charged oligonucleotides.111,112,308,309 The main problem is in the area
of data interpretation, where multiple charge states and adduct formation can
complicate even the simplest of CID spectra. Previous reports have centered on
the CID of small, multiply-charged negative ion DNA fragments (up to 8 base
pairs) using electrospray ionization/ion trap mass spectrometry, and linker DNA
8-mers using FTICR and photodissociation.111,308 Both techniques show a great
deal of promise for the analysis of larger oligonucleotides.
The fragmentation nomenclature employed for tandem MS/MS studies of
oligonucleotides was developed by McLuckey et al.308 Shown in figure 5-17 is a
triply-charged tetranucleotide with bases B, through B4. Cleavage points along
the phosphodiester backbone are indicated by the lower case letters a, b, c, and


Figure 4-12:
The aluminum mounting plate assembly shown in a series of five diagrams: (1) general
dimensions for the mounting plates, (2) ceramic mounting holes to hold the ion trap assembly; the
holes are offset 13.3 from center to facilitate alignment of the multipass ring electrode with the
laser, (3) assembly pin hole dimensions for the stainless steel rods used to hold the analyzer
assembly together, (4) holes for the optical rail assembly, and (5) relief holes for the pulsed-valve
and electrical feedthroughs.


13
trap. These parameters determine whether or not ion motion is stable or
unstable. For motion in both the r and z direction the values of a^ and qrz
become:
az = -2ar=
-8eU
m(ro+2z02)s
Qz = 2qr=
4eV
(1-8)
(1-9)
m(ro+2Zo )2
The Mathieu stability parameters au and qu are directly proportional to the applied
dc and rf voltages on the ring electrode. Since a and qu are inversely
proportional to m/z, the motion of all ions can be expressed in terms of these
Mathieu stability parameters.23
In equations 1-8 and 1-9, the relationship the between axial (zj and radial
(r0) dimensions of the ion trap are defined theoretically by z0=r0/21/2. However, all
commercially manufactured ion traps to date posses a stretched geometry in the
z-direction. Therefore, instead of a theoretically defined Zq distance of 0.707 cm,
the ion trap is stretched axially 10.7% with a new value of 0.783 cm.53 In a
previous report by Johnson et al.54, the general form of equations 1-8 and 1-9
(also known as the Knight equations) were found to be a good approximation for
experimental measurements to determine qr and a,z values made with the
stretched ion trap geometry.
The stability diagram, which specifies stable ion trajectories, can then be
defined as the intersection of the solutions to the Mathieu equation in the z-


Figure 2-4:
Kinetic data from the selective mass storage of the 79Br isotope from the molecular ion of allyl
bromide, m/z 120. A plot of In (l/l0) versus time yields a straight line, with the slope equivalent to
the rate coefficient of the ion-molecule reaction of m/z 120 with neutral allyl bromide.


42
IRMPD process and to outline the applications of this process to large biological
ions in the gas phase.
The majority of investigations into the mechanism of the photon absorption
process have utilized pulsed or continuous wave (cw) lasers operating in the
wavelength range 9.5 to 10.5 /m. The fundamental model of photon absorption
and subsequent dissociation is based on the extensive studies of the SF6 system
at 10.6 //m.144 The basic concepts of this model also apply to the IRMPD process
of other gas-phase systems.
In the model proposed by Black et al., there are three successive regions
for photon absorption as the energy deposited in the ion increases.144 A
schematic energy diagram in figure 1-6 shows the three individual regions (I,II,
and III) corresponding to coherent multiphoton interaction, incoherent single
photon interaction, and dissociation threshold/channels. In the coherent
multiphoton interaction region (I), vibrational states are essentially discrete and the
corresponding absorption of a single IR photon is a straightforward process.
However, successive transitions to higher vibrational levels are not possible since
the vibrational spacing is now out of resonance with the monochromatic IR
photons. Therefore, in order to further excite the ion of interest other mechanisms
which compensate for these anharmonicities must be in operation.145
For any polyatomic ion the density of the vibrational levels increases with
internal energy. The rate of this increase depends upon the magnitude of the
frequencies of the individual modes and the molecular complexity of the ion.




To Computer
cn
o


110
_ d[C5H,,02r s kl[C6H,503]- k2[C5H02]* (2'13>
at
By substituting for [C6H1503]+ in equation 2-13 with equation 2-12, the following
expression is obtained:
dickey*
dt
MC6H,503];e-M
Subsequent integration of equation 2-14 yields:
MC5H02]- (2-14)
[C5H02]- = [C,H,50,li(e-M -e-k!') (2-15)
K2 K-,
Therefore, the equation for the variation of the third sequential reaction product
[C3H70]+, can be obtained from the sum of the total ion current; assuming no
other routes for ion losses:
[CeH,s03]- [C5H02r [C3H70]* = [C3H1503]J (2-16)
so that:
[C3H70]- = [C6H,503] [C6H,503]- [C5H02J* (2-17)
Inserting the expressions for [C6H1503]+ (equation 2-12) and [C5H1102]+ (equation
2-15) into equation 2-17 yields:
[c3H7or =
k2 k.,
[k2(1 e'k,t) k,(1 e*kzt)] (2_18)
As seen in figure 2-14, the ion counts for [C6H1503]+ (m/z 135) fall
exponentially while the counts for [C5H1102]+ (m/z 103) approach a maximum.


Figure 2-9:
Effect of the applied excitation signal (e.g dipolar or quadrupolar) on ion motion in the quadrupole
ion trap. The symbol ujz refers to the fundamental frequency of motion of an ion in the z-direction.
For maximum power absorption under dipolar excitation conditions, the ions frequency should
match that of the applied supplementary ac signal as seen in (a). The negative cycles of the
supplementary ac coincide with the ion frequency in the z-direction, thus leading to an increase
in amplitude for the ion trajectory and corresponding increase in ion kinetic energy. For
quadrupolar excitation where the applied supplementary ac signal is in phase with respect to both
endcaps, the applied frequency must be two times that of the fundamental ion frequency to obtain
maximum power absorption. An applied supplementary ac at the identical frequency of motion
of an ion in the z-direction leads to a significantly reduced power absorption spectrum (as seen
experimentally in figure 2-10).


Intensity
Quadrupolar Excitation Frequency in kHz


O)


87
The power absorption spectrum of protonated 12-crown-4 ether (no He
buffer gas present) for qz=0.3 and az=0 is shown in figure 2-10. Three main
frequency bands were observed at a/r (59.2 kHz), 2u/r (126.0 kHz), and 2uiz (239.8
kHz). Two of these absorption bands, at 2ior and 2uz, were offscale, while the wr
band showed limited absorption. The magnitude of the 2a/r and 2uz bands was
approximately equal, which corresponds well with results reported by Eades et
al.162,164 Also, the magnitudes of all absorption bands observed agree well with
the theoretical values calculated by March.57,156,157-162 166 Note that the absorption
bands at 2wz and 2wr are not directly related to frequencies of ion motion, but
instead represent only the absorption of power. This can be explained by the
frequency doubling of the quadrupolar excitation signal applied to the endcaps,
such that the charge on an endcap is of the appropriate polarity to obtain
maximum power absorption (figure 2-9). Therefore, this absorption of power at
2a/r and 2wz represents actual ion motion at wT and a/z, respectively.
To examine the effect of coupled ion motion using quadrupolar excitation,
the appropriate excitation signal was applied at the three frequencies observed
(a/r, 2u)v and 2uz) with concurrent laser irradiation, as mentioned previously in this
chapter. A plot of the photodissociation efficiency versus the quadrupolar
excitation voltage applied to the endcaps yields the three curves shown in figure
2-11. As expected for the 2a/z band (representing u>z frequency component),
there is observed a steep drop-off in photodissociation efficiency as the axial
excursions of the ions exceed the laser beam width (for the reasons mentioned


4-21261 and 4-22261. For the case where no resonant ejection frequency is applied,
the instrument operates in the traditional mass selective instability scan mode.
When this occurs, the larger mass ions (e.g., > m/z 650) are never "scanned out"
of the ion trap because the applied rf voltage is not high enough to bring these
ions to the z-stability edge of the Mathieu diagram (see figure 4-21)261. Instead,
they essentially drift out of the analyzer region when the applied rf voltage goes
to zero at the end of the analytical scan. To efficiently detect the larger mass
ions, the z-stability edge of the Mathieu diagram must be moved in" so that ions
of higher m/z are detected. Moving the z-stability edge is accomplished by
applying an auxiliary ac frequency (i.e., the resonant ejection frequency) on the
endcap electrodes, which is equivalent to an ions given frequency of motion in
the z-direction. As the rf voltage is ramped during the analytical scan, the ions
come into resonance with the applied frequency and are ejected from the ion trap
and detected by the electron multiplier. This phenomenon of mass range
extension is shown in figure 4-22 where the points A, B, and C represent the
application of the resonant ejection frequency for qeject = 0.906, 0.454, and
0.0908, thereby extending the mass range to 670,1300, and 6500, respectively.261


spherically symmetric mirrors were placed in the radial plane of the ion trap ring
electrode to increase the photoabsorption pathlength of the incident radiation.
The performance of this multipass ring electrode was characterized by conducting
experiments on ion motion, kinetics, consecutive reactions, spectroscopy and
buffer gas pressure to examine their effects on photodissociation efficiency of a
series of model compounds.
Next, a novel ion injection system was designed and built in order to
transfer biological ions from the electrospray source to the ion trap mass
spectrometer in the most efficient manner possible. The instrumentation
consisted of a modified electrospray interface coupled to an rf-only octopole,
which transmitted ions from a high pressure region near the ion source to the
lower pressure region of the ion trap analyzer. The theory, design, and
characterization of the rf-only octopole is discussed, along with an analysis of the
ion transmission properties of the device.
Finally, a detailed investigation into the photodissociation of ions from
carbohydrates, peptides/proteins and oligonucleotides is presented. Sensitivity
considerations, fragmentation patterns, and comparison studies to collisional
activation are discussed. An analysis of the type of structural information
obtained from photodissociation showed its applicability in solving biological
problems.
IX


Figure 5-18: Photodissociation spectra (5 laser shots) of the negatively
charged RNA dimers adenyl (ApA) adenosine (top) and adenyl
(ApC) cytidine (bottom). The individual structures and
fragmentations are also shown. Bond cleavage for the
photodissociation process occurs at the phosphodiester bond
PO where there is a relatively high photoabsorption cross-
section in the IR region.


Intensity (Arbitrary Units)
800 i
600
400
200
K+
in/z 39
C7H123
m/z 144
v+
(M+H) +
m/z 265 (M+K)+
m/z 303
1|III|IIIpi1T1
50 100 150 200
m/z
250
300
350
337


75
include the amplitude of the applied excitation signal (frequency shifts to higher
values with increasing amplitude), He buffer gas pressure as described by the
physics of the damped harmonic oscillator model (frequency shifts to lower values
with increasing gas pressure), and space charge or ion-ion interaction
considerations (frequency shifts to lower values with increasing ion population).
In addition, arbitrary shifts in an ions frequency can also occur due to the
uncertainty of the exact RF voltage applied to the ring electrode of the ion trap.
Consequently, the experimentally determined ion component frequencies are only
qualitative estimates of the true ion frequencies.
The power absorption spectrum of protonated diglyme (no He buffer gas
present) for qz=0.3 and a^O is shown in figure 2-7. Three main frequency bands
were observed at wz (118.3 kHz), 2aiz (235.9 kHz), and 3a/z (349.8 kHz). The
broad frequency band at 55.1 kHz could correspond to the wz/2 band, however
the broad bandwidth precludes direct frequency assignment at this time. An
unknown frequency band at 446.0 kHz is also present. The largest absorption
band (the ions secular frequency of motion in the z-direction) at a/z was very
broad due to the large excitation amplitude used. Observation of the smaller
bands at 2o/z and 3a/z can be attributed partly to higher-order field effects
(hexapolar) on ion motion and have been successfully predicted using forced
Mathieu equations as developed by Williams et al.165 To account for power
absorption at these frequencies, several different explanations (or combinations
thereof) are plausible, including: 1) the presence of higher order fields


270
The amplitude of this resonant ejection frequency must be great enough
to cause ejection of the high mass ions within a few rf cycles after the ions come
into resonance. Typical values are in the 7.5 V^p range, which means that ions
gain sufficient kinetic energy quickly enough to be ejected from the trap before
they are dissociated. Considerations of indefinite mass range extension, mass
resolution issues, and electrometer modifications are negligible for the operational
mass range of the experiments described in chapters 4 and 5. A detailed
discussion of the principles and operational parameters can be found in the
literature.73259"261
Basic ESI Operation
For ESI operation, the El source and corresponding flange extension were
removed and the modified ESI source and Conflat adapter flange were installed.
The system was then cabled as described in the system interconnect section (see
figure 4-19). The compounds used to evaluate system performance of the ESI/ion
trap were horse muscle apomyoglobin, bovine heart cytochrome c, and bovine
insulin (Sigma Chemical Company, St. Louis, MO). These peptides/proteins were
chosen because they are frequently used as standards for evaluating ESI/mass
spectrometer performance. Protein/peptide samples were prepared in 50:50
methanol:water solutions with a 0.1% acetic acid content. Unless specified
otherwise, the sample of interest was infused directly into the instrument at a flow
rate of 3 //L/min using a syringe pump (described previously). The ion trap was


Figure 2-2:
Adjustable mounting bracket for the ion trap analyzer assembly. The three slotted holes allow for
rotation of the analyzer to facilitate laser alignment with the entrance aperture on the ring electrode.


Figure 5-5:
Structure of the simple monosaccharides 2-deoxy-D-glucose and 1 -0-methyl-/?-D-glucopyranosside
used for evaluation of the IRMPD technique for ring cleavage fragmentation.


Intensity
381
700


275
extensive fragmentation of the higher-charged states can occur. Therefore the
selection of the appropriate capillary-skimmer bias is usually a compromise
between the desolvation and adduct formation effects.84,106
Other parameters that can affect adduct formation include the capillary
temperature and the dry gas flow. For the apomyoglobin sample in figure 4-23,
the capillary temperature was 180 C, which provided the most efficient
desolvation (solvent molecules) while not producing any unwanted thermal
degradation effects. Since the gas flow in the electrospray source was not
heated, some condensation effects caused by gas expansion in the vacuum
region could contribute significantly to adduct formation. This can be typically
offset by increased capillary temperatures, increased capillary-skimmer bias, or
excitation (e.g., CID of adducted ions) in the ion trap analyzer to produce
desolvated [M+nH]+ ions (where n=1,2,3...).
During the electrospray process, changes in the protein environment (e.g
pH effects, addition of organic solvent, addition of modifiers such as urea, and
temperature effects) can dramatically affect the observed charge state distribution.
As shown in figure 4-24 for bovine heart cytochrome c, a bimodal charge state
distribution is observed for an acetic acid content of 0.1% (pH=3.27). At lower
pH (and/or high organic solvent content), cytochrome c denatures from a tightly
folded complex to a random coil state.262,263 This protein "unfolding" essentially
makes basic amino acid residues normally unavailable for protonation (e.g.,
effectively buried in the internal globular structure of the protein) readily


8 Inch Conflat
248


141
considerations, the octopole electrode structure must have a two-fold rotational
axis of symmetry where the device can be rotated about the axis rr/4 radians, and
the new configuration is indistinguishable from the old. Since the potential along
the z-axis is zero, the hyperbolic rod potential relative the z-axis is 0J2 and -0J2
for adjacent rod pairs.219,237'238
Octopole Field Potentials
The potential field generated by an octopole device can be derived by
starting with the LaPlace differential equation defined in terms of polar coordinates
where 0=0(r,0,t) is equivalent to the rectangular coordinate system with
0=0(x,y,t). Polar coordinates will be used to solve the equation for the field
potential since the mathematics involved are much less complex and time-
consuming. At the end of the derivation, the polar coordinate system will be
converted back to traditional rectangular coordinates. Laplaces equation takes
the form in polar coordinates of:
V4, = ^*^ = ^f = 0 (3-22)
dx2 dy2 dr r2 302
Since the octopole electrode structure displays circular cylindrical symmetry,
circular cylindrical coordinates can be used to seek the solution of the Laplace
equation, which takes the form:
= R(r) 0(6) 0(t) (3-23)
where each of the functions R(r), 0(0), and 0o(t) has only one independent


256


359
For the case of single-frequency CID, it is critical that a constant ion population
is present so that re-tuning of the instrumental parameters involved for the CID
process (e.g., 20 min even for an experienced operator) need not occur (see
figure 5-1). Also, in order to gain the same information obtained from the IRMPD
experiment in figure 5-7, a third stage of mass analysis must be performed which
means that the efficiency of the MS3 step is directly influenced by the parent ion
population frequency shifts, which translates down to the MS/MS efficiency, which
then can drastically affect the formation of the parent species for the MS3 process.
In addition, the time required for tuning of the ion trap for an MS3 process can
take on the order of 30 min. More importantly, for the case of real-time sample
concentration profiles observed in gas or liquid chromatography, single-frequency
CID experiments could produce a limited amount of structural information due to
shifting peak concentration profiles. For the IRMPD process, there is an initial
time investment for turning on and tuning the laser; however, since the
dissociation event is independent of ion population (e.g., only depends on the
ability of functional groups to absorb photons), MS/MS experiments with
photodissociation can be carried out for real-time chromatographic analysis. In
addition, "multiple" stages of mass spectrometry for the IRMPD process can be
obtained by simply increasing the irradiance time for the parent/product ion
species. The one potential drawback of the IRMPD process could be the
formation of low intensity structurally significant ions which are never observed
because the laser irradiation period is set such that upon fragment ion formation,


347
sequence information have also been successfully performed using FAB and
electrospray ionization techniques for generation of abundant parent ion
populations.302'304
In this section, an alternative approach to traditional CID for determining
carbohydrate structure is presented. Photodissociation in the IR region is a
natural choice for cleaving glycosidic bonds, identifying ring structure, or possibly
determining "post translational" modifications to monosaccharide residues. This
is due mainly to the high photon absorption cross-section seen for C0C
linkages (see chapter 2). Although the appropriate derivatization chemistry and
enzymatic data are needed to help determine internal oligosaccharide sequence
information, the use of photodissociation in combination with electrospray
ionization can provide a basis for tackling the more difficult problem of anomeric
confirmation in the gas-phase.
The first portion of this section discusses the determination of the number
of hydroxy substituents on monosaccharide residues. The initial studies
performed were done using a solids-probe and ammonia chemical ionization in
order to test the feasibility of the photodissociation technique. Original
comparisons with CID data showed increased dissociation efficiencies, and
reduced analysis times for IR photodissociation experiments. The second portion
of this section covers the use of photodissociation for cleaving the glycosidic
bonds of straight chain oligosaccharides (e.g., ammonium adducts of raffinose
and stachyose) using the instrument described in chapter 4.


14
direction with those in the r-direction. In this overlap region, ions are stable in
both the z and r directions and can be stored in an ion trap. The overlap region
which is closest to the a, q origin is of the most practical significance, since the
voltages are less and thus the electronics needed to apply the appropriate rf and
dc voltages to the ring electrode can be readily designed. Figure 1 -2 shows the
stability diagram for the three-dimensional quadrupole ion trap.23 The ordinate
and abscissa are expressed in terms of the dimensionless quantities a and q
respectively.
Ions within the stability region exist in a pseudopotential well, with a depth
of about 1 eV near the origin and 10 eV near the right hand (z) stability edge.
The lines drawn across the stability diagram are called iso-6, lines which describe
detailed trajectories of ions at that particular point. Iso-6 lines essentially
determine the frequency of ion motion; they are found in the general solution to
the Mathieu equation:
OO CO
u(0=A£ C2ncos(2n + 6)£+ B £ C2nsin(2n+6K <110)
n=- n=-
where the coefficients are the amplitude components of oscillation and the
(2n+6) terms represent the frequency components of ion motion. These
quantities can be calculated for the corresponding values of a and q using
recurrence relations/continued fractions.55 The relationship between ion frequency
(ft/n J and 6 can be defined as follows:


Table 3-1. Parameters for the radial potential energy equation.
Device
n
e
w (MHz)
r0 (mm)
V0 (V)
m (g/mol)
Quadrupole
2
+5
1.659
2.94
250
5733
Hexapole
3
+ 5
1.659
2.94
250
5733
Octopole
4
+5
1.659
2.94
250
5733
158


Figure 2-15: Reaction mechanism tor the IRMPD of protonated diglyme using
a cw C02 laser. The low energy reaction channel corresponds
to the absorption of a minimum of three photons (e.g. formation
of m/z 103), while the high-energy reaction channel corresponds
to the absorption of at least an additional 6 photons, for a total
of 9 photons absorbed by the protonated species. All parts of
the reaction mechanism were confirmed by MS" experiments.


Figure 4-10: A plot of detected rf versus mass set voltage. The second y-axls on the right is the corresponding
rf output voltage. Since the scope probe has some finite capacitance, the rf output voltage is only
an approximation of the real output voltage.


Ring Electrode
cw CO2
Laser Beam
1
Spherical
Mirrors
r=2.0 cm
Spherical
Mirror
r=2.0 cm
1 Pass
3 Passes
Hi 2 Passes
4 Passes
0


115
l^6^133]+ ~ [^10^21^5)0
ki
k2 k1
(e
-k,t g -k2t
(2-27)
[C4H902]+ [C10H21O5]0
k,k2e
k1k2e"kl
-k2t
[C2H50]* = c6h2105]
k, k3e
-k2t
(k2 -
M(k3 -
ki)
k1 k2e"k3t
'
(ki
~ k3)(k2 -
CO
k2k3e R
CM
1
- ki)(k3 -
-k,)
k1 k2e"k
3^
(2-28)
(2-29)
(ki k2)(k3 k2) (k1 k3)(k2 k3)
As seen in figure 2-16, the ion counts for [C10H21O5]+ (m/z 221) fall
exponentially (kn= 104 s'1) while the counts for [C6H13OJ+ (m/z 133) approach a
maximum. Because the rate of formation of [C4H902]+ (m/z 89) is proportional to
the ion counts for [C6H1303]+ (m/z 133) the rate is zero at time and has a
maximum value when [C6H1303]+ (m/z 133) reaches a maximum. At this
maximum point, the value of Ic, was calculated as 32 s'1 (see derivation of
equations 2-19 and 2-20). The same derivation which was applied to the
consecutive reaction sequence for protonated diglyme and the third sequential
reaction product of protonated 15-crown-5 ether was also applied to the fourth
sequential reaction product of protonated 15-crown-5 ether, which gave k3=19 s'1.
In figure 2-16 is seen the complete consecutive reaction sequence for protonated
15-crown-5 ether. There also was a small degree of back reaction observed for


342
helium buffer gas pressures during the laser irradiation period were reduced to
below 4.0x1 O'6 torr. However, significant reductions of pressure in the ion trap
analyzer not only adversely affect peak shape and resolution, but more
importantly also reduce ion injection efficiency and thus sensitivity. For the
angiotensin I studies, ion injection times were increased from the 1 to 3 ms range
(when operating with a helium buffer gas pressure of 1.0x10"* torr) to the 68 to 75
ms range with the reduced operating pressures (e.g., 2x1 O'5 torr). The addition
of a pulsed-valve (as discussed in chapter 2 and chapter 6) in order to separate
the ion injection, photodissociation, and detection events in time from the helium
pulse should significantly improve photodissociation efficiency without sacrificing
sensitivity.
Gramicidin D
The second peptide investigated using the IRMPD process was the
antibiotic gramicidin D. The electrospray and instrumental conditions employed
were similar to those reported in the previous chapter. Photodissociation was first
observed for the singly-charged ion of gramicidin D with a helium buffer gas
pressure of 1.7x1 O'5 torr and an irradiance time of 85 ms. Ion injection conditions
were set such that the majority of the signal observed was that of the singly-
charged species, with only a small portion of the doubly-charged ion stored for
instrument calibration purposes. Since no isolation was performed, the [M+H]+,
[M+Na]+, and [M+K]+ (as well as the low intensity [M+2H]+2 ion) were all


Figure 5-7: IRMPD spectra of the ammonium adduct of 2-deoxy-D-glucose
at 85 ms and 120 ms irradiance time. High dissociation
efficiencies were observed for IRMPD, since the kinetic energy
of ion motion does not need to be converted to the internal
vibrational energy needed for fragmentation. The number of
hydroxyl substituents on the ring can be determined much easier
with IRMPD, since dissociation efficiency is independent of ion
population.


146
A pictorial representation of component and unit vector system used in this
derivation can be seen in figure 3-3. To determine the electric field vector in the
x-direction, equation 3-30 and the following identity are used:
(n 2k)
k = 0,1,2,... s n = 4 3*32)
From the previous section (not shown), the cos n0term for equation 3-30 can be
derived as:
cos(n6) =
1
/ \
n
: V (2
Xn-2y2 +
xn 4 y 4
(3-33)
rn
Differentiating equation 3-33 with respect to x and (substituting for n=4 and
evaluating k=1 to 4) and applying equations 3-30 and 3-32, the vector component
E, is defined as:
Ex = 4>0 r3 cos 30 (3-34)
ro
For the component vector in the y-direction Ey, equation 3-30 and the following
identity are used:
2k
= n
n -1
2k -1
k = 1,2,... s n =4
(3-35)
From the previous section (not shown), the sin n0 term for equation 3-30 can be
derived as:


116
m/z 45 and m/z 89 with neutral 15-crown-5 ether. Again, the rate coefficients for
the two reactions (to form protonated 15-crown-5 ether) were less than 0.5 s'1 for
a pressure of 3.6x1 O'7 torr. Only at longer irradiance times (t>50 ms), where the
ion population of the protonated 15-crown-5 ether (m/z 221) is small, will the
effect of the back reaction be significant.
Buffer Gas Effects
Collisional deactivation studies of polyatomic species typically provide
energy transfer information on a broad scale. These gross changes in energy
levels are represented by changes in observed fragmentation or dissociation of
given polyatomic species. Previous studies of collisional quenching (pulsed C02
lasers) carried out in the ICR examined the efficiency of the quenching
process.129,131,182'184 For the case of low power cw C02 lasers, collisional
deactivation successfully competes with photodissociation at sufficiently low
photon fluxes and relatively high neutral pressures.185 Therefore, the amount of
photodissociation observed relates directly to the collisional deactivation efficiency
of the neutral buffer gas.
In figure 2-17, the effect of collisions on the IRMPD of protonated diglyme
is shown. The general trend observed for all buffer gases was a decrease in the
photodissociation efficiency with increasing buffer gas pressure. The observed
trend for quenching efficiency was N2>Ar>He. These results correlated well with
the previous studies of polyatomic and diatomic ions using the same buffer


A NOVEL ELECTROSPRAY ION TRAP MASS SPECTROMETER FOR
PHOTODISSOCIATION OF BIOLOGICAL MOLECULES
LD
1780
159^
Si%
JAMES LEE STEPHENSON, JR.
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
1995
UNIVERSITY OF FLORIDA LIBRAS

"I have little patience with scientists who take a board of wood, look for its
thinnest part, and drill a great number of holes where drilling is easy."
Albert Einstein in "Einsteins Philosphy of Science"
Reviews of Modern Physics 1949, 21, no. 3.
"Just work hard and be honest, and everything else will take care of itself..."
Orson Calvin "O.C." Pearson ....my Grandfather

ACKNOWLEDGMENTS
I would like to begin by expressing my unfeigned gratitude to Dr. Richard
A. Yost for allowing me to pursue my own research ideas (no matter how bad)
over the last five years. It has been a most satisfying experience personally and
professionally. I also would like to thank the following members of my committee:
Dr. John Gander for his insightful comments on carbohydrate chemistry; Dr. Jim
Winefordner for the endless jokes; Dr. Bob Kennedy for the career discussions on
academic life; Dr. John Eyler for convincing me that a modified ring electrode
would actually work; and Dr. Dave Powell for those grueling 10 mile Sunday runs
and brutal Tuesday night track workouts. My survival here was also due in no
small part to Jeanne Karably, Susan Ciccarone, and Donna Balkcom, all of whom
pointed me in the right direction, provided me with the right paperwork, and told
me where to be in order to graduate.
Financial support for this research came from a variety of sources including
the University of Florida Division of Sponsored Research, the Office of Naval
Research, the ACS Analytical Division Fellowship (sponsored by The Procter &
Gamble Company), and a Dissertation Fellowship from the College of Liberal Arts
and Sciences at the University of Florida.
in

A great deal of credit for the success of this project goes to Matt Booth
who has been my collaborator, co-worker, and good friend; from San Jose to
Gainesville, Matt always had an idea, a beer, or a word of encouragement that
kept me going. This dissertation also would not have been possible without the
Herculean design and machining effort put forth by Joe Shalosky; he is truly an
artist. In addition, I would like to acknowledge Scott Quarmby for assistance with
electronic design of the rf circuitry and Stephen Bou for the set-up/operation of
the pulsed C02 laser.
My Yost Group" experience was made possible by a whole cast of
characters including Uli Bernier, Tim Griffin, Tracie Williams, Shannan Carlson,
and John Laycock, just to name a few. To the "Night Shift of the Yost Lab,"
Nathan Yates, Don Eades, Jodie Johnson, and Brad Coopersmith, I am truly
grateful for the late-night beers at the Salty Dog and the "Hut" experience, which
made writing this dissertation a much less painful process.
I wish to thank my parents, James L and Vivian P. Stephenson, for
instilling in me the value of education, hard work, and humility. Their
unconditional support during the last five years has been invaluable.
Lastly, I wish to thank my wife Tracy (a.k.a Chica) for her love, support, and
understanding. She organized the weekend getaways, the La Chua trail walks,
B.E. breaks, and parties which added spice to our lives and memories to hang
on to.
IV

TABLE OF CONTENTS
ACKNOWLEDGMENTS ¡
ABSTRACT vi
CHAPTERS
1 INTRODUCTION 1
The Quadrupole Ion Trap Mass Spectrometer 3
History 3
Theoretical Aspects of Ion Motion 7
Ion Activation 17
General Operation 21
Electrospray Ionization 28
Basic Principles of Ion Formation 30
Molecular Weight Determination 35
Photodissociation 36
Infrared Multiple Photon Dissociation (IRMPD) 38
The Photon Absorption Process 41
2 FUNDAMENTAL INVESTIGATIONS OF IRMPD IN THE
QUADRUPOLE ION TRAP 47
Instrumentation 47
Experimental Design 48
Ion Trap Operation without Helium Buffer Gas 57
The Multi-Pass Ring Electrode 63
Effects of Ion Motion on Photodissociation
Efficiency 70
Dipolar Excitation 72
Quadrupolar Excitation 82
The Photon Absorption Process 93
IRMPD Kinetics 97
Consecutive Reactions 100
Buffer Gas Effects 116
v

Wavelength Dependence/Infrared Spectroscopy of
Gas-Phase Ions 120
3 THE RF-ONLY OCTOPOLE ION TRANSMISSION GUIDE 129
Ion Behavior in Electromagnetic Fields 131
Octopole Electrode Arrangement 138
Octopole Reid Potentials 141
Electric Field Strength Calculations 145
Electrode Geometry 149
Equations of Motion 150
Octopole Design Considerations 155
Effective Trapping Potential 157
Assembly and Construction 161
Factors Affecting Ion Transmission 169
Initial Entry Angle Conditions 172
RF Amplitude 180
RF Frequency 183
Kinetic Energy 189
Collisional Focusing 194
4 ELECTROSPRAY/ION TRAP INSTRUMENTATION: DESIGN
AND OPERATION 196
General Overview 196
Instrument Design 196
Vacuum Manifold and Pumping System 197
Electrospray Ion Source 204
RF-Only Octopole 210
Analyzer Assembly 226
Detector Assembly 243
Photodissociation Set-Up 246
System Interconnections 254
Instrument Characterization 258
Ion Trap High Mass Theory/Operation 263
Basic ESI Operation 270
Octopole RF Level 282
Octopole Offset 287
Ion Gate Lens 291
RF Level/lon Injection 294
Ion Isolation 303
Collision-induced Dissociation (CID) 309
Negative Ion Mode 322
VI

5 PHOTODISSOCIATION OF BIOLOGICALLY IMPORTANT
MOLECULES: PROTEINS, CARBOHYDRATES,
AND OLIGONUCLEOTIDES 328
General Overview of Structural Elucidation 328
Peptides and Proteins 338
Human Angiotensin I 338
Gramicidin D 342
Carbohydrates and Oligosaccharides 343
Monosaccharide Cleavage 348
Raffinose 363
Stachyose 374
Oligonucleotides 377
RNA Dimers 382
Carbohydrate Antibiotics 389
Macrolide Antibiotics Erythromycin 390
6 CONCLUSIONS AND FUTURE WORK 399
REFERENCE LIST 405
BIOGRAPHICAL SKETCH 427
vii

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
A NOVEL ELECTROSPRAY ION TRAP MASS SPECTROMETER FOR
PHOTODISSOCIATION OF BIOLOGICAL MOLECULES
By
James L. Stephenson, Jr.
December, 1995
Chairperson: Dr. Richard A. Yost
Major Department: Chemistry
The combined techniques of photodissociation and mass spectrometry
have been used extensively to study the fundamental aspects of gas-phase ion
chemistry. Within the last several years, a great deal of interest has been shown
in employing photodissociation as an analytical tool for the structural elucidation
of biological molecules, due in part to the limitations of traditional tandem mass
spectrometrictechniques (e.g. collision-induced dissociation) in providing relevant
structural information. This dissertation presents the design and characterization
of an electrospray ion trap mass spectrometer, capable of performing
photodissociation experiments on a wide range of biological molecules.
Previous photodissociation studies have been limited to fundamental
investigations due mainly to the limited photoabsorption cross-sections observed
for organic ions. In order to increase the photodissociation efficiency, three
Vlll

spherically symmetric mirrors were placed in the radial plane of the ion trap ring
electrode to increase the photoabsorption pathlength of the incident radiation.
The performance of this multipass ring electrode was characterized by conducting
experiments on ion motion, kinetics, consecutive reactions, spectroscopy and
buffer gas pressure to examine their effects on photodissociation efficiency of a
series of model compounds.
Next, a novel ion injection system was designed and built in order to
transfer biological ions from the electrospray source to the ion trap mass
spectrometer in the most efficient manner possible. The instrumentation
consisted of a modified electrospray interface coupled to an rf-only octopole,
which transmitted ions from a high pressure region near the ion source to the
lower pressure region of the ion trap analyzer. The theory, design, and
characterization of the rf-only octopole is discussed, along with an analysis of the
ion transmission properties of the device.
Finally, a detailed investigation into the photodissociation of ions from
carbohydrates, peptides/proteins and oligonucleotides is presented. Sensitivity
considerations, fragmentation patterns, and comparison studies to collisional
activation are discussed. An analysis of the type of structural information
obtained from photodissociation showed its applicability in solving biological
problems.
IX

CHAPTER 1
INTRODUCTION
Over the last decade, perhaps the most important advancement in
quadrupole ion trap mass spectrometry has been that of tandem mass
spectrometry or MS" for structural elucidation of organic ions. The most
frequently used method for the activation of these ions has been collisional
activation, commonly known as collision-induced dissociation (CID).1 The major
factors that contribute to the success of CID experiments in the quadrupole ion
trap mass spectrometer (QITMS) include the ability to perform tandem-in-time as
opposed to tandem-in-space MS/MS experiments, the efficient conversion of
parent ions to product ions (typically 10-50%), and, most importantly, the high
collision cross sectional area observed for CID (on the order of 10 to 200 A2).1-3
These advantages arise in part because in the quadrupole ion trap, uniquely
amongst tandem mass spectrometers, kinetic energy is imparted to the parent
ions only between collisions.
To date, the majority of the published applications using the QITMS for
structural elucidation has employed collisional-activation as the method of choice
for ion activation.4,5 More recently, the attention of many researchers has focused
on the fundamental understanding of the collisional-activation process in the
QITMS.6 Even though collisional-activation cannot provide all the answers in
1

2
tandem (MS/MS") mass spectrometry, there has only been a limited effort to
investigate alternative techniques for ion activation in trapping instruments
(including ion cyclotron resonance and quadrupole ion trap mass spectrometers).
This dissertation presents an alternative method for the activation of
polyatomic ions in the gas phase, that of photon absorption or what is frequently
called photo-induced dissociation (PID). Photodissociation has been used
extensively by physical chemists to study fundamental properties of gas-phase
ions. The combination of mass spectrometry (employing both ion trap and ion
cyclotron resonance instruments) and photodissociation has been used
successfully to investigate the chemical kinetics, reactivity, and spectroscopy of
various ionic species. The long storage times and instrumental configuration of
trapping instruments are ideally suited for photodissociation experiments. Some
advantages of trapping instruments include the measurement of photon-induced
ion decay as a function of laser irradiance time, the use of the multiphoton
absorption processes to study fragmentation, and the use of the
photodissociation spectrum as a fingerprint for determination of isomeric ion
structures.
This introductory chapter begins with the relevant history of the quadrupole
ion trap mass spectrometer, followed by a brief discussion of ion motion, the
principles of ion activation, and general QITMS operation procedures. Since the
technique of electrospray ionization (ESI) is used to generate gas-phase ions for
the biological studies presented in this dissertation, an introduction to the basic

3
principles of ESI also is included. The chapter concludes with a discussion of
photodissociation that covers the basics of the photon absorption mechanism and
provides an introduction to the infrared multiple photon dissociation (IRMPD)
process.
The Quadrupole Ion Trap Mass Spectrometer
History
The origin of the quadrupole ion trap mass spectrometer dates back to the
original patent of Paul and Steinwedel, which disclosed the operation of the
quadrupole mass filter and the quadrupole ion trap.7 The first studies by Paul and
coworkers centered on the use of the device for ion storage over long periods of
time.8 Early detection procedures measured the power absorption of stored ions,
which utilized an rf voltage applied to both endcaps. In 1959, the original mass
selective detection studies were performed by Fischer, who demonstrated the
ability of the device to obtain unit mass resolution for a series of krypton
isotopes.9 In the early to mid sixties, spectroscopic studies of ions by Dehmelt
and Majors showed that high resolution studies were possible for all ground state,
metastable, atomic, and molecular ions.10-12 The first mass discrimination
experiments were performed in 1968 by Dawson and Whetten.13 This led to the
first use of the quadrupole ion trap as a true mass spectrometer. These
experiments utilized an external detector (electron multiplier) for the detection of
ions ejected through holes in the endcap electrodes. In 1971, it was discovered

4
that a combination of both rf and dc voltages applied to the ring electrode
produced trapping conditions that were favorable for mass-selective storage.14
The storage conditions used in these experiments were defined by the upper and
lower apices of the Mathieu stability diagram.14
The combination of a quadrupole ion trap (used as an ion source) and a
quadrupole mass filter was used extensively by Todd, Lawson, and Bonner for the
analysis of ejected ions from the ion trap, the general characterization/behavior
of trapped ions, chemical ionization, and ion molecule kinetics.15-17 This hybrid
instrument, called the QUISTOR (or QUadrupole Ion STORe), was operated with
the endcaps held at ground potential and a combination of rf and dc voltages
applied to the ring electrode. An electron gate was used to pulse electrons into
the center volume of the ion trap where ionization of the neutral sample gas(es)
occurred. The detection event consisted of a dc pulse applied to one or both
endcap electrodes, which extracted ions out towards the quadrupole mass filter.
The late 1970s also saw the development of mass-selective isolation by Fulford
and March.18 Here, the ion of interest was moved to the apex point in the stability
diagram where all ions of smaller m/z were unstable in the axial z-direction and
all ions of higher m/z were unstable in the radial r-direction. In 1979, Fulford et
al. performed the first experiments that involved resonance excitation at a given
ion populations unique frequency of motion. This caused either ion ejection from
the trap or collision-induced dissociation of the ions of interest.19

5
In the early 1980s, a landmark series of papers by Hughes, March, and
Young first demonstrated the use of IR photodissociation in a quadrupole ion
trap.20"22 The technique of infrared multiple photon dissociation was used for
studying the gas-phase ion chemistry of the proton-bound dimer of 2-propanol.
Other early IRMPD experiments focused on wavelength dependence, collisional
cooling, dissociation efficiency, and analyte pressure dependence of various gas-
phase systems.23 A more detailed discussion of the instrumentation, experimental
parameters, and theoretical aspects of IRMPD in the quadrupole ion trap appears
in the photodissociation section of this chapter.
Perhaps the most important development in quadrupole ion trap mass
spectrometry was the mass selective instability scan developed by Stafford et
al.24,25 This mode of operation was radically different from previous modes in that
the rf amplitude applied to the ring electrode was ramped linearly with respect to
time. This enabled ions of increasingly higher mass-to-charge ratio (m/z) values
to be sequentially ejected from the ion trap. A mass spectrum could then be
recorded as a function of the time it takes ions of various m/z to be ejected to the
detector. Furthermore, by pressurizing the ion trap analyzer with approximately
10"3 torr of a light buffer gas (helium or hydrogen), drastic improvements in mass
resolution, sensitivity, and dynamic range were obtained for ion trap operation.26
The mechanism of action for these improvements involved the collision of ions
with the relatively slow and much less massive background gas atoms or

6
molecules. These collisions caused viscous damping of ionic motion, thereby
focusing the ion cloud to the center of the trap.
These improvements led to the first commercially available ion trap detector
(ITD 700) developed by Finnigan MAT. This product was designed as a low cost
benchtop GC/MS detector, which gained popularity for trace level environmental
and clinical analysis. Further improvements in dynamic range were obtained with
automatic gain control (AGC), which regulated the number of ions stored in the
trap as a function of sample concentration.27 This limited space charge effects
and ion-molecule reactions typically seen with constant ionization time
experiments.
The first fully functional research-grade ion trap mass spectrometer was
developed by Kelley et al. in 1985.28,29 The Finnigan MAT Ion Trap Mass
Spectrometer (ITMS) possessed a wide range of experimental capabilities
including electron ionization (El), chemical ionization (Cl), mass isolation (e.g.,
apex isolation), tandem MS", and user-defined software. With the advent of
resonance ejection (axial modulation) in 1988, further improvements in resolution
and dynamic range were achieved.30 The technique of resonance ejection was
also responsible for the extension of the mass range of the quadrupole ion trap
to well beyond m/z 50,000.
Rapid growth in the field of ion trap mass spectrometry over the last six to
seven years has been driven by the coupling of external ion sources to the
device. Some of the external ion sources coupled to the ITMS include fast atom

7
bombardment (FAB)31, glow discharge (GD)32'33, electron and chemical ionization
(El/Cl)34, and electrospray ionization (ESI).35 The discovery of high resolution ion
trap mass spectrometry by Schwartz et al., combined with advances in high mass
analysis published previously, has enabled ion trap mass spectrometry to take a
lead role in the analysis of biomolecules.3637 The advances discussed in this
dissertation address many of the issues currently limiting ion trap mass
spectrometry in the analysis of biomolecules. Other recent advances in the field
include the use of broadband waveforms (stored waveform inverse Fourier
transforms, SWIFT) for MS/MS analysis and for mass isolation.38'39 In addition,
alternative techniques to scanning and detection of ions from those traditionally
used over the last 10 years have been developed.40"13
Several reviews published over the last few years cover many of the
aforementioned developments in greater detail.44-46 Several books encompassing
ion trap mass spectrometry as well as biological mass spectrometry, cover a
range of topics from basic instrumental principles to applications development.47-50
Theoretical Aspects of Ion Motion
A fundamental understanding of ion motion in the quadrupole ion trap is
important in evaluating various ion activation techniques (collisional activation or
surface-induced dissociation, frequently called SID). Some of these evaluation
criteria include dissociation efficiencies, understanding ion-neutral collisions, and
collisions of ions with surfaces. Therefore, a fundamental comprehension of ion

8
motion is important for describing the conversion of an ions kinetic energy (for
CID and SID) of motion into the vibrational energy needed to accomplish
fragmentation in an MS/MS experiment. For efficient photodissociation, the
amount of time required for sufficient absorption of a photon(s) is directly related
to the overlap of the ion trajectory with the incident radiation. Consequently, it is
important to grasp the basic concepts of ion motion and how they might be
applied to the interaction between ions and light.
The motion of an ion in a quadrupole field can be described
mathematically by the Mathieu equation, a second-order linear differential
equation originally used to characterize the vibrational motion of a stretched skin
or membrane.51 The derivation begins by assuming the presence of an ideal
quadrupolar field (no space charge effects due to other ions), defined by 0, the
potential at any point (x,y,z) in that field:
4> = -^Ux2+0y2+Yz2) (1-1)
ro
where 0O is the applied electric potential, A, a, and y are a weighting constants
in the x, y and z directions respectively, and r0 is the inscribed radius of the ring
electrode. Any oscillatory and dc potentials applied to the ring electrode can be
represented mathematically in the form:

9
4>0 = U-V0_p cost (1-2)
where Q is the angular frequency of the rf trapping field applied to the ring
electrode (in rad s'1, which is equal to 2/rf where f is the frequency in Hz), U is the
applied dc voltage, is the zero-to-peak amplitude of the rf voltage, and t is the
time variable.23
For any quadrupolar device, the field is uncoupled in the three coordinate
directions, so that the forces acting on an ion are independent of one another.
For this condition to hold, equation 1-1 must satisfy the LaPlace condition where
the field strength at the center of the ion trap must equal zero. Therefore, when
the potential 0O is applied to the ring electrode (xy plane) and the two hyperbolic
endcaps are held at ground potential, the potential at any point 0 within the
device is represented by equation 1 -1.23
Because the geometry of the quadrupole ion trap has cylindrical symmetry,
the x and y components are combined to give a single radial component defined
by x2+y2=r2. The orientation in space of the ring electrode and two hyperbolic
endcaps needed to define a quadrupolar trapping field is r2=2z02, where r0 is the
radius of the ring electrode and z is the center-to-endcap distance.52 The
hyperbolic shape of the electrodes in this geometrical configuration can be seen
in figure 1-1. The equations defining the hyperbolic shape of the ring electrode
and endcap electrodes are given as:

Figure 1-1:
The geometric configuration of the quadrupole ion trap. Shown are the three hyperbolic electrodes
which comprise the analyzer. The center-to-endcap and the center-to-ring distances are indicated
by z0 and r0 respectively.

Filament Endcap
Ring Electrode
Exit Endcap

12
4(r2-2z2H 0-3)
To
!~{r2-2z2) = -1 (1-4)
2Zo2
The equations of motion for an ion in both the r- and z-directions can be
derived from the forces exerted on the ion independently in each direction. By
applying Newtons second law of motion, substituting equation 1-2 for 0O in
equation 1-1, applying the condition x2+y2=r2, and differentiating, the equations
of motion in the r- and z-direction are obtained:
d2r
dt2
2e
2mr02
(U-VcosQt)r=0
(1-5)
+ (U-VcosQt)z = 0 (1-6)
dt2 2mr02
where m is the mass of the ion of interest and t is the time variable. These
equations are examples of the Mathieu equations, whose solutions have been
studied extensively.23 The general form of the Mathieu equation can be expressed
as:
4%*(a-2quCos25)u = 0 (1-7)
at
where u represents r or z, and f=Qt/2. By performing a series of operations and
substitutions, the parameters au and qu can be evaluated for the quadrupole ion

13
trap. These parameters determine whether or not ion motion is stable or
unstable. For motion in both the r and z direction the values of a^ and qrz
become:
az = -2ar=
-8eU
m(ro+2z02)s
Qz = 2qr=
4eV
(1-8)
(1-9)
m(ro+2Zo )2
The Mathieu stability parameters au and qu are directly proportional to the applied
dc and rf voltages on the ring electrode. Since a and qu are inversely
proportional to m/z, the motion of all ions can be expressed in terms of these
Mathieu stability parameters.23
In equations 1-8 and 1-9, the relationship the between axial (zj and radial
(r0) dimensions of the ion trap are defined theoretically by z0=r0/21/2. However, all
commercially manufactured ion traps to date posses a stretched geometry in the
z-direction. Therefore, instead of a theoretically defined Zq distance of 0.707 cm,
the ion trap is stretched axially 10.7% with a new value of 0.783 cm.53 In a
previous report by Johnson et al.54, the general form of equations 1-8 and 1-9
(also known as the Knight equations) were found to be a good approximation for
experimental measurements to determine qr and a,z values made with the
stretched ion trap geometry.
The stability diagram, which specifies stable ion trajectories, can then be
defined as the intersection of the solutions to the Mathieu equation in the z-

14
direction with those in the r-direction. In this overlap region, ions are stable in
both the z and r directions and can be stored in an ion trap. The overlap region
which is closest to the a, q origin is of the most practical significance, since the
voltages are less and thus the electronics needed to apply the appropriate rf and
dc voltages to the ring electrode can be readily designed. Figure 1 -2 shows the
stability diagram for the three-dimensional quadrupole ion trap.23 The ordinate
and abscissa are expressed in terms of the dimensionless quantities a and q
respectively.
Ions within the stability region exist in a pseudopotential well, with a depth
of about 1 eV near the origin and 10 eV near the right hand (z) stability edge.
The lines drawn across the stability diagram are called iso-6, lines which describe
detailed trajectories of ions at that particular point. Iso-6 lines essentially
determine the frequency of ion motion; they are found in the general solution to
the Mathieu equation:
OO CO
u(0=A£ C2ncos(2n + 6)£+ B £ C2nsin(2n+6K <110)
n=- n=-
where the coefficients are the amplitude components of oscillation and the
(2n+6) terms represent the frequency components of ion motion. These
quantities can be calculated for the corresponding values of a and q using
recurrence relations/continued fractions.55 The relationship between ion frequency
(ft/n J and 6 can be defined as follows:

Figure 1-2: Mathieu stability diagram for the quadrupole ion trap. The
dimensionless quantity qz is directly proportional to the applied
rf voltage, while the dimensionless quantity is directly
proportional to the applied dc voltage. The 62 lines are a direct
relation to the secular frequency of motion for a given ion.
Adapted from reference 23.


17
where 0<6U<1 and n=0, 1, 2, 3 Consequently, the main secular frequency
(n=0) is calculated as 6UC1/2.23
As mentioned earlier, the fundamental driving force behind the derivation
of the field equations for the quadrupole ion trap is the fact the motions in the r-
and z-directions are independent of one another. This concept enabled Paul and
others to describe the motion of ions with a simplified mathematical treatise.55 As
a better understanding (both theoretical and experimental) of ion motion in the
quadrupole ion trap was gained, it was realized that coupled motion between the
r- and z-directions was indeed possible. In 1989, March et al. published a
theoretical derivation of this coupled motion (discussed later in chapter 2) under
resonance excitation conditions.57 This dissertation presents a novel way to
experimentally confirm the presence of coupled motion in the quadrupole ion trap,
employing the technique of photodissociation.
Ion Activation
The most common result of ion activation of any polyatomic species in the
gas phase is unimolecular dissociation. Unimolecular dissociation in trapping
instruments typically occurs from a stable ion which has been "activated" and
made unstable. The fragments observed from this dissociation depend on the
structure of the parent ion and can thus provide structural information on the

18
parent species itself. The ion activation techniques which drive the unimolecular
dissociation process are usually evaluated by a series of four criteria: (1) energy
deposition in the ion of interest; (2) energy distribution or bandwidth of the
deposition process; (3) variability of the deposited energy; and (4) reaction cross-
section for the ion activation technique.1
Several different types of reactions in tandem mass spectrometry (MS/MS)
can be used to bring about the ion activation process. Collisional activation is the
most widely used method for ion activation, operating at translational energies
ranging from approximately 10 eV to about the 10 KeV range. Fragmentation of
polyatomic species via collisional activation is called collision-induced dissociation
(CID).58"61 Another increasingly popular method of ion activation is surface-
induced dissociation (SID). Energy transfer in the SID process can be quite large
(on the order of about 8 eV), where the amount of energy transferred is controlled
by the translational energy of the incident ions.62-63 A third method for the
activation of polyatomic ions involves the interaction of the parent ion with a beam
of electrons, a process called electron excitation. Product ion spectra generated
by electron excitation are characterized by broad energy distributions with an
upper limit approaching the energy of the electrons used for ion activation.6^66
The final method typically employed for ion activation involves the absorption of
photons from a light source (e.g., laser irradiation), with the resulting
fragmentation process referred to as photodissociation.67'69 Photodissociation
holds several advantages over the other ion activation techniques, including the

19
ability to impart a wide range of well defined energies to the parent ion of interest.
The ability to finely control the energy deposition process with photon absorption,
and thus the fragmentation process, is the central focus of this dissertation.
It is beyond the scope of this dissertation to discuss in detail the
fundamental aspects of the ion activation techniques other than photodissociation
(e.g., collisional activation, surface-induced dissociation, and electron excitation).
Instead, a general overview of the unimolecular dissociation process is presented
in order to give the reader a better understanding of the results achieved with ion
activation (e.g., fragmentation process).
Unimolecular reactions in the gas phase have been studied extensively and
can provide a large amount of information concerning rates of dissociation,
energy partitioning, and development of models to describe a particular gas-
phase reaction system. Mass spectrometry provides a unique environment for
evaluating gas-phase systems since reactions can take place in collision-free
conditions. In addition, both parent and product ions can be isolated and
identified with high specificity. Unimolecular dissociations have been used
successfully to explain electron ionization of simple molecules, and the theory is
now currently being adapted to help predict the products from simple collisional
activation experiments.70 This theory, termed the quasiequilibrium theory (QET),
has associated with it four basic assumptions important not only for the statistical
theory of electron ionization, but for general considerations of unimolecular
reactions in the gas phase. The assumptions of the basic theory are as follows:

20
(1) the dissociation event is long when compared to the time needed for
ionization or excitation; (2) the rate of the dissociation event is much slower than
the corresponding rate for the redistribution of the energy deposited during
ionization/activation; (3) fragmentation is the result of a series of competing and
consecutive reactions; and (4) an internal energy equilibrium is achieved where
the energy is randomized over all internal states with equal probability.1
After the initial energy deposition, the weaker bonds are preferentially
broken, thus revealing mass fragments indicative of the parent ion. This would
mean that a mass spectrum or MS/MS spectrum depends only on the amount of
energy deposited into the ion and not on how that energy was deposited. For a
typical MS/MS experiment, energy can be deposited during the ionization event
and/or in a subsequent ion activation reaction.1
The product ions produced in an MS/MS experiment are determined by
several factors in addition to the amount of internal energy deposited into the
system. These include both the time frame for the reaction to occur and the
individual microscopic rate constants for each dissociation pathway. Since there
exists an inverse relationship between ion lifetimes and the rate of dissociation of
an ion, examining the plot of mass spectra as a function of ion lifetime can yield
information similar to that obtained with traditional breakdown curves. Early
studies by Morgan et al. have reported that rate constants in excess of 2 x1 O'7 s1
were required for daughter ions to be dissociated in the collision cell of a reverse
geometry (magnetic sector followed by electrostatic sector, BE) double-focusing

21
instrument.71 By applying the basic ideas of transition state theory to individual
microscopic rate constants, the expression for the intemal-energy-dependent rate
constant becomes:
1 W*(E-E0) 1 W*(E-E0)
h dW(E)/dE h p(E)
(1-12)
where W(E) is the number of energy states of the ion with energy less than or
equal to E, p is the energy level density (dW/dE), W* is the same as W except
that the transition state is assumed for the ion, E0 is the ion activation energy and
therefore (E-E0) can be defined as the internal energy of the ion. A schematic
representation of the relevant energetics72 is shown in figure 1 -3.
General Operation
The quadrupole ion trap instrumentation used and developed for this
dissertation was based on an early design by Kelley et al. of Finnigan MAT (San
Jose, CA).27'28 A schematic of the Finnigan MAT ITMS is shown in figure 1-4.
In the normal mode of operation, electrons are gated into the central volume of
the analyzer (through a 1/16" hole) by pulsing a gate electrode to +180 V for a
specified ionization time period. The ions formed by electron ionization are
trapped by an rf voltage applied to the ring electrode of the analyzer (while the
endcap electrodes are held at ground potential). The detection event is
accomplished by ramping the amplitude of the rf voltage applied to the ring

Figure 1-3: Schematic representation of the energy terms relating to the ionization (or dissociation) of a
polyatomic molecule: l2, Ionization energy of the molecule P; Elh, thermal excitation of a molecule
P prior to ionization; E,fd, energy transferred to P by the incident electron, photon, or collision; E,
resulting internal energy in the ion; q1t reaction coordinate for P+ -* A+ + B; >c(, the activation
energy for P+ A+ + B; e^,, activation energy for the reverse reaction A+ + B -* P+; AH0. AH298,
standard enthalpy changes for P A+ + B + e at 0 K and 298 K, respectively; D0=AH0, standard
enthalpy change for the dissociation P -* A + B; lz(A), ionization energy of the fragment A. Only
two of the many vibrational degrees of freedom of P, P+, P+*, and A+ are represented; the potential
surfaces of A and B are omitted. The spacing of vibrational levels is greatly exaggerated relative
to D0 and eac{: \z is typically 2-4 times D0. The level (a) represents the energy of the dissociated
neutral system, A + B, with the species in their ground states. Most of these energy terms refer
to the differences in energy between specific levels; the exceptions are AH298 and AH0, which are
the usual thermodynamic quantities. Adapted from reference 72.

23

Schematic diagram of the Finnigan MAT Ion Trap Mass Spectrometer (ITMS). The drive
frequency applied to the ring electrode is 1.1 MHz. The supplementary rf generator can be used
to apply both dipolar (as illustrated in the figure) and quadrupolar excitation signals to the endcap
electrodes.
Figure 1-4:

Filament
End Cap
Electron Multiplier Detector
To Preamplifier
(Ion Signal)
Amplifier and
RF Generator,
Fundamental
RF Voltage
£
Scan Acquisition
Processor
(Computer)
£
- ¡
Ring Electrode ^
End Cap
V
JT
Amplifier and
RF Generator,
Supplementary
RF Voltage

26
electrode, thus causing ions of successively higher m/z ratios to be ejected from
the ion trap to the detector (mass selective instability).24,25
Increased resolution and sensitivity can be obtained when axial modulation
is used in conjunction with the mass selective instability scan. Axial modulation
is accomplished by applying an auxiliary ac voltage to the endcap electrodes 180
out of phase.30 This dipolar resonance excitation frequency is set to
approximately 525 kHz just below the 6Z=1 boundary (at 550 kHz) of the stability
diagram. The improved peak shape and resolution obtained is due to presence
of a uniform electric field in the z-direction, which reduces ion shielding and space
charge effects normally seen with the mass selective instability scan where the
field strength at the center of the trap is zero.23,30 This technique has also been
used successfully to extend the mass range of the quadrupole ion trap.73 A
discussion of mass range extension will follow in chapter four of this dissertation.
Other applications for the technique of resonance excitation include notch
filtering27, high resolution36,74, ion isolation74, and collision-induced dissociation75
studies.
Ion isolation can be accomplished using a variety of methods. A
combination of rf and dc voltages applied to the ring electrode can be used for
either apex or two-step isolation.76-79 These two methods utilize the edges of the
stability diagram to preferentially isolate the ion of interest. Resonance excitation
using forward and reverse scans is also used for ion isolation.74 By applying a
high frequency dipolar excitation signal to the endcap electrodes and increasing

27
the rf amplitude applied to the ring electrode, all m/z values below a specific ion
of interest are ejected from the trap. Next, by lowering the dipolar frequency and
decreasing the rf amplitude, masses of higher m/z values are ejected, thus storing
the desired m/z ion or range of ions.
Broadband techniques have also been employed for ion isolation studies.
The various methods available include stored waveform inverse Fourier transform
(SWIFT)39, the use of multiple single discrete frequencies36 80 81, and random
noise.82 Perhaps the most effective means for ejecting a large range of ions is
that of SWIFT or random noise. These techniques have proven to be very
successful, since a large number of signals (of various frequencies) can be
applied simultaneously to the endcap electrodes. In essence, sensitivity is
increased because the ion trap is selectively filled with the ion(s) of interest and
not with unwanted matrix or background ions.
The ability to implement user-defined scanning strategies through the Ion
Catcher Mass Spectrometer (ICMS) software developed in this laboratory was
critical for the success of the experiments in this dissertation.83 Specifically, the
development of ICMS software allowed for the computer control (TTL signal) of
external devices such as a pulsed C02 laser, a continuous wave (cw) C02 laser,
and a pulsed-valve controller.83 The FORTH programming option permitted the
design of intricate experiments involving resonant excitation frequency, laser
control, pulsed-valve control, and ion cool time (vibrational relaxation). In

28
addition, scan table times could be adjusted up to 1 s, to facilitate various
photodissociation experiments.
Electrosprav Ionization
Perhaps the most successful liquid chromatography/mass spectrometry
(LC/MS) interface to date is that of the electrospray (ESI) ion source. Over the
last seven to eight years, the volumes of research in the field have produced an
abundance of methodologies and applications for the identification and
quantitation of biological species. Some of the reasons for the rapid growth rate
of ESI include the speed with which commercial instrumentation was developed,
the ability to couple ESI with microscale capillary separation techniques, and the
ability to perform tandem mass spectrometry on multiply charged species.84
The first use of electrospray as an ionization technique for biological
species was reported over 25 years ago by Dole and coworkers who defined
many of the operational parameters used today.85,86 Doles original detection
scheme involved ion mobility and ion retardation methods, since an appropriate
mass spectrometer was not available. Dole was also the first to report the
multiple charging effect seen with large biological species.87 The first reports of
ESI combined with mass spectrometry were by Fenn and Aleksandrov et al. in
1984.88,89 Aleksandrovs group was also the first to interface an LC to an
electrospray ionization source connected to a magnetic sector mass
spectrometer.90

29
The most important aspect of ESI is that of multiple charging. In a
landmark paper published in 1988, Fenn and coworkers reported as many as 45
charges attached to proteins of molecular weight 40,000 da.91 This work was
quickly duplicated and extended by both Smith and Covey.9293 The multiple
charging effect permitted the analysis of high molecular weight biological species
using existing quadrupole instrumentation, due to the lower mass-to-charge ratios
that can be obtained with multiple charging. Another advantage of the multiple
charging process is that a more accurate molecular weight determination can be
obtained from a distribution of multiply charged peaks.47 In addition, tandem
mass spectrometry has been shown to be useful for the dissociation of multiply
charged species. The most frequently used instrument for structural elucidation
studies of electrosprayed ions is the triple quadrupole mass spectrometer. One
of the most notable applications of tandem mass spectrometry and ESI ionization
has been the work of Don Hunt and colleagues.94"97 The Hunt group successfully
used only femtomoles of material to identify histocompatibility complex-bound
peptides. This work ultimately led to the identification of a peptide with high
binding affinity for cytotoxic killer T cells.94
The use of tandem mass spectrometry (triple quadrupole) in conjunction
with ESI is not the only active area of research in the ESI field. Other areas of
great interest include mechanistic investigations98"100, surface-induced
dissociation101,102, ion-molecule reactions103, non-covalent interactions104,105,

30
collisional activation106-117, and photodissociation.111,112 Several recent reviews on
ESI coupled with a variety of applications can be found in the literature.84,113-115
The purpose of this section is to provide the reader with a general
understanding of the electrospray process and examine some of the more recent
advances in the field. A brief discussion of charged droplet formation is followed
by a section on the chemistry of multiply charged ions to give the reader
knowledge of the basic physical principles of ESI.
Basic Principles of Ion Formation
The production of ions in electrospray mass spectrometry is comprised of
two steps: the production of highly charged droplets with their dispersal at
atmospheric pressure and the evaporation of these droplets to produce multiply
charged ions.115 The production of highly charged droplets begins with a small
flow of liquid through a simple metal capillary (stainless steel needle) which
operates at an elevated electric potential relative to a counter electrode. The
potential of this electric field on the capillary is typically between 3 and 6 kV
relative to the counter electrode placed about 0.3 to 2 cm away. The counter
electrode has an orifice where charged clusters, ions, or droplets are passed into
the mass spectrometer. Charge accumulations occur at the liquid surface due to
the application of the electric field. Therefore, flow rate, solution resistivity, and
surface tension are important variables in droplet production. The bias of the
capillary needle relative to the counter electrode can be selected to produce

31
either positively or negatively charged droplets. The electric potential essentially
disrupts the flow of liquid from the capillary tip resulting in production of the
charged droplets. To aid in the desolvation process, the droplets are typically
entrained in a nebulization gas such as nitrogen, oxygen, or even SF6. Oxygen
and SF6 act as electron scavengers for negative ion electrospray or when
spraying pure water solutions.84
Solution resistivities on the order of < 10'5 Q are required for stable spray
conditions at room temperature. This corresponds to a solution conductivity of
aqueous electrolytes of approximately 1CT* normal (N), where normality is defined
as one gram molecular weight of the dissolved substance divided by the
hydrogen equivalent of the substance per liter of solution. The higher the surface
tension (the higher the aqueous fraction in the solution), the higher the threshold
voltage needed for the onset of the electrospray process. The relationships
between the electrospray onset voltage V0, the surface tension T the needle or
capillary radius r, and the needle to counter electrode distance h can be shown
to approximate:
V0(Tsr)1sln (1-13)
r
Because the droplet size decreases with increasing solution conductivity, lower
flow rates are required for solutions which are highly conductive. The
dependence of solution conductivity (a) on ESI ion current (I) is relatively weak
(l
32
The presence of a dry nebulization gas at approximately 80 C and the use
of a heated capillary (counter electrode) can aid significantly in the desolvation
process of the charged droplet. The formation of multiply charged ions from the
charged droplet is an area of great debate. The droplets eventually reach a point
where the repulsive coulombic forces approach those of the cohesive forces
(surface tension) that hold the droplet together. At this point the droplet may form
an ion by one of two proposed mechanisms: droplet fission at the Raleigh limit
or direct field evaporation of the droplet.116,117 A discussion of the relevant
thermodynamics of these two processes is beyond the scope of this dissertation.
However, it is the belief of this author that the direct evaporation theory as set
forth by Iribarne and Thomson applies to most situations.118,119
Once the charged droplets/ions pass through the counter electrode region,
they proceed through a differentially pumped region containing one or two
skimmer cones. A schematic diagram of the typical electrospray source
developed by Fenn and coworkers can be seen in figure 1-5, showing the
electrospray needle, counter electrode, differentially pumped region, skimmer
cones, and dc lens injection system.120 Details of the various ion injection
systems used are discussed in chapters 3 and 4, along with details of a new ion
injection system (for quadrupole ion traps) for electrosprayed ions using an rf-only
octopole beam guide.

Figure 1-5:
Typical ESI ion source used in mass spectrometry. The glass capillary can be replaced by a
heated stainless steel capillary. The skimmer and ion source lens assembly design will vary
depending upon the instrumental configuration.

Cylindrical
Electrode
Glass
Capillary
Skimmer
Ion Source
Lenses
Baffle
I
)
Turbo
Pump
Analyzer
Turbo
Pump

35
Molecular Weight Determination
Molecular weight data from ESI spectra can easily be obtained due to the
charge state distribution associated with the ionization process. Typically, the
width of the charge state distribution is approximately half that of the highest
charge state, although the effects of the various factors (e.g., pH, applied
potential) are not yet well understood.121,122 The adjacent peaks (of the multiply
charged ion distribution) in the spectrum of positively charged biopolymers
usually vary by one charge. Therefore, in order to determine molecular weight
(Mr) from an ESI spectrum (where the charge varies on adjacent peaks by the
addition or subtraction of one proton), the following expressions are used:
p^-M, + Maz1 = Mr + 1.0079z1 (1-14)
where p, is the m/z of interest and z, is the charge on p1t and Ma is assumed to
be the charge carrying species (proton). By examining another peak in the
charge distribution spectrum, another equation can be generated for a m/z higher
than the previous example (p2>p1) that is j peaks away from pr
p2(z1-j) = Mr+1.0079(z1-j)
Equations (1-14) and (1-15) can then be solved for p1 yielding:
_ j(p2-1.0079)
1 (P2-P1)
(1-15)
(1-16)
The value of the molecular weight is then calculated by evaluating zn to the
nearest whole number.91,123 Improved precision can be obtained by performing

36
the same calculation for the entire series of multiply charged ions. The
accuracies which have been reported to date for proteins and other biopolymers
over 100 kDa (molecular weight) are better than 0.005%M One of the best mass
accuracy measurements to date was recorded for myoglobin, with an observed
error of less than 1 ppm obtain using a FTICR mass spectrometer.124 A new
method of molecular weight determination developed by Hagen and Monning
uses a multiplicative correlation algorithm for processing charge distribution
data.125 The ability of the technique to accurately determine molecular weight
increases as the (M+H)+ signal is spread out over larger and larger charge state
distributions.
Photodissociation
Photo-induced dissociation is the next most frequently used method for
activation of polyatomic ions after collisional activation. The range of internal
energies present after the photon absorption process is much narrower than that
obtained with collisional energy transfer. Therefore, the usefulness of PID for the
study of ion structures is greatly enhanced. However, the reduced absorption
cross-sections observed with photodissociation (10'2 2) compared to those of
collision-induced dissociation (10 to 200 2) can limit this technique for analytical
applications. The recent availability of higher powered light sources over a wider
range of wavelengths should provide greater flexibility for photodissociation as a
routine analytical technique.1,67

37
The process of photodissociation for a positive ion can be described by the
following equation:
nhv
A+ A** P + N (1-17)
relaxation dissociation
where A+ is the ion of interest, n is the number of photons absorbed, hv is the
photon energy, A+* is the excited state, and P+ represents the product ion (with
loss of neutral N). For photodissociation to occur several prerequisites must be
met. The most important criteria include the absorption of photons with energy
hv, the existence of excited states above the dissociation threshold, a slow
relaxation rate compared to light absorption (multiphoton processes), and
dissociation rates which are fast on the time scale of the type of mass
spectrometer employed.167
The information obtained from a photodissociation experiment can address
a variety of gas-phase chemistry issues. One of the most important issues is the
difference observed in fragmentation spectra between PID and CID. The narrow
well defined energy transfer distribution step in PID typically leads to the
dissociation process via the fragmentation pathway with the lowest activation
energy (especially for visible and infrared wavelengths).68,126127
In addition, wavelength-dependent spectra can be obtained as long as the
internal energy of the ion population is above the dissociation threshold for the
wavelength of interest. The photodissociation spectrum can then be compared
with the typical absorption spectrum of the neutral molecules, provided that light

38
absorption can occur for the given structure of the ion. The information obtained
from these experiments can be used as a fingerprint in the determination of ion
structures.126,127
In trapping instruments, photo-induced ion decay can be measured as a
function of irradiance time. These data can be used effectively to distinguish
isomeric ion structures in the gas-phase. Kinetic energy release data can also be
used (below 100 mV) to add important ion fragmentation information.128
Infrared Multiple Photon Dissociation (IRMPD)
Infrared multiple photon dissociation is particularly well suited for use with
trapping instruments. The ability to trap ions for extended periods of time at low
pressures in both the ion cyclotron resonance (ICR) cell and the quadrupole ion
trap mass spectrometer allows for the sequential absorption of infrared photons
via low intensity infrared radiation. The IRMPD process was first characterized by
Beauchamp and co-workers for positive ions using an ICR mass
spectrometer.129,130 These and other studies focused primarily on the chemical
kinetics, reactivity, and spectroscopy of various ionic species. Brauman and co
workers were the first to study the vibrational relaxation of gas-phase ions and the
accompanying physics of pulsed megawatt infrared multiphoton dissociation
using a C02 laser.131 For the case of trapped negative ions, irradiation can
produce both a photodissociation and an electron photodetachment spectrum.132'
134

39
As with the ICR technique, the quadrupole ion trap is capable of storing
ions for long periods of time. The storage capability makes the quadrupole ion
trap mass spectrometer very compatible with a wide range of experiments using
light. One of the first successful uses of the ion trap in conjunction with
photodissociation involved the study of the proton-bound dimer of 2-propanol
utilizing a cw infrared laser.20 2-Propanol was chosen for study since its gas-
phase ion chemistry is well known and the formation of the protonated dimer is
easily accomplished. The instrumental configuration consisted of an ion trap
connected directly to the ion source of a quadrupole mass filter (QUISTOR mode
of operation). A detailed description of the optimization of ion trapping
characteristics for studies of ion photodissociation is a QUISTOR can be found
in March and Hughes.21
These earlier studies of the IRMPD process in the ion trap were performed
with a single-pass ring electrode design, with a 3 mm diameter hole on the center
axis of the ring electrode as the entrance aperture for the low power cw C02 laser
beam. Upon reaching the other side of the ring electrode, a portion of the beam
passed through a 0.8 mm diameter hole and through a NaCI window, where the
laser power was monitored externally. The remainder of the laser beam was
reflected by the ring electrode throughout the QUISTOR. The pressure of 2-
propanol was adjusted to 5 mPa so that photodissociation of the proton-bound
dimer, (2M+H)+, at m/z 121 could occur at an appreciable rate. At pressures
optimum for the formation of the proton-bound dimer (13 mPa), no laser-induced

40
dissociation was observed due to collisional deactivation of the vibrationally
excited proton-bound dimer.20'21 At the low pressures used in these experiments,
the dissociation rate constant kD was related to the phenomenologically defined
cross section aD and the photon flux 0 by the following equation:
kD= o4> d-18)
The highest absorption cross section for 2-propanol was found to be at a
wavenumber of 944 cm'1, with the corresponding absorption of 10 photons. The
dissociation rate constant kD was determined to be 2.2 s'1, assuming first order
dependence on photon flux.
Photodissociation experiments for the proton-bound dimer of 2-propanol
(m/z 121) showed three different photoreaction channels open for the IRMPD
process. March and Hughes give a detailed description for verification of the
various reaction pathways, the photodissociation of the various isotopic analogues
of 2-propanol, and the ion relaxation processes involved.22,23
The same experimental apparatus has also been used to investigate the
gas-phase ion chemistry of ethanethiol, 1- and 2-propanethiol, and 1-
hydroxyethanethiol.135,136 The collisionally-cooled, proton-bound dimers of
ethanethiol, 1-propanethiol, and 2-propanethiol were unaffected by laser
irradiation at 944 cm'1. However, the proton-bound dimer of 2-hydroxyethanethiol
was thought to contain a SH+S linkage which was shown to be transparent
at the same wavelength. Isomer differentiation by multiphoton dissociation of the
proton-bound dimer of propanone (m/z 117) and protonated diacetone alcohol

41
(m/z 117) has also been demonstrated.137'139 The authors were able to
differentiate the two isomeric m/z 177 ions by an aldol condensation reaction
observed with the fragmentation of protonated diacetone alcohol. This reaction
was not observed with the propanone dimer in the gas phase.
Investigation into the wavelength dependence of IRMPD photodissociation
efficiency of the proton-bound dimer of ethanol using the QUISTOR demonstrated
the observable frequency shifts of the C-0 stretch in the IR region.23140
Application of the quistor technique to the study of photodissociation rates by
varying relaxation time, buffer gas pressure, and analyte pressure has yielded
data for the proton-bound dimers of isopropanol, 2-dT-2-propanol, and ethanol.
The authors also studied the effect of collision rate on the defined
photodissociation cross section. The results obtained for the fully relaxed proton-
bound dimer population showed access to the lowest Ea pathway, thus
demonstrating the only variable observed was that of the collisional deactivation
process (corresponding to higher collision rates > 5 ms'1).23
The Photon Absorption Process
The interaction of gas-phase molecules with infrared light is currently one
of the most rapidly developing fields in chemical physics. Since the first
discovery of infrared photon absorption by ions in the gas phase, the field has
attracted the attention of researchers from a variety of disciplines.141'143 This
section is intended to give the reader a general overview of the mechanism of the

42
IRMPD process and to outline the applications of this process to large biological
ions in the gas phase.
The majority of investigations into the mechanism of the photon absorption
process have utilized pulsed or continuous wave (cw) lasers operating in the
wavelength range 9.5 to 10.5 /m. The fundamental model of photon absorption
and subsequent dissociation is based on the extensive studies of the SF6 system
at 10.6 //m.144 The basic concepts of this model also apply to the IRMPD process
of other gas-phase systems.
In the model proposed by Black et al., there are three successive regions
for photon absorption as the energy deposited in the ion increases.144 A
schematic energy diagram in figure 1-6 shows the three individual regions (I,II,
and III) corresponding to coherent multiphoton interaction, incoherent single
photon interaction, and dissociation threshold/channels. In the coherent
multiphoton interaction region (I), vibrational states are essentially discrete and the
corresponding absorption of a single IR photon is a straightforward process.
However, successive transitions to higher vibrational levels are not possible since
the vibrational spacing is now out of resonance with the monochromatic IR
photons. Therefore, in order to further excite the ion of interest other mechanisms
which compensate for these anharmonicities must be in operation.145
For any polyatomic ion the density of the vibrational levels increases with
internal energy. The rate of this increase depends upon the magnitude of the
frequencies of the individual modes and the molecular complexity of the ion.

Figure 1-6: Schematic energy level diagram as originally applied to the
IRMPD of SF6. In region I, coherent multiphoton interaction
describes one of the mechanisms (intensity-dependent power
broadening effects) whereby non-resonant absorption can take
place in a sparse region of vibrational levels. In region II, the
quasicontinuum, resonant absorption steps are always possible,
and can be treated as stepwise (incoherent) excitation processes.
Region III lies above the dissociation threshold. Adapted from
reference 145.

44
Dissociation
channels
Dissociation threshold
Incoherent
single-photon
interaction
Quasi-continuum
I
Coherent
multiphoton
interaction

45
Eventually, these vibrational levels merge to form what is termed the
quasicontinuum or the incoherent single-photon interaction region (II). Ions in
region II are characterized by their ability to undergo the resonant absorption
process for a given monochromatic laser frequency (although there may be some
structure within the quasicontinuum itself). If an ion can be excited through the
discrete vibrational levels in region I, then there always exists a path for the
absorption process to occur through the quasicontinuum. Polyatomic ions can
be excited into the quasicontinuum by a variety of mechanisms including thermal
excitation, collisional or particle excitation, exothermic chemical reactions, or
electronic excitation followed by internal conversion or forced multiphoton
excitation. For the biological ions used in this dissertation, no excitation methods
are needed to push ions into the quasicontinuum, since these species are
sufficiently large that they already exist in the quasicontinuum as a result of
internal thermal energy content.130
As the ion continues to absorb energy, ft eventually reaches a point where
it obtains enough energy to dissociate (region III). If randomization of this excess
internal energy above the dissociation threshold occurs at a much faster rate than
the dissociation process itself, then the traditional statistical theories associated
with unimolecular dissociation (the QET theory discussed previously), and that of
Rice-Ramsperger-Kassel-Marcus (RRKM theory, i.e. the application of transition-
state theory to unimolecular reactions) can be applied to the photodissociation
process.145 Applications of RRKM theory include relating the excess energy

46
absorbed above the dissociation threshold to the lifetime of the excited species,
and estimation of the energy distributions within the product ions from the IRMPD
process.

CHAPTER 2
FUNDAMENTAL INVESTIGATIONS OF IRMPD IN THE
QUADRUPOLE ION TRAP
The second chapter of this dissertation focuses on the fundamental aspects
of the photodissociation (IRMPD) process in the quadrupole ion trap mass
spectrometer. A large part of the discussion centers around the design of a novel
multipass optical arrangement for use with IRMPD. This design circumvents
previous problems of limited IR laser power, small IR absorption cross sections
for many polyatomic organic species, and the limited ion statistics of trapping and
detection of ions for IRMPD in the quadrupole ion trap. Fundamental
investigations (using model compounds formed by El/Cl) into consecutive
reactions, ion motion effects, the photon absorption process, kinetics, buffer gas
effects, and infrared spectroscopy employing a cw C02 laser are presented.14"148
This work is the basis for future applied studies addressed in chapters 4 and 5
of this dissertation.
Instrumentation
The instrumentation addressed in this section covers the experimental
design (Finnigan MAT ITMS parameters), ion trap operation without He buffer gas,
and a detailed description of the multipass optical arrangement necessary for all
47

48
subsequent experiments performed in this and all of the following dissertation
chapters. The basis for the experimental design is derived from previous studies
by Watson et al. in the ICR mass spectrometer.149 A detailed description of the
instrumentation employed has been published.147,148
Experimental Design
All experiments were performed on a Finnigan MAT (San Jose, CA) ion trap
mass spectrometer (ITMS). Except for the pulsed-valve experiments, the ion trap
was operated with no He buffer gas in the vacuum chamber. The base pressure
of the instrument was 3.5x1 O'8 torr (uncorrected), as indicated on a Bayard-Alpert
ion gauge (Granville-Phillips, Boulder, CO) mounted on the vacuum manifold.
The Teflon ring electrode spacers were found to absorb strongly at 944 cm'1
during the laser irradiation period, which desorbed both neutral and ionic species
that affected both ion storage efficiency and detection; they were therefore
removed. The vacuum manifold was equipped with a modified flange containing
a ZnSe window to pass IR radiation. The modified software used (ICMS) provided
two TTL pulses for computer control of both the cw C02 laser and the pulsed-
valve apparatus.83 The experimental arrangement can be seen in figure 2-1.
The cw C02 laser employed was an Apollo Model 570 that is line tunable
over a wavelength range of 1099-924 cm'1 and has a beam diameter of
approximately 1 cm. The maximum laser power obtainable was 50 W at 944 cm'1.
The laser power supply was modified with electronics that transformed an

Figure 2-1:
Instrumental configuration for IRMPD in the ion trap.

To Computer
cn
o

51
incoming TTL signal to the CMOS logic used by the laser. A spectrum analyzer
(Optical Engineering, Model 16-A, Santa Rosa, CA) was placed directly in-line with
the cw C02 laser beam for wavelength measurements. All laser energy
measurements (Coherent Radiation Model 410 power meter, Boulder, CO) were
taken inside the ITMS vacuum manifold to correct for beam loss at the turning
mirror surface and ZnSe window. Because of space considerations, the ion trap
analyzer was removed from the ITMS vacuum manifold for the energy
measurements. The power meter was then positioned at the exact location where
the laser beam would enter the ion trap analyzer. The laser was placed parallel
to the ITMS manifold with the beam reflected at a 90 angle by a gold coated
mirror as seen in figure 2-1. Laser beam alignment was accomplished with a He-
Ne laser placed in-line with the cw C02 laser. The analyzer of the ITMS was fitted
with a modified mounting bracket to allow for rotation of the ion trap analyzer from
its original fixed position to facilitate alignment (figure 2-2).
The pulsed-valve used in the experiments (Series 9, General Valve
Corporation, Fairfield, NJ) was mounted on the opposite flange from the ZnSe
window (see figure 2-1) and was used to pulse He into the ion trap to increase
trapping efficiency. The pulsed valve was placed 0.5 cm from the outer diameter
of the ring electrode. Its horizontal position was between the modified ring
electrode and the entrance endcap. The pulsed-valve controller (built at the
University of Florida) was controlled by an external TTL pulse generated by the
ITMS electronics. The timing diagram shown in figure 2-3 displays both the laser

Figure 2-2:
Adjustable mounting bracket for the ion trap analyzer assembly. The three slotted holes allow for
rotation of the analyzer to facilitate laser alignment with the entrance aperture on the ring electrode.

0,745"
Center nark 2,750' dlanet

Figure 2-3:
ITMS scan function and timing diagram for a typical IRMPD experiment (figure not to scale). 1,
pre-ionization/pulsed-valve on; 2, ionization-chemical self-ionization reaction; 3, two-step mass
isolation; 4, vibrational relaxation; 5, laser on; 6, laser decay; 7, acquisition.

RF
Laser
TTL 1
Pulsed
Valve
TTL 2
Resonant
Excitation
Frequency

56
control TTL pulse and the pulsed-valve TTL control pulse for a given scan
function.
Bis(2-methoxyethyl)ether (diglyme; ACS reagent grade purchased from
Fisher Scientific, Fairlawn, NJ), 3-bromo-1-propene (allyl bromide; ACS reagent
grade purchased from Aldrich, Milwaukee, Wl), 12-crown-4 ether, and 15-crown-5
ether (ACS reagent grade purchased from Sigma, St. Louis, MO) were introduced
via a Granville-Phillips (Boulder, CO) fine metering valve system directly into the
ion trap manifold. Depending on the experiment, the sample pressure ranged
from 1.1 to 3.6x1 O'7 torr. Formation of the protonated diglyme, 12-crown-4 ether,
or 15-crown-5 ether was accomplished by self chemical ionization, predominantly
due to the reaction of the low mass even-electron fragment ions from electron
ionization (El) with the neutral molecules for approximately 400 ms. The resulting
[M+H]+ ion of interest was then mass isolated by a two-step rf/dc isolation
routine.150 Next a 400 ms delay was included to allow for removal of any excess
internal energy by radiative and/or collisional cooling. The [M+H]+ ion was then
irradiated for a specified time period with the cw C02 laser. After the laser
irradiation period, a 10 ms time period was incorporated for decay of the laser
output when the high voltage was turned off. For all photodissociation efficiency
measurements, the first sequential product ion from the IRMPD of the [M+H]+
was resonantly ejected (q2 = 0.3, frequency = 118.1 kHz, amplitude = 100 mV,
values approximate) during the laser irradiation period (periods 5 and 6 in figure
2-3). This prevented the possible occurrence of any ion molecule reactions as the

57
result of the reaction of the sequential absorption product(s) with neutral sample
molecules. The remaining ions were then mass analyzed via resonance ejection
with qzeject = 0.89 and an amplitude of 1.5 V (zero to peak).26,30
Ion Trap Operation without Helium Buffer Gas
The presence of a light buffer gas (He or H2) at a relatively high pressure
(1 mtorr) has been shown to enhance resolution, sensitivity, and improve
detection limits associated with operation of the quadrupole ion trap.26
Unfortunately, the presence of a buffer gas at 1 mtorr can significantly decrease
photodissociation efficiencies in the infrared region. Since the time between
absorption of consecutive photons for the IRMPD process is on the millisecond
time scale, ions which have acquired some fixed amount of internal energy from
the photon absorption process (but not enough to reach the dissociation
threshold) can undergo collisional damping, thus reducing the photodissociation
efficiency of IRMPD.151,152 Therefore, in order to properly evaluate the performance
of the newly designed multipass-ring electrode, the ITMS was operated without
the presence of He buffer gas so as not to interfere with the photodissociation
process. This section presents a practical dialogue on ion trap operation without
buffer gas and discusses the relevant instrument parameters (without He buffer
gas) which affect trapping efficiency, resolution, and mass range.
One of the most important aspects of no He ion trap operation is trapping
efficiency. The two instrument parameters which affect trapping efficiency the

58
most are the ionization and cool times. Since there is no appreciable
concentration of buffer gas (only sample gas pressure in the 10'7 torr range),
longer ionization times are needed to generate large numbers of "reagent" ions
for the self-chemical ionization process or for the production of molecular ions
[M+ ]. Typical ionization times for experiments with no He buffer gas are on the
order of 60 to 100 ms. Under normal He buffer gas conditions, 60-100 ms
ionization times would severely space charge the ion trap for sample pressures
in the 10'7 torr range. However, since trapping efficiency is severely reduced with
no buffer gas, a larger number of ions must be made (necessitating a longer
ionization time) to trap enough ions so as to limit the large statistical variations
associated with low ion densities.
Perhaps even more important than generating large numbers of ions is the
cool time associated with the experiment. Wu and Brodbelt showed that for low
pressure operation of the ion trap (pressures < 9x105 torr), increased cool times
led to very efficient storage conditions similar to those with 1 mtorr of He buffer
gas.153 Over any given time period, it was shown that about 50-100 collisions
were needed to cool ions to the center of the ion trap. The authors reported that
for shorter storage times and lower pressures, the majority of the ions in the
expanded ion cloud are likely to be accelerated into an endcap rather than
through the endcap exit holes during the mass-selective instability scan. Also, it
has been postulated that ions on the outer fringes of the ion cloud could receive

59
too much kinetic energy from the rf field for coherent ejection during the analytical
____ 153,154
scan.
For many of the experiments discussed in this chapter, the cool time and
the reaction time for the formation of the [M+H]+ ion for diglyme and the crown
ethers was combined (since during any reaction time, collisional cooling can also
occur). In many instances, the cool/reaction times were on the order of 400 ms.
At these extended cool times, ion stability (ion-molecule reactions) can become
problematic. For the case of an even-electron species, such as protonated
diglyme, [M+H]+, at m/z 135, storage times of over one minute showed no
appreciable dissociation/reaction of the parent ion species. However, for an odd-
electron species like ionized allyl bromide ([M+] m/z 120 and m/z 122), long
storage/cool times can produce unwanted ion-molecule reaction products. As
seen in figure 2-4, an appreciable rate of reaction (2.5 s'1) of m/z 120 (the M+ of
allyl bromide) with the neutral allyl bromide present in the trap produced a
decrease in the parent ion signal intensity over a 100 ms time period. The
reaction of the parent species with the neutral allyl bromide produced the allyl
carbocation at m/z 41. The mechanism of this reaction was thought to occur
through a collision between the [M+ ] ions at m/z 120 and 122 with the neutral
species; this energetic collision (which has a substantially higher average kinetic
energy than that observed when He buffer gas is present) initiates charge-site
migration to the allyl portion of the molecule and loss of the bromine radical as
shown in equation 2-1.

Figure 2-4:
Kinetic data from the selective mass storage of the 79Br isotope from the molecular ion of allyl
bromide, m/z 120. A plot of In (l/l0) versus time yields a straight line, with the slope equivalent to
the rate coefficient of the ion-molecule reaction of m/z 120 with neutral allyl bromide.

O)

62
[C3H5Br]*- + C3H5Br [C3H5]+ + Br* + C3H5Br (2-1)
After the desired cool/reaction time for a given ion population, ion isolation
was accomplished by either apex or two-step isolation.767,9 The tuning of the
necessary rf/dc voltage combinations to store only ions of a single m/z required
fine adjustments of 0.1 to 0.5 V. This meticulous procedure was required
because under conditions of no He buffer gas, the stability edges of the Mathieu
diagram become very steep. These steep edges were found to be extremely
sensitive to ion population; without a fine rf/dc voltage control, a large percentage
of the parent ion population of interest could be ejected during the isolation event.
During the mass selective instability scan (no helium buffer gas), the
optimum resonance ejection parameters were found at a qzeject=0.89 and an
amplitude of 1.5 V(0_P). These values gave peak widths slightly smaller than 0.5
mass units at full width half maximum (FWHM). This observation was consistent
with that of the high resolution mode of operation, where reduced peak widths
were observed only using resonance ejection techniques.36 Operation of the ion
trap without helium buffer gas (no resonance ejection scan), produced larger
peak widths at FWHM (>0.5 mass units) and reduced signal-to-noise (s/n) ratios
for the peaks of interest. However, ion statistics and signal reproducibility were
found to be somewhat more reliable without using the resonance ejection scan.
For verification of the various reaction pathways, notch filter ejection
experiments were performed during the laser irradiation period. Frequency
probes for protonated diglyme (m/z 135) without He buffer gas required

63
significantly less resonant excitation amplitude (55 mV for no He) compared with
that of the same experiment with He buffer gas (245 mV) (see figure 2-5). In
addition, large shifts in the maximum ejection frequency were observed. These
shifts were attributed to a lack of helium buffer gas, which was needed for a well
defined frequency distribution. Also observed in figure 2-5, frequency bandwidths
were much greater (approximately 0.4 kHz with He) for the case where no He
buffer gas was present (2 kHz). All frequency data in the experiment were taken
for a constant number of ions stored in the trap (as measured by the electron
multiplier and detector circuits) so as to minimize bandwidth and frequency shift
phenomena typically observed with different ion populations. Although the
absolute number of counts for protonated diglyme (m/z 135) was matched in the
profile scan mode for the case of He buffer gas versus no He buffer gas, there
could be substantially more ions present which were not detected for the
aforementioned reasons in this section.
The Multi-Pass Ring Electrode
The 1.0-cm-diameter IR laser beam (Gaussian profile) was attenuated by
a 0.3 cm (1/8") entrance aperture on the ring electrode. The laser was aligned
such that the center portion of the Gaussian beam profile passed through the
entrance aperture. The ring electrode was modified by incorporation of three
polished stainless steel spherical concave mirrors (radius of curvature = 2.0 cm)
mounted on the inner surface of the ring, as shown in figure 2-6. The

Frequency optimization curves for protonated diglyme (with and without He buffer gas) using
dipolar excitation. The frequency was incremented every 0.1 kHz. For the case where He buffer
gas was present, a well defined optimum frequency was obtained due to the well defined
trajectories of the ions confined to the center of the ion trap.
Figure 2-5:

4001
£
c
4-
s
114.0 115.0 116.0 117.0 118.0
Frequency in kHz
O)
Ol

Modified ring electrode for multipass IRMPD experiments. Mirror positions and the eight laser
passes across the radial plane of the ring electrode, along with approximate photon density int eh
radial plane of the ring electrode, are shown. The positions of mirrors A and B determine the
number of laser transversals across the radial plane of the ring electrode.
Figure 2-6:

Ring Electrode
cw CO2
Laser Beam
1
Spherical
Mirrors
r=2.0 cm
Spherical
Mirror
r=2.0 cm
1 Pass
3 Passes
Hi 2 Passes
4 Passes
0

68
approximate photon density (assuming constant intensity across the attenuated
beam width) observed in the radial plane of the ring electrode can be seen in
figure 2-6. The most critical adjustment of the mirror system was the separation
of the centers of curvature of the mirrors labeled A and B. This separation
distance determines the number of beam transversals across the ring electrode:
4,8,12, or any other multiple of 4. The mirrors were mounted on the ring
electrode such that the centers of curvature of Mirrors A and B were on the front
surface of mirror C, and the center of curvature of mirror C was halfway between
mirrors A and B.155 This method of mirror alignment establishes a system of
conjugate foci on the reflecting surfaces of mirrors A,B, and C. Consequently,
light leaving the surface of mirror A is focused by mirror C on the surface of mirror
B, and the light leaving mirror B is then focused back to the original point on
mirror A. Similarly, any light leaving mirror C and going to either mirror A or B is
focused back to mirror C at some point offset from the original one15S (see figure
2-6).
This technique for extending the optical pathlength in restricted volumes
has many advantages over previous designs which incorporate flat mirror systems
or a spherical mirror and a truncated prism scheme.155 One advantage is the
ease in making adjustments since all tolerances but the horizontal angles of
mirrors A and B are usually quite large. Other inaccuracies which are introduced
are small and not cumulative. Another advantage is that light losses on mirrors
surfaces are kept to a minimum. Since there are only two reflections (at normal

69
incidence), spots, dust or pinholes on mirrors A and B have a much less serious
effect since the light from any point on the mirror surface always goes back to the
same point; therefore, if there is a spot on the mirror surface, the light falling on
the spot is lost but on the second reflection from that mirror no more light is lost.
Yet another advantage is that there is only one transmission of light through an
entrance aperture; no light travels through any glass or other optical material
where losses due to reflection can occur.155
Each mirror (and its mounting bracket) was constructed from a single piece
of stainless steel. At one end of each piece, the radius of curvature was cut (r =
2.0 cm) and the surface was highly polished. The three mirrors were mounted
into precision-drilled holes in the ring electrode, positioned such that the
alignment was automatic; no realignment has been needed since the original
assembly of the multipass ring electrode (36 months). A small machine screw
was used to hold each mirror-mirror mount assembly in place on the ring
electrode. Laser alignment was set such that the center portion of the 1-cm
Gaussian beam profile was transmitted through the entrance aperture, thus
yielding high photodissociation efficiency. With efficiencies already greater than
those published for previous designs on a QUISTOR trap or ICR cell, condensing
the beam down to 0.3 cm was considered less important than examining the
numerous ways in which gas-phase ion chemistry can be studied via IRMPD with
this unique design. Furthermore, with the beam focused to 0.3 cm, damage
could possibly occur on the surface of the ion trap mirror when the laser is tuned

70
to a strong IR emission line (e.g., 10.59 //m). With continued use over a 36-month
period, no degradation of the ion trap mirror surfaces has been observed with the
unfocused laser beam.147
Effects of Ion Motion on Photodissociation Efficiency
The effect of resonance excitation on ion motion in the quadrupole ion trap
is a rapidly evolving area of research, encompassing both simulation and
experimental investigations.57,156'160 The original reports of resonant excitation of
ions at their secular frequency of motion (where wz = the fundamental axial
frequency and tur = the fundamental radial frequency) due to an externally
applied ac field on the endcap electrodes was reported by Paul and Fischer.8,9
It was originally thought that radial excitation (o/r) would increase ion trajectories
only in the r direction and that axial excitation (wj would increase ion trajectories
in the z direction. However, with the nonideal (stretched) quadrupole ion trap, ion
motion in the r- and z-direction are coupled, leading to a complex series of
interactions between ion motion in both the r- and z-directions.161 Today,
resonance excitation of stored ions in the quadrupole ion trap is accomplished
by applying an auxiliary ac signal to the endcap electrodes. The frequency of the
ac signal applied corresponds to a particular frequency component of ion motion;
the greater the contribution of the component frequency to ion motion, the greater
the ability of the ion to absorb power from the applied field. The larger the power

71
absorption from the applied field, the greater the increase of an ions kinetic
energy (trajectory) for a particular frequency component.
The two most common types of resonance excitation are the dipolar and
quadrupolar techniques. Dipolar excitation (standard on all existing commercial
instrumentation) is performed by applying two auxiliary ac signals 180 out of
phase to the endcaps. Whereas quadrupolar excitation is performed by
application of the same ac signal to both endcaps. Experimentally, both dipolar
and quadrupolar resonance excitation have been used for CID studies, SID
studies, and resonance ejection (axial modulation) in the quadrupole ion trap.
More recently, there has been a concerted effort to understand the nuances of ion
motion during the excitation period.57,156'160
In this dissertation, a unique method using IRMPD for the evaluation of the
coupled motion phenomena observed with resonance excitation is reported. The
basic theory behind ion motion has been reported in chapter one of this work and
elsewhere. The two subsections entitled Dipolar Excitation and Quadrupolar
Excitation briefly discuss the theoretical aspects of resonance excitation, as well
as present relevant photodissociation studies concerning coupled ion motion.
This technique takes advantage of the high photon density in the radial plane of
the ring electrode, which can be used to determine the presence of coupled axial
excitation (a/J while performing radial (o/r) excitation experiments.
All coupled motion data presented in this study are plotted as
photodissociation efficiency versus excitation voltage (dipolar or quadrupolar

72
excitation). The photodissociation efficiency (PD) for a given experiment is defined
as the fraction of the original ion population photodissociated over a given
exposure time for a specified laser irradiance:
PD=1-
(2-2)
V 'o/
where I is the signal intensity of the dissociating ion at the end of the exposure
period and l0 is the signal intensity after the same period without irradiation. This
definition of l0 corrects for any unimolecular or collision-induced dissociation that
may occur.
Dipolar Excitation
Dipolar excitation is the most common form of resonance excitation
employed today. The techniques of collision-induced dissociation, notch filtering,
and axial modulation all use commercially available dipolar excitation circuits and
have been well characterized experimentally. Part of the reason for the success
of the aforementioned techniques is due to the asymmetric nature of the dipolar
field. Since the signals applied to the endcap electrodes are 180 out of phase,
axial motion is favored over radial motion during the excitation period.57 This can
be understood by examining the relationship between ion motion and the dipolar
mode of excitation. As an ion approaches each endcap, the potential on that
endcap needs to be of the appropriate polarity to obtain maximum power
absorption. With the application of a 180 out-of-phase signal, the polarity on

73
each endcap alternates at the same frequency as the ions secular frequency in
the axial direction.162
The trapping conditions observed with the application of a dipolar
excitation signal can be significantly different than those seen by direct application
of a quadrupolar trapping field to the ring electrode. Therefore, the resulting ion
trajectories no longer follow traditional Mathieu parameters, since ion motion is
now controlled by both the quadrupolar trapping field and the dipolar excitation
field. Previously reported simulation(s) for the application of dipolar fields for
resonance excitation typically employ numerical methods and simplified models
to describe ion motion.163 Exact solutions cannot be obtained because of
electrode surface geometry considerations (machining of hyperbolic surfaces,
truncation effects, and endcap holes) and the presence of higher-order fields
which result in a perturbed quadrupolar trapping field, therefore affecting ion
motion. The aforementioned reasoning results in coupled motion in the r- and z-
directions, thus making exact solutions difficult to solve mathematically.
A preliminary investigation to determine the influence of storage
conditions (under the influence of dipolar excitation) on photodissociation
efficiency was undertaken with protonated diglyme. To minimize the effects of
collisions on photodissociation efficiency, the ion trap was again operated with no
He buffer gas. A detailed description on the instrumentation employed and the
procedures for data acquisition, can be found in the Experimental Design section
of this chapter.

74
To determine the appropriate dipolar excitation frequencies for the ion
motion studies, the (M+H)+ ion of protonated diglyme was stored for a period of
10 ms at a qz value of 0.3. During the storage period, a dipolar excitation signal
was applied to the endcaps with an amplitude of 6 Vp_p. The frequency range
probed was from 25 kHz to 500 kHz, with the frequency incremented at intervals
of 10 Hz. A plot of the intensity of m/z 135 of protonated diglyme versus the
excitation frequency yielded a series of absorption bands with the center
frequency assigned as the minimum intensity value of m/z 135 for a given
absorption band. For absorption bands which were offscale for a 6 VP.P dipolar
excitation amplitude, the center frequency was taken as the center point of the
FWHM for the band. To assign the ions frequency from an absorption band as
accurately as possible, several instrumental parameters must be strictly controlled.
These include varying only one instrumental parameter per scan table and
allowing appropriate stabilization times for changes in the various instrumental
parameters (rf voltages, dc voltages, dipolar frequency, etc.). A detailed
description of all the factors involved has been published by Eades et al.162-164
Typically, the minimum value of frequency optimization curve is taken as
a close approximation of an ions component frequency of motion. As previously
mentioned, shifts in the observed frequency due to the instrumental variations
mentioned above can cause inaccuracies in proper frequency assignment.
However, other experimental factors can also contribute to inaccuracies in
frequency assignment for a given component of ion motion. These factors

75
include the amplitude of the applied excitation signal (frequency shifts to higher
values with increasing amplitude), He buffer gas pressure as described by the
physics of the damped harmonic oscillator model (frequency shifts to lower values
with increasing gas pressure), and space charge or ion-ion interaction
considerations (frequency shifts to lower values with increasing ion population).
In addition, arbitrary shifts in an ions frequency can also occur due to the
uncertainty of the exact RF voltage applied to the ring electrode of the ion trap.
Consequently, the experimentally determined ion component frequencies are only
qualitative estimates of the true ion frequencies.
The power absorption spectrum of protonated diglyme (no He buffer gas
present) for qz=0.3 and a^O is shown in figure 2-7. Three main frequency bands
were observed at wz (118.3 kHz), 2aiz (235.9 kHz), and 3a/z (349.8 kHz). The
broad frequency band at 55.1 kHz could correspond to the wz/2 band, however
the broad bandwidth precludes direct frequency assignment at this time. An
unknown frequency band at 446.0 kHz is also present. The largest absorption
band (the ions secular frequency of motion in the z-direction) at a/z was very
broad due to the large excitation amplitude used. Observation of the smaller
bands at 2o/z and 3a/z can be attributed partly to higher-order field effects
(hexapolar) on ion motion and have been successfully predicted using forced
Mathieu equations as developed by Williams et al.165 To account for power
absorption at these frequencies, several different explanations (or combinations
thereof) are plausible, including: 1) the presence of higher order fields

The power absorption spectrum of protonated diglyme for an applied dipolar excitation signal (6
Vp.p) at qz=0.3 and az=0 (no He buffer gas present). The applied dipolar signal was started at a
frequency of 25 kHz and incremented at 0.1 kHz intervals to 500 kHz. The symbol u/z refers to the
fundamental frequency of motion for an ion in the z-direction.
Figure 2-7:

Intensity
Dipolar Excitation Frequency in kHz

78
(hexapolar); 2) harmonic considerations of the applied dipole field; or 3) direct
contributions to ion motion. A detailed investigation into the origin of these
absorption bands has been performed by Eades162 and Vedel.166 These two
authors were independently able to verify the existence of the 2a;z frequency band
as a component of ion motion and not just part of the harmonics from the
excitation field. Other absorption bands observed by Eades and Vedel which
were not seen in these experiments include wJ2 and aiz+u)r Several plausible
explanations for this observation include the qz value used, the large a/z band
width, the absence of He buffer gas, and the sensitivity of the chemical probe
(protonated diglyme).
To examine the effect of dipolar excitation on ion (axial) motion, the
protonated diglyme ions at m/z 135 were irradiated with a cw C02 laser during the
dipolar excitation (10 ms) period. Since the time between consecutive photon
absorption for the IR laser in this experiment is on the ms time scale,
instantaneous ion trajectories or velocities cannot be determined. However, C02
lasers can be used to evaluate the time-averaged behavior of the ion (cloud)
population. The multipass optical design used in these experiments produces a
high photon density in the radial plane of the ring electrode as shown by figure
2-6. Any ions whose axial excursions exceed the beam width of the laser (3 mm)
should show a marked decrease in photodissociation efficiency.167
To test this theory, protonated diglyme ions were excited at a dipolar
frequency of 118 kHz (o/J, which was determined from the power absorption

79
spectrum. The effect on the photodissociation efficiency of exciting the [M+H]+
ions of diglyme (no He present) using dipolar resonant excitation (during period
5 in figure 2-3) is seen in figure 2-8. With increasing dipolar resonant excitation
voltage applied to the endcaps, the axial excursions of the ions increase,
decreasing the fraction of time the ions spend in the radial plane of the ring
electrode which, in turn, decreases the extent of interaction between the stored
ions and photons. The decrease in photodissociation efficiency can be further
rationalized by considering the inset on figure 2-8. As long as the ions maximum
excursion (trajectory) in the axial or z-direction does not exceed the laser beam
width (dashed line figure 2-8), the photodissociation efficiency remains relatively
constant. However, if the ions gain enough kinetic energy from the applied
dipolar field so that ion trajectories exceed the laser beam width in the axial
direction, a significant decrease in photodissociation efficiency is observed. This
decrease in efficiency can be explained by considering instantaneous ion velocity
arguments. As an ion moves away from the center of any rf-only device
(quadrupole or ion trap), the magnitude of the restoring forces become larger, as
for any harmonic oscillator. Therefore, as an ion reaches its maximum
displacement from the center of the ring electrode, its instantaneous velocity is
very slow compared to that of the same ion as it accelerates through the center
of the device. This means that when the peak-to-peak excursions of the individual
ion trajectories are larger than the laser beam width in the axial direction, the ions
spend most of their time outside the radial plane of the ring electrode where the

Figure 2-8:
Effect of resonant excitation (w2=118.3 kHz, qz=0.3, 6^=0) voltage on photodissociation efficiency.
When the axial excursions of the ions exceed the width of the C02 laser beam (as indicated in the
figure inset by the solid line), photodissociation efficiency drops off dramatically, since the ions
spend a significant amount of time at their maximum excursions outside the beam width. Error
bars are defined as the standard deviation of the mean.

Photodissociation Efficiency 1-(I/I0)
Laser Beam
Width
$
Ion Trajectory
| i i i i | i i i i | i i r~r | i i i i |
10 20 30 40 50 60
Dipolar Resonant Excitation Voltage in mV
00

82
photon density is highest (see solid line figure 2-8). This displacement results in
the observed sharp decrease in photodissociation efficiency for a dipolar
excitation amplitude of 15 mV.167
The unique advantage of the multi-pass ring electrode system (as shown
above) is its sensitivity in detecting the presence of axial excitation. By exciting
ions in the radial or r-direction (using quadrupolar excitation), the multi-pass ring
electrode system should be able to detect the presence of coupled motion in the
z-direction. However, this system would not be useful in detecting coupled ion
motion in the r-direction since there is a large photon density in the radial plane
of the ring electrode, which would not result in a drop-off in photodissociation
efficiency if the ions were excited radially. In the next section, a series of
experiments is described in which quadrupolar excitation/photodissociation
experiments are used to detect axial excitation (z-direction) when a given ion
population is excited only in the radial or r-direction.167
Quadrupolar Excitation
Over the last several years, the use of quadrupolar excitation as an
experimental tool has drawn increasing attention.166,168,169 The basic premise of
quadrupolar excitation is the application of the same auxiliary ac potential (in-
phase) to both endcap electrodes. The application of this in-phase ac signal
creates a symmetric field similar to the quadrupolar rf trapping field on the ring
electrode. Therefore, quadrupolar excitation does not favor either radial (r-

83
direction) or axial (z-direction) excitation of a given ion population with the
application of the excitation signal. This phenomenon can be understood by
examining the relationship between ion motion and the quadrupoiar mode of
excitation. As an ion approaches each endcap, the potential on that endcap
needs to be of the appropriate polarity to obtain maximum power absorption.
With the application of an in-phase auxiliary ac potential, the excitation signal
must be applied at twice the ions axial or radial frequency. Figure 2-9
demonstrates the sequence of events needed for maximum power absorption
using quadrupoiar excitation compared with that of dipolar excitation. Previously
published theoretical simulations have predicted strong power absorption at the
2uu frequencies (where u is either for the r- or z-directions) for quadrupoiar
excitation.57,156,157,162,166
The trapping conditions observed with the application of a quadrupoiar
excitation signal can be significantly different than those seen by direct application
of a quadrupoiar trapping field to the ring electrode. Therefore, the resulting ion
trajectories no longer follow traditional Mathieu parameters, since ion motion is
now controlled by both the quadrupoiar trapping field and the excitation field.
Due to the quadrupoiar nature of the excitation field, exact solutions to the
equations of motion can be calculated (March et al.)158. In this report, March and
coworkers were able to predict the presence of the 2uu bands, with the 2wz band
having a somewhat greater power absorption than the 2wr band.57,156-158

Figure 2-9:
Effect of the applied excitation signal (e.g dipolar or quadrupolar) on ion motion in the quadrupole
ion trap. The symbol ujz refers to the fundamental frequency of motion of an ion in the z-direction.
For maximum power absorption under dipolar excitation conditions, the ions frequency should
match that of the applied supplementary ac signal as seen in (a). The negative cycles of the
supplementary ac coincide with the ion frequency in the z-direction, thus leading to an increase
in amplitude for the ion trajectory and corresponding increase in ion kinetic energy. For
quadrupolar excitation where the applied supplementary ac signal is in phase with respect to both
endcaps, the applied frequency must be two times that of the fundamental ion frequency to obtain
maximum power absorption. An applied supplementary ac at the identical frequency of motion
of an ion in the z-direction leads to a significantly reduced power absorption spectrum (as seen
experimentally in figure 2-10).

Dipolar Excitation (entrance endcap)
Dipolar Excitation (Exit Endcap)
Quadrupolar Excitation (Entrance Endcap)
Qnadrtipolar Excitation (Exit Endcap)

86
An investigation to determine the influence of storage conditions (under the
influence of quadrupolar excitation and thus determine the presence of coupled
ion motion in the axial or z-direction while exciting the ions radially) was
undertaken using protonated 12-crown-4 ether. As mentioned previously, once
the ion trajectories exceed the 3 mm laser beam width in the axial or z-direction,
a marked drop off in photodissociation efficiency will occur, indicating the
presence of excitation in the axial direction. To minimize the effects of collisions
on photodissociation efficiency, the ion trap was again operated with no He buffer
gas. A detailed description of the instrumentation employed, limitations of
frequency measurements, and the procedures for data acquisition can be found
in the Experimental Design and Dipolar Excitation sections of this chapter.
To determine the appropriate quadrupolar excitation frequencies for the ion
motion studies, the [M+H]+ ion of protonated 12-crown-4 ether was stored and
analyzed as described previously in the Dipolar Excitation section of this chapter.
The frequency range probed was from 25 kHz to 500 kHz, with the frequency
incremented at intervals of 10 Hz. A plot of the intensity of m/z 177 of protonated
12-crown-4 ether versus the excitation frequency yielded a series of absorption
bands with the center frequency assigned as the minimum intensity value of m/z
177 for a given absorption band. For absorption bands which were offscale for
a 6 Vp_p quadrupolar excitation amplitude, the center frequency was taken as the
center point of the FWHM for the band.167

87
The power absorption spectrum of protonated 12-crown-4 ether (no He
buffer gas present) for qz=0.3 and az=0 is shown in figure 2-10. Three main
frequency bands were observed at a/r (59.2 kHz), 2u/r (126.0 kHz), and 2uiz (239.8
kHz). Two of these absorption bands, at 2ior and 2uz, were offscale, while the wr
band showed limited absorption. The magnitude of the 2a/r and 2uz bands was
approximately equal, which corresponds well with results reported by Eades et
al.162,164 Also, the magnitudes of all absorption bands observed agree well with
the theoretical values calculated by March.57,156,157-162 166 Note that the absorption
bands at 2wz and 2wr are not directly related to frequencies of ion motion, but
instead represent only the absorption of power. This can be explained by the
frequency doubling of the quadrupolar excitation signal applied to the endcaps,
such that the charge on an endcap is of the appropriate polarity to obtain
maximum power absorption (figure 2-9). Therefore, this absorption of power at
2a/r and 2wz represents actual ion motion at wT and a/z, respectively.
To examine the effect of coupled ion motion using quadrupolar excitation,
the appropriate excitation signal was applied at the three frequencies observed
(a/r, 2u)v and 2uz) with concurrent laser irradiation, as mentioned previously in this
chapter. A plot of the photodissociation efficiency versus the quadrupolar
excitation voltage applied to the endcaps yields the three curves shown in figure
2-11. As expected for the 2a/z band (representing u>z frequency component),
there is observed a steep drop-off in photodissociation efficiency as the axial
excursions of the ions exceed the laser beam width (for the reasons mentioned

Figure 2-10: The power absorption spectrum of protonated 12-crown-4 ether for an applied quadrupolar
excitation signal (6 Vp.p) at qz=0.3 and az=0 (no He buffer gas present). The applied quadrupolar
signal was started at a frequency of 25 kHz and incremented at 0.1 kHz intervals to 500 kHz. The
symbols u/z and cu, refer to the fundamental frequencies of motion for an ion in the z- and r-direction
respectively.

Intensity
Quadrupolar Excitation Frequency in kHz

Figure 2-11: Application of quadrupolar excitation signals at 59.2 kHz (a/r), 126.0 kHz (2wr), and 239.8 kHz (2coz)
to determine the presence of coupled motion observed in the z-direction with quadrupolar
excitation. The observed decrease in photodissociation efficiency for the 2a/r band cannot be
explained by radial excitation phenomena due to the presence of high photon density across radial
plane of the ring electrode.

Photodissociation Efficiency (1-I/Iq)
Quadrupolar Resonant Excitation Voltage in mV
CD

92
in the Dipolar Excitation section). For the wr band, a small decrease in
photodissociation efficiency is observed at 2750 mV, which could correspond to
a limited amount of coupled motion in the axial or z-direction. In the case of the
2a/r band (representing the tur frequency component), a sharp decrease in
photodissociation efficiency was observed at 750 mV. This decrease cannot be
attributed to excitation in the radial direction, since there exists a high photon
density throughout the radial plane of the multi-pass ring electrode. Therefore,
the only plausible explanation for the observed drop-off in photodissociation
efficiency for excitation at 2a/r is the coupling of ion motion in the axial direction.
Experimental evidence suggests that the majority of the coupled motion observed
(at qz=0.3, az=0) was due mainly to excitation in the z-direction and not
necessarily ejection or collisions of ions with the endcaps. This conclusion was
drawn from the fact that as the quadrupolar excitation amplitude was increased
(laser off) past the point where photodissociation efficiency dropped off (750 mV),
there was only a small reduction in the parent ion population. At an excitation
voltage of 1000 mV, a significant reduction in parent ion population occurred (as
evidenced by the increased variation of photodissociation efficiency values
beginning at 1000 mV), which likely represents radial "ejection" (collisions with the
ring electrode) of the ions and not complementary axial ejection.167 The results
obtained experimentally using the multi-pass ring electrode agree quite well with
those predicted by March et al.57'156,157 For qz=0.2940 and 3^=0 of the m/z 134
ion of n-butylbenzene, March and coworkers reported that radial ejection

93
predominates, but there exists a significant amount of axial excitation associated
with the application of the 2u, frequency in the quadrupolar excitation
mode.57156'157
The Photon Absorption Process
The minimum number of photons (n) needed to reach the thermodynamic
threshold for dissociation via a given reaction channel was calculated by dividing
the enthalpy change by the photon energy:
A H = nhv Na (2-3)
where AH is the enthalpy change, n is the number of photons absorbed, h is
Planks constant, v is the frequency of the radiation used, and NA is Avogadros
number.
The consecutive reaction sequence for protonated diglyme as it absorbs
IR photons is seen in equations 2-4 and 2-5:
[C6H1503r nhv [C5H02r CHjOH (2-4)
[C5H02]* + nhv [CgHyO]- C2H40 (2-5)
The AHf values for neutral methanol, neutral acetaldehyde, and protonated
diglyme were obtained directly from the literature.170 The AHf values for the
photofragments m/z 103 and 59 were estimated by using the quantum
mechanical AM1 (Austin Model 1) calculations based on the neglect of diatomic
differential overlap approximation.171'172 The accuracy of the AM1 calculations was

94
checked by comparing the value obtained in the literature (114 kJ/mol) for
protonated diglyme (m/z 135) against that of the AM1 routine (100 kJ/mol). A
summary of the AHf values for the two consecutive reactions of diglyme is given
in table 2-1. The errors observed in the tabulated values (e.g., AHf C5H1102+ and
AHf C3H70+) were on the order of 15 kJ/mol as determined by Katritzky et al. for
14 different cationic species by using the AM1 calculation.1'3 The results
observed for the reaction channels were n > 3 for the lower energy process and
n > 6 for the higher energy process as shown in table 2-1. The total number of
photons absorbed by the C-0 stretches in protonated diglyme was n > 9 for
formation of the m/z 59 fragment. The calculated photon energy at 944 cm"1 was
0.117 eV.
For the case of allyl bromide, only a one-step reaction for the photon
absorption process is involved (at 944 cm"1), as shown in equation 2-6:
[C3H5Br]+ + nhv -> [C3H5]+ + Br (2-6)
The AHf values for neutral bromine, protonated allyl bromide, and the allyl cation
were obtained directly from the literature.174 The result observed for the single
reaction channel was n > 5 for the formation of the allyl cation, as shown in table
2-2. The larger number of photons needed to reach the dissociation threshold for
the single allyl bromide reaction channel compared to that of the lower energy
channel for protonated diglyme was confirmed experimentally by the longer
induction periods associated with the allyl bromide reaction channel. These
results also agreed well with the smaller photoabsorption cross-sections expected

Table 2-1. Thermochemical data for the photodissociation of protonated diglyme.
Reaction
^^f(reactlon)
kJ/mol
^^((products)
kJ/mol
^^f(reactants)
kJ/mol
Photons
C6H1503+ + nhv -* C5H1102+ + CH3OH
37
C5H1102+ = 352*
CH3OH = -201b
C0H15O3+ = 114b
n ;> 3
C5H1102+ + nhv -*> C3H70+ + C2H40
68
C3H70+ = 586*
C2H40 = -166b
C6H1102+ = 352*
n 2r 6
a AHf values obtained by AM1 calculations references 167, 168.
b AHf values obtained from reference 166.
CO
OI

Table 2-2. Thermochemical data for the photodissociation of allyl bromide.
Reaction
^^f(reaction)
kJ/mol
AHf(products)
kJ/mol
^Hf(reactants)
kJ/mol
Photons
C3H5Br+ + nhv C3H5+ + Br
57.6
C3H5+ = 957.7
Br = 117.9
C3H5Br+ = 1018
n ^ 5
a AH, values obtained from reference 170.
to
O)

97
f0r the CBr stretch (as compared to the COC stretch) observed in the gas-
phase IR spectroscopy.
IRMPD Kinetics
As mentioned previously, the photodissociation efficiency (PD) for a given
experiment is defined as the fraction of the original ion population
photodissociated over a given exposure time for a specified laser irradiance:
PD=1-
' n
*0,
(2-2)
where I is the signal intensity of the dissociating ion at the end of the exposure
L f'
period and l0 is the signal intensity after the same period without irradiation. This
definition of l0 corrects for any unimolecular or collision-induced dissociation that
may occur. The photodissociation yield observed (PD) is generally dependent on
the wavelength of the laser, while the observation of a particular reaction route is
independent of laser wavelength. From equation 2-2, a first order relationship for
IRMPD kinetics (equation 2-8) was obtained:
In (l/l0) = -kDt (2-8)
where kD is defined as the photodissociation rate coefficient.130 By plotting ln(l/l0)
versus t (figure 2-12) the rate coefficient obtained for the diglyme [M+H]+ ion was
kD = 97.2 1.9 s1.146,147 This value was significantly higher than the 2-30 s'1
values obtained at higher laser irradiances reported in the literature.23,129,130
Clearly the multipass ring electrode enhances photodissociation efficiency to a

Figure 2-12:
Determination of the rate coefficient for protonated diglyme at a pressure of 1.1x107 torr by using
weighted linear regression, kD=97.2 1.9 s'1 (slope 95% confidence interval of the slope). Error
bars are defined as the standard deviation of the mean.

(Vi) i
-0.51
-1.0
-1.5
-2.0
-2.5
10.0
1 1 1 1 1 1 1 1 1 1 1 1 I
15.0 20.0 25.0
1
30.0
Irradiance Time in ms

100
great extent. Table 2-3 shows a summary of the rate coefficients for the
protonated species diglyme, 12-crown-4 ether, 15-crown-5 ether, and the
molecular ion of allyl bromide. As expected, the kD values observed for the C O
stretch (diglyme, 12-crown-4, 15-crown-5) were higher than the kD value obtained
for allyl bromide due to the reduced photoabsorption cross section for the C Br
stretch. Due to the linearity of kD, the rate coefficient can be expressed in terms
of the phenomenological cross section aD:
kD = oD$ (2-9)
where

Typical values of aD are 0.5-2x1CT20 cm2.130
Consecutive Reactions
Figure 2-13a shows the El mass spectrum of diglyme at a sample pressure
of 3.6x1 O'7 torr. Typically no molecular ion M+ at m/z 134 is present. Instead, a
fragment ion at m/z 89, formed by the cleavage of the carbon-carbon bond, and
a fragment ion at m/z 59, formed by either loss of formaldehyde from m/z 89 or
loss of C3H702 from the molecular ion, predominate the spectrum. This fact,
coupled with the ability to produce protonated diglyme (m/z 135) easily via ion-
molecule reactions, led to the use of the protonated molecule for IRMPD studies.
Characteristic photodissociation spectra of the [M+H]+ ion of diglyme at
944 cm-1 are shown in figures 2-13b and 2-13c. At lower irradiation energies
(irradiation time = 10 ms), the single reaction channel observed was the formation

Table 2-3. kD values for the first consecutive reaction channels of protonated diglyme, 12-crown-4 ether, 15-crown-5 ether,
and allyl bromide.
Compound
kD (s'1)
Pressure (torr)
Back Reaction
Diglyme
97.2 1.9
3.6x1 O'7
yes
12-Crown-4 Ether
89.1 2.9
4.2x1 O7
yes
15-Crown-5 Ether
104 2.6
4.2x1 O'7
yes
Allyl Bromide
59.6 2.2
2.9x1 O'7
yes
o

(a) El mass spectrum of diglyme at a pressure of 3.6x1 O'7 torr and 15 ms ionization time (no He
present), (b) IRMPD spectrum of protonated diglyme at 944 cm1 and 10 ms irradiance time
(energy = 0.201 J). (c) IRMPD spectrum of protonated diglyme at 944 cm'1 and 40 ms irradiance
time (energy = 0.852 J).
Figure 2-13:

Intensity
103
miz 59
C3H.O+
miz 103

c5h,a+
miz 135

c6h,A+
11111II1111111
y! 11111II11 rl111111111
Y n | i ri i | i i i i | i iTi |
30 40 50 60 70 80 90 100 110 120 130 140
miz
o

104
of m/z 103 with corresponding loss of neutral methanol as shown in figure 2-13b.
At higher irradiation energies (irradiation time = 40 ms), the dominant reaction
channel involved the loss of an acetaldehyde neutral from the m/z 103 ion shown
in figure 2-13c. The presence of a small peak at m/z 59 at lower irradiance times
suggested a competitive reaction mechanism for the formation of the product ion
species from the m/z 135 parent ion. Alternatively, when ion intensity was plotted
as a function of laser irradiance time (figure 2-14), the appearance was that of a
series of consecutive reactions.147
To determine if the reaction mechanism was competitive or consecutive,
a series of MSn experiments were conducted. In the first experiment, the
formation and mass isolation of protonated diglyme (m/z 135) ions were achieved
as described earlier. A tandem mass spectrometry experiment was then
performed with a 10 ms laser irradiance pulse followed by mass isolation of the
m/z 103 product ion. A series of MS3 experiments on the m/z 103 ion performed
by varying the laser irradiance times from 1 to 80 ms (increased energy input)
produced the ion growth curve for the m/z 59 product ion shown in figure 2-14.
To verify the exclusive formation of the m/z 59 product ion from the m/z 103
parent, a second tandem mass spectrometry experiment was performed on the
protonated diglyme parent ion (m/z 135) with concurrent notch-filter ejection of
the m/z 103 product ion. As before, the laser irradiance time was varied from 1
to 80 ms. Results from this experiment produced a decrease in the protonated
diglyme parent ion, but no formation of the m/z 59 product ion. These

Figure 2-14: IRMPD ion growth curves for protonated diglyme as a function of laser irradiance time. Two
reaction channels were observed: one for the formation of the m/z 103 fragment from the (M+H)+
ion and the other for the formation of the m/z 59 fragment from m/z 103. Some back reaction of
m/z 59 with neutral diglyme to reform the (m/z 135) (M+H)+ Ion was observed at longer irradiance
times. Error bars are defined as the standard deviation of the mean.

Ion Current
500=i_
Irradiance Time in ms
106

107
experiments confirmed the formation of the m/z 59 product ion from the m/z 103
precursor, thus verifying the presence of a consecutive reaction mechanism of the
type m/z 135 -+ m/z 103 -* m/z 59.147 The proposed mechanism of this
consecutive reaction, seen in figure 2-15, corresponds well to results reported
previously for the IRMPD of protonated diglyme in the ICR cell.175
The equations for the consecutive reactions of protonated diglyme can be
derived from the original work of Harcourt and Esson where the reaction of
protonated diglyme follows the simple consecutive reaction mechanism of:176-178
nhv nhv
[c6H15o3r [CsH^r [c3H7or
(2-10)
If the initial ion population of the protonated diglyme is [C6H15O3]0+ and the ion
population at any time t is [C6H1503]+, then the rate equation for the first
sequential reaction becomes:
_d[CH,5031- (2-11)
Applying the boundary condition where [C6H1503]+ = [C6H15OJ0+ when t = 0 and
integrating gives:
[C6H150,]- = [C6H,5O3]0e-k'' (2-12)
For the second sequential reaction step the net rate of formation for [C5H1102]+
is:

Figure 2-15: Reaction mechanism tor the IRMPD of protonated diglyme using
a cw C02 laser. The low energy reaction channel corresponds
to the absorption of a minimum of three photons (e.g. formation
of m/z 103), while the high-energy reaction channel corresponds
to the absorption of at least an additional 6 photons, for a total
of 9 photons absorbed by the protonated species. All parts of
the reaction mechanism were confirmed by MS" experiments.

109
H | m/z 135
A
ch2CH2ch2och2ch2och3
jj nhv n=3
ch3dch2ch2ch2ch2och3
CH?OH + CH2 CH2OCH?CH7CH
2 v_.a2
H /H i
Q
i2 ch2 och3
H H
m/z 103

ch3ch=o-ch2ch2
nhv n=6
n
ch3ch=och2ch2och3
ch3 -L ccbch3
H H

CH2 CH=0 CH3 I m/z 59

110
_ d[C5H,,02r s kl[C6H,503]- k2[C5H02]* (2'13>
at
By substituting for [C6H1503]+ in equation 2-13 with equation 2-12, the following
expression is obtained:
dickey*
dt
MC6H,503];e-M
Subsequent integration of equation 2-14 yields:
MC5H02]- (2-14)
[C5H02]- = [C,H,50,li(e-M -e-k!') (2-15)
K2 K-,
Therefore, the equation for the variation of the third sequential reaction product
[C3H70]+, can be obtained from the sum of the total ion current; assuming no
other routes for ion losses:
[CeH,s03]- [C5H02r [C3H70]* = [C3H1503]J (2-16)
so that:
[C3H70]- = [C6H,503] [C6H,503]- [C5H02J* (2-17)
Inserting the expressions for [C6H1503]+ (equation 2-12) and [C5H1102]+ (equation
2-15) into equation 2-17 yields:
[c3H7or =
k2 k.,
[k2(1 e'k,t) k,(1 e*kzt)] (2_18)
As seen in figure 2-14, the ion counts for [C6H1503]+ (m/z 135) fall
exponentially while the counts for [C5H1102]+ (m/z 103) approach a maximum.

111
Because the rate of formation of [C3H70]+ (m/z 59) is proportional to the ion
counts for [C5H1102]+ (m/z 103) the rate is zero at time ^ and has a maximum
value when [C5H1102]+ (m/z 103) reaches a maximum. Atthis maximum point, the
value of k, can be obtained:
-[C5;2r = 0 k,[C6H,503]* k2[C5H02]* (2-19)
at
^ ^[^6^15^3] (2-20)
2 [CsH.Ar
For protonated diglyme, k, was found to be 35 s'1. The derivation procedure does
not take into account the small back reaction (k<0.5 s'1) of m/z 59 with neutral
diglyme at a pressure of 3.6x1 O'7 torr to form [M+H]+ at m/z 135. In this case the
rate of formation of protonated diglyme from the back reaction only effects the
photodissociation of [M+H]+ from diglyme (formed via ion-molecule reactions) at
longer irradiance times (t>40 ms) where the reaction approaches completion.
The consecutive reaction sequence for 15-crown-5 ether (figure 2-16) was
also examined. To form a significant number of [M+H]+ ions at m/z 221, neutral
15-crown-5 ether was allowed to react with low molecular weight even-electron
species generated from El of the neutral. Reaction times for the protonated 15-
crown-5 ether were on the order of 400 ms. Following mass isolation of m/z 221,
the laser irradiance time was varied from 0 to 100 ms to track the series of
consecutive reactions. All reaction channels were verified using MS" experiments

Figure 2-16: IRMPD ion growth curves for protonated 15-crown-5 ether as a function of laser irradiance time.
Three reaction channels were observed with the following sequence: m/z 221 -* m/z 133 (loss of
C4H802) -* m/z 89 (loss of acetaldehyde) -* m/z 45 (loss of acetaldehyde). Error bars are defined
as the standard deviation of the mean.

2400
1800
1200
600
0
m/z 221
m/z 45
100
Irradiance Time in ms
w

114
as described previously. The results obtained gave a series of irreversible
consecutive reactions of the form:
[C10H21O5r [C6H1303]+ [C4H902]+ ^ [C2H50]*
m/z 221 m/z 133 m/z 89 m/z 45 (2_21)
The sequential reaction equations are derived as before, applying the boundary
conditions at t=0 mentioned above. The rate equations for the sequential four-
stage reaction are seen below:179'181
-dlC^O-ij.'. = k,(C10H2tO5l* (2-22)
dJCeH^OjT = ki(CioH2io5]* k2[CH,303]* (2-23)
dt
- d [C4H302]* = k2[CeH,303]* k3[C4H9o2]* (2-24)
d|Cy k3[C4H902]- (2-25)
By multiplying through using the appropriate integrating factor(s) (for equation 2-
23, ux=ek(2)t, and for equation 2-24, ux=ek(3)t) to form exact differentials,
subsequent integration/substitution (as described previously) gives:
[Ci0H2iO5]- = [CmH^Ojlie"1''1
(2-26)

115
l^6^133]+ ~ [^10^21^5)0
ki
k2 k1
(e
-k,t g -k2t
(2-27)
[C4H902]+ [C10H21O5]0
k,k2e
k1k2e"kl
-k2t
[C2H50]* = c6h2105]
k, k3e
-k2t
(k2 -
M(k3 -
ki)
k1 k2e"k3t
'
(ki
~ k3)(k2 -
CO
k2k3e R
CM
1
- ki)(k3 -
-k,)
k1 k2e"k
3^
(2-28)
(2-29)
(ki k2)(k3 k2) (k1 k3)(k2 k3)
As seen in figure 2-16, the ion counts for [C10H21O5]+ (m/z 221) fall
exponentially (kn= 104 s'1) while the counts for [C6H13OJ+ (m/z 133) approach a
maximum. Because the rate of formation of [C4H902]+ (m/z 89) is proportional to
the ion counts for [C6H1303]+ (m/z 133) the rate is zero at time and has a
maximum value when [C6H1303]+ (m/z 133) reaches a maximum. At this
maximum point, the value of Ic, was calculated as 32 s'1 (see derivation of
equations 2-19 and 2-20). The same derivation which was applied to the
consecutive reaction sequence for protonated diglyme and the third sequential
reaction product of protonated 15-crown-5 ether was also applied to the fourth
sequential reaction product of protonated 15-crown-5 ether, which gave k3=19 s'1.
In figure 2-16 is seen the complete consecutive reaction sequence for protonated
15-crown-5 ether. There also was a small degree of back reaction observed for

116
m/z 45 and m/z 89 with neutral 15-crown-5 ether. Again, the rate coefficients for
the two reactions (to form protonated 15-crown-5 ether) were less than 0.5 s'1 for
a pressure of 3.6x1 O'7 torr. Only at longer irradiance times (t>50 ms), where the
ion population of the protonated 15-crown-5 ether (m/z 221) is small, will the
effect of the back reaction be significant.
Buffer Gas Effects
Collisional deactivation studies of polyatomic species typically provide
energy transfer information on a broad scale. These gross changes in energy
levels are represented by changes in observed fragmentation or dissociation of
given polyatomic species. Previous studies of collisional quenching (pulsed C02
lasers) carried out in the ICR examined the efficiency of the quenching
process.129,131,182'184 For the case of low power cw C02 lasers, collisional
deactivation successfully competes with photodissociation at sufficiently low
photon fluxes and relatively high neutral pressures.185 Therefore, the amount of
photodissociation observed relates directly to the collisional deactivation efficiency
of the neutral buffer gas.
In figure 2-17, the effect of collisions on the IRMPD of protonated diglyme
is shown. The general trend observed for all buffer gases was a decrease in the
photodissociation efficiency with increasing buffer gas pressure. The observed
trend for quenching efficiency was N2>Ar>He. These results correlated well with
the previous studies of polyatomic and diatomic ions using the same buffer

Figure 2-17:
Effect of collisions on the photodissociation rate of protonated diglyme using N, Ar, He, and neutral
diglyme as target gases. All pressure measurements are corrected values. Error bars are defined
as the standard deviation of the mean.

Rate Coefficient (s )
Pressure xlO 6torr
00

119
gases.185-189 Traditionally, collisional deactivation efficiency increases with mass
and polarizability of the neutral molecule. For collisions involving He and Ar, the
loss of vibrational energy from the protonated diglyme can occur only by a
vibrational to translational energy transfer. For the N2 collision partner, vibrational
energy transfer is possible but not realistic due to the large difference in
vibrational frequencies. In all probability, the greater collisional quenching
efficiency of N2 is due to its greater polarizability as compared to that of Ar. The
collisional deactivation process was most efficient when the protonated diglyme
molecules collided with the diglyme neutral. In this case, proton transfer (i.e.
symmetric charge transfer) occurs, and the close match between the vibrational
frequencies of the colliding ion and molecule facilitates intermolecular vibrational
energy transfer.130
Pulsed valve experiments were conducted to determine if the increase in
trapping efficiency during ionization associated with the addition of He buffer gas
to the ion trap analyzer would interfere with the collision-free requirements
subsequently needed for IRMPD. A 1.6 ms pulse of He gas (10 psi He back
pressure) was found to trap the maximum number of diglyme (M+H)+ ions. The
signal intensity of protonated diglyme (m/z 135) by using a 1.6 ms pulse of He
gas during the pre-ionization period was higher by a factor of seven than when
He was not used in conjunction with the ionization event. The ionization time for
both experiments was 1.0 ms.146,147

120
For photodissociation to occur, the rate of photon absorption must be
greater than that of collisional and radiative deactivation. Thus for pulsed-valve
experiments, long ion storage times (several seconds) were required before
triggering the laser so that the He buffer gas initially used to efficiently trap and
damp the diglyme ions needed for the formation of the protonated (m/z 135)
species could be pumped away. Storage efficiency measurements show an
approximate loss of 2% of the original ion signal after 70 ms of storage. The
photodissociation efficiency as a function of the laser delay time is shown in figure
2-18. After a 2 s delay between the He gas pulse and the laser trigger, the
photodissociation efficiency levels off at approximately 90%.146,147
Wavelength Dependence/Infrared Spectroscopy of Gas-Phase Ions
In figure 2-19, the wavelength dependence of the IRMPD spectrum of
protonated diglyme is shown. The wavelength was varied from 933 to 953 cm"1.
The spectrum was normalized to an irradiation energy of 0.252 J. The reaction
channel was not found to be dependent of the laser wavelength used; only
photodissociation efficiency was affected by varying the laser wavelength. The
maximum photodissociation efficiency was found at 944 cm"1. A higher degree
of photodissociation efficiency was observed in the ion trap compared to an eight-
pass ICR cell (==90% at >tmax versus =55% at ^max)147. The difference in these
values was attributed to the easier alignment of the ion trap system. The width
of the absorption peak for protonated diglyme in the ICR cell was approximately

Figure 2-18:
Photodissociation efficiency as a function of post pulse-valve laser delay. Photodissociation
efficiency remains constant at over 90% for delay times > 2s. Error bars are defined as the
standard deviation of the mean.

Photodissociation Efficiency l-(I/Io)
Laser Delay Time in s

Figure 2-19:
Photodissociation efficiency as a function of laser wavelength for protonated diglyme at an energy
of 0.252 J (1.1x107 torr diglyme). Error bars are defined as the standard deviation of the mean.

Photodissociation Efficiency (l-I/Io)
l.Oi
0.8
0.6
0.4
| 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
932 936 940 944 948 952 956
Wavenumbers in cm'1 ^

125
7 cm'1 compared to 12 cm'1 for the ion trap.147,189 This difference may arise from
the difference between ion temperatures in the QITMS and the ICR. Ion
translational temperatures (measured with argon-nitrogen kinetic thermometer
reactions) in the QITMS were found to lie in the range 1700-3300 K, whereas
those in the ICR were found to be 500-880 K.190191 The higher kinetic energy of
the ions in the QITMS may correlate with a higher internal energy content, thus
leading to a larger absorption bandwidth because of a higher percentage of
molecules in the vibrational quasicontinuum.
A direct comparison between the infrared spectrum (wavelength
dependence) of protonated diglyme and the corresponding neutral spectrum may
not be feasible due to the presence of the extra OH bond observed with the
protonated species. This extra OH bond could result in a change of the
bonding orbitals and thus affect the COC stretch. However, any direct
comparison between IRMPD and neutral IR spectra must be done with care due
to resolution considerations and the limited spectral coverage of the C02 laser.189
Although direct assignment of various P,Q, or R branches may be difficult, a
general pattern or profile for the gas-phase absorption spectrum is possible.
A direct comparison between gas-phase neutral and positive ion IR
spectrum can be obtained for allyl bromide.148189 Since during the ionization
event a nonbonding electron is removed from the bromine substituent of the
molecule, there is a negligible change in the bonding orbitals which control the
CBr force constant. Therefore, a comparison of the neutral IR gas-phase

126
spectrum with that of the positive ion IRMPD spectrum is feasible. To obtain the
gas-phase neutral spectrum of allyl bromide (3-bromopropene), a Nicolet 7199
FTIR spectrometer equipped with a quartz cell and KBr windows (10 cm
pathlength) was used. The cell was filed with allyl bromide to a pressure of 0.45
torr. In figure 2-20 is shown the gas-phase neutral spectrum (850-1350 cm'1) and
the IRMPD spectrum of allyl bromide. The widths of the doublet peaks at 944 cm'
1 and 976 cm'1 are significantly smaller than those of the corresponding neutral.
The allyl bromide ions stored in the ion trap undergo at most one to two collisions
during the laser irradiance period, meaning the spectrum observed will not be
collisionally broadened. For the case of the gas-phase neutral, at a pressure of
0.45 torr collisional broadening is observed as evidenced by the approximate 90
cm'1 bandwidth of the doublet peak at 920 cm'1. As mentioned previously, the
widths of the peaks obtained with the ion trap were again somewhat larger than
those observed with the ICR cell (which follows the same reasoning stated
above). The individual absorption peaks observed in the IRMPD spectrum are
due to the overlap of sharp C02 laser lines with the sharp molecular absorption
bands. For larger more complex molecules/ions, where the absorption of the first
photon may be in the vibrational quasicontinuum, a more broad featureless
spectrum may be expected.148,189

Figure 2-20:
Gas-phase spectroscopic study of allyl bromide neutrals and ions. The spectrum (FTIR) on the
left is for neutral allyl bromide at. a pressure of 0.45 torr; on the right hand side is shown the
spectrum obtained using IRMPD at a pressure of 1.1x1 O'7 torr. A comparison between the neutral
and IRMPD spectra is feasible since a non-bonding electron is removed from the bromine atom
(of allyl bromide) during the ionization event, therefore not affecting the force constant of the CBr
stretch. Error bars are defined as the standard deviation of the mean.

Absorbance
Wavenumbers in cm *
1.00-,
0.80-
0.60-
0.40-
0.20
I I I I j I I I I j I I I I | 1 I I I | I I I I | I I I "l~|
930 940 950 960 970 980 990
Wavenumbers in cm1
128

CHAPTER 3
THE RF-ONLY OCTOPOLE ION TRANSMISSION GUIDE
Over the last eight years, ion injection into the quadrupole ion trap has
become one of the most popular areas of research in mass spectrometry. Since
the original report of ion injection using an external El ion source by Louris et
al.192, a myriad of different ion sources have been interfaced with the ion trap.
These include but are not limited to El/Cl192'194, fast atom bombardment (FAB)195,
particle beam196, thermospray197, electrospray35,198 202, glow-discharge203205,
atmospheric pressure206,207, inductively coupled plasma (ICP)208, laser
desorption209-213, super critical fluid214, and resonance enhanced multiphoton
ionization (REMPI)215"216 sources. In all the aforementioned literature, every
external ionization source has utilized some form of dc lens system for ion
injection into the ion trap. Although dc lens systems may be the simplest to
design and construct for an ion injection system, they are not necessarily the
most efficient way to transfer ions from an external source to the ion trap analyzer.
A stack of dc lenses has a poor conductance for letting neutrals introduced along
with the ions to be pumped away. This is especially true for high pressure
ionization sources such as electrospray or atmospheric pressure ionization where
a large number of collisions between ionized species and gas-phase neutrals can
occur, effectively scattering a large portion of the ions. Since dc lenses serve
129

130
only to focus the ion beam and not to recapture" scattered ions for ion injection,
an alternative method which could recapture scattered ions and focus them into
appropriate trajectories for ion injection would be desirable.
One technique which fulfills this requirement is the rf-only multipole. RF-
only multipoles have been used extensively in both analytical and physical mass
spectrometry217,218 The ability of these devices to focus ions in a high pressure
environment can be understood by examining the forces exerted on a given ion
population as it moves through an rf-only device. As an ion is displaced from the
center axis of the device, the restoring force acting upon that ion (in an rf-only
multipole with 2n electrodes) is proportional to the (n-l)* power of that
displacement, where n=2,3, and 4 for a quadrupole, hexapole, and octopole,
respectively219 Therefore, as ions are displaced from the center of the rf-only
device due to collisions with neutral atoms or molecules, the restoring forces
present recapture the displaced ions and successfully transfer them from a high
pressure region (e.g., electrospray ion source) to a lower pressure region (e.g.,
ion trap analyzer) with minimal scattering losses. Based on the above discussion,
the octopole would be the logical choice for an ion injection device due to the
greater restoring forces present for n=4 case.
Previously, the rf-only octopole has been used as an ion-molecule reaction
cell220,221, as a collision cell in tandem quadrupole instruments222, as an ion
injection device for triple quadrupole instruments223, and as a device for
determining ion-molecule reaction cross sections and energetics (i.e., translational

131
energy dependence, product branching ratios, collision-cross sections) of ion-
molecule reactions.224-227 More recently, the rf-only octopole has been employed
in the development of commercial (liquid chromatography/mass spectrometry) ion
trap instrumentation as an ion injection device (Finnigan MAT LCQ ).2ZB Other rf-
only devices have been designed and used (e.g., hexapoles and quadrupoles)
for the aforementioned purposes and have been employed as ion transmission
devices.229-232
In this chapter, the theory, design, and implementation of an rf-only
octopole for use as an ion injection device is discussed. A general overview of
ion behavior in electromagnetic fields is presented followed by a discussion of
multipole field potentials. Calculations for electric field strength, electrode
geometry, and the equations of motion are then introduced to give the reader a
general understanding of the physics of rf-only (octopole) devices. The next
section presents a discussion on the design considerations for the rf-only
octopole used in the electrospray/ion trap instrument built in our laboratory. The
chapter concludes with a discussion of the factors which contribute to the ion
transmission properties of the rf-only octopole (initial entry angle, rf frequency, rf
amplitude, collisions factors, and kinetic energy considerations).
Ion Behavior in Electromagnetic Fields
This section is intended to introduce from first principles the differential
equation governing the motion of an ion through an electromagnetic field with any

132
number of electrodes (pole pieces). The derivation in this section represents a
summary of the work of Friedman et al.232 While not intended to cover in detail
all the mathematical nuances of the derivation, this qualitative discussion should
provide the reader with the necessary background to understand the basic
physics behind ion motion.
From Newtons law of motion the differential equation which describes the
motion of a charged particle through any field is defined as:
F=m (3-D
dt
where F is the force on the ion, m is the mass of the ion, and r is the radius
vector. For an electromagnetic field with components E for the electric field and
B for the magnetic field, the force on an ion is given by the Lorentz force law
where:
F = e(E + v x B) i3"2)
where e is the charge on the ion and v is the ion velocity. Since the rf-only field
applied to the multipole device is a time-varying field, the general form of the
Maxwell equations can be used to compute the E and B fields:
V D = p v x E =
at
(3-3)
V B=0 vxH=J+-^
at
here D is the electric displacement vector, p is the charge density in the medium,

133
H is the magnetic field strength vector and J is the current per unit area in the
medium. The del operator (V) was defined in vector space as:
V = k (3-4)
3x 8y 3z
where T J and k are unit vectors along the x, y, and z, axes respectively.232
The divergence, curl and time derivative of the Maxwell equations 3-3 can
be computed for an electromagnetic field provided that the charge density
p(x,y,z,t) and the current density J(x,y,z,t) are known for all space and time. In
addition, the values E, D, B, and H of must satisfy equation 3-3 and cannot be
chosen arbitrarily. The Maxwell equations can be rewritten in terms of the vector
quantities B and E from the Lorentz force law in equation 3-2, providing that some
basic assumptions are made about the material from which the pole pieces are
constructed (linear, homogenous, isotropic materials) and the medium (vacuum)
in which the ion moves. With these assumptions the Maxwell equations
become:233,234
V (eE ) = p VxE = ^
3t
(3-5)
V B = 0 vxlj + M)
ii at
where e is the permitivity of free space and /# is the permeability of free space.
For purposes of computing the E and B fields acting on an ion in equation 3-5,
the associated ion charge(s) and current density can be neglected as the ion

134
moves through a vacuum (free space). The Maxwell equations now become:
V e = 0 v x E =
at
V B = 0
v B a(e0E)
F0 at
(3-6)
where under vacuum conditions (approximately free space), the values of e and
// become the constants e0 and /0. Therefore, for a low enough density of ions
(p and J approach zero), the Maxwell equations in 3-6 are a good approximation
in the vacuum region enclosed by any number of pole pieces. When the E and
B fields in the Maxwell equations are not changing with time (e.g., instantaneous
values), the Maxwell equations are evaluated as:
V E = 0 V x E=0
V B = 0 V x B = 0
(3-7)
The applicability of the Maxwell equations in equation 3-7 to the case of an
enclosed vacuum region (any number of pole pieces) where the electromagnetic
fields are changing, can be explained by the speed of an ion as it moves through
the device compared to that of the speed of light. From Maxwells equations 3-6,
B0=E0/c where B0 and E0 are the maximum amplitudes of the E and B fields233,
and c is the speed of light with c={ejioy'12. Since the fastest speed of an ion in
a quadrupole instrument235-236 is only v/cdO"4, the B term in the Lorentz force law
is negligible and equation 3-2 reduces to:

135
F = eE
and the Maxwell equations 3-7 reduce to:
V E = 0
(3-8)
Vx E = 0
(3-9)
with the right hand side of equation 3-9 valid for an applied potential which has
a low enough frequency so that:
A = -l
(3-10)
where the wavelength of the applied potential (il) must be significantly larger than
the length (I) of the multipole pieces.
An electric field E which satisfies Maxwells relationship in equation(s) 3-9
can be derived by employing two mathematical theorems which combined yield
the basic equation for an applied potential to any multipole device in a vacuum
system. The first theorem is that of Stake's, which for any continuous
differentiable vector field A is defined by
V x A ds = j A dr
(3-11)
where dr is the vector on the circumference of the closed area S moving in the
direction of integration.232 By Applying Stakes theorem to the right hand equation
from Maxwells relationships in equation 3-9 yields:
E d?=0 (3-12)
for any closed path where for any surface VxE=0. The second mathematical

136
theorem, states that for any vector field around a closed path which gives a value
of zero (as in equation 3-12), the vector field may be represented as the
divergence of a scalar field with:
E = -V(x,y,z) (3'13)
The proof for equation 3-13 can be found in Friedman et al.232 Substituting
equation 3-13 into the first Maxwell equation of 3-9 gives:
V V<|>(x,y,z) =0 (3_14)
the V*V term is called the LaPlacian as is defined as V2. This transforms
equation 3-14 to:
V2(x,y,z) = 0 i315)
Putting Laplaces equation 3-15 into rectangular coordinates gives:
V2 = V V
8 ? a ,* 3 '
i +1 + k
k dX dy dz
(3-16)
- JL + Ji- + JL
ax2 + ay2 + az2
Since equation 3-15 is a result of the Maxwell relationships from equation 3-9, any
function of 0(x,y,z) which is a solution of equation 3-15 (with an E field defined
by equation 3-13) satisfies the requirements of Maxwells relationships in equation
3-9 232
The differential equation for ion motion for any configuration (e.g., any

137
number of pole pieces) can now be found. Thus, for any ion moving in a field
with n conductors and constant voltage 0¡(¡=1,2,3....n), the equation for the
surface of the ith conductor is z=f¡(x,y). The equation of motion 0(x,y,z) is then
found by satisfying LaPlaces equation 3-15 and meeting the boundary conditions:
(x,y,f (x,y)) = ¡ (i = 1,2,3, n) 3-17
Therefore, the potential applied to any of the electrodes (pole pieces) at their
surfaces must equal the defined potential function.
Once the potential equation is known, Newtons law of motion can be used
to find the differential equation of motion by:
F = ^ = -e V<}>(x,y,z) (3-18)
dr2
the method is approximately valid for time varying potentials giving:
4>¡ = 4*DC ~ (3-19)
which is subject to the limiting conditions imposed from equation 3-10, where the
applied potential does not change too rapidly for the length of the multipole
pieces. From equation 3-19, rods and 0AC refers to the amplitude of the time-varying potential of frequency ()
applied to the same multipole rods. Equation 3-19 can also be rewritten into the
more familiar form where for the applied potential 0O:
4>0 = U V cos where U is the applied DC voltage and V is the amplitude of the time-varying

138
potential.
Octopole Electrode Arrangement
The derivation for the octopole potential in this section and those that
follow is based on the general derivation for a multipole device by Szabo et al.219
The first step in understanding the mathematics behind the rf-only octopole is an
examination of the octopole electrode arrangement. For any time-varying two-
dimensional field with cylindrical symmetry, the electrode structure is defined by
the nth order, where 2n equals the number of hyperbolic electrodes. In the case
of the octopole, the applied electric field is 4th order with 2n=8. The two
dimensional octopolar field is generated by the eight parallel/hyperbolic rods
shown in figure 3-1. Ideally the rods should be hyperbolic in nature, with an
applied potential between opposite pairs of rods defined by equation 3-20 from
the previous section:
0 = U V cos w (t t0) (3-20)
with the independent time variable denoted as t and the initial phase defined by
V The frequency w in equation 3-20 is defined in angular units by:
co = 2 it f (3-21)
where f is the applied frequency.219
The x-axis in figure 3-1 is chosen so that it bisects (lies in the plane of
symmetry) one of the positively charged electrodes. For symmetry

Figure 3-1: The electrode structure of the ideal octopole (n=4). The contour lines of the hyperbola-like cross-
sections of the electrodes are given by a polynomial of the 4th order. The distribution of the
electrode potentials for the octopole electrode system is also shown. Adapted from reference 219.

140
+

141
considerations, the octopole electrode structure must have a two-fold rotational
axis of symmetry where the device can be rotated about the axis rr/4 radians, and
the new configuration is indistinguishable from the old. Since the potential along
the z-axis is zero, the hyperbolic rod potential relative the z-axis is 0J2 and -0J2
for adjacent rod pairs.219,237'238
Octopole Field Potentials
The potential field generated by an octopole device can be derived by
starting with the LaPlace differential equation defined in terms of polar coordinates
where 0=0(r,0,t) is equivalent to the rectangular coordinate system with
0=0(x,y,t). Polar coordinates will be used to solve the equation for the field
potential since the mathematics involved are much less complex and time-
consuming. At the end of the derivation, the polar coordinate system will be
converted back to traditional rectangular coordinates. Laplaces equation takes
the form in polar coordinates of:
V4, = ^*^ = ^f = 0 (3-22)
dx2 dy2 dr r2 302
Since the octopole electrode structure displays circular cylindrical symmetry,
circular cylindrical coordinates can be used to seek the solution of the Laplace
equation, which takes the form:
= R(r) 0(6) 0(t) (3-23)
where each of the functions R(r), 0(0), and 0o(t) has only one independent

142
variable, r, 0, and t, respectively.219
If the potential impressed between the octopole electrodes and the
potential of the oscillatory electric field generated between the electrodes (by the
individual octopole electrode potentials) have only one sinusoidal component
(e.g., assuming the electrode shape and alignment are perfect), then the general
solution to the Laplace equation may be written as:
(r,e,t) = (Akrk + Bkr"k)(Cksink0 + Dkcosk6)[U Vcos where Bk, Ck, and Dk are constants and (r,0) are defined as in figure 3-2.
Since 0(r,0,t) must be an even function (Ck=0) and must satisfy the following
boundary conditions:
= -^4>o = -^[u Vcoso)(t t0)] if r = r0 and 6 = 0 (3_25)
= 0 if 6 = ; 3; 5;...
8 8 8
Then the potential of an octopolar field with eight poles becomes:
cos 46
(3-26)
withp0 (the electrode potential) defined as in equation 3-20. To convert equation
3-26 from polar to rectangular coordinates, Moivres formula and the binomial
expansion can be used with:

Figure 3-2:
Rectangular and polar coordinates used to calculate the potential field of an octopole electrode
configuration with 8 identical electrodes. The x-axis is chosen so as to coincide with the plane of
symmetry of one of those electrodes which has the potential +J2. Theta (O) is the angle of
rotation, r is the displacement from the center axis, and r0 is the inscribed radius. Adapted from
reference 219.

144
X

145
eine = (cosn6 + isinn6) = (cose + isin6)n
= (¡j)(cos0)"-k (i sin 0)k
where n=4 for the case of an octopole and:
(3-27)
n!
k!(n k)
for k = 0,1,2,... s n = 0,1,2,
(3-28)
By substituting:
cos0 = and sin0 = ^ (3-29)
r r
(the definitions for cos and sin) into Moivres formula and applying the binomial
expansion (see equation 3-27), the equation for the field potential in rectangular
coordinates is obtained:219
4> = -^-0(x4 6x2y2 y4)
2r0
(3-30)
Electric Field Strength Calculations
To determine the strength of the electric field vector E and its component
vectors Ex (x-axis direction) and Ey (y-axis direction) for an octopolar field, the
vector equation may be written as:
E = ux Ex + Uv Ev = ux Uv-^- = grad
y y 0X 1 dy
(3-31)

146
A pictorial representation of component and unit vector system used in this
derivation can be seen in figure 3-3. To determine the electric field vector in the
x-direction, equation 3-30 and the following identity are used:
(n 2k)
k = 0,1,2,... s n = 4 3*32)
From the previous section (not shown), the cos n0term for equation 3-30 can be
derived as:
cos(n6) =
1
/ \
n
: V (2
Xn-2y2 +
xn 4 y 4
(3-33)
rn
Differentiating equation 3-33 with respect to x and (substituting for n=4 and
evaluating k=1 to 4) and applying equations 3-30 and 3-32, the vector component
E, is defined as:
Ex = 4>0 r3 cos 30 (3-34)
ro
For the component vector in the y-direction Ey, equation 3-30 and the following
identity are used:
2k
= n
n -1
2k -1
k = 1,2,... s n =4
(3-35)
From the previous section (not shown), the sin n0 term for equation 3-30 can be
derived as:

Unit vectors and components of the electric field strength in rectangular and polar coordinates.
UK1 Uy, Ex, and Ey are unit vectors and transposed vectors in the rectangular coordinate system
respectively. U* Url E*, and Er are the angle of rotation and displacement from center vectors in
the polar coordinate system respectively. Adapted from reference 219.
Figure 3-3:

148
X

149
(3-36)
Differentiating equation 3-36 with respect to y and (substituting for n=4 and
evaluating k=1 to 4) and applying equations 3-30 and 3-35, the vector component
Ey is defined as:219
Ey = 4>0 r3 sin 30
(3-37)
Electrode Geometry
If the electrode potentials of the adjacent octopole electrodes are defined
as +0/2 and -0/2 respectively (see figure 3-1), and 0O is defined as in equation
3-20, then the equation of the contour line on the hyperbolic surface of the
octopole electrode may be obtained. The equation for the contour surface
derived here will represent the xy surface, with a corresponding substitution of the
condition 0=0O into equation 3-26 (the field potential in polar coordinates). This
substitution for an octopole geometry gives:
(3-38)
cos 46 = 1
For rectangular coordinates, the expression obtained from equation 3-30 is:
To convert the contour equations to the per-unit system where the spatial

150
(x4 6x2y2 + y4) = 1 (3-39)
ro
coordinates are defined in dimensionless units, the three variables x, y, and r
become:
X = ; Y = -2-; R = (3-40)
r r0 r0
Therefore, the equation of the electrode contour line (from equation 3-39)
becomes:
X4 6X2Y2 + Y4 = 1 (3_41)
Accordingly, if a charged particle were to hit an electrode surface then the values
of X, Y, and R must be > 1219
Equations of Motion
The derivation for the equations of motion in an octopolar field need only
be considered in the xy plane. The force acting on an ion in the z-direction, or
along what is frequently termed the beam axis, is always zero. If the pole pieces
for the octopole are perfectly parallel and fringing fields are neglected, then the
ion will move along the axis of the octopole with a constant velocity determined
by its kinetic energy (offset potential) and entrance angle when it enters the
octopole. From Newtons law of motion, the z term is defined as:

151
(3-42)
which has the general solution:
Z = Z0 + v2(t t0)
(3-43)
with z defined as the initial position in the z-direction and vz the velocity in the
z-direction.
The force acting on a charged particle in the xy plane of an octopolar field
is obtained by combining Newtons law of motion with the Lorentz force law
(magnetic field strength of zero). The total force F is defined as:
F = m = ma = eE
dt
(3-44)
where v is the velocity in the xy plane, e is the charge on the ion, and E is the
electric field strength vector as defined in equation 3-31. Using the component
vectors of the electric field strength, equation 3-44 can be rewritten in rectangular
coordinates as:
,¡ = iF(¡iE
dt2 dt2 m x m
(3-45)
substituting for E, (equation 3-34) and Ey (equation 3-37), the equations of motion
in the x- and y-direction in rectangular coordinates are written as:

152
d2x
dt2
[ U V cos to (t t0) ] (x 3 3xy2)
mr04
(3-46)
= -^%[U Vcoso>(t-t0)](3x2y y3) (3-47)
dt2 mr0
Extension of the Mathieu equations (discussed in chapter two for a quadrupolar
field, n=2) to that of an octopolar field described here gives the following
relationships:
+ [an 2qn cos dt2
+ [an 2qn cos w(t t0)](3x2y y3) = 0 (3-49)
with n=4 for the case of the octopole, the a and qn parameters are evaluated as:
n3eU 0 32eU
= a4 =
2mw2r02 m (3-50)
q
n
n3eV
4mco2r02
16eV
_ 2,2
m (3-51)
Equations 3-48 and 3-49 show that motion in the octopole is coupled in the
xy plane. This has two important ramifications: (1) ion trajectories for the
octopole will be much more complex than the case for a quadrupole where
motion is independent in the x- and y-direction; and (2) the size and shape of the

153
corresponding stability diagram is now dependent on the initial conditions of ion
motion.
Equations 3-48 and 3-49 can be converted into their dimensionless form
by substituting equation 3-40 for the x, y, and r values and subsequently
transforming the time variable to a dimensionless quantity. The conversion of the
equations of motion to this form means that all numerical calculations relating to
the octopole device are in dimensionless units. The advantages of the
dimensionless equation of motions are twofold: (1) the constants and variables
of the octopole will lie in a somewhat more narrow numerical range compared to
their true values; and (2) keeping track of exact units and numerical values of the
system parameters (which can be quite complex especially for the case involving
coupled ion motion observed with the octopole) is simplified.219
In order to understand the restoring force generated by the octopole
electrode arrangement, the equation of motion in the complex form must be
generated. Each of the two dimensional vectors involved in ion motion (position
vector(s), electric field vectors, velocity vector, and acceleration vector) are
functions of the real time variable t in the complex xy plane. Taking this into
account, the equation of motion in the complex form is:
Iff* Jt[U-Veo..(1-t0)](Z)-,.0 (3-52)
dt mr0
where Z is the complex conjugate of Z which has the following definition:

154
Z = x iy = re "ie = r cos 6 ir sin 0
Applying the constraints of the Mathieu equation to equation 3-52 and multiplying
through by m (with the appropriate rearrangement) gives:
m-^-^ [- ma4(Z)n_1 ] = 2mq4[cos4(t to)](Z)n~1 (3-54)
dT 2
The motion of an ion in an octopolar field is defined essentially as an
undamped non-linear forced oscillation. This motion is undamped because (in
the term on the left hand side of equation 3-54) there are no additive terms for the
dZ/dt independent variable. In this case, the combined term m(dZ/dt) refers to
the inertia force while -ma4(Z)lv1 is the restoring or spring force. As an ion
increases its displacement (Z) from the center of the octopole, the stiffness" of
the spring can be defined as the first derivative of the restoring force. The spring
is referred to as "hard" if the restoring forces increase as the ion increases its
displacement from the center of the octopole. For the case of the quadrupole
analyzer where n=2, the restoring force is a linear function of the ions
displacement (Z). If n > 3 then the restoring force is non-linear (hexapole,
octopole, dodecapole, etc...) and is considered a "hard" spring. The final term on
the right hand side of equation 3-54, 2mq4[cos n(t-t0)](Z)rv1, is the applied external
force with corresponding time-varying amplitude and is also a non-linear function

155
of the displacement (Z).219
Octopole Design Considerations
This section considers the relative design and assembly procedures used
to construct the rf-only octopole for the ESI/ion trap instrument. A discussion of
the basic considerations for round pole pieces versus hyperbolic pole pieces, an
evaluation of the effective trapping potential, and assembly procedures are
presented.
The best performance of any rf-only octopole device will be achieved when
the shape (cross section contour) of the machined electrodes exactly matches
that obtained by theory (i.e., hyperbolic). Previous methods of manufacture of
these devices have centered on using a single piece of pressed ceramic made
to the exact length specifications of the device.239,240 The inner surface of the
ceramic consists of eight inwardly protruding sections with these surfaces cut to
the exact equation of the octopole contour. The inner portion of the ceramic is
metal coated so as to produce an octopolar field by application of the appropriate
rf frequency and voltage. Various other techniques for fabricating and designing
monopoles, quadrupoles, and other multipoles developed over the past thirty
years are primarily for the design of mass filter systems (e.g., application of both
rf and dc voltages to the multipole).239'249 These designs are somewhat complex
and cumbersome due to the strict requirements needed for mass filter systems
which can obtain a mass resolution of unit mass or better. For the rf only

156
octopole (or any rf-only device), the most critical adjustment is ensuring the pole
pieces (in the xy plane) are all of the same length and no one end of any pole
piece protrudes excessively from the xy plane. This requirement is necessary to
minimize fringing field effects which can introduce forced motion in the z-direction.
Otherwise, considerations such as r0 requirements and consistencies in electrode
shape are not as critical due to the operation of the device in the rf-only mode
(i.e., operation as an ion transmission device as opposed to a mass filter).
Hyperbolic shaped rods for the octopole assembly can be difficult to
machine and design as well as to mount accurately. To build the octopole device
in a timely and accurate configuration, round rods can be used in place of
hyperbolic rods to produce an octopolar field comparable to that generated by
a pure hyperbolic contour surface. To minimize the distortions generated by
round rods, the relationship between (the rod radius) and r0 (the inscribed
radius) is:220
1 = 0.37 (3-55)
r
The original relationship for ajr0 was developed for a quadrupole analyzer by
Dayton et al.250 The accuracy of the technique used to calculate the ratio was
251
later refined by Denison.

157
Effective Trapping Potential
To understand the effect of trapping potential on multipole device design,
a discussion of the forces acting on an ion and thus displacing it from the center
axis of the device is necessary. The general equation which describes this
displacement (for the effective radial potential energy; xy plane) for any multipole
device is defined as:
(3-56)
where V#ff is the effective trapping potential in the xy plane, n is the number of
poles, e is the charge on the ion, V0 is the amplitude of the rf voltage (zero-to-
peak), m is the mass of the ion, u is the rf frequency, r0 is the inscribed radius,
and r is the displacement distance of the ion from the center axis of the
device.221227
A plot of effective trapping potential (Vctf) versus ion displacement from the
central axis (r) of three different rf-only devices (quadrupole n=2, hexapole n=3,
and octopole n=4) is shown in figure 3-4. The octopole parameters used are
seen in table 3-1. These parameters are the actual design specifications of the
rf-only octopole developed in our laboratory. All parameters for the three
multipole systems in table 3-1 are constant except for the value of n. The plot
shown in figure 3-4 is for the +3 charge state of bovine insulin. The radial
potential of the octopole has steep repulsive walls which approximate the ideal

Table 3-1. Parameters for the radial potential energy equation.
Device
n
e
w (MHz)
r0 (mm)
V0 (V)
m (g/mol)
Quadrupole
2
+5
1.659
2.94
250
5733
Hexapole
3
+ 5
1.659
2.94
250
5733
Octopole
4
+5
1.659
2.94
250
5733
158

Graph of the trapping potential function (Veff versus r the ion displacement) for a quadrupole (n=2),
hexapole (n=3), and octopole (n=4). The curves represent the +5 charge state of bovine insulin
(average molecular weight 5733 g/mol) where V0 is 250 V, w is 1.659 MHz, and r0 is 2.94 mm for
each multipole device.
Figure 3-4:

120
100
80
60
40
20
0
Ion Displacement from the Central Axis (mm)
160

161
case of the square well. This means ions of differing m/z ratios can be
transmitted to the ion trap at a constant translational energy (due to the large flat
bottom" portion of the well, where ion kinetic energies have only small
perturbations), which yields higher transmission and trapping efficiencies. For the
quadrupole and hexapole, the radial potential is more triangular in shape and,
therefore, not as many ions can be transferred at constant kinetic energy thus
leading to a decrease in ion injection efficiency. This phenomena can be
explained by the dependence of the radial potential on the normalized ion
displacement (r/r0)2"'2. The effective radial potential in an octopole is proportional
to r6, which provides a large trapping volume for the ions of interest.221-227 In the
case of the other extreme where n=2 for the quadrupole, the radial potential is
proportional to r2 with a maximum trapping energy one-fourth that of the octopole.
Other advantages of the rf-only octopole include the ability to shield low energy
ions from stray electric fields (or contact potentials) and reduced effects from
space charging.221-227
Assembly and Construction
A unique method for the assembly and construction of the rf-only octopole
was developed by Joe Shalosky and the author at the University of Florida. The
four solid support pieces for the octopole (see figure 3-5) were made from
Delrin. Delrin was used because the material can be threaded and tapped (e.g.,
machinable) and the surface friction generated by Delrin as it rubs against

Figure 3-5: Top and side views of the Delrin supports used in the construction of the rf-only octopole. The
outer diameter of the supports fit the baffle wall, ESI interface and the analyzer/octopole support
assembly.

Delrin Supports
0,200
~r
1,061
1
//
chamfer
ends

164
stainless steel or anodized aluminum is negligible. The ability of the material to
slide easily against these surfaces is important because of the design (discussed
in detail in chapter four) of the ion trap analyzer assembly, which can be rotated
to facilitate laser alignment. Caution must be taken with Delrin in terms of
manifold heating since the material loses its rigidity at 80 C. The four solid
support pieces were positioned on the ion trap analyzer side, at the baffle wall,
for the electrical connections and at the electrospray interface.
Each of the eight stainless steel mounting assemblies was made from a
single piece of stainless steel, as shown in figure 3-6. The contact points for the
octopole rods were machined to the exact radius of the rods to facilitate the spot
welding process, which therefore limits the variability of r0 for the 30 cm length of
the rods. Two stainless steel mounting assemblies were placed on each Delrin
(front and back) support, offset by 45. Each assembly contained four contact
points so that opposing pairs of rods would have the appropriate rf potentials
(e.g., phase) applied. Each stainless steel mounting assembly was attached to
the Delrin support via four stainless steel screws.
The Delrin support and stainless steel mounting assemblies were then put
together. The eight rods (all of length 30 cm) were then positioned on the contact
points by sliding the completed mounting assembly over the rods where a teflon
plug/spacer was used to push the rods securely against the contact point
surfaces (figure 3-7). To ensure that the ends of the octopole rods were aligned
in the same plane, an alignment jig with eight precision drilled holes was used on

Stainless steel mounting assembly design used for alignment/assembly of the octopole rods. The
extruded tabs are for attachment of the individual rods (e.g spot welded) for generation of the
correct octopolar field. This condition is obtained when the stainless steel mounting assemblies
are offset 45 from one another on the front/back surface of the Delrin support pieces.
Figure 3-6:

Stainless Steel Mounts
Weld Lip
00,418
00,079
ctopole Rods
ctopole Rods
Attached on
the Opposite Side
00,079"
Mounting Holes
00,073"
Rod Through
Holes
00,163"
Side View (After Initial Cut)
-0,062'
00,938"
0,418" Weld Lip
*-0.100"

Figure 3-7:
Schematic diagram of the teflon plug spacer, brass alignment jig, and brass alignment ring used
for proper alignment of the eight octopole rods during the spot welding process.

Brass Ring
Top View
Side View i0.125
0.500"
I
Teflon Plug
Top View
Q-00.234
Chamfer
r \
1.000"
v. 2
l | 0.234"
End View
Alignment Jig
168

169
one end of the assembly. Rods that are out of alignment (e.g., ends not in the
same plane) can introduce fringing fields and forces in the z-direction of ion
motion, thus reducing ion transmission efficiencies. The other end of the rods
had a brass ring (placed approximately 1 inch from the end-plane) with eight
precision drilled holes to ensure the correct spacing of the rods during the
welding process.
To apply the rf field to the rod assembly, two stainless steel pins were
welded onto the solid support assembly. These pins were designed to accept
copper-beryllium connectors from the rf supply/flange. A complete diagram of the
octopole assembly can be seen in figure 3-8.
Factors Affecting Ion Transmission
A theoretical discussion of the factors that influence ion transmission is
particularly important when evaluating and designing an rf-only octopole.
Parameters such as initial entry angle conditions, rf frequency, rf amplitude,
kinetic energy (transverse direction or xy plane) and pressure must be considered
when interfacing an ion source to a mass spectrometer via an rf-only multipole.
Although the author spent a copious amount of time deriving and understanding
the equations of motion for the octopole, to date there is no simulation program
for which our design can be fully evaluated. With the upcoming introduction of
Simion ver. 6.0 for Windows252, the design parameters and equations from the
above section could be used to develop a rather versatile simulation program.

Figure 3-8:
Schematic diagram of the complete octopole assembly with Delrin supports, stainless steel
mounting assemblies, and octopole rods. The inscribed radius (r0) is 2.94 mm.

Octopole Assembly

172
Simulation results from an rf-only collision cell designed by Davis and
Wright are presented in order to demonstrate the ion transmission properties of
the rf-only octopole compared to those of the traditional rf-only quadrupole.253
The validity of the computer modeling program used by Davis and Wright was
tested by evaluating the standard configuration (for the rf/dc quadrupole) of the
Concept ISQ mass spectrometer from Kratos Analytical (Manchester, UK) and the
rf-only collision cell of a ZAB-EQ mass spectrometer (VG Fisions, Manchester,
UK). The dimensions of the quadrupole and octopole discussed below are r0=4.0
mm (inscribed radius) with a total rod length of 20 cm. The multipole coordinate
system used for these studies is shown in figure 3-9.253
Initial Entry Angle Conditions
The trajectories for an ion of m/z 1000 which enters the multipole device(s)
1 mm off axis at angles of 0, 2, and 4 degrees (i9 relative the x-axis is 45), can
be seen in figure 3-10253. In this plot the abscissa is defined as the displacement
(| Z |) of the ion from the central axis and is plotted against the ions longitudinal
position in the device. From figure 3-10a (rf-only quadrupole) the amplitude of the
trajectories increase with increasing entry angle, but the wavelength of secular
motion and phase of the trajectories are similar. For the case of the rf-only
octopole shown in figure 3-10b, entry angle dependence on ion trajectory
amplitude is approximately the same as that observed for the rf-only quadrupole.
However, the effects on the wavelength of secular motion and corresponding

Multipole coordinate system used for simulation studies: (a) designation of r0 for a multipole (e.g.
quadrupole) device, (b) coordinate axis system (rectangular) with the xy plane denoted in the plane
of the page, (c) off-axis angle for ion entry into the multipole device, 0 is relative to the x-axis (xy
plane). Adapted from reference 253.
Figure 3-9:

174

Figure 3-10: (a) Entry angle dependence (at 0, 2, and 4) in an rf-only quadrupole for m/z 1000 with Vo p=400
V and ion entry 1 mm off axis, (b) Entry angle dependence (at 0, 2, and 4) in an rf-only
octopole, conditions same as in (a). Adapted from reference 253.

Displacement (mm) |Z|
176

177
phase are much more pronounced. The smaller the angle of entry for an ion into
the octopole, the longer the wavelength for the secular motion of an ion as it
traverses the length of the octopole.253
Figure 3-1 1253 shows the dependence of ion trajectories as they enter the
multipole devices from different distances off the center axis. As with the previous
example, the wavelength and phase of secular motion in the rf-only quadrupole
are approximately the same, while for the case of the rf-only octopole longer
wavelengths of secular motion are observed for an ion which enters the octopole
closer to the center axis.
The authors reported that the results of these two experiments on the beam
shape of an ion packet as it exits these multipole devices are twofold.253 For the
rf-only quadrupole, where the fundamental wavelength and phase of secular
motion show little variation, the net secular motion exhibits some form of coherent
oscillation. Therefore, the image the ion packet makes as it exits from the
quadrupole will depend on the average distance the ions (of a particular m/z) are
away from the central axis of the device. The smaller the image, the better the
focusing properties of the rf-only quadrupole for efficiently transferring the ion
packet to another analyzer (e.g., quadrupole mass filter with rf/dc voltages or
quadrupole ion trap). The ion beam image for a series of different m/z ions will
be much more complex than that of the single m/z case. Here, the wavelength
of secular motion is different for the different m/z ions and will produce a much
more complex beam image at the exit of the quadrupole. These different beam

Figure 3-11: (a) Ion displacement dependence (at 1, 2, and 3 mm) in an rf-only quadrupole for m/z 1000 with
Vop=400 V and an ion entry angle of 0. (b) Ion displacement dependence (at 1, 2, and 3 mm)
in an rf-only octopole, conditions same as in (a). Adapted from reference 253.

Displacement (mm) |Z|
179

180
images for different m/z values may lead to mass discrimination, which can have
a profound effect on total ion transmission.253,254 Since the wavelength of secular
motion in the rf-only octopole varies widely with entrance angle and position, the
image created by either single or multiple m/z values will have a "smeared"
appearance due to the noncoherence of the secular motion. This noncoherent
secular motion does not produce the mass discrimination effects which can be
observed for the rf-only quadrupole.253
RF Amplitude
A plot of ion displacement versus multipole length for varying rf voltages
is seen in figure 3-12.253 The secular wavelength of ion motion for both the rf-only
quadrupole and octopole is shown to decrease for increasing values of rf voltage.
This result was expected theoretically since with increasing rf voltages both q and
P values increase, while the secular wavelength {A) decreases. This means the
ion will increase its trajectory to the point where it will become unstable (as
defined by the Mathieu stability diagram for the 4-pole or 8-pole device). The
limiting point for the rf-only quadrupole is q2=0.908, whereas in the case of the
octopole222,237, q470. Therefore, the octopole can be operated at much greater
rf voltages and requires very little, if any, tuning for a wide range of m/z ratios
(see figure 3-12). However, for larger entrance distances (>0.5 mm) off axis, the
tendency for ejection of ion from the octopole approaches that of the quadrupole.
This phenomenon can be explained by examining the force on a given ion in the

Figure 3-12:
(a) RF voltage dependence (v=100, 200, and 300 V) in an rf-only quadrupole, q2 values are 0.433,
0.866, and 1.30 respectively. Conditions for m/z 100 are ion entry at 0.25 mm off axis at an angle
of 0.5 and an rf frequency of 1.2 MHz. (b) RF voltage dependence (V=100, 200, 300, 500, and
1000 V) in an rf-only octopole, q4 values are 3.46, 6.92, 10.4, 17.3 and 34.5 respectively.
Remaining conditions as in (a). Adapted from reference 253.

Displacement (mm) |Z|
182
0,- r0 | (a)
¡ 300 V
S f
i
2-0 h I

183
various multipoles. The force on the ion in the quadrupole has a linear
dependence on displacement from the center of the device. The force exerted
on the ions in the octopole, however, is cubed with respect to the distance off the
center axis of the device. At large distances off the center axis, trajectories can
become quite large and ion motion can become unstable as shown in figure 3-
13.253
In this case, the authors report that the best possible operating parameters
for the rf-only octopole would be for low ion entry angles and small off-axis
entrance distances.253 The instrument designed in our laboratory has the
octopole placed as close as possible (0.015" away from the skimmer cone) to the
electrospray interface. By placing the rf-only octopole this close to the skinner
cone, a larger cross section of the beam image is near the center axis of the
device and the divergence angle of the entering ions is minimized. This, coupled
with the incoherent oscillations observed for octopole operation at various m/z
ratios, make it an excellent alternative to traditional dc lens or rf-only quadrupole
ion injection systems.
RF Frequency
As the rf frequency applied to the octopole is increased, the wavelength of
the secular motion also increases as seen in figure 3-14. This agrees well with
the fundamental relationship observed between the secular wavelength of motion
and the applied rf frequency. The fundamental wavelength of motion for an

Figure 3-13: (a) RF voltage dependence (V=100 and 150 V) in a quadrupole cell, q2 values are 0.43 and 0.65
respectively, m/z 100, ion entry 1 mm/3, 1.2 MHz. (b) RF voltage dependence (V=100 and 150
V) in an octopole cell, q4 values are 3.46 and 5.20 respectively. Remaining conditions as in (a).
Adapted from reference 253.

Displacement (mm) | Z |
185
(a)
(b)
4-0
2 0
50
100
200
Length (mm)
150

Figure 3-14: RF frequency dependence (w=0.6, 1.2, and 2.4 MHz) in an rf-only octopole; q4 values are 13.9,
3.47, and 0.87 respectively. Ion entry conditions are for m/z 100 at 1 mm off axis and a 1 angle,
rf voltage was 100 V0P. Adapted from reference 253.

Length (mm)
150
200
187

188
octopole seems to follow the same relationship(s) as those of the rf-only
quadrupole.253,254 In this case, the fundamental wavelength of motion (4^) is
defined as:
lab
sec
f
(3-57)
sec
where vlab is the longitudinal velocity and is the secular frequency of motion.
The secular frequency of motion for a quadrupole is related to the fundamental
applied rf frequency by:
= f
'sec
q2
2y/2
(3-58)
with f defined as the applied rf frequency and q2 defined for a quadrupole as:
=
4eV
_ 22
mo) r0
(3-59)
Combining equations 3-57, 3-58, and 3-59 yields for the fundamental secular
wavelength:
'sec
= /mE
lab
4u2r02f
eV
(3-60)
here is derived from the standard relationship of kinetic energy and velocity
with v= v,ab.222'253'255 Since the square of the ions secular frequency is inversely
proportional to the value of q (equation 3-59), a significant reduction in rf drive
frequency can lead to unstable ion trajectories as seen in figure 3-14.253

189
Kinetic Energy
The kinetic energy arguments given here are relative to the transverse
kinetic energy (in the xy plane), which depends on how far the ions move off the
center axis of the rf-only device 253 In figure 3-15253 is shown a plot of the secular
wavelength of motion in an rf-only octopole, with an accompanying plot of
transverse kinetic energy of that motion as a function of distance traveled down
the octopole. As the ion moves to a point farthest away from the center axis of
the octopole, a maximum (for the oscillations) in transverse kinetic energy is
observed (see figure 3-15).253 These observed oscillations at large excursions
arise from contributions to ion motion due to higher order harmonics and not from
the ions fundamental (secular) frequency of motion256
When compared to the rf-only quadrupole (all other conditions being the
same), the magnitude of the transverse kinetic energy oscillations in the rf-only
octopole is much less than in the quadrupole (see figure 3-16). For the constant
conditions of rf voltage and frequency reported in figure 3-16 by Davis and
Wright253, the ion in the rf-only quadrupole approaches its limiting value with
respect to the Mathieu stability parameter at q2=0.908. For the rf-only octopole,
the limiting q4 values from the Mathieu stability diagram237 is approximately 70;
therefore, a significant reduction in transverse kinetic energy is observed even
with operation at higher rf voltages (e.g., 1000 V rf still gives transverse kinetic
energy values less than 1 eV for the simulation shown in figure 3-16).253 It is
advantageous to keep transverse kinetic energy to a minimum so as to keep the

Figure 3-15: (a) Secular motion of an ion in an rf-only octopole, m/z 1000, V0.P=400 V, ion entry 0.5 mm, angle
of 3, w=1.2 MHz. (b) Transverse kinetic energy in an rf-only octopole, conditions as in (a).
Adapted from reference 253.

Transverse Kinetic Energy (eV)
O O
200
Displacement (mm) \Z\
*!

CO

Figure 3-16: Ion entry dependence (at 0.5 and 1.0 mm) in an rf-only quadrupole, m/z 100, V0P= 100 V, ion entry
0, io= 1.2 MHz. (b) Ion entry dependence (at 0.5 and 1.0 mm) in an rf-only octopole, Remaining
conditions as in (a). Adapted from reference 253.

Transverse Kinetic Energy (eV)
193
(a)
(b)

194
beam image exiting from the octopole as small as possible. This reduces the exit
angles of the ions as they leave the device, thus leading to improved ion
transmission properties and a well defined translational ion energy.253
Collisional Focusing
Collisional focusing of ions in an rf-only device was first reported by
Douglas and French.257 In their experiments, an rf-only quadrupole was used to
inject ions from both atmospheric and electrospray ionization sources into a mass
analyzing (rf/dc) quadrupole. For ion injection energies on the order of 1 to 30
eV, ion transmission efficiencies were found to increase through the rf-only
quadrupole at pressures up to 8x1 O'3 torr. The increased ion transmission
observed was found in direct contrast with those predicted by traditional
scattering models. The collisional focusing capabilities of the rf-only quadrupole
were found to be quite similar to those observed in the three dimensional
quadrupole ion trap.257
Douglas and French reported that collisional focusing improved with
increasing ion mass and not mass-to-charge ratio. The improved ion transmission
was attributed mainly to a significant loss in the ions axial kinetic energy. In
addition, higher mass ions (at the same q value of lower mass ions) were
transmitted with higher rf voltages applied to the rods, thus leading to a greater
well depth and better confinement along the central axis of the rf-only quadrupole.
Monte Carlo simulations of this phenomenon provided good agreement with

195
experimental values for reasonable collision cross sections.257
Extension of this work to nonlinear devices such as the rf-only hexapole
and octopole, has not been accomplished. Results from the rf-only quadrupole
study suggest strongly the coupling of atmospheric pressure ionization sources
with rf-only devices for ion injection. Future studies in our laboratory will focus
on the effect of collisions on ion transmission in rf-only multipoles with n>2.

CHAPTER 4
ELECTROSPRAY/ION TRAP INSTRUMENTATION:
DESIGN AND OPERATION
General Overview
This chapter presents the general design and characterization of a novel
electrospray ionization/ion trap mass spectrometer for the photodissociation of
biological macromolecular ions. The first section discusses design considerations
(vacuum manifold, ESI ion source, ion injection, and analyzer/detector assembly)
associated with the electrospray/ion trap system. This is followed by a discussion
of the experiments used to characterize system performance, including an
evaluation of the rf-only octopole ion injection design, absolute sensitivity,
collision-induced dissociation, mass isolation, and negative ion operation.
Instrument Design
This section contains the design considerations necessary for the
construction and operation of an rf-only octopole ion injection system for use with
an electrospray/ion trap mass spectrometer. Also, special considerations
concerning vacuum system design for use with the IRMPD process are
addressed. Pertinent modifications to an Analytica ESI source for coupling to the
rf-only octopole ion injection system are discussed in order to understand the ion
196

197
transmission properties of this unique arrangement. Next, a short review of the
pertinent design considerations for rf-only octopoles (from chapter 3) is presented
along with the exact specifications for operation of the rf-only octopole used in
this instrument. The final parts of this section cover the design and construction
of the analyzer and detector assemblies.
Vacuum Manifold and Pumping System
A Finnigan MAT ITMS frame was used as a base for construction of the
vacuum manifold for the ESI/ion trap. The standard Finnigan MAT manifold and
wooden table top were removed from the frame assembly to allow for placement
of the new vacuum manifold. Three new table top sections and one side panel
were machined from 0.5" thick aluminum panels to construct the new frame
assembly. A series of 1" square aluminum rods were then used to elevate and
allow level placement of the vacuum manifold on the instrument frame. This level
placement (parallel to the floor within 0.02") was necessary to help eliminate
any laser alignment problems associated with future photodissociation
experiments. The table top panel used for mounting the vacuum manifold was
constructed with three precision-machined holes/slots to accommodate two
turbomolecular pumps and allow for placement of a Conflat adapter flange
needed for the electrospray interface. A schematic diagram of the table top
design is shown in figure 4-1.

Figure 4-1:
Mounting plate design for attachment of the vacuum manifold to the instrument support table.
Circular portions indicate 8" outline of 500 L/s turbomolecular pumps.

Mounting Plate for
Top View of Mounting Plate
(Vertical Dimensions)
Manifold Assembly
Top View of Mounting Plate
(Horizontal Dimensions)
units are in inches
199

200
The vacuum manifold was assembled from pieces of Finnigan MAT 4500
and TSQ 46 vacuum manifolds. The manifold was made entirely of stainless
steel, with the majority of flange connections made with Conflat (copper
gasket/knife edge) style matings to achieve the highest vacuum possible (needed
to do IR photodissociation). A total of 10 connection ports were made to the
vacuum manifold. The Conflat connections included an 8.0" port for the
analyzer/detector assembly, a 4.5" port for the rf (drive frequency) feedthrough,
a spare 4.5" port for a second ion gauge, a 6.0" port for a Conflat adapter flange
needed for the ESI interface, another 6.0" port for electrical connections to the ESI
interface (two heater connections, two temperature sensor connections, and a
heated capillary offset connection) and rf-only octopole (two low voltage rf
connections, <5kV), and a 2.75" port for a removable laser window system. 0-
ring connections included two large 8.0" ports (located on the bottom of the
manifold) for attachment of two 500 L/s turbomolecular pumps and two 5.0" ports
(top of manifold) for glass windows to view the differentially pumped regions of
the system. A schematic diagram of the vacuum manifold is presented in figure
4-2.
The pumping system for the manifold consisted of two 500 LVs
turbomolecular pumps (model TPH 510, Balzers, Hudson, NH) mounted to two
turbomolecular pump compensators (Balzers, Hudson, NH) in order to damp the
vibrations associated with turbomolecular pump operation. These pump
vibrations were minimized to limit their effects on any optical lens/mirror systems

Figure 4-2:
Top view of the vacuum manifold design for the ESI/ion trap instrument. The center line of the
manifold was accurate to within 0.010" from end to end.

Window Flange
ZnSe Window Flange
Blank Flange

203
mounted to the ITMS table/frame (for use in photodissociation experiments). The
pumps and associated compensators were then mounted to anodized aluminum
o-ring adapter flanges attached to the main vacuum chamber. The
turbomolecular pumps were controlled by two Balzers TCP 300 power supplies
wired directly into the ITMS electronics. Two Alcatel (Alcatel Corporation,
Hingham, MA) mechanical pumps (model 2012A) rated at 300 L/min were used
as fore pumps for the turbomolecular pump system.
A Bayard-Alpert ion gauge (Granville-Phillips, Boulder, CO) was connected
to a modified anodized aluminum o-ring adapter flange located directly under the
analyzer side of the vacuum manifold. A modified Cajon fitting was used to hold
the ion gauge in place directly under the rear part of the ITMS frame to protect
it from damage. The ion gauge controller was a Granville-Phillips model 280
gauge controller (Boulder, CO) equipped with an out-gas control used directly
after system pump-down.
To facilitate alignment of the ESI source, rf-only octopole, and the
analyzer/detector assembly, two precision machined centering rings were welded
inside the vacuum chamber. The front centering ring supports the baffle wall
assembly used both to support the rf-only octopole and to allow differential
pumping (by the two 500 L/s pumps) between the ESI source and the
analyzer/detector assembly. The second centering ring functions as a
support/center for the analyzer/detector assembly (relative to the ESI source).
The accuracy of the center axis of the manifold running from the front 6.0" flange

204
to the rear 8.0" flange was 0.10" (see figure 4-2). Later experimental evidence
and theoretical support from chapter 3 suggested that this alignment accuracy
contributed greatly to the success of this project, especially after ion trap analyzer
cleaning or any other process which required disassembly/reassembly of the
major system components (ESI source, rf-only octopole, analyzer/detector
assembly). Figure 4-3 shows the complete vacuum manifold assembly including
ESI source, rf-only octopole, and analyzer/detector assembly.
Initial system vacuum tests to check the base pressure were performed
over a one-week period (initial pump-down), with the manifold heated to 100 C.
The observed base pressure after this time was 1.9x1 O'8 torr. Under normal
operating conditions, the vacuum manifold was not heated due to the presence
of the Delrin supports used in the construction of the rf-only octopole.
Electrosprav Ion Source
The electrospray ionization source used with the ESI/ion trap system was
a standard Analytica (Analytica Inc., Branford, CT) source equipped with a
stainless steel heated capillary (modification performed at Finnigan MAT) as seen
in figure 4-4. The purpose of the heated capillary is to help droplet desolvation
for the electrospray ionization process. The heated capillary was 4.580" long with
an outer diameter of 0.308" and an inner diameter of 0.020". The heated capillary
temperature was controlled by an Omega 6000 (Omega Engineering, Stamford,
CT) series temperature controller with a 220 V output, previously used for

Figure 4-3:
A simplified schematic diagram (top view) of the ESI/ion trap system including the electrospray
source, octopole assembly, ion trap analyzer and detector assembly.

ZnSe Window Blank Flange
206

Figure 4-4:
Analytica electrospray ionization source equipped with a heated capillary (as modified by Mark Hail
at Finnigan MAT, San Jose, CA).

Vacuum Connection
208

209
controlling the Finnigan MAT ITMS manifold temperature. The 220 V output was
stepped down via a transformer to 24 V (< 2 amp output current). Typical
operating temperatures were in the 170 to 215 C range. For this temperature
range, little if any thermal degradation or induced fragmentation of the [M+nH]n+
(where n=1, 2, 3...) ion(s) was observed.
The electrospray source was pumped by two 500 L/min rotary pumps
(model UNO 016B, Balzers Inc., Hudson, NH). A Harvard Apparatus model 22
syringe pump (South Natick, MA) was used for the direct infusion of samples into
the electrospray source.
To facilitate coupling of the rf-only octopole to the ESI source, several
design changes were made. First, to simplify the ion optics and obtain maximum
ion transfer efficiencies, the second skimmer cone, lens L2 and lens L3 were
removed from the electrospray source (see figure 4-4). An adapter ring was then
constructed from stainless steel to extend the remaining skimmer cone 0.250
forward. The adapter ring was mounted between the base plate of the
electrospray head and the skimmer cone, employing o-ring seals between the
skimmer cone and base plate assemblies. The alignment tool used to determine
the exact distance from the heated capillary exit to the skimmer cone was then
modified to compensate for the stainless steel adapter ring to maintain a constant
distance between the heated capillary and the skimmer cone at 3.5 to 4.0 mm.
The electrospray needle was also moved forward to compensate for the adjusted
heated capillary position. In figure 4-5 is shown the modified electrospray source

210
described above.
The tube lens located at the end of the heated capillary (see figure 4-5)
was used to gate ions into the rf-only octopole (via the skimmer cone). Control
of the voltage(s) applied to the tube/gate lens was accomplished by using the
gate control circuit from the ITMS electronics. The circuit was modified so that
both positive and negative ions could be gated efficiently. For the analysis of
positive ions, a variable positive voltage is applied to the tube/gate lens (10-120
V) to focus the ion beam. To control the pulse width of the beam, a negative -180
V potential is applied to the tube/gate lens to stop ion transmission. For negative
ions, a variable negative voltage (-10 to -100 V) is used for ion focusing and a
+ 180 V signal is used to stop ion transmission (e.g., control the pulse width).
All vacuum electrical connections (capillary heaters/temperature sensors,
capillary offset voltage) for the ESI source were made through a 6" Conflat flange
(equipped with an Amphenol connector) located to the left of the ion source (see
figure 4-3). High voltage for the electrospray needle and drying gas (N2, air, or
02) were controlled manually by an external power/gas distribution unit (Finnigan
MAT, San Jose, CA).
RF-Onlv Octopole
In the previous chapter a detailed description of the theoretical and
practical design considerations of rf-only octopoles was given. This section

Figure 4-5: The modified Analytica electrospray ion source showing the elimination of the second skimmer
cone, lens L2 and lens L3. Also shown is the stainless steel adaptor ring used to extend the first
skimmer cone region 0.250" forward for coupling to the rf-only octopole.

Vacuum Connection
ESI Tube Lens
ESI Needle
Heated Capillary
212

213
focuses on the tuning of the rf matching circuit needed for operation of the
octopole device. It is assumed the reader is familiar with basic electronics and
the principles of octopole operation and design. The procedures employed for
determination of the octopole operating frequency are based on the Standing
Wave Ratio (SWR bridge) method as developed by William J. Fies, Jr. at Finnigan
MAT Corporation (San Jose, CA).258
The rf power supply and corresponding components used to drive the rf-
only octopole were obtained from a Finnigan MAT (San Jose, CA) 4500 single
quadrupole mass spectrometer. The basic rf circuitry from the Finnigan MAT
4500 system consists of an rf amplifier with a low-pass filter to reduce harmonic
frequencies which may be present (see figure 4-6). The amplifier and low-pass
filter are designed to drive a 50 Q load for the tuned circuit. The purpose of the
rf circuit is to produce a specific rf voltage for mass analysis (for the tuned circuit)
with minimum power consumption. The circuit consists of a coil with inductance
L and specified capacitance Cr (corresponding power loss specified by Rn which
occurs almost entirely in the coil), and matching capacitor(s) Cm. Tuning of the
resonant frequency for the system is accomplished by adjusting either the
inductance L or the capacitance Cr. The circuit is considered tuned" when the
circuit is resonant at the frequency of the rf driving power and R, is matched to
the generator resistance of the power amplifier.258
For the case of the rf-only octopole designed here at the University of
Florida, the capacitance of the device is unknown. For the purpose of

Figure 4-6:
Simplified schematic of the Finnigan MAT 4500 rf circuit showing the rf power amplifier, low-pass
filter, matching capacitor, and remote rf tuning unit. Adapted from reference 258.

rf Power Amplifier
50 ohm cable
Low-Pass
Filter
rf Timing Unit
Remot-rf
215

216
determining the appropriate operating rf frequency the load was assumed to be
50 £2. Since the capacitance of the device is unknown (e.g., depends on the
material the electrodes are made from, electrode spacing, and electrode surface
area), there could be a large amount of rf power dissipated from the rf coil (tuned
circuit not optimized) leading to a significantly reduced maximum power output
from the rf amplifier. Since the octopole will be operated in the rf only mode,
maximum voltage levels of over 2000 V^p are not needed, so an optimized
matching circuit is not critical.
A general schematic of the SWR bridge circuit used for tuning is shown in
figure 4-7. The circuit consists of a Wheatstone bridge composed of three equal
resistors and the octopole with an unknown impedance (Z). The resistor values
are chosen at 50 £2 so that the impedance through the octopole will be 50 £2 at
the appropriate operating frequency (e.g., current through the circuit is 0).
To determine the proper operating frequency, a function generator is
needed which can scan a range of frequencies typically from 0.5 to 4.0 MHz. In
addition, the function generator should be able to produce a marker TTL signal
for determination of the optimum rf frequency. A dual channel oscilloscope is
used to monitor the frequency and marker TTL signals. A schematic diagram of
the test set-up is shown in figure 4-8. A Stanford Research System DS345
arbitrary waveform generator (Sunnyvale, CA) was set to scan from 0.5 MHz to
2.0 MHz at a 50 ms scan speed in order to determine a coarse adjustment range
for the optimum frequency. When the frequency synthesizer is scanned, the SWR
trace appears on the oscilloscope with a "dip" at the frequency that represents the

Figure 4-7:
General schematic of the SWR bridge circuit. The three resistors (labeled R) are of equal value.
The impedance of the octopole (z) is unknown, and is defined as the complex impedance of the
input to the rf tuned circuit. For a 50 n load, the values of R in the bridge are 50 Q, thus giving
0 current through the detector when the rf circuit is in tune. Adapted From Reference 258

218

Figure 4-8:
Block diagram of the SWR bridge test set-up. The Stanford Research Systems function generator
produces the frequency sweep and marker TTL signals used to determine the operating frequency
of the rf-only octopole.

x drive
220

221
correct tune for the rf circuit. Once a coarse frequency reading of 1.6 MHz was
obtained (see figure 4-9a), the function generator was reset to scan from 1.6 to
1.7 MHz. The marker frequency was then adjusted to match the lowest point on
the SWR trace which indicated an optimum operating frequency of 1.659 MHz
(figure 4-9b).
To run the rf-only octopole at the highest frequency possible (for reasons
discussed in the previous chapter), the taps of the rf-coils were moved in 18 turns
on the 32 turn coils. This significantly increased the optimum rf frequency from
1.2 MHz to the 1.659 MHz value obtained above. The frequency increase can be
explained by the inverse relationship between the applied rf and the inductance
of the coil (e.g., a reduction of the coil inductance leads to an increase in the
applied rf frequency). Although this modification significantly reduces the gain of
the rf, the voltage requirements needed for operation in the rf-only mode are
significantly reduced compared to mass filter operation. A plot of the detected
rf versus the measured rf output (figure 4-10) shows a maximum rf amplitude of
approximately 2000 V^p. The rf output was measured with a standard scope
probe which has a significant amount of capacitance that will affect the tuning of
the rf circuit; therefore the 2000 V^p output is only an approximation of V^.
Another more complex solution to increase the frequency is to either adjust the
capacitance of the octopole device, or the matching capacitors or the rf circuit,
since there exists an inverse relationship between the applied rf frequency and
capacitance.

Figure 4-9:
(a) Scope trace from the frequency scan of the function generator in the SWR bridge test set-up.
The display range shown is from 1.520 MHz to 1.720 MHz. (b) Expanded view of the scope trace
from part (a). Channel 2 represents the marker frequency (TTL). The marker frequency indicates
the frequency of the function generator when the bottom of the sweep peak is aligned with the
marker TTL signal.

C \
Scale
vertical 50 mV/div
horizontal 20 kHz/div
v /
1.520
units of MHz
1.720
(b)
/ \
Scale
vertical 50 mV/div
horizontal 5 kHz/div
\ J
1.640
units of MHz
1.690
223

Figure 4-10: A plot of detected rf versus mass set voltage. The second y-axls on the right is the corresponding
rf output voltage. Since the scope probe has some finite capacitance, the rf output voltage is only
an approximation of the real output voltage.

Detected RF (V)
Mass Set (V)
RF Voltage (V

226
In figure 4-11 is shown a schematic diagram of the complete rf circuitry
used to control the rf-only octopole. The octopole offset (extraction voltage) and
mass set (a zero to ten volt set point for producing the rf amplitude) were
obtained from a Finnigan MAT TSQ 46 lens voltage controller. A Stanford
Research System model DS345 (used above) arbitrary waveform generator was
used to produce the 1.659 MHz frequency for the rf amplifier circuit.
Analyzer Assembly
The analyzer assembly was designed to fit on an 8" rotatable Conflat
flange. The rotatable flange was used to facilitate the alignment of the multipass
ring electrode with the incoming laser (pulsed or continuous wave) radiation.
Welded to the rotatable flange were four stainless steel rods which were used as
an optical rail system to facilitate movement of the analyzer assembly relative to
the laser window and detector arrangement. A 1 /4 Swagelock union was welded
into the 8" Conflat for direct coupling of a He buffer gas line or connection to a
pulsed-valve as described in chapter two of this dissertation. A Negretti valve
(Negretti LTD, Southampton, UK) was used to control the back pressure of He for
constant pressure or pulsed valve experiments. The 8" Conflat was also equipped
with three straight high voltage feedthroughs (rated to 20 kV) and 7 single-ended
MHV feedthroughs rated to 5 kV (Insulator Seal Incorporated, Hayward, CA). One
of the high voltage feedthroughs was used for the 20 kV applied to the off-axis
conversion dynode, while the seven MHV feedthroughs were used for the detector

Figure 4-11:
Octopole rf circuit for the ESI/ion trap instrument. The mass set and detected rf voltages are sent
through a comparitor which determines the power output of the rf to the rod assembly. The
octopole offset is applied to all eight rods through the rf circuit.

Amplified RF
Remote RF
Octopole Offset
Detected RF
Octopole
4500 QEM
TSQ 46 Lens
Voltage Supply
228

229
signal, electron multiplier voltage, exit tube lens, applied (e.g., quadrupolar or
dipolar) excitation signal (2) and pulsed-valve control (2). Four 0.5" diameter 4"
long brass rods were attached to the non-vacuum side of the 8" Conflat flange to
allow for proper clearance of the aforementioned feedthroughs when the analyzer
assembly was set upright on a lab bench for maintenance.
Two aluminum mounting plates were used to contain the ion trap analyzer,
exit tube lens and octopole mounting assembly. The mounting plates were drilled
out to accommodate the optical rail assembly and to pass the electrical
connections to the endcap electrodes and exit tube lens. The three holes drilled
for the ion trap analyzer were offset 13.3 from center, so that alignment of the
entrance aperture of the multipass ring electrode was parallel to the incoming
laser irradiation. This condition was achieved when the 8" Conflat was turned
such that (from a top view) opposing pairs of rods were in line with the plane of
the table top. The position of the mounting plates, and therefore the analyzer
assembly, was fixed by two set screws on each plate. A schematic diagram of
the aluminum mounting plate can be seen in figure 4-12.
The ion trap analyzer was from a standard Finnigan MAT ITMS, with the
entrance endcap replaced with an exit endcap. The two endcaps were positioned
such that the seven holes on both exit endcaps were symmetrically aligned. In
principle, this alignment will reduce the hexapolar non-linear field contributions to
the quadrupolar trapping field. In addition, the use of an exit endcap for an
entrance endcap allows placement of the rf-only octopole to within 0.050" of the

Figure 4-12:
The aluminum mounting plate assembly shown in a series of five diagrams: (1) general
dimensions for the mounting plates, (2) ceramic mounting holes to hold the ion trap assembly; the
holes are offset 13.3 from center to facilitate alignment of the multipass ring electrode with the
laser, (3) assembly pin hole dimensions for the stainless steel rods used to hold the analyzer
assembly together, (4) holes for the optical rail assembly, and (5) relief holes for the pulsed-valve
and electrical feedthroughs.

Analyzer Mounti
01.850'
Top View
Plate (Bottom)
/- 03.750'
Dimensions for
Inner and Outer
I.D., and Cut-outs
for Set Screws
2-56 tap, 0.0B9" drill
Number: 1
Material: Aluminum
Modifications: Anodize
All Units are in Inches
Page 1 of 5

Analyzer
Vlounting Plate
(Bottom)
Dimensions of Ceramic Mounting Holes
Number: 1
Material: Aluminum
Modifications: Anodize
All Units are in Inches
Page 2 of 5
232

Analyzer Mounting Plate
Assembly Pin Hole Dimensions
(Bottom
. 125
233

Analyzer Mounting Plate (Bottom)
Dimensions for
Mounting Rod
Holes
Number: 1
Material: Aluminum
Modifications: Anodize
All Units are in Inches
Page 4 of 5
Figure 12 continued
234

Analyzer Mounting Plate (Bottom)
Dimensions of PulsedVaive/Feed Through Holes
Number: 1
Material: Aluminum
Modifications: Anodize
All Units are in Inches
Page 5 of 5
235

236
endcap, which facilitates efficient transfer of ions from the octopole to the ion
trap).
The octopole was positioned by an aluminum mounting piece with an
internal diameter designed to brace the Delrin support at the end of the octopole
rod assembly. The aluminum mounting piece was designed with three large cut
outs to aid pumping around the analyzer region. The piece was mounted on top
of the ion trap analyzer assembly using three alumina ceramic supports. A
schematic diagram of the octopole support can be seen in figure 4-13.
The next part of the analyzer assembly was an exit tube lens made from
stainless steel to transfer the ions from the mass analyzer to the detector. This
exit lens consisted of a single piece of stainless steel with a reduced orifice tube
lens used to reduce field penetration effects from the 20 kV dynode. The back
side of this assembly was threaded to accept a variety of tube lens extensions.
The tube lens extension was needed to help transfer ions from the analyzer to the
detector because the physical constraints of the manifold would not allow
placement of the detector assembly within 2" of the exit endcap. The exit tube
lens was mounted to the bottom of the analyzer assembly again using alumina
ceramic supports. The exit tube lens assembly and corresponding extension
piece are shown in figure 4-14.
The analyzer stack was assembled in the following manner: (1) aluminum
mounting plate (bottom); (2) exit tube lens; (3) exit endcap; (4) multipass ring
electrode, (5) exit endcap (e.g., ion entrance), (6) octopole mounting assembly;

Figure 4-13: Octopole mount assembly for interfacing the octopole with the ion trap analyzer. The cut-out slots
are for increased pumping in the analyzer region.

Top V¡ew(s)
Octopole Mount
(Analyzer Assembly)
Side View(s)
All Units are in Inches
238

Three-part figure showing: (1) the tube lens assembly with beveled ends to limit field penetration
from the 20 kV dynode, and to aid in ion transfer to the detector, (2) variable tube lens extension
piece complete with threaded end for extension to the 20 kV dynode detector, and (3) the tube
lens electrical connection.
Figure 4-14:

Top V¡ew(s)
Tube Lens Assembly
Side View(s)
0.326
L
0.531
| j 0,389'
ifl iu
-ve=>
C
r540*
T tf=3
0.560'
Thread to 0,215
appropriate size
7T
All Units are in Inches
240

Tube Lens Extension
-3.000"
0.560"
R0.050"
1
Thread to
appropriate size
All units are in inches
Material: Stainless Steel
Figure 14 continued

Figure 14continued
242

243
and (7) aluminum mounting plate (top). The assembly was held together by three
stainless steel rods which were bolted through the two aluminum mounting plates.
Aluminum was used in the design of the mounting plates and octopole mounts
in order to reduce the strain (e.g., from weight) on the optical rail system. All
aluminum pieces were anodized with a 0.002" natural hardcoat finish. The
aluminum was anodized for several reasons including: (1) to ensure that all
support pieces would become non-conductors; (2) to reduce pump-down time
associated with trapped water and oxygen found on porous pure aluminum
solids; and (3) to increase the ease of movement of the analyzer assembly
against the stainless steel optical rail. A complete diagram of the analyzer
assembly is shown in figure 4-15.
The final piece of the analyzer assembly was an anodized aluminum
centering ring which mates directly with the centering ring in the vacuum
manifold. This allows for proper alignment of the ESI-octopole-analyzer assembly
of the instrument and provides additional mechanical support for the optical rail
system (see figure 4-15).
Detector Assembly
The detector used in the instrument consisted of a 20 kV off-axis dynode
and a continuous electron multiplier. This detector assembly was purchased
directly from Finnigan MAT (San Jose, CA) and modified to fit the ESI/ion trap
instrument. A stainless steel mounting bracket was used to hold the rear base

Figure 4-15:
Complete drawing of the analyzer assembly including the alignment ring, octopole mount,
aluminum mounting plate, ion trap, tube lens, and detector assembly.

Electron
245

246
plate (of the detector assembly) to the 8" Conflat flange. The mounting bracket
was drilled out to allow high energy ions and neutrals to pass through the
detector region and strike the surface of the 8" Conflat, thus reducing the
background noise (from stray ions) associated with the detection event. A Bertan
model 205B power supply (Hicksville, NY) was used for the 20 kV applied to the
dynode. The electron multiplier was controlled by the ITMS electronics. The
anode lead (signal out) from the multiplier was directed into the ITMS electrometer
circuit. In addition, the anode lead was made as short as possible and was
shielded with copper foil to reduce spurious noise in the detector signal. A
schematic diagram of the set-up used is shown in figure 4-16.
Photodissociation Set-Up
Two different C02 lasers were employed for the photodissociation
experiments described in chapter 5 of this dissertation. The first was a
continuous wave C02 laser described previously, and the second was a pulsed
C02 laser (Lumonics Series TE-860-4 Excimer, Ottawa, Ontario, Canada) capable
of a 3 J pulse (at 10.60 //m) for an approximate 3 ns pulse width. The duty cycle
for full power operation of the laser was 20 Hz (e.g., 50 ms). Each laser required
a different optical set-up depending on the type of photodissociation experiment
employed. For optics mounted on the instrument table, two compensators
(described previously) were used to damp vibrations generated by the two 500
L/s turbomolecular pumps.

Figure 4-16: Detector assembly with wire leads attached (indicated by dashed lines). The anode lead was
shielded with a copper foil (grounded to the manifold) so as to reduce background noise in the
electrometer.

8 Inch Conflat
248

249
The vacuum-air interface for both the pulsed and cw lasers consisted of a
2.75" flange modified to accept a 1.5" ZnSe window (Melles Griot, Irvine, CA). The
vacuum seal was made by two teflon rings placed on either side of the ZnSe
window.
The photodissociation set-up for the cw laser can be seen in figure 4-17.
A 13" x 20" optical table was constructed and used for mounting the various ion
optics. The 1 cm diameter unfocused beam was passed through a beam selector
(designed to pass IR radiation) and reflected at a 90 angle off a gold plated
mirror. The beam then entered the mass spectrometer through the ZnSe window.
A helium-neon (632.8 nm) laser (model 05 LLR 851, 5 mW power output, from
Melles Griot, Irvine, CA) was used to expedite the alignment of the cw C02 laser.
The helium-neon laser was placed on top of the instrument table and aimed
directly at the 1" beam selector (Melles Griot, Irvine, CA) which reflected the 632.8
nm light and passed IR (9.0 to 11.0 //m) radiation. This greatly simplified the
alignment process since the helium-neon laser did not need to be placed in line
with the cw C02 laser for beam alignment.
In the case of the pulsed laser (as seen in figure 4-18), the beam selector
was not used since the beam shape from the laser (approximately a 1" x 1.25"
square beam) exceeded the dimensions of the beam selector. As mentioned
previously, the beam was first reflected at a 90 angle off the surface of a gold
plated mirror. To focus the beam down to the appropriate size to pass through
the entrance aperture of the multipass ring electrode, a convex focusing mirror

Figure 4-17:
Photodissociation set-up for the continuous wave-laser. The beam selector allows 10.6 //m
radiation to pass directly through to the gold plated reflecting mirror, while the output of the
Helium-Neon laser (632.8) is reflected off the front surface of the beam selector.

Vacuum Manifold
[
y
J Ion Trap
-j J
1 1-
ir
\
\
Instrument Table Helium-Neon
co2
Gold Plated
Laser Beam
Reflecting Mirror
/ /
/
/
f
/
/
*
Optical Table
111 11 i 1111 1111111 11 1111111111 i 1111111
1111111 \.
/
s
Beam Selector
Apollo
cw CO 2
Laser

Figure 4-18:
Photodissociation set-up for the pulsed-laser. The beam selector was removed since the beam
width of the pulse laser exceeded that of the beam selector. Alignment was accomplished by
using a Helium-Neon laser and an additional gold plated turning mirror. A focusing lens was
added to make the beam diameter 0.2 cm as it enters the multipass ring electrode.

Vacuum Manifold
L
]d
J Ion Trap
[
"i ^ r
Focusing Lens
1
_
1
\
\
\
%
Instrument Table Helium-Neon
co2
Gold Plated
Laser Beam
Reflecting Mirror
/ /
/
/
*
*
/
/
Optical Table
111111 111 11111 i 1111 i i 111 i 1111 i 1111 i
111 111
/
/
Removable /
Gold Plated
Reflecting Mirror
C
!b
Lumonics
Pulsed CO 2
Laser
253

254
with a focal length of 4" was placed in line with the beam just in front of the ZnSe
window (see figure 4-18). This lens focused the beam down to a 0.5 cm spot
size. A coarse alignment was obtained by placing the helium-neon laser in line
with the laser beam. Fine adjustment of the alignment was performed by placing
burn-paper at the entrance aperture of the multipass ring electrode to determine
spot size and position. The major disadvantage of using the pulsed laser is in the
alignment procedure, since the instrument must be vented to ensure proper beam
position.
System Interconnections
The system interconnect diagram for the ESI/ion trap instrument is shown
in figure 4-19. The ion trap offset voltage was controlled by the selective mass
storage unit of the ITMS electronics. The balun circuit was floated and modified
to output 18 V^p. The mass set, octopole offset, endcap offset, exit tube lens, ion
gate, and ESI capillary offset voltages were controlled by a Finnigan MAT TSQ 46
voltage power supply. The ion gate voltage was controlled by a bipolar supply,
which was variable for focusing either negative or positive ions (depending on the
switch settings for the ion gate control board built and attached to the Power
Control PCB of the ITMS electronics) into the ion trap. The voltage ranges and
calibration data (using a Fluke 75 DVM, Everett, WA) for the power supply are
shown in table 4-1.

Figure 4-19:System interconnect diagram for the ESI/ion trap system. All electrical connections are shown.

256

Table 4-1. Voltage power supply and calibration data for ESI/ion trap instrument.
Device
Calibration (y=actual V, x=display V)
Font Panel Label
Voltage Range (V)
Mass Set

Quad 3 Offset
o
7
o
Octopole Offset
y=1.05(x) + 0.0
Quad 2 Offset
+ 12.1 <*-11.5
Endcap Offset
y = 1.04(x) + 0.0
2L3
+ 42 -42
Exit Tube Lens
y=1.02(x) + 0.001
3L2
+ 79 <* -150
ESI Capillary Offset
y=1.00(x) 0.009
3L1
+ 68 <* -145
Ion Gate
y=1.01 (x) + 0.006
3L3
+ 129 <* -131
Axial Modulation
y=3.58(x) 0.049

0 <* 21 (P-P0)*
* ac volatage
257

258
Instrument Characterization
This section describes the initial experiments performed on the ESI/ion trap
instrument used to evaluate system performance. It begins with a brief discussion
of the first test (El mode) of the rf-only octopole, analyzer, and detector
assemblies. An introduction to high mass analysis (mass range extension) follows
with a few remarks on instrument operation and theory. From there, a discussion
on ESI operation ensues with a few brief remarks on absolute sensitivity. The
next few sections cover characterization of various instrument parameters such
as octopole rf voltage, octopole offset, ion gate lens voltage, ion isolation, and rf
injection level (qmject). The chapter concludes with an MSn study on the
neuropeptide angiotensin I and a brief dialogue on negative ion operation.
Initial system tests of the rf-only octopole, analyzer assembly and detector
assembly were performed using electron ionization (El) mass spectrometry via an
external ion source. To perform these tests, the electrospray and corresponding
Conflat adapter flange were removed from the instrument and a Rnnigan MAT
4500 El/Cl ion source and corresponding flange extension were added to the
system. The filament and lenses for the El source were controlled via the 4500
quadrupole electronics module used for operation of the rf-only octopole. Lens
assembly L3 (quadrupole entrance lens) was modified with a tube lens extension
to act as an ion gate for El operation with the ion trap.
The first mass spectrum (El mode) of perfluorotributylamine (PFTBA)
obtained with the new instrument is shown in figure 4-20. The instrument

Figure 4-20:
El mass spectrum of PFTBA taken with the rf-only octopole/ion trap system. The high mass end
of the spectrum is very intense compared to that obtained from quadrupole instruments. The
increased sensitivity on the high mass side is characteristic of ion trap spectra in general.

800
600
400
200
0
m/z 131
m/z 69
m/z 100
m/z 502
11111111111 M 1111111111 r i| n n 11111111 n 11 in 11111111111111
50
100 150 200 250
300 350
m/z
400 450 500 550 600
260

261
parameters for the spectrum are shown in table 4-2. The acquisition was for a
low-mass cut-off m/z 30 and a He bath gas pressure of 1.0x10^ torr. The
spectrum represents the average of 5 microscans. The PFTBA spectrum
obtained is typical of ion trap spectra in general, which show enhanced sensitivity
for the high mass end of the spectrum.
In the case of El operation, the ion gate (quadrupole entrance lens L3) is
at a negative voltage for pulsing ions into the octopole and at a high positive
voltage to prevent ions from entering the octopole. This is the opposite case for
electrospray operation (discussed below) where a positive voltage is used to gate
positive ions into the octopole and a high negative voltage is used to stop the
ions from entering the octopole. For the case of electrospray, ions are
undergoing a rapid jet expansion going from the liquid phase to the gas phase.
This means the dispersion of the ions is much greater. To effectively inject ions
as close as possible to the center axis of the octopole, the voltage must be
positive in order to focus the ion beam into the octopole. Since the El ionization
process occurs at sample pressures well below 1x1 O'6 torr, the ion beam is more
effectively extracted and focused from the ion source with a negative voltage as
opposed to a positive voltage where ions would strike the wall of the ion gate
tube lens.

262
Table 4-2. Instrument parameters for El, octopole, and ion trap operation for the
PFTBA spectrum shown in figure 4-20.
Instrument Parameter
Value
Filament Current
35 fjA
Electron Energy
70 eV
Extractor Lens (L1)
-8 V
Lens (L2)
-105 V
Quadrupole Entrance Lens (Ion Gate)
-55 V (gate on) +180 V (gate off)
Mass Set
0.3 V detected RF
Octopole Offset
-3.0 V
Ion Trap Offset (ring and endcaps)
-5.0 V
Exit Tube Lens
-120 V
Dynode
-15 kV
Electron Multiplier
-1250 V

263
Ion Trap High Mass Theorv/Qperation
Since the majority of the compounds under study in chapters 4 and 5
exceed the mass range of the standard Finnigan MAT ITMS electronics
(m/zmax=650), the mass range of the instrument was extended to approximately
m/z 2500. Any of the following techniques can be used to extend the mass range
of the instrument: (1) reducing the radius of the ring electrode; (2) reducing the
drive frequency of the main rf; and (3) applying a resonant ejection frequency to
the endcap electrodes to reduce q^ct-73-259,260 Increasing the mass range of the
ESI/ion trap by a reduction in the radius (r0) of the ring electrode requires a
redesign of both the ring electrode and the multipass optical system, which could
be very complicated. In addition, reduced size ring electrodes are more
susceptible to space charging effects at lower ion concentrations due to lower
trapping volume. The second method, involving reduction of the rf drive
frequency can significantly degrade mass resolution for the large drop in rf
frequency (e.g., below 0.6 MHz) needed to reach a mass range of at least m/z
2500. These first two methods require physical modifications to the instrument
to obtain high mass data. The third mass range extension technique, called
resonance ejection, requires only a modest change in the software to facilitate
mass range extension.73,259'260 The mass range extension method employed in this
dissertation is that of resonant ejection, where reducing the resonant ejection
frequency reduces qeject, thereby increasing the mass range. In equation 4-1 are
shown the relationships that govern the mass range extension process in the

264
quadrupole ion trap:
(4-1)
where e is the charge on the ion, Vmax is the zero-to-peak voltage of the drive rf,
r0 is the radius of the ion trap ring electrode, and q#jact is the Mathieu stability
parameter directly related to the resonant ejection ion frequency applied to the
endcap electrodes. The desired m/z^ defines the mass range extension by
determining the value of qeject in equation 4-1 and, therefore, the corresponding
6Z value where the appropriate resonant ejection frequency is generated. Various
methods exist for calculating 6Z, including recurrence relations/continued fractions
which are the most accurate.55 However, for qeject values less than 0.4, the
approximate relationship:
(4-2)
can be used. Once the value of 6eject is calculated, the applied resonant ejection
frequency is found by:
(4-3)
where 0<6U<1, Cl denotes the rf drive frequency, and n=0, 1, 2, 3... For
calculating w (the qeject frequency), n=0 and equation 4-3 then reduces to:
The application of the resonant ejection frequency (calculated from
equation 4-4) to the endcap electrodes can be understood by examining figures

4-21261 and 4-22261. For the case where no resonant ejection frequency is applied,
the instrument operates in the traditional mass selective instability scan mode.
When this occurs, the larger mass ions (e.g., > m/z 650) are never "scanned out"
of the ion trap because the applied rf voltage is not high enough to bring these
ions to the z-stability edge of the Mathieu diagram (see figure 4-21)261. Instead,
they essentially drift out of the analyzer region when the applied rf voltage goes
to zero at the end of the analytical scan. To efficiently detect the larger mass
ions, the z-stability edge of the Mathieu diagram must be moved in" so that ions
of higher m/z are detected. Moving the z-stability edge is accomplished by
applying an auxiliary ac frequency (i.e., the resonant ejection frequency) on the
endcap electrodes, which is equivalent to an ions given frequency of motion in
the z-direction. As the rf voltage is ramped during the analytical scan, the ions
come into resonance with the applied frequency and are ejected from the ion trap
and detected by the electron multiplier. This phenomenon of mass range
extension is shown in figure 4-22 where the points A, B, and C represent the
application of the resonant ejection frequency for qeject = 0.906, 0.454, and
0.0908, thereby extending the mass range to 670,1300, and 6500, respectively.261

Mathieu stability diagram for the quadrupole ion trap. Operated in the mass selective instability
mode ions of successively higher m/z ratio are moved along the a=o line and "pushed over" the
z-stability edge. Higher mass ions (indicated by the larger black circles) > m/z 650 are not
scanned out of the trap during the acquisition period due to the limited rf voltage output (from
reference 261).
Figure 4-21:

0.2-
-0.2-
-0.4-
-0.6-
-0.8
z stable
r stable
C| eject 0.908
I I
q oc RF Trapping
z Voltage
Amplitude
a oc DC Voltage
z Amplitude
0.0 0.5 1.0 1.5
q
z
267

Figure 4-22:
Mathieu stability diagram indicating the resonance ejection points for various degrees of mass
range extension (from reference 261).

269

270
The amplitude of this resonant ejection frequency must be great enough
to cause ejection of the high mass ions within a few rf cycles after the ions come
into resonance. Typical values are in the 7.5 V^p range, which means that ions
gain sufficient kinetic energy quickly enough to be ejected from the trap before
they are dissociated. Considerations of indefinite mass range extension, mass
resolution issues, and electrometer modifications are negligible for the operational
mass range of the experiments described in chapters 4 and 5. A detailed
discussion of the principles and operational parameters can be found in the
literature.73259"261
Basic ESI Operation
For ESI operation, the El source and corresponding flange extension were
removed and the modified ESI source and Conflat adapter flange were installed.
The system was then cabled as described in the system interconnect section (see
figure 4-19). The compounds used to evaluate system performance of the ESI/ion
trap were horse muscle apomyoglobin, bovine heart cytochrome c, and bovine
insulin (Sigma Chemical Company, St. Louis, MO). These peptides/proteins were
chosen because they are frequently used as standards for evaluating ESI/mass
spectrometer performance. Protein/peptide samples were prepared in 50:50
methanol:water solutions with a 0.1% acetic acid content. Unless specified
otherwise, the sample of interest was infused directly into the instrument at a flow
rate of 3 //L/min using a syringe pump (described previously). The ion trap was

271
set-up for high mass operation as described in the previous section. In figures
4-23 and 4-24 are seen the ESI/ion trap mass spectra of (500 fmol///L) horse
muscle apomyoglobin and (350 fmol///L) bovine heart cytochrome c, respectively.
A summary of the operating parameters for the electrospray source can be found
in table 4-3.
The spectrum of myoglobin (figure 4-23) shows a charge state distribution
ranging from +11 (m/z 1542) to +27 (m/z 628). Beginning at charge state +16
(m/z 1060), a small shoulder begins to appear on the right-hand side of the peaks
ranging up to charge state +11 (m/z 1542). This can be attributed to either
resonance ejection voltage optimization effects or ESI adduct ion formation (or a
combination of both). In terms of resonance ejection voltage, larger amplitudes
are needed to efficiently eject higher m/z ions from the ion trap. Many times in
high mass operation, the chosen resonance ejection voltage is a compromise
over the mass range of interest (e.g., not optimum for lower or higher mass ions,
typically observed for m/z ranges over 500). A possible solution to this problem
might involve the ramping of the resonance ejection voltage over the mass range
scanned.
For the other scenario (where adduct ions with the solvent could be
forming), this could be explained by the low capillary-skimmer bias of +15 V.
Low bias values typically result in minimization of collisional desolvation in the
atmosphere-vacuum interface. Characteristically, collisional desolvation is
greatest for higher charge states, thus giving lower residual adduction (as
observed in figure 4-23). For capillary-skimmer biases greater than +25 V,

272
Table 4-3. Electrospray instrument parameters
Electrospray Parameter
Value (units variable)
Sample Flow Rate
3 //L/min
Gas Flow Rate
30 psi (NJ
Needle Voltage
+4500 V
Capillary Offset
+ 15 V
Capillary Temperature
180 C

Figure 4-23:
ESI spectrum of horse muscle apomyoglobin showing charge states ranging from +27 to +11.
The sample concentration was 700 fmol///L at a flow rate of 3 /^L/min.

10000
8000
6000
4000
2000
0
+19
+20
+21
+22
+23
+24
+26 i
+18
+17
mk
+16
+15
+12
L +11
fnrrp-f
VJ\J
I II | I I I | I I I | I T I | II 1 | 1 I I | I 1 I
0 600 700 800 900 1000 1100 1200 1300 1400 1500
m/z
1600
274

275
extensive fragmentation of the higher-charged states can occur. Therefore the
selection of the appropriate capillary-skimmer bias is usually a compromise
between the desolvation and adduct formation effects.84,106
Other parameters that can affect adduct formation include the capillary
temperature and the dry gas flow. For the apomyoglobin sample in figure 4-23,
the capillary temperature was 180 C, which provided the most efficient
desolvation (solvent molecules) while not producing any unwanted thermal
degradation effects. Since the gas flow in the electrospray source was not
heated, some condensation effects caused by gas expansion in the vacuum
region could contribute significantly to adduct formation. This can be typically
offset by increased capillary temperatures, increased capillary-skimmer bias, or
excitation (e.g., CID of adducted ions) in the ion trap analyzer to produce
desolvated [M+nH]+ ions (where n=1,2,3...).
During the electrospray process, changes in the protein environment (e.g
pH effects, addition of organic solvent, addition of modifiers such as urea, and
temperature effects) can dramatically affect the observed charge state distribution.
As shown in figure 4-24 for bovine heart cytochrome c, a bimodal charge state
distribution is observed for an acetic acid content of 0.1% (pH=3.27). At lower
pH (and/or high organic solvent content), cytochrome c denatures from a tightly
folded complex to a random coil state.262,263 This protein "unfolding" essentially
makes basic amino acid residues normally unavailable for protonation (e.g.,
effectively buried in the internal globular structure of the protein) readily

Figure 4-24: ESI spectrum of bovine heart cytochrome c showing a bimodal charge state distribution indicative
of a 50:50 methanol:water and 0.1% acetic acid solution. Sample concentration was 350 fmol//yl_
at a flow rate of 3 /yL/min.

5000
4000
3000
2000
1000
0
+9
+15
+16
+13
+19
+20
i- T-r 1^1 V[
+10
+12 +11
LL
ftrri'rrrfr
+8
rpm"|"i
Mii
P|
600 700 800
900
1000 1100 1200 1300 1400 1500 1600
m/z
277

278
accessible for protonation. This effect explains the presence of the higher charge
state ions (+20 to +12) of the protein in a random coil confirmation (see figure
4-24).262-264 A summary of the instrumental parameters used to obtain the spectra
in figures 4-23 and 4-24 is seen in table 4-4.
Perhaps the most important observation in these initial studies is the low
concentration of sample used to tune the instrument (e.g., 500 fmol///L for
apomyoglobin and 350 fmol///L for horse heart cytochrome c). These values are
typically one order of magnitude lower than those used by standard
ESI/quadrupole systems. This sensitivity advantage can be attributed directly to
the rf-only octopole ion injection system of the ESI/ion trap instrument.
The initial test of absolute sensitivity was conducted with a 5 fmol/j/L
sample of bovine insulin. In figure 4-25 is seen the ESI/ion trap spectra of bovine
insulin at 5 pmol///L and 5 fmol///L, respectively. The difference in the two spectra
was the ion injection time for each sample. The instrumental conditions for
acquisition of multiply charge bovine insulin were very similar to those of
apomyoglobin and cytochrome c seen in table 4-4. The detection limit in the full
scan mode was taken at the point where all three charge state ions of bovine
insulin were still visible above the baseline with a signal-to-noise ratio of
approximately 3:1. The amount of sample consumed during the ion injection
period for 5 fmol///L of bovine insulin can be calculated by the following
equation:35

Table 4-4. Instrumental parameters for ESI spectra of apomyoglobln and cytochrome c.
Instrument Parameter
Apomyoglobin Value (units
variable)
Cytochrome C
(units variable)
Octopole RF
0.3 V detected RF
0.3 V detected RF
Octopole Offset
-3.0 V
-3.0 V
Ion Trap Offset
-5.0 V
-5.0 V
He Pressure
1.0x104 torr (uncorr.)
1.0x104 torr (uncorr.)
Ion Gate
+ 75 V
+68 V
RF Level (qin|ect)
35 (1 ms); 55 (1 ms)
35 (1 ms); 60 (1 ms)
Scan Range (qeject)
to m/z 1750 (134 kHz)
to m/z 1875 (125 kHz)
Resonance Ejection
Amplitude
7.5 V0 P
8.2 V0P
Ion Injection Time
2 ms
2 ms
Dynode
-10 kV
-10 kV
Electron Multiplier
-1300 V
-1300 V
279

Figure 4-25:
Bovine insulin spectrum at 5 pmol///L and 5 fmol///L respectively.

Intensity (Arbitrary Units)
281
16000
12000
8000
4000
0
+4
+5
5 pmol/fiL
5 ms ion injection
+3
Wijriiyrttfi ivpiritpMj |rnq
m/z

282
sample used = (sample cone.) (flow rate) (ion injection time) f4-5)
where sample concentration, flow rate, and ion injection time are in units of
(n,p,f)mol///L, //LVmin, and ms, respectively. For the case of the 5 fmol//L sample
in figure 4-25, the total sample consumed was 50 atomoles consumed during the
200 ms ion injection period.
Octopole RF Level
To test the affect of the applied rf voltage, octopole offset and ion gate lens
on ion transmission, the intensity of the ion signal for multiply-charged yeast
ubiquitin was plotted as a function of rf voltage, octopole offset voltage, and ion
gate voltage, respectively. Yeast ubiquitin was obtained from ICN
Pharmaceuticals Inc. (Costa Mesa, CA) and was prepared in a 50:50
methanol:water solution containing 0.1 % acetic acid. Addition of methanol to the
ubiquitin standard causes unfolding of the protein from its native form.265"270 This
lowering of the dielectric constant (e.g., addition of alcohol) increases the
percentage of the protein in the cr-helical confirmation.267,269'270 The observed
charge states of ubiquitin in the methanolic solution range from +6 (aqueous) to
the +10 to +12 charge state (due to the binding of methanol to the protein
surface thereby unfolding the protein). The high ion intensity of these charge
states make ubiquitin an attractive choice to study the efficiency of ion
transmission in the ESI/ion trap instrument.
A plot of ion intensity for the + 6 (m/z 1417), +8(m/z1063), +10(m/z851),

283
and +12 (m/z 709) charge states of ubiquitin as a function of the rf voltage
applied to the octopole is shown in figure 4-26. The even charge states were
chosen to simplify the visual appearance of the various plots. In each case the
ion intensity increased sharply at 0.2 V detected rf, where all the curves began to
level off. At the maximum rf voltage (2000 V^,), there is no significant decrease
in the ion intensity of any of the charge states. This behavior can be explained
by the triangular-like extension of the lowest stability region for an octopolar field.
The value of the Mathieu parameter q4 for an octopolar field can be calculated
from the following relationship:237
16eV0_p
_ 2*2
m to r0
(4-6)
Due to coupled motion in the x-y plane of the octopole, the shape (e.g.,
boundaries) of the stability diagram depends on the ions initial entry conditions,
as described in chapter 3. Although a single stability diagram cannot be
constructed for the rf-only octopole, the stability limit for q4 (directly proportional
to rf voltage) is > 50 for ions traversing the octopole. For the +6 (m/z 1417), +8
(m/z 1063), +10 (m/z 851), and +12 (m/z 709) charge states of ubiquitin, the
values of q4 are seen in table 4-5. For the condition where 7^=2000, none of the
q4 values exceed the approximate stability limit of the octopole stability diagram.
The flat response of the ion intensity curves in figure 4-26 demonstrates the
limited amount of tuning needed for operation of the rf-only octopole ion guide.

Figure 4-26:
Effect of rf voltage on ion transmission efficiency. Past a detected rf value of 0.2 V, the signals for
all the charge states of ubiquitin level off and are independent of the rf voltage. The extended
stability region of the octopole helps make the flat response curves possible.

14000
12000
10000
8000
6000
4000
2000
0
RF Voltage (V)
500 1000 1500 2000
J I I I I I I I I I I I I I I I I I I I
0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
Detected RF (absolute value, V)
285

286
Table 4-5. Calculated q4 values for the +6 (m/z 1417), +8 (m/z 1063), +10 (m/z
851), and +12 (m/z 709) charge states of ubiquitin at 2000 V^.
m/z
Charge State
q4
Detected rf
1417
+6
2.320
2.0 V
1063
+ 8
3.094
2.0 V
851
+ 10
3.867
2.0 V
709
+ 12
4.641
2.0 V

287
In almost all applications, the rf voltage applied to the octopole can remain
constant while not significantly affecting the ion transmission properties of a wide
range of ions. When compared to an rf-only quadrupole system (with the same
system parameters as for the octopole and r0 adjusted to produce a quadrupolar
field), all the calculated q2 values exceed the stability limit of 0.908 (see table 4-6).
The value of the Mathieu parameter q2 for a quadrupolar field can be calculated
from the following relationship:13,241
2eV0_P
_ 2,2
(Du r0
(4-7)
Therefore, in order to observe a spectrum over a wide mass range, the rf-only
quadrupole might need an rf voltage ramp (during the ion injection period) in
order to observe all the ions of interest in a given full scan spectrum.
Octopole Offset
The octopole offset can be defined as the potential applied to the octopole
rods in order to extract ions from the skimmer cone region of the ion source. For
the positive ion mode the value of the offset voltage is negative, ranging from a
few hundred millivolts (negative) to approximately -15 volts. The dc offset voltage
is applied equally to all eight pole pieces of the octopole via the rf voltage circuit.
The octopole offset is typically set one to two volts higher than the corresponding
ion trap offset voltage (positive ion mode). This allows for sufficient deceleration
of the ions as they are injected into the mass analyzer region. In figure 4-27 is

288
Table 4-6. Calculated q2 values for the +6 (m/z 1417), +8 (m/z 1063), +10 (m/z
851), and +12 (m/z 709) charge states of ubiquitin at 2000 V^,.
m/z
Charge State
q2
1417
+6
3.309
1063
+8
4.412
851
+ 10
5.515
709
+ 12
6.618

Figure 4-27: Effect of octopole offset voltage on ion transmission efficiency for the peptide ubiquitin. As
expected, higher voltages are need to extract higher mass ions from the electrospray source
region.

14000
12000
10000
8000
6000
4000
2000
0
Ion Trap Offset -5V
Octopole Offset Voltage (V)
290

291
shown a plot of ion intensity versus octopole offset for the even charge states of
the protein ubiquitin. The electrospray conditions for ubiquitin were the same as
in the previous section (see table 4-4). The trend observed in figure 4-27 gave
an optimum ion transmission for the smaller m/z 709 (+12 charge state) ranging
from -0.2 V to -2.0 V (e.g., flat portion of the curve). For m/z 1417 (+6 charge
state), the optimum voltage for ion transmission was -2.8 V. As expected, more
negative voltages were need to maximize the injection efficiency of the larger m/z
ions. The relatively small voltage optimization range (2.5 V) for the multiply-
charged ions of ubiquitin could be attributed to the flat portion of the radial
potential well of the rf-only octopole (see chapter 3). This flat portion of the well
allows a large number of different m/z ions (all at the same kinetic energy) to be
transmitted efficiently through the device.
Ion Gate Lens
The function of the ion gate lens, which is located between the ESI heated
capillary and skimmer cone (see figure 4-3), is to pulse ions through the skimmer
cone region and into the rf-only octopole for transmission to the ion trap analyzer.
In figure 4-28 is shown a plot of ion intensity versus the electrospray gate lens
voltage for the even positive ion charge states of the protein ubiquitin. The
electrospray conditions for ubiquitin were the same as in the Basic ESI Operation
section of this chapter (see table 4-4). Typical voltages applied during the ion

Figure 4-28:
Effect of the electrospray gate lens on ion transmission. The observed maximum transmission
efficiency for the even charge states of ubiquitin varies only by 20 V across a mass range of 700.

6000
5000
4000
3000
2000
1000
0
r

+12 (m/z 709)

+10 (m/z 851)

+8 (m/z 1063)
A
+6 (m/z 1417)
/
I I I I I I MI [I I I I | I 1TI |TT1T| TI I I | I I II | II I I | I I I I | I I II | III I | II I I |
20 30 40 50 60 70 80 90 100 110 120 130
Electrospray Gate Lens Voltage (V)
293

294
injection period to gate ions through the skimmer cone region range from +50
to +100 V. Positive voltages are used to gate ions because of the rapid
expansion from the liquid phase to the gas phase (e.g., large ion dispersion
angles and multiple collision events) and the relatively low energy/velocity of the
ions as they exit the heated capillary. To control the pulse width of positive ions,
a -180 V potential is applied to the ion gate lens to prevent ion transmission
through the skimmer cone region. Ion injection time periods varied from a few
hundred microseconds to hundreds of milliseconds (for low concentration
samples).
The optimum voltage range for transmission of the positive ion even charge
states of ubiquitin was +20 V, with a +60 V and +80 V potential for optimum ion
signal of m/z 709 (+12 charge state) and m/z 1417 (+6 charge state),
respectively. The relatively small ion gate voltage optimization range for the
multiply-charged ions of ubiquitin could be attributed to the efficient focusing of
the ion beam through the skimmer cone region, and therefore the injection of ions
close to the center axis of the octopole.
RF Level/lon Injection
The one ESI/ion trap parameter that does have a profound effect on the
observed mass spectrum is the ion trap rf level during the ion injection period
(frequently termed qinject). Early studies of the ion injection process into the
quadrupole ion trap recognized that a relationship existed among the ion trap

295
offset, the m/z of the ion(s) of interest, and the minimum rf level of the ion
trap.192,271'272
To fully appreciate the effects of rf level, m/z ratio and ion trap offset on ion
injection efficiency, a minimum rf level study was carried out on the peptide MRFA
(methionine-arginine-phenylalanine-alanine), which under electrospray conditions
gives both doubly and singly charged species at m/z 262 and 524, respectively.
The MRFA standard (Sigma Chemical Company, St. Louis, MO) was dissolved in
a 50:50 methanol:water solution with 0.1% acetic acid. Electrospray conditions
and flow rate were the same as in the Basic ESI Operation section of this chapter
(see table 4-4). The ion trap offset was set to -5 V with an octopole offset voltage
of -2.2 V.
Figure 4-29 shows two plots: one with the ion intensity of the peptide MRFA
as a function of qinject, and the other the intensity as a function of the minimum rf
level or exclusion limit. The ion injection data indicate that with increasing m/z,
the minimum rf level needed to trap the ions increases. Correspondingly, the q
value needed for efficient trapping decreases slightly. This is in good agreement
with the original studies of Louris et al.192 The behavior of the ions in this manner
can be supported by the original work of Major and Dehmelt.11 Here, if the
average energy of the ions is less than some arbitrary value W (a<<1 and
q<<1), then the ions move within the confines of the trap as determined by
equation 4-8:

Figure 4-29:
Ion injection efficiency as a function of the minimum rf level, expressed in terms of q,n]9Ct and
minimum rf level (given as the low-mass cutoff). The dashed lines indicate points at which spectra
were taken to show the effect of ion injection rf, on spectral quality (see figure 4-30).

RF Level 25
Met-Arg-Phe-Ala
2
a
£
2
u
<
1600
1400
1200
1000
800
5 600
400
200
+2 Charge State
m/z 262
+1 Charge State
m/z 524
I I | I I I | I I I | I II | I I rj I II |
0.00 0.04 0.08 0.12 0.16 0.20 0.24 0.28
^inject
RF Level 38
TT I I | I I II | I I I I | I I I I |
40 50 60 70
RF Level (m/z)
20
297

298
a2 c2 8W
Or + 6, =
r 2 mQ2
(4-8)
where r and z represent the radial and axial displacements, m is the mass of the
ion, Q is the angular frequency of the rf drive, and q, a, and 6 are trapping
parameters. The maximum kinetic energy any ion can have is defined by the
physical constraints of the device where r0 and are the maximum amplitudes
of the radial and axial displacements, respectively. Assuming that ion motion
during the injection process is limited to the z-direction, with 6z=(qz/21/2), then qz
can be defined as:
q2
4 W
z0Q \ m
(4-9)
As the ions undergo collisions, energy is redistributed in the r-direction; when
sufficient energy is removed, equation 4-9 is satisfied. Thus, for low qz values, the
minimum rf level is inversely proportional to the square root of the mass.11,192 In
table 4-7 the minimum rf level, and threshold qz value (both obtained
experimentally from figure 4-29), and the theoretically calculated qz value
(obtained by using an arbitrary constant) are shown. As seen from table 4-7, the
experimental qz values agreed well with the theoretically calculated values.
To obtain a representative ESI spectrum (based on the aforementioned
arguments) of the compound of interest, it may be necessary to increment the
minimum rf level during the ion injection period. To fully understand this concept,
ESI spectra were taken at rf cut-off levels of m/z 25 and m/z 38 respectively (see

299
Table 4-7. Experimental and theoretical threshold values of qz for the peptide
MRFA.
m/z
RF Level (experimental)
qz (experimental)0
qz (theory)b
262
20
0.070
0.070
524
n "I
29
0.049
0.045
Experimental value determined for the half-height of the appearance of trapped
ions.
Theoretical values calculated with an arbitrary constant of 1.3; qz=(1.3/m)1/2.

300
the dashed lines in figure 4-29). The representative spectra are shown in figure
4-30. For an rf level of m/z 25, the doubly-charged peak at m/z 262 predominates
with only a small contribution from the singly-charged species at m/z 524. At a
higher rf level of m/z 38, the singly-charged [M+H]+ ion is approximately 33% of
the m/z 262 (+2 charge state) base peak. It is clear from this simple low mass
range example that a series of ion injection steps at various rf levels would be
appropriate in order to obtain a "truer" representative ESI mass spectrum. For the
horse muscle apomyoglobin and bovine heart cytochrome c samples, discussed
in the Basic ESI Operation section, the rf level was incremented during the ion
injection period (see table 4-4) to obtain the spectra seen in figures 4-23 and 4-
24.
Another interesting aspect of figure 4-30 is the presence of the m/z 393
peak (or the y3 ion), indicating the loss of methionine from either the singly-
charged ion at m/z 524 or the doubly charged ion at m/z 263. This peak
essentially arises from the ion trap offset value of -5 V (translational energy),
although preliminary evidence indicates that injection at somewhat higher q2
values can contribute to fragmentation. It should also be noted that some ions
may be eliminated entirely from the spectrum if CID occurs before or during the
ion injection into the analyzer. Therefore, to determine the importance of the CID
process during ion injection becomes increasingly difficult, especially for multiply-
charged systems where coulombic repulsion forces are strong and the
corresponding energy deposition for CID is small.

Figure 4-30: Spectra of the peptide MRFA taken at an rf level (low-mass cut
off) of m/z 25 and m/z 38. This experiment demonstrates the
need for either scanning or incrementing the rf level during ion
injection in order to obtain a representative ESI spectrum.

Intensity (Arbitrary Units)
302
m/z

303
Another observation of these experiments was the increase in qinject needed
for higher ion translational energies during the ion injection process. For
translational energies in the 10 V range, qjnject optimum was 0.2 to 0.3 for both the
singly and doubly-charged ions of MRFA.
Ion Isolation
A variety of techniques has been used successfully to perform ion isolation
studies on the quadrupole ion trap. Some of these include apex14,76,77, two-step78'
80, SWIFT (stored waveform inverse Fourier transform)38,39, random noise82,
forward-reverse scans74,273, and FNF (filtered noise field)81 techniques. The apex
and two-step isolation routines are effective, but are limited in mass range to
approximately m/z 600 or lower. This limitation is due to the high dc voltages
required to be applied to the ring electrode (physical constraints of the ion trap),
the cost of high voltage equipment, and stability diagram considerations for
higher masses. In this section, a brief discussion of the forward-reverse scan
method is presented to demonstrate the high mass isolation techniques needed
during ESI/ion trap operation. Other more complex techniques such as FNF,
random noise, and SWIFT are currently under investigation in our laboratory.
The original forward-reverse scan technique, as described by Kaiser et
al.273, begins by placing a resonance ejection voltage at the frequency of interest
on the endcap electrodes. Next, the ion of interest is placed just below the
resonance ejection frequency via an rf amplitude ramp (see figure 4-31). As the

Figure 4-31: Sequential steps of the forward-reverse scan isolation process.
To eject unwanted low mass ions the mass of interest is ramped
forward towards the resonance point. To eject unwanted ions of
higher m/z, the resonance frequency is turned off and the ions
are moved past the resonance point by using an rf ramp. Next,
the resonance frequency is turned back ion and the ions are
scanned in reverse towards the resonance point to eject the high
mass ions (from reference 74).

305
Mass of
Resonance
Interest Point
QOO#

OOO*
000

Stability
Limit
OQO*
OOO*
0
q
0.908

306
ion of interest is ramped to its position just below the resonant frequency, ions of
lower m/z are resonantly ejected from the trap. To resonantly eject ions of higher
m/z, the resonant excitation voltage is turned off and the rf amplitude is ramped
so that the ion of interest is located between the resonant ejection frequency and
the z-stability edge of the Mathieu diagram (see figure 4-31). The resonance
ejection voltage is turned on again, and the rf amplitude is ramped in the reverse
or downward direction so that ions of higher m/z than the ion of interest are
resonantly ejected. In this case, the ion of interest is located just above the
applied resonant ejection frequency and is the only remaining ion (m/z) in the
trap.
This exact procedure with isolation of the +18 charge state (m/z 943) of
horse muscle apomyoglobin is shown in figure 4-32. The chosen resonant
ejection frequency for this experiment was 60 kHz at 6.3 V^P. In the top portion
of figure 4-32 is shown the elimination of the lower mass fragments (higher
charge states) of apomyoglobin. The middle section demonstrates the reverse
scan process where higher m/z ions (lower charge states) are eliminated. The
last portion of this figure shows the complete isolation of the +18 charge state of
apomyoglobin achieved when the forward and reverse scans are combined. This
technique is very efficient with isolation efficiencies much greater than those
obtained with traditional rf/dc isolation routines. Also, this technique can be used
quite effectively for high resolution (e.g., slow scanning) isolation experiments.

Figure 4-32: Results of the forward, reverse, and combined forward-reverse
scans for the isolation of the +18 charge state of horse muscle
apomyoglobin. The unknown fragment peak in the spectrum of
the mass-isolated +18 charge state is due to CID induced by the
increase in kinetic energy of the ion as it approaches the
resonance point.

Intensity
308
4000 i
3000
2000
+18
Mass Isolated Ion
1000
Unknown Fragment
'I TpTTTJTT I | I rV'j fTTTTflTTTT
600 700 800 900 1000 1100 1200 1300 1400 1500
m/z
0

309
One limitation of this technique (especially for ESI mass spectrometry) is
the possibility of induced fragmentation of the isolated product (CID). This can
occur when the ion of interest gains enough kinetic energy from the isolation
process (e.g., approaches the resonance ejection frequency too closely) that
fragmentation (CID) is induced. As seen in figure 4-32, an unknown fragment ion
occurs at m/z 1472, which is a result of this process. The amount of
fragmentation can be controlled somewhat by reducing the amplitude of the
resonant ejection voltage. However, if the voltage is reduced too far, inefficient
ion ejection at the resonance point could occur, resulting in poor isolation
efficiency. Particular attention must be paid to electrospray ions where large
numbers of charges on an ion can induce coulombic repulsion forces, reducing
the collision energy needed for the CID process.
Collision-induced Dissociation (CID)
Collision-induced dissociation (CID) has seen a great deal of development
for biological applications in mass spectrometry.94'97,108'109,114-274'278 Applications
have ranged from determination of side chain fragmentations of leucine and
isoleucine for isomer differentiation, to determination of glycopeptide structures
using LC/MS/MS in triple quadrupole instruments.27^277 Ion trap mass
spectrometry (employing CID) has also seen rapid development in the analysis
of biological materials, particularly peptides and proteins.273,274
To evaluate the MS" capabilities of the ESI/ion trap instrument, a series of

310
tandem mass spectrometry experiments were conducted on the peptide
angiotensin I. Human angiotensin I is a decapeptide (DRVYIHPFHL), which is a
pressor substance converted to its active form (angiotensin II) by cleavage of the
HL residues from the carboxy terminus. The human angiotensin I was obtained
from ICN Pharmaceuticals (Costa Mesa, CA), with 700 fmol/pL standards
dissolved in a 50:50 methanol:water solution with 0.4% acetic acid. The
electrospray and instrumental parameters were as described in the Basic ESI
Operation section of this chapter. The ESI spectrum of angiotensin I produced
an intense triply-charged ion at m/z 433, which was mass-isolated using forward-
reverse scans as described in the Ion Isolation section. The MS/MS parameters
for the CID of the triply-charged m/z 433 ion are shown in table 4-8. Figure 4-33
shows the MS/MS spectrum of the triply-charged m/z 433 ion of angiotensin I.
The two major types of fragment ions observed were b ions (charge retention on
the carboxy terminus) and y ions (charge retention on the amino terminus). This
naming convention follows that originally developed by Reopstorff and
Fohlman.279 The primary mechanism of formation for these ions is shown in figure
4-34. In general, formation of either b or y type ions involves cleavage of the
amide bonds along the peptide backbone (see figure 4-34). The proton attached
to the amide nitrogen is typically free to move along the peptide chain. Formation
of b type ions occurs by 6-cleavage and charge migration to the carbonyl end of
the amino terminus fragment. The b ion in figure 4-34 is termed b2 because it
contains two amino acid residues from the original peptide structure (counting

Figure 4-33: The MS/MS spectrum of the +3 charge state of the peptide angiotensin I. Observed fragmentation
includes both doubly and singly charged b and y ions, singly charged a ions, and some sequence
specific internal fragment ions (due to the presence of proline).

3000
2500
2000
1500
1000
500
0
Asp-Arg-Val-Tyr-IIe-His-Pro-Phe-His-Leu
PFHorHPF
a.
t t f 'ftMyy4HLrf4l
10 300 400 500 600 700 800 900 1000
Ill/z
1100
312

313
Table 4-8. Instrumental parameters for the MS/MS of the triply-charged m/z 433
ion of human angiotensin I.
Parameter
value
Qz
0.3
MS/MS Frequency
118 kHz
Resonance Excitation Amplitude
1.5 Vp.p
Resonance Excitation Time
20 ms
He Buffer Gas Pressure
1.0x1 O'4 torr (uncorr.)

Figure 4-34: Simple fragmentation scheme for the formation of b and y Ions. The starting peptide contains 5
amino acids with 6-cleavage and charge migration responsible for b ion formation (charge retained
on the amino terminus), and hydrogen migration from the a-carbon responsible for y ion formation
(charge retained on the carboxy terminus).

Peptide Fragmentation Scheme
P
-Cleavage and Charge Migration
Hydrogen Migration
(a- Carbon)
Ion Type b2
Ion Type y3

316
from the amino terminus). The complementary reaction (which involves charge
retention on the amide nitrogen) for the formation of y type ions is characterized
by a hydrogen ion migration, typically from the a-carbon. The y ion in figure 4-34
is termed y3 because it contains three amino acid residues from the original
peptide structure (counting from the carboxy terminus).280'282
The interpretation of the MS/MS spectrum in figure 4-33 can be explained
by the charge localizations associated with the triply-charged parent ion at m/z
433. The fragmentation indicates that protons are located on Arg2 and His9, with
the third proton free to move along the peptide backbone between these two
residues. This leads to the large amount of (acylium ion type) mid-range
fragmentation observed (b4, b5, b6, and b8). The corresponding series of ions (a4,
ag, ag, and aj arise from the loss of neutral carbon monoxide from the b series
of ions. Some doubly-charged ions are also observed (b6+2 and b9+2), associated
with basic residues His6 and His9 respectively. A complementary series of y ions
also appear (y2, y3, y4) in the spectrum with an unusually high intensity observed
for the y3 ion (proline effect). Attempts of MS3 experiments on the y3 ion to verify
its structure produced no observable fragmentation due to the stability of the ion
in the gas phase. In addition, no a7-b7 pair of ions is observed presumably due
to the proline effect.280'282
The peaks labeled RV, HPF/PFH, and PFHL in figure 4-33 are internal
sequence y type ions, with the latter two possibly driven by the presence of the
pro7 residue. These peaks arise due to the high basicity of proline in the gas-

317
phase (although internal sequence ions from histidine and arginine can also
appear in the spectrum due to their large gas phase basicities). Of all the basic
residues, proline has the highest proton affinity, even exceeding those of arginine
and lysine. The mechanism of formation for these ions follows the same
convention as y type ions with charge retention on the amide nitrogen and
subsequent hydrogen ion migration to the highly basic proline residue (see figure
4-35).280 Internal sequence type ions have been observed with peptides that
contain proline residues in ion trap mass spectrometry. The high ion abundance
of these peaks is probably due to the timescale of the ion trap MS/MS
experiment, which takes place on the millisecond as opposed to the microsecond
timescale found in either magnetic sector or triple quadrupole instruments. Over
the millisecond timeframe, ions in the trap will tend to go towards their lowest
energy state, if given the opportunity.
MS3 experiments were performed on both the b5 and b9+2 ions from the
MS/MS spectrum. Figure 4-36 shows the MS3 spectrum obtained from the b5 ion
of angiotensin I. The b5 ion was mass isolated using forward and reverse scans
and fragmented at a qz of 0.3 (amplitude=2.0 Vp_p). Typically, singly-charged b
ions fragment easily to produce lower mass b ions and some a ions, as seen in
figure 4-36. There is a great deal of interest in generating a large number of high
mass b ions, so as to evaluate MS" techniques for sequencing peptides one
residue at a time. From figure 3-36, the formation of the a^ b4, and a4 verifies the
loss of isoleucine (or leucine which has the same residual mass at 113) from the

Figure 4-35:
Mechanism for formation of internal y type ions from peptides which have proline within the peptide
backbone (e.g. not located on a terminus).

Internal Cleavage at Either Pro or His
Hydrogen Migration (a-Carbon)
Ion Type y4
OH
319

Figure 4-36:
MS3 spectrum of the b5 ion from angiotensin I. The predominant ions in the spectrum are the next
lowest b ion in the series (b4) and the corresponding a ions which arise from the loss of carbon
monoxide from the acylium portion of the b ion.

Intensity
m/z
CO
JO

322
b5 fragment.
MS3 data from the fragmentation of the b9+z ion produced a series of
singly-charged higher m/z b ions, as shown in figure 4-37. It is interesting to note
little or no complementary singly-charged y ions were observed. However, there
was a series of four unknown peaks (labeled with question marks in figure 4-37),
which may be possible cyclization rearrangements driven by the presence of the
pro7 residue and the long reaction times associated with ion trap experiments.
Further MS3 experiments utilizing labeled standards would be necessary to verify
these ion fragmentations.
Attempts using MS3 to determine the possibility of a y8+2 in the y4 peak in
figure 4-33 yielded no data which could be interpreted to either confirm or deny
the presence of this ion. This may result in part (as mentioned earlier) from the
high stability of the y series ions in general.
Negative Ion Mode
The majority of ESI-MS analysis performed to date has been in the positive
ion mode. This is especially the case for peptide and protein samples, which
contain a high number of basic residues (histidine, lysine, arginine) with high
dissociation constants that produce multiple sites for protonation in acidic
solutions. Negative ion electrospray of these compounds in an acidic solution is
not practical because the pKa values of the corresponding acidic amino acids
(aspartic acid and glutamic acid) in a protein structure are 5 or less.283

Figure 4-37: MS3 spectrum of the b9+2 ion from angiotensin I. The predominant ions observed are the higher
m/z ions b6 and b8. The unknown peaks may correspond to cyclization for the peptide backbone
structure; however, without proper labeling experiments this cannot be confirmed.

2000
1500
1000
500
0
Asp-Arg-Val-Tyr-IIe-His-Pro-Phe-His (b*2 fragment)
I I 1 I I I l l I | I I I l| 1 I I I | l l I I | I I I I | I l M |
500 600 700 800 900 1000 1100
m/z
324

325
To evaluate the potential of the instrument for negative ion electrospray
analysis, the peptide bradykinin (APPGFSPPA) was used. Bradykinin is
vasodilator belonging to a group of hypotensive peptides called plasma kinins284
To form negative ions, the sample (20 pmol//^l) was dissolved in a 3 mm NH4OH
solution which removes a proton from the carboxy terminus. The polarity of the
dynode, ion trap offset, octopole offset, ion gate lens, and electrospray needle
voltages were reversed for negative ion operation. Oxygen was used for the
counter-current gas flow as an electron scavenger. In figure 4-38 is shown the
negative ion spectrum of Bradykinin. The only ion present is the [M-H]' from the
loss of a proton on the carboxy terminus. The ion injection time was 3.0 ms with
^injectset at 0-2. The sensitivity in the negative ion mode compared to that of the
positive ion mode was a factor of 10 to 100 less. Several possible reasons for
this include source design, peptide stability in highly basic solutions, and
ionization efficiency.

Figure 4-38: Negative ion ESI spectrum of bradykinin at 20 pmol///L in a 3 mM NH4OH solution. The polarity
of the dynode, octopole offset, ion trap offset, ion gate lens, and electrospray needle were reversed
in order to acquire the spectrum.

12000
10000
8000
6000
4000
2000
0
(M-H)"
m/z 1059
rftVpiVf rrqprVTT pt ft i jrtrrp^r
l
r
1
400 500 600 700 800
m/z
900
1000 1100 1200
327

CHAPTER 5
PHOTODISSOCIATION OF BIOLOGICALLY IMPORTANT MOLECULES:
PROTEINS, CARBOHYDRATES, AND OLIGONUCLEOTIDES
General Overview of Structural Elucidation
The use of tandem mass spectrometry for the structural elucidation of
biological molecules has grown at an exponential pace with the advent of the
electrospray ionization technique. The majority of the work to date has focused
on sequencing oligopeptides and phospholipids (ionized by electrospray and
FAB), employing collisional-activation as the ion activation technique of
choice.2812a5,2B6 The simplest and most economical method to obtain structural
information by electrospray mass spectrometry is that of nozzle-skimmer
dissociation as developed by Loo et al.287 This technique, which involves
dissociation of multiply-charged ions in the high pressure source region, has been
performed on molecules as large as serum albumins.288 However, this technique
is only usable for pure samples and cannot be applied to mixture analysis, since
product ions cannot be matched to their parent ions.
The presence of multiple charges on ionic species can have a significant
affect on the ability to fragment the species via CID to obtain structurally relevant
information. Unusually high fragmentation efficiencies, site-specific cleavages,
and fragmentation driven by non-basic functional groups are just a few of the
328

329
different results that can be seen with CID of multiply-charged species.108 On the
other hand, highly basic peptides such as the neuropeptides of the dynorphin
series exhibit very little, if any, fragmentation (typically immonium ions and small
N-terminal fragments). Even for highly-charged basic fragments (with protons
anchored into position) where coulombic repulsions are strong, fragmentation is
still limited.110
In addition, the broad energy distribution associated with the CID process
can further complicate the spectral interpretation process. Also, for singly-
charged ions of m/z > 2500, kinetic and collisional effects limit the amount of
structural information available.289,290 Therefore, it is desirable to explore
alternative methods for the activation of biological species in the gas-phase.
Photon absorption is one possible alternative that overcomes many of the
aforementioned short-comings of collisional activation. A well-defined energy
deposition process (independent of converting an ions translational energy into
the vibrational energy needed for fragmentation) and the long ion-storage times
associated with trapping instruments make photodissociation an ideal choice for
structural elucidation of biological molecules.
Early studies using photodissociation for structural studies of biological
species in the ICR employed a wide range of wavelengths from infrared (IR) to
ultraviolet (UV).1111291,292 Williams et al. reported photodissociation efficiencies on
the order of 100% for the peptide alamethicin using 193 nm light when the ions
were confined to the beam path.291 Corresponding CID experiments involving

330
pulsed-gas introduction of collision gas, which makes the MS/MS experiment
much more complicated, produced dissociation efficiencies on the order of 15%,
which is significantly lower than for photodissociation291 Comparison between
photodissociation and surface-induced dissociation of porphyrins in the ICR by
Castro et al. yielded higher dissociation efficiencies for the photodissociation
process. Longer irradiance times (e.g., with no parent ion selection after the first
MS/MS step), produced more fragment ions at higher abundances than those
obtained with SID. At very long irradiance times (both UV and UV-vis
wavelengths), the new fragment ions were produced at the expense of
diagnostically significant ions. These new fragment ions produced were of no
value for the structural elucidation process.292
Recently, IRMPD has been employed in the ICR for photodissociation of
biological species in the gas-phase. Little et al. demonstrated the capability of
IRMPD to obtain sequence information for peptides/proteins and
oligonucleotides.111 The IRMPD of carbonic anhydrase produced fragment ions
similar to, but with valuable additions, fragmentation obtained by other methods
(e.g., CID and SID). Optimization of irradiance times varied widely for
peptides/proteins (from 50 to over 300 ms), indicating a greater range of ion
stabilities than originally believed from previous CID data. Irradiance times for
oligonucleotides (negative ion mode) were significantly less (e.g., 10-30 ms) than
those of the peptide/protein experiments. This difference could be attributed to
the photon resonance of the PO stretching frequency. More importantly, IRMPD

331
was shown to have greater selectivity, have less mass discrimination, and could
dissociate much more stable ions than the corresponding CID process.111
This chapter begins with a survey of both pulsed and continuous wave
lasers for photodissociation of ions from peptides, oligosaccharides, and
oligonucleotides. This is followed with a discussion of the relative merits of
photodissociation versus the CID process. In addition, the very first ESI-
photodissociation spectrum is shown; the remaining data in this chapter represent
a combination of the fundamental studies presented in chapter 2, the theoretical
and design principles in chapter 3, and the instrumentation considerations of
chapter 4. The multipass ring electrode discussed in chapter 2 is the heart of
these experiments, using the increased photoabsorption pathlength to counter the
effects of higher pressure operation in the ion trap which can quench
photodissociation at these wavelengths. The data shown in this chapter are the
first ever reported for photodissociation of biological species in the quadruple
ion trap. The high sensitivity associated with the ion trap, combined with an
efficient ion injection system for transport of ions from an electrospray LC/MS
interface, makes the ion trap an obvious choice for the analysis of micro
quantities of material from biological extracts. Combined with photodissociation
for structural elucidation, an instrument of this design has the potential to solve
a variety of problems in the field of biochemistry.
As mentioned earlier, one major advantage of photodissociation is that an
ions kinetic energy does not have to be converted into internal energy to effect

332
fragmentation. In addition, several tuning parameters including helium buffer gas
pressure, ion frequency, and ion population, which further complicate the single
frequency CID experiment even for an experienced user are not concerns in PID
experiments. Even when broadband excitation techniques are employed to
minimize ion frequency and ion population effects, dissociation efficiencies
observed are somewhat lower than in the corresponding single-frequency
experiments, and can be drastically lower than in photodissociation experiments
when a sufficient photoabsorption cross-section exists for the ions of interest.
The effect of ion population on the dissociation efficiency of protonated 12-
crown-4 ether produced by Cl is shown in figure 5-1. Ionization time (shown on
the x-axis) refers directly to stored ion population in the ion trap. For the case of
the single-frequency CID experiment (where the tuning parameters were optimized
for a low number of ions stored in the ion trap), a significant decrease in
photodissociation efficiency was observed as the ionization time and hence the
ion population increased. Typically, an increase in ion population results in a shift
of the fundamental frequency of ion motion to lower frequency, thus resulting in
a decrease in dissociation efficiency for the given CID tune parameters (see figure
5-1 ).293 In the case of broadband excitation, where a whole range of frequencies
are excited over a given time period, the dissociation efficiency is independent of
ion population. However, a decrease is observed in dissociation efficiency
compared to that of the optimum single frequency results.

Figure 5-1: Evaluation of the various MS/MS techniques used with the quadrupole ion trap. The dissociation
efficiency of photodissociation, single frequency CID, and broadband excitation are plotted versus
ionization time. The ionization time is directly proportional to the trapped ion population.

Dissociation Efficiency
Ionization Time in ms
334

335
Photodissociation efficiency seen is independent of ion population.
However, the overall dissociation efficiency is substantially higher than for either
CID technique (approximately 90% compared with 65% for the optimum single
frequency tune and 45% for the broadband excitation shown in figure 5-1). These
results can be attributed to two factors: (1) the increased photoabsorption
pathlength of the multipass ring electrode, which compensates for the reduced
photoabsorption cross-sections of organic species; and (2) the elimination of
collisions for transfer of translational energy to vibrational energy, where ion
stability can cause competition between resonance ejection and CID during an
MS/MS experiment.
The first continuous wave C02 laser photodissociation experiment on
electrosprayed ions was that of [M+K]+ and [M+H]+ ions from 18-crown-6 ether.
The photodissociation set-up has been described in chapter 4 of this dissertation.
Electrospray conditions consisted of a 5 pmol///L solution of 18-crown-6 ether in
a 50:50 methanol:water solution with 3 mM potassium chloride added for adduct
formation. Helium buffer gas pressure was 4.5x1 O'5 torr (uncorrected), which was
the pressure where the onset of photodissociation was first observed. The
spectrum shown in figure 5-2 has both the [M+H]+ and [M+K]+ ions stored in the
trap during the laser irradiance period. The two fragment ions produced arise
from direct dissociation of the adduct ion (m/z 303) to K+ (m/z 39) and the
corresponding neutral 18-crown-6, and the dissociation of the protonated parent
(m/z 265) ion to produce C7H1203+ (m/z 144).

Figure 5-2:
Initial photodissociation spectrum of the potassium adduct and protonated 18-crown-6 ether
obtained on the electrospray/ion trap instrument. The peak at m/z 39 represents the direct
dissociation of the potassium adduct, and the peak at m/z 144 the dissociation of the [M+H]+ ion.

Intensity (Arbitrary Units)
800 i
600
400
200
K+
in/z 39
C7H123
m/z 144
v+
(M+H) +
m/z 265 (M+K)+
m/z 303
1|III|IIIpi1T1
50 100 150 200
m/z
250
300
350
337

338
Peptides and Proteins
This section discusses the photodissociation of the triply-charged ion of
human angiotensin I and the singly-charged ion of the peptide antibiotic
gramicidin D. All experiments in this section were performed with a continuous
wave C02 laser as described previously. The fragmentation nomenclature follows
that of Roepstorff discussed in the CID section of chapter 4.279 Unfortunately, after
preliminary data was obtained for these compounds, a series of malfunctions in
the cw lasers power supply halted all future studies with this laser on peptides,
oligosaccharides, and oligonucleotides. Further experiments were performed with
a pulsed C02 laser.
Human Angiotensin I
One of the first peptides investigated with the new instrument was human
angiotensin I. A continuous wave photodissociation set-up (as described in
chapter 4) was employed for the initial peptide experiments. The electrospray
conditions used were the same as those indicated in chapter 4 for the CID
studies. The m/z 433 ion (triply-charged) of angiotensin I was mass-selected in
the ion trap as described previously. The helium buffer gas pressure was initially
1.0x10^ torr (uncorrected). For this buffer gas pressure, no fragmentation was
observed with the continuous wave C02 laser (for irradiation times up to 400 ms).
This indicated that the large number of collisions of the triply-charged parent ion
with helium buffer gas deactivated (via collisional cooling) the parent ion

339
sufficiently to prevent the multiple photon absorption process from promoting the
ion to the dissociation threshold energy. When the buffer gas pressure was
reduced to 2.0x1 O'5 torr, the appearance of a photodissociation spectrum (85 ms
irradiance time) was observed (see figure 5-3).
The fragmentation observed included the mid range b fragments indicative
of the +3 charge state (b5 and b6), the internal y ion series perhaps generated by
the presence of pro7 (HPF/PFH and PFHL fragments), and the y4-y8+2 peak(s).
These fragment ions compare well with what was generated using single
frequency CID experiments discussed in the previous chapter (figure 4-33). One
major difference observed between photodissociation and CID was the presence
of a b9 ion in the photodissociation spectrum. The singly-charged b9 ion is of
particular importance since it could be mass-isolated and a series of MS" studies
performed, where individual amino acids from the peptide chain could be
sequenced in succession by either further photodissociation or CID experiments
as demonstrated in chapter 4. Therefore, generating high mass b ions is of
particular importance for accomplishing a ladder sequencing experiment (removal
of one consecutive amino acid from the peptide chain for each dissociation step)
in order to verify peptide sequences in mass spectrometry and Edman
degradation.
The negative aspect of this experiment is the need for reduced pressure
of the helium buffer gas to obtain a photodissociation spectrum. Indeed,
improved fragmentation efficiencies could be observed for the IRMPD process if

Figure 5-3:
IRMPD of the triply charged m/z 433 ion of human angiotensin I. The spectrum bears a
resemblance to the corresponding CID spectrum reported in chapter 4. The peaks at PFH and
PFHL represent internal y sequence ions generated by the presence of proline at residue 7.

Intensity
3000 i
2500
2000
1500
1000
500
(M+3H)
+3
HPF or PFH
Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-His-Leii
200
300
400
500
600
700
m/z
800
900
1000 1100 1200
341

342
helium buffer gas pressures during the laser irradiation period were reduced to
below 4.0x1 O'6 torr. However, significant reductions of pressure in the ion trap
analyzer not only adversely affect peak shape and resolution, but more
importantly also reduce ion injection efficiency and thus sensitivity. For the
angiotensin I studies, ion injection times were increased from the 1 to 3 ms range
(when operating with a helium buffer gas pressure of 1.0x10"* torr) to the 68 to 75
ms range with the reduced operating pressures (e.g., 2x1 O'5 torr). The addition
of a pulsed-valve (as discussed in chapter 2 and chapter 6) in order to separate
the ion injection, photodissociation, and detection events in time from the helium
pulse should significantly improve photodissociation efficiency without sacrificing
sensitivity.
Gramicidin D
The second peptide investigated using the IRMPD process was the
antibiotic gramicidin D. The electrospray and instrumental conditions employed
were similar to those reported in the previous chapter. Photodissociation was first
observed for the singly-charged ion of gramicidin D with a helium buffer gas
pressure of 1.7x1 O'5 torr and an irradiance time of 85 ms. Ion injection conditions
were set such that the majority of the signal observed was that of the singly-
charged species, with only a small portion of the doubly-charged ion stored for
instrument calibration purposes. Since no isolation was performed, the [M+H]+,
[M+Na]+, and [M+K]+ (as well as the low intensity [M+2H]+2 ion) were all

343
irradiated. In figure 5-4 is shown the corresponding photodissociation spectrum.
The major fragment ion observed was a rearrangement loss of the modified
carboxy terminus of the peptide. The [M-61]+ ion at m/z 1818 is formed by
hydrogen migration from somewhere along the peptide backbone (preferentially
on the or carbon relative to the tryptophan residue in position 15), and subsequent
cleavage of the amide bond between the modified carboxy terminus and the
tryptophan residue in position 15. The two peaks to the immediate right of the
[M+H]+ ion (m/z 1879) are the sodium and potassium adducts of gramicidin D.
Photodissociation of these species yielded no loss of the modified carboxy
terminus. The appearance of the [M-61]+ ion as the major fragment peak in the
spectrum suggests that protonation occurs preferentially on the modified carboxy
terminus of the peptide (e.g., the most basic site) and charge site driven
fragmentation results. The only other structurally relevant peak observed was the
b14 ion at m/z 1636. As seen with the previous IRMPD experiment, a sequence-
specific b ion is produced which could be used to obtain ladder sequence
information to verify peptide sequences.
Carbohydrates and Oligosaccharides
One of the biggest challenges in biochemistry is the identification and
sequence analysis of complex carbohydrates and oligosaccharides. Perhaps the
most difficult problem in carbohydrate chemistry is the investigation of biologically
active binding-recognition systems. The biologically active portion of a

Figure 5-4: IRMPD of the singly charged [M+H]+ ion of gramicidin D. The [M+2H]+ ion is present for mass
calibration purposes. The two peaks just above the [M+H]+ ion represent the sodium and
potassium adducts of gramicidin D respectively. The peak at [M-61]+ represents cleavage of the
modified carboxy terminus with subsequent hydrogen ion migration.

345

346
carbohydrate or oligosaccharide can also be linked to a variety of other
biochemically important compounds to form glycoproteins, peptidoglycans,
lipopolysaccharides, and glycoshpingolipids. To determine the structure of the
oligosaccharide, or oligosaccharide portion of a complex biomolecule, several
steps are necessary to establish structure: (1) determination of the reducing and
nonreducing ends of branched or straight chain oligosaccharides, (2)
determination of monosaccharide ratios and ring sizes, (3) determination of the
internal sequence of individual monosaccharides, (4) determination of branching
points (if any), (5) anomeric data relating to linkage type and bonding, (6)
presence of modified sugars possibly containing acyl groups, phosphates,
sulfates, pyruvates, cyclic acetals, or taurine moieties. To obtain the
aforementioned information, a myriad of techniques are employed, including wet
chemical, enzymatic, antibody or lectin affinity, chromatography (thin layer, gas,
liquid, paper), nuclear magnetic resonance, circular dichroism, and mass
spectrometry.47,294'297
The major contribution of mass spectrometry to carbohydrate analysis has
been that of high sensitivity (nanomole to femtomole) studies which have included
selective detection of glycopeptides in protein digests, multi-residue confirmation
of aminoglycoside antibiotics, sequencing of cationized carbohydrate antibiotics,
identification and linkage position determination of reducing ends in
oligosaccharides, and the structural analysis of steroidal oligoglycosides.275,298"301
Collision-induced dissociation studies needed to help determine oligosaccharide

347
sequence information have also been successfully performed using FAB and
electrospray ionization techniques for generation of abundant parent ion
populations.302'304
In this section, an alternative approach to traditional CID for determining
carbohydrate structure is presented. Photodissociation in the IR region is a
natural choice for cleaving glycosidic bonds, identifying ring structure, or possibly
determining "post translational" modifications to monosaccharide residues. This
is due mainly to the high photon absorption cross-section seen for C0C
linkages (see chapter 2). Although the appropriate derivatization chemistry and
enzymatic data are needed to help determine internal oligosaccharide sequence
information, the use of photodissociation in combination with electrospray
ionization can provide a basis for tackling the more difficult problem of anomeric
confirmation in the gas-phase.
The first portion of this section discusses the determination of the number
of hydroxy substituents on monosaccharide residues. The initial studies
performed were done using a solids-probe and ammonia chemical ionization in
order to test the feasibility of the photodissociation technique. Original
comparisons with CID data showed increased dissociation efficiencies, and
reduced analysis times for IR photodissociation experiments. The second portion
of this section covers the use of photodissociation for cleaving the glycosidic
bonds of straight chain oligosaccharides (e.g., ammonium adducts of raffinose
and stachyose) using the instrument described in chapter 4.

348
Monosaccharide Cleavage
The first step in obtaining relevant carbohydrate sequence data employing
photodissociation is understanding the ring fragmentations of simple
monosaccharides subunits. The goal was to determine the number of ring
substituents on an individual monosaccharide. In this section the ring
dissociations of 2-deoxy-D-glucoseand 1 -O-methyl-D-glucopyranoside (structures
shown in figure 5-5) are discussed, which were used as a feasibility study for
more advanced applications (e.g., glycosidic bond cleavages). The samples were
introduced via a solids probe into a Finnigan MAT ITMS mass spectrometer set
up for photodissociation experiments as described (see figure 2-1) in chapter 2.
Sample ionization was accomplished by ammonia Cl with reaction times on the
order of 65 to 90 ms and ammonia pressures of 5.5x1 O'6 torr. For CID
experiments, the helium buffer gas pressure was 1.0x10^ torr (uncorrected). To
maximize photodissociation efficiency, the ion trap was operated without helium
buffer gas.
The CID mass spectrum of the ammonium adduct of 2-deoxy-D-glucose
(m/z 182) is shown in figure 5-6. The observed dissociation efficiency of the
ammonium adduct ion was 90%. The major peak produced at m/z 164 involves
the loss of neutral H20 from the 2-deoxy-D-glucose ring. The peak at m/z 147
also showed the loss of neutral H20 with an additional loss of ammonia. Low
intensity ions at m/z 146 [M+NH4+-2H20] and m/z 129 [M-2H20-NH3]+ indicated
the loss of a second neutral H20 from the monosaccharide ring. To determine

Figure 5-5:
Structure of the simple monosaccharides 2-deoxy-D-glucose and 1 -0-methyl-/?-D-glucopyranosside
used for evaluation of the IRMPD technique for ring cleavage fragmentation.

H
2-deoxy-D-Glucose
mw=164.0
H
1-O-Methyl-D-Glucopyranoside
mw=194.1
350

Figure 5-6: The MS2 and MS3 IRMPD spectra of the ammonium adduct of 2-
deoxy-D-glucose. In the first experiment the [M+NH4+] ion at
m/z 182 was mass isolated and fragmented at a q of 0.3. For
the MS3 experiment the first sequential product ion from the MS2
process at m/z 164 was mass isolated and fragmented at a q of
0.3. A significant reduction in CID efficiency marked the MS3
spectrum.

Intensity Intensity
352
3000 i
2500
MS/MS Spectrum
q=0.3, Amplitude=135 mV
Time=10 ms
(M+NH^-HjO)
m/z 164
2000
1500
1000
500 m/z 129
0
(M+NH4-H20-NH3)
m/z 147
(M+NH4+-2HjO)
m/z 146
z (M-2H20-NH3)
\
1
(M+NH4+)
m/z 182
r
TT|
t
400 i
300
200
100
MS/MS/MS Spectrum
q=0.3, Amplitude=250 mV
Time=10 ms
(M+NH4-H20)+
m/z 164
(M+NH4-3HjO-NBLj)+
m/z 111
(M+NH.-HjO-NHj)
m/z 147
(M+NH4-2H20)
m/z 146
\
(M+NH4-2HjO-NHj)+
m/z 129
(M+NH4-NH3)
m/z 165
100
120
140
m/z
160
180
0

353
the presence of other ring substitutions (e.g., more OH or other functional
groups) a third sequential step of MS was performed on the m/z 164 ion
[M+NH/-H20].
The MS3 spectrum of the ammonium adduct of 2-deoxy-D-glucose is also
shown in figure 5-5. Due to the stability of the m/z 164, dissociation efficiencies
were only *=13%. Diagnostic ions indicating the loss of one and two neutral H20
molecules from the ring at m/z 147, m/z 146, and m/z 127 were present in the
MS3 spectrum as well as the MS2 spectrum. An additional ion at m/z 111
[M+NH4+-3H20-NH3] signified the presence of a third hydroxy substituent on the
monosaccharide ring. There was no indication of cleavage of the hydroxyl group
attached to position 6 of the pyranose ring, nor was there any evidence for the
loss of methanol (e.g., cleavage of the C5C6 bond) from the pyranose ring.
In the IRMPD spectrum of the ammonium adduct of 2-deoxy-D-glucose
(figure 5-7), the m/z 164, 147, 146, 129, and 111 peaks (indicating the losses of
one, two and three hydroxyl moieties from the pyranose ring) were all present in
a single MS/MS experiment. This was due to the formation of product ions from
the m/z 182 parent species during the laser irradiance time (85 ms). As the first
sequential products were formed, they could undergo further dissociation to yield
more fragment ions, since there was no mass isolation step during the laser
irradiance period. The dissociation efficiency of the m/z 182 ammonium adduct
was *=100%. As the laser irradiance time was increased (figure 5-7), this
essentially drove the series of consecutive reactions for the ammonium adduct of

Figure 5-7: IRMPD spectra of the ammonium adduct of 2-deoxy-D-glucose
at 85 ms and 120 ms irradiance time. High dissociation
efficiencies were observed for IRMPD, since the kinetic energy
of ion motion does not need to be converted to the internal
vibrational energy needed for fragmentation. The number of
hydroxyl substituents on the ring can be determined much easier
with IRMPD, since dissociation efficiency is independent of ion
population.

m/z
o
o
K)
O
4-
O
o\
O
00
o
Intensity
'M
O
O
O
o
o
'Jl
o
o
O
o
o
J I L
J 1L
I I I
J I L
S-S
s"
II
z
* j
o
li
2
+
o
i!
00
Irradian ce Time=120 ms
(M+NH4-3H,o-NHj)+ Energy=2.0 J
m/z 111
(M+NH4-2HjO-NH,)
m/z 129
Intensity
i
h-
K
K>
u>
o
Ul
O
o
o
O
o
O
o
o
o
o
o
o
> I I I I > 1 *I
+
355

356
2-deoxy-D-glucose to its final product ion, the m/z 111 species [M+NH4+-3H20-
NH3], The consecutive reaction curves in figure 5-8 for the IRMPD of the
ammonium adduct of 2-deoxy-D-glucose show that only 61% of the fragment ions
produced from the photodissociation process are structurally significant ions; the
remaining 39% represent the direct dissociation of the ammonium adduct
complex to form the ammonium ion and corresponding 2-deoxy-D-glucose
neutral. The ammonium ion intensity curve in figure 5-8 was incomplete due to
the complex series of back reactions observed for the NH4+ ion with neutrals
present in the trapping region during the laser irradiance period (e.g., longer
irradiance times correspond to longer reaction times for the dissociated NH4+ ion).
The trace is present only to confirm the presence of this dissociation pathway; the
experimental conditions were altered in a separate experiment to efficiently trap
the m/z 18 NH4+ ion.
The bottom half of figure 5-8 shows the individual structurally diagnostic
portion of the consecutive reaction curves. It is clearly seen that a series of
complex consecutive and competitive reactions occur during the
photodissociation process. The presence of the competitive reaction curves (e.g.,
m/z 164 and m/z 165) could involve reactions where the critical energy of the two
reaction channels are equivalent.
Although the types of fragment ions produced by CID and PID in this
example were the same (which might be expected since both are low energy
processes), the time needed to produce each spectrum was radically different.

Figure 5-8: Consecutive reaction curves for the IRMPD of the ammonium
adduct of 2-deoxy-D-g!ucose. The top half of the figure shows
the sum of the major product ion intensities, plus the presence
of the direct dissociation of the complex to showing the
ammonium species at m/z 18 (dashed line). The bottom portion
of the figure represents the individual curves comprising the
major product ion intensity curve in the top half of the figure.

Intensity Intensity
358
250 i
200
150
100
50
Product Ion Growth Curves
m/z 165 m/zl46
m/z 164 + m/z 129
m/z 147 m/z 111
0 20 40 60 80 100 120 140 160 180 200
Laser Irradiance Time in ms

359
For the case of single-frequency CID, it is critical that a constant ion population
is present so that re-tuning of the instrumental parameters involved for the CID
process (e.g., 20 min even for an experienced operator) need not occur (see
figure 5-1). Also, in order to gain the same information obtained from the IRMPD
experiment in figure 5-7, a third stage of mass analysis must be performed which
means that the efficiency of the MS3 step is directly influenced by the parent ion
population frequency shifts, which translates down to the MS/MS efficiency, which
then can drastically affect the formation of the parent species for the MS3 process.
In addition, the time required for tuning of the ion trap for an MS3 process can
take on the order of 30 min. More importantly, for the case of real-time sample
concentration profiles observed in gas or liquid chromatography, single-frequency
CID experiments could produce a limited amount of structural information due to
shifting peak concentration profiles. For the IRMPD process, there is an initial
time investment for turning on and tuning the laser; however, since the
dissociation event is independent of ion population (e.g., only depends on the
ability of functional groups to absorb photons), MS/MS experiments with
photodissociation can be carried out for real-time chromatographic analysis. In
addition, "multiple" stages of mass spectrometry for the IRMPD process can be
obtained by simply increasing the irradiance time for the parent/product ion
species. The one potential drawback of the IRMPD process could be the
formation of low intensity structurally significant ions which are never observed
because the laser irradiation period is set such that upon fragment ion formation,

360
the newly formed product ion absorbs enough photons to quickly dissociate to
the next sequential product ion.
For the other monosaccharide investigated in this study (1-0-methyl-/?-D-
glucopyranoside), the pyranose ring contained a methoxy substitution at position
one. The reason for examining the fragmentation of this compound was to
investigate the ability of IRMPD to cleave substituted groups on the pyranose ring
structure (e.g., naturally occurring as well as permethylated derivatives). The
methoxy substituent is the simplest case and serves quite well for evaluation
purposes. The conditions for instrument operation were the same as described
above for 2-deoxy-D-glucose.
The IRMPD spectrum of 1 -0-methyl-/?-D-glucopyranoside is shown in figure
5-9. For an irradiance time of 100 ms, structurally diagnostic ions at m/z 194
[M+NH4+-H20] indicating the loss of one hydroxyl group, m/z 180 [M+NH4+-
CH3OH] indicating the loss of the substituted methoxy group, m/z 162 [M+NH4+-
CH30H-H20] showing the loss of methoxy and one hydroxyl, m/z 144 [M+NH4+-
CH30H-2H20] showing the loss of methoxy and two hydroxyl groups, and m/z
126 [M+NH4+-CH30H-3H20] marking the methoxy loss plus 3 hydroxyl groups
(i.e., removed of all ring substituents) are observed.
Although photodissociation experiments have been shown to remove all
substituents from the pyranose ring structure to aid in ring identification, it does
not indicate the substitution position for each individual loss. Perhaps the most
important information provided by photodissociation of the ammonium adducts
of monosaccharides will be the identification of "post translationally" modified

Figure 5-9: IRMPD spectra of the ammonium adduct of 1 -0-methyl-/?-D-
glucopyranoside at irradiance time of 100 and 150 ms
respectively. The peak at m/z 180 is the loss of methanol,
indicating the presence of a modified group on the ring structure
of the monosaccharide.

Intensity Intensity
362
(M+NHJ+
100 120 140 160
200 220
m/z
180

363
substituents.
Raffinose
To determine the applicability of photodissociation to sequencing
oligosaccharides, two straight chain oligosaccharides, raffinose (O-cr-D-
galactopyranosyl[1-6]-cr-D-glucopyranosyl-/?-D-fructofuranoside) and stachyose (a-
D-galactopyranosyl-[1 -6]-cr-D-galactopyranosyl-[1 -6]-o-D-glucopyranosyl-[1 -2]-/?-D-
fructofuranoside), were chosen for study. In figure 5-10 are shown the structures
of these two compounds. This section will focus on the photodissociation and
fragmentation of raffinose, an oligosaccharide typically used in tissue culture
media.
A 20 pmol///L solution of raffinose was prepared in a 50:50 methanol:water
solution with 3 mM NH4OH. Samples were directly infused through the
electrospray interface as described previously in chapter 4. For photodissociation
experiments, a pulsed C02 excimer laser was employed as described in the
photodissociation set-up section in chapter 4. The laser was capable of an
approximate 3-5 ns pulse with an energy of 1.1 J (at 944 cm'1). Due to the short
pulse of laser energy and the high photon absorption cross section for the
COC stretch, the ion trap was operated with buffer gas at a pressure of
1.0x1 O'4 torr (uncorrected) for both CID and photodissociation experiments. The
major (parent) ion produced from the electrospray process was the ammonium
adduct of raffinose at m/z 522.
The fragmentation nomenclature employed for the spectral interpretation

Figure 5-10: Structures of the oligosaccharides raffinose (0-a-D-galactopyranosyl[1 -6]-or-D-glucopyranosyl-/?-D
fructofuranoside) and stachyose (cr-D-galactopyranosyl-[1-6]-cr-D-galactopyranosyl-[1-6]-a-D
glucopyranosyl-[1-2]-/?-D-fructofuranoside).

D-Galactose
ch2oh
oh ) o
OH
D-Galactose
o'
OH
D-Galactose
CH2
o
OH
OH
O'
,CH
D-Fructose
ch2oh
OH
2 D-Glucose
o
OH
o-
OH
k3
ch2oh
OH
CH2OH 9
OH
D-Fructose
HO
3
ch2oh
Raffinose
mw=504.2
OH
Stachyose
mw=660.2
365

366
of the raffinose and stachyose data was that of Domon and Costello.305 In figure
5-11 is shown a simple fragmentation of a model disaccharide with cleavage
points following the Domon and Costello nomenclature. Charge retention on the
reducing portion of the sugar is indicated by the fragments X (ring opening
cleavage), Y and Z. For charge retention on the nonreducing terminus of the
sugar, the fragments are designated A (ring opening cleavage), B, and C. The
superscripts in figure 5-11 indicate the exact bond cleavage positions, while the
presence of subscript numbers identify the residue number. Branches are labeled
or, p, y where a is the branch with the highest mass.305
In figure 5-12 are shown both the CID and PID (pulsed C02 laser) spectra
of the ammonium adduct of raffinose. The major product ion from the single
frequency CID process was m/z 505 [M+H]+ arising from a loss of neutral
ammonia from the adducted species. Although this ion gave no structural
information, it could be used to help confirm molecular weight. The presence of
the peaks at m/z 343 and m/z 326 indicate the loss of one of the terminal
monosaccharide groups (D-fructose from the reducing end or D-galactose from
the nonreducing end), with the peak at m/z 326 showing an additional loss of
ammonia. Assignment of these ions as either B2a or is not possible since D-
galactose and D-fructose have the same mass. However, from a biochemical
standpoint, the most labile bond in the system is the glucopyranosyl-jff-D-
fructofuranoside bond, which would indicate that the ions formed in this mass
spectrum are B type ions with charge retention on the nonreducing terminus.

Figure 5-11:
Carbohydrate fragmentation nomenclature as determined by Domon and Costello (from reference
305). Charge retention on the reducing terminal is indicated by the X, Y, and Z fragmentation.
Charge retention on the nonreducing end is represented by the A, B, and C fragments. Ring
opening cleavage fragments are X and A with superscripts denoting bond breaking points on the
ring.

Oligosaccharide Fragmentation Nomenclature
Nonreducing End
1,5
- X
ch2oh/
- Y
/ CH2OH
368

Figure 5-12: CID and photodissociation spectra of the ammonium adduct of
raffinose. The major fragmentation observed in the CID
spectrum is the [M+H]+ at m/z 505 which essentially confirms
the molecular weight, but provides no structural information.
The photodissociation spectrum shows complete fragmentation
information for the trisaccharide.

Intensity
370
m/z

371
These results would also be consistent with other low-energy CID results for
straight chain oligosaccharides reported in the literature.47 The fragmentation
analysis for raffinose is shown in figure 5-13.
Also shown in figure 5-12 is the photodissociation spectrum (2-laser pulses)
of the ammonium adduct of raffinose. Compared to the CID spectrum, PID
produced only a small [M+H]+ peak, with the majority of the ion signal producing
structurally diagnostic ions. The two largest ions in the spectrum were the m/z
326 and m/z 343 ions indicative of the loss of a terminal monosaccharide (as
seen with the CID spectrum). Also present were a series of peaks indicating
successive losses of H20 at m/z 308, m/z 290, and m/z 272 from the ion at
m/z 326. An additional peak at m/z 164 represents the cleavage of the internal
monosaccharide residue (D-glucose) to form either the B1a or Z1o ion.
For determination of an actual unknown sample, identification of the
internal sequence ions of an oligosaccharide would be extremely difficult using
mass spectrometry alone (even using the appropriate derivatization chemistry),
since anomeric and linkage position data are stereospecific. It is important to
point out that with photodissociation, the entire sequence of the trisaccharide was
obtained without complicated instrument tuning procedures. In contrast for CID,
a third stage of mass spectrometry would have to be performed in order to obtain
the same information (e.g., cleavage of the internal residue) as from the
photodissociation experiment. This means that for photodissociation another
sample could be analyzed immediately after the first, since no tuning of

Glycosidic bond cleavages of the ammonium adduct of raffinose for the collisional and photon
activation processes.
Figure 5-13:

Raffnose
Bia
m/z 164 (-NH3)
m/z 343 Z2a
m/z 326 (-NH3)
D-Glalactose
D-Glucose
m/z 343
m/z 326 (-NH3)
D-Fructose
m/z 164 (-NH3)
Zia
373

374
instrument parameters is needed to induce fragmentation (only an initial time
investment at the beginning of the days analysis).
Stachvose
The next compound studied was the tetrasaccharide stachyose (see figure
5-10 for structure). Stachyose (a-D-galactopyranosyl-[1 -6]-or-D-galactopyranosyl-
[1 -6]-cr-D-glucopyranosyl-[1 -2]-/?-D-fructofuranoside) differs from raffinose by the
one additional D-galactose on the nonreducing end and a [1-2] /? linkage from
glucose to fructofuranoside. Electrospray and instrumental conditions were the
same as for raffinose in the previous section. The sample concentration was 20
pmol///L infused directly through the ESI source.
The single-frequency CID spectrum of the ammonium adduct of stachyose
(m/z 684) is shown in figure 5-14. The major fragment peaks in the spectrum
were observed at m/z 505 and m/z 488 (B^ or ZgJ, indicating the loss of the
nonreducing galactosyl terminal or the fructosyl residue which has its reducing
end bonded /?-[1-2] to the glucopyranosyl moiety. The peak at m/z 488 shows
the same cleavage with loss of ammonia. The peaks observed at m/z 343 and
m/z 326 indicate the loss of an internal sugar residue, either D-galactose or D-
glucose (B^ or ZJ. The small peak at approximately m/z 589 may indicate a
ring-opening reaction which cleaves in the 3 and 5 position of D-galactose to form
the 3,5X3a ion or cleaves in the 2 and 4 position of D-fructose to produce the 2,4A3a
ion. Isotopic labeling experiments would be required for direct mass assignment
of this ion. The peak at m/z 680 was an occasional artifact peak observed, and

Figure 5-14: CID spectrum of the ammonium adduct of stachyose. Only the loss of two monosaccharide units
is observed under ICD conditions (one terminal and one internal residue). A small degree of ring
cleavage on a terminal monosaccharide unit is observed by the presence of the z,4A3o or 3,6X3a peak.

Intensity
376

377
depended on the amplitude of the resonance ejection frequency and single
frequency CID signal. In figure 5-15 is shown the fragmentation scheme for the
ammonium adduct ion of stachyose. As was the case for raffinose, another stage
of mass spectrometry is needed to obtain entire sequence information.
The photodissociation spectra for 1,2 and 3 laser pulses of the ammonium
adduct of stachyose are seen in figure 5-16. After two laser pulses, the entire
sequence of the tetrasaccharide is obtained, with the peak at m/z 164 indicating
either the nonreducing terminus (B1o) ion or the blocked reducing end D-fructose
ion (Z^). For the three laser pulse spectrum, peaks at m/z 146 and m/z 128
indicate the successive losses of H20. Interestingly enough, the small peak
indicating ring opening cleavages at approximately m/z 589 is not observed in the
photodissociation spectra. However, an unknown peak at m/z 640 is observed
in all three spectra (structure unknown). As was the case for raffinose, the [1-2]
/? linkage between the D-glucosyl and D-fructofuranosyl is the most labile
glycosidic bond in stachyose. Therefore, the sequential product ions in figure 5-
16 probably represent charge retention the nonreducing terminus, again
consistent with previously reported low-energy CID data47 Fragmentation
patterns for the ammonium adduct of stachyose are found in figure 5-15.
Oligonucleotides
The first real advances in mass spectrometry for oligonucleotide analysis
came with the advent of laser desorption and electrospray ionization

Figure 5-15:
Glycosidic bond cleavages of the ammonium adduct of stachyose for the collisional and photon
activation processes. The ring cleavages were only observed in the CID spectrum.

35X3
CH2OH /
^3a
m/z 505
m/z 488 (-NH3)
la
m/z 164 (-NH3)
2a
m/z 343
m/z 326 (-NH3)
^2a
m/z 343
m/z 326 (-NH )
OH
MWAlWiWW
3a
m/z 505
m/z 488 (-NH )
CH20H 0
HO'i m/z 164 (-NHL)
ch2oh
OH
2,4A
3a
379

Figure 5-16: Photodissociation spectra of the ammonium adduct of stachyose
for one, two, and three laser pulses. After two laser pulses The
entire sequence for stachyose was obtained. With the addition
of a third laser pulse, ring fragmentations (e.g. loss of H20
indicating the presence of the hydroxyl group) were observed for
the terminal monosaccharide residue (m/z 128 and m/z 146).

Intensity
381
700

382
methods.306,307 Recent successes in the field of electrospray have enabled
ionization of oligonucleotides up to 76 base pairs.115 Typical electrospray spectra
include sodium attachment to the multiply-charged anions of the oligonucleotides;
this can complicate molecular weight analysis and has proved to be the single
most limiting factor in electrospray mass spectrometry of this compound class.115
In addition, the weaker signals generated by the oligonucleotide anions can limit
sensitivity of the electrospray technique (compared to the more abundant positive
ion signal for peptides and proteins).
Perhaps the most difficult challenge of negative ion electrospray for the
analysis of oligonucleotides is to obtain structural information from MS/MS
experiments. To date, only a handful of articles has appeared on the MS/MS of
negatively-charged oligonucleotides.111,112,308,309 The main problem is in the area
of data interpretation, where multiple charge states and adduct formation can
complicate even the simplest of CID spectra. Previous reports have centered on
the CID of small, multiply-charged negative ion DNA fragments (up to 8 base
pairs) using electrospray ionization/ion trap mass spectrometry, and linker DNA
8-mers using FTICR and photodissociation.111,308 Both techniques show a great
deal of promise for the analysis of larger oligonucleotides.
The fragmentation nomenclature employed for tandem MS/MS studies of
oligonucleotides was developed by McLuckey et al.308 Shown in figure 5-17 is a
triply-charged tetranucleotide with bases B, through B4. Cleavage points along
the phosphodiester backbone are indicated by the lower case letters a, b, c, and

383
d for charge retention on the 5 end and w, x, y, and z for charge retention on the
3 end of the oligonucleotide. Subscripts indicate the number of bases contained
in the fragment. Fragmentation of the nucleoside bases is represented by an
upper case B, where B is the individual base A, G, C, T, or U. A subscript is
assigned to the B symbol to represent the position of the base from the 5 end
of the molecule. Bases are represented parenthetically to avoid confusion with
the normal sequence terms (e.g., B^A)).
The merits of photodissociation for DNA/RNA sequence analysis are
discussed in this section. Two sets of RNA dimers were studied in order to
determine the feasibility of the technique (pulsed-C02 laser photodissociation) for
future sequencing experiments.
RNA Dimers
Two RNA dimers, adenyl adenosine (ApA) and adenyl cytidine (ApC), were
used for the photodissociation experiments. The samples were obtained from the
Core Biotechnology Facility at the University of Florida, and were purified to
remove sodium salts. Sample concentrations were 2 pmol///L in a 70:30
methanol:water solution. Electrospray conditions and instrumental parameters
were set as described in chapter 4 of this dissertation. The photodissociation set
up employed the pulsed laser described in chapter 4.
The lone parent species produced for the negative ion electrospray for the
two dimers was a singly-charged negative ion (M-H)' at m/z 595 for ApA and m/z

Figure 5-17:
Oligonucleotide fragmentation scheme as defined by McLuckey et al. (from reference 308). The
letter Bn represents the individual nucleoside bases, with position one defined form the 5 end. The
w, x, y, and z fragments have the charge retained on the 3 end while the a, b, c, and d fragments
have the charge retained on the 5 terminus.

Oligonucleotide Fragmentation Nomenclature
HO
N
w3 x3 y3 z3
\
o
r"o '

/ /
/
A
o
w2 x2 y2 z2
/
t
O
-o
w, x, y!
/ / / /
\
O
-o
\
OH
a.
/ / r r r r r r r r r
b] Cj dj a2 b2 c2 d2 a3 b3 c3 d3
385

386
571 for ApC. No sodium adducts were observed in the spectra. The negative
charge is located on the phosphodiester bridge, as shown in figure 5-17. The
maximum number of negative charges attached to any oligonucleotide is equal
to the number of phosphodiester bridges present, plus any phosphate groups
attached to the free 3 or 5 positions on the ribose sugar backbone. For a DNA
tetramer (ApApApA) run on the electrospray ion trap, the highest charge state
observed was that of (M-3H)'3, indicating the presence of the three
phosphodiester bridges involved in the ionization process.112,308
The photodissociation spectrum, structure and cleavage points of adenyl
adenosine and adenyl cytidine is shown in figure 5-18. The photodissociation
spectra represent 5 laser shots from the pulsed C02 laser. In the top spectrum
is shown the fragmentation of the anion of ApA, with two characteristic peaks at
m/z 134 and m/z 329. The peak at m/z 134 represents loss of the adenine base
from either position one or position two; therefore, no subscript for the Bn(A) peak
can be assigned. The second peak at m/z 329 indicated direct cleavage of the
phosphodiester bond (PO), to form either a C,' ion (charge retention on the 5
side) or a XT ion (charge retention on the 3 side).
The fragmentation of the anion ApC is shown in the bottom portion of
figure 5-18. As with the previous example, cleavage of the adenine base is
observed and the fragment at m/z 134 can be assigned to the B/ peak. The
second peak observed at m/z 329 represents cleavage of the phosphodiester
bond to form the C/ peak containing the ribose sugar and cytosine base.

Figure 5-18: Photodissociation spectra (5 laser shots) of the negatively
charged RNA dimers adenyl (ApA) adenosine (top) and adenyl
(ApC) cytidine (bottom). The individual structures and
fragmentations are also shown. Bond cleavage for the
photodissociation process occurs at the phosphodiester bond
PO where there is a relatively high photoabsorption cross-
section in the IR region.

Intensity
388

389
Compared with the spectra generated from the CID of oligonucleotides, the
types of ions observed in photodissociation spectra differed markedly.308 For CID
spectra, preferential cleavage occurs at the CO of the sugar (charge retention
of the 3 end) to form w-type fragment ions. Complementary cleavages to form
ions with a (a-B(A)) type fragmentation were also observed for oligonucleotides
containing adenine base(s). The photodissociation spectra presented in this
section show preferential cleavage at the PO, which has a fairly high
photoabsorption cross-section; these data correspond well with those obtained
using IRMPD in the FTICR.111
Carbohydrate Antibiotics
Carbohydrate antibiotics are some of the most important compounds in all
of biochemistry. In the next decade, and increasing emphasis will be placed on
discovery of new carbohydrate antibiotics, as microorganisms continue to develop
resistance to current antibiotic treatments. Over the years, one of the most
important classes of carbohydrate antibiotics has been those of macrolide
antibiotics.310 Macrolide antibiotics all contain a large lactone ring (aglycone of
12 to 22 atoms) with no nitrogen atoms and few double bonds. Linked to the
aglycone ring are one or more sugars which can be nitrogen-containing. These
sugar linkages to the aglycone ring are critical for biological activity. The
structural determination of the various functional groups comprising these
compounds and other carbohydrate antibiotics has long been of

390
interest.298*99302311-322
Perhaps the most widely used macrolide antibiotic is erythromycin. This
antibiotic is typically used to control respiratory infections, and is particularly
effective against Gram-positive organisms involved in streptococcal,
staphylococcal, and pneumococcal infections.310 The mechanism of action
involves binding of the antibiotic to the 50S subunit of the bacterial ribosome, thus
blocking the action of peptidyl transferase in the peptide elongation process. In
figure 5-19 is shown the structure of erythromycin, with the aglycone moiety and
two attached monosaccharides (D-desosamine and L-cladinose).
It is the purpose of this section to evaluate the ability of the
photodissociation technique for structural elucidation of carbohydrate antibiotics.
Studies were performed with protonated erythromycin as a model compound in
order to understand the fragmentations observed with basic macrolide antibiotics.
Macrolide Antibiotics Erythromycin
The erythromycin standard was obtained from ICN Pharmaceutical (Costa
Mesa, CA). Solutions were made up in 50:50 methanol:water with 0.2% acetic
acid. The concentration of the electrospray standard was 5 pmol///L. All
electrospray and instrumental conditions were the same as described in chapter
4. The photodissociation set-up employed a pulsed C02 laser as discussed in
chapter 4 of this dissertation.
The preferential site of protonation is the tertiary amine group on the D

Figure 5-19:
Structure of the macrolide antibiotic erythromycin showing the lactone ring (aglycone moiety), the
amino sugar D-desosamine and nonamino sugar L-cladinose.

Erythromycin
raw = 733.4
L-Cladinose
392

393
-desosamine sugar. Therefore, it would be predicted that charge-site
fragmentation should be driven from this portion of the ion. In figure 5-20 is seen
the photodissociation spectrum (7 laser pulses) of protonated erythromycin. The
low mass peaks at m/z 100 f-5AJ and m/z 116 (l2AJ are charge-site driven
fragmentations for the ring opening of the D-desosamine sugar. Direct cleavage
on either side of the glycosidic bond (with accompanying hydrogen transfers) of
the D-desosamine sugar is indicated by the two fragments m/z 158 (B1a) and m/z
174 (C1a). The appearance of the small peak at m/z 174 signifies the transfer of
two hydrogens from the amino sugar to the aglycone-L-cladinose neutral. The
unusual aspect of this fragment is the fact that it is primarily observed in the
charge-remote fragmentation process, indicative of high-energy collision
298,322
processes.
Even more astounding is the presence of a large peak at m/z 576 (Yop),
indicating loss of the L-cladinose sugar from the aglycone ring (cleavage on the
L-cladinose side of the glycosidic bond). The peak at m/z 558 could either be
cleavage of the glycosidic bond on the aglycone side to form the Zoe ion, or loss
of water from the m/z 576 ion. The peaks at m/z 540 and m/z 522 are
consecutive water losses occurring from the aglycone ring (e.g., three hydroxyl
substituents present). Around the protonated erythromycin region (m/z 734, m/z
716 and m/z 698) is observed two consecutive losses of water. This indicates
that some form of charge migration or charge-remote fragmentation mechanism
is in place since there is only one hydroxyl group present on D-desosamine (e.g.,

Figure 5-20:
Photodissociation spectrum (7 laser pulses) oftheprotonated [M+H]+ ion m/z734 of erythromycin.
The mechanism and composition of the fragments labeled aglycone ring cleavages are unknown.
Many of the fragments observed (e.g. C1fl and Y0/J) correspond to data obtained using high energy
CID as seen on magnetic sector instruments.

100
80
60
40
20
0
la
m/z 158
,5A
la
m/z 100
/
0JA
m/z 116
\
la
m/z 174
(M+H)+
m/z 734
Zp or Y0p- h2o
m/z 558
Consecutive Water Losses
m/z 540 and m/z 522
Aglycone Ring Cleavages
Y
*op
m/z 576
(m-h2o>
m/z 716
+
too
200 300
400 500
m/z
600
700
800
395

396
where protonation occurs on the tertiary nitrogen). The fragmentation pathways
observed for protonated erythromycin are shown in figure 5-21.
The photodissociation spectrum of protonated erythromycin bears an
uncanny resemblance to high-energy CID spectra taken on magnetic sector
instruments.284,308 Protonated species under high-energy collisions give
information on the monosaccharide residues present on the aglycone ring moiety.
These fragmentations typically arise from a charge-remote mechanism. One other
interesting note is the formation of a series of ions (from m/z 200 to m/z 500) in
the photodissociation spectrum indicating ring opening reactions of the aglycone
moiety. At this time reasonable neutral losses and peak assignments have not
been made, and will require an extensive amount of investigation. It is theorized
that these ring opening cleavages are perhaps driven by the photoactivation of
the COC ether group in the aglycone ring (with some form of accompanying
charge migration or charge-remote fragmentation mechanism). No ring
fragmentations have been observed in previously published high-energy CID
spectra. The peak observed at m/z 403 could be the loss of D-desosamine from
the Y0fi at m/z 576 to form the S type ion.

Figure 5-21: Fragmentation of the macrolide antibiotic erythromycin obtained with photon activation (7 laser
pulses). The fragmentation observed mimicked that of the high energy CID process where charge
remote fragmentation mechanisms are common.

Aglycone Moeity
Erythromycin
mw = 733.4
398

CHAPTER 6
CONCLUSIONS AND FUTURE WORK
As illustrated by the overall success of this project to date, the future for the
ESI/ion trap is very promising. Alternative ion activation techniques for structural
elucidation such as photodissociation show a great deal of promise in tackling the
difficult problems associated with analytical biochemistry. The major
accomplishment of this work is reflected not just in low detection limits, but also
in the application of innovative new ideas in order to push ion trap mass
spectrometry to its limits.
In order to accomplish the major goal of the project; photodissociation of
biological molecules, a very methodical course was taken. It began with
increasing the photoabsorption pathlength of the photodissociation process, to
overcome the problems (e.g. low absorption cross-sections for organic species,
and limited power and tunability of the C02 laser output) traditionally associated
with IRMPD. This increase in pathlength was accomplished by mounting three
spherically asymmetric concave mirrors in the radial plane of the ring electrode,
thereby producing a multipass cell. It was demonstrated that the rate coefficients
for various photodissociation processes were increased by a factor of 8 to 10 over
the traditional single-pass experiments. Other fundamental investigations included
the photon absorption process, IRMPD kinetics, the study of consecutive
399

400
reactions, and the effects of buffer gas on photodissociation efficiency. The
information obtained in these fundamental studies, was used successfully to
optimize the photodissociation/instrument parameters used in the study of
peptides/proteins, carbohydrates, and oligonucleotides described in chapter 5.
The techniques used to acquire the wavelength-dependent spectra
obtained for the allyl bromide ion and protonated diglyme may provide a method
for examining the gas-phase IR absorption profiles of various ionic species.
Although the fine structure is not available from these experiments, general
features can be observed indicating the presence of various functional moieties.
The unique design of the multipass ring electrode also lends itself to
detection of coupled motion (r- and z-directions) in the quadrupole ion trap.
Since a high photon density is observed in the radial plane of the ring electrode,
any excitation signal applied to the endcaps which increases the axial excursions
of the ions can be detected by a sharp decrease in photodissociation efficiency
of the parent species. The observed decrease in photodissociation efficiency is
due to the reduced amount of time the ions spend in the radial plane of the ring
electrode.
The next step of the project involved the design and construction of an ion
transmission device (the rf-only octopole) to transport ions from the LC/MS
interface region (ESI source) to the ion trap analyzer in the most efficient manner
possible. The major advantage of this device is the cubic dependence of the rf
restoring force for a given ion as it moves off axis. This is particularly important

401
when ions pass from the high pressure ESI source region (multiple collision
conditions) into the octopole. The octopole is placed as close as possible to the
exit of the ESI source (skimmer cone region) where dispersion angles of the ions
are minimized. The flat-bottomed radial potential well for the octopole illustrates
another important advantage of the device. In the case of a wide range of m/z
ions, the flat bottom of the radial potential well allows the efficient transfer of these
ions all at the same energy, eliminating mass discrimination effects observed with
traditional dc lens systems. Another important advantage of the rf-only octopole
arises from the nonlinear fields associated with the device. These nonlinear fields
allow for the exit of a uniformly sized ion beam from the device. Therefore as ions
traverse the octopole, their frequency of oscillation is dependent on the ion entry
angle, the ion off-axis distance, and rf voltage, leading to a uniform ion beam
image independent of rf-only octopole parameters.
Also described in this section is a novel approach to the assembly of an
rf multipole device. The technique developed here at the University of Florida is
simple, inexpensive, and extremely accurate. The technique minimizes errors in
the r0 distance for the length of the device and eliminates errors associated with
fringing fields due to rod end misalignment (ends not in the same plane).
The most critical step in the process was the design and assembly of the
entire ESI/ion trap instrument. Great care was taken from the table top design to
the implementation of the copper foil grounding shield for the detector anode
lead. Two photodissociation set-ups were employed, one for the continuous wave

402
C02 laser the other for pulsed C02 laser operation. Instrumental characterization
was first performed on an El source in order to test the integrity of the
octopole/ion trap design.
Tuning and operation of the ESI/ion trap system was carried out on a
variety of biochemical compounds. The concentrations used were approximately
10 to 100 times lower than those employed by other ion trap or triple quadrupole
instruments. The absolute detection limit observed was 5 fmol/j/L for bovine
insulin. Characterization of the octopole rf amplitude confirmed what was
predicted theoretically in chapter 3, with a flat response for the various m/z range
under investigation. Other instrumental parameter investigated included the
octopole offset, the ion gate lens, ion isolation, and the rf level during the ion
injection period. The results of the ion trap rf level studies indicated that to get
a true representation of the ion signal generated for the electrospray process, the
rf level during the ion injection should be ramped so that ion injection
discrimination effects are minimized.
One important element for all ion trap experiments is that of time for
analysis or duty cycle. Since the operation of the ion trap is on the millisecond
timescale, ions have time to obtain their minimum energy configuration. This is
especially evident in the CID spectra of the triply charged ion of the peptide
angiotensin I, where rearrangement peaks driven by the presence of proline
produce internal y sequence ions (due to the highly basic nature of proline in the
gas phase).

403
Photodissociation of ionized peptides/proteins, carbohydrates, and
oligonucleotides was demonstrated as a viable alternative to traditional CID
techniques. The photoabsorption cross-section for these classes of compounds
can be ranked in the following order: carbohydrates > oligonucleotides >
peptides/proteins. Evidence for this ranking can be explained by the reduced
amount of photodissociation observed for multiply protonated peptides/proteins,
a moderate amount of photodissociation for negatively charged oligonucleotides
(5 laser pulses), and copious amounts of fragmentation for positive ions from
carbohydrates (1 to 3 laser pulses) using a pulsed C02 laser.
For the case of multiply protonated peptides and proteins,
photodissociation spectra are similar to CID spectra with the exception of the
presence of high-mass b ions in the photodissociation spectra. One plausible
explanation of this behavior may be the collisional stabilization of the product
species at helium buffer gas pressures in the low 1x1 O'5 torr range. As the
pulsed-valve comes on-line with ion injection event, more detailed and quantitative
studies can be performed in order to better understand this behavior.
In oligonucleotide analysis, the information provided by photodissociation
represents direct dissociation of the phosphodiester bridge between individual
nucleotides. To date, the largest oligonucleotide analyzed by photodissociation
in the instrument was a DNA 14-mer. Due to calibration issues and peak
resolution problems from the large scan range (the limited data acquisition ADC
steps result in reduced peak definition, i.e. less ADC steps per peak the larger the

404
mass range), all attempts to interpret the data have been unsuccessful. However,
with a digital signal processing card (DSP) and a reduced rf ramp rate for high
resolution, these problems could be overcome.
The area of greatest potential is the analysis of carbohydrate linkages using
IR photodissociation. The high photoabsorption cross-section of the COC
linkage makes this technique an ideal tool for linkage analysis. As demonstrated
with protonated erythromycin, the first steps to the identification of carbohydrate
moieties attached to biochemical molecules could be accomplished more quickly
and with less sample than other more established mass spectrometric/organic
derivatization procedures. Perhaps the most important future study in this area
would involve establishing fragmentation rules for the photodissociation process;
only when this is accomplished will the true potential of the technique be known.

REFERENCE LIST
1. Busch, K.L; Glish, G.L; McLuckey, S.A. in Mass Spectrometry/Mass
Spectrometry: Techniques and Applications of Tandem Mass Spectrometry
VCH: New York, 1988, pp 87-90.
2. Johnson, J.V.; Yost, R.A.; Kelley, P.E.; Bradford, D.C. Anal. Chem. 1990,
62, 2162-2172.
3. McLuckey, S.A.; Van Berkel, G.J.; Goeringer, D.E.; Glish, G.L. Anal. Chem.
1994, 66, 689A-696A.
4. McLuckey, S.A.; Glish, G.L.; Van Berkel, G.J. Int. J. Mass Spectrom. Ion
Processes 1991, 106, 213-235.
5. McLuckey, S.A. J. Am. Soc. Mass Spectrom. 1992, 3, 599-614.
6. Herron, W.; Goeringer, D.E.; McLuckey, S.A. in Proceedings of the 43rd
ASMS Conference on Mass Spectrometry and Allied Topics Atlanta, GA,
1995, 410.
7. Paul, W.; Steinwedel, H. Apparatus for Separating Charged Particles of
Different Specific Charges" German Patent 944,900, 1956; United States
Patent 2,939,952 1960.
8. Paul, W.; Reinhard, H.P.; Zahn, U. 2. Phys. 1958, 152, 143-182.
9. Fischer, E. Z. Phys. 1959, 156, 1-26.
10. Dehmelt, H.G. Adv. At. Mol. Phys. 1967, 3, 53-72.
11. Major, F.G.; Dehmelt, H.G. Phys. Rev. 1968, 179, 91-107.
12. Dehmelt, H.G. Adv. At. Mol. Phys. 1969, 5, 109.
13. Dawson, P.H.; Whetten, N.R. J. Vac. Sci. Technol. 1968, 5, 11-18.
14. Dawson, P.H.; Whetten, N.R. "Three-Dimensional Mass Spectrometer and
405

406
Gauge" British Patent 1,225,272, 1971.
15. Todd, J.F.J; Bonner, R.F.; Lawson, G. J. Chem. Soc. Chem. Commun.
1972, 1179-1180.
16. Lawson, G.; Bonner, R.F.; Mather, R.E.; Todd, J.F.J; March, R.E. J. Chem.
Soc., Faraday Trans. 1 1976, 73, 545-557.
17. Bonner, R.F.; Lawson, G.; Todd, J.F.J.; March, R.E. Adv. Mass Spectrom.
1974, 6, 377-384.
18. Fulford, J.E.; March, R.E. Int. J. Mass Spectrom. Ion Phys. 1978, 26, 155-
162.
19. Fulford, J.E.; Hoa, D.N.; Hughes, R.J.; March, R.E.; Bonner, R.F.; Wong,
G.J. J. Vac. Sci. Technol. 1980, 17, 829-835.
20. Hughes, R.J.; March, R.E.; Young, A.B. Int. J. Mass Spectrom. Ion Phys.
1982, 42, 255-263.
21. Hughes, R.J.; March, R.E.; Young, A.B. Can J. Chem. 1983, 61, 824-833.
22. Hughes, R.J.; March, R.E.; Young, A.B. Can. J. Chem. 1983, 61, 834-845.
23. March, R.E.; Hughes, R.J. in Quadrupole Storage Mass Spectrometry John
Wiley & Sons: New York, 1989.
24. Stafford, G.C.; Kelley, P.E.; Reynolds, W.E.; Syka, J.E.P. in Proceedings of
the 31st ASMS Conference on Mass Spectrometry and Allied Topics
Boston, MA, 1983, 48-49.
25. Stafford, G.C.; Kelley, P.E.; Bradford, D.C. Am. Lab. 1983, 51-57.
26. Stafford, G.C.; Kelley, P.E.; Syka, J.E.P.; Reynolds, W.E.; Todd, J.F.J. Int.
J. Mass Spectrom. Ion Processes 1984, 60, 85-98.
27. Kelley, P.E.; Stafford, G.C.; Syka, J.E.P.; Reynolds, W.E. Louris, J.N.; Amy,
J.W.; Todd, J.F.J. in Proceedings of the 33rd ASMS conference on Mass
Spectrometry and Allied Topics San Diego, CA, 1985, 707-708.
Kelley, P.E.; Syka, J.E.P.; Ceja, P.C.; Stafford, G.C.; Louris, J.N.;
Grutzmacher, H.F.; Kuck, D.; Todd, J.F.J. in Proceedings of the 34th ASMS
conference on Mass Spectrometry and Allied Topics Cincinnati, OH, 1986,
963-964.
28.

407
29. Louris, J.N; Amy, J.; Ridley, T.; Kascheres, C.; Cooks, R.G. in Proceedings
of the 35th ASMS Conference on Mass Spectrometry and Allied Topics
Denver, CO, 1987, 766-767.
30. Weber-Grabau, M.; Kelley, P.E.; Syka, J.E.P.; Bradshaw, S.C.; Hokeman,
D. J. in Proceedings of the 36th ASMS Conference on Mass Spectrometry
and Allied Topics San Francisco, CA, 1988, 1106-1107.
31. Louris, J.N.; Brodbelt-Lustig, J.S.; Kaiser, R.E.; Cooks, R.G. in Proceedings
of the 36th ASMS Conference on Mass Spectrometry and Allied Topics San
Francisco, CA, 1988, 968-969.
32. Duckworth, D.C.; Barshick, C.M.; Smith, D.H.; McLuckey, S.A. Anal. Chem.
1994, 66, 92-98.
33. Glish, G.L; Goeringer, D.E.; Asano, K.G.; McLuckey, S.A. Int. J. Mass
Spectrom. Ion Processes 1989, 94, 15-24.
34. Weber-Grabau, M.; Kelley, P.E.; Syka, J.E.P.; Bradshaw, S.C.; Brodbelt,
J.S. in Proceedings of the 35th ASMS Conference on Mass Spectrometry
and Allied Topics Denver, CO, 1987, 769-770.
35. Van Berkel, F.J.; Glish, G.L.; McLuckey, S.A. Anal. Chem. 1990, 62, 1284-
1295.
36. Schwartz, J.C.; Syka, J.E.P.; Jardine, I. J. Am. Soc. Mass Spectrom. 1991,
2, 198-204.
37. Williams, J.D.; Cox, K.; Morand, K.L.; Cooks, R.G.; Julian, Jr., R.K. in
Proceedings of the 39th ASMS Conference on Mass Spectrometry and
Allied Topics Nashville, TN, 1991, 1481-1482.
38. Yates, N.A.; Griffin, T.P.; Yost, R.A. in Proceedings of the 41st ASMS
Conference on Mass Spectrometry and Allied Topics San Francisco, CA,
1992, 444a-444b.
39. Soni, M.H.; Cooks, R.G. Anal. Chem. 1994, 66, 2488-2496.
40. Todd, J.F.J; Penmann, A.D.; Thorner, D.A.; Smith, R.D. in Proceedings of
the 38th ASMS Conference on Mass Spectrometry and Allied Topics
Tuscon, AZ, 1990, 532-533.
41. Johnson, J.V.; Pedder, R.E.; Yost, R.A. Int. J. Mass Spectrom. Ion
Processes 1991, 106, 197-212.

408
42. Kaia, S.M.S.; Andre, J.; Zerega, Y.; Brincourt, G.; Catella, R. Int. J. Mass
Spectrom. Ion Processes 1991, 107, 191-203.
43. Kelley, P.E.; Hokeman, D.J.; Bradshaw, S.C.; Jones, L.B.; Louris, J.N.
Wilson, R.W. in Proceedings of the 42nd ASMS Conference on Mass
Spectrometry and Allied Topics Chicago, IL, 1994, 709.
44. Griffiths, I.W. Rapid Comm. Mass Spectrom. 1990, 4, 69-73.
45. Todd, J.F.J. Mass Spectrom. Reviews 1991, 10, 3-52.
46. March, R.E. Int. J. Mass Spectrom. Ion Processes 1992, 118/119, 71-135.
47. McCloskey, J.A., ed. in Methods in Enzymology, Volume 193, Mass
Spectrometry Academic Press: San Diego, 1990.
48. Matsuo, T., ed. Biological Mass Spectrometry: Present and Future Wiley:
New York, 1994.
49. Desiderio, D.M., ed. Mass Spectrometry of Peptides CRC Press: Boca
Raton, FL, 1991.
50. McEwen, C.N.; Larsen, B.S., eds. Mass Spectrometry of Biological Materials
Marcel Decker: New York, 1990.
51. Mathieu, J. Math Purs Appi (J. Liouville), 1868, 13, 137.
52. Knight, R.D. Int. J. Mass Spectrom. Ion Phys. 1983, 51, 127-131.
53. Louris, J.; Stafford, g.; Syka, J.; Taylor, D. in Proceedings of the 40th
Conference on Mass Spectrometry and Allied Topics Washington, DC,
1992, 1003.
54. Johnson, J.V.; Pedder, R.E.; Yost, R.A. Rapid Comm. Mass Spectrom.
1992, 6, 760-764.
55. McLachlan, N.W. in Theory and Applications of Mathieu Functions
Claredon: Oxford, 1947.
56. Paul, W. "Historical Development of the Quadrupole Ion Trap" presented at
the 40th ASMS Conference on Mass Spectrometry and Allied Topics
Washington, D.C., May 31 June 5, 1995.
57. March, R.E.; McMahon, A.W.; Loundry, F.A.; Alfred, R.L.; Todd, J.F.J.;

409
Vedel, F. Int. J. Mass Spectrom. Ion Processes 1989, 95, 119-156.
58. Haddon, W.F.; McLafferty, F.W. J. Am. Chem. Soc. 1968, 90, 4745.
59. Jennings, K.R. Int. J. Mass Spectrom. Ion Phys. 1968, 1, 227.
60. Rosenstock, H.M.; Melton, C.E. J. Chem. Phys. 1957, 26, 314.
61. Kupriyanov, S.E.; Perov, A.A. Russ. J. Phys. Chem. 1965, 39, 871.
62. Cooks, R.G.; Terwilliger, D.T.; Ast, T.; Beynon, J.H.; Keough, T.J. J. Am.
Chem Soc. 1975, 97, 1583.
63. Cooks, R.G.; Ast, T.; Beynon, J.H. Int. J. Mass Spectrom. Ion Phys. 1975,
16, 348.
64. Cody, R.B.; Freiser, B.S. Anal. Chem. 1979, 51, 547.
65. Freiser, B.S. Int. J. Mass Spectrom. Ion Phys. 1978, 26, 79.
66. Fedor, D.W.; Cody, R.B.; Burinsky, D.J.; Freiser, B.S.; Cooks, R.G. Int. J.
Mass Spectrom. Ion Phys. 1979, 31, 27.
67. Dunbar, R.C. in Gas Phase Ion Chemistry, Vol. 2 Bowers, M.T., ed.;
Academic Press: London, 1979, pp 182-217.
68. Dunbar, R.C. in Gas Phase Ion Chemistry, Vol. 3 Bowers, M.T., ed.;
Academic Press: London, 1984, pp 130-164.
69. Dunbar, R.C. in Molecular Ions: Spectroscopy Structure and Chemistry
Miller, T.A., ed.; North-Holland: Amsterdam, 1983.
70. Goeringer, D.E.; McLuckey, S.A. J. Chem. Phys. 1995, in press.
71. Morgan, R.P.; Brenton, A.G.; Beynon, J.H. Int. J. Mass Spectrom. Ion Phys.
1979, 29, 195.
72. Futrell, J.H. in Gaseous Ion Chemistry and Mass Spectrometry Futrell, J.H.,
ed.; John Wiley & Sons: New York, 1986.
73. Kaiser, Jr., R.E.; Cooks, R.G.; Stafford, Jr. G.C.; Syka, J.E.P.; Hemberger,
P.H. Int. J. Mass spectrom. Ion Processes, 1991, 106, 79-115.
74. Schwartz, J.C.; Jardine, I. Rapid Comm. Mass Spectrom. 1992, 6, 313-317.

410
75. Louris, J.N.; Cooks, R.G.; Syka, J.E.P.; Kelley, P.E.; Stafford, G.C.; Todd,
J.F.J. Anal. Chem. 1987, 59, 1677-1685.
76. Todd, J.F.J.; Bexon, J.J.; Smith, R.D.; Weber-Grabau, M.; Kelley, P.E.;
Syka, J.E.P.; Stafford, G.C.; Bradshaw, S.C. in Proceedings of the 16th
Meeting of the British Mass Spectrometry Society York, U.K., 1987, 206-
209.
77. Tucker, D.B.; Hamiester, C.H.; Bradshaw, S.C.; Hokeman, D.J.; Weber-
Grabau, M. in Proceedings of the 36th ASMS Conference on Mass
Spectrometry and Allied Topics San Francisco, CA, 1988, 662-663.
78. Gronowska, I.W.; Paradisi, C.; Traldi, P.; Vettori, U. Rapid Comm. Mass
Spectrom., 1990, 4, 306-313.
79. Yates, N.A.; Yost, R.A.; Bradshaw, S.C.; Tucker, D.B. in Proceedings of the
39th ASMS Conference on Mass Spectrometry and Allied Topics Nashville,
TN, 1991, 132-133.
80. Eades, D.M.; Yates, N.A.; Yost, R.A. in Proceedings of the 39th ASMS
Conference on Mass Spectrometry and Allied Topics Nashville, TN, 1991,
1491-1492.
81. Kelley, P.E.; Hoekman, D.J.; Bradshaw, S.C. in Proceedings of the 41st
ASMS Conference on Mass Spectrometry and Allied Topics San Francisco,
CA, 1993, 453a-453b.
82. McLuckey, S.A.; Goeringer, D.E.; Glish, G.L. Anal. Chem. 1992, 64, 1455-
1460.
83. Yates, N.A. ICMS Software, Department of Chemistry, University of Florida,
1993.
84. Smith, R.D.; Loo, J.A.; Ogorzalek, L; Busman, M.; Udseth, H.R. Mass
Spectrom. Reviews 1991, 10, 359-451.
85. Dole, M.; Mack, L.L; Hines, R.L.; Mobley, R.C.; Ferguson, L.D.; Alice, M.B.
J. Chem. Phys. 1968, 49, 2240-2249.
86. Mack, L.L; Kralik, P.; Rheude, A.; Dole, M. J. Chem. Phys. 1970, 52, 4977-
4986.
87. Gieniec, J.; Mack, L.L; Nakamae, K.; Gupta, C.; Kumar, V.; Dole, M.
Biomed. Mass Spectrom. 1984, 11, 259-268.

411
88. Yamashita, M.; Fenn, J.B. J. Phys. Chem. 1984, 88, 4451-4459.
89. Aleksandrov, M.L; Gall, L.N.; Krasnov, V.N.; Nikolaev, V.I.; Pavlenko, V.A.;
Shkurov, V.A. Dokl. Akad. Nauk. SSSR 1984, 277, 379-383.
90. Aleksandrov, M.L; Gall, L.N.; Krasnov, V.N.; Nikolaev, V.I.; Pavlenko, V.A;
Shkurov, V.A.; Baram, G.I.; Gracher, M.A.; Knorre, V.D.; Kusner, Y.S.
Bioorg. Kim. 1985, 11, 705-708.
91. Meng, C.K.; Mann, M.; Fenn, J.B. Phys. D. 1988, 10, 361-368.
92. Loo, J.A.; Udseth, H.R.; Smith, R.D. Biomed. Environ. Mass Spectrom.
1988, 17, 411-414.
93. Covey, T.R.; Bonner, R.F.; Shushan, B.I.; Henion, J.D. Rapid Comm. Mass
Spectrom. 1988, 2, 249-256.
94. Cox, A.L; Skipper, J.; Chen, Y.; Henderson, R.A.; Darrow, T.L;
Shabanowitz, J.; Englehard, V.H.; Hunt, D.F.; Slingluff, C.L Science 1994,
264, 716-719.
95. Hunt, D.F.; Michel, H.; Dickerson, T.A.; Shabanowitz, J.; Cox, A.L;
Sakaguchi, K.; Appella, E.; Grey, H.M.; Sette, A. Science 1992, 256, 1817-
1820.
96. Hunt, D.F.; Henderson, R.A.; Shabanowitz, J.; Sakaguchi, K.; Michel, H.;
Sevilir, N.; Cox, A.L; Appella, E.; Englehard, V.H. Science 1992,255,1261-
1263.
97. Henderson, R.A.; Michel, H.; Sakaguchi, K.; Shabanowitz, J.; Appella, E.;
Hunt, D.F.; Englehard, V.H. Science 1992, 255, 1264-1266.
98. Ikonomou, M.G.; Blades, A.T. Kebarle, P.AnaJ. Chem. 1991,63,1989-1998.
99. Kebarle, P.; Tang, L. Anal. Chem. 1993, 65, 972A-986A.
100. Fenn, J.B. J. Am. Soc. Mass Spectrom. 1993, 4, 524-535.
101. Chorush, P.A.; Little, D.D.; Beu, S.C.; Wood, T.D.; McLafferty, F.W. Anal.
Chem. 1995, 67, 1042-1046.
102. Jones, J.C.; Dongue, A.R.; Sonogyi, A.; Wysocki, V.H. J. Am. Chem. Soc.
1994, 116, 8368-8369.

412
103. Hemling, M.E.; Conboy, J.J.; Bean, M.F.; Mentzer, M.; Carr, S.A. J. Am.
Soc. Mass Spectrom. 1994, 5, 434-442.
104. Ganem, B.; Li, Y.T.; Henion, J.D. J.Am. Chem. Soc. 1991, 113, 6294-6296.
105. Light-Wahl, K.J.; Schwartz, B.L.; Smith, R.D. J.Am. Chem. Soc. 1994, 116,
5271-5278.
106. Busman, M.; Rockwood, A.L; Smith, R.D. J. Phys. Chem. 1992, 96, 2397-
2400.
107. Senko, M.W.; Speir, J.P.; McLaferty, F.W. Anal. Chem. 1994, 66, 2801-
2808.
108. Ishikawa, K.; Nishimura, T.; Koga, Y.; Niwa, Y. Rapid Comm. Mass
Spectrom. 1994, 8, 933-938.
109. Kilby, G.W.; Sheil, M.M. Org. Mass Spectrom. 1993, 28, 1417-1423.
110. Tang, X.J.; Thibault, P.; Boyd, R.K. Anal. Chem. 1993, 65, 2824-2834.
111. Little, D.P.; Speir, P.J.; Senko, M.W.; OConnor, P.B.; McLafferty, F.W. Anal.
Chem. 1994, 66, 2809-2815.
112. Stephenson, Jr., J.L.; Booth, M.M.; Boue, S.M.; Eyler, J.R.; Yost, R.A. in
Biological and Biotechnical Applications of ESI-MS American Chemical
Society: Washington, D.C. 1995 in press.
113. Siuzdak, G. Proc. Natl. Acad. Sci. 1994, 91. 11290-11297.
114. Senko, M.W.; McLafferty, F.W. Annu. Rev. Biophys. Biomol. Struct. 1994,
23, 763-785.
115. Smith, R.D.; Loo, J.A.; Edmonds, C.G.; Barinaga, C.J.; Udseth, H.R. Anal.
Chem. 1990, 62, 882-899.
116. Rayleigh, J.W.S. S. Philos. Mag. 1882, 14, 184.
117. Muller, E.W. Phys. Rev. 1956, 102, 618-624.
118. Iribarne, J.V.; Thomson, B.A. J. Chem. Phys. 1976, 64, 2287-2284.
Thomson, B.A.; Iribarne, J.V. J. Chem. Phys. 1979, 71, 4451-5563.
119.

413
120. Fenn, J.B. Anal. Chem. 1985, 57, 675-679.
121. Loo, J.A.; Udseth, H.R.; Smith, R.D. Anal. Biochem. 1989, 179, 404-412.
122. Guevremont, R.; Siu, K.W.M.; LeBlanc, J.C.Y.; Berman, S.S. J. Am. Soc.
Mass Spectrom. 1990, 1,
123. Mann, M.; Meng, C.K.; Fenn, J.B. Anal. Chem. 1989, 61, 1702-1708.
124. Beu, S.C.; Senko, M.W.; Quinn, J.P.; Wampler, F.M. Ill; McLafferty, F.W. J.
Am. Soc. Mass Spectrom. 1993, 4, 557-565.
125. Hagen, J.J; Monning, C.A. Anal. Chem. 1994, 66, 1877-1883.
126. van der Hart, W.J. Mass spectrom. Reviews 1989, 8, 237-268.
127. van der Hart, W.J. Int. J. Mass Spectrom. Ion Processes 1991, 118/119,
617-633.
128. Dunbar, R.C.; Weddle, G.H. J. Phys. Chem. 1988, 92, 5706.
129. Bomse, D.S.; Woodin, R.L.; Beauchamp, J.L. J. Am. Chem. Soc. 1979,101,
5503-5512.
130. Thorne, L.R.; Beauchamp, J.L. in Gas Phase Ion Chemistry, Vol. 3: Ions
and Light Bowers, M.T., ed.; Academic Press: London, 1984, pp 42-95.
131. Rosenfeld, R.N.; Jasinski, J.M.; Brauman, J.l.J.Am. Chem. Soc. 1979,101,
5503-5512.
132. Turnas, W.; Foster, R.F.; Brauman, J.l. Isr. J. Chem. 1984, 24, 223-231.
133. Drzaic, P.S.; Marks, J.; Brauman, J.l. in Gas Phase Ion Chemistry, Vol. 3:
Ions and Light Bowers, M.T., ed.; Academic Press: London, 1984, pp 168-
207.
134. Mead, R.D.; Lineberger, W.C.; in Gas Phase ion Chemistry, Vol. 3: Ions
and Light: Bowers, M.T., ed.; Academic Press: London, 1984, pp 214-246.
135. March, R.E.; Kamar, A.; Young, A.B. in Advances in Mass Spectrometry
1985: Proc. 10th Int. Mass Spectrom. Confer. Wales, 1986, 949-950.
136. Kamar, A. Application of the Quadrupole Ion Storage Trap (QUISTOR) to the
study of Gas Phase lon/Molecule Reactions Ph.D. Dissertation, Queens

414
Univ., Kingston, Ont., Canada, 1985.
137. March, R.E.; Young, A.B.; Hughes, R.J.; Kamar, A.; Baril, M. Spectrosc. Int.
J. 1984, 3, 17-32.
138. Kamar, A.; Young, A.B.; March, R.E. Can J. Chem. 1986, 64, 1979-1988.
139. Young, A.B.; March, R.E.; Hughes, R.J. Can. J. Chem. 1985, 63, 2324-
2331.
140. March, R.E.; Hughes, R.J.; Young, A.B. in Proc. 13th Meeting Brit. Mass
Spectrom. Soc. Warwick, U.K., 1983, 77-79.
141. Isenor, N.R.; Richardson, M.C. Appl. Phys. Lett. 1971, 18, 224.
142. Letokhov, V.S.; Ryabov, E.A. Tumanov, O.A. Optics Commun. 1972,5,168.
143. Isenor, N.R.; Merchant, V.; Hallsworth, R.S.; Richardson, M.C. Can. J. Phys.
1973, 51, 1281.
144. Black, J.G.; Yablonovitch, N.; Bloembergen and Mukamel, S. Phys. Rev.
Lett. 1977, 38, 131.
145. Ashfold, N.R.; Hancock, G. in Gas Kinetics and Energy Transfer
Spottiswoode Ballantyne Ltd.: London, 1981, 73-116.
146. Stephenson Jr.; Booth, M.M.; Shalosky, J.A.; Eyler, J.R.; Yost, R.A. in
Proceedings of the 41st ASMS Conference on Mass Spectrometry and
Allied Topics San Francisco, CA, 1993, 445a-445b.
147. Stephenson Jr., J.L.; Booth, M.M.; Shalosky, J.A.; Eyler, J.R.; Yost, R.A. J.
Am. Soc. Mass Spectrom. 1994, 5, 886-893.
148. Stephenson Jr., J.L.; Booth, M.M.; Eyler, J.R.; Yost, R.A. in Proceedings of
the 42nd ASMS Conference on Mass Spectrometry and Allied Topics
Chicago, IL, 1994, 32.
149. Watson, C.H.; Zimmerman, J.A.; Bruce, J.E.; Eyler, J.R. J. Phys. Chem.
1991, 95, 6081-6086.
150. Yates, N.A.; Yost, R.A. in Proceedings of the 39th ASMS Conference on
Mass Spectrometry and Allied Topics Nashville, TN, 1991, 1489-1490.
151. Quigley, G.P. in Chemical Physics 6: Laser Induced-Processes in

415
Molecules Kompa, K.L.; Smith, S.D., eds.; Springer-Vertag: New York,
1979, 134-137.
152. Quigley, G.P. in Chemical Physics 3, Advances in Laser Chemistry Zewail,
A.H., ed.; Springer-Verlag: New York, 1978, 374-383.
153. Wu, H.F.; Brodbelt, J.S. Int J. Mass Spectrom. Ion Processes 1992, 115,
67-81.
154. Reiser, H.-P, Kaiser, R.E.; Savickas, P.J.; Cooks, R.G. Int. J. Mass
Spectrom Ion Processes 1991, 106, 237.
155. White, J.U. J. Opt. Soc. Am. 1942, 32, 285-288.
156. March, R.E.; McMahon, A.W. Allinson, E.T.; Londry, F.A.; Alfred, R.L; Todd,
J.F.J.; Vedel, F. Int. J. Mass Spectrom. Ion Processes 1990, 99, 109-124.
157. March, R.E.; Londry, F.A.; Alfred, R.L.; Todd, J.F.J.; Penman, A.D.; Vedel,
F.; Vedel, M. Int. J. Mass Spectrom. Ion Processes 1991, 110, 159-178.
158. Alfred, R.L.; Londry, F.A.; March, R.E. Int. J. Mass Spectrom. Ion Processes
1993, 125, 171-185.
159. Hemberger, P.H.; Nogar, N.S.; Williams, J.D.; Cooks, R.G.; Syka, J.E.P.
Chem. Phys. Lett. 1992, 191, 405-410.
160. Williams, J.D.; Cooks, R.G.; Syka, J.E.P.; Hemberger, P.H.; Nogar, N.S. J.
Am. Soc. Mass Spectrom. 1993, 4, 792-292.
161. Wang, Y. Fundamental Study of Magntica! and Electrical Ion Traps Ph.D.
Dissertation, Bremen, Germany, 1992.
162. Eades, D.M. Higher Order Field Effects on Stored Ions in a Quadrupole Ion
Trap, Ph.D. Dissertation, University of Florida, Gainesville, FL, 1994.
163. Reiser, H-P; Julian Jr., R.K.; Cooks, R.G. Int. J. Mass spectrom. Ion
Processes 1992, 121, 49-63.
164. Eades, D.M.; Johnson, J.V.; Yost, R.A. in Proceedings of the 41st ASMS
Conference on Mass Spectrometry and Allied Topics San Francisco, CA,
1993, 693a-693b.
165. Williams, J.D.; Reiser, H-P.; Kaiser, R.E.; Cooks, R.G. Int. J. Mass
Spectrom. Ion Processes 1991, 108, 199-219.

416
166. Vedel, F.; Vedel, M.; March, R.E. Int. J. Mass Spectrom. Ion Processes
1990, 99, 125-138.
167. Stephenson Jr., J.L.; Booth, M.M.; Eyler, J.R.; Yost, R.A. submitted to R.
Comm. Mass Spectrom.
168. Vedel, F. Int. J. Mass Spectrom. Ion Processes 1991, 106, 33-61.
169. Vedel, F.; Vedel, M.; March, R.E. Int. J. Mass Spectrom. Ion Processes
1991, 108, R11-R20.
170. Lias, S.G.; Bartmess, J.E.; Liebman, J.F.; Holmes, J.L.; Levin, R.D.; Mallard,
W.G. J. Phys. Chem. Ref. Data 1988, 17, Suppl. 1, 74, 90, 274.
171. Dewar, M.J.S.; Zoebisch, E.G.; Healy, E.F.; Stewart, J.J.P. J. Am. Chem.
Soc. 1990, 107, 3902.
172. Coaxian, J.; Stephenson Jr., J.L. AM1 Calculations, University of Florida.
173. Katritzky, A.R.; Watson, C.H.; Dega-Szafran, Z.; Eyler, J.R. J. Am. Chem.
Soc. 1990, 112, 2476.
174. Lias, S.G.; Bartmess, J.E.; Liebman, J.F.; Holmes, J.L.; Levin, R.D.; Mallard,
W.G. J. Phys. Chem. Ref. Data 1988, 17, Suppl. 1, 52, 114.
175. Baykut, G.; Watson, C.H.; Weller, R.R.; Eyler, J.R. J. Am. Chem. Soc. 1985,
107, 8036-8042.
176. Harcourt, A.V.; Esson, W. Proc. R. Soc. London 1865, 14, 470.
177. Harcourt, A.V.; Esson, W. Philos. Trans. 1866, 156, 193.
178. Harcourt, A.V.; Esson, W. Philos. Trans. 1867, 157, 117.
179. Capellos, C.; Bielski, B.H.J. Kinetic Systems: Mathematical Descriptions of
Chemical Kinetics in Solution Wiley-lnterscience: New York, 1972, 46-49.
180. Laidler, K.J. in Chemical Kinetics Harper & Row: New York, 1987, 278-279.
181. Moore, J.W.; Pearson, R.G. in Kinetics and Mechanism John Wiley & Sons:
New York, 1981, 290-296.
182. Woodin, R.L.; Bomse, D.A.; Beauchamp, J.L. Chem. Phys. Lett. 1979, 63,
630-636.

417
183. Jasinski, J.M.; Rosenfeld, R.N.; Meyer, F.K. Brauman, J.l. J. Am. Chem.
Soc. 1982, 104, 652-658.
184. Rosenfeld, R.N.; Jasinski, J.M.; Brauman, J.l. J. Am. Chem. Soc. 1982,104,
659-663.
185. Watson, C. H. Infrared Multiple Photon Dissociation of Gaseous Ions Studied
by Fourier Transform Ion Cyclotron Resonance Mass Spectrometry Ph.D.
Dissertation, University of Florida, Gainesville, FL, 1986.
186. Ibuki, T.; Sugita, N. J. Chem. Phys. 1983, 79, 5392-5395.
187. Dobler, W.; Ramler, H.; Vinninger, H.; Howorka, F.; Undinger, W. Chem.
Phys. Lett. 1983, 97, 553-556.
188. Freguson, E.E.; Adams, N.G.; Simth, D.; Alge, E.J. J. Chem. Phys. 1984,
80, 6095-6098.
189. Peiris, D.L; Cheeseman, M.A.; Ramanathan, R.; Eyler, J.R. J. Phys. Chem.
1993, 97, 7839-7843.
190. Basic, C.; Eyler, J.R.; Yost, R.A. J. Am. Soc. Mass Spectrom. 1992, 3, 716-
726.
191. Bruce, J.E.; Eyler, J.R. J. Am. Soc. Mass Spectrom. 1992, 3, 727-733.
192. Louris, J.N.; Amy, J.W.; Ridley, T.Y.; Cooks, R.G. Int. J. Mass Spectrom. Ion
Processes 1989, 88, 97-111.
193. Pedder, R.E.; Yost, R.A.; Weber-Grabau, M. in Proceedings of the 37th
ASMS Conference on Mass Spectrometry and Allied Topics Miami, FL,
1989, 468-469.
194. Schwartz, J.C.; Cooks, R.G. in Proceedings of the 36th ASMS Conference
on Mass Spectrometry and Allied Topics San Francisco, CA, 1988, 634-
635.
195. Soni, M.; Cooks, R.G. Anal. Chem. 1994, 66, 2488-2496.
196. Bier, M.E.; Hartford, R.E.; Herron, J.R.; Stafford, G.C. in Proceedings of the
39th ASMS Conference on Mass Spectrometry and Allied Topics Nashville,
TN, 1991, 538-539.
197. Kaiser, R.E.; Williams, J.D.; Schwartz, J.C.; Lammert, S.A.; Cooks, R.G. in

418
Proceedings of the 37th ASMS Conference on Mass Spectrometry and
Allied Topics Miami, FL, 1989, 369-370.
198. Mordehai, A.V.; Hopfgartner, G.; Huggins, T.G.; Henion, J.D. Rapid
Commun. Mass Spectrom. 1992, 6, 508-516.
199. Bou, S.M.; Stephenson, J.L.; Yost, R.A. in Proceedings of the 43rd ASMS
Conference on Mass Spectrometry and Allied Topics Atlanta, GA, 1995,
408.
200. Doktycz, M.J.; Habibigoudarz, S.; McLuckey, S.A. Anal. Chem. 1994, 66,
3416-3422.
201. Bou, S.M.; Jones, J.A.; Yost, R.A. in Proceedings of the 42nd ASMS
Conference on Mass Spectrometry and Allied Topics Chicago, IL, 1994,
218.
202. Stephenson Jr., J.L.; Booth, M.M.; Shalosky, J.A.; Eyler, J.R.; Yost, R.A. Int.
J. Mass Spectrom. Ion Processes 1995, submitted.
203. Asano, K.G.; Glish, G.L.; McLuckey, S.A. in Proceedings of the 36th ASMS
Conference on Mass Spectrometry and Allied Topics San Francisco, CA,
1988, 636-637.
204. McLuckey, S.A.; Glish, G.L.; Asano, K.G. Anal. Chem. Acta. 1989, 225, 25-
35.
205. Duckworth, D.C.; Barshick, C.M.; Simth, D.H.; McLuckey, S.A. Anal. Chem.
1994, 66, 92-98.
206. McLuckey, S.A.; Glish, G.L.; Van Berkel, G.J. in Proceedings of the 38th
ASMS Conference on Mass Spectrometry and Allied Topics Tuscon, AZ,
1990, 512-513.
207. Chien, B.M.; Michael, S.M.; Lubman, D.M. Anal. Chem. 1993, 65, 1916-
1924.
208. Koppenaal, D.W.; Barinaga, C.J.; Smith, M.R. J. Anal. Atomic Spectrom.
1994, 9, 1053-1058.
209. Louris, J.N.; Brodbelt-Lustig, J.S.; Kaiser, R.E.; Cooks, R.G. in Proceedings
of the 36th ASMS Conference on Mass Spectrometry and Allied Topics San
Francisco, CA, 1988, 968-969.

419
210. Eiden, G.G.; Garrett, A.W.; Cisper, M.E.; Nogar, N.S.; Hemberger, P.H. Int.
J. Mass Spectrom. Ion Processes 1994, 136, 119-141.
211. Booth, M.M.; Stephenson Jr., J.L: Yost, R.A. in Proceedings of the 42nd
ASMS Conference on Mass Spectrometry and Allied Topics Chicago, IL,
1994, 693.
212. Yang, M.; Dale, J.M.; Whitten, W.B. Anal. Chem. 1995, 67, 1021-1025.
213. Dale, J.M.; Yang, M.; Whiten, W.B. Anal. Chem. 1994, 66, 3431-3435.
214. Pinkston, J.D.; Delaney, T.E.; Morand, K.L Anal. Chem. 1992, 64, 1571-
1577.
215. Mikami, N.; Sato, S.; Ishigaki, M.; Toshiki, S. Chem. Phys. Lett. 1990, 166,
470-474.
216. Mikami, N.; Sato, S.; Ishigaki, M. Chem. Phys. Lett. 1991, 180, 431-435.
217. Yost, R.A.; Enke, C.G. Anal. Chem. 1979, 51, 1251A.
218. Teloy, E.; Gerlich, D. Chem. Phys. 1974, 4, 417-427.
219. Szabo, I. Int. J. Mass Spectrom. Ion Processes 1986, 73, 197-235.
220. Guettler, R.D.; Jones, G.C.; Posey, L.A.; Kirchner, N.J.; Keller, B.A. Zare,
R.N. J. Chem. Phys. 1994, 101, 3763-3771.
221. Ervin, K.M.; Amnentrout, P.B. J. Chem. Phys. 1985, 83, 166-189.
222. Syka, J.E.P. Szabo, I. in Proceedings of the 36th ASMS Conference on
Mass Spectrometry and Allied Topics San Francisco, CA, 1988,1328-1329.
223. Hail, M.; Mylchreest, I. in Proceedings of the 41st ASMS Conference on
Mass Spectrometry and Allied Topics San Francisco, CA, 1993, 745a.
224. Shao, J.D.; NG, C.Y. Chem. Phys. Lett. 1985, 118, 481-485.
225. Gerlich, D.; Disch, R.; Scherbarth, S. J. Chem. Phys. 1987, 87, 350-359.
226. Posey, L.A.; Guettler, R.D.; Kirchner, N.J.; Zare, R.N. J. Chem. Phys. 1994,
101, 3772-3786.
227. Anderson, S.L.; Houle, F.A.; Gerlich, D.; Lee, Y.T. J. Chem. Phys. 1981, 75,

420
2153-2162.
228. Bier, M.E.; Schwartz, J.C.; Zhou, J.; Taylor, D.; Syka, J.E.P.; James, M.;
Fies, W.; Stafford, G. in Proceedings of the 43rd ASMS Conference on
Mass Spectrometry and Allied Topics Atlanta, GA, 1995, 1117.
229. Miller, P.E.; Denton, B.M. Int. J. Mass Spectrom. Ion Processes 1986, 72,
223-238
230. Everdij, J.J; Huijser, A.; Verster, N.F. Rev. Sci. Instrum. 1973, 44, 721-725.
231. Hgg, A.; Szabo, I. IntJ. Mass Spectrom. Ion Processes 1986, 73, 237-275.
232. Friedman, M.H.; Yergy, A.L; Campana, J.E. Rev. Sci. Instrum. 1982, 15,
53-61.
233. Pugh, E.M.; Pugh E.W. in Principles of Electricity and Magnetism Addison-
Wesley: Reading, MA, 1960.
234. Ramo, S.; Whinnery, J.R.; Van Duzer, T. in Fields and Waves in
Communication Electronics Wiley: New York, 1984.
235. Campana, J.E. Int. J. Mass Spectrom. Ion Processes 1980, 33, 101-117.
236. Campana, J.E.; Jurs, P.C. Int. J. Mass Spectrom. Ion Processes 1980, 33,
119-137.
237. Hgg, C.; Szabo, I. Int. J. Mass Spectrom. Ion Processes 1986, 73, 277-
294.
238. Hgg, C.; Szabo, I. Int. J. Mass Spectrom. Ion Processes 1986, 73, 295-
312.
239. Ball, G.W.; Lawson, G.; Todd, J.F.J. in Dynamic Mass Spectrometry vol. 3
Price, D., ed.; Heydon: London, 1972, p. 99.
240. Hlein, W. U.S. Pat. 3,238,146 1967.
241. Dawson, P.H., ed. Quadrupole Mass Spectrometry and its Applications
Elsevier: Amsterdam, 1976.
242. von Zahn, U. Rev. Sci. Instrum. 1963, 34, 1.

421
243. von Zahn, U. "Method and Apparatus for Separating Ions of Respectively
Different Specific Ionic Charges" U.S.Pat. 3,197,633, 1965.
244. Gnther, K.G.; Freiler, H. "The Electric Mass Filter" U.S. Pat. 3,105,899,
1963.
245. Brubaker, W.M. "Auxiliary Electrodes for Quadrupole Mass Filters" U.S. Pat.
3,129,327, 1964.
246. Uthe, P.M. "Quadrupole Mass Analyzer" U.S. Pat. 3,457,404, 1969.
247. Uthe, P.M. "Quadrupole in which the Pole Electrodes Comprise Metallic
Rods Whose Mounting Surfaces Coincide with Those of the Mounting
Means" U.S. Pat. 3,553,451, 1971.
248. McGinnis P.F. "Mass Filter Electrode" U.S. Pat. 3,699,330, 1972.
249. Turner, W.R. "Mass Filter with Artifact Reducing Electrode Structure" U.S.
Pat. 3,725,700, 1973.
250. Dayton, I.F.; Shoemaker, F.C.; Mozley, R.F. Rev. Sci. Instrum. 1954, 25,
485.
251. Denison, D.R. J. Vac. Sci. Technol. 1971, 8, 266.
252. Dahl, D.A. in Proceedings of the 43rd ASMS Conference on Mass
Spectrometry and Allied Topics Atlanta, GA, 1995, 717.
253. Davis, S.C.; Wright, B. Rapid Comm. Mass Spectrom. 1990, 4, 186-197.
254. Alexander, A.J.; Dyer, E.W.; Boyd, R.K. Rapid Comm. Mass Spectrom.
1989, 3, 364-372.
255. Schoen, A.E.; Syka, J.E.P. in Proceedings of the 34th ASMS Conference
on Mass Spectrometry and Allied Topics Cincinnati, OH, 1986, 722-723.
256. Pedder, R.E. Fundamental Studies in the Quadrupole Ion Trap Mass
Spectrometer, Ph.D. Dissertation, University of Florida, Gainesville, FL,
1994.
257. Douglas, D.J.; French, J.B. J. Am. Soc. Mass Spectrom. 1992, 3, 398-408.
258. Fies, W.J. "The Standing Wave Ratio Method for Tuning RF Resonant
Circuits" internal report, Finnigan MAT corporation, 1985.

422
259. Kaiser, R.E.; Cooks, R.G.; Moss, J.; Hemberger, P.H. RapidCommun. Mass
Spectrom. 1989, 3, 50-53.
260. Kaiser, R.E.; Cooks, R.G.; Stafford, G.C.; Syka, J.E.P.; Hemberger, P.H. Int.
J. Mass Spectrom. Ion Processes 1991, 106, 79-115.
261. Booth, M. M.; Stephenson, Jr. J. L; Johnson, J. V.; Yost, R. A.
Fundamental and Practical Considerations of Mass Range Extension in a
Quadrupole Ion Trap Mass Spectrometer Presented to the Florida Section
of the American Chemical Society, April 3 May 2, 1992, Coco Beach, FL.
262. Chowdhury, S.K.; Katta, V.; Chait, B.T. J.Am. Chem. Soc. 1990, 112, 9012-
9013.
263. Loo, J.A.; Ogorazlek Loo, R.R.; Udseth, H.R.; Edmonds, C.G.; Smith, R.D.
Rapid Commun. Mass Spectrom. 1991, 5, 101-105.
264. McDonald, C.C.; Phillips, W.D.; Glickson, J.D. J. Am. Chem. Soc. 1971,93,
235-246.
265. Creighton, T.E. in Proteins, Structures and Molecular Principles W.H.
Freeman: New York, 1984.
266. Tanford, C. in Advances in Protein Chemistry Anfinsen, C.G.; Anson, M.L.;
Edsall, J.T.; Richards, R.M., eds.; Academic Press: New York, 1968, vol.
23, p.121.
267. Timasheff, S.N. Acc. Chem. Res. 1970, 3, 62.
268. Franks, R.; England, D. CRC Critical Rev. Biochem. 1975, 3, 165.
269. Herskovits, T.T.; Gadgefbeku, B.; Jaillet, H.T. J. Biol. Chem. 1970, 245,
2599.
270. Wilkinson, K.D.; Mayer, A.N. Arch. Biochem. Biophys. 1986, 250, 390-399.
271. Kishore, M.N.; Ghosh, P.K. Int. J. Mass spectrom. Ion Phys. 1979, 29, 345-
350.
272. Rajagopal, C.; Ghosh, P.K. Int. J. Mass Spectrom. Ion Phys. 1981, 41, 1-6.
273. Kaiser, R.E.; Cooks, R.G.; Syka, J.E.P.; Stafford, G.C. Rapid Commun.
Mass Spectrom. 1990, 4, 30.

423
274. Cox, K.A.; Williams, J.D.; Kaiser, R.E.; Cooks, R.G. Biol Mass Spectrom.
1992, 21, 226-241.
275. Huddleston, M.J.; Beam, M.F.; Carr, S.A. Anal. Chem. 1993, 65, 877-884.
276. Johnson, R.S.; Martin, S.A.; Bieman, K.; Stults, J.T.; Watson, J.T. Anal.
Chem. 1987, 59, 2621-2625.
277. Johnson, R.S.; Martin, S.A. Bieman, K. Int. J. Mass Spectrom. Ion
Processes 1988, 86, 137-154.
278. Baldwin, M.A. Nat. Prod. Rep., 1995, 12, 33-44.
279. Reopstorff, P.; Fohlman, J. Biomed. and Environ. Mass Spectrom. 1984,11,
601.
280. Hunt, D.F.; Yates, J.R. Ill; Shabonowitz, J.; Winston, S.; Hauer, C.R. Proc.
Natl. Acad. Sci. U.S.A. 1986, 83, 6233-6237.
281. Bieman, K.; Martin, S.A. Mass Spectrom. Rev. 1987, 6, 1-75.
282. Tang, X.J.; Boyd, R.K. Rapid Commun. Mass Spectrom. 1992, 6, 651-657.
283. Loo, J.A.; Ogorzalek, R.R.; Light, K.J.; Edmonds, C.G.; Smith, R.D. Anal.
Chem. 1992, 64, 81-86.
284. Windholz, M.; Budavari, S.; Blumetti, R.F.; Otterbein, E.S. eds. The Merk
Index, Tenth Edition, Merk & Company Inc.: Pahway, NJ, 1983, p.188.
285. Hunt, D.F.; Shabanowitz, J.; Yates J.R., III; Griffin, P.R.; Zhu, N.Z. in
Methods in Protein Sequence Ana/ys/s Wittmann-Liebold, B., ed.; Springer-
Verlag: New York, 1989, 183-190.
286. Jenson, N.J.; Gross, M.L. Mass Spectrom. Rev. 1988, 7, 41-69.
287. Loo, J.A.; Udseth, H.R.; Smith, R.D. Rapid Commun. Mass Spectrom. 1988,
2, 207-210.
288. Loo, J.A.; Edmonds, C.G.; Smith, R.D. Anal. Chem. 1991, 63, 2488-2499.
289. Sheil, M.M.; Derrick, P.J. Org. Mass Spectrom. 1988, 23, 429^35.
290. Bricker, D.L.; Russell, D.H. J. Am. Chem. Soc. 1986, 108, 6174-6179.

424
291. Castro, J.A.; Nuwaysir, L.M.; Ijames, C.F.; Wilkins, C.L Anal. Chem. 1992,
64, 2238-2243.
292. Williams, E.R.; Furlong, J.J.P.; McLafferty, F.W. J. Am. Soc. Mass
Spectrom. 1990, 1, 288-294.
293. Yates, N.A. Methods for Gas Chromatography-Tandem Mass Spectrometry
on the Quadrupole Ion Trap Ph.D. Dissertation University of Florida,
Gainesville, FL, 1994.
294. Hellerqvist, C.G.; Sweetman, B.J. in Biomedical Applications of Mass
Spectrometry, vol. 34 John Wiley & Sons: New York, 1989, 177-244.
295. Kent, P.W. Pestic. Sci. 1994, 41, 209-238.
296. Reinhold, V.; Reinhold, B. Anal. Chem. 1995, 67, 1772-1784.
297. Kamerling, J.P.; Vleigenthart, J.F.G. in Clinical Biochemistry, Principles,
Methods and Applications: Mass spectrometry A.M. Lawson, ed.; Walter
de Gruyter: New York, 1989, 177-244.
298. Florencio, M.H.; Despeyroux, D.; Jennings, K.R. Org. Mass Spectrom.
1994, 29, 483-490.
299. McLaughlin, LG.; Henion, J.D.; Kijak, P.J. Biol. Mass Spectrom. 1994, 23,
417-429.
300. Garozzo, D.; Guiffrida, M.; Impallomeni, G. Anal. Chem. 1990, 62, 279-286.
301. Chen, Y.; Chen, N,; Li, H.; Zhao, F.; Chen, N. Biomed. Environ. Mass
Spectrom. 1987, 14, 9-15.
302. Schneider, R.P.; Lynch, M.J.; Ericson, J.F.; Fouda, H.G. Anal. Chem. 1991,
63, 1789-1794.
303. Laine, R.A.; Pamidimukkala, K.M.; French, A.D.; Hall, R.W.; Abbas, S.A.;
Jain, R.K.; Matta, K.L. J. Am. Chem. Soc. 1988, 110, 6931-6939.
304. Gu, J.; Hiraga, T.; Wada, Y. Biol. Mass Spectrom. 1994, 23, 212-217.
305. Domon, B.; Costello, C.E. Glycoconjugate J. 1988, 5, 397.
306. Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299.

425
307. Fenn, J.B.; Mann, M.; Meng, C.K.; Wong, S.F.; Whitehouse, C.M. Science
1989, 264, 64.
308. McLuckey, S.A.; Van Berkel, G.J.; Glish, G.L. J. Am. Soc. Mass Spectrom.
1992, 3, 60-70.
309. Edmonds, C.G.; Barinaga, C.J.; Loo, J.A.; Udseth, H.R.; Smith, R.D. in
Proceedings of the 37th ASMS Conference on Mass Spectrometry and
Allied Topics Miami Beach, FL, 1989, 844.
310. Corcoran, J.W.; Hahn, F.E., eds. Antibiotics III: Mechanism of Action of
Antimicrobial and Antitumor Agents Springer-Verlag: New York, 1975,396-
479.
311. Roberts, G.D.; Carr, S.A.; Christensen, S.B. in Proceedings of the 35th
ASMS Conference on Mass Spectrometry and Allied Topics Denver, CO,
1987, 933.
312. David, L; Scanzi, E.; Fraisse, D.; Tabet, J.C. Tetrahedron 1982, 38, 1619.
313. Barbalas, M.P.; McLafferty, F.W. Occolowitz, J.L Biomed. Mass Spectrom.
1982, 10, 258.
314. Cooper, R.; Unger, S.E. J. Antibiot. 1985, 38, 24.
315. Siegel, M.M.; McGahren, W.J.; Tomer, K.B.; Chang, T.T. Biomed. Environ.
Mass Spectrom. 1987, 14, 29.
316. Holzman, G.; Ostwald, U.; Nickel, P.; Haack, H.J.; Widjaja, H.; Arduny, U.
Biomed. Mass Spectrom. 1985, 12, 659.
317. Nelson, C.C.; McCloskey, J.A.; Isono, K. in Proceedings of the 37th ASMS
Conference on Mass Spectrometry and Allied Topics Miami, FL, 1989, 724.
318. Curtis, J.M.; Bradley, B.; Derrick, P.J.; Sheil, M.M. Org. Mass Spectrom.
1992, 27, 502.
319. Vincenti, M.; Guglielmetti, G.; Andriollo, N.; Cassani, G. Biomed. Environ.
Mass Spectrom. 1990, 19, 240.
320. Edwards, D.M.F.; Selva, E.; Stella, S.; Zerilli, L.F.; Gallo, G.G. Biol. Mass
Spectrom. 1992, 21, 51.
321. Straub, R.; Linder, M.; Voyksner, R.D. Anal. Chem. 1994, 66, 3651-3658.

426
322. Cerny, R.L; MacMillan, D.K.; Gross, M.L.; Mallams, A.K.; Pramanik, B.N.
J. Am. Soc. Mass Spectrom. 1994, 5, 152-158.

BIOGRAPHICAL SKETCH
James L. Stephenson, Jr. was bom in Richmond, Virginia, on October 18,
1961, to James L. and Vivian P. Stephenson. He spent his childhood building
tree forts, playing football, and riding/wrecking a wide variety of bicycles. In 1980,
he graduated from Hermitage High School and went on to attend college at East
Carolina University in Greenville, North Carolina. While attending East Carolina
University, he was active in Phi Sigma Pi Fraternity and spent an excess amount
of time taking math courses (as evidenced by chapter 3 of this dissertation). He
was awarded the Claude Pennock Todd Fellowship his senior year and graduated
cum laude in 1984 with a B.S. in biochemistry.
His first real job (i.e., not a lifeguard) was as an analytical chemist for
California Analytical Laboratories in Richmond, Virginia. It was here that his
interest in mass spectrometry was kindled. In the fall of 1985, he met his future
wife, Tracy A. Reitz of Alexandria, Virginia; the two were married in October 1987.
As his interest in mass spectrometry grew, he went to work as an Field Engineer
for Finnigan MAT in Washington, D.C. During this period, he was fortunate
enough to work with Dr. Henry Fates of the National Institutes of Health, where he
learned a great deal about mass spectrometry and life in general.
Jim and Tracy moved to San Jose, California, in 1988 due to Jims
427

428
promotion within Finnigan MAT to Applications Development Chemist (working
with the ion trap team). At exactly 5:04 P.M. on October 17, 1989, while
discussing the merits of going back to graduate school with Bob Finnigan, Jim
had the dubious distinction of hiding under Bobs desk during the Loma Prieta
earthquake. Despite the words "Rick Yost" being uttered just before the quake,
Jim and Tracy left California for graduate school (Jim in the Chemistry Department
and Tracy in the Journalism Department) at the University of Florida in August of
1990.
While in graduate school, Jim was awarded a departmental teaching
award, an Analytical Division Fellowship from the American Chemical Society, and
a Dissertation Fellowship From the College of Liberal Arts and Sciences at the
University of Florida. Upon graduation, Jim will work with Dr. Scott McLuckey as
a postdoctoral fellow at Oak Ridge National Laboratories in Oak Ridge,
Tennessee.

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.
-f
Richard A. Yosl/Chairman
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.
John R. Eyler j
Pressor 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.
(/ Professor of Microbiology and Cell
Science
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.
Robert T. Kennedy
Assistant 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.
David H. Powell
Associate Scientist 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.
fames D. Winefopdner
Braduate Research Professor of
Chemistry
This dissertation was submitted to the Graduate Faculty of the Department
of Chemistry in the College of Liberal Arts and Sciences and to the Graduate
School and was accepted as partial fulfillment of the requirements for the degree
of Doctor of Philosophy.
December, 1995
Dean, Graduate School



180
images for different m/z values may lead to mass discrimination, which can have
a profound effect on total ion transmission.253,254 Since the wavelength of secular
motion in the rf-only octopole varies widely with entrance angle and position, the
image created by either single or multiple m/z values will have a "smeared"
appearance due to the noncoherence of the secular motion. This noncoherent
secular motion does not produce the mass discrimination effects which can be
observed for the rf-only quadrupole.253
RF Amplitude
A plot of ion displacement versus multipole length for varying rf voltages
is seen in figure 3-12.253 The secular wavelength of ion motion for both the rf-only
quadrupole and octopole is shown to decrease for increasing values of rf voltage.
This result was expected theoretically since with increasing rf voltages both q and
P values increase, while the secular wavelength {A) decreases. This means the
ion will increase its trajectory to the point where it will become unstable (as
defined by the Mathieu stability diagram for the 4-pole or 8-pole device). The
limiting point for the rf-only quadrupole is q2=0.908, whereas in the case of the
octopole222,237, q470. Therefore, the octopole can be operated at much greater
rf voltages and requires very little, if any, tuning for a wide range of m/z ratios
(see figure 3-12). However, for larger entrance distances (>0.5 mm) off axis, the
tendency for ejection of ion from the octopole approaches that of the quadrupole.
This phenomenon can be explained by examining the force on a given ion in the


246
plate (of the detector assembly) to the 8" Conflat flange. The mounting bracket
was drilled out to allow high energy ions and neutrals to pass through the
detector region and strike the surface of the 8" Conflat, thus reducing the
background noise (from stray ions) associated with the detection event. A Bertan
model 205B power supply (Hicksville, NY) was used for the 20 kV applied to the
dynode. The electron multiplier was controlled by the ITMS electronics. The
anode lead (signal out) from the multiplier was directed into the ITMS electrometer
circuit. In addition, the anode lead was made as short as possible and was
shielded with copper foil to reduce spurious noise in the detector signal. A
schematic diagram of the set-up used is shown in figure 4-16.
Photodissociation Set-Up
Two different C02 lasers were employed for the photodissociation
experiments described in chapter 5 of this dissertation. The first was a
continuous wave C02 laser described previously, and the second was a pulsed
C02 laser (Lumonics Series TE-860-4 Excimer, Ottawa, Ontario, Canada) capable
of a 3 J pulse (at 10.60 //m) for an approximate 3 ns pulse width. The duty cycle
for full power operation of the laser was 20 Hz (e.g., 50 ms). Each laser required
a different optical set-up depending on the type of photodissociation experiment
employed. For optics mounted on the instrument table, two compensators
(described previously) were used to damp vibrations generated by the two 500
L/s turbomolecular pumps.


409
Vedel, F. Int. J. Mass Spectrom. Ion Processes 1989, 95, 119-156.
58. Haddon, W.F.; McLafferty, F.W. J. Am. Chem. Soc. 1968, 90, 4745.
59. Jennings, K.R. Int. J. Mass Spectrom. Ion Phys. 1968, 1, 227.
60. Rosenstock, H.M.; Melton, C.E. J. Chem. Phys. 1957, 26, 314.
61. Kupriyanov, S.E.; Perov, A.A. Russ. J. Phys. Chem. 1965, 39, 871.
62. Cooks, R.G.; Terwilliger, D.T.; Ast, T.; Beynon, J.H.; Keough, T.J. J. Am.
Chem Soc. 1975, 97, 1583.
63. Cooks, R.G.; Ast, T.; Beynon, J.H. Int. J. Mass Spectrom. Ion Phys. 1975,
16, 348.
64. Cody, R.B.; Freiser, B.S. Anal. Chem. 1979, 51, 547.
65. Freiser, B.S. Int. J. Mass Spectrom. Ion Phys. 1978, 26, 79.
66. Fedor, D.W.; Cody, R.B.; Burinsky, D.J.; Freiser, B.S.; Cooks, R.G. Int. J.
Mass Spectrom. Ion Phys. 1979, 31, 27.
67. Dunbar, R.C. in Gas Phase Ion Chemistry, Vol. 2 Bowers, M.T., ed.;
Academic Press: London, 1979, pp 182-217.
68. Dunbar, R.C. in Gas Phase Ion Chemistry, Vol. 3 Bowers, M.T., ed.;
Academic Press: London, 1984, pp 130-164.
69. Dunbar, R.C. in Molecular Ions: Spectroscopy Structure and Chemistry
Miller, T.A., ed.; North-Holland: Amsterdam, 1983.
70. Goeringer, D.E.; McLuckey, S.A. J. Chem. Phys. 1995, in press.
71. Morgan, R.P.; Brenton, A.G.; Beynon, J.H. Int. J. Mass Spectrom. Ion Phys.
1979, 29, 195.
72. Futrell, J.H. in Gaseous Ion Chemistry and Mass Spectrometry Futrell, J.H.,
ed.; John Wiley & Sons: New York, 1986.
73. Kaiser, Jr., R.E.; Cooks, R.G.; Stafford, Jr. G.C.; Syka, J.E.P.; Hemberger,
P.H. Int. J. Mass spectrom. Ion Processes, 1991, 106, 79-115.
74. Schwartz, J.C.; Jardine, I. Rapid Comm. Mass Spectrom. 1992, 6, 313-317.


135
F = eE
and the Maxwell equations 3-7 reduce to:
V E = 0
(3-8)
Vx E = 0
(3-9)
with the right hand side of equation 3-9 valid for an applied potential which has
a low enough frequency so that:
A = -l
(3-10)
where the wavelength of the applied potential (il) must be significantly larger than
the length (I) of the multipole pieces.
An electric field E which satisfies Maxwells relationship in equation(s) 3-9
can be derived by employing two mathematical theorems which combined yield
the basic equation for an applied potential to any multipole device in a vacuum
system. The first theorem is that of Stake's, which for any continuous
differentiable vector field A is defined by
V x A ds = j A dr
(3-11)
where dr is the vector on the circumference of the closed area S moving in the
direction of integration.232 By Applying Stakes theorem to the right hand equation
from Maxwells relationships in equation 3-9 yields:
E d?=0 (3-12)
for any closed path where for any surface VxE=0. The second mathematical


Delrin Supports
0,200
~r
1,061
1
//
chamfer
ends


5000
4000
3000
2000
1000
0
+9
+15
+16
+13
+19
+20
i- T-r 1^1 V[
+10
+12 +11
LL
ftrri'rrrfr
+8
rpm"|"i
Mii
P|
600 700 800
900
1000 1100 1200 1300 1400 1500 1600
m/z
277


The power absorption spectrum of protonated diglyme for an applied dipolar excitation signal (6
Vp.p) at qz=0.3 and az=0 (no He buffer gas present). The applied dipolar signal was started at a
frequency of 25 kHz and incremented at 0.1 kHz intervals to 500 kHz. The symbol u/z refers to the
fundamental frequency of motion for an ion in the z-direction.
Figure 2-7:


Table 2-3. kD values for the first consecutive reaction channels of protonated diglyme, 12-crown-4 ether, 15-crown-5 ether,
and allyl bromide.
Compound
kD (s'1)
Pressure (torr)
Back Reaction
Diglyme
97.2 1.9
3.6x1 O'7
yes
12-Crown-4 Ether
89.1 2.9
4.2x1 O7
yes
15-Crown-5 Ether
104 2.6
4.2x1 O'7
yes
Allyl Bromide
59.6 2.2
2.9x1 O'7
yes
o


Vacuum Connection
ESI Tube Lens
ESI Needle
Heated Capillary
212


14000
12000
10000
8000
6000
4000
2000
0
RF Voltage (V)
500 1000 1500 2000
J I I I I I I I I I I I I I I I I I I I
0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
Detected RF (absolute value, V)
285


29
The most important aspect of ESI is that of multiple charging. In a
landmark paper published in 1988, Fenn and coworkers reported as many as 45
charges attached to proteins of molecular weight 40,000 da.91 This work was
quickly duplicated and extended by both Smith and Covey.9293 The multiple
charging effect permitted the analysis of high molecular weight biological species
using existing quadrupole instrumentation, due to the lower mass-to-charge ratios
that can be obtained with multiple charging. Another advantage of the multiple
charging process is that a more accurate molecular weight determination can be
obtained from a distribution of multiply charged peaks.47 In addition, tandem
mass spectrometry has been shown to be useful for the dissociation of multiply
charged species. The most frequently used instrument for structural elucidation
studies of electrosprayed ions is the triple quadrupole mass spectrometer. One
of the most notable applications of tandem mass spectrometry and ESI ionization
has been the work of Don Hunt and colleagues.94"97 The Hunt group successfully
used only femtomoles of material to identify histocompatibility complex-bound
peptides. This work ultimately led to the identification of a peptide with high
binding affinity for cytotoxic killer T cells.94
The use of tandem mass spectrometry (triple quadrupole) in conjunction
with ESI is not the only active area of research in the ESI field. Other areas of
great interest include mechanistic investigations98"100, surface-induced
dissociation101,102, ion-molecule reactions103, non-covalent interactions104,105,


Photodissociation Efficiency (l-I/Io)
l.Oi
0.8
0.6
0.4
| 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
932 936 940 944 948 952 956
Wavenumbers in cm'1 ^


269


278
accessible for protonation. This effect explains the presence of the higher charge
state ions (+20 to +12) of the protein in a random coil confirmation (see figure
4-24).262-264 A summary of the instrumental parameters used to obtain the spectra
in figures 4-23 and 4-24 is seen in table 4-4.
Perhaps the most important observation in these initial studies is the low
concentration of sample used to tune the instrument (e.g., 500 fmol///L for
apomyoglobin and 350 fmol///L for horse heart cytochrome c). These values are
typically one order of magnitude lower than those used by standard
ESI/quadrupole systems. This sensitivity advantage can be attributed directly to
the rf-only octopole ion injection system of the ESI/ion trap instrument.
The initial test of absolute sensitivity was conducted with a 5 fmol/j/L
sample of bovine insulin. In figure 4-25 is seen the ESI/ion trap spectra of bovine
insulin at 5 pmol///L and 5 fmol///L, respectively. The difference in the two spectra
was the ion injection time for each sample. The instrumental conditions for
acquisition of multiply charge bovine insulin were very similar to those of
apomyoglobin and cytochrome c seen in table 4-4. The detection limit in the full
scan mode was taken at the point where all three charge state ions of bovine
insulin were still visible above the baseline with a signal-to-noise ratio of
approximately 3:1. The amount of sample consumed during the ion injection
period for 5 fmol///L of bovine insulin can be calculated by the following
equation:35


403
Photodissociation of ionized peptides/proteins, carbohydrates, and
oligonucleotides was demonstrated as a viable alternative to traditional CID
techniques. The photoabsorption cross-section for these classes of compounds
can be ranked in the following order: carbohydrates > oligonucleotides >
peptides/proteins. Evidence for this ranking can be explained by the reduced
amount of photodissociation observed for multiply protonated peptides/proteins,
a moderate amount of photodissociation for negatively charged oligonucleotides
(5 laser pulses), and copious amounts of fragmentation for positive ions from
carbohydrates (1 to 3 laser pulses) using a pulsed C02 laser.
For the case of multiply protonated peptides and proteins,
photodissociation spectra are similar to CID spectra with the exception of the
presence of high-mass b ions in the photodissociation spectra. One plausible
explanation of this behavior may be the collisional stabilization of the product
species at helium buffer gas pressures in the low 1x1 O'5 torr range. As the
pulsed-valve comes on-line with ion injection event, more detailed and quantitative
studies can be performed in order to better understand this behavior.
In oligonucleotide analysis, the information provided by photodissociation
represents direct dissociation of the phosphodiester bridge between individual
nucleotides. To date, the largest oligonucleotide analyzed by photodissociation
in the instrument was a DNA 14-mer. Due to calibration issues and peak
resolution problems from the large scan range (the limited data acquisition ADC
steps result in reduced peak definition, i.e. less ADC steps per peak the larger the


149
(3-36)
Differentiating equation 3-36 with respect to y and (substituting for n=4 and
evaluating k=1 to 4) and applying equations 3-30 and 3-35, the vector component
Ey is defined as:219
Ey = 4>0 r3 sin 30
(3-37)
Electrode Geometry
If the electrode potentials of the adjacent octopole electrodes are defined
as +0/2 and -0/2 respectively (see figure 3-1), and 0O is defined as in equation
3-20, then the equation of the contour line on the hyperbolic surface of the
octopole electrode may be obtained. The equation for the contour surface
derived here will represent the xy surface, with a corresponding substitution of the
condition 0=0O into equation 3-26 (the field potential in polar coordinates). This
substitution for an octopole geometry gives:
(3-38)
cos 46 = 1
For rectangular coordinates, the expression obtained from equation 3-30 is:
To convert the contour equations to the per-unit system where the spatial


390
interest.298*99302311-322
Perhaps the most widely used macrolide antibiotic is erythromycin. This
antibiotic is typically used to control respiratory infections, and is particularly
effective against Gram-positive organisms involved in streptococcal,
staphylococcal, and pneumococcal infections.310 The mechanism of action
involves binding of the antibiotic to the 50S subunit of the bacterial ribosome, thus
blocking the action of peptidyl transferase in the peptide elongation process. In
figure 5-19 is shown the structure of erythromycin, with the aglycone moiety and
two attached monosaccharides (D-desosamine and L-cladinose).
It is the purpose of this section to evaluate the ability of the
photodissociation technique for structural elucidation of carbohydrate antibiotics.
Studies were performed with protonated erythromycin as a model compound in
order to understand the fragmentations observed with basic macrolide antibiotics.
Macrolide Antibiotics Erythromycin
The erythromycin standard was obtained from ICN Pharmaceutical (Costa
Mesa, CA). Solutions were made up in 50:50 methanol:water with 0.2% acetic
acid. The concentration of the electrospray standard was 5 pmol///L. All
electrospray and instrumental conditions were the same as described in chapter
4. The photodissociation set-up employed a pulsed C02 laser as discussed in
chapter 4 of this dissertation.
The preferential site of protonation is the tertiary amine group on the D


12000
10000
8000
6000
4000
2000
0
(M-H)"
m/z 1059
rftVpiVf rrqprVTT pt ft i jrtrrp^r
l
r
1
400 500 600 700 800
m/z
900
1000 1100 1200
327


Intensity
m/z
CO
JO


306
ion of interest is ramped to its position just below the resonant frequency, ions of
lower m/z are resonantly ejected from the trap. To resonantly eject ions of higher
m/z, the resonant excitation voltage is turned off and the rf amplitude is ramped
so that the ion of interest is located between the resonant ejection frequency and
the z-stability edge of the Mathieu diagram (see figure 4-31). The resonance
ejection voltage is turned on again, and the rf amplitude is ramped in the reverse
or downward direction so that ions of higher m/z than the ion of interest are
resonantly ejected. In this case, the ion of interest is located just above the
applied resonant ejection frequency and is the only remaining ion (m/z) in the
trap.
This exact procedure with isolation of the +18 charge state (m/z 943) of
horse muscle apomyoglobin is shown in figure 4-32. The chosen resonant
ejection frequency for this experiment was 60 kHz at 6.3 V^P. In the top portion
of figure 4-32 is shown the elimination of the lower mass fragments (higher
charge states) of apomyoglobin. The middle section demonstrates the reverse
scan process where higher m/z ions (lower charge states) are eliminated. The
last portion of this figure shows the complete isolation of the +18 charge state of
apomyoglobin achieved when the forward and reverse scans are combined. This
technique is very efficient with isolation efficiencies much greater than those
obtained with traditional rf/dc isolation routines. Also, this technique can be used
quite effectively for high resolution (e.g., slow scanning) isolation experiments.


Figure 5-20:
Photodissociation spectrum (7 laser pulses) oftheprotonated [M+H]+ ion m/z734 of erythromycin.
The mechanism and composition of the fragments labeled aglycone ring cleavages are unknown.
Many of the fragments observed (e.g. C1fl and Y0/J) correspond to data obtained using high energy
CID as seen on magnetic sector instruments.


9
4>0 = U-V0_p cost (1-2)
where Q is the angular frequency of the rf trapping field applied to the ring
electrode (in rad s'1, which is equal to 2/rf where f is the frequency in Hz), U is the
applied dc voltage, is the zero-to-peak amplitude of the rf voltage, and t is the
time variable.23
For any quadrupolar device, the field is uncoupled in the three coordinate
directions, so that the forces acting on an ion are independent of one another.
For this condition to hold, equation 1-1 must satisfy the LaPlace condition where
the field strength at the center of the ion trap must equal zero. Therefore, when
the potential 0O is applied to the ring electrode (xy plane) and the two hyperbolic
endcaps are held at ground potential, the potential at any point 0 within the
device is represented by equation 1 -1.23
Because the geometry of the quadrupole ion trap has cylindrical symmetry,
the x and y components are combined to give a single radial component defined
by x2+y2=r2. The orientation in space of the ring electrode and two hyperbolic
endcaps needed to define a quadrupolar trapping field is r2=2z02, where r0 is the
radius of the ring electrode and z is the center-to-endcap distance.52 The
hyperbolic shape of the electrodes in this geometrical configuration can be seen
in figure 1-1. The equations defining the hyperbolic shape of the ring electrode
and endcap electrodes are given as:


23


Transverse Kinetic Energy (eV)
O O
200
Displacement (mm) \Z\
*!

CO


Figure 4-4:
Analytica electrospray ionization source equipped with a heated capillary (as modified by Mark Hail
at Finnigan MAT, San Jose, CA).


39
As with the ICR technique, the quadrupole ion trap is capable of storing
ions for long periods of time. The storage capability makes the quadrupole ion
trap mass spectrometer very compatible with a wide range of experiments using
light. One of the first successful uses of the ion trap in conjunction with
photodissociation involved the study of the proton-bound dimer of 2-propanol
utilizing a cw infrared laser.20 2-Propanol was chosen for study since its gas-
phase ion chemistry is well known and the formation of the protonated dimer is
easily accomplished. The instrumental configuration consisted of an ion trap
connected directly to the ion source of a quadrupole mass filter (QUISTOR mode
of operation). A detailed description of the optimization of ion trapping
characteristics for studies of ion photodissociation is a QUISTOR can be found
in March and Hughes.21
These earlier studies of the IRMPD process in the ion trap were performed
with a single-pass ring electrode design, with a 3 mm diameter hole on the center
axis of the ring electrode as the entrance aperture for the low power cw C02 laser
beam. Upon reaching the other side of the ring electrode, a portion of the beam
passed through a 0.8 mm diameter hole and through a NaCI window, where the
laser power was monitored externally. The remainder of the laser beam was
reflected by the ring electrode throughout the QUISTOR. The pressure of 2-
propanol was adjusted to 5 mPa so that photodissociation of the proton-bound
dimer, (2M+H)+, at m/z 121 could occur at an appreciable rate. At pressures
optimum for the formation of the proton-bound dimer (13 mPa), no laser-induced


Figure 5-2:
Initial photodissociation spectrum of the potassium adduct and protonated 18-crown-6 ether
obtained on the electrospray/ion trap instrument. The peak at m/z 39 represents the direct
dissociation of the potassium adduct, and the peak at m/z 144 the dissociation of the [M+H]+ ion.


377
depended on the amplitude of the resonance ejection frequency and single
frequency CID signal. In figure 5-15 is shown the fragmentation scheme for the
ammonium adduct ion of stachyose. As was the case for raffinose, another stage
of mass spectrometry is needed to obtain entire sequence information.
The photodissociation spectra for 1,2 and 3 laser pulses of the ammonium
adduct of stachyose are seen in figure 5-16. After two laser pulses, the entire
sequence of the tetrasaccharide is obtained, with the peak at m/z 164 indicating
either the nonreducing terminus (B1o) ion or the blocked reducing end D-fructose
ion (Z^). For the three laser pulse spectrum, peaks at m/z 146 and m/z 128
indicate the successive losses of H20. Interestingly enough, the small peak
indicating ring opening cleavages at approximately m/z 589 is not observed in the
photodissociation spectra. However, an unknown peak at m/z 640 is observed
in all three spectra (structure unknown). As was the case for raffinose, the [1-2]
/? linkage between the D-glucosyl and D-fructofuranosyl is the most labile
glycosidic bond in stachyose. Therefore, the sequential product ions in figure 5-
16 probably represent charge retention the nonreducing terminus, again
consistent with previously reported low-energy CID data47 Fragmentation
patterns for the ammonium adduct of stachyose are found in figure 5-15.
Oligonucleotides
The first real advances in mass spectrometry for oligonucleotide analysis
came with the advent of laser desorption and electrospray ionization


rf Power Amplifier
50 ohm cable
Low-Pass
Filter
rf Timing Unit
Remot-rf
215


195
experimental values for reasonable collision cross sections.257
Extension of this work to nonlinear devices such as the rf-only hexapole
and octopole, has not been accomplished. Results from the rf-only quadrupole
study suggest strongly the coupling of atmospheric pressure ionization sources
with rf-only devices for ion injection. Future studies in our laboratory will focus
on the effect of collisions on ion transmission in rf-only multipoles with n>2.


Stainless steel mounting assembly design used for alignment/assembly of the octopole rods. The
extruded tabs are for attachment of the individual rods (e.g spot welded) for generation of the
correct octopolar field. This condition is obtained when the stainless steel mounting assemblies
are offset 45 from one another on the front/back surface of the Delrin support pieces.
Figure 3-6:


Figure 2-12:
Determination of the rate coefficient for protonated diglyme at a pressure of 1.1x107 torr by using
weighted linear regression, kD=97.2 1.9 s'1 (slope 95% confidence interval of the slope). Error
bars are defined as the standard deviation of the mean.


Transverse Kinetic Energy (eV)
193
(a)
(b)


416
166. Vedel, F.; Vedel, M.; March, R.E. Int. J. Mass Spectrom. Ion Processes
1990, 99, 125-138.
167. Stephenson Jr., J.L.; Booth, M.M.; Eyler, J.R.; Yost, R.A. submitted to R.
Comm. Mass Spectrom.
168. Vedel, F. Int. J. Mass Spectrom. Ion Processes 1991, 106, 33-61.
169. Vedel, F.; Vedel, M.; March, R.E. Int. J. Mass Spectrom. Ion Processes
1991, 108, R11-R20.
170. Lias, S.G.; Bartmess, J.E.; Liebman, J.F.; Holmes, J.L.; Levin, R.D.; Mallard,
W.G. J. Phys. Chem. Ref. Data 1988, 17, Suppl. 1, 74, 90, 274.
171. Dewar, M.J.S.; Zoebisch, E.G.; Healy, E.F.; Stewart, J.J.P. J. Am. Chem.
Soc. 1990, 107, 3902.
172. Coaxian, J.; Stephenson Jr., J.L. AM1 Calculations, University of Florida.
173. Katritzky, A.R.; Watson, C.H.; Dega-Szafran, Z.; Eyler, J.R. J. Am. Chem.
Soc. 1990, 112, 2476.
174. Lias, S.G.; Bartmess, J.E.; Liebman, J.F.; Holmes, J.L.; Levin, R.D.; Mallard,
W.G. J. Phys. Chem. Ref. Data 1988, 17, Suppl. 1, 52, 114.
175. Baykut, G.; Watson, C.H.; Weller, R.R.; Eyler, J.R. J. Am. Chem. Soc. 1985,
107, 8036-8042.
176. Harcourt, A.V.; Esson, W. Proc. R. Soc. London 1865, 14, 470.
177. Harcourt, A.V.; Esson, W. Philos. Trans. 1866, 156, 193.
178. Harcourt, A.V.; Esson, W. Philos. Trans. 1867, 157, 117.
179. Capellos, C.; Bielski, B.H.J. Kinetic Systems: Mathematical Descriptions of
Chemical Kinetics in Solution Wiley-lnterscience: New York, 1972, 46-49.
180. Laidler, K.J. in Chemical Kinetics Harper & Row: New York, 1987, 278-279.
181. Moore, J.W.; Pearson, R.G. in Kinetics and Mechanism John Wiley & Sons:
New York, 1981, 290-296.
182. Woodin, R.L.; Bomse, D.A.; Beauchamp, J.L. Chem. Phys. Lett. 1979, 63,
630-636.


134
moves through a vacuum (free space). The Maxwell equations now become:
V e = 0 v x E =
at
V B = 0
v B a(e0E)
F0 at
(3-6)
where under vacuum conditions (approximately free space), the values of e and
// become the constants e0 and /0. Therefore, for a low enough density of ions
(p and J approach zero), the Maxwell equations in 3-6 are a good approximation
in the vacuum region enclosed by any number of pole pieces. When the E and
B fields in the Maxwell equations are not changing with time (e.g., instantaneous
values), the Maxwell equations are evaluated as:
V E = 0 V x E=0
V B = 0 V x B = 0
(3-7)
The applicability of the Maxwell equations in equation 3-7 to the case of an
enclosed vacuum region (any number of pole pieces) where the electromagnetic
fields are changing, can be explained by the speed of an ion as it moves through
the device compared to that of the speed of light. From Maxwells equations 3-6,
B0=E0/c where B0 and E0 are the maximum amplitudes of the E and B fields233,
and c is the speed of light with c={ejioy'12. Since the fastest speed of an ion in
a quadrupole instrument235-236 is only v/cdO"4, the B term in the Lorentz force law
is negligible and equation 3-2 reduces to:


Brass Ring
Top View
Side View i0.125
0.500"
I
Teflon Plug
Top View
Q-00.234
Chamfer
r \
1.000"
v. 2
l | 0.234"
End View
Alignment Jig
168


Vacuum Manifold
L
]d
J Ion Trap
[
"i ^ r
Focusing Lens
1
_
1
\
\
\
%
Instrument Table Helium-Neon
co2
Gold Plated
Laser Beam
Reflecting Mirror
/ /
/
/
*
*
/
/
Optical Table
111111 111 11111 i 1111 i i 111 i 1111 i 1111 i
111 111
/
/
Removable /
Gold Plated
Reflecting Mirror
C
!b
Lumonics
Pulsed CO 2
Laser
253


A great deal of credit for the success of this project goes to Matt Booth
who has been my collaborator, co-worker, and good friend; from San Jose to
Gainesville, Matt always had an idea, a beer, or a word of encouragement that
kept me going. This dissertation also would not have been possible without the
Herculean design and machining effort put forth by Joe Shalosky; he is truly an
artist. In addition, I would like to acknowledge Scott Quarmby for assistance with
electronic design of the rf circuitry and Stephen Bou for the set-up/operation of
the pulsed C02 laser.
My Yost Group" experience was made possible by a whole cast of
characters including Uli Bernier, Tim Griffin, Tracie Williams, Shannan Carlson,
and John Laycock, just to name a few. To the "Night Shift of the Yost Lab,"
Nathan Yates, Don Eades, Jodie Johnson, and Brad Coopersmith, I am truly
grateful for the late-night beers at the Salty Dog and the "Hut" experience, which
made writing this dissertation a much less painful process.
I wish to thank my parents, James L and Vivian P. Stephenson, for
instilling in me the value of education, hard work, and humility. Their
unconditional support during the last five years has been invaluable.
Lastly, I wish to thank my wife Tracy (a.k.a Chica) for her love, support, and
understanding. She organized the weekend getaways, the La Chua trail walks,
B.E. breaks, and parties which added spice to our lives and memories to hang
on to.
IV


20
(1) the dissociation event is long when compared to the time needed for
ionization or excitation; (2) the rate of the dissociation event is much slower than
the corresponding rate for the redistribution of the energy deposited during
ionization/activation; (3) fragmentation is the result of a series of competing and
consecutive reactions; and (4) an internal energy equilibrium is achieved where
the energy is randomized over all internal states with equal probability.1
After the initial energy deposition, the weaker bonds are preferentially
broken, thus revealing mass fragments indicative of the parent ion. This would
mean that a mass spectrum or MS/MS spectrum depends only on the amount of
energy deposited into the ion and not on how that energy was deposited. For a
typical MS/MS experiment, energy can be deposited during the ionization event
and/or in a subsequent ion activation reaction.1
The product ions produced in an MS/MS experiment are determined by
several factors in addition to the amount of internal energy deposited into the
system. These include both the time frame for the reaction to occur and the
individual microscopic rate constants for each dissociation pathway. Since there
exists an inverse relationship between ion lifetimes and the rate of dissociation of
an ion, examining the plot of mass spectra as a function of ion lifetime can yield
information similar to that obtained with traditional breakdown curves. Early
studies by Morgan et al. have reported that rate constants in excess of 2 x1 O'7 s1
were required for daughter ions to be dissociated in the collision cell of a reverse
geometry (magnetic sector followed by electrostatic sector, BE) double-focusing


177
phase are much more pronounced. The smaller the angle of entry for an ion into
the octopole, the longer the wavelength for the secular motion of an ion as it
traverses the length of the octopole.253
Figure 3-1 1253 shows the dependence of ion trajectories as they enter the
multipole devices from different distances off the center axis. As with the previous
example, the wavelength and phase of secular motion in the rf-only quadrupole
are approximately the same, while for the case of the rf-only octopole longer
wavelengths of secular motion are observed for an ion which enters the octopole
closer to the center axis.
The authors reported that the results of these two experiments on the beam
shape of an ion packet as it exits these multipole devices are twofold.253 For the
rf-only quadrupole, where the fundamental wavelength and phase of secular
motion show little variation, the net secular motion exhibits some form of coherent
oscillation. Therefore, the image the ion packet makes as it exits from the
quadrupole will depend on the average distance the ions (of a particular m/z) are
away from the central axis of the device. The smaller the image, the better the
focusing properties of the rf-only quadrupole for efficiently transferring the ion
packet to another analyzer (e.g., quadrupole mass filter with rf/dc voltages or
quadrupole ion trap). The ion beam image for a series of different m/z ions will
be much more complex than that of the single m/z case. Here, the wavelength
of secular motion is different for the different m/z ions and will produce a much
more complex beam image at the exit of the quadrupole. These different beam


Intensity
376


Analyzer Mounting Plate
Assembly Pin Hole Dimensions
(Bottom
. 125
233


Figure 4-31: Sequential steps of the forward-reverse scan isolation process.
To eject unwanted low mass ions the mass of interest is ramped
forward towards the resonance point. To eject unwanted ions of
higher m/z, the resonance frequency is turned off and the ions
are moved past the resonance point by using an rf ramp. Next,
the resonance frequency is turned back ion and the ions are
scanned in reverse towards the resonance point to eject the high
mass ions (from reference 74).


Figure 4-20:
El mass spectrum of PFTBA taken with the rf-only octopole/ion trap system. The high mass end
of the spectrum is very intense compared to that obtained from quadrupole instruments. The
increased sensitivity on the high mass side is characteristic of ion trap spectra in general.


Figure 2-20:
Gas-phase spectroscopic study of allyl bromide neutrals and ions. The spectrum (FTIR) on the
left is for neutral allyl bromide at. a pressure of 0.45 torr; on the right hand side is shown the
spectrum obtained using IRMPD at a pressure of 1.1x1 O'7 torr. A comparison between the neutral
and IRMPD spectra is feasible since a non-bonding electron is removed from the bromine atom
(of allyl bromide) during the ionization event, therefore not affecting the force constant of the CBr
stretch. Error bars are defined as the standard deviation of the mean.


6
molecules. These collisions caused viscous damping of ionic motion, thereby
focusing the ion cloud to the center of the trap.
These improvements led to the first commercially available ion trap detector
(ITD 700) developed by Finnigan MAT. This product was designed as a low cost
benchtop GC/MS detector, which gained popularity for trace level environmental
and clinical analysis. Further improvements in dynamic range were obtained with
automatic gain control (AGC), which regulated the number of ions stored in the
trap as a function of sample concentration.27 This limited space charge effects
and ion-molecule reactions typically seen with constant ionization time
experiments.
The first fully functional research-grade ion trap mass spectrometer was
developed by Kelley et al. in 1985.28,29 The Finnigan MAT Ion Trap Mass
Spectrometer (ITMS) possessed a wide range of experimental capabilities
including electron ionization (El), chemical ionization (Cl), mass isolation (e.g.,
apex isolation), tandem MS", and user-defined software. With the advent of
resonance ejection (axial modulation) in 1988, further improvements in resolution
and dynamic range were achieved.30 The technique of resonance ejection was
also responsible for the extension of the mass range of the quadrupole ion trap
to well beyond m/z 50,000.
Rapid growth in the field of ion trap mass spectrometry over the last six to
seven years has been driven by the coupling of external ion sources to the
device. Some of the external ion sources coupled to the ITMS include fast atom


x drive
220


Figure 1-1:
The geometric configuration of the quadrupole ion trap. Shown are the three hyperbolic electrodes
which comprise the analyzer. The center-to-endcap and the center-to-ring distances are indicated
by z0 and r0 respectively.


D-Galactose
ch2oh
oh ) o
OH
D-Galactose
o'
OH
D-Galactose
CH2
o
OH
OH
O'
,CH
D-Fructose
ch2oh
OH
2 D-Glucose
o
OH
o-
OH
k3
ch2oh
OH
CH2OH 9
OH
D-Fructose
HO
3
ch2oh
Raffinose
mw=504.2
OH
Stachyose
mw=660.2
365


424
291. Castro, J.A.; Nuwaysir, L.M.; Ijames, C.F.; Wilkins, C.L Anal. Chem. 1992,
64, 2238-2243.
292. Williams, E.R.; Furlong, J.J.P.; McLafferty, F.W. J. Am. Soc. Mass
Spectrom. 1990, 1, 288-294.
293. Yates, N.A. Methods for Gas Chromatography-Tandem Mass Spectrometry
on the Quadrupole Ion Trap Ph.D. Dissertation University of Florida,
Gainesville, FL, 1994.
294. Hellerqvist, C.G.; Sweetman, B.J. in Biomedical Applications of Mass
Spectrometry, vol. 34 John Wiley & Sons: New York, 1989, 177-244.
295. Kent, P.W. Pestic. Sci. 1994, 41, 209-238.
296. Reinhold, V.; Reinhold, B. Anal. Chem. 1995, 67, 1772-1784.
297. Kamerling, J.P.; Vleigenthart, J.F.G. in Clinical Biochemistry, Principles,
Methods and Applications: Mass spectrometry A.M. Lawson, ed.; Walter
de Gruyter: New York, 1989, 177-244.
298. Florencio, M.H.; Despeyroux, D.; Jennings, K.R. Org. Mass Spectrom.
1994, 29, 483-490.
299. McLaughlin, LG.; Henion, J.D.; Kijak, P.J. Biol. Mass Spectrom. 1994, 23,
417-429.
300. Garozzo, D.; Guiffrida, M.; Impallomeni, G. Anal. Chem. 1990, 62, 279-286.
301. Chen, Y.; Chen, N,; Li, H.; Zhao, F.; Chen, N. Biomed. Environ. Mass
Spectrom. 1987, 14, 9-15.
302. Schneider, R.P.; Lynch, M.J.; Ericson, J.F.; Fouda, H.G. Anal. Chem. 1991,
63, 1789-1794.
303. Laine, R.A.; Pamidimukkala, K.M.; French, A.D.; Hall, R.W.; Abbas, S.A.;
Jain, R.K.; Matta, K.L. J. Am. Chem. Soc. 1988, 110, 6931-6939.
304. Gu, J.; Hiraga, T.; Wada, Y. Biol. Mass Spectrom. 1994, 23, 212-217.
305. Domon, B.; Costello, C.E. Glycoconjugate J. 1988, 5, 397.
306. Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299.


Mounting Plate for
Top View of Mounting Plate
(Vertical Dimensions)
Manifold Assembly
Top View of Mounting Plate
(Horizontal Dimensions)
units are in inches
199


8
motion is important for describing the conversion of an ions kinetic energy (for
CID and SID) of motion into the vibrational energy needed to accomplish
fragmentation in an MS/MS experiment. For efficient photodissociation, the
amount of time required for sufficient absorption of a photon(s) is directly related
to the overlap of the ion trajectory with the incident radiation. Consequently, it is
important to grasp the basic concepts of ion motion and how they might be
applied to the interaction between ions and light.
The motion of an ion in a quadrupole field can be described
mathematically by the Mathieu equation, a second-order linear differential
equation originally used to characterize the vibrational motion of a stretched skin
or membrane.51 The derivation begins by assuming the presence of an ideal
quadrupolar field (no space charge effects due to other ions), defined by 0, the
potential at any point (x,y,z) in that field:
4> = -^Ux2+0y2+Yz2) (1-1)
ro
where 0O is the applied electric potential, A, a, and y are a weighting constants
in the x, y and z directions respectively, and r0 is the inscribed radius of the ring
electrode. Any oscillatory and dc potentials applied to the ring electrode can be
represented mathematically in the form:


Intensity Intensity
358
250 i
200
150
100
50
Product Ion Growth Curves
m/z 165 m/zl46
m/z 164 + m/z 129
m/z 147 m/z 111
0 20 40 60 80 100 120 140 160 180 200
Laser Irradiance Time in ms


Figure 5-15:
Glycosidic bond cleavages of the ammonium adduct of stachyose for the collisional and photon
activation processes. The ring cleavages were only observed in the CID spectrum.


Figure 3-12:
(a) RF voltage dependence (v=100, 200, and 300 V) in an rf-only quadrupole, q2 values are 0.433,
0.866, and 1.30 respectively. Conditions for m/z 100 are ion entry at 0.25 mm off axis at an angle
of 0.5 and an rf frequency of 1.2 MHz. (b) RF voltage dependence (V=100, 200, 300, 500, and
1000 V) in an rf-only octopole, q4 values are 3.46, 6.92, 10.4, 17.3 and 34.5 respectively.
Remaining conditions as in (a). Adapted from reference 253.


144
X


169
one end of the assembly. Rods that are out of alignment (e.g., ends not in the
same plane) can introduce fringing fields and forces in the z-direction of ion
motion, thus reducing ion transmission efficiencies. The other end of the rods
had a brass ring (placed approximately 1 inch from the end-plane) with eight
precision drilled holes to ensure the correct spacing of the rods during the
welding process.
To apply the rf field to the rod assembly, two stainless steel pins were
welded onto the solid support assembly. These pins were designed to accept
copper-beryllium connectors from the rf supply/flange. A complete diagram of the
octopole assembly can be seen in figure 3-8.
Factors Affecting Ion Transmission
A theoretical discussion of the factors that influence ion transmission is
particularly important when evaluating and designing an rf-only octopole.
Parameters such as initial entry angle conditions, rf frequency, rf amplitude,
kinetic energy (transverse direction or xy plane) and pressure must be considered
when interfacing an ion source to a mass spectrometer via an rf-only multipole.
Although the author spent a copious amount of time deriving and understanding
the equations of motion for the octopole, to date there is no simulation program
for which our design can be fully evaluated. With the upcoming introduction of
Simion ver. 6.0 for Windows252, the design parameters and equations from the
above section could be used to develop a rather versatile simulation program.


21
instrument.71 By applying the basic ideas of transition state theory to individual
microscopic rate constants, the expression for the intemal-energy-dependent rate
constant becomes:
1 W*(E-E0) 1 W*(E-E0)
h dW(E)/dE h p(E)
(1-12)
where W(E) is the number of energy states of the ion with energy less than or
equal to E, p is the energy level density (dW/dE), W* is the same as W except
that the transition state is assumed for the ion, E0 is the ion activation energy and
therefore (E-E0) can be defined as the internal energy of the ion. A schematic
representation of the relevant energetics72 is shown in figure 1 -3.
General Operation
The quadrupole ion trap instrumentation used and developed for this
dissertation was based on an early design by Kelley et al. of Finnigan MAT (San
Jose, CA).27'28 A schematic of the Finnigan MAT ITMS is shown in figure 1-4.
In the normal mode of operation, electrons are gated into the central volume of
the analyzer (through a 1/16" hole) by pulsing a gate electrode to +180 V for a
specified ionization time period. The ions formed by electron ionization are
trapped by an rf voltage applied to the ring electrode of the analyzer (while the
endcap electrodes are held at ground potential). The detection event is
accomplished by ramping the amplitude of the rf voltage applied to the ring


331
was shown to have greater selectivity, have less mass discrimination, and could
dissociate much more stable ions than the corresponding CID process.111
This chapter begins with a survey of both pulsed and continuous wave
lasers for photodissociation of ions from peptides, oligosaccharides, and
oligonucleotides. This is followed with a discussion of the relative merits of
photodissociation versus the CID process. In addition, the very first ESI-
photodissociation spectrum is shown; the remaining data in this chapter represent
a combination of the fundamental studies presented in chapter 2, the theoretical
and design principles in chapter 3, and the instrumentation considerations of
chapter 4. The multipass ring electrode discussed in chapter 2 is the heart of
these experiments, using the increased photoabsorption pathlength to counter the
effects of higher pressure operation in the ion trap which can quench
photodissociation at these wavelengths. The data shown in this chapter are the
first ever reported for photodissociation of biological species in the quadruple
ion trap. The high sensitivity associated with the ion trap, combined with an
efficient ion injection system for transport of ions from an electrospray LC/MS
interface, makes the ion trap an obvious choice for the analysis of micro
quantities of material from biological extracts. Combined with photodissociation
for structural elucidation, an instrument of this design has the potential to solve
a variety of problems in the field of biochemistry.
As mentioned earlier, one major advantage of photodissociation is that an
ions kinetic energy does not have to be converted into internal energy to effect


408
42. Kaia, S.M.S.; Andre, J.; Zerega, Y.; Brincourt, G.; Catella, R. Int. J. Mass
Spectrom. Ion Processes 1991, 107, 191-203.
43. Kelley, P.E.; Hokeman, D.J.; Bradshaw, S.C.; Jones, L.B.; Louris, J.N.
Wilson, R.W. in Proceedings of the 42nd ASMS Conference on Mass
Spectrometry and Allied Topics Chicago, IL, 1994, 709.
44. Griffiths, I.W. Rapid Comm. Mass Spectrom. 1990, 4, 69-73.
45. Todd, J.F.J. Mass Spectrom. Reviews 1991, 10, 3-52.
46. March, R.E. Int. J. Mass Spectrom. Ion Processes 1992, 118/119, 71-135.
47. McCloskey, J.A., ed. in Methods in Enzymology, Volume 193, Mass
Spectrometry Academic Press: San Diego, 1990.
48. Matsuo, T., ed. Biological Mass Spectrometry: Present and Future Wiley:
New York, 1994.
49. Desiderio, D.M., ed. Mass Spectrometry of Peptides CRC Press: Boca
Raton, FL, 1991.
50. McEwen, C.N.; Larsen, B.S., eds. Mass Spectrometry of Biological Materials
Marcel Decker: New York, 1990.
51. Mathieu, J. Math Purs Appi (J. Liouville), 1868, 13, 137.
52. Knight, R.D. Int. J. Mass Spectrom. Ion Phys. 1983, 51, 127-131.
53. Louris, J.; Stafford, g.; Syka, J.; Taylor, D. in Proceedings of the 40th
Conference on Mass Spectrometry and Allied Topics Washington, DC,
1992, 1003.
54. Johnson, J.V.; Pedder, R.E.; Yost, R.A. Rapid Comm. Mass Spectrom.
1992, 6, 760-764.
55. McLachlan, N.W. in Theory and Applications of Mathieu Functions
Claredon: Oxford, 1947.
56. Paul, W. "Historical Development of the Quadrupole Ion Trap" presented at
the 40th ASMS Conference on Mass Spectrometry and Allied Topics
Washington, D.C., May 31 June 5, 1995.
57. March, R.E.; McMahon, A.W.; Loundry, F.A.; Alfred, R.L.; Todd, J.F.J.;


Figure 5-11:
Carbohydrate fragmentation nomenclature as determined by Domon and Costello (from reference
305). Charge retention on the reducing terminal is indicated by the X, Y, and Z fragmentation.
Charge retention on the nonreducing end is represented by the A, B, and C fragments. Ring
opening cleavage fragments are X and A with superscripts denoting bond breaking points on the
ring.


3000
2500
2000
1500
1000
500
0
Asp-Arg-Val-Tyr-IIe-His-Pro-Phe-His-Leu
PFHorHPF
a.
t t f 'ftMyy4HLrf4l
10 300 400 500 600 700 800 900 1000
Ill/z
1100
312


407
29. Louris, J.N; Amy, J.; Ridley, T.; Kascheres, C.; Cooks, R.G. in Proceedings
of the 35th ASMS Conference on Mass Spectrometry and Allied Topics
Denver, CO, 1987, 766-767.
30. Weber-Grabau, M.; Kelley, P.E.; Syka, J.E.P.; Bradshaw, S.C.; Hokeman,
D. J. in Proceedings of the 36th ASMS Conference on Mass Spectrometry
and Allied Topics San Francisco, CA, 1988, 1106-1107.
31. Louris, J.N.; Brodbelt-Lustig, J.S.; Kaiser, R.E.; Cooks, R.G. in Proceedings
of the 36th ASMS Conference on Mass Spectrometry and Allied Topics San
Francisco, CA, 1988, 968-969.
32. Duckworth, D.C.; Barshick, C.M.; Smith, D.H.; McLuckey, S.A. Anal. Chem.
1994, 66, 92-98.
33. Glish, G.L; Goeringer, D.E.; Asano, K.G.; McLuckey, S.A. Int. J. Mass
Spectrom. Ion Processes 1989, 94, 15-24.
34. Weber-Grabau, M.; Kelley, P.E.; Syka, J.E.P.; Bradshaw, S.C.; Brodbelt,
J.S. in Proceedings of the 35th ASMS Conference on Mass Spectrometry
and Allied Topics Denver, CO, 1987, 769-770.
35. Van Berkel, F.J.; Glish, G.L.; McLuckey, S.A. Anal. Chem. 1990, 62, 1284-
1295.
36. Schwartz, J.C.; Syka, J.E.P.; Jardine, I. J. Am. Soc. Mass Spectrom. 1991,
2, 198-204.
37. Williams, J.D.; Cox, K.; Morand, K.L.; Cooks, R.G.; Julian, Jr., R.K. in
Proceedings of the 39th ASMS Conference on Mass Spectrometry and
Allied Topics Nashville, TN, 1991, 1481-1482.
38. Yates, N.A.; Griffin, T.P.; Yost, R.A. in Proceedings of the 41st ASMS
Conference on Mass Spectrometry and Allied Topics San Francisco, CA,
1992, 444a-444b.
39. Soni, M.H.; Cooks, R.G. Anal. Chem. 1994, 66, 2488-2496.
40. Todd, J.F.J; Penmann, A.D.; Thorner, D.A.; Smith, R.D. in Proceedings of
the 38th ASMS Conference on Mass Spectrometry and Allied Topics
Tuscon, AZ, 1990, 532-533.
41. Johnson, J.V.; Pedder, R.E.; Yost, R.A. Int. J. Mass Spectrom. Ion
Processes 1991, 106, 197-212.


Figure 4-17:
Photodissociation set-up for the continuous wave-laser. The beam selector allows 10.6 //m
radiation to pass directly through to the gold plated reflecting mirror, while the output of the
Helium-Neon laser (632.8) is reflected off the front surface of the beam selector.


Figure 14continued
242


120
For photodissociation to occur, the rate of photon absorption must be
greater than that of collisional and radiative deactivation. Thus for pulsed-valve
experiments, long ion storage times (several seconds) were required before
triggering the laser so that the He buffer gas initially used to efficiently trap and
damp the diglyme ions needed for the formation of the protonated (m/z 135)
species could be pumped away. Storage efficiency measurements show an
approximate loss of 2% of the original ion signal after 70 ms of storage. The
photodissociation efficiency as a function of the laser delay time is shown in figure
2-18. After a 2 s delay between the He gas pulse and the laser trigger, the
photodissociation efficiency levels off at approximately 90%.146,147
Wavelength Dependence/Infrared Spectroscopy of Gas-Phase Ions
In figure 2-19, the wavelength dependence of the IRMPD spectrum of
protonated diglyme is shown. The wavelength was varied from 933 to 953 cm"1.
The spectrum was normalized to an irradiation energy of 0.252 J. The reaction
channel was not found to be dependent of the laser wavelength used; only
photodissociation efficiency was affected by varying the laser wavelength. The
maximum photodissociation efficiency was found at 944 cm"1. A higher degree
of photodissociation efficiency was observed in the ion trap compared to an eight-
pass ICR cell (==90% at >tmax versus =55% at ^max)147. The difference in these
values was attributed to the easier alignment of the ion trap system. The width
of the absorption peak for protonated diglyme in the ICR cell was approximately


0,745"
Center nark 2,750' dlanet


70
to a strong IR emission line (e.g., 10.59 //m). With continued use over a 36-month
period, no degradation of the ion trap mirror surfaces has been observed with the
unfocused laser beam.147
Effects of Ion Motion on Photodissociation Efficiency
The effect of resonance excitation on ion motion in the quadrupole ion trap
is a rapidly evolving area of research, encompassing both simulation and
experimental investigations.57,156'160 The original reports of resonant excitation of
ions at their secular frequency of motion (where wz = the fundamental axial
frequency and tur = the fundamental radial frequency) due to an externally
applied ac field on the endcap electrodes was reported by Paul and Fischer.8,9
It was originally thought that radial excitation (o/r) would increase ion trajectories
only in the r direction and that axial excitation (wj would increase ion trajectories
in the z direction. However, with the nonideal (stretched) quadrupole ion trap, ion
motion in the r- and z-direction are coupled, leading to a complex series of
interactions between ion motion in both the r- and z-directions.161 Today,
resonance excitation of stored ions in the quadrupole ion trap is accomplished
by applying an auxiliary ac signal to the endcap electrodes. The frequency of the
ac signal applied corresponds to a particular frequency component of ion motion;
the greater the contribution of the component frequency to ion motion, the greater
the ability of the ion to absorb power from the applied field. The larger the power


REFERENCE LIST
1. Busch, K.L; Glish, G.L; McLuckey, S.A. in Mass Spectrometry/Mass
Spectrometry: Techniques and Applications of Tandem Mass Spectrometry
VCH: New York, 1988, pp 87-90.
2. Johnson, J.V.; Yost, R.A.; Kelley, P.E.; Bradford, D.C. Anal. Chem. 1990,
62, 2162-2172.
3. McLuckey, S.A.; Van Berkel, G.J.; Goeringer, D.E.; Glish, G.L. Anal. Chem.
1994, 66, 689A-696A.
4. McLuckey, S.A.; Glish, G.L.; Van Berkel, G.J. Int. J. Mass Spectrom. Ion
Processes 1991, 106, 213-235.
5. McLuckey, S.A. J. Am. Soc. Mass Spectrom. 1992, 3, 599-614.
6. Herron, W.; Goeringer, D.E.; McLuckey, S.A. in Proceedings of the 43rd
ASMS Conference on Mass Spectrometry and Allied Topics Atlanta, GA,
1995, 410.
7. Paul, W.; Steinwedel, H. Apparatus for Separating Charged Particles of
Different Specific Charges" German Patent 944,900, 1956; United States
Patent 2,939,952 1960.
8. Paul, W.; Reinhard, H.P.; Zahn, U. 2. Phys. 1958, 152, 143-182.
9. Fischer, E. Z. Phys. 1959, 156, 1-26.
10. Dehmelt, H.G. Adv. At. Mol. Phys. 1967, 3, 53-72.
11. Major, F.G.; Dehmelt, H.G. Phys. Rev. 1968, 179, 91-107.
12. Dehmelt, H.G. Adv. At. Mol. Phys. 1969, 5, 109.
13. Dawson, P.H.; Whetten, N.R. J. Vac. Sci. Technol. 1968, 5, 11-18.
14. Dawson, P.H.; Whetten, N.R. "Three-Dimensional Mass Spectrometer and
405


62
[C3H5Br]*- + C3H5Br [C3H5]+ + Br* + C3H5Br (2-1)
After the desired cool/reaction time for a given ion population, ion isolation
was accomplished by either apex or two-step isolation.767,9 The tuning of the
necessary rf/dc voltage combinations to store only ions of a single m/z required
fine adjustments of 0.1 to 0.5 V. This meticulous procedure was required
because under conditions of no He buffer gas, the stability edges of the Mathieu
diagram become very steep. These steep edges were found to be extremely
sensitive to ion population; without a fine rf/dc voltage control, a large percentage
of the parent ion population of interest could be ejected during the isolation event.
During the mass selective instability scan (no helium buffer gas), the
optimum resonance ejection parameters were found at a qzeject=0.89 and an
amplitude of 1.5 V(0_P). These values gave peak widths slightly smaller than 0.5
mass units at full width half maximum (FWHM). This observation was consistent
with that of the high resolution mode of operation, where reduced peak widths
were observed only using resonance ejection techniques.36 Operation of the ion
trap without helium buffer gas (no resonance ejection scan), produced larger
peak widths at FWHM (>0.5 mass units) and reduced signal-to-noise (s/n) ratios
for the peaks of interest. However, ion statistics and signal reproducibility were
found to be somewhat more reliable without using the resonance ejection scan.
For verification of the various reaction pathways, notch filter ejection
experiments were performed during the laser irradiation period. Frequency
probes for protonated diglyme (m/z 135) without He buffer gas required


356
2-deoxy-D-glucose to its final product ion, the m/z 111 species [M+NH4+-3H20-
NH3], The consecutive reaction curves in figure 5-8 for the IRMPD of the
ammonium adduct of 2-deoxy-D-glucose show that only 61% of the fragment ions
produced from the photodissociation process are structurally significant ions; the
remaining 39% represent the direct dissociation of the ammonium adduct
complex to form the ammonium ion and corresponding 2-deoxy-D-glucose
neutral. The ammonium ion intensity curve in figure 5-8 was incomplete due to
the complex series of back reactions observed for the NH4+ ion with neutrals
present in the trapping region during the laser irradiance period (e.g., longer
irradiance times correspond to longer reaction times for the dissociated NH4+ ion).
The trace is present only to confirm the presence of this dissociation pathway; the
experimental conditions were altered in a separate experiment to efficiently trap
the m/z 18 NH4+ ion.
The bottom half of figure 5-8 shows the individual structurally diagnostic
portion of the consecutive reaction curves. It is clearly seen that a series of
complex consecutive and competitive reactions occur during the
photodissociation process. The presence of the competitive reaction curves (e.g.,
m/z 164 and m/z 165) could involve reactions where the critical energy of the two
reaction channels are equivalent.
Although the types of fragment ions produced by CID and PID in this
example were the same (which might be expected since both are low energy
processes), the time needed to produce each spectrum was radically different.


Figure 4-25:
Bovine insulin spectrum at 5 pmol///L and 5 fmol///L respectively.


258
Instrument Characterization
This section describes the initial experiments performed on the ESI/ion trap
instrument used to evaluate system performance. It begins with a brief discussion
of the first test (El mode) of the rf-only octopole, analyzer, and detector
assemblies. An introduction to high mass analysis (mass range extension) follows
with a few remarks on instrument operation and theory. From there, a discussion
on ESI operation ensues with a few brief remarks on absolute sensitivity. The
next few sections cover characterization of various instrument parameters such
as octopole rf voltage, octopole offset, ion gate lens voltage, ion isolation, and rf
injection level (qmject). The chapter concludes with an MSn study on the
neuropeptide angiotensin I and a brief dialogue on negative ion operation.
Initial system tests of the rf-only octopole, analyzer assembly and detector
assembly were performed using electron ionization (El) mass spectrometry via an
external ion source. To perform these tests, the electrospray and corresponding
Conflat adapter flange were removed from the instrument and a Rnnigan MAT
4500 El/Cl ion source and corresponding flange extension were added to the
system. The filament and lenses for the El source were controlled via the 4500
quadrupole electronics module used for operation of the rf-only octopole. Lens
assembly L3 (quadrupole entrance lens) was modified with a tube lens extension
to act as an ion gate for El operation with the ion trap.
The first mass spectrum (El mode) of perfluorotributylamine (PFTBA)
obtained with the new instrument is shown in figure 4-20. The instrument


Figure 2-8:
Effect of resonant excitation (w2=118.3 kHz, qz=0.3, 6^=0) voltage on photodissociation efficiency.
When the axial excursions of the ions exceed the width of the C02 laser beam (as indicated in the
figure inset by the solid line), photodissociation efficiency drops off dramatically, since the ions
spend a significant amount of time at their maximum excursions outside the beam width. Error
bars are defined as the standard deviation of the mean.


46
absorbed above the dissociation threshold to the lifetime of the excited species,
and estimation of the energy distributions within the product ions from the IRMPD
process.


Internal Cleavage at Either Pro or His
Hydrogen Migration (a-Carbon)
Ion Type y4
OH
319


Figure 4-2:
Top view of the vacuum manifold design for the ESI/ion trap instrument. The center line of the
manifold was accurate to within 0.010" from end to end.


303
Another observation of these experiments was the increase in qinject needed
for higher ion translational energies during the ion injection process. For
translational energies in the 10 V range, qjnject optimum was 0.2 to 0.3 for both the
singly and doubly-charged ions of MRFA.
Ion Isolation
A variety of techniques has been used successfully to perform ion isolation
studies on the quadrupole ion trap. Some of these include apex14,76,77, two-step78'
80, SWIFT (stored waveform inverse Fourier transform)38,39, random noise82,
forward-reverse scans74,273, and FNF (filtered noise field)81 techniques. The apex
and two-step isolation routines are effective, but are limited in mass range to
approximately m/z 600 or lower. This limitation is due to the high dc voltages
required to be applied to the ring electrode (physical constraints of the ion trap),
the cost of high voltage equipment, and stability diagram considerations for
higher masses. In this section, a brief discussion of the forward-reverse scan
method is presented to demonstrate the high mass isolation techniques needed
during ESI/ion trap operation. Other more complex techniques such as FNF,
random noise, and SWIFT are currently under investigation in our laboratory.
The original forward-reverse scan technique, as described by Kaiser et
al.273, begins by placing a resonance ejection voltage at the frequency of interest
on the endcap electrodes. Next, the ion of interest is placed just below the
resonance ejection frequency via an rf amplitude ramp (see figure 4-31). As the


Figure 4-16: Detector assembly with wire leads attached (indicated by dashed lines). The anode lead was
shielded with a copper foil (grounded to the manifold) so as to reduce background noise in the
electrometer.


Intensity Intensity
362
(M+NHJ+
100 120 140 160
200 220
m/z
180


213
focuses on the tuning of the rf matching circuit needed for operation of the
octopole device. It is assumed the reader is familiar with basic electronics and
the principles of octopole operation and design. The procedures employed for
determination of the octopole operating frequency are based on the Standing
Wave Ratio (SWR bridge) method as developed by William J. Fies, Jr. at Finnigan
MAT Corporation (San Jose, CA).258
The rf power supply and corresponding components used to drive the rf-
only octopole were obtained from a Finnigan MAT (San Jose, CA) 4500 single
quadrupole mass spectrometer. The basic rf circuitry from the Finnigan MAT
4500 system consists of an rf amplifier with a low-pass filter to reduce harmonic
frequencies which may be present (see figure 4-6). The amplifier and low-pass
filter are designed to drive a 50 Q load for the tuned circuit. The purpose of the
rf circuit is to produce a specific rf voltage for mass analysis (for the tuned circuit)
with minimum power consumption. The circuit consists of a coil with inductance
L and specified capacitance Cr (corresponding power loss specified by Rn which
occurs almost entirely in the coil), and matching capacitor(s) Cm. Tuning of the
resonant frequency for the system is accomplished by adjusting either the
inductance L or the capacitance Cr. The circuit is considered tuned" when the
circuit is resonant at the frequency of the rf driving power and R, is matched to
the generator resistance of the power amplifier.258
For the case of the rf-only octopole designed here at the University of
Florida, the capacitance of the device is unknown. For the purpose of


17
where 0<6U<1 and n=0, 1, 2, 3 Consequently, the main secular frequency
(n=0) is calculated as 6UC1/2.23
As mentioned earlier, the fundamental driving force behind the derivation
of the field equations for the quadrupole ion trap is the fact the motions in the r-
and z-directions are independent of one another. This concept enabled Paul and
others to describe the motion of ions with a simplified mathematical treatise.55 As
a better understanding (both theoretical and experimental) of ion motion in the
quadrupole ion trap was gained, it was realized that coupled motion between the
r- and z-directions was indeed possible. In 1989, March et al. published a
theoretical derivation of this coupled motion (discussed later in chapter 2) under
resonance excitation conditions.57 This dissertation presents a novel way to
experimentally confirm the presence of coupled motion in the quadrupole ion trap,
employing the technique of photodissociation.
Ion Activation
The most common result of ion activation of any polyatomic species in the
gas phase is unimolecular dissociation. Unimolecular dissociation in trapping
instruments typically occurs from a stable ion which has been "activated" and
made unstable. The fragments observed from this dissociation depend on the
structure of the parent ion and can thus provide structural information on the


28
addition, scan table times could be adjusted up to 1 s, to facilitate various
photodissociation experiments.
Electrosprav Ionization
Perhaps the most successful liquid chromatography/mass spectrometry
(LC/MS) interface to date is that of the electrospray (ESI) ion source. Over the
last seven to eight years, the volumes of research in the field have produced an
abundance of methodologies and applications for the identification and
quantitation of biological species. Some of the reasons for the rapid growth rate
of ESI include the speed with which commercial instrumentation was developed,
the ability to couple ESI with microscale capillary separation techniques, and the
ability to perform tandem mass spectrometry on multiply charged species.84
The first use of electrospray as an ionization technique for biological
species was reported over 25 years ago by Dole and coworkers who defined
many of the operational parameters used today.85,86 Doles original detection
scheme involved ion mobility and ion retardation methods, since an appropriate
mass spectrometer was not available. Dole was also the first to report the
multiple charging effect seen with large biological species.87 The first reports of
ESI combined with mass spectrometry were by Fenn and Aleksandrov et al. in
1984.88,89 Aleksandrovs group was also the first to interface an LC to an
electrospray ionization source connected to a magnetic sector mass
spectrometer.90


59
too much kinetic energy from the rf field for coherent ejection during the analytical
____ 153,154
scan.
For many of the experiments discussed in this chapter, the cool time and
the reaction time for the formation of the [M+H]+ ion for diglyme and the crown
ethers was combined (since during any reaction time, collisional cooling can also
occur). In many instances, the cool/reaction times were on the order of 400 ms.
At these extended cool times, ion stability (ion-molecule reactions) can become
problematic. For the case of an even-electron species, such as protonated
diglyme, [M+H]+, at m/z 135, storage times of over one minute showed no
appreciable dissociation/reaction of the parent ion species. However, for an odd-
electron species like ionized allyl bromide ([M+] m/z 120 and m/z 122), long
storage/cool times can produce unwanted ion-molecule reaction products. As
seen in figure 2-4, an appreciable rate of reaction (2.5 s'1) of m/z 120 (the M+ of
allyl bromide) with the neutral allyl bromide present in the trap produced a
decrease in the parent ion signal intensity over a 100 ms time period. The
reaction of the parent species with the neutral allyl bromide produced the allyl
carbocation at m/z 41. The mechanism of this reaction was thought to occur
through a collision between the [M+ ] ions at m/z 120 and 122 with the neutral
species; this energetic collision (which has a substantially higher average kinetic
energy than that observed when He buffer gas is present) initiates charge-site
migration to the allyl portion of the molecule and loss of the bromine radical as
shown in equation 2-1.


Figure 5-19:
Structure of the macrolide antibiotic erythromycin showing the lactone ring (aglycone moiety), the
amino sugar D-desosamine and nonamino sugar L-cladinose.


30
collisional activation106-117, and photodissociation.111,112 Several recent reviews on
ESI coupled with a variety of applications can be found in the literature.84,113-115
The purpose of this section is to provide the reader with a general
understanding of the electrospray process and examine some of the more recent
advances in the field. A brief discussion of charged droplet formation is followed
by a section on the chemistry of multiply charged ions to give the reader
knowledge of the basic physical principles of ESI.
Basic Principles of Ion Formation
The production of ions in electrospray mass spectrometry is comprised of
two steps: the production of highly charged droplets with their dispersal at
atmospheric pressure and the evaporation of these droplets to produce multiply
charged ions.115 The production of highly charged droplets begins with a small
flow of liquid through a simple metal capillary (stainless steel needle) which
operates at an elevated electric potential relative to a counter electrode. The
potential of this electric field on the capillary is typically between 3 and 6 kV
relative to the counter electrode placed about 0.3 to 2 cm away. The counter
electrode has an orifice where charged clusters, ions, or droplets are passed into
the mass spectrometer. Charge accumulations occur at the liquid surface due to
the application of the electric field. Therefore, flow rate, solution resistivity, and
surface tension are important variables in droplet production. The bias of the
capillary needle relative to the counter electrode can be selected to produce


Figure 4-8:
Block diagram of the SWR bridge test set-up. The Stanford Research Systems function generator
produces the frequency sweep and marker TTL signals used to determine the operating frequency
of the rf-only octopole.


428
promotion within Finnigan MAT to Applications Development Chemist (working
with the ion trap team). At exactly 5:04 P.M. on October 17, 1989, while
discussing the merits of going back to graduate school with Bob Finnigan, Jim
had the dubious distinction of hiding under Bobs desk during the Loma Prieta
earthquake. Despite the words "Rick Yost" being uttered just before the quake,
Jim and Tracy left California for graduate school (Jim in the Chemistry Department
and Tracy in the Journalism Department) at the University of Florida in August of
1990.
While in graduate school, Jim was awarded a departmental teaching
award, an Analytical Division Fellowship from the American Chemical Society, and
a Dissertation Fellowship From the College of Liberal Arts and Sciences at the
University of Florida. Upon graduation, Jim will work with Dr. Scott McLuckey as
a postdoctoral fellow at Oak Ridge National Laboratories in Oak Ridge,
Tennessee.


386
571 for ApC. No sodium adducts were observed in the spectra. The negative
charge is located on the phosphodiester bridge, as shown in figure 5-17. The
maximum number of negative charges attached to any oligonucleotide is equal
to the number of phosphodiester bridges present, plus any phosphate groups
attached to the free 3 or 5 positions on the ribose sugar backbone. For a DNA
tetramer (ApApApA) run on the electrospray ion trap, the highest charge state
observed was that of (M-3H)'3, indicating the presence of the three
phosphodiester bridges involved in the ionization process.112,308
The photodissociation spectrum, structure and cleavage points of adenyl
adenosine and adenyl cytidine is shown in figure 5-18. The photodissociation
spectra represent 5 laser shots from the pulsed C02 laser. In the top spectrum
is shown the fragmentation of the anion of ApA, with two characteristic peaks at
m/z 134 and m/z 329. The peak at m/z 134 represents loss of the adenine base
from either position one or position two; therefore, no subscript for the Bn(A) peak
can be assigned. The second peak at m/z 329 indicated direct cleavage of the
phosphodiester bond (PO), to form either a C,' ion (charge retention on the 5
side) or a XT ion (charge retention on the 3 side).
The fragmentation of the anion ApC is shown in the bottom portion of
figure 5-18. As with the previous example, cleavage of the adenine base is
observed and the fragment at m/z 134 can be assigned to the B/ peak. The
second peak observed at m/z 329 represents cleavage of the phosphodiester
bond to form the C/ peak containing the ribose sugar and cytosine base.


Tube Lens Extension
-3.000"
0.560"
R0.050"
1
Thread to
appropriate size
All units are in inches
Material: Stainless Steel
Figure 14 continued


36
the same calculation for the entire series of multiply charged ions. The
accuracies which have been reported to date for proteins and other biopolymers
over 100 kDa (molecular weight) are better than 0.005%M One of the best mass
accuracy measurements to date was recorded for myoglobin, with an observed
error of less than 1 ppm obtain using a FTICR mass spectrometer.124 A new
method of molecular weight determination developed by Hagen and Monning
uses a multiplicative correlation algorithm for processing charge distribution
data.125 The ability of the technique to accurately determine molecular weight
increases as the (M+H)+ signal is spread out over larger and larger charge state
distributions.
Photodissociation
Photo-induced dissociation is the next most frequently used method for
activation of polyatomic ions after collisional activation. The range of internal
energies present after the photon absorption process is much narrower than that
obtained with collisional energy transfer. Therefore, the usefulness of PID for the
study of ion structures is greatly enhanced. However, the reduced absorption
cross-sections observed with photodissociation (10'2 2) compared to those of
collision-induced dissociation (10 to 200 2) can limit this technique for analytical
applications. The recent availability of higher powered light sources over a wider
range of wavelengths should provide greater flexibility for photodissociation as a
routine analytical technique.1,67


7
bombardment (FAB)31, glow discharge (GD)32'33, electron and chemical ionization
(El/Cl)34, and electrospray ionization (ESI).35 The discovery of high resolution ion
trap mass spectrometry by Schwartz et al., combined with advances in high mass
analysis published previously, has enabled ion trap mass spectrometry to take a
lead role in the analysis of biomolecules.3637 The advances discussed in this
dissertation address many of the issues currently limiting ion trap mass
spectrometry in the analysis of biomolecules. Other recent advances in the field
include the use of broadband waveforms (stored waveform inverse Fourier
transforms, SWIFT) for MS/MS analysis and for mass isolation.38'39 In addition,
alternative techniques to scanning and detection of ions from those traditionally
used over the last 10 years have been developed.40"13
Several reviews published over the last few years cover many of the
aforementioned developments in greater detail.44-46 Several books encompassing
ion trap mass spectrometry as well as biological mass spectrometry, cover a
range of topics from basic instrumental principles to applications development.47-50
Theoretical Aspects of Ion Motion
A fundamental understanding of ion motion in the quadrupole ion trap is
important in evaluating various ion activation techniques (collisional activation or
surface-induced dissociation, frequently called SID). Some of these evaluation
criteria include dissociation efficiencies, understanding ion-neutral collisions, and
collisions of ions with surfaces. Therefore, a fundamental comprehension of ion


Analyzer Mounting Plate (Bottom)
Dimensions for
Mounting Rod
Holes
Number: 1
Material: Aluminum
Modifications: Anodize
All Units are in Inches
Page 4 of 5
Figure 12 continued
234


109
H | m/z 135
A
ch2CH2ch2och2ch2och3
jj nhv n=3
ch3dch2ch2ch2ch2och3
CH?OH + CH2 CH2OCH?CH7CH
2 v_.a2
H /H i
Q
i2 ch2 och3
H H
m/z 103

ch3ch=o-ch2ch2
nhv n=6
n
ch3ch=och2ch2och3
ch3 -L ccbch3
H H

CH2 CH=0 CH3 I m/z 59


421
243. von Zahn, U. "Method and Apparatus for Separating Ions of Respectively
Different Specific Ionic Charges" U.S.Pat. 3,197,633, 1965.
244. Gnther, K.G.; Freiler, H. "The Electric Mass Filter" U.S. Pat. 3,105,899,
1963.
245. Brubaker, W.M. "Auxiliary Electrodes for Quadrupole Mass Filters" U.S. Pat.
3,129,327, 1964.
246. Uthe, P.M. "Quadrupole Mass Analyzer" U.S. Pat. 3,457,404, 1969.
247. Uthe, P.M. "Quadrupole in which the Pole Electrodes Comprise Metallic
Rods Whose Mounting Surfaces Coincide with Those of the Mounting
Means" U.S. Pat. 3,553,451, 1971.
248. McGinnis P.F. "Mass Filter Electrode" U.S. Pat. 3,699,330, 1972.
249. Turner, W.R. "Mass Filter with Artifact Reducing Electrode Structure" U.S.
Pat. 3,725,700, 1973.
250. Dayton, I.F.; Shoemaker, F.C.; Mozley, R.F. Rev. Sci. Instrum. 1954, 25,
485.
251. Denison, D.R. J. Vac. Sci. Technol. 1971, 8, 266.
252. Dahl, D.A. in Proceedings of the 43rd ASMS Conference on Mass
Spectrometry and Allied Topics Atlanta, GA, 1995, 717.
253. Davis, S.C.; Wright, B. Rapid Comm. Mass Spectrom. 1990, 4, 186-197.
254. Alexander, A.J.; Dyer, E.W.; Boyd, R.K. Rapid Comm. Mass Spectrom.
1989, 3, 364-372.
255. Schoen, A.E.; Syka, J.E.P. in Proceedings of the 34th ASMS Conference
on Mass Spectrometry and Allied Topics Cincinnati, OH, 1986, 722-723.
256. Pedder, R.E. Fundamental Studies in the Quadrupole Ion Trap Mass
Spectrometer, Ph.D. Dissertation, University of Florida, Gainesville, FL,
1994.
257. Douglas, D.J.; French, J.B. J. Am. Soc. Mass Spectrom. 1992, 3, 398-408.
258. Fies, W.J. "The Standing Wave Ratio Method for Tuning RF Resonant
Circuits" internal report, Finnigan MAT corporation, 1985.


104
of m/z 103 with corresponding loss of neutral methanol as shown in figure 2-13b.
At higher irradiation energies (irradiation time = 40 ms), the dominant reaction
channel involved the loss of an acetaldehyde neutral from the m/z 103 ion shown
in figure 2-13c. The presence of a small peak at m/z 59 at lower irradiance times
suggested a competitive reaction mechanism for the formation of the product ion
species from the m/z 135 parent ion. Alternatively, when ion intensity was plotted
as a function of laser irradiance time (figure 2-14), the appearance was that of a
series of consecutive reactions.147
To determine if the reaction mechanism was competitive or consecutive,
a series of MSn experiments were conducted. In the first experiment, the
formation and mass isolation of protonated diglyme (m/z 135) ions were achieved
as described earlier. A tandem mass spectrometry experiment was then
performed with a 10 ms laser irradiance pulse followed by mass isolation of the
m/z 103 product ion. A series of MS3 experiments on the m/z 103 ion performed
by varying the laser irradiance times from 1 to 80 ms (increased energy input)
produced the ion growth curve for the m/z 59 product ion shown in figure 2-14.
To verify the exclusive formation of the m/z 59 product ion from the m/z 103
parent, a second tandem mass spectrometry experiment was performed on the
protonated diglyme parent ion (m/z 135) with concurrent notch-filter ejection of
the m/z 103 product ion. As before, the laser irradiance time was varied from 1
to 80 ms. Results from this experiment produced a decrease in the protonated
diglyme parent ion, but no formation of the m/z 59 product ion. These


145
eine = (cosn6 + isinn6) = (cose + isin6)n
= (¡j)(cos0)"-k (i sin 0)k
where n=4 for the case of an octopole and:
(3-27)
n!
k!(n k)
for k = 0,1,2,... s n = 0,1,2,
(3-28)
By substituting:
cos0 = and sin0 = ^ (3-29)
r r
(the definitions for cos and sin) into Moivres formula and applying the binomial
expansion (see equation 3-27), the equation for the field potential in rectangular
coordinates is obtained:219
4> = -^-0(x4 6x2y2 y4)
2r0
(3-30)
Electric Field Strength Calculations
To determine the strength of the electric field vector E and its component
vectors Ex (x-axis direction) and Ey (y-axis direction) for an octopolar field, the
vector equation may be written as:
E = ux Ex + Uv Ev = ux Uv-^- = grad
y y 0X 1 dy
(3-31)


TABLE OF CONTENTS
ACKNOWLEDGMENTS ¡
ABSTRACT vi
CHAPTERS
1 INTRODUCTION 1
The Quadrupole Ion Trap Mass Spectrometer 3
History 3
Theoretical Aspects of Ion Motion 7
Ion Activation 17
General Operation 21
Electrospray Ionization 28
Basic Principles of Ion Formation 30
Molecular Weight Determination 35
Photodissociation 36
Infrared Multiple Photon Dissociation (IRMPD) 38
The Photon Absorption Process 41
2 FUNDAMENTAL INVESTIGATIONS OF IRMPD IN THE
QUADRUPOLE ION TRAP 47
Instrumentation 47
Experimental Design 48
Ion Trap Operation without Helium Buffer Gas 57
The Multi-Pass Ring Electrode 63
Effects of Ion Motion on Photodissociation
Efficiency 70
Dipolar Excitation 72
Quadrupolar Excitation 82
The Photon Absorption Process 93
IRMPD Kinetics 97
Consecutive Reactions 100
Buffer Gas Effects 116
v


Cylindrical
Electrode
Glass
Capillary
Skimmer
Ion Source
Lenses
Baffle
I
)
Turbo
Pump
Analyzer
Turbo
Pump


Figure 3-5: Top and side views of the Delrin supports used in the construction of the rf-only octopole. The
outer diameter of the supports fit the baffle wall, ESI interface and the analyzer/octopole support
assembly.


Table 4-4. Instrumental parameters for ESI spectra of apomyoglobln and cytochrome c.
Instrument Parameter
Apomyoglobin Value (units
variable)
Cytochrome C
(units variable)
Octopole RF
0.3 V detected RF
0.3 V detected RF
Octopole Offset
-3.0 V
-3.0 V
Ion Trap Offset
-5.0 V
-5.0 V
He Pressure
1.0x104 torr (uncorr.)
1.0x104 torr (uncorr.)
Ion Gate
+ 75 V
+68 V
RF Level (qin|ect)
35 (1 ms); 55 (1 ms)
35 (1 ms); 60 (1 ms)
Scan Range (qeject)
to m/z 1750 (134 kHz)
to m/z 1875 (125 kHz)
Resonance Ejection
Amplitude
7.5 V0 P
8.2 V0P
Ion Injection Time
2 ms
2 ms
Dynode
-10 kV
-10 kV
Electron Multiplier
-1300 V
-1300 V
279


Modified ring electrode for multipass IRMPD experiments. Mirror positions and the eight laser
passes across the radial plane of the ring electrode, along with approximate photon density int eh
radial plane of the ring electrode, are shown. The positions of mirrors A and B determine the
number of laser transversals across the radial plane of the ring electrode.
Figure 2-6:


RF
Laser
TTL 1
Pulsed
Valve
TTL 2
Resonant
Excitation
Frequency


2
tandem (MS/MS") mass spectrometry, there has only been a limited effort to
investigate alternative techniques for ion activation in trapping instruments
(including ion cyclotron resonance and quadrupole ion trap mass spectrometers).
This dissertation presents an alternative method for the activation of
polyatomic ions in the gas phase, that of photon absorption or what is frequently
called photo-induced dissociation (PID). Photodissociation has been used
extensively by physical chemists to study fundamental properties of gas-phase
ions. The combination of mass spectrometry (employing both ion trap and ion
cyclotron resonance instruments) and photodissociation has been used
successfully to investigate the chemical kinetics, reactivity, and spectroscopy of
various ionic species. The long storage times and instrumental configuration of
trapping instruments are ideally suited for photodissociation experiments. Some
advantages of trapping instruments include the measurement of photon-induced
ion decay as a function of laser irradiance time, the use of the multiphoton
absorption processes to study fragmentation, and the use of the
photodissociation spectrum as a fingerprint for determination of isomeric ion
structures.
This introductory chapter begins with the relevant history of the quadrupole
ion trap mass spectrometer, followed by a brief discussion of ion motion, the
principles of ion activation, and general QITMS operation procedures. Since the
technique of electrospray ionization (ESI) is used to generate gas-phase ions for
the biological studies presented in this dissertation, an introduction to the basic


Oligonucleotide Fragmentation Nomenclature
HO
N
w3 x3 y3 z3
\
o
r"o '

/ /
/
A
o
w2 x2 y2 z2
/
t
O
-o
w, x, y!
/ / / /
\
O
-o
\
OH
a.
/ / r r r r r r r r r
b] Cj dj a2 b2 c2 d2 a3 b3 c3 d3
385


426
322. Cerny, R.L; MacMillan, D.K.; Gross, M.L.; Mallams, A.K.; Pramanik, B.N.
J. Am. Soc. Mass Spectrom. 1994, 5, 152-158.


Figure 3-16: Ion entry dependence (at 0.5 and 1.0 mm) in an rf-only quadrupole, m/z 100, V0P= 100 V, ion entry
0, io= 1.2 MHz. (b) Ion entry dependence (at 0.5 and 1.0 mm) in an rf-only octopole, Remaining
conditions as in (a). Adapted from reference 253.


133
H is the magnetic field strength vector and J is the current per unit area in the
medium. The del operator (V) was defined in vector space as:
V = k (3-4)
3x 8y 3z
where T J and k are unit vectors along the x, y, and z, axes respectively.232
The divergence, curl and time derivative of the Maxwell equations 3-3 can
be computed for an electromagnetic field provided that the charge density
p(x,y,z,t) and the current density J(x,y,z,t) are known for all space and time. In
addition, the values E, D, B, and H of must satisfy equation 3-3 and cannot be
chosen arbitrarily. The Maxwell equations can be rewritten in terms of the vector
quantities B and E from the Lorentz force law in equation 3-2, providing that some
basic assumptions are made about the material from which the pole pieces are
constructed (linear, homogenous, isotropic materials) and the medium (vacuum)
in which the ion moves. With these assumptions the Maxwell equations
become:233,234
V (eE ) = p VxE = ^
3t
(3-5)
V B = 0 vxlj + M)
ii at
where e is the permitivity of free space and /# is the permeability of free space.
For purposes of computing the E and B fields acting on an ion in equation 3-5,
the associated ion charge(s) and current density can be neglected as the ion


Intensity
103
miz 59
C3H.O+
miz 103

c5h,a+
miz 135

c6h,A+
11111II1111111
y! 11111II11 rl111111111
Y n | i ri i | i i i i | i iTi |
30 40 50 60 70 80 90 100 110 120 130 140
miz
o


40
dissociation was observed due to collisional deactivation of the vibrationally
excited proton-bound dimer.20'21 At the low pressures used in these experiments,
the dissociation rate constant kD was related to the phenomenologically defined
cross section aD and the photon flux 0 by the following equation:
kD= o4> d-18)
The highest absorption cross section for 2-propanol was found to be at a
wavenumber of 944 cm'1, with the corresponding absorption of 10 photons. The
dissociation rate constant kD was determined to be 2.2 s'1, assuming first order
dependence on photon flux.
Photodissociation experiments for the proton-bound dimer of 2-propanol
(m/z 121) showed three different photoreaction channels open for the IRMPD
process. March and Hughes give a detailed description for verification of the
various reaction pathways, the photodissociation of the various isotopic analogues
of 2-propanol, and the ion relaxation processes involved.22,23
The same experimental apparatus has also been used to investigate the
gas-phase ion chemistry of ethanethiol, 1- and 2-propanethiol, and 1-
hydroxyethanethiol.135,136 The collisionally-cooled, proton-bound dimers of
ethanethiol, 1-propanethiol, and 2-propanethiol were unaffected by laser
irradiation at 944 cm'1. However, the proton-bound dimer of 2-hydroxyethanethiol
was thought to contain a SH+S linkage which was shown to be transparent
at the same wavelength. Isomer differentiation by multiphoton dissociation of the
proton-bound dimer of propanone (m/z 117) and protonated diacetone alcohol


68
approximate photon density (assuming constant intensity across the attenuated
beam width) observed in the radial plane of the ring electrode can be seen in
figure 2-6. The most critical adjustment of the mirror system was the separation
of the centers of curvature of the mirrors labeled A and B. This separation
distance determines the number of beam transversals across the ring electrode:
4,8,12, or any other multiple of 4. The mirrors were mounted on the ring
electrode such that the centers of curvature of Mirrors A and B were on the front
surface of mirror C, and the center of curvature of mirror C was halfway between
mirrors A and B.155 This method of mirror alignment establishes a system of
conjugate foci on the reflecting surfaces of mirrors A,B, and C. Consequently,
light leaving the surface of mirror A is focused by mirror C on the surface of mirror
B, and the light leaving mirror B is then focused back to the original point on
mirror A. Similarly, any light leaving mirror C and going to either mirror A or B is
focused back to mirror C at some point offset from the original one15S (see figure
2-6).
This technique for extending the optical pathlength in restricted volumes
has many advantages over previous designs which incorporate flat mirror systems
or a spherical mirror and a truncated prism scheme.155 One advantage is the
ease in making adjustments since all tolerances but the horizontal angles of
mirrors A and B are usually quite large. Other inaccuracies which are introduced
are small and not cumulative. Another advantage is that light losses on mirrors
surfaces are kept to a minimum. Since there are only two reflections (at normal


Displacement (mm) |Z|
182
0,- r0 | (a)
¡ 300 V
S f
i
2-0 h I


Amplified RF
Remote RF
Octopole Offset
Detected RF
Octopole
4500 QEM
TSQ 46 Lens
Voltage Supply
228


425
307. Fenn, J.B.; Mann, M.; Meng, C.K.; Wong, S.F.; Whitehouse, C.M. Science
1989, 264, 64.
308. McLuckey, S.A.; Van Berkel, G.J.; Glish, G.L. J. Am. Soc. Mass Spectrom.
1992, 3, 60-70.
309. Edmonds, C.G.; Barinaga, C.J.; Loo, J.A.; Udseth, H.R.; Smith, R.D. in
Proceedings of the 37th ASMS Conference on Mass Spectrometry and
Allied Topics Miami Beach, FL, 1989, 844.
310. Corcoran, J.W.; Hahn, F.E., eds. Antibiotics III: Mechanism of Action of
Antimicrobial and Antitumor Agents Springer-Verlag: New York, 1975,396-
479.
311. Roberts, G.D.; Carr, S.A.; Christensen, S.B. in Proceedings of the 35th
ASMS Conference on Mass Spectrometry and Allied Topics Denver, CO,
1987, 933.
312. David, L; Scanzi, E.; Fraisse, D.; Tabet, J.C. Tetrahedron 1982, 38, 1619.
313. Barbalas, M.P.; McLafferty, F.W. Occolowitz, J.L Biomed. Mass Spectrom.
1982, 10, 258.
314. Cooper, R.; Unger, S.E. J. Antibiot. 1985, 38, 24.
315. Siegel, M.M.; McGahren, W.J.; Tomer, K.B.; Chang, T.T. Biomed. Environ.
Mass Spectrom. 1987, 14, 29.
316. Holzman, G.; Ostwald, U.; Nickel, P.; Haack, H.J.; Widjaja, H.; Arduny, U.
Biomed. Mass Spectrom. 1985, 12, 659.
317. Nelson, C.C.; McCloskey, J.A.; Isono, K. in Proceedings of the 37th ASMS
Conference on Mass Spectrometry and Allied Topics Miami, FL, 1989, 724.
318. Curtis, J.M.; Bradley, B.; Derrick, P.J.; Sheil, M.M. Org. Mass Spectrom.
1992, 27, 502.
319. Vincenti, M.; Guglielmetti, G.; Andriollo, N.; Cassani, G. Biomed. Environ.
Mass Spectrom. 1990, 19, 240.
320. Edwards, D.M.F.; Selva, E.; Stella, S.; Zerilli, L.F.; Gallo, G.G. Biol. Mass
Spectrom. 1992, 21, 51.
321. Straub, R.; Linder, M.; Voyksner, R.D. Anal. Chem. 1994, 66, 3651-3658.


412
103. Hemling, M.E.; Conboy, J.J.; Bean, M.F.; Mentzer, M.; Carr, S.A. J. Am.
Soc. Mass Spectrom. 1994, 5, 434-442.
104. Ganem, B.; Li, Y.T.; Henion, J.D. J.Am. Chem. Soc. 1991, 113, 6294-6296.
105. Light-Wahl, K.J.; Schwartz, B.L.; Smith, R.D. J.Am. Chem. Soc. 1994, 116,
5271-5278.
106. Busman, M.; Rockwood, A.L; Smith, R.D. J. Phys. Chem. 1992, 96, 2397-
2400.
107. Senko, M.W.; Speir, J.P.; McLaferty, F.W. Anal. Chem. 1994, 66, 2801-
2808.
108. Ishikawa, K.; Nishimura, T.; Koga, Y.; Niwa, Y. Rapid Comm. Mass
Spectrom. 1994, 8, 933-938.
109. Kilby, G.W.; Sheil, M.M. Org. Mass Spectrom. 1993, 28, 1417-1423.
110. Tang, X.J.; Thibault, P.; Boyd, R.K. Anal. Chem. 1993, 65, 2824-2834.
111. Little, D.P.; Speir, P.J.; Senko, M.W.; OConnor, P.B.; McLafferty, F.W. Anal.
Chem. 1994, 66, 2809-2815.
112. Stephenson, Jr., J.L.; Booth, M.M.; Boue, S.M.; Eyler, J.R.; Yost, R.A. in
Biological and Biotechnical Applications of ESI-MS American Chemical
Society: Washington, D.C. 1995 in press.
113. Siuzdak, G. Proc. Natl. Acad. Sci. 1994, 91. 11290-11297.
114. Senko, M.W.; McLafferty, F.W. Annu. Rev. Biophys. Biomol. Struct. 1994,
23, 763-785.
115. Smith, R.D.; Loo, J.A.; Edmonds, C.G.; Barinaga, C.J.; Udseth, H.R. Anal.
Chem. 1990, 62, 882-899.
116. Rayleigh, J.W.S. S. Philos. Mag. 1882, 14, 184.
117. Muller, E.W. Phys. Rev. 1956, 102, 618-624.
118. Iribarne, J.V.; Thomson, B.A. J. Chem. Phys. 1976, 64, 2287-2284.
Thomson, B.A.; Iribarne, J.V. J. Chem. Phys. 1979, 71, 4451-5563.
119.


Figure 4-33: The MS/MS spectrum of the +3 charge state of the peptide angiotensin I. Observed fragmentation
includes both doubly and singly charged b and y ions, singly charged a ions, and some sequence
specific internal fragment ions (due to the presence of proline).


37
The process of photodissociation for a positive ion can be described by the
following equation:
nhv
A+ A** P + N (1-17)
relaxation dissociation
where A+ is the ion of interest, n is the number of photons absorbed, hv is the
photon energy, A+* is the excited state, and P+ represents the product ion (with
loss of neutral N). For photodissociation to occur several prerequisites must be
met. The most important criteria include the absorption of photons with energy
hv, the existence of excited states above the dissociation threshold, a slow
relaxation rate compared to light absorption (multiphoton processes), and
dissociation rates which are fast on the time scale of the type of mass
spectrometer employed.167
The information obtained from a photodissociation experiment can address
a variety of gas-phase chemistry issues. One of the most important issues is the
difference observed in fragmentation spectra between PID and CID. The narrow
well defined energy transfer distribution step in PID typically leads to the
dissociation process via the fragmentation pathway with the lowest activation
energy (especially for visible and infrared wavelengths).68,126127
In addition, wavelength-dependent spectra can be obtained as long as the
internal energy of the ion population is above the dissociation threshold for the
wavelength of interest. The photodissociation spectrum can then be compared
with the typical absorption spectrum of the neutral molecules, provided that light


Figure 4-36:
MS3 spectrum of the b5 ion from angiotensin I. The predominant ions in the spectrum are the next
lowest b ion in the series (b4) and the corresponding a ions which arise from the loss of carbon
monoxide from the acylium portion of the b ion.


Intensity (Arbitrary Units)
302
m/z


4
that a combination of both rf and dc voltages applied to the ring electrode
produced trapping conditions that were favorable for mass-selective storage.14
The storage conditions used in these experiments were defined by the upper and
lower apices of the Mathieu stability diagram.14
The combination of a quadrupole ion trap (used as an ion source) and a
quadrupole mass filter was used extensively by Todd, Lawson, and Bonner for the
analysis of ejected ions from the ion trap, the general characterization/behavior
of trapped ions, chemical ionization, and ion molecule kinetics.15-17 This hybrid
instrument, called the QUISTOR (or QUadrupole Ion STORe), was operated with
the endcaps held at ground potential and a combination of rf and dc voltages
applied to the ring electrode. An electron gate was used to pulse electrons into
the center volume of the ion trap where ionization of the neutral sample gas(es)
occurred. The detection event consisted of a dc pulse applied to one or both
endcap electrodes, which extracted ions out towards the quadrupole mass filter.
The late 1970s also saw the development of mass-selective isolation by Fulford
and March.18 Here, the ion of interest was moved to the apex point in the stability
diagram where all ions of smaller m/z were unstable in the axial z-direction and
all ions of higher m/z were unstable in the radial r-direction. In 1979, Fulford et
al. performed the first experiments that involved resonance excitation at a given
ion populations unique frequency of motion. This caused either ion ejection from
the trap or collision-induced dissociation of the ions of interest.19


156
octopole (or any rf-only device), the most critical adjustment is ensuring the pole
pieces (in the xy plane) are all of the same length and no one end of any pole
piece protrudes excessively from the xy plane. This requirement is necessary to
minimize fringing field effects which can introduce forced motion in the z-direction.
Otherwise, considerations such as r0 requirements and consistencies in electrode
shape are not as critical due to the operation of the device in the rf-only mode
(i.e., operation as an ion transmission device as opposed to a mass filter).
Hyperbolic shaped rods for the octopole assembly can be difficult to
machine and design as well as to mount accurately. To build the octopole device
in a timely and accurate configuration, round rods can be used in place of
hyperbolic rods to produce an octopolar field comparable to that generated by
a pure hyperbolic contour surface. To minimize the distortions generated by
round rods, the relationship between (the rod radius) and r0 (the inscribed
radius) is:220
1 = 0.37 (3-55)
r
The original relationship for ajr0 was developed for a quadrupole analyzer by
Dayton et al.250 The accuracy of the technique used to calculate the ratio was
251
later refined by Denison.


82
photon density is highest (see solid line figure 2-8). This displacement results in
the observed sharp decrease in photodissociation efficiency for a dipolar
excitation amplitude of 15 mV.167
The unique advantage of the multi-pass ring electrode system (as shown
above) is its sensitivity in detecting the presence of axial excitation. By exciting
ions in the radial or r-direction (using quadrupolar excitation), the multi-pass ring
electrode system should be able to detect the presence of coupled motion in the
z-direction. However, this system would not be useful in detecting coupled ion
motion in the r-direction since there is a large photon density in the radial plane
of the ring electrode, which would not result in a drop-off in photodissociation
efficiency if the ions were excited radially. In the next section, a series of
experiments is described in which quadrupolar excitation/photodissociation
experiments are used to detect axial excitation (z-direction) when a given ion
population is excited only in the radial or r-direction.167
Quadrupolar Excitation
Over the last several years, the use of quadrupolar excitation as an
experimental tool has drawn increasing attention.166,168,169 The basic premise of
quadrupolar excitation is the application of the same auxiliary ac potential (in-
phase) to both endcap electrodes. The application of this in-phase ac signal
creates a symmetric field similar to the quadrupolar rf trapping field on the ring
electrode. Therefore, quadrupolar excitation does not favor either radial (r-


401
when ions pass from the high pressure ESI source region (multiple collision
conditions) into the octopole. The octopole is placed as close as possible to the
exit of the ESI source (skimmer cone region) where dispersion angles of the ions
are minimized. The flat-bottomed radial potential well for the octopole illustrates
another important advantage of the device. In the case of a wide range of m/z
ions, the flat bottom of the radial potential well allows the efficient transfer of these
ions all at the same energy, eliminating mass discrimination effects observed with
traditional dc lens systems. Another important advantage of the rf-only octopole
arises from the nonlinear fields associated with the device. These nonlinear fields
allow for the exit of a uniformly sized ion beam from the device. Therefore as ions
traverse the octopole, their frequency of oscillation is dependent on the ion entry
angle, the ion off-axis distance, and rf voltage, leading to a uniform ion beam
image independent of rf-only octopole parameters.
Also described in this section is a novel approach to the assembly of an
rf multipole device. The technique developed here at the University of Florida is
simple, inexpensive, and extremely accurate. The technique minimizes errors in
the r0 distance for the length of the device and eliminates errors associated with
fringing fields due to rod end misalignment (ends not in the same plane).
The most critical step in the process was the design and assembly of the
entire ESI/ion trap instrument. Great care was taken from the table top design to
the implementation of the copper foil grounding shield for the detector anode
lead. Two photodissociation set-ups were employed, one for the continuous wave


Ion Current
500=i_
Irradiance Time in ms
106


Figure 3-10: (a) Entry angle dependence (at 0, 2, and 4) in an rf-only quadrupole for m/z 1000 with Vo p=400
V and ion entry 1 mm off axis, (b) Entry angle dependence (at 0, 2, and 4) in an rf-only
octopole, conditions same as in (a). Adapted from reference 253.


10000
8000
6000
4000
2000
0
+19
+20
+21
+22
+23
+24
+26 i
+18
+17
mk
+16
+15
+12
L +11
fnrrp-f
VJ\J
I II | I I I | I I I | I T I | II 1 | 1 I I | I 1 I
0 600 700 800 900 1000 1100 1200 1300 1400 1500
m/z
1600
274


294
injection period to gate ions through the skimmer cone region range from +50
to +100 V. Positive voltages are used to gate ions because of the rapid
expansion from the liquid phase to the gas phase (e.g., large ion dispersion
angles and multiple collision events) and the relatively low energy/velocity of the
ions as they exit the heated capillary. To control the pulse width of positive ions,
a -180 V potential is applied to the ion gate lens to prevent ion transmission
through the skimmer cone region. Ion injection time periods varied from a few
hundred microseconds to hundreds of milliseconds (for low concentration
samples).
The optimum voltage range for transmission of the positive ion even charge
states of ubiquitin was +20 V, with a +60 V and +80 V potential for optimum ion
signal of m/z 709 (+12 charge state) and m/z 1417 (+6 charge state),
respectively. The relatively small ion gate voltage optimization range for the
multiply-charged ions of ubiquitin could be attributed to the efficient focusing of
the ion beam through the skimmer cone region, and therefore the injection of ions
close to the center axis of the octopole.
RF Level/lon Injection
The one ESI/ion trap parameter that does have a profound effect on the
observed mass spectrum is the ion trap rf level during the ion injection period
(frequently termed qinject). Early studies of the ion injection process into the
quadrupole ion trap recognized that a relationship existed among the ion trap


Erythromycin
raw = 733.4
L-Cladinose
392


402
C02 laser the other for pulsed C02 laser operation. Instrumental characterization
was first performed on an El source in order to test the integrity of the
octopole/ion trap design.
Tuning and operation of the ESI/ion trap system was carried out on a
variety of biochemical compounds. The concentrations used were approximately
10 to 100 times lower than those employed by other ion trap or triple quadrupole
instruments. The absolute detection limit observed was 5 fmol/j/L for bovine
insulin. Characterization of the octopole rf amplitude confirmed what was
predicted theoretically in chapter 3, with a flat response for the various m/z range
under investigation. Other instrumental parameter investigated included the
octopole offset, the ion gate lens, ion isolation, and the rf level during the ion
injection period. The results of the ion trap rf level studies indicated that to get
a true representation of the ion signal generated for the electrospray process, the
rf level during the ion injection should be ramped so that ion injection
discrimination effects are minimized.
One important element for all ion trap experiments is that of time for
analysis or duty cycle. Since the operation of the ion trap is on the millisecond
timescale, ions have time to obtain their minimum energy configuration. This is
especially evident in the CID spectra of the triply charged ion of the peptide
angiotensin I, where rearrangement peaks driven by the presence of proline
produce internal y sequence ions (due to the highly basic nature of proline in the
gas phase).


154
Z = x iy = re "ie = r cos 6 ir sin 0
Applying the constraints of the Mathieu equation to equation 3-52 and multiplying
through by m (with the appropriate rearrangement) gives:
m-^-^ [- ma4(Z)n_1 ] = 2mq4[cos4(t to)](Z)n~1 (3-54)
dT 2
The motion of an ion in an octopolar field is defined essentially as an
undamped non-linear forced oscillation. This motion is undamped because (in
the term on the left hand side of equation 3-54) there are no additive terms for the
dZ/dt independent variable. In this case, the combined term m(dZ/dt) refers to
the inertia force while -ma4(Z)lv1 is the restoring or spring force. As an ion
increases its displacement (Z) from the center of the octopole, the stiffness" of
the spring can be defined as the first derivative of the restoring force. The spring
is referred to as "hard" if the restoring forces increase as the ion increases its
displacement from the center of the octopole. For the case of the quadrupole
analyzer where n=2, the restoring force is a linear function of the ions
displacement (Z). If n > 3 then the restoring force is non-linear (hexapole,
octopole, dodecapole, etc...) and is considered a "hard" spring. The final term on
the right hand side of equation 3-54, 2mq4[cos n(t-t0)](Z)rv1, is the applied external
force with corresponding time-varying amplitude and is also a non-linear function


100
great extent. Table 2-3 shows a summary of the rate coefficients for the
protonated species diglyme, 12-crown-4 ether, 15-crown-5 ether, and the
molecular ion of allyl bromide. As expected, the kD values observed for the C O
stretch (diglyme, 12-crown-4, 15-crown-5) were higher than the kD value obtained
for allyl bromide due to the reduced photoabsorption cross section for the C Br
stretch. Due to the linearity of kD, the rate coefficient can be expressed in terms
of the phenomenological cross section aD:
kD = oD$ (2-9)
where

Typical values of aD are 0.5-2x1CT20 cm2.130
Consecutive Reactions
Figure 2-13a shows the El mass spectrum of diglyme at a sample pressure
of 3.6x1 O'7 torr. Typically no molecular ion M+ at m/z 134 is present. Instead, a
fragment ion at m/z 89, formed by the cleavage of the carbon-carbon bond, and
a fragment ion at m/z 59, formed by either loss of formaldehyde from m/z 89 or
loss of C3H702 from the molecular ion, predominate the spectrum. This fact,
coupled with the ability to produce protonated diglyme (m/z 135) easily via ion-
molecule reactions, led to the use of the protonated molecule for IRMPD studies.
Characteristic photodissociation spectra of the [M+H]+ ion of diglyme at
944 cm-1 are shown in figures 2-13b and 2-13c. At lower irradiation energies
(irradiation time = 10 ms), the single reaction channel observed was the formation


264
quadrupole ion trap:
(4-1)
where e is the charge on the ion, Vmax is the zero-to-peak voltage of the drive rf,
r0 is the radius of the ion trap ring electrode, and q#jact is the Mathieu stability
parameter directly related to the resonant ejection ion frequency applied to the
endcap electrodes. The desired m/z^ defines the mass range extension by
determining the value of qeject in equation 4-1 and, therefore, the corresponding
6Z value where the appropriate resonant ejection frequency is generated. Various
methods exist for calculating 6Z, including recurrence relations/continued fractions
which are the most accurate.55 However, for qeject values less than 0.4, the
approximate relationship:
(4-2)
can be used. Once the value of 6eject is calculated, the applied resonant ejection
frequency is found by:
(4-3)
where 0<6U<1, Cl denotes the rf drive frequency, and n=0, 1, 2, 3... For
calculating w (the qeject frequency), n=0 and equation 4-3 then reduces to:
The application of the resonant ejection frequency (calculated from
equation 4-4) to the endcap electrodes can be understood by examining figures


Stainless Steel Mounts
Weld Lip
00,418
00,079
ctopole Rods
ctopole Rods
Attached on
the Opposite Side
00,079"
Mounting Holes
00,073"
Rod Through
Holes
00,163"
Side View (After Initial Cut)
-0,062'
00,938"
0,418" Weld Lip
*-0.100"


Figure 5-21: Fragmentation of the macrolide antibiotic erythromycin obtained with photon activation (7 laser
pulses). The fragmentation observed mimicked that of the high energy CID process where charge
remote fragmentation mechanisms are common.


420
2153-2162.
228. Bier, M.E.; Schwartz, J.C.; Zhou, J.; Taylor, D.; Syka, J.E.P.; James, M.;
Fies, W.; Stafford, G. in Proceedings of the 43rd ASMS Conference on
Mass Spectrometry and Allied Topics Atlanta, GA, 1995, 1117.
229. Miller, P.E.; Denton, B.M. Int. J. Mass Spectrom. Ion Processes 1986, 72,
223-238
230. Everdij, J.J; Huijser, A.; Verster, N.F. Rev. Sci. Instrum. 1973, 44, 721-725.
231. Hgg, A.; Szabo, I. IntJ. Mass Spectrom. Ion Processes 1986, 73, 237-275.
232. Friedman, M.H.; Yergy, A.L; Campana, J.E. Rev. Sci. Instrum. 1982, 15,
53-61.
233. Pugh, E.M.; Pugh E.W. in Principles of Electricity and Magnetism Addison-
Wesley: Reading, MA, 1960.
234. Ramo, S.; Whinnery, J.R.; Van Duzer, T. in Fields and Waves in
Communication Electronics Wiley: New York, 1984.
235. Campana, J.E. Int. J. Mass Spectrom. Ion Processes 1980, 33, 101-117.
236. Campana, J.E.; Jurs, P.C. Int. J. Mass Spectrom. Ion Processes 1980, 33,
119-137.
237. Hgg, C.; Szabo, I. Int. J. Mass Spectrom. Ion Processes 1986, 73, 277-
294.
238. Hgg, C.; Szabo, I. Int. J. Mass Spectrom. Ion Processes 1986, 73, 295-
312.
239. Ball, G.W.; Lawson, G.; Todd, J.F.J. in Dynamic Mass Spectrometry vol. 3
Price, D., ed.; Heydon: London, 1972, p. 99.
240. Hlein, W. U.S. Pat. 3,238,146 1967.
241. Dawson, P.H., ed. Quadrupole Mass Spectrometry and its Applications
Elsevier: Amsterdam, 1976.
242. von Zahn, U. Rev. Sci. Instrum. 1963, 34, 1.


72
excitation). The photodissociation efficiency (PD) for a given experiment is defined
as the fraction of the original ion population photodissociated over a given
exposure time for a specified laser irradiance:
PD=1-
(2-2)
V 'o/
where I is the signal intensity of the dissociating ion at the end of the exposure
period and l0 is the signal intensity after the same period without irradiation. This
definition of l0 corrects for any unimolecular or collision-induced dissociation that
may occur.
Dipolar Excitation
Dipolar excitation is the most common form of resonance excitation
employed today. The techniques of collision-induced dissociation, notch filtering,
and axial modulation all use commercially available dipolar excitation circuits and
have been well characterized experimentally. Part of the reason for the success
of the aforementioned techniques is due to the asymmetric nature of the dipolar
field. Since the signals applied to the endcap electrodes are 180 out of phase,
axial motion is favored over radial motion during the excitation period.57 This can
be understood by examining the relationship between ion motion and the dipolar
mode of excitation. As an ion approaches each endcap, the potential on that
endcap needs to be of the appropriate polarity to obtain maximum power
absorption. With the application of a 180 out-of-phase signal, the polarity on


Intensity Intensity
352
3000 i
2500
MS/MS Spectrum
q=0.3, Amplitude=135 mV
Time=10 ms
(M+NH^-HjO)
m/z 164
2000
1500
1000
500 m/z 129
0
(M+NH4-H20-NH3)
m/z 147
(M+NH4+-2HjO)
m/z 146
z (M-2H20-NH3)
\
1
(M+NH4+)
m/z 182
r
TT|
t
400 i
300
200
100
MS/MS/MS Spectrum
q=0.3, Amplitude=250 mV
Time=10 ms
(M+NH4-H20)+
m/z 164
(M+NH4-3HjO-NBLj)+
m/z 111
(M+NH.-HjO-NHj)
m/z 147
(M+NH4-2H20)
m/z 146
\
(M+NH4-2HjO-NHj)+
m/z 129
(M+NH4-NH3)
m/z 165
100
120
140
m/z
160
180
0


155
of the displacement (Z).219
Octopole Design Considerations
This section considers the relative design and assembly procedures used
to construct the rf-only octopole for the ESI/ion trap instrument. A discussion of
the basic considerations for round pole pieces versus hyperbolic pole pieces, an
evaluation of the effective trapping potential, and assembly procedures are
presented.
The best performance of any rf-only octopole device will be achieved when
the shape (cross section contour) of the machined electrodes exactly matches
that obtained by theory (i.e., hyperbolic). Previous methods of manufacture of
these devices have centered on using a single piece of pressed ceramic made
to the exact length specifications of the device.239,240 The inner surface of the
ceramic consists of eight inwardly protruding sections with these surfaces cut to
the exact equation of the octopole contour. The inner portion of the ceramic is
metal coated so as to produce an octopolar field by application of the appropriate
rf frequency and voltage. Various other techniques for fabricating and designing
monopoles, quadrupoles, and other multipoles developed over the past thirty
years are primarily for the design of mass filter systems (e.g., application of both
rf and dc voltages to the multipole).239'249 These designs are somewhat complex
and cumbersome due to the strict requirements needed for mass filter systems
which can obtain a mass resolution of unit mass or better. For the rf only


Figure 5-1: Evaluation of the various MS/MS techniques used with the quadrupole ion trap. The dissociation
efficiency of photodissociation, single frequency CID, and broadband excitation are plotted versus
ionization time. The ionization time is directly proportional to the trapped ion population.


Intensity (Arbitrary Units)
281
16000
12000
8000
4000
0
+4
+5
5 pmol/fiL
5 ms ion injection
+3
Wijriiyrttfi ivpiritpMj |rnq
m/z


263
Ion Trap High Mass Theorv/Qperation
Since the majority of the compounds under study in chapters 4 and 5
exceed the mass range of the standard Finnigan MAT ITMS electronics
(m/zmax=650), the mass range of the instrument was extended to approximately
m/z 2500. Any of the following techniques can be used to extend the mass range
of the instrument: (1) reducing the radius of the ring electrode; (2) reducing the
drive frequency of the main rf; and (3) applying a resonant ejection frequency to
the endcap electrodes to reduce q^ct-73-259,260 Increasing the mass range of the
ESI/ion trap by a reduction in the radius (r0) of the ring electrode requires a
redesign of both the ring electrode and the multipass optical system, which could
be very complicated. In addition, reduced size ring electrodes are more
susceptible to space charging effects at lower ion concentrations due to lower
trapping volume. The second method, involving reduction of the rf drive
frequency can significantly degrade mass resolution for the large drop in rf
frequency (e.g., below 0.6 MHz) needed to reach a mass range of at least m/z
2500. These first two methods require physical modifications to the instrument
to obtain high mass data. The third mass range extension technique, called
resonance ejection, requires only a modest change in the software to facilitate
mass range extension.73,259'260 The mass range extension method employed in this
dissertation is that of resonant ejection, where reducing the resonant ejection
frequency reduces qeject, thereby increasing the mass range. In equation 4-1 are
shown the relationships that govern the mass range extension process in the


283
and +12 (m/z 709) charge states of ubiquitin as a function of the rf voltage
applied to the octopole is shown in figure 4-26. The even charge states were
chosen to simplify the visual appearance of the various plots. In each case the
ion intensity increased sharply at 0.2 V detected rf, where all the curves began to
level off. At the maximum rf voltage (2000 V^,), there is no significant decrease
in the ion intensity of any of the charge states. This behavior can be explained
by the triangular-like extension of the lowest stability region for an octopolar field.
The value of the Mathieu parameter q4 for an octopolar field can be calculated
from the following relationship:237
16eV0_p
_ 2*2
m to r0
(4-6)
Due to coupled motion in the x-y plane of the octopole, the shape (e.g.,
boundaries) of the stability diagram depends on the ions initial entry conditions,
as described in chapter 3. Although a single stability diagram cannot be
constructed for the rf-only octopole, the stability limit for q4 (directly proportional
to rf voltage) is > 50 for ions traversing the octopole. For the +6 (m/z 1417), +8
(m/z 1063), +10 (m/z 851), and +12 (m/z 709) charge states of ubiquitin, the
values of q4 are seen in table 4-5. For the condition where 7^=2000, none of the
q4 values exceed the approximate stability limit of the octopole stability diagram.
The flat response of the ion intensity curves in figure 4-26 demonstrates the
limited amount of tuning needed for operation of the rf-only octopole ion guide.


152
d2x
dt2
[ U V cos to (t t0) ] (x 3 3xy2)
mr04
(3-46)
= -^%[U Vcoso>(t-t0)](3x2y y3) (3-47)
dt2 mr0
Extension of the Mathieu equations (discussed in chapter two for a quadrupolar
field, n=2) to that of an octopolar field described here gives the following
relationships:
+ [an 2qn cos dt2
+ [an 2qn cos w(t t0)](3x2y y3) = 0 (3-49)
with n=4 for the case of the octopole, the a and qn parameters are evaluated as:
n3eU 0 32eU
= a4 =
2mw2r02 m (3-50)
q
n
n3eV
4mco2r02
16eV
_ 2,2
m (3-51)
Equations 3-48 and 3-49 show that motion in the octopole is coupled in the
xy plane. This has two important ramifications: (1) ion trajectories for the
octopole will be much more complex than the case for a quadrupole where
motion is independent in the x- and y-direction; and (2) the size and shape of the


CHAPTER 4
ELECTROSPRAY/ION TRAP INSTRUMENTATION:
DESIGN AND OPERATION
General Overview
This chapter presents the general design and characterization of a novel
electrospray ionization/ion trap mass spectrometer for the photodissociation of
biological macromolecular ions. The first section discusses design considerations
(vacuum manifold, ESI ion source, ion injection, and analyzer/detector assembly)
associated with the electrospray/ion trap system. This is followed by a discussion
of the experiments used to characterize system performance, including an
evaluation of the rf-only octopole ion injection design, absolute sensitivity,
collision-induced dissociation, mass isolation, and negative ion operation.
Instrument Design
This section contains the design considerations necessary for the
construction and operation of an rf-only octopole ion injection system for use with
an electrospray/ion trap mass spectrometer. Also, special considerations
concerning vacuum system design for use with the IRMPD process are
addressed. Pertinent modifications to an Analytica ESI source for coupling to the
rf-only octopole ion injection system are discussed in order to understand the ion
196


335
Photodissociation efficiency seen is independent of ion population.
However, the overall dissociation efficiency is substantially higher than for either
CID technique (approximately 90% compared with 65% for the optimum single
frequency tune and 45% for the broadband excitation shown in figure 5-1). These
results can be attributed to two factors: (1) the increased photoabsorption
pathlength of the multipass ring electrode, which compensates for the reduced
photoabsorption cross-sections of organic species; and (2) the elimination of
collisions for transfer of translational energy to vibrational energy, where ion
stability can cause competition between resonance ejection and CID during an
MS/MS experiment.
The first continuous wave C02 laser photodissociation experiment on
electrosprayed ions was that of [M+K]+ and [M+H]+ ions from 18-crown-6 ether.
The photodissociation set-up has been described in chapter 4 of this dissertation.
Electrospray conditions consisted of a 5 pmol///L solution of 18-crown-6 ether in
a 50:50 methanol:water solution with 3 mM potassium chloride added for adduct
formation. Helium buffer gas pressure was 4.5x1 O'5 torr (uncorrected), which was
the pressure where the onset of photodissociation was first observed. The
spectrum shown in figure 5-2 has both the [M+H]+ and [M+K]+ ions stored in the
trap during the laser irradiance period. The two fragment ions produced arise
from direct dissociation of the adduct ion (m/z 303) to K+ (m/z 39) and the
corresponding neutral 18-crown-6, and the dissociation of the protonated parent
(m/z 265) ion to produce C7H1203+ (m/z 144).


45
Eventually, these vibrational levels merge to form what is termed the
quasicontinuum or the incoherent single-photon interaction region (II). Ions in
region II are characterized by their ability to undergo the resonant absorption
process for a given monochromatic laser frequency (although there may be some
structure within the quasicontinuum itself). If an ion can be excited through the
discrete vibrational levels in region I, then there always exists a path for the
absorption process to occur through the quasicontinuum. Polyatomic ions can
be excited into the quasicontinuum by a variety of mechanisms including thermal
excitation, collisional or particle excitation, exothermic chemical reactions, or
electronic excitation followed by internal conversion or forced multiphoton
excitation. For the biological ions used in this dissertation, no excitation methods
are needed to push ions into the quasicontinuum, since these species are
sufficiently large that they already exist in the quasicontinuum as a result of
internal thermal energy content.130
As the ion continues to absorb energy, ft eventually reaches a point where
it obtains enough energy to dissociate (region III). If randomization of this excess
internal energy above the dissociation threshold occurs at a much faster rate than
the dissociation process itself, then the traditional statistical theories associated
with unimolecular dissociation (the QET theory discussed previously), and that of
Rice-Ramsperger-Kassel-Marcus (RRKM theory, i.e. the application of transition-
state theory to unimolecular reactions) can be applied to the photodissociation
process.145 Applications of RRKM theory include relating the excess energy


366
of the raffinose and stachyose data was that of Domon and Costello.305 In figure
5-11 is shown a simple fragmentation of a model disaccharide with cleavage
points following the Domon and Costello nomenclature. Charge retention on the
reducing portion of the sugar is indicated by the fragments X (ring opening
cleavage), Y and Z. For charge retention on the nonreducing terminus of the
sugar, the fragments are designated A (ring opening cleavage), B, and C. The
superscripts in figure 5-11 indicate the exact bond cleavage positions, while the
presence of subscript numbers identify the residue number. Branches are labeled
or, p, y where a is the branch with the highest mass.305
In figure 5-12 are shown both the CID and PID (pulsed C02 laser) spectra
of the ammonium adduct of raffinose. The major product ion from the single
frequency CID process was m/z 505 [M+H]+ arising from a loss of neutral
ammonia from the adducted species. Although this ion gave no structural
information, it could be used to help confirm molecular weight. The presence of
the peaks at m/z 343 and m/z 326 indicate the loss of one of the terminal
monosaccharide groups (D-fructose from the reducing end or D-galactose from
the nonreducing end), with the peak at m/z 326 showing an additional loss of
ammonia. Assignment of these ions as either B2a or is not possible since D-
galactose and D-fructose have the same mass. However, from a biochemical
standpoint, the most labile bond in the system is the glucopyranosyl-jff-D-
fructofuranoside bond, which would indicate that the ions formed in this mass
spectrum are B type ions with charge retention on the nonreducing terminus.


CHAPTER 5
PHOTODISSOCIATION OF BIOLOGICALLY IMPORTANT MOLECULES:
PROTEINS, CARBOHYDRATES, AND OLIGONUCLEOTIDES
General Overview of Structural Elucidation
The use of tandem mass spectrometry for the structural elucidation of
biological molecules has grown at an exponential pace with the advent of the
electrospray ionization technique. The majority of the work to date has focused
on sequencing oligopeptides and phospholipids (ionized by electrospray and
FAB), employing collisional-activation as the ion activation technique of
choice.2812a5,2B6 The simplest and most economical method to obtain structural
information by electrospray mass spectrometry is that of nozzle-skimmer
dissociation as developed by Loo et al.287 This technique, which involves
dissociation of multiply-charged ions in the high pressure source region, has been
performed on molecules as large as serum albumins.288 However, this technique
is only usable for pure samples and cannot be applied to mixture analysis, since
product ions cannot be matched to their parent ions.
The presence of multiple charges on ionic species can have a significant
affect on the ability to fragment the species via CID to obtain structurally relevant
information. Unusually high fragmentation efficiencies, site-specific cleavages,
and fragmentation driven by non-basic functional groups are just a few of the
328


322
b5 fragment.
MS3 data from the fragmentation of the b9+z ion produced a series of
singly-charged higher m/z b ions, as shown in figure 4-37. It is interesting to note
little or no complementary singly-charged y ions were observed. However, there
was a series of four unknown peaks (labeled with question marks in figure 4-37),
which may be possible cyclization rearrangements driven by the presence of the
pro7 residue and the long reaction times associated with ion trap experiments.
Further MS3 experiments utilizing labeled standards would be necessary to verify
these ion fragmentations.
Attempts using MS3 to determine the possibility of a y8+2 in the y4 peak in
figure 4-33 yielded no data which could be interpreted to either confirm or deny
the presence of this ion. This may result in part (as mentioned earlier) from the
high stability of the y series ions in general.
Negative Ion Mode
The majority of ESI-MS analysis performed to date has been in the positive
ion mode. This is especially the case for peptide and protein samples, which
contain a high number of basic residues (histidine, lysine, arginine) with high
dissociation constants that produce multiple sites for protonation in acidic
solutions. Negative ion electrospray of these compounds in an acidic solution is
not practical because the pKa values of the corresponding acidic amino acids
(aspartic acid and glutamic acid) in a protein structure are 5 or less.283


Figure 4-24: ESI spectrum of bovine heart cytochrome c showing a bimodal charge state distribution indicative
of a 50:50 methanol:water and 0.1% acetic acid solution. Sample concentration was 350 fmol//yl_
at a flow rate of 3 /yL/min.


0.2-
-0.2-
-0.4-
-0.6-
-0.8
z stable
r stable
C| eject 0.908
I I
q oc RF Trapping
z Voltage
Amplitude
a oc DC Voltage
z Amplitude
0.0 0.5 1.0 1.5
q
z
267


Frequency optimization curves for protonated diglyme (with and without He buffer gas) using
dipolar excitation. The frequency was incremented every 0.1 kHz. For the case where He buffer
gas was present, a well defined optimum frequency was obtained due to the well defined
trajectories of the ions confined to the center of the ion trap.
Figure 2-5:


Figure 5-17:
Oligonucleotide fragmentation scheme as defined by McLuckey et al. (from reference 308). The
letter Bn represents the individual nucleoside bases, with position one defined form the 5 end. The
w, x, y, and z fragments have the charge retained on the 3 end while the a, b, c, and d fragments
have the charge retained on the 5 terminus.


Vacuum Connection
208


Figure 2-3:
ITMS scan function and timing diagram for a typical IRMPD experiment (figure not to scale). 1,
pre-ionization/pulsed-valve on; 2, ionization-chemical self-ionization reaction; 3, two-step mass
isolation; 4, vibrational relaxation; 5, laser on; 6, laser decay; 7, acquisition.


H
2-deoxy-D-Glucose
mw=164.0
H
1-O-Methyl-D-Glucopyranoside
mw=194.1
350


Graph of the trapping potential function (Veff versus r the ion displacement) for a quadrupole (n=2),
hexapole (n=3), and octopole (n=4). The curves represent the +5 charge state of bovine insulin
(average molecular weight 5733 g/mol) where V0 is 250 V, w is 1.659 MHz, and r0 is 2.94 mm for
each multipole device.
Figure 3-4:


Figure 2-18:
Photodissociation efficiency as a function of post pulse-valve laser delay. Photodissociation
efficiency remains constant at over 90% for delay times > 2s. Error bars are defined as the
standard deviation of the mean.


Schematic diagram of the Finnigan MAT Ion Trap Mass Spectrometer (ITMS). The drive
frequency applied to the ring electrode is 1.1 MHz. The supplementary rf generator can be used
to apply both dipolar (as illustrated in the figure) and quadrupolar excitation signals to the endcap
electrodes.
Figure 1-4:


345


418
Proceedings of the 37th ASMS Conference on Mass Spectrometry and
Allied Topics Miami, FL, 1989, 369-370.
198. Mordehai, A.V.; Hopfgartner, G.; Huggins, T.G.; Henion, J.D. Rapid
Commun. Mass Spectrom. 1992, 6, 508-516.
199. Bou, S.M.; Stephenson, J.L.; Yost, R.A. in Proceedings of the 43rd ASMS
Conference on Mass Spectrometry and Allied Topics Atlanta, GA, 1995,
408.
200. Doktycz, M.J.; Habibigoudarz, S.; McLuckey, S.A. Anal. Chem. 1994, 66,
3416-3422.
201. Bou, S.M.; Jones, J.A.; Yost, R.A. in Proceedings of the 42nd ASMS
Conference on Mass Spectrometry and Allied Topics Chicago, IL, 1994,
218.
202. Stephenson Jr., J.L.; Booth, M.M.; Shalosky, J.A.; Eyler, J.R.; Yost, R.A. Int.
J. Mass Spectrom. Ion Processes 1995, submitted.
203. Asano, K.G.; Glish, G.L.; McLuckey, S.A. in Proceedings of the 36th ASMS
Conference on Mass Spectrometry and Allied Topics San Francisco, CA,
1988, 636-637.
204. McLuckey, S.A.; Glish, G.L.; Asano, K.G. Anal. Chem. Acta. 1989, 225, 25-
35.
205. Duckworth, D.C.; Barshick, C.M.; Simth, D.H.; McLuckey, S.A. Anal. Chem.
1994, 66, 92-98.
206. McLuckey, S.A.; Glish, G.L.; Van Berkel, G.J. in Proceedings of the 38th
ASMS Conference on Mass Spectrometry and Allied Topics Tuscon, AZ,
1990, 512-513.
207. Chien, B.M.; Michael, S.M.; Lubman, D.M. Anal. Chem. 1993, 65, 1916-
1924.
208. Koppenaal, D.W.; Barinaga, C.J.; Smith, M.R. J. Anal. Atomic Spectrom.
1994, 9, 1053-1058.
209. Louris, J.N.; Brodbelt-Lustig, J.S.; Kaiser, R.E.; Cooks, R.G. in Proceedings
of the 36th ASMS Conference on Mass Spectrometry and Allied Topics San
Francisco, CA, 1988, 968-969.


309
One limitation of this technique (especially for ESI mass spectrometry) is
the possibility of induced fragmentation of the isolated product (CID). This can
occur when the ion of interest gains enough kinetic energy from the isolation
process (e.g., approaches the resonance ejection frequency too closely) that
fragmentation (CID) is induced. As seen in figure 4-32, an unknown fragment ion
occurs at m/z 1472, which is a result of this process. The amount of
fragmentation can be controlled somewhat by reducing the amplitude of the
resonant ejection voltage. However, if the voltage is reduced too far, inefficient
ion ejection at the resonance point could occur, resulting in poor isolation
efficiency. Particular attention must be paid to electrospray ions where large
numbers of charges on an ion can induce coulombic repulsion forces, reducing
the collision energy needed for the CID process.
Collision-induced Dissociation (CID)
Collision-induced dissociation (CID) has seen a great deal of development
for biological applications in mass spectrometry.94'97,108'109,114-274'278 Applications
have ranged from determination of side chain fragmentations of leucine and
isoleucine for isomer differentiation, to determination of glycopeptide structures
using LC/MS/MS in triple quadrupole instruments.27^277 Ion trap mass
spectrometry (employing CID) has also seen rapid development in the analysis
of biological materials, particularly peptides and proteins.273,274
To evaluate the MS" capabilities of the ESI/ion trap instrument, a series of


Absorbance
Wavenumbers in cm *
1.00-,
0.80-
0.60-
0.40-
0.20
I I I I j I I I I j I I I I | 1 I I I | I I I I | I I I "l~|
930 940 950 960 970 980 990
Wavenumbers in cm1
128


423
274. Cox, K.A.; Williams, J.D.; Kaiser, R.E.; Cooks, R.G. Biol Mass Spectrom.
1992, 21, 226-241.
275. Huddleston, M.J.; Beam, M.F.; Carr, S.A. Anal. Chem. 1993, 65, 877-884.
276. Johnson, R.S.; Martin, S.A.; Bieman, K.; Stults, J.T.; Watson, J.T. Anal.
Chem. 1987, 59, 2621-2625.
277. Johnson, R.S.; Martin, S.A. Bieman, K. Int. J. Mass Spectrom. Ion
Processes 1988, 86, 137-154.
278. Baldwin, M.A. Nat. Prod. Rep., 1995, 12, 33-44.
279. Reopstorff, P.; Fohlman, J. Biomed. and Environ. Mass Spectrom. 1984,11,
601.
280. Hunt, D.F.; Yates, J.R. Ill; Shabonowitz, J.; Winston, S.; Hauer, C.R. Proc.
Natl. Acad. Sci. U.S.A. 1986, 83, 6233-6237.
281. Bieman, K.; Martin, S.A. Mass Spectrom. Rev. 1987, 6, 1-75.
282. Tang, X.J.; Boyd, R.K. Rapid Commun. Mass Spectrom. 1992, 6, 651-657.
283. Loo, J.A.; Ogorzalek, R.R.; Light, K.J.; Edmonds, C.G.; Smith, R.D. Anal.
Chem. 1992, 64, 81-86.
284. Windholz, M.; Budavari, S.; Blumetti, R.F.; Otterbein, E.S. eds. The Merk
Index, Tenth Edition, Merk & Company Inc.: Pahway, NJ, 1983, p.188.
285. Hunt, D.F.; Shabanowitz, J.; Yates J.R., III; Griffin, P.R.; Zhu, N.Z. in
Methods in Protein Sequence Ana/ys/s Wittmann-Liebold, B., ed.; Springer-
Verlag: New York, 1989, 183-190.
286. Jenson, N.J.; Gross, M.L. Mass Spectrom. Rev. 1988, 7, 41-69.
287. Loo, J.A.; Udseth, H.R.; Smith, R.D. Rapid Commun. Mass Spectrom. 1988,
2, 207-210.
288. Loo, J.A.; Edmonds, C.G.; Smith, R.D. Anal. Chem. 1991, 63, 2488-2499.
289. Sheil, M.M.; Derrick, P.J. Org. Mass Spectrom. 1988, 23, 429^35.
290. Bricker, D.L.; Russell, D.H. J. Am. Chem. Soc. 1986, 108, 6174-6179.


Figure 5-9: IRMPD spectra of the ammonium adduct of 1 -0-methyl-/?-D-
glucopyranoside at irradiance time of 100 and 150 ms
respectively. The peak at m/z 180 is the loss of methanol,
indicating the presence of a modified group on the ring structure
of the monosaccharide.


286
Table 4-5. Calculated q4 values for the +6 (m/z 1417), +8 (m/z 1063), +10 (m/z
851), and +12 (m/z 709) charge states of ubiquitin at 2000 V^.
m/z
Charge State
q4
Detected rf
1417
+6
2.320
2.0 V
1063
+ 8
3.094
2.0 V
851
+ 10
3.867
2.0 V
709
+ 12
4.641
2.0 V


138
potential.
Octopole Electrode Arrangement
The derivation for the octopole potential in this section and those that
follow is based on the general derivation for a multipole device by Szabo et al.219
The first step in understanding the mathematics behind the rf-only octopole is an
examination of the octopole electrode arrangement. For any time-varying two-
dimensional field with cylindrical symmetry, the electrode structure is defined by
the nth order, where 2n equals the number of hyperbolic electrodes. In the case
of the octopole, the applied electric field is 4th order with 2n=8. The two
dimensional octopolar field is generated by the eight parallel/hyperbolic rods
shown in figure 3-1. Ideally the rods should be hyperbolic in nature, with an
applied potential between opposite pairs of rods defined by equation 3-20 from
the previous section:
0 = U V cos w (t t0) (3-20)
with the independent time variable denoted as t and the initial phase defined by
V The frequency w in equation 3-20 is defined in angular units by:
co = 2 it f (3-21)
where f is the applied frequency.219
The x-axis in figure 3-1 is chosen so that it bisects (lies in the plane of
symmetry) one of the positively charged electrodes. For symmetry


Table 2-2. Thermochemical data for the photodissociation of allyl bromide.
Reaction
^^f(reaction)
kJ/mol
AHf(products)
kJ/mol
^Hf(reactants)
kJ/mol
Photons
C3H5Br+ + nhv C3H5+ + Br
57.6
C3H5+ = 957.7
Br = 117.9
C3H5Br+ = 1018
n ^ 5
a AH, values obtained from reference 170.
to
O)


Detected RF (V)
Mass Set (V)
RF Voltage (V


79
spectrum. The effect on the photodissociation efficiency of exciting the [M+H]+
ions of diglyme (no He present) using dipolar resonant excitation (during period
5 in figure 2-3) is seen in figure 2-8. With increasing dipolar resonant excitation
voltage applied to the endcaps, the axial excursions of the ions increase,
decreasing the fraction of time the ions spend in the radial plane of the ring
electrode which, in turn, decreases the extent of interaction between the stored
ions and photons. The decrease in photodissociation efficiency can be further
rationalized by considering the inset on figure 2-8. As long as the ions maximum
excursion (trajectory) in the axial or z-direction does not exceed the laser beam
width (dashed line figure 2-8), the photodissociation efficiency remains relatively
constant. However, if the ions gain enough kinetic energy from the applied
dipolar field so that ion trajectories exceed the laser beam width in the axial
direction, a significant decrease in photodissociation efficiency is observed. This
decrease in efficiency can be explained by considering instantaneous ion velocity
arguments. As an ion moves away from the center of any rf-only device
(quadrupole or ion trap), the magnitude of the restoring forces become larger, as
for any harmonic oscillator. Therefore, as an ion reaches its maximum
displacement from the center of the ring electrode, its instantaneous velocity is
very slow compared to that of the same ion as it accelerates through the center
of the device. This means that when the peak-to-peak excursions of the individual
ion trajectories are larger than the laser beam width in the axial direction, the ions
spend most of their time outside the radial plane of the ring electrode where the


Aglycone Moeity
Erythromycin
mw = 733.4
398


92
in the Dipolar Excitation section). For the wr band, a small decrease in
photodissociation efficiency is observed at 2750 mV, which could correspond to
a limited amount of coupled motion in the axial or z-direction. In the case of the
2a/r band (representing the tur frequency component), a sharp decrease in
photodissociation efficiency was observed at 750 mV. This decrease cannot be
attributed to excitation in the radial direction, since there exists a high photon
density throughout the radial plane of the multi-pass ring electrode. Therefore,
the only plausible explanation for the observed drop-off in photodissociation
efficiency for excitation at 2a/r is the coupling of ion motion in the axial direction.
Experimental evidence suggests that the majority of the coupled motion observed
(at qz=0.3, az=0) was due mainly to excitation in the z-direction and not
necessarily ejection or collisions of ions with the endcaps. This conclusion was
drawn from the fact that as the quadrupolar excitation amplitude was increased
(laser off) past the point where photodissociation efficiency dropped off (750 mV),
there was only a small reduction in the parent ion population. At an excitation
voltage of 1000 mV, a significant reduction in parent ion population occurred (as
evidenced by the increased variation of photodissociation efficiency values
beginning at 1000 mV), which likely represents radial "ejection" (collisions with the
ring electrode) of the ions and not complementary axial ejection.167 The results
obtained experimentally using the multi-pass ring electrode agree quite well with
those predicted by March et al.57'156,157 For qz=0.2940 and 3^=0 of the m/z 134
ion of n-butylbenzene, March and coworkers reported that radial ejection


393
-desosamine sugar. Therefore, it would be predicted that charge-site
fragmentation should be driven from this portion of the ion. In figure 5-20 is seen
the photodissociation spectrum (7 laser pulses) of protonated erythromycin. The
low mass peaks at m/z 100 f-5AJ and m/z 116 (l2AJ are charge-site driven
fragmentations for the ring opening of the D-desosamine sugar. Direct cleavage
on either side of the glycosidic bond (with accompanying hydrogen transfers) of
the D-desosamine sugar is indicated by the two fragments m/z 158 (B1a) and m/z
174 (C1a). The appearance of the small peak at m/z 174 signifies the transfer of
two hydrogens from the amino sugar to the aglycone-L-cladinose neutral. The
unusual aspect of this fragment is the fact that it is primarily observed in the
charge-remote fragmentation process, indicative of high-energy collision
298,322
processes.
Even more astounding is the presence of a large peak at m/z 576 (Yop),
indicating loss of the L-cladinose sugar from the aglycone ring (cleavage on the
L-cladinose side of the glycosidic bond). The peak at m/z 558 could either be
cleavage of the glycosidic bond on the aglycone side to form the Zoe ion, or loss
of water from the m/z 576 ion. The peaks at m/z 540 and m/z 522 are
consecutive water losses occurring from the aglycone ring (e.g., three hydroxyl
substituents present). Around the protonated erythromycin region (m/z 734, m/z
716 and m/z 698) is observed two consecutive losses of water. This indicates
that some form of charge migration or charge-remote fragmentation mechanism
is in place since there is only one hydroxyl group present on D-desosamine (e.g.,


Rate Coefficient (s )
Pressure xlO 6torr
00


Vacuum Manifold
[
y
J Ion Trap
-j J
1 1-
ir
\
\
Instrument Table Helium-Neon
co2
Gold Plated
Laser Beam
Reflecting Mirror
/ /
/
/
f
/
/
*
Optical Table
111 11 i 1111 1111111 11 1111111111 i 1111111
1111111 \.
/
s
Beam Selector
Apollo
cw CO 2
Laser


RF Level 25
Met-Arg-Phe-Ala
2
a
£
2
u
<
1600
1400
1200
1000
800
5 600
400
200
+2 Charge State
m/z 262
+1 Charge State
m/z 524
I I | I I I | I I I | I II | I I rj I II |
0.00 0.04 0.08 0.12 0.16 0.20 0.24 0.28
^inject
RF Level 38
TT I I | I I II | I I I I | I I I I |
40 50 60 70
RF Level (m/z)
20
297


800
600
400
200
0
m/z 131
m/z 69
m/z 100
m/z 502
11111111111 M 1111111111 r i| n n 11111111 n 11 in 11111111111111
50
100 150 200 250
300 350
m/z
400 450 500 550 600
260


6000
5000
4000
3000
2000
1000
0
r

+12 (m/z 709)

+10 (m/z 851)

+8 (m/z 1063)
A
+6 (m/z 1417)
/
I I I I I I MI [I I I I | I 1TI |TT1T| TI I I | I I II | II I I | I I I I | I I II | III I | II I I |
20 30 40 50 60 70 80 90 100 110 120 130
Electrospray Gate Lens Voltage (V)
293


Figure 5-12: CID and photodissociation spectra of the ammonium adduct of
raffinose. The major fragmentation observed in the CID
spectrum is the [M+H]+ at m/z 505 which essentially confirms
the molecular weight, but provides no structural information.
The photodissociation spectrum shows complete fragmentation
information for the trisaccharide.


A NOVEL ELECTROSPRAY ION TRAP MASS SPECTROMETER FOR
PHOTODISSOCIATION OF BIOLOGICAL MOLECULES
LD
1780
159^
Si%
JAMES LEE STEPHENSON, JR.
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
1995
UNIVERSITY OF FLORIDA LIBRAS


299
Table 4-7. Experimental and theoretical threshold values of qz for the peptide
MRFA.
m/z
RF Level (experimental)
qz (experimental)0
qz (theory)b
262
20
0.070
0.070
524
n "I
29
0.049
0.045
Experimental value determined for the half-height of the appearance of trapped
ions.
Theoretical values calculated with an arbitrary constant of 1.3; qz=(1.3/m)1/2.


288
Table 4-6. Calculated q2 values for the +6 (m/z 1417), +8 (m/z 1063), +10 (m/z
851), and +12 (m/z 709) charge states of ubiquitin at 2000 V^,.
m/z
Charge State
q2
1417
+6
3.309
1063
+8
4.412
851
+ 10
5.515
709
+ 12
6.618


330
pulsed-gas introduction of collision gas, which makes the MS/MS experiment
much more complicated, produced dissociation efficiencies on the order of 15%,
which is significantly lower than for photodissociation291 Comparison between
photodissociation and surface-induced dissociation of porphyrins in the ICR by
Castro et al. yielded higher dissociation efficiencies for the photodissociation
process. Longer irradiance times (e.g., with no parent ion selection after the first
MS/MS step), produced more fragment ions at higher abundances than those
obtained with SID. At very long irradiance times (both UV and UV-vis
wavelengths), the new fragment ions were produced at the expense of
diagnostically significant ions. These new fragment ions produced were of no
value for the structural elucidation process.292
Recently, IRMPD has been employed in the ICR for photodissociation of
biological species in the gas-phase. Little et al. demonstrated the capability of
IRMPD to obtain sequence information for peptides/proteins and
oligonucleotides.111 The IRMPD of carbonic anhydrase produced fragment ions
similar to, but with valuable additions, fragmentation obtained by other methods
(e.g., CID and SID). Optimization of irradiance times varied widely for
peptides/proteins (from 50 to over 300 ms), indicating a greater range of ion
stabilities than originally believed from previous CID data. Irradiance times for
oligonucleotides (negative ion mode) were significantly less (e.g., 10-30 ms) than
those of the peptide/protein experiments. This difference could be attributed to
the photon resonance of the PO stretching frequency. More importantly, IRMPD


131
energy dependence, product branching ratios, collision-cross sections) of ion-
molecule reactions.224-227 More recently, the rf-only octopole has been employed
in the development of commercial (liquid chromatography/mass spectrometry) ion
trap instrumentation as an ion injection device (Finnigan MAT LCQ ).2ZB Other rf-
only devices have been designed and used (e.g., hexapoles and quadrupoles)
for the aforementioned purposes and have been employed as ion transmission
devices.229-232
In this chapter, the theory, design, and implementation of an rf-only
octopole for use as an ion injection device is discussed. A general overview of
ion behavior in electromagnetic fields is presented followed by a discussion of
multipole field potentials. Calculations for electric field strength, electrode
geometry, and the equations of motion are then introduced to give the reader a
general understanding of the physics of rf-only (octopole) devices. The next
section presents a discussion on the design considerations for the rf-only
octopole used in the electrospray/ion trap instrument built in our laboratory. The
chapter concludes with a discussion of the factors which contribute to the ion
transmission properties of the rf-only octopole (initial entry angle, rf frequency, rf
amplitude, collisions factors, and kinetic energy considerations).
Ion Behavior in Electromagnetic Fields
This section is intended to introduce from first principles the differential
equation governing the motion of an ion through an electromagnetic field with any


404
mass range), all attempts to interpret the data have been unsuccessful. However,
with a digital signal processing card (DSP) and a reduced rf ramp rate for high
resolution, these problems could be overcome.
The area of greatest potential is the analysis of carbohydrate linkages using
IR photodissociation. The high photoabsorption cross-section of the COC
linkage makes this technique an ideal tool for linkage analysis. As demonstrated
with protonated erythromycin, the first steps to the identification of carbohydrate
moieties attached to biochemical molecules could be accomplished more quickly
and with less sample than other more established mass spectrometric/organic
derivatization procedures. Perhaps the most important future study in this area
would involve establishing fragmentation rules for the photodissociation process;
only when this is accomplished will the true potential of the technique be known.


262
Table 4-2. Instrument parameters for El, octopole, and ion trap operation for the
PFTBA spectrum shown in figure 4-20.
Instrument Parameter
Value
Filament Current
35 fjA
Electron Energy
70 eV
Extractor Lens (L1)
-8 V
Lens (L2)
-105 V
Quadrupole Entrance Lens (Ion Gate)
-55 V (gate on) +180 V (gate off)
Mass Set
0.3 V detected RF
Octopole Offset
-3.0 V
Ion Trap Offset (ring and endcaps)
-5.0 V
Exit Tube Lens
-120 V
Dynode
-15 kV
Electron Multiplier
-1250 V


363
substituents.
Raffinose
To determine the applicability of photodissociation to sequencing
oligosaccharides, two straight chain oligosaccharides, raffinose (O-cr-D-
galactopyranosyl[1-6]-cr-D-glucopyranosyl-/?-D-fructofuranoside) and stachyose (a-
D-galactopyranosyl-[1 -6]-cr-D-galactopyranosyl-[1 -6]-o-D-glucopyranosyl-[1 -2]-/?-D-
fructofuranoside), were chosen for study. In figure 5-10 are shown the structures
of these two compounds. This section will focus on the photodissociation and
fragmentation of raffinose, an oligosaccharide typically used in tissue culture
media.
A 20 pmol///L solution of raffinose was prepared in a 50:50 methanol:water
solution with 3 mM NH4OH. Samples were directly infused through the
electrospray interface as described previously in chapter 4. For photodissociation
experiments, a pulsed C02 excimer laser was employed as described in the
photodissociation set-up section in chapter 4. The laser was capable of an
approximate 3-5 ns pulse with an energy of 1.1 J (at 944 cm'1). Due to the short
pulse of laser energy and the high photon absorption cross section for the
COC stretch, the ion trap was operated with buffer gas at a pressure of
1.0x1 O'4 torr (uncorrected) for both CID and photodissociation experiments. The
major (parent) ion produced from the electrospray process was the ammonium
adduct of raffinose at m/z 522.
The fragmentation nomenclature employed for the spectral interpretation


27
the rf amplitude applied to the ring electrode, all m/z values below a specific ion
of interest are ejected from the trap. Next, by lowering the dipolar frequency and
decreasing the rf amplitude, masses of higher m/z values are ejected, thus storing
the desired m/z ion or range of ions.
Broadband techniques have also been employed for ion isolation studies.
The various methods available include stored waveform inverse Fourier transform
(SWIFT)39, the use of multiple single discrete frequencies36 80 81, and random
noise.82 Perhaps the most effective means for ejecting a large range of ions is
that of SWIFT or random noise. These techniques have proven to be very
successful, since a large number of signals (of various frequencies) can be
applied simultaneously to the endcap electrodes. In essence, sensitivity is
increased because the ion trap is selectively filled with the ion(s) of interest and
not with unwanted matrix or background ions.
The ability to implement user-defined scanning strategies through the Ion
Catcher Mass Spectrometer (ICMS) software developed in this laboratory was
critical for the success of the experiments in this dissertation.83 Specifically, the
development of ICMS software allowed for the computer control (TTL signal) of
external devices such as a pulsed C02 laser, a continuous wave (cw) C02 laser,
and a pulsed-valve controller.83 The FORTH programming option permitted the
design of intricate experiments involving resonant excitation frequency, laser
control, pulsed-valve control, and ion cool time (vibrational relaxation). In


Figure 4-7:
General schematic of the SWR bridge circuit. The three resistors (labeled R) are of equal value.
The impedance of the octopole (z) is unknown, and is defined as the complex impedance of the
input to the rf tuned circuit. For a 50 n load, the values of R in the bridge are 50 Q, thus giving
0 current through the detector when the rf circuit is in tune. Adapted From Reference 258


35X3
CH2OH /
^3a
m/z 505
m/z 488 (-NH3)
la
m/z 164 (-NH3)
2a
m/z 343
m/z 326 (-NH3)
^2a
m/z 343
m/z 326 (-NH )
OH
MWAlWiWW
3a
m/z 505
m/z 488 (-NH )
CH20H 0
HO'i m/z 164 (-NHL)
ch2oh
OH
2,4A
3a
379


Figure 3-2:
Rectangular and polar coordinates used to calculate the potential field of an octopole electrode
configuration with 8 identical electrodes. The x-axis is chosen so as to coincide with the plane of
symmetry of one of those electrodes which has the potential +J2. Theta (O) is the angle of
rotation, r is the displacement from the center axis, and r0 is the inscribed radius. Adapted from
reference 219.


Figure 2-16: IRMPD ion growth curves for protonated 15-crown-5 ether as a function of laser irradiance time.
Three reaction channels were observed with the following sequence: m/z 221 -* m/z 133 (loss of
C4H802) -* m/z 89 (loss of acetaldehyde) -* m/z 45 (loss of acetaldehyde). Error bars are defined
as the standard deviation of the mean.


400
reactions, and the effects of buffer gas on photodissociation efficiency. The
information obtained in these fundamental studies, was used successfully to
optimize the photodissociation/instrument parameters used in the study of
peptides/proteins, carbohydrates, and oligonucleotides described in chapter 5.
The techniques used to acquire the wavelength-dependent spectra
obtained for the allyl bromide ion and protonated diglyme may provide a method
for examining the gas-phase IR absorption profiles of various ionic species.
Although the fine structure is not available from these experiments, general
features can be observed indicating the presence of various functional moieties.
The unique design of the multipass ring electrode also lends itself to
detection of coupled motion (r- and z-directions) in the quadrupole ion trap.
Since a high photon density is observed in the radial plane of the ring electrode,
any excitation signal applied to the endcaps which increases the axial excursions
of the ions can be detected by a sharp decrease in photodissociation efficiency
of the parent species. The observed decrease in photodissociation efficiency is
due to the reduced amount of time the ions spend in the radial plane of the ring
electrode.
The next step of the project involved the design and construction of an ion
transmission device (the rf-only octopole) to transport ions from the LC/MS
interface region (ESI source) to the ion trap analyzer in the most efficient manner
possible. The major advantage of this device is the cubic dependence of the rf
restoring force for a given ion as it moves off axis. This is particularly important


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 EF6OWXE3T_M2P140 INGEST_TIME 2015-08-10T19:12:52Z PACKAGE AA00029830_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


48
subsequent experiments performed in this and all of the following dissertation
chapters. The basis for the experimental design is derived from previous studies
by Watson et al. in the ICR mass spectrometer.149 A detailed description of the
instrumentation employed has been published.147,148
Experimental Design
All experiments were performed on a Finnigan MAT (San Jose, CA) ion trap
mass spectrometer (ITMS). Except for the pulsed-valve experiments, the ion trap
was operated with no He buffer gas in the vacuum chamber. The base pressure
of the instrument was 3.5x1 O'8 torr (uncorrected), as indicated on a Bayard-Alpert
ion gauge (Granville-Phillips, Boulder, CO) mounted on the vacuum manifold.
The Teflon ring electrode spacers were found to absorb strongly at 944 cm'1
during the laser irradiation period, which desorbed both neutral and ionic species
that affected both ion storage efficiency and detection; they were therefore
removed. The vacuum manifold was equipped with a modified flange containing
a ZnSe window to pass IR radiation. The modified software used (ICMS) provided
two TTL pulses for computer control of both the cw C02 laser and the pulsed-
valve apparatus.83 The experimental arrangement can be seen in figure 2-1.
The cw C02 laser employed was an Apollo Model 570 that is line tunable
over a wavelength range of 1099-924 cm'1 and has a beam diameter of
approximately 1 cm. The maximum laser power obtainable was 50 W at 944 cm'1.
The laser power supply was modified with electronics that transformed an


317
phase (although internal sequence ions from histidine and arginine can also
appear in the spectrum due to their large gas phase basicities). Of all the basic
residues, proline has the highest proton affinity, even exceeding those of arginine
and lysine. The mechanism of formation for these ions follows the same
convention as y type ions with charge retention on the amide nitrogen and
subsequent hydrogen ion migration to the highly basic proline residue (see figure
4-35).280 Internal sequence type ions have been observed with peptides that
contain proline residues in ion trap mass spectrometry. The high ion abundance
of these peaks is probably due to the timescale of the ion trap MS/MS
experiment, which takes place on the millisecond as opposed to the microsecond
timescale found in either magnetic sector or triple quadrupole instruments. Over
the millisecond timeframe, ions in the trap will tend to go towards their lowest
energy state, if given the opportunity.
MS3 experiments were performed on both the b5 and b9+2 ions from the
MS/MS spectrum. Figure 4-36 shows the MS3 spectrum obtained from the b5 ion
of angiotensin I. The b5 ion was mass isolated using forward and reverse scans
and fragmented at a qz of 0.3 (amplitude=2.0 Vp_p). Typically, singly-charged b
ions fragment easily to produce lower mass b ions and some a ions, as seen in
figure 4-36. There is a great deal of interest in generating a large number of high
mass b ions, so as to evaluate MS" techniques for sequencing peptides one
residue at a time. From figure 3-36, the formation of the a^ b4, and a4 verifies the
loss of isoleucine (or leucine which has the same residual mass at 113) from the


Multipole coordinate system used for simulation studies: (a) designation of r0 for a multipole (e.g.
quadrupole) device, (b) coordinate axis system (rectangular) with the xy plane denoted in the plane
of the page, (c) off-axis angle for ion entry into the multipole device, 0 is relative to the x-axis (xy
plane). Adapted from reference 253.
Figure 3-9:


419
210. Eiden, G.G.; Garrett, A.W.; Cisper, M.E.; Nogar, N.S.; Hemberger, P.H. Int.
J. Mass Spectrom. Ion Processes 1994, 136, 119-141.
211. Booth, M.M.; Stephenson Jr., J.L: Yost, R.A. in Proceedings of the 42nd
ASMS Conference on Mass Spectrometry and Allied Topics Chicago, IL,
1994, 693.
212. Yang, M.; Dale, J.M.; Whitten, W.B. Anal. Chem. 1995, 67, 1021-1025.
213. Dale, J.M.; Yang, M.; Whiten, W.B. Anal. Chem. 1994, 66, 3431-3435.
214. Pinkston, J.D.; Delaney, T.E.; Morand, K.L Anal. Chem. 1992, 64, 1571-
1577.
215. Mikami, N.; Sato, S.; Ishigaki, M.; Toshiki, S. Chem. Phys. Lett. 1990, 166,
470-474.
216. Mikami, N.; Sato, S.; Ishigaki, M. Chem. Phys. Lett. 1991, 180, 431-435.
217. Yost, R.A.; Enke, C.G. Anal. Chem. 1979, 51, 1251A.
218. Teloy, E.; Gerlich, D. Chem. Phys. 1974, 4, 417-427.
219. Szabo, I. Int. J. Mass Spectrom. Ion Processes 1986, 73, 197-235.
220. Guettler, R.D.; Jones, G.C.; Posey, L.A.; Kirchner, N.J.; Keller, B.A. Zare,
R.N. J. Chem. Phys. 1994, 101, 3763-3771.
221. Ervin, K.M.; Amnentrout, P.B. J. Chem. Phys. 1985, 83, 166-189.
222. Syka, J.E.P. Szabo, I. in Proceedings of the 36th ASMS Conference on
Mass Spectrometry and Allied Topics San Francisco, CA, 1988,1328-1329.
223. Hail, M.; Mylchreest, I. in Proceedings of the 41st ASMS Conference on
Mass Spectrometry and Allied Topics San Francisco, CA, 1993, 745a.
224. Shao, J.D.; NG, C.Y. Chem. Phys. Lett. 1985, 118, 481-485.
225. Gerlich, D.; Disch, R.; Scherbarth, S. J. Chem. Phys. 1987, 87, 350-359.
226. Posey, L.A.; Guettler, R.D.; Kirchner, N.J.; Zare, R.N. J. Chem. Phys. 1994,
101, 3772-3786.
227. Anderson, S.L.; Houle, F.A.; Gerlich, D.; Lee, Y.T. J. Chem. Phys. 1981, 75,


(Vi) i
-0.51
-1.0
-1.5
-2.0
-2.5
10.0
1 1 1 1 1 1 1 1 1 1 1 1 I
15.0 20.0 25.0
1
30.0
Irradiance Time in ms


Figure 5-3:
IRMPD of the triply charged m/z 433 ion of human angiotensin I. The spectrum bears a
resemblance to the corresponding CID spectrum reported in chapter 4. The peaks at PFH and
PFHL represent internal y sequence ions generated by the presence of proline at residue 7.


Window Flange
ZnSe Window Flange
Blank Flange


140
+


107
experiments confirmed the formation of the m/z 59 product ion from the m/z 103
precursor, thus verifying the presence of a consecutive reaction mechanism of the
type m/z 135 -+ m/z 103 -* m/z 59.147 The proposed mechanism of this
consecutive reaction, seen in figure 2-15, corresponds well to results reported
previously for the IRMPD of protonated diglyme in the ICR cell.175
The equations for the consecutive reactions of protonated diglyme can be
derived from the original work of Harcourt and Esson where the reaction of
protonated diglyme follows the simple consecutive reaction mechanism of:176-178
nhv nhv
[c6H15o3r [CsH^r [c3H7or
(2-10)
If the initial ion population of the protonated diglyme is [C6H15O3]0+ and the ion
population at any time t is [C6H1503]+, then the rate equation for the first
sequential reaction becomes:
_d[CH,5031- (2-11)
Applying the boundary condition where [C6H1503]+ = [C6H15OJ0+ when t = 0 and
integrating gives:
[C6H150,]- = [C6H,5O3]0e-k'' (2-12)
For the second sequential reaction step the net rate of formation for [C5H1102]+
is:


216
determining the appropriate operating rf frequency the load was assumed to be
50 £2. Since the capacitance of the device is unknown (e.g., depends on the
material the electrodes are made from, electrode spacing, and electrode surface
area), there could be a large amount of rf power dissipated from the rf coil (tuned
circuit not optimized) leading to a significantly reduced maximum power output
from the rf amplifier. Since the octopole will be operated in the rf only mode,
maximum voltage levels of over 2000 V^p are not needed, so an optimized
matching circuit is not critical.
A general schematic of the SWR bridge circuit used for tuning is shown in
figure 4-7. The circuit consists of a Wheatstone bridge composed of three equal
resistors and the octopole with an unknown impedance (Z). The resistor values
are chosen at 50 £2 so that the impedance through the octopole will be 50 £2 at
the appropriate operating frequency (e.g., current through the circuit is 0).
To determine the proper operating frequency, a function generator is
needed which can scan a range of frequencies typically from 0.5 to 4.0 MHz. In
addition, the function generator should be able to produce a marker TTL signal
for determination of the optimum rf frequency. A dual channel oscilloscope is
used to monitor the frequency and marker TTL signals. A schematic diagram of
the test set-up is shown in figure 4-8. A Stanford Research System DS345
arbitrary waveform generator (Sunnyvale, CA) was set to scan from 0.5 MHz to
2.0 MHz at a 50 ms scan speed in order to determine a coarse adjustment range
for the optimum frequency. When the frequency synthesizer is scanned, the SWR
trace appears on the oscilloscope with a "dip" at the frequency that represents the


371
These results would also be consistent with other low-energy CID results for
straight chain oligosaccharides reported in the literature.47 The fragmentation
analysis for raffinose is shown in figure 5-13.
Also shown in figure 5-12 is the photodissociation spectrum (2-laser pulses)
of the ammonium adduct of raffinose. Compared to the CID spectrum, PID
produced only a small [M+H]+ peak, with the majority of the ion signal producing
structurally diagnostic ions. The two largest ions in the spectrum were the m/z
326 and m/z 343 ions indicative of the loss of a terminal monosaccharide (as
seen with the CID spectrum). Also present were a series of peaks indicating
successive losses of H20 at m/z 308, m/z 290, and m/z 272 from the ion at
m/z 326. An additional peak at m/z 164 represents the cleavage of the internal
monosaccharide residue (D-glucose) to form either the B1a or Z1o ion.
For determination of an actual unknown sample, identification of the
internal sequence ions of an oligosaccharide would be extremely difficult using
mass spectrometry alone (even using the appropriate derivatization chemistry),
since anomeric and linkage position data are stereospecific. It is important to
point out that with photodissociation, the entire sequence of the trisaccharide was
obtained without complicated instrument tuning procedures. In contrast for CID,
a third stage of mass spectrometry would have to be performed in order to obtain
the same information (e.g., cleavage of the internal residue) as from the
photodissociation experiment. This means that for photodissociation another
sample could be analyzed immediately after the first, since no tuning of


Figure 2-10: The power absorption spectrum of protonated 12-crown-4 ether for an applied quadrupolar
excitation signal (6 Vp.p) at qz=0.3 and az=0 (no He buffer gas present). The applied quadrupolar
signal was started at a frequency of 25 kHz and incremented at 0.1 kHz intervals to 500 kHz. The
symbols u/z and cu, refer to the fundamental frequencies of motion for an ion in the z- and r-direction
respectively.


78
(hexapolar); 2) harmonic considerations of the applied dipole field; or 3) direct
contributions to ion motion. A detailed investigation into the origin of these
absorption bands has been performed by Eades162 and Vedel.166 These two
authors were independently able to verify the existence of the 2a;z frequency band
as a component of ion motion and not just part of the harmonics from the
excitation field. Other absorption bands observed by Eades and Vedel which
were not seen in these experiments include wJ2 and aiz+u)r Several plausible
explanations for this observation include the qz value used, the large a/z band
width, the absence of He buffer gas, and the sensitivity of the chemical probe
(protonated diglyme).
To examine the effect of dipolar excitation on ion (axial) motion, the
protonated diglyme ions at m/z 135 were irradiated with a cw C02 laser during the
dipolar excitation (10 ms) period. Since the time between consecutive photon
absorption for the IR laser in this experiment is on the ms time scale,
instantaneous ion trajectories or velocities cannot be determined. However, C02
lasers can be used to evaluate the time-averaged behavior of the ion (cloud)
population. The multipass optical design used in these experiments produces a
high photon density in the radial plane of the ring electrode as shown by figure
2-6. Any ions whose axial excursions exceed the beam width of the laser (3 mm)
should show a marked decrease in photodissociation efficiency.167
To test this theory, protonated diglyme ions were excited at a dipolar
frequency of 118 kHz (o/J, which was determined from the power absorption


14000
12000
10000
8000
6000
4000
2000
0
Ion Trap Offset -5V
Octopole Offset Voltage (V)
290


Figure 4-29:
Ion injection efficiency as a function of the minimum rf level, expressed in terms of q,n]9Ct and
minimum rf level (given as the low-mass cutoff). The dashed lines indicate points at which spectra
were taken to show the effect of ion injection rf, on spectral quality (see figure 4-30).


ACKNOWLEDGMENTS
I would like to begin by expressing my unfeigned gratitude to Dr. Richard
A. Yost for allowing me to pursue my own research ideas (no matter how bad)
over the last five years. It has been a most satisfying experience personally and
professionally. I also would like to thank the following members of my committee:
Dr. John Gander for his insightful comments on carbohydrate chemistry; Dr. Jim
Winefordner for the endless jokes; Dr. Bob Kennedy for the career discussions on
academic life; Dr. John Eyler for convincing me that a modified ring electrode
would actually work; and Dr. Dave Powell for those grueling 10 mile Sunday runs
and brutal Tuesday night track workouts. My survival here was also due in no
small part to Jeanne Karably, Susan Ciccarone, and Donna Balkcom, all of whom
pointed me in the right direction, provided me with the right paperwork, and told
me where to be in order to graduate.
Financial support for this research came from a variety of sources including
the University of Florida Division of Sponsored Research, the Office of Naval
Research, the ACS Analytical Division Fellowship (sponsored by The Procter &
Gamble Company), and a Dissertation Fellowship from the College of Liberal Arts
and Sciences at the University of Florida.
in


282
sample used = (sample cone.) (flow rate) (ion injection time) f4-5)
where sample concentration, flow rate, and ion injection time are in units of
(n,p,f)mol///L, //LVmin, and ms, respectively. For the case of the 5 fmol//L sample
in figure 4-25, the total sample consumed was 50 atomoles consumed during the
200 ms ion injection period.
Octopole RF Level
To test the affect of the applied rf voltage, octopole offset and ion gate lens
on ion transmission, the intensity of the ion signal for multiply-charged yeast
ubiquitin was plotted as a function of rf voltage, octopole offset voltage, and ion
gate voltage, respectively. Yeast ubiquitin was obtained from ICN
Pharmaceuticals Inc. (Costa Mesa, CA) and was prepared in a 50:50
methanol:water solution containing 0.1 % acetic acid. Addition of methanol to the
ubiquitin standard causes unfolding of the protein from its native form.265"270 This
lowering of the dielectric constant (e.g., addition of alcohol) increases the
percentage of the protein in the cr-helical confirmation.267,269'270 The observed
charge states of ubiquitin in the methanolic solution range from +6 (aqueous) to
the +10 to +12 charge state (due to the binding of methanol to the protein
surface thereby unfolding the protein). The high ion intensity of these charge
states make ubiquitin an attractive choice to study the efficiency of ion
transmission in the ESI/ion trap instrument.
A plot of ion intensity for the + 6 (m/z 1417), +8(m/z1063), +10(m/z851),


"I have little patience with scientists who take a board of wood, look for its
thinnest part, and drill a great number of holes where drilling is easy."
Albert Einstein in "Einsteins Philosphy of Science"
Reviews of Modern Physics 1949, 21, no. 3.
"Just work hard and be honest, and everything else will take care of itself..."
Orson Calvin "O.C." Pearson ....my Grandfather


172
Simulation results from an rf-only collision cell designed by Davis and
Wright are presented in order to demonstrate the ion transmission properties of
the rf-only octopole compared to those of the traditional rf-only quadrupole.253
The validity of the computer modeling program used by Davis and Wright was
tested by evaluating the standard configuration (for the rf/dc quadrupole) of the
Concept ISQ mass spectrometer from Kratos Analytical (Manchester, UK) and the
rf-only collision cell of a ZAB-EQ mass spectrometer (VG Fisions, Manchester,
UK). The dimensions of the quadrupole and octopole discussed below are r0=4.0
mm (inscribed radius) with a total rod length of 20 cm. The multipole coordinate
system used for these studies is shown in figure 3-9.253
Initial Entry Angle Conditions
The trajectories for an ion of m/z 1000 which enters the multipole device(s)
1 mm off axis at angles of 0, 2, and 4 degrees (i9 relative the x-axis is 45), can
be seen in figure 3-10253. In this plot the abscissa is defined as the displacement
(| Z |) of the ion from the central axis and is plotted against the ions longitudinal
position in the device. From figure 3-10a (rf-only quadrupole) the amplitude of the
trajectories increase with increasing entry angle, but the wavelength of secular
motion and phase of the trajectories are similar. For the case of the rf-only
octopole shown in figure 3-10b, entry angle dependence on ion trajectory
amplitude is approximately the same as that observed for the rf-only quadrupole.
However, the effects on the wavelength of secular motion and corresponding


353
the presence of other ring substitutions (e.g., more OH or other functional
groups) a third sequential step of MS was performed on the m/z 164 ion
[M+NH/-H20].
The MS3 spectrum of the ammonium adduct of 2-deoxy-D-glucose is also
shown in figure 5-5. Due to the stability of the m/z 164, dissociation efficiencies
were only *=13%. Diagnostic ions indicating the loss of one and two neutral H20
molecules from the ring at m/z 147, m/z 146, and m/z 127 were present in the
MS3 spectrum as well as the MS2 spectrum. An additional ion at m/z 111
[M+NH4+-3H20-NH3] signified the presence of a third hydroxy substituent on the
monosaccharide ring. There was no indication of cleavage of the hydroxyl group
attached to position 6 of the pyranose ring, nor was there any evidence for the
loss of methanol (e.g., cleavage of the C5C6 bond) from the pyranose ring.
In the IRMPD spectrum of the ammonium adduct of 2-deoxy-D-glucose
(figure 5-7), the m/z 164, 147, 146, 129, and 111 peaks (indicating the losses of
one, two and three hydroxyl moieties from the pyranose ring) were all present in
a single MS/MS experiment. This was due to the formation of product ions from
the m/z 182 parent species during the laser irradiance time (85 ms). As the first
sequential products were formed, they could undergo further dissociation to yield
more fragment ions, since there was no mass isolation step during the laser
irradiance period. The dissociation efficiency of the m/z 182 ammonium adduct
was *=100%. As the laser irradiance time was increased (figure 5-7), this
essentially drove the series of consecutive reactions for the ammonium adduct of


73
each endcap alternates at the same frequency as the ions secular frequency in
the axial direction.162
The trapping conditions observed with the application of a dipolar
excitation signal can be significantly different than those seen by direct application
of a quadrupolar trapping field to the ring electrode. Therefore, the resulting ion
trajectories no longer follow traditional Mathieu parameters, since ion motion is
now controlled by both the quadrupolar trapping field and the dipolar excitation
field. Previously reported simulation(s) for the application of dipolar fields for
resonance excitation typically employ numerical methods and simplified models
to describe ion motion.163 Exact solutions cannot be obtained because of
electrode surface geometry considerations (machining of hyperbolic surfaces,
truncation effects, and endcap holes) and the presence of higher-order fields
which result in a perturbed quadrupolar trapping field, therefore affecting ion
motion. The aforementioned reasoning results in coupled motion in the r- and z-
directions, thus making exact solutions difficult to solve mathematically.
A preliminary investigation to determine the influence of storage
conditions (under the influence of dipolar excitation) on photodissociation
efficiency was undertaken with protonated diglyme. To minimize the effects of
collisions on photodissociation efficiency, the ion trap was again operated with no
He buffer gas. A detailed description on the instrumentation employed and the
procedures for data acquisition, can be found in the Experimental Design section
of this chapter.


Intensity
308
4000 i
3000
2000
+18
Mass Isolated Ion
1000
Unknown Fragment
'I TpTTTJTT I | I rV'j fTTTTflTTTT
600 700 800 900 1000 1100 1200 1300 1400 1500
m/z
0


218


18
parent species itself. The ion activation techniques which drive the unimolecular
dissociation process are usually evaluated by a series of four criteria: (1) energy
deposition in the ion of interest; (2) energy distribution or bandwidth of the
deposition process; (3) variability of the deposited energy; and (4) reaction cross-
section for the ion activation technique.1
Several different types of reactions in tandem mass spectrometry (MS/MS)
can be used to bring about the ion activation process. Collisional activation is the
most widely used method for ion activation, operating at translational energies
ranging from approximately 10 eV to about the 10 KeV range. Fragmentation of
polyatomic species via collisional activation is called collision-induced dissociation
(CID).58"61 Another increasingly popular method of ion activation is surface-
induced dissociation (SID). Energy transfer in the SID process can be quite large
(on the order of about 8 eV), where the amount of energy transferred is controlled
by the translational energy of the incident ions.62-63 A third method for the
activation of polyatomic ions involves the interaction of the parent ion with a beam
of electrons, a process called electron excitation. Product ion spectra generated
by electron excitation are characterized by broad energy distributions with an
upper limit approaching the energy of the electrons used for ion activation.6^66
The final method typically employed for ion activation involves the absorption of
photons from a light source (e.g., laser irradiation), with the resulting
fragmentation process referred to as photodissociation.67'69 Photodissociation
holds several advantages over the other ion activation techniques, including the


Figure 4-6:
Simplified schematic of the Finnigan MAT 4500 rf circuit showing the rf power amplifier, low-pass
filter, matching capacitor, and remote rf tuning unit. Adapted from reference 258.


Dipolar Excitation (entrance endcap)
Dipolar Excitation (Exit Endcap)
Quadrupolar Excitation (Entrance Endcap)
Qnadrtipolar Excitation (Exit Endcap)


44
Dissociation
channels
Dissociation threshold
Incoherent
single-photon
interaction
Quasi-continuum
I
Coherent
multiphoton
interaction


295
offset, the m/z of the ion(s) of interest, and the minimum rf level of the ion
trap.192,271'272
To fully appreciate the effects of rf level, m/z ratio and ion trap offset on ion
injection efficiency, a minimum rf level study was carried out on the peptide MRFA
(methionine-arginine-phenylalanine-alanine), which under electrospray conditions
gives both doubly and singly charged species at m/z 262 and 524, respectively.
The MRFA standard (Sigma Chemical Company, St. Louis, MO) was dissolved in
a 50:50 methanol:water solution with 0.1% acetic acid. Electrospray conditions
and flow rate were the same as in the Basic ESI Operation section of this chapter
(see table 4-4). The ion trap offset was set to -5 V with an octopole offset voltage
of -2.2 V.
Figure 4-29 shows two plots: one with the ion intensity of the peptide MRFA
as a function of qinject, and the other the intensity as a function of the minimum rf
level or exclusion limit. The ion injection data indicate that with increasing m/z,
the minimum rf level needed to trap the ions increases. Correspondingly, the q
value needed for efficient trapping decreases slightly. This is in good agreement
with the original studies of Louris et al.192 The behavior of the ions in this manner
can be supported by the original work of Major and Dehmelt.11 Here, if the
average energy of the ions is less than some arbitrary value W (a<<1 and
q<<1), then the ions move within the confines of the trap as determined by
equation 4-8:


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.
fames D. Winefopdner
Braduate Research Professor of
Chemistry
This dissertation was submitted to the Graduate Faculty of the Department
of Chemistry in the College of Liberal Arts and Sciences and to the Graduate
School and was accepted as partial fulfillment of the requirements for the degree
of Doctor of Philosophy.
December, 1995
Dean, Graduate School


204
to the rear 8.0" flange was 0.10" (see figure 4-2). Later experimental evidence
and theoretical support from chapter 3 suggested that this alignment accuracy
contributed greatly to the success of this project, especially after ion trap analyzer
cleaning or any other process which required disassembly/reassembly of the
major system components (ESI source, rf-only octopole, analyzer/detector
assembly). Figure 4-3 shows the complete vacuum manifold assembly including
ESI source, rf-only octopole, and analyzer/detector assembly.
Initial system vacuum tests to check the base pressure were performed
over a one-week period (initial pump-down), with the manifold heated to 100 C.
The observed base pressure after this time was 1.9x1 O'8 torr. Under normal
operating conditions, the vacuum manifold was not heated due to the presence
of the Delrin supports used in the construction of the rf-only octopole.
Electrosprav Ion Source
The electrospray ionization source used with the ESI/ion trap system was
a standard Analytica (Analytica Inc., Branford, CT) source equipped with a
stainless steel heated capillary (modification performed at Finnigan MAT) as seen
in figure 4-4. The purpose of the heated capillary is to help droplet desolvation
for the electrospray ionization process. The heated capillary was 4.580" long with
an outer diameter of 0.308" and an inner diameter of 0.020". The heated capillary
temperature was controlled by an Omega 6000 (Omega Engineering, Stamford,
CT) series temperature controller with a 220 V output, previously used for


164
stainless steel or anodized aluminum is negligible. The ability of the material to
slide easily against these surfaces is important because of the design (discussed
in detail in chapter four) of the ion trap analyzer assembly, which can be rotated
to facilitate laser alignment. Caution must be taken with Delrin in terms of
manifold heating since the material loses its rigidity at 80 C. The four solid
support pieces were positioned on the ion trap analyzer side, at the baffle wall,
for the electrical connections and at the electrospray interface.
Each of the eight stainless steel mounting assemblies was made from a
single piece of stainless steel, as shown in figure 3-6. The contact points for the
octopole rods were machined to the exact radius of the rods to facilitate the spot
welding process, which therefore limits the variability of r0 for the 30 cm length of
the rods. Two stainless steel mounting assemblies were placed on each Delrin
(front and back) support, offset by 45. Each assembly contained four contact
points so that opposing pairs of rods would have the appropriate rf potentials
(e.g., phase) applied. Each stainless steel mounting assembly was attached to
the Delrin support via four stainless steel screws.
The Delrin support and stainless steel mounting assemblies were then put
together. The eight rods (all of length 30 cm) were then positioned on the contact
points by sliding the completed mounting assembly over the rods where a teflon
plug/spacer was used to push the rods securely against the contact point
surfaces (figure 3-7). To ensure that the ends of the octopole rods were aligned
in the same plane, an alignment jig with eight precision drilled holes was used on


Figure 5-14: CID spectrum of the ammonium adduct of stachyose. Only the loss of two monosaccharide units
is observed under ICD conditions (one terminal and one internal residue). A small degree of ring
cleavage on a terminal monosaccharide unit is observed by the presence of the z,4A3o or 3,6X3a peak.


272
Table 4-3. Electrospray instrument parameters
Electrospray Parameter
Value (units variable)
Sample Flow Rate
3 //L/min
Gas Flow Rate
30 psi (NJ
Needle Voltage
+4500 V
Capillary Offset
+ 15 V
Capillary Temperature
180 C


Figure 1-6: Schematic energy level diagram as originally applied to the
IRMPD of SF6. In region I, coherent multiphoton interaction
describes one of the mechanisms (intensity-dependent power
broadening effects) whereby non-resonant absorption can take
place in a sparse region of vibrational levels. In region II, the
quasicontinuum, resonant absorption steps are always possible,
and can be treated as stepwise (incoherent) excitation processes.
Region III lies above the dissociation threshold. Adapted from
reference 145.


Figure 4-38: Negative ion ESI spectrum of bradykinin at 20 pmol///L in a 3 mM NH4OH solution. The polarity
of the dynode, octopole offset, ion trap offset, ion gate lens, and electrospray needle were reversed
in order to acquire the spectrum.


Intensity
370
m/z


12
4(r2-2z2H 0-3)
To
!~{r2-2z2) = -1 (1-4)
2Zo2
The equations of motion for an ion in both the r- and z-directions can be
derived from the forces exerted on the ion independently in each direction. By
applying Newtons second law of motion, substituting equation 1-2 for 0O in
equation 1-1, applying the condition x2+y2=r2, and differentiating, the equations
of motion in the r- and z-direction are obtained:
d2r
dt2
2e
2mr02
(U-VcosQt)r=0
(1-5)
+ (U-VcosQt)z = 0 (1-6)
dt2 2mr02
where m is the mass of the ion of interest and t is the time variable. These
equations are examples of the Mathieu equations, whose solutions have been
studied extensively.23 The general form of the Mathieu equation can be expressed
as:
4%*(a-2quCos25)u = 0 (1-7)
at
where u represents r or z, and f=Qt/2. By performing a series of operations and
substitutions, the parameters au and qu can be evaluated for the quadrupole ion


Figure 4-11:
Octopole rf circuit for the ESI/ion trap instrument. The mass set and detected rf voltages are sent
through a comparitor which determines the power output of the rf to the rod assembly. The
octopole offset is applied to all eight rods through the rf circuit.


Figure 4-27: Effect of octopole offset voltage on ion transmission efficiency for the peptide ubiquitin. As
expected, higher voltages are need to extract higher mass ions from the electrospray source
region.


19
ability to impart a wide range of well defined energies to the parent ion of interest.
The ability to finely control the energy deposition process with photon absorption,
and thus the fragmentation process, is the central focus of this dissertation.
It is beyond the scope of this dissertation to discuss in detail the
fundamental aspects of the ion activation techniques other than photodissociation
(e.g., collisional activation, surface-induced dissociation, and electron excitation).
Instead, a general overview of the unimolecular dissociation process is presented
in order to give the reader a better understanding of the results achieved with ion
activation (e.g., fragmentation process).
Unimolecular reactions in the gas phase have been studied extensively and
can provide a large amount of information concerning rates of dissociation,
energy partitioning, and development of models to describe a particular gas-
phase reaction system. Mass spectrometry provides a unique environment for
evaluating gas-phase systems since reactions can take place in collision-free
conditions. In addition, both parent and product ions can be isolated and
identified with high specificity. Unimolecular dissociations have been used
successfully to explain electron ionization of simple molecules, and the theory is
now currently being adapted to help predict the products from simple collisional
activation experiments.70 This theory, termed the quasiequilibrium theory (QET),
has associated with it four basic assumptions important not only for the statistical
theory of electron ionization, but for general considerations of unimolecular
reactions in the gas phase. The assumptions of the basic theory are as follows:


Figure 4-34: Simple fragmentation scheme for the formation of b and y Ions. The starting peptide contains 5
amino acids with 6-cleavage and charge migration responsible for b ion formation (charge retained
on the amino terminus), and hydrogen migration from the a-carbon responsible for y ion formation
(charge retained on the carboxy terminus).


Mathieu stability diagram for the quadrupole ion trap. Operated in the mass selective instability
mode ions of successively higher m/z ratio are moved along the a=o line and "pushed over" the
z-stability edge. Higher mass ions (indicated by the larger black circles) > m/z 650 are not
scanned out of the trap during the acquisition period due to the limited rf voltage output (from
reference 261).
Figure 4-21:


Displacement (mm) |Z|
179


26
electrode, thus causing ions of successively higher m/z ratios to be ejected from
the ion trap to the detector (mass selective instability).24,25
Increased resolution and sensitivity can be obtained when axial modulation
is used in conjunction with the mass selective instability scan. Axial modulation
is accomplished by applying an auxiliary ac voltage to the endcap electrodes 180
out of phase.30 This dipolar resonance excitation frequency is set to
approximately 525 kHz just below the 6Z=1 boundary (at 550 kHz) of the stability
diagram. The improved peak shape and resolution obtained is due to presence
of a uniform electric field in the z-direction, which reduces ion shielding and space
charge effects normally seen with the mass selective instability scan where the
field strength at the center of the trap is zero.23,30 This technique has also been
used successfully to extend the mass range of the quadrupole ion trap.73 A
discussion of mass range extension will follow in chapter four of this dissertation.
Other applications for the technique of resonance excitation include notch
filtering27, high resolution36,74, ion isolation74, and collision-induced dissociation75
studies.
Ion isolation can be accomplished using a variety of methods. A
combination of rf and dc voltages applied to the ring electrode can be used for
either apex or two-step isolation.76-79 These two methods utilize the edges of the
stability diagram to preferentially isolate the ion of interest. Resonance excitation
using forward and reverse scans is also used for ion isolation.74 By applying a
high frequency dipolar excitation signal to the endcap electrodes and increasing


210
described above.
The tube lens located at the end of the heated capillary (see figure 4-5)
was used to gate ions into the rf-only octopole (via the skimmer cone). Control
of the voltage(s) applied to the tube/gate lens was accomplished by using the
gate control circuit from the ITMS electronics. The circuit was modified so that
both positive and negative ions could be gated efficiently. For the analysis of
positive ions, a variable positive voltage is applied to the tube/gate lens (10-120
V) to focus the ion beam. To control the pulse width of the beam, a negative -180
V potential is applied to the tube/gate lens to stop ion transmission. For negative
ions, a variable negative voltage (-10 to -100 V) is used for ion focusing and a
+ 180 V signal is used to stop ion transmission (e.g., control the pulse width).
All vacuum electrical connections (capillary heaters/temperature sensors,
capillary offset voltage) for the ESI source were made through a 6" Conflat flange
(equipped with an Amphenol connector) located to the left of the ion source (see
figure 4-3). High voltage for the electrospray needle and drying gas (N2, air, or
02) were controlled manually by an external power/gas distribution unit (Finnigan
MAT, San Jose, CA).
RF-Onlv Octopole
In the previous chapter a detailed description of the theoretical and
practical design considerations of rf-only octopoles was given. This section


396
where protonation occurs on the tertiary nitrogen). The fragmentation pathways
observed for protonated erythromycin are shown in figure 5-21.
The photodissociation spectrum of protonated erythromycin bears an
uncanny resemblance to high-energy CID spectra taken on magnetic sector
instruments.284,308 Protonated species under high-energy collisions give
information on the monosaccharide residues present on the aglycone ring moiety.
These fragmentations typically arise from a charge-remote mechanism. One other
interesting note is the formation of a series of ions (from m/z 200 to m/z 500) in
the photodissociation spectrum indicating ring opening reactions of the aglycone
moiety. At this time reasonable neutral losses and peak assignments have not
been made, and will require an extensive amount of investigation. It is theorized
that these ring opening cleavages are perhaps driven by the photoactivation of
the COC ether group in the aglycone ring (with some form of accompanying
charge migration or charge-remote fragmentation mechanism). No ring
fragmentations have been observed in previously published high-energy CID
spectra. The peak observed at m/z 403 could be the loss of D-desosamine from
the Y0fi at m/z 576 to form the S type ion.


120
100
80
60
40
20
0
Ion Displacement from the Central Axis (mm)
160


Figure 5-10: Structures of the oligosaccharides raffinose (0-a-D-galactopyranosyl[1 -6]-or-D-glucopyranosyl-/?-D
fructofuranoside) and stachyose (cr-D-galactopyranosyl-[1-6]-cr-D-galactopyranosyl-[1-6]-a-D
glucopyranosyl-[1-2]-/?-D-fructofuranoside).


Glycosidic bond cleavages of the ammonium adduct of raffinose for the collisional and photon
activation processes.
Figure 5-13:


57
result of the reaction of the sequential absorption product(s) with neutral sample
molecules. The remaining ions were then mass analyzed via resonance ejection
with qzeject = 0.89 and an amplitude of 1.5 V (zero to peak).26,30
Ion Trap Operation without Helium Buffer Gas
The presence of a light buffer gas (He or H2) at a relatively high pressure
(1 mtorr) has been shown to enhance resolution, sensitivity, and improve
detection limits associated with operation of the quadrupole ion trap.26
Unfortunately, the presence of a buffer gas at 1 mtorr can significantly decrease
photodissociation efficiencies in the infrared region. Since the time between
absorption of consecutive photons for the IRMPD process is on the millisecond
time scale, ions which have acquired some fixed amount of internal energy from
the photon absorption process (but not enough to reach the dissociation
threshold) can undergo collisional damping, thus reducing the photodissociation
efficiency of IRMPD.151,152 Therefore, in order to properly evaluate the performance
of the newly designed multipass-ring electrode, the ITMS was operated without
the presence of He buffer gas so as not to interfere with the photodissociation
process. This section presents a practical dialogue on ion trap operation without
buffer gas and discusses the relevant instrument parameters (without He buffer
gas) which affect trapping efficiency, resolution, and mass range.
One of the most important aspects of no He ion trap operation is trapping
efficiency. The two instrument parameters which affect trapping efficiency the


189
Kinetic Energy
The kinetic energy arguments given here are relative to the transverse
kinetic energy (in the xy plane), which depends on how far the ions move off the
center axis of the rf-only device 253 In figure 3-15253 is shown a plot of the secular
wavelength of motion in an rf-only octopole, with an accompanying plot of
transverse kinetic energy of that motion as a function of distance traveled down
the octopole. As the ion moves to a point farthest away from the center axis of
the octopole, a maximum (for the oscillations) in transverse kinetic energy is
observed (see figure 3-15).253 These observed oscillations at large excursions
arise from contributions to ion motion due to higher order harmonics and not from
the ions fundamental (secular) frequency of motion256
When compared to the rf-only quadrupole (all other conditions being the
same), the magnitude of the transverse kinetic energy oscillations in the rf-only
octopole is much less than in the quadrupole (see figure 3-16). For the constant
conditions of rf voltage and frequency reported in figure 3-16 by Davis and
Wright253, the ion in the rf-only quadrupole approaches its limiting value with
respect to the Mathieu stability parameter at q2=0.908. For the rf-only octopole,
the limiting q4 values from the Mathieu stability diagram237 is approximately 70;
therefore, a significant reduction in transverse kinetic energy is observed even
with operation at higher rf voltages (e.g., 1000 V rf still gives transverse kinetic
energy values less than 1 eV for the simulation shown in figure 3-16).253 It is
advantageous to keep transverse kinetic energy to a minimum so as to keep the


Table 4-1. Voltage power supply and calibration data for ESI/ion trap instrument.
Device
Calibration (y=actual V, x=display V)
Font Panel Label
Voltage Range (V)
Mass Set

Quad 3 Offset
o
7
o
Octopole Offset
y=1.05(x) + 0.0
Quad 2 Offset
+ 12.1 <*-11.5
Endcap Offset
y = 1.04(x) + 0.0
2L3
+ 42 -42
Exit Tube Lens
y=1.02(x) + 0.001
3L2
+ 79 <* -150
ESI Capillary Offset
y=1.00(x) 0.009
3L1
+ 68 <* -145
Ion Gate
y=1.01 (x) + 0.006
3L3
+ 129 <* -131
Axial Modulation
y=3.58(x) 0.049

0 <* 21 (P-P0)*
* ac volatage
257


414
Univ., Kingston, Ont., Canada, 1985.
137. March, R.E.; Young, A.B.; Hughes, R.J.; Kamar, A.; Baril, M. Spectrosc. Int.
J. 1984, 3, 17-32.
138. Kamar, A.; Young, A.B.; March, R.E. Can J. Chem. 1986, 64, 1979-1988.
139. Young, A.B.; March, R.E.; Hughes, R.J. Can. J. Chem. 1985, 63, 2324-
2331.
140. March, R.E.; Hughes, R.J.; Young, A.B. in Proc. 13th Meeting Brit. Mass
Spectrom. Soc. Warwick, U.K., 1983, 77-79.
141. Isenor, N.R.; Richardson, M.C. Appl. Phys. Lett. 1971, 18, 224.
142. Letokhov, V.S.; Ryabov, E.A. Tumanov, O.A. Optics Commun. 1972,5,168.
143. Isenor, N.R.; Merchant, V.; Hallsworth, R.S.; Richardson, M.C. Can. J. Phys.
1973, 51, 1281.
144. Black, J.G.; Yablonovitch, N.; Bloembergen and Mukamel, S. Phys. Rev.
Lett. 1977, 38, 131.
145. Ashfold, N.R.; Hancock, G. in Gas Kinetics and Energy Transfer
Spottiswoode Ballantyne Ltd.: London, 1981, 73-116.
146. Stephenson Jr.; Booth, M.M.; Shalosky, J.A.; Eyler, J.R.; Yost, R.A. in
Proceedings of the 41st ASMS Conference on Mass Spectrometry and
Allied Topics San Francisco, CA, 1993, 445a-445b.
147. Stephenson Jr., J.L.; Booth, M.M.; Shalosky, J.A.; Eyler, J.R.; Yost, R.A. J.
Am. Soc. Mass Spectrom. 1994, 5, 886-893.
148. Stephenson Jr., J.L.; Booth, M.M.; Eyler, J.R.; Yost, R.A. in Proceedings of
the 42nd ASMS Conference on Mass Spectrometry and Allied Topics
Chicago, IL, 1994, 32.
149. Watson, C.H.; Zimmerman, J.A.; Bruce, J.E.; Eyler, J.R. J. Phys. Chem.
1991, 95, 6081-6086.
150. Yates, N.A.; Yost, R.A. in Proceedings of the 39th ASMS Conference on
Mass Spectrometry and Allied Topics Nashville, TN, 1991, 1489-1490.
151. Quigley, G.P. in Chemical Physics 6: Laser Induced-Processes in


310
tandem mass spectrometry experiments were conducted on the peptide
angiotensin I. Human angiotensin I is a decapeptide (DRVYIHPFHL), which is a
pressor substance converted to its active form (angiotensin II) by cleavage of the
HL residues from the carboxy terminus. The human angiotensin I was obtained
from ICN Pharmaceuticals (Costa Mesa, CA), with 700 fmol/pL standards
dissolved in a 50:50 methanol:water solution with 0.4% acetic acid. The
electrospray and instrumental parameters were as described in the Basic ESI
Operation section of this chapter. The ESI spectrum of angiotensin I produced
an intense triply-charged ion at m/z 433, which was mass-isolated using forward-
reverse scans as described in the Ion Isolation section. The MS/MS parameters
for the CID of the triply-charged m/z 433 ion are shown in table 4-8. Figure 4-33
shows the MS/MS spectrum of the triply-charged m/z 433 ion of angiotensin I.
The two major types of fragment ions observed were b ions (charge retention on
the carboxy terminus) and y ions (charge retention on the amino terminus). This
naming convention follows that originally developed by Reopstorff and
Fohlman.279 The primary mechanism of formation for these ions is shown in figure
4-34. In general, formation of either b or y type ions involves cleavage of the
amide bonds along the peptide backbone (see figure 4-34). The proton attached
to the amide nitrogen is typically free to move along the peptide chain. Formation
of b type ions occurs by 6-cleavage and charge migration to the carbonyl end of
the amino terminus fragment. The b ion in figure 4-34 is termed b2 because it
contains two amino acid residues from the original peptide structure (counting


Wavelength Dependence/Infrared Spectroscopy of
Gas-Phase Ions 120
3 THE RF-ONLY OCTOPOLE ION TRANSMISSION GUIDE 129
Ion Behavior in Electromagnetic Fields 131
Octopole Electrode Arrangement 138
Octopole Reid Potentials 141
Electric Field Strength Calculations 145
Electrode Geometry 149
Equations of Motion 150
Octopole Design Considerations 155
Effective Trapping Potential 157
Assembly and Construction 161
Factors Affecting Ion Transmission 169
Initial Entry Angle Conditions 172
RF Amplitude 180
RF Frequency 183
Kinetic Energy 189
Collisional Focusing 194
4 ELECTROSPRAY/ION TRAP INSTRUMENTATION: DESIGN
AND OPERATION 196
General Overview 196
Instrument Design 196
Vacuum Manifold and Pumping System 197
Electrospray Ion Source 204
RF-Only Octopole 210
Analyzer Assembly 226
Detector Assembly 243
Photodissociation Set-Up 246
System Interconnections 254
Instrument Characterization 258
Ion Trap High Mass Theory/Operation 263
Basic ESI Operation 270
Octopole RF Level 282
Octopole Offset 287
Ion Gate Lens 291
RF Level/lon Injection 294
Ion Isolation 303
Collision-induced Dissociation (CID) 309
Negative Ion Mode 322
VI


71
absorption from the applied field, the greater the increase of an ions kinetic
energy (trajectory) for a particular frequency component.
The two most common types of resonance excitation are the dipolar and
quadrupolar techniques. Dipolar excitation (standard on all existing commercial
instrumentation) is performed by applying two auxiliary ac signals 180 out of
phase to the endcaps. Whereas quadrupolar excitation is performed by
application of the same ac signal to both endcaps. Experimentally, both dipolar
and quadrupolar resonance excitation have been used for CID studies, SID
studies, and resonance ejection (axial modulation) in the quadrupole ion trap.
More recently, there has been a concerted effort to understand the nuances of ion
motion during the excitation period.57,156'160
In this dissertation, a unique method using IRMPD for the evaluation of the
coupled motion phenomena observed with resonance excitation is reported. The
basic theory behind ion motion has been reported in chapter one of this work and
elsewhere. The two subsections entitled Dipolar Excitation and Quadrupolar
Excitation briefly discuss the theoretical aspects of resonance excitation, as well
as present relevant photodissociation studies concerning coupled ion motion.
This technique takes advantage of the high photon density in the radial plane of
the ring electrode, which can be used to determine the presence of coupled axial
excitation (a/J while performing radial (o/r) excitation experiments.
All coupled motion data presented in this study are plotted as
photodissociation efficiency versus excitation voltage (dipolar or quadrupolar


Figure 4-32: Results of the forward, reverse, and combined forward-reverse
scans for the isolation of the +18 charge state of horse muscle
apomyoglobin. The unknown fragment peak in the spectrum of
the mass-isolated +18 charge state is due to CID induced by the
increase in kinetic energy of the ion as it approaches the
resonance point.


339
sufficiently to prevent the multiple photon absorption process from promoting the
ion to the dissociation threshold energy. When the buffer gas pressure was
reduced to 2.0x1 O'5 torr, the appearance of a photodissociation spectrum (85 ms
irradiance time) was observed (see figure 5-3).
The fragmentation observed included the mid range b fragments indicative
of the +3 charge state (b5 and b6), the internal y ion series perhaps generated by
the presence of pro7 (HPF/PFH and PFHL fragments), and the y4-y8+2 peak(s).
These fragment ions compare well with what was generated using single
frequency CID experiments discussed in the previous chapter (figure 4-33). One
major difference observed between photodissociation and CID was the presence
of a b9 ion in the photodissociation spectrum. The singly-charged b9 ion is of
particular importance since it could be mass-isolated and a series of MS" studies
performed, where individual amino acids from the peptide chain could be
sequenced in succession by either further photodissociation or CID experiments
as demonstrated in chapter 4. Therefore, generating high mass b ions is of
particular importance for accomplishing a ladder sequencing experiment (removal
of one consecutive amino acid from the peptide chain for each dissociation step)
in order to verify peptide sequences in mass spectrometry and Edman
degradation.
The negative aspect of this experiment is the need for reduced pressure
of the helium buffer gas to obtain a photodissociation spectrum. Indeed,
improved fragmentation efficiencies could be observed for the IRMPD process if


203
mounted to the ITMS table/frame (for use in photodissociation experiments). The
pumps and associated compensators were then mounted to anodized aluminum
o-ring adapter flanges attached to the main vacuum chamber. The
turbomolecular pumps were controlled by two Balzers TCP 300 power supplies
wired directly into the ITMS electronics. Two Alcatel (Alcatel Corporation,
Hingham, MA) mechanical pumps (model 2012A) rated at 300 L/min were used
as fore pumps for the turbomolecular pump system.
A Bayard-Alpert ion gauge (Granville-Phillips, Boulder, CO) was connected
to a modified anodized aluminum o-ring adapter flange located directly under the
analyzer side of the vacuum manifold. A modified Cajon fitting was used to hold
the ion gauge in place directly under the rear part of the ITMS frame to protect
it from damage. The ion gauge controller was a Granville-Phillips model 280
gauge controller (Boulder, CO) equipped with an out-gas control used directly
after system pump-down.
To facilitate alignment of the ESI source, rf-only octopole, and the
analyzer/detector assembly, two precision machined centering rings were welded
inside the vacuum chamber. The front centering ring supports the baffle wall
assembly used both to support the rf-only octopole and to allow differential
pumping (by the two 500 L/s pumps) between the ESI source and the
analyzer/detector assembly. The second centering ring functions as a
support/center for the analyzer/detector assembly (relative to the ESI source).
The accuracy of the center axis of the manifold running from the front 6.0" flange


417
183. Jasinski, J.M.; Rosenfeld, R.N.; Meyer, F.K. Brauman, J.l. J. Am. Chem.
Soc. 1982, 104, 652-658.
184. Rosenfeld, R.N.; Jasinski, J.M.; Brauman, J.l. J. Am. Chem. Soc. 1982,104,
659-663.
185. Watson, C. H. Infrared Multiple Photon Dissociation of Gaseous Ions Studied
by Fourier Transform Ion Cyclotron Resonance Mass Spectrometry Ph.D.
Dissertation, University of Florida, Gainesville, FL, 1986.
186. Ibuki, T.; Sugita, N. J. Chem. Phys. 1983, 79, 5392-5395.
187. Dobler, W.; Ramler, H.; Vinninger, H.; Howorka, F.; Undinger, W. Chem.
Phys. Lett. 1983, 97, 553-556.
188. Freguson, E.E.; Adams, N.G.; Simth, D.; Alge, E.J. J. Chem. Phys. 1984,
80, 6095-6098.
189. Peiris, D.L; Cheeseman, M.A.; Ramanathan, R.; Eyler, J.R. J. Phys. Chem.
1993, 97, 7839-7843.
190. Basic, C.; Eyler, J.R.; Yost, R.A. J. Am. Soc. Mass Spectrom. 1992, 3, 716-
726.
191. Bruce, J.E.; Eyler, J.R. J. Am. Soc. Mass Spectrom. 1992, 3, 727-733.
192. Louris, J.N.; Amy, J.W.; Ridley, T.Y.; Cooks, R.G. Int. J. Mass Spectrom. Ion
Processes 1989, 88, 97-111.
193. Pedder, R.E.; Yost, R.A.; Weber-Grabau, M. in Proceedings of the 37th
ASMS Conference on Mass Spectrometry and Allied Topics Miami, FL,
1989, 468-469.
194. Schwartz, J.C.; Cooks, R.G. in Proceedings of the 36th ASMS Conference
on Mass Spectrometry and Allied Topics San Francisco, CA, 1988, 634-
635.
195. Soni, M.; Cooks, R.G. Anal. Chem. 1994, 66, 2488-2496.
196. Bier, M.E.; Hartford, R.E.; Herron, J.R.; Stafford, G.C. in Proceedings of the
39th ASMS Conference on Mass Spectrometry and Allied Topics Nashville,
TN, 1991, 538-539.
197. Kaiser, R.E.; Williams, J.D.; Schwartz, J.C.; Lammert, S.A.; Cooks, R.G. in


291
shown a plot of ion intensity versus octopole offset for the even charge states of
the protein ubiquitin. The electrospray conditions for ubiquitin were the same as
in the previous section (see table 4-4). The trend observed in figure 4-27 gave
an optimum ion transmission for the smaller m/z 709 (+12 charge state) ranging
from -0.2 V to -2.0 V (e.g., flat portion of the curve). For m/z 1417 (+6 charge
state), the optimum voltage for ion transmission was -2.8 V. As expected, more
negative voltages were need to maximize the injection efficiency of the larger m/z
ions. The relatively small voltage optimization range (2.5 V) for the multiply-
charged ions of ubiquitin could be attributed to the flat portion of the radial
potential well of the rf-only octopole (see chapter 3). This flat portion of the well
allows a large number of different m/z ions (all at the same kinetic energy) to be
transmitted efficiently through the device.
Ion Gate Lens
The function of the ion gate lens, which is located between the ESI heated
capillary and skimmer cone (see figure 4-3), is to pulse ions through the skimmer
cone region and into the rf-only octopole for transmission to the ion trap analyzer.
In figure 4-28 is shown a plot of ion intensity versus the electrospray gate lens
voltage for the even positive ion charge states of the protein ubiquitin. The
electrospray conditions for ubiquitin were the same as in the Basic ESI Operation
section of this chapter (see table 4-4). Typical voltages applied during the ion


Top V¡ew(s)
Octopole Mount
(Analyzer Assembly)
Side View(s)
All Units are in Inches
238


Analyzer Mounting Plate (Bottom)
Dimensions of PulsedVaive/Feed Through Holes
Number: 1
Material: Aluminum
Modifications: Anodize
All Units are in Inches
Page 5 of 5
235


Figure 4-26:
Effect of rf voltage on ion transmission efficiency. Past a detected rf value of 0.2 V, the signals for
all the charge states of ubiquitin level off and are independent of the rf voltage. The extended
stability region of the octopole helps make the flat response curves possible.


(a) El mass spectrum of diglyme at a pressure of 3.6x1 O'7 torr and 15 ms ionization time (no He
present), (b) IRMPD spectrum of protonated diglyme at 944 cm1 and 10 ms irradiance time
(energy = 0.201 J). (c) IRMPD spectrum of protonated diglyme at 944 cm'1 and 40 ms irradiance
time (energy = 0.852 J).
Figure 2-13:


Oligosaccharide Fragmentation Nomenclature
Nonreducing End
1,5
- X
ch2oh/
- Y
/ CH2OH
368


Figure 4-28:
Effect of the electrospray gate lens on ion transmission. The observed maximum transmission
efficiency for the even charge states of ubiquitin varies only by 20 V across a mass range of 700.


200
The vacuum manifold was assembled from pieces of Finnigan MAT 4500
and TSQ 46 vacuum manifolds. The manifold was made entirely of stainless
steel, with the majority of flange connections made with Conflat (copper
gasket/knife edge) style matings to achieve the highest vacuum possible (needed
to do IR photodissociation). A total of 10 connection ports were made to the
vacuum manifold. The Conflat connections included an 8.0" port for the
analyzer/detector assembly, a 4.5" port for the rf (drive frequency) feedthrough,
a spare 4.5" port for a second ion gauge, a 6.0" port for a Conflat adapter flange
needed for the ESI interface, another 6.0" port for electrical connections to the ESI
interface (two heater connections, two temperature sensor connections, and a
heated capillary offset connection) and rf-only octopole (two low voltage rf
connections, <5kV), and a 2.75" port for a removable laser window system. 0-
ring connections included two large 8.0" ports (located on the bottom of the
manifold) for attachment of two 500 L/s turbomolecular pumps and two 5.0" ports
(top of manifold) for glass windows to view the differentially pumped regions of
the system. A schematic diagram of the vacuum manifold is presented in figure
4-2.
The pumping system for the manifold consisted of two 500 LVs
turbomolecular pumps (model TPH 510, Balzers, Hudson, NH) mounted to two
turbomolecular pump compensators (Balzers, Hudson, NH) in order to damp the
vibrations associated with turbomolecular pump operation. These pump
vibrations were minimized to limit their effects on any optical lens/mirror systems


63
significantly less resonant excitation amplitude (55 mV for no He) compared with
that of the same experiment with He buffer gas (245 mV) (see figure 2-5). In
addition, large shifts in the maximum ejection frequency were observed. These
shifts were attributed to a lack of helium buffer gas, which was needed for a well
defined frequency distribution. Also observed in figure 2-5, frequency bandwidths
were much greater (approximately 0.4 kHz with He) for the case where no He
buffer gas was present (2 kHz). All frequency data in the experiment were taken
for a constant number of ions stored in the trap (as measured by the electron
multiplier and detector circuits) so as to minimize bandwidth and frequency shift
phenomena typically observed with different ion populations. Although the
absolute number of counts for protonated diglyme (m/z 135) was matched in the
profile scan mode for the case of He buffer gas versus no He buffer gas, there
could be substantially more ions present which were not detected for the
aforementioned reasons in this section.
The Multi-Pass Ring Electrode
The 1.0-cm-diameter IR laser beam (Gaussian profile) was attenuated by
a 0.3 cm (1/8") entrance aperture on the ring electrode. The laser was aligned
such that the center portion of the Gaussian beam profile passed through the
entrance aperture. The ring electrode was modified by incorporation of three
polished stainless steel spherical concave mirrors (radius of curvature = 2.0 cm)
mounted on the inner surface of the ring, as shown in figure 2-6. The