UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Flourescence spectroscopy of trapped molecular ions Wright, Kenneth Charles 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2004-902889.pdf [ 11.55MB ]
Metadata
JSON: 831-1.0061168.json
JSON-LD: 831-1.0061168-ld.json
RDF/XML (Pretty): 831-1.0061168-rdf.xml
RDF/JSON: 831-1.0061168-rdf.json
Turtle: 831-1.0061168-turtle.txt
N-Triples: 831-1.0061168-rdf-ntriples.txt
Original Record: 831-1.0061168-source.json
Full Text
831-1.0061168-fulltext.txt
Citation
831-1.0061168.ris

Full Text

Fluorescence Spectroscopy of Trapped Molecular Ions by Kenneth Charles Wright B.Sc, Dalhousie University, 1993 A THESIS SUBMITTED IN PARTIAL F U L F I L M E N T OF T H E REQUIREMENTS FOR T H E D E G R E E OF DOCTOR OF PHILOSOPHY in The Faculty of Graduate Studies (Department of Chemistry) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA December 22, 2003 © Kenneth Charles Wright, 2003 Abstract This thesis describes the development of a unique instrument capable of detecting fluorescence emission from large gas phase molecular ions trapped in a three-dimensional quadrupole ion trap. The hypothesis that has formed the basis of this work is the belief that fluorescence spectroscopy can be combined with ion trap mass spectrometry to probe the structure of gas phase molecular ions. The ion trap provides a rarefied environment where fluorescence experiments can be conducted without interference from solvent molecules or impurities. Although fluorescence was not detected during preliminary experiments, two significant experimental challenges associated with detecting the gas phase fluorescence of ions were discovered. First, gas phase ions were vul-nerable to photodissociation and low laser powers were necessary to avoid photodissociation. Since fluorescence emission is directly proportional to laser intensity, a lower laser power limits the fluorescence signal. Second, the fluorescence emission was not significantly Stokes shifted from the excitation. The lack of Stokes shift meant the small fluorescence signal must be detected in the presence of a large amount of background scatter generated by the excitation. Initially, this background was seven orders of magnitude higher than the analytical signal ultimately detected. A specially designed fiber optic probe was inserted between the electrodes of the ion trap to stop light scattered off the outside surfaces of the trap from reaching the detector. The inside surfaces of the ion trap were coated black to further reduce the amount of scattered light collected. These innovations helped reduced the background by six orders of magnitude and fluorescence emission from rhodamine-6G was detected. Pulse counting experiments were used to optimize fluorescence detection. The effects of trapping level, laser power, and irradiation time were investi-Abstract iii gated and optimized. The instrument developed in this work not only allows for the detection of fluorescent photons, but the sensitivity is high enough for the light to be dispersed and an emission spectrum recorded. The emission spectra of rhodamine-6G and 5-carboxyrhodamine-6G ions reported in this thesis represent the first spectra recorded from large molecular ions confined in a quadrupole ion trap. Finally, anti-Stokes fluorescence from rhodamine-6G was also detected. iv Table of Contents Abstract ii Table of Contents iv List of Tables x List of Figures xi List of Abbreviations xv Acknowledgements xviii 1 Introduction 1 1.1 Thesis Objective 1 1.2 Background 3 1.3 Thesis Overview 8 2 Mass Spectrometry • 10 Table of Contents v 2.1 Quadrupole Ion Trap Theory 12 2.1.1 The Mathieu Equation 15 2.1.2 Regions of Ion Trajectory Stability 21 2.1.3 Secular Frequencies and Resonant Excitation 23 2.1.4 Stretched Ion Traps 25 2.2 Ion Trap Operation 27 2.2.1 Mass Selective Instability 28 2.2.2 Buffer Gas and Collisional "Cooling" 32 2.2.3 Mass Selective Ion Isolation 33 2.2.4 CID and MS/MS 36 2.3 Linear Quadrupoles . 37 2.3.1 Linear Quadrupole Theory 37 2.3.2 Linear Quadrupole Operation 40 2.4 The Ion Source 41 2.4.1 Thermal Desorption and Electron Impact Ionization . . 43 2.4.2 Electrospray Ionization 45 2.5 MS Instrumentation 47 2.5.1 Vacuum System 48 2.5.2 Ion Optics 50 Table of Contents vi 2.5.3 Ion Trap Control 53 3 Spectroscopy and Optics 54 3.1 Optical Access to the Trapping Volume 56 3.2 Optical Detection 58 3.3 First Experiments with PAH Cations 59 3.3.1 Methods 61 3.3.2 Photodissociation 65 3.3.3 Background Scatter 68 3.3.4 Unwanted Neutrals 70 3.3.5 Results from PAH Experiments 73 3.4 Preliminary Experiments Using ESI 75 3.4.1 Ionic Dyes 76 3.4.2 Photodissociation Studies 79 3.4.3 Fluorescently Tagged Peptides 89 3.5 Requirements for a Successful Fluorescence Experiment . . . . 92 4 Pulse Counting Experiments 95 4.1 Fiber Optic Probe 99 4.1.1 P M T Detector for the Fiber Optic Probe 104 Table of Contents vii 4.2 Coating the Ion Trap Electrodes 105 4.3 Pulse Counting Instrumentation 109 4.3.1 Data Collection for Pulse Counting 112 4.3.2 P M T Cooling . 114 4.4 Fluorescence vs. Laser Power 115 4.4.1 Instrumental Parameters 116 4.4.2 Results and Discussion 116 4.5 Fluorescence vs. Trapping Level (qz) 122 4.5.1 Results and Discussion 122 4.6 Fluorescence vs. Laser Power and qz 124 4.6.1 Results and Discussion 125 4.7 Irradiation Time 127 4.7.1 Results and Discussion 127 4.8 Final Optimization 129 4.8.1 Complete Optimization at Two Pressures 130 4.8.2 Results and Discussion 131 4.9 Fluorescence From Fragments 135 4.10 Chopping Experiments 137 4.11 Conclusions 138 Table of Contents viii 4.12 Perspective 140 5 Fluorescence Emission Spectroscopy 143 5.1 Instrumentation 146 5.1.1 Spectrograph and ICCD 146 5.1.2 Methods 151 5.1.3 Solution Spectra 152 5.2 Emission Spectra with Notch Filter 154 5.2.1 Discussion 155 5.3 Emission Spectra without Filter 161 5.3.1 Discussion 163 5.4 Anti-Stokes Fluorescence 166 5.4.1 Discussion 168 5.5 Conclusions 172 5.6 Perspective 173 6 Summary and Conclusion 175 6.1 Summary 175 6.2 Conclusion 178 6.3 Future Directions 182 Table of Contents ix 6.3.1 Improvements to the Instrument 182 6.3.2 Applications 184 6.3.3 Concluding Remarks 185 Bibliography 186 List of Tables 2.1 Suitable operational voltages for singly charged ions. 4.1 Instrument settings used for pulse counting experiments. xi List of Figures 2.1 Ideal hyperbolic electrode shapes for the 3D quadrupole and the linear quadrupole (identical field radii (r0) shown) 12 2.2 Ion trap electrodes 13 2.3 Cross-section of ion trap electrodes 14 2.4 Regions of stability defined by a and q 23 2.5 Stability defined by the overlap of the r-stability and z-stability. 24 2.6 Stretched ion trap 26 2.7 Linear quadrupole rod set 38 2.8 Round rods approximate the hyperbolic shape of ideal quadrupole electrodes 39 2.9 Thermal desorption probe. 44 2.10 ESI Front-end 46 2.11 Mass spectrometer 48 List of Figures xii 3.1 Instrumentation used to study PAH cations 60 3.2 The matrix isolated cation absorption spectra of J3enzo(ghi)perylene and Tetrabenzo(acfh)napthalene 62 3.3 Timing diagram for early optical experiments on PAH cations. 64 3.4 Photodissociation of Tetrabenzo(acfh)napthalene 66 3.5 Photodissociation of Benzo(ghi)perylene 67 3.6 Chaotic background from laser scatter and neutral emission recorded during PAH experiments 72 3.7 Structures of 5-CR 6G and R6G 76 3.8 The solution and gas phase absorbance of 5-CR 6G 78 3.9 Photodissociation of 5-CR 6G 82 3.10 Photodissociation of the molecular ion and first fragment of 5-CR 6G 83 3.11 Buffer gas pressure and photodissociation 84 3.12 Heavier buffer gases and photodissociation 87 3.13 Chopping the laser beam and photodissociation 89 3.14 Structures of fluorescently tagged peptides FITC-Gly-Gly-His and FITC-Gly-His 91 3.15 Photodissociation and CID of FITC-Gly-His 92 List of Figures x i i i 4.1 Fiber optic probe 103 4.2 Carbon coating of electrodes 107 4.3 Instrumental configuration for pulse counting experiment. . . . 110 4.4 Instrumental block diagram for pulse counting experiment. . .111 4.5 Timing diagram for pulse counting experiments 114 4.6 Fluorescence signal vs. laser power for qz — 0.5 118 4.7 Fluorescence signal, background signal, and S/N ratio 120 4.8 Background subtracted fluorescence signal from 5-CR 6G vs. qz for a fixed laser power 123 4.9 Optimizing laser power and ion cloud size 125 4.10 Fluorescence measured in 1 s intervals of a 10 s experiment. . 128 4.11 Optimization of qz at 1 mtorr and 0.55 mtorr 133 4.12 Timing diagram to test fragment ions for fluorescence 137 5.1 Final version of the instrument 147 5.2 Photograph of the instrument with the P M T attached to the fiber optic probe and spectrograph ready to be connected. . . 149 5.3 Solution absorbance spectra of R 6 G and 5-CR 6G 153 5.4 Solution emission spectra of R6G and 5-CR 6G 153 5.5 Emission spectra collected with the notch filter 156 List of Figures xiv 5.6 R6G emission (left axis) and notch filter transmittance (right axis) 158 5.7 R6G emission spectra with and without photodissociation. . . 160 5.8 R6G and 5-CR 6G emission spectra 160 5.9 Fluorescence emission spectra of R6G and 5-CR 6G ions. . . . 162 5.10 R6G fluorescence emission spectrum without background sub-traction 165 5.11 R6G fluorescence emission spectrum without background sub-traction (expanded vertical scale) 165 5.12 Effect of a long pass filter 167 5.13 Anti-Stokes fluorescence 169 5.14 R6G ion emission from 488.0 nm and 496.6 nm excitation (left axis). The notch filter transmittance shown for reference (right axis) 171 6.1 Energy level diagrams for fluorescence experiments 180 List of Abbreviations Abbreviation Description 5-CR 6G 5-carboxyrhodamine 6G AC alternating current ADC analog to digital converter ASMS American Society for Mass Spectrometry CCD charged coupled device CID collision-induced dissociation CPS counts per second C W continuous wave DAC digital to analog converter DC direct current DIB diffuse interstellar band ECL emitter coupled logic EHD electrohydrodynamic List of Abbreviations xvi Abbreviation Description ESI electrospray ionization F C 4 3 perflourotributylamine FITC-Gly-Gly-His fluorescien-glycine-glycine-histidine FITC-Gly-His fluorescien-glycine-histidine fl focal length FNF filtered noise field F R E T fluorescence resonance energy transfer FT-ICR fourier transform ion cyclotron resonance GFP green fluorescent protein HPLC high performance liquid chromatography ICCD intensified charged coupled device ICR ion cyclotron resonance IVR internal vibrational redistribution MCP microchannel plate MS/MS two successive mass selective operations MSn n mass selective operations m/z mass/charge List of Abbreviations xvii Abbreviation Description O.D. optical density PAH polycyclic aromatic hydrocarbon P M T photomultiplier tube R6G rhodamine 6G RF radio frequency S/N signal to noise ratio TOF time of flight TTL transistor transistor logic UBC University of British Columbia xvm Acknowledgements Many thanks are extended to my academic colleagues and mentors. I wish to thank my supervisor Mike Blades for all his assistance. I also want to express my appreciation to Don Douglas who has been very generous with his time and laboratory equipment. Special thanks must go to John Hepburn, for providing many informative tutorials on molecular spectroscopy, allowing me to participate in group meetings, and especially for giving me the idea to coat the ion trap electrodes with carbon. Bruce Todd also deserves a special mention. Bruce has been a great source of technical information over the years. I want to express my gratitude to the University of British Columbia and the Natural Science and Engineering Research Council for funding my work. I also acknowledge the significant contribution provided by other members of the UBC Chemistry Department for their professional support and friend-ship during my years at UBC: John Ellis, Lance Jones, Elizabeth Varty, Brian Ditchburn, Elena Polishchuk, Ben Clifford, Tim Daynard and especially Lani Collins (sorry about that email!) you are wonderful people to know and work with. For me, there cannot possibly be a better place to build an analytical instrument than the UBC Chemistry Department. The shops and service labs are staffed by truly excellent people. The members of the Electronic and Mechanical Engineering Services shops deserve a great deal of credit for the instrumental achievements described in this thesis. I will forever be indebted to the outstanding members of the Mechanical shop: Des Lovrity, Oskar Greiner, Ed Gomm, Ken Love, Ron Marwick, and Brian Snapkauskas. Truly kings among men. From the Electronics shop I want to thank Dave Bains, Jason Gozjolko, Roily Chan, Mike Hatton, Milan Coschizza, Martin Carlisle and especially the dynamic duo of Brian Greene and Dave Tonkin. Acknowledgements xix "Greener" and "Fish" are not only excellent technicians and teachers, but have become great friends. Finishing my work would have been very difficult without the superb treatment provided by my Dream Team of medical professionals: physicians Dr. Peter Grantham and Dr. Judy Stringer, rehabilitation specialist Dr. John Le Nobel, physiotherapist Tyler Dumont and chiropractor Dr. David Kavlheim. My lawyers Anthony Vecchio and Art Vertlieb also deserve thanks for their role in assembling this medical team. Thanks must also go to my family and friends for their constant support. In particular I want to thank my parents, Don and Shirley, who have been very patient and my sister Elaine for all her hard work proofreading my thesis. Shane Greek has also been a great friend to have around. Saving the best for last, I want to thank my wife Holly. Without Holly's emotional and financial support I could not have finished my work. Holly has been "plunking quarters" into me the past few years hoping that I would eventually "pay off". I may have hit a few snags along the way, but in the end I have achieved my goals: to make a significant contribution to science, secure a position at an excellent research facility, and most importantly to have a truly wonderful wife. X X For Holly 1 Chapter 1 Introduction 1.1 Thesis Objective Instrumentation development has long been a cornerstone of analytical chem-istry. Critics of analytical chemistry often confuse instrumentation science with Edisonian tinkering. In fact, many instrument developments are hypothesis-driven. Hypotheses are formulated from the detailed knowledge of how in-struments function and by having a fundamental understanding of the chem-ical and physical phenomena selected for study. The hypothesis that has driven this work is the belief that fluorescence spectroscopy can be combined with ion trap mass spectrometry to probe the structure of gas phase molecular ions. The ion trap mass spectrome-ter provides a rarefied environment where fluorescence experiments can be conducted without interference from solvent molecules or impurities. Investi-gating the molecular fluorescence of trapped ions will be useful in many areas: Chapter 1. Introduction 2 studying the conformation of biomolecules, identifying potential diffuse in-terstellar band (DIB) carriers, and for providing information that would be complementary to other molecular spectroscopies. The gas phase three-dimensional structure of biomolecules could be in-vestigated by monitoring conformationally dependent native fluorescence of proteins. Fluorescence resonance energy transfer (FRET) L ' 2 experiments to probe the structure of gas phase biomolecules has also been proposed.3 A F R E T experiment involves the transfer of electronic excitation energy from a donor molecule to an acceptor chromophore. The energy transfer occurs by dipole-dipole resonance interactions and depends on the separation distance and relative orientation of the donor and acceptor molecules (FRET pairs). By selectively tagging a biomolecule with a FRET pair and monitoring the fluorescence emission, the three-dimensional structure of the biomolecule can be investigated. Fluorescence spectroscopy and ion trap mass spectrometry could also be used to evaluate potential DIB carriers. The astrophysics community is in-terested in identifying the molecules responsible for the broad emission bands observed in the interstellar medium. Polycyclic aromatic hydrocarbon (PAH) cations are considered to be potential DIB candidates.4'5 An ion trap could Chapter 1. Introduction 3 be used to simulate interstellar conditions and investigate the fluorescence emission of PAH cations. The advantages of combining ion trap mass spectrometry and fluorescence emission spectroscopy cannot be applied to any investigation without first developing a suitable instrument. The objective of this work was to develop an instrument capable of detecting fluorescence emission from large gas phase molecular ions confined in an ion trap mass spectrometer. 1.2 Background The origin of mass spectrometry comes from J.J. Thomson's work at Cam-bridge University in the early twentieth century.6 From his discovery of the electron (using a cathode ray tube) to his investigations using beams of pos-itive ions, Thompson set the groundwork for the field of mass spectrometry and predicted its impact on chemistry many years before it was realized. "I have described at some lengths the applications of positive rays to chemical analysis: one of the main reasons for writing this book was the hope that it might induce others and especially chemists, to try this method of analysis. I feel sure that there are Chapter 1. Introduction 4 many problems in chemistry which could be solved with much greater ease by this than by any other method. This method is surprisingly sensitive - more so even than that of spectrum analysis, requires an infinitesimal amount of material and does not require this to be especially purified; the technique is not difficult if appliances for producing high vacua are available " J.J. Thompson6 F.W. Aston, a student of Thomson, built a mass spectrograph based on velocity focusing to study stable isotopes. Around the same time (1914-1918) A.J . Dempster, a Canadian working at the University of Chicago, developed the first sector mass spectrometer. Mass spectrometric techniques were soon applied to important problems such as the hydrocarbon analysis of aircraft fuel during World War Two. As scientists became aware of the potential of mass spectrometric analysis, fundamental studies in the field thrived and new instruments were developed. In the 1950's Wolfgang Paul and co-workers invented a new type of mass analyzer called a quadrupole mass spectrometer.7 Chapter 1. Introduction 5 Although 3D quadrupole ion traps (also known as Paul traps) are primar-ily used as mass spectrometers they have also been used to confine ions for spectroscopic interrogation.8 - 1 3 It has been argued that trapped gas phase ions are close to the ideal spectroscopic sample of an isolated species at rest.14 A quadrupole ion trap offers mass selectivity and excellent spatial and tempo-ral confinement of ions without introducing perturbations that would impact negatively on the spectroscopy.14 While these capabilities have been exploited to conduct fluorescence experiments with atomic and small molecular ions, little progress has been made with larger molecular ions. For the purpose of this work molecules in the range of hundreds Dal-tons are considered to be large molecules. Many significant experimental challenges must be overcome in order to realize the advantages of using a quadrupole ion trap to confine large molecular ions for spectroscopic inter-rogation. One problem common to all types of ions confined in an ion trap is low ion density (typically, [ion] < 5 x 1 0 _ 1 0 M ) . In order to detect fluorescence from such a small number of ions the fluorescence must be efficiently sep-arated from the excitation. Atomic ions can be excited by a multiphoton process that would result in the resonant fluorescence having a wavelength Chapter 1. Introduction 6 (A) much less than the excitation (for a two photon excitation, emission A = \ excitation A). Single atomic ions have been detected this way.8 Emis-sion from atomic ions can also be detected by monitoring a transition lower in energy than the excitation.3 Alternatively, since atomic ions and small molecules have excited state lifetimes ranging from several tens of nanosec-onds to seconds, a pulsed excitation source can be used. The collection of the fluorescence emission is delayed allowing the small fluorescence signal to be detected without interference from the excitation pulse.1 5 These spectro-scopic techniques used to study atomic ions and small molecular ions trapped in the gas phase are not appropriate for larger ions. An excitation scheme similar to the multiphoton process used for atomic ions leads to rapid heat-ing and photodissociation of molecular ions. Larger molecules typically have excited state lifetimes of a few nanoseconds and it is therefore very difficult to resolve temporally their emission from the excitation.1 6 In condensed phase experiments fluorescence detection is limited by the background emission from impurities and the elastic or inelastic scattering from the solvent or matrix. 1 7 An ion trap mass spectrometer can eliminate solvent and impurities from the probe volume. However, the lack of solva-tion presented unexpected difficulties. In condensed phase experiments any Chapter 1. Introduction 7 excess vibrational energy acquired by a molecule is efficiently transferred to the surrounding medium. In gas phase experiments, however, internal vi-brational redistribution (IVR) redistributes excess internal energy, but the molecule remains vibrationally excited. Vibrational cooling of the ions can take place through collisions with buffer gas molecules or by infrared emis-sion but these processes are very slow and would likely not occur between absorption and emission events. Continuous excitation, such as that provided by a continuous wave (CW) laser, causes rapid heating of the ions and leads to photodissociation. To avoid unwanted photodissociation, laser irradiances are typically four orders of magnitude lower than those that can be used in solution experiments. Since the molecule cannot relax to the lowest vibrational level of the excited stated before emission occurs, the fluorescence will not be significantly Stokes (wavelength) shifted from the excitation. For solution experiments, filters that have sufficient optical density at the excitation wavelength and good transmission at the emission maximum are common. As the Stokes shift decreases it is more difficult to find a filter with enough attenuation at the excitation wavelength without also reducing the total fluorescence detected. The gas phase absorption spectrum of a molecule Chapter 1. Introduction 8 may also be different from its solution absorbance spectrum. While a shift in absorbance maximum is an important effect to consider, it is a relatively minor problem to solve compared to those already mentioned. The two main problems of photodissociation and the lack of a significant Stokes shift in the fluorescence emission were not fully appreciated when the instrument was first designed. While initial attempts to detect fluorescence from trapped ions were unsuccessful, many crucial lessons were learned. The knowledge acquired from these initial experiments ultimately led to a better understanding of ion trap instrumentation and the photophysics and spec-troscopy of gas phase molecular ions. It was then possible to successfully redesign the instrument and detect fluorescence from large molecular ions. This new instrumentation not only allows for the collection of fluorescent photons, but the sensitivity is high enough for the light to be dispersed and a fluorescence spectrum to be collected. 1.3 Thesis Overview This thesis describes the development of a unique instrument capable of detecting fluorescence emission from large gas phase molecular ions trapped Chapter 1. Introduction 9 in a three-dimensional quadrupole ion trap. Chapter 2 reviews the theory of quadrupole ion traps and details the development of the ion trap mass spectrometer used to confine ions. Chapter 3 provides a basic discussion of luminescence spectroscopy and describes the development of the optical system. It also contains a description of early experiments with PAH cations and ionic dyes. Although those exper-iments did not generate a detectable fluorescence signal, the vital knowledged acquired was used to rationalize improvements to the instrument. Chapter 4 outlines the instrumental modifications and methods used in the first successful fluorescence experiments. Pulse counting experiments used to optimize fluorescence detection are also described. In addition, the effects of trapping level, laser power, and irradiation time are reported. Chapter 5 details the instrumentation and methods used to collect the emission spectra of rhodamine 6G and 5-carboxyrhodamine 6G ions. Fi-nally, Chapter 6 provides a summary of the thesis and describes potential applications of the instrument developed. Chapter 2 Mass Spectrometry 10 A mass spectrometer consists of a source of gas phase ions, an analyzer that separates ions according to their mass/charge (m/z) ratio and a detector that determines the fraction of the total ion current carried by each of the ions. The mass/charge ratio of an ion is defined as: which has units of Thomson (Th). There are many different combinations of ion sources, mass analyzers and detectors available. The mass analyzer and detector regions of a mass spectrometer are contained in a vacuum system while the ion source may be external to the vacuum chamber. A trapping device, the quadrupole ion trap, was chosen to mass select and confine molecular ions for optical spectroscopy. The final version of the instrument also employs two linear quadrupoles for m/z mass(amu) elementary charge Chapter 2. Mass Spectrometry 11 ion storage and ion transmission. The ion source originally utilized a thermal desorption probe and electron impact ionization but interference from neutral molecules motivated the change to an electrospray ionization (ESI) source. According to Earnshaw's theorem no combination of static electric fields can be used to confine an electric charge. Quadrupoles are dynamic trapping devices that are capable of confining ions because they use time varying electric fields. Since the three-dimensional quadrupole and linear quadrupole can be used to confine ions they both can be called ion traps. In this thesis the three-dimensional quadrupole will be referred to as a quadrupole ion trap, ion trap, 3D quadrupole, or Paul trap. The linear quadrupole will be referred to as a linear trap or linear quadrupole. Of course, the 3D quadrupole and linear quadrupole are very similar. The quadrupole potential used for mass selection and ion confinement is generated by applying radio frequency voltages (Vrf) to hyperbolic elec-trodes. The ideal electrode shapes are shown in Figure 2.1. These truncated electrodes will produce an almost perfect quadrupole field near the centre of the trapping volume. Chapter 2. Mass Spectrometry 12 3 D Quadrupole Linear Quadrupole Figure 2.1: Ideal hyperbolic electrode shapes for the 3D quadrupole and the linear quadrupole (identical field radii (r0) shown). 2.1 Quadrupole Ion Trap Theory A quadrupole ion trap can be used as a mass selective ion storage device or as a versatile mass spectrometer. Although the ion trap may seem to be a very simple device to understand and operate, there are many subtleties that require closer attention. Descriptions of quadrupole ion trap theory and operation are reviewed in several textbooks and papers. 1 8 - 2 0 A qualitative description of the ion trap and its operation will precede a review of the relevant theory. Chapter 2. Mass Spectrometry 13 Figure 2.2: Ion trap electrodes. Endcap and ring electrode shown together on left. Single endcap electrode shown on right. White quartz spacers define electrode spacing when assembled. Figure 2.2 shows the electrodes of the ion trap used in the instrument. Two identical endcaps are positioned symmetrically on either side of the ring electrode. A radio frequency (RF) potential (K / ) is applied to the ring electrode while the endcaps are essentially grounded. Figure 2.3 shows a cross sectional view through the centre of symmetry of the ion trap used in this thesis. The force (F) acting on an ion of charge e in an electric field (£?) is defined as: Fu = -et = -e dcp du (2.2) Chapter 2. Mass Spectrometry 14 ring entrance exit endcap endcap Figure 2.3: Cross-sectional view of ion trap electrodes. RF-potential (Vrj) applied to ring while endcaps are grounded. where </> is the electric potential and u represents any direction. Knowing Newton's second law: Fu = mau (2.3) one can easily predict the direction an ion will move in an electric field. When ring electrode has a positive potential, a positive ion will experience a force directed away from the ring and toward one of the endcaps. When the ring has a negative potential, the same ion will then be attracted to Chapter 2. Mass Spectrometry 15 the ring and move away from the endcaps. Trapped ions will experience small deviations in their trajectories, commonly referred to as micromotions, with the same frequency as the applied RF potential. This micromotion is superimposed on a larger trajectory as the ion completes a three-dimensional "figure-of-eight".21 Ions will experience forces proportional to the charge they carry and the acceleration experienced as a result of these forces will depend inversely on their mass. Ions of different m/z ratios will complete figure-of -eights at unique secular frequencies. Ions can be ejected from the trap based on their secular frequencies and an external detector can be used to record a mass spectrum. 2.1.1 The Mathieu Equation The general operating principles of a quadrupole mass spectrometer can be best understood after a formal treatment of the physics involved. The motion of ions in quadrupole fields are described by second-order differential equa-tions. Conveniently, in 1868 E. Mathieu solved a second-order differential equation while studying the resonant frequencies of drums.2 2 Mathieu's so-lution can be used by simply rearranging the equation describing the motion of ions into the canonical form of the Mathieu equation: Chapter 2. Mass Spectrometry 16 d2u — + (au - 2qu cos 2 t> = 0 (2.4) The potential (</>) at any point within the quadrupole field can be de-scribed by the equation: K,y,z — o-x2 + ay2 + bz2 (2.5) where a is the weighting constant for the equivalent x and y coordinates and b is the weighting constant for the z coordinate. Any electric field that does not contain point charges must be linear in all directions and satisfy the Laplace condition (\/2<p = 0)- Therefore: 5x2 by2 8z2 From substituting Equation 2.5 into Equation 2.6: V2c£ = 2 a + 2 a + 2& = 0 (2.7) 4a+26 = 0 (2.8) Chapter 2. Mass Spectrometry 17 6 = - 2 a (2.9) and by substitution back into Equation 2.5 the potential at any point be-comes: <frx,y,z = 0 x 2 + ay2 ~~ 2az2 (2-10) At this point it is convenient to make the usual substitutions (x = r cos 9 and y = rsm9) to convert Equation 2.10 to cylindrical coordinates. The cylindrical coordinates are defined in Figure 2.3. 4>rt2 = a(r cos Of + a(r sin <9)2 - 2az2 (2.11) </vi2 = ar2(cos2 d + sin 2 9) - 2az2 (2.12) Since cos2 9 + sin 2 9 — 1 Equation 2.12 becomes: </»r,2 = ar2 - 2az2 (2.13) If the ion trap was operated by applying the potential ((f)) to the ring and the same potential 180° out of phase (—</>) on the endcaps. The weighting fac-tor (a) could be determined by substituting the known boundary conditions Chapter 2. Mass Spectrometry 18 into Equation 2.13. Since the endcaps are grounded, a constant (w) must be added to Equation 2.13 in order to satisfy the boundary conditions. It is important to note that this constant (w) does not appear in the equations of motion. 4>r,z = or2 - 2az2 + w (2.14) The endcap spacing is defined as 2za; r0 is inner radius of the the ring elec-trode and 4>0 is the electric potential applied to the ring electrode while the endcaps are grounded. Therefore: <l>(ro,0) =4>o = ar20 + w (2.15) </>{0,zo) = 0 = -2az2Q + w (2.16) Subtracting Equation 2.16 from Equation 2.15 gives: b0 = ar20 + 2az2Q a = r2 + 2z2 ' O ' O (2.17) (2.18) Chapter 2. Mass Spectrometry 19 Substituting Equation 2.18 into Equation 2.15 gives: ™=4^4-2 (2-19) rl + 2zl V ' The potential in the ion trap can be finally expressed by the following equa-tion: ^ - 2 z 2 ) + ^4r9 (2.20) Since there are no terms in Equation 2.20 that include both r and z, the equations of motions in each direction can be treated independently. The force experienced by an ion in any direction («) can be expressed as: Fu = mau = -eEu (2.21) The equations of motion are therefore: Chapter 2. Mass Spectrometry 20 The applied potential will now be defined as a combination of a DC potential, U, and an RF potential, VcosQt, where Q is the angular frequency of the RF potential (Q — 2irf). Here, V is the 0-peak voltage of the RF potential. <f>0 = U + Vcosnt (2.25) After substituting Equation 2.25 into Equations 2.23 and 2.24 and rearrange-ment, the equations of motion become: d2r ( 2eU 2eVcosttt\ n . „. d2z ( 4eU 4eVcosQt\ n / n n n . + ( 2 - 2 7 > Substitutions with the dimensionless Mathieu parameter £ are now necessary. , , d2 tt2 d2 Therefore: / 8e(7 -8eVcos 2c; . d~e + \m(r20 - 2z20)Q2 ~ m{r2 + 2z*)Q2) T ~ ( ' Chapter 2. Mass Spectrometry 21 d2z I -16eU J ^ e K c c 6 2 j _ \ de + \m{r20 - 2z2)tt2 ~ m(r20 + 2z20)W ) Z ~ ( ' The equations describing the motions of ions in a quadrupole ion trap can now be arranged into the canonical form of the Mathieu equation: d2r — + (ar - 2gr cos 2<f) r = 0 (2.31) dfz d2z — + (az - 2qz cos 20 z = 0 (2.32) where _ 8eU _ -4eV a r ~ m(r20 + 2z20)W Qr ~ m(r20 + 2z2)Q2 { 3 ) _ -Well _ 8eV a z ~ m(rl + 2z20)n2 q z ~ m(r20 + 2z20)Q2 { ' 2.1.2 R e g i o n s o f I o n T r a j e c t o r y S t a b i l i t y The solution to the Mathieu equation can now be used to describe the motion of ions using the values for the Mathieu parameters au and qu. The solutions Chapter 2. Mass Spectrometry 22 to the Mathieu equation are shown graphically in Figure 2.4. Figures 2.4 and 2.5 were adapted from a March paper.2 3 Ions with qz and az values falling within the shaded regions have trajectories that are stable in the z-dimension. The same is true for the r-dimension. Ions falling outside of these regions of stability will have trajectories that will increase until the ion hits an electrode or is ejected from the trap. An ion that is stable in the z-dimension but unstable in the r-dimension will likely hit the ring electrode. An ion which is stable in the r-dimension but becomes unstable in the z-dimension will be ejected from the trap through holes in the endcap electrodes. Ions which are stable in both dimensions will remain trapped. To determine which ions are trapped it is useful to overlap the regions shown in Figure 2.4. The area labelled "Region I" in Figure 2.5 highlights a region where ions are stable in both dimensions. Since most ion traps have no DC potential applied to the electrodes (i.e. U = 0) they operate on the o-axis of this stability region. Figure 2.5 also shows that the right boundary of this stability region is defined by qz = 0.908. Chapter 2. Mass Spectrometry 23 Figure 2.4: Regions of stability defined by a and q. "z-stability" on left and "r-stability" on right. Adapted from a March paper.2 3 2.1.3 Secular Frequencies and Resonant Excitation Trapped ions complete their stable trajectories with angular frequencies, called secular frequencies (wu), defined by the following relationship: (2.35) where: 0u 1 (q < 0.4 and (7 = 0) (2.36) Equation 2.34 shows that qu depends on the mass (m) and charge (e) of the ion, the amplitude (V) and frequency (Q) of the applied voltage, and the Chapter 2. Mass Spectrometry 24 region I Figure 2.5: Stability in both dimensions defined by the overlap of r-stability and z-stability. The right boundary of Region I is defined by j3z = 1 (or qz = 0.908 when U = 0). Adapted from a March paper. dimensions of the ion trap (rQ and z0). For a given set of trapping parameters ions will therefore have unique frequencies depending on their m/z ratio. Ions with a specific m/z can be manipulated by applying a dipolar excitation across the endcap electrodes at a frequency corresponding to their unique secular frequency. This is called resonant excitation. Resonant excitation is the basis for many ion manipulation techniques that make ion trap mass spectrometry so powerful. Resonant excitation can be used to improve the mass resolution24 and extend the mass range25 of an Chapter 2. Mass Spectrometry 25 ion trap mass spectrometer. 2.1.4 Stretched Ion Traps In the solution to the Mathieu equation outlined in Section 2.1.1 the rela-tionship between the electrode dimensions r0 and z0 were not defined. For a pure quadrupole field one must use perfectly shaped hyperbolic electrodes and the following relationship must be maintained: Real electrodes must be truncated adding higher order fields (hexapolar and octupolar) to the mainly quadrupolar field experienced by the ions. Contri-butions to higher order fields from truncating electrodes are very small and usually do not affect the operation of the ion trap. It was discovered that increasing the distance between the endcaps would add a significant amount of higher order fields(1 — 2% of V) and that these higher order fields could be used to improve ion trap performance.26 An ion trap with a z0 greater than that dictated by Equation 2.37 is said to be "stretched". The ion trap used in the instrument was manufactured by Teledyne r 2 = (2.37) Chapter 2. Mass Spectrometry 26 Figure 2.6: The endcaps of the Teledyne 3DQ Discovery are stretched to improve performance. The red line defines "ideal" electrode ge-ometry. The internal diameter of the ring electrode is 2 cm. (Mountain View, USA). Figure 2.6 shows the ideal electrode shape and spac-ing compared with the actual spacing used in the Teledyne 3DQ Discovery. Chapter 2. Mass Spectrometry 27 The electric potential of a stretched trap is still mostly quadrupolar, es-pecially near the centre of the trap. Ion trajectories and secular frequen-cies in a stretched trap can be described adequately by considering only the quadrupolar component of the electric field. The higher order fields become more significant as an ion moves farther away from the centre of the trap. When the trajectory of an ion increases it will experience greater effects of hexapolar and octupolar fields. The trajectory of any ion can increase either by coming into resonance with the applied RF-potential (near qz = 0.908) or as the result of resonance excitation. These higher order field contributions are exploited by instrument manufactures to improve the resolution of ion trap devices. The role of higher order fields and other aspects of quadrupole theory will be used to explain the operation of a quadrupole ion trap. 2.2 Ion Trap Operation Quadrupole ion traps are most commonly used as mass spectrometers. To produce a mass spectrum the majority of commercial traps use some version of a mass selective instability mode. The selectivity of an ion trap can be improved by using a powerful technique called collision-induced dissociation Chapter 2. Mass Spectrometry 28 (CID) to preform tandem mass spectrometry. A relatively high pressure buffer gas is used in ion trap mass spectrometers; this is avoided in other mass spectrometers. 2.2.1 Mass Selective Instability The majority of ion trap mass spectrometers are operated in what is called a "mass selective instability mode".2 7 Trapped ions are sequentially ejected from the ion trap based on their m/z. The ion current measured with an electron multiplier external to the trap is plotted against m/z to produce a mass spectrum. This mode of operation can be understood by considering the stability diagram (Figure 2.5) and that the qz value of an ion is directly proportional to the applied RF-potential V (Equation 2.34). For an ion trap operated with only an RF-potential, the ion qz value will determine its stability. Ions with qz values below 0.908 will be trapped. To eject an ion from the trap the RF-potential applied to the ring is increased until that ion reaches the stability boundary (/3Z = 1 line, Figure 2.5). At [3Z — 1 the secular frequency of an ion in the z-dimension will come into resonance with the applied potential and its trajectory will quickly increase until it is ejected from the trap through Chapter 2. Mass Spectrometry 29 holes in the endcap electrodes. Lower mass ions have higher q values for a given V so low mass ions will be ejected before heavier ions. Linear ramping of V will produce a signal linear in m/z at the detector, which is a mass spectrum. The mass resolution will depend on how quickly ions come into resonance and the difference in secular frequencies between mass units (n Th and (n + 1) Th). The difference in secular frequencies between mass units depends on the frequency of the RF-potential, trap size, and the qz at which the ions are ejected. Ion traps are usually designed to give unit mass resolution from 0 — 650 Th and the frequency of the RF-potential is normally close to 1 MHz. The RF-potential of ion trap used in this work has a frequency of 909090.90 Hz. As a comparison, a ring potential at 60 Hz would be ap-propriate for the trapping of micron sized particles. Resonance excitation (Section 2.1.3) can be used with a mass selective instability scan to improve the mass resolution or to extend the mass range of the ion trap. The Teledyne 3DQ ion trap has a potential applied across the endcaps to excite ions just before they reach the stability boundary. This pre-ejection excitation increases the trajectory of the ions and keeps them farther away from the centre of the trap where space charge from heavier ions may Chapter 2. Mass Spectrometry 30 affect their trajectories and secular frequencies. The electric field experienced by the ions farther away from the centre of the trap is also stronger and this helps to eject all ions of a particular m/z faster. The Teledyne 3DQ can easily achieve unit mass resolution from 0 — 650 Th. M a s s R a n g e E x t e n s i o n The mass range of the ion trap used can be extended by ejecting ions at lower qz values. Ejection at a lower qz is accomplished by applying a differ-ent frequency across the endcaps during the mass selective instability scan. Instead of ions being ejected at qz = 0.908, ions can be ejected at a lower RF-voltage. Therefore, for a given RF-power supply the mass range of an ion trap can be increased. Mass range extension is not without problems. First, ejecting heavier ions requires a greater amplitude for the endcap ex-citation. If operational amplifiers are used to supply the endcap excitation then the mass limit will depend on their maximum output. Even if there is no issue with the power supply, the resolution will degrade as the ejection qz is lowered because the secular frequencies of adjacent mass units will be closer together. Currently, ion trap manufacturers extend the mass range of their ion traps Chapter 2. Mass Spectrometry 31 by lowering the frequency of the RF-potential applied to the ring electrode. In this way the resolution is not necessarily degraded as the mass range is extended. The frequency of the RF-potential is fixed for the ion trap used in this work. Mass Calibration Mass calibration is accomplished by using known standards. The initial version of the instrument used an electron gun to ionize a calibration gas (FC43) that was leaked into the vacuum chamber. When the instrument was converted to include an ESI source, appropriate calibration solutions had to be sprayed in. These calibration methods are very straightforward. The most important aspect of mass calibration is appreciating just how many operational parameters can affect it. While conducting ion trap experiments it is often difficult to change only one parameter without changing another. Scan Speed (Th/s), resonant excitation (determined by: frequency, mag-nitude, rate of change), trap pressure, electrode temperature, and the total number of ions trapped all affect the mass calibration. Some of these pa-rameters are merely hardware and software effects. However, the effects of background pressure and ion density are fundamental to ion trap mass spec-trometry. Chapter 2. Mass Spectrometry 32 2.2.2 Buffer Gas and Collisional "Cooling" Quadrupole ion traps are operated at a relatively high pressure for a mass spectrometer. Sector, time of flight (TOF), Fourier transform ion cyclotron resonance (FT-ICR), and even linear analyzing quadrupole mass spectrom-eters are designed to minimize ion-neutral collisions. Ion traps are typically operated with a helium buffer gas pressure in the 0.01 — 1 mtorr range. This is one to three orders of magnitude higher than pressures used in other mass spectrometers. Ion-neutral collisions are necessary for ions to be efficiently trapped and for optimizing ion trap performance.27 The exact ion-neutral collision frequency will depend on the buffer gas pressure and the collisional cross-sections of the ion and the buffer gas. For a molecule in the hundreds of Dalton range the mean time between collisions could be as short as lps. An ion will lose a small amount of kinetic energy during each of these collisions. Ions may be created in the trapping volume or they may be injected into the trapping volume from an external source. Ions further away from the trap centre are more likely to be lost from the trap even if they should be stable. The initial position and kinetic energy Chapter 2. Mass Spectrometry 33 of an ion is a critical factor in determining if it will be trapped. Collisions with the buffer gas help trap ions and keep them closer to the centre of the ion trap. This decrease in trajectory amplitude and kinetic energy is often referred to as "collisional cooling". It is important to note that these ions are not rotationally or vibrationally "cool". Trapped ions are considered to be at the same temperature as the buffer gas.28 A buffer gas improves ion trap performance by concentrating ions near the centre of the trap. Ions a particular m/z, near the centre of the trap, will experience the same fields and therefore have the same secular frequency. As outlined in Section 2.2.1, ions are separated based on their secular frequen-cies. 2.2.3 Mass Selective Ion Isolation Filtered Noise Fields Section 2.1.3 describes how an ion can be ejected from the ion trap with a dipolar excitation applied across the endcaps at the secular frequency of the ion. It is also possible to isolate an ion of a particular m/z while ejecting all others by applying a waveform which Teledyne called a filtered noise field (FNF). A FNF is made up of many frequencies except the secular frequency Chapter 2. Mass Spectrometry 34 corresponding to the ion which is to be isolated. The ability to isolate ions of a single m/z and trap them for an almost indefinite period of time is the principal advantage of using an ion trap for optical experiments. Pure solutions were used for most of the experiments described in this thesis so mass selective ion isolation was usually unnecessary. Only the ion of inter-est was usually trapped. This technique could be used if the molecule to be interrogated with optical experiments was contained in a more complex mixture. Trapping Times Theoretically an ion can be trapped indefinitely. Unfortunately, time limita-tions are imposed by the commercial ion trap software. The Teledyne 3DQ was manufactured in the early 1990's and the data acquisition hardware avail-able at that time had some limitations. Synchronous voltage scanning and data collection was limited by the buffer capacities of the digital to analog and analog to digital converters (DAC's and ADC's). While the Teledyne software is very versatile, the maximum trapping time is only 15 seconds. This 15 second time limit was a severe handicap during earlier stages of the instrument development. Chapter 2. Mass Spectrometry 35 Mass Range for Ion Confinement The mass range of the ion trap discussed in Section 2.2.1 refers to the mass range of the ion trap as a mass spectrometer. The largest m/z ion that can be ejected from the trap and detected represents the high mass limit of the spectrometer. As an ion storage device the mass limit of the trap is much greater. The stability region shown in Figure 2.5 asymptotically approaches the origin where qz = 0. An ion trap may therefore appear to have no mass limit and that infinitely heavy ions could be trapped. The trapping capacity of an ion trap can be thought of as a potential well. A deep potential well would result in strong focusing of the ions. Ions with very low q experience a shallow potential well and are not trapped effectively. Another problem with very massive and/or multiplied charged ions is their kinetic energy. Heavier ions require a greater number of collisions with buffer gas neutrals before they are slowed enough to be trapped. The trapping efficiency of large ions therefore approaches 0% as qz approaches zero. Chapter 2. Mass Spectrometry 36 2.2.4 Collision-Induced Dissociation and Tandem Mass Spectrometry Collision-induced dissociation can be used to determine the structures of ions and to increase the selectivity of ion identification. CID of an ion in a quadrupole ion trap is another powerful technique based on resonance exci-tation. Resonance excitation is used to increase the trajectory of an isolated ion. As the ion moves away from the centre of the trap it will experience an increasing electric field and therefore greater acceleration. Collisions with the buffer gas will be more energetic and the internal energy of the ion will slowly increase until dissociation occurs. The fragmentation pattern of an ion can be used to elucidate structure. Selectivity of ion identification can be enhanced by using CID for tandem mass spectrometry. Tandem mass spectrometry is the process of completing two or more mass selective operations in series before the ions are ejected from the trap and detected. Two successive mass selective operations are called " M S / M S " and n mass selective operations are called " M 5 n " . Tandem mass spectrom-etry can be used to differentiate ions which have the same m/z but different fragmentation patterns. The molecular ion (precursor) is first isolated and Chapter 2. Mass Spectrometry 37 then CID is used to create fragments known as product ions which are then mass analyzed to produce an MS/MS spectrum. If MS/MS is not selec-tive enough for a particular application, a product ion can be isolated and fragmented (MS3) to further increase the selectivity of the technique. When combined with chromatography, MS/MS usually provides sufficient selec-tivity for most applications. Police and customs agencies around the world routinely use tandem mass spectrometry to screen for drugs and explosives. 2.3 Linear Quadrupoles Linear quadrupoles can be used as ion guides, ion storage devices, and mass analyzers. In the instrument developed, two linear quadrupoles are the key elements of the ion optic system that moves ions from the atmospheric elec-trospray ionization (ESI) source to the ion trap. 2.3.1 Linear Quadrupole Theory-Most linear quadrupoles consist of four symmetrically positioned rods as shown in Figure 2.7. Round rods can be used to approximate the ideal Chapter 2. Mass Spectrometry 38 Figure 2.7: Linear quadrupole rod set. 2ro is the distance between poles and r is the rod radius. hyperbolic shape using the following relationship: r = 1.1487r0 (2.38) where r is the rod radius and 2rG is the distance between the electrodes. Figure 2.8 demonstrates how rods having a circular cross-section can ap-proximate the ideal hyperbolic shape. Analyzing quadrupoles must be ma-chined to very high tolerances if reasonable resolution is to be obtained. Linear quadrupoles used as storage devices or for ion transmission can be constructed with far less precision. Chapter 2. Mass Spectrometry 39 Linear Quadrupole Figure 2.8: End view of linear quadrupole rod set and "ideal" hyperbolic shape. Round rods are used to approximate the hyperbolic shape of ideal quadrupole electrodes. An electric potential (c/>0) is applied to one set of rods while the same potential 180° out of phase (—<f>0) is applied to the other rod set (Figure 2.8). The Mathieu equation and solution are also used to describe the linear quadrupole. The Mathieu parameters for the linear quadrupole are: Chapter 2. Mass Spectrometry 40 The x-dimension can be defined by a line passing through the centres of the rod ends labeled A in Figure 2.7 and the y-dimension would then be defined by a line passing through the centres of the rod ends labeled B in Figure 2.7. The z-dimension is therefore parallel to the rods. Since there is no quadrupole field in the z-dimension the linear quadrupole does not focus ions in the z-dimension. Linear quadrupoles operated at high vacuum pressures will quickly focus a beam of ions towards the centre of the device. This property, first discovered by Douglas2 9, is extremely useful for ion transmission. Collisional focusing in the linear quadrupole concentrates the diffuse ion beam from the ion source so that it may be efficiently focused through small apertures. 2.3.2 Linear Quadrupole Operation Only an RF-potential was applied between the poles of the quadrupoles used in our instrument (i.e. [7 = 0). RF-only linear quadrupoles should transmit a wide mass range of ions. Lenses positioned at both ends of the linear quadrupole are used to trap ions. To trap positive ions, the lenses are biased with a higher DC voltage than the quadrupole rods. Al l four rods of the linear quadrupole are biased with the same potential so as not to affect the Chapter 2. Mass Spectrometry 41 quadrupole field (i.e. (7 = 0). Ions are pulsed out of the linear quadrupole by lowering the voltage on one lens (exit lens). To use a linear quadrupole in ion transmission mode the entrance lens, quadrupole rods, and exit lens are biased with decreasing voltages so positive ions will be accelerated in the desired direction. Although RF-only linear quadrupoles can be operated at high vacuum pressures they must be kept clean. This may seem to be an obvious point but poor practices in this area are common. Sometimes plastics are used as insulators in high pressure regions of a mass spectrometer because the outgassing of these materials has little effect on the measured vacuum pres-sure. However, contaminants from these plastics can lead to the formation of unwanted adducts when ions are stored for long periods of time (> 1 s). Con-taminated electrodes and mounting hardware can also hold a static charge and can have a negative impact on ion storage and transmission. 2.4 The Ion Source Many different ion sources can be coupled to an ion trap mass spectrometer. In early versions of the instrument an electron gun was used to ionize neutral Chapter 2. Mass Spectrometry 42 molecules, which had been thermally desorbed from a heated tungsten fila-ment. The final version of the instrument uses electrospray ionization (ESI) to generate ions external to the ion trap. Although the thermal desorption probe ultimately had to be replaced, much was learned about the perfor-mance of the ion trap and the photophysics of trapped gas phase molecular ions. Analyte solutions used in this work were prepared using HPLC grade sol-vents. For the PAH experiments saturated solutions were used to load analyte molecules onto the thermal desorption probe. For the ESI experiments dilute solutions were prepared in three dilution steps. While volumetric flasks were always used to prepare the first two stock solutions the final solutions used for ESI were prepared less precisely. Of course, the integrity of the molecules was confirmed by mass spectrometric analysis for each optical experiment. Also, the exact composition of the stock or ESI solutions are not relevant to the optical experiments preformed on trapped ions. Only the ions confined in the ion trap and neutral molecules present in the vacuum are important. Chapter 2. Mass Spectrometry 43 2.4.1 Thermal Desorption and Electron Impact Ionization The thermal desorption probe designed and built for the initial experiments consisted of a long stainless steel tube used to support two wire electrodes (see Figure 2.9). A tungsten filament was spot welded between the wires at the vacuum end of the probe. To load the probe with molecules the filament was soaked in an analyte solution and allowed to dry. The probe was then inserted into the vacuum system through a ball valve and the tungsten filament was positioned between the exit endcap and the ring electrode. A current was passed through the wire electrodes to heat the tungsten filament, thus desorbing the analyte molecules into the internal volume of the ion trap. Before the vacuum system pumped away all the neutrals, an electron gun was used to pulse a beam of electrons into the ion trap ionizing some fraction of the molecules. This electron impact ionization can be represented by the following equation: M + e" -> M+ + 2e~ (2.40) This ion source was very easy to use and efficiently filled the ion trap Chapter 2. Mass Spectrometry 44 Thermal D e s o r p t i o n P r o b e r f i l amen t r p l a s t i c g r i p ceramic s p a c e r s t a i n l e s s s t e e l t u b e r r f i lanent 1 I va . . . — s p o t Weld vacuum seal (g lue) f i l a n e n t 20 ga # i r e s dimensions in cm _L 1.2500 0'300. -23.0000 I p r o b e t i p e x p a n d e d <4X) Figure 2.9: Thermal desorption probe. with ions. Unfortunately, thermal desorption also created a significant back-ground of neutral molecules in the vacuum system. It was suspected that any emission from trapped ions was buried under the emission from neutrals (gas phase or adsorbed to surfaces). Changing to an external ESI source eliminated the problem of a neutral background and allowed a new class of molecules to be investigated. Chapter 2. Mass Spectrometry 45 2.4.2 Electrospray Ionization Electrospray is a method used to transfer ions present in solution to the gas phase. Many articles on ESI fundamentals have been written, 3 0 - 3 2 and a book edited by Cole 3 3 is an excellent reference. The general design of the ESI front-end was provided by Dr. Don Douglas (UBC Chemistry) as a part of a collaboration. Figure 2.10 is a schematic of the ESI front-end used here. A glass capillary carries a solution of ions from the syringe pump to the stainless steel sprayer where a high potential induces electrohydrodynamic (EHD) spraying. 3 4 - 3 6 Air is used to pneumatically assist the spraying process when solutions hav-ing a high surface tension (i.e. mostly water) are used. Pneumatically as-sisted electrospray is called ionspray. Typically, a potential of 3000V-5000V between the the sprayer and the curtain plate will produce a spray containing highly charged droplets. Some of these droplets evaporate and pass through the sampling orifice. Evaporation, enhanced by using a counterflow of dry nitrogen gas, will eventually lead to the formation of gas phase ions. Ions and neutrals now pass through the orifice and into the vacuum system where the expansion is sampled by the skimmer. Bare ions (no solvent) are then confined in the first RF-only linear quadrupole (QQ). Chapter 2. Mass Spectrometry 46 Sampling Orifice v Skimmer Syringe Pump Figure 2.10: ESI Front-end. Chapter 2. Mass Spectrometry 47 Electrospray ionization is a "soft" ionization technique that allows ther-mally labile ions to be transferred to the gas phase without falling apart. Besides eliminating neutral analytes from the ion trap the ESI source can be used to generate ions from wide variety of molecules including dyes, peptides and proteins. 2.5 MS Instrumentation A commercial ion trap (Teledyne 3DQ Discovery) was significantly modified to provide the ion trap mass spectrometer component of the instrument. Al l of the initial modifications involved gaining optical access to the trapping volume of the ion trap. The final version of the instrument has an external electrospray ion source. Although a great deal of time was needed to couple this external ion source to the Teledyne system, successful completion of the optical experiment may not have been possible without this new ion source. Figure 2.11 is a diagram of the mass spectrometer used to confine ions for optical spectroscopy. Ions generated using ESI are transmitted to the ion trap using two RF-only linear quadrupoles (Qo and Q\). Transmission of ions from the ESI source to the ion trap requires a large pressure drop and Chapter 2. Mass Spectrometry 48 C U R T A I N P L A T E S A M P L I N G O R F I C E S K I M M E R S Y R I N G E P U M P Figure 2.11: Mass spectrometer, ion optics to accelerate the ions towards the trap. 2.5 .1 V a c u u m S y s t e m For optical experiments the ion trap is usually held close to 1 mtorr. The pressure drop from atmosphere is achieved in three stages. The area between the sampling orifice and skimmer is pumped using a Leybold D8A roughing pump providing 190 L/min of pumping speed. The pressure achieved in this region is approximately 950 mtorr. The next stage of pumping uses a 50L/s Chapter 2. Mass Spectrometry 49 turbomolecular pump (Leybold Turbovac 50) to keep Qo at a pressure of approximately 7 mtorr. Finally, the vacuum chamber containing Q\ and the ion trap was pumped with another 50 L/s Leybold Turbovac 50. Both turbo pumps were backed using an Edwards E2M-18 2-stage roughing pump (343 L/min). The pressure in the ion trap is increased for optical experiments by leaking in a helium buffer gas. A Granville Phillips (Boulder, USA) variable leak valve was used to control the pressure of helium. Between optical experiments the helium pressure is decreased to prolong the life of the turbo pumps. Some helium (1 to 5 x 10~5torr) is still leaked into the system when the instrument is not in use to prevent the backtreaming of pump oil which would otherwise contaminate the vacuum system. Surfaces contaminated with pump oil, or other organic molecules, can hold a static charge. A severely contaminated vacuum system can degrade the performance of a mass spectrometer. The transmission efficiency of the electrospray front-end was especially vulnerable to contamination. In fact, the ion trap could not be filled to capacity in 15 s (time limit of ion trap software) until all of the plastic was removed and a thorough cleaning of the front-end was completed. This cleaning of the front-end also included an Chapter 2. Mass Spectrometry 50 ESI k. Ion Guide Operating Voltages Component Voltage Power Supply Sprayer 3500 Kepco #188-0021 Curtain Plate 195 V G Micromass PQ Lens Supply Sampling Orifice 80 V G Micromass PQ Lens Supply Skimmer 46 V G Micromass PQ Lens Supply Q0 Offset 35 Xantrex LXQ 60-1 Gate 39 Chem. E-Shop Lens Supply #00728 Q i Offset 30 Xantrex LXQ 60-1 LI 30 Chem. E-Shop Lens Supply #00728 L2 30 Chem. E-Shop Lens Supply #00728 Table 2.1: Suitable operational voltages for singly charged ions. improvement of the ion optics. 2.5.2 I o n Optics The ion optics used to transmit ions from the ESI source to the ion trap are very straightforward. To accelerate a cation, an electric field must be established in the desired direction. Table 2.1 shows the voltages used to trap a 450 Th singularly charged cation. The voltage drop between the first few ion optics is large because of the high pressures in these regions. Once past the skimmer, voltage drops are much lower. For most experiments, ions were accumulated in Q0 by keeping the Gate voltage 3 — 5 V higher than the Q0 offset. The offsets applied to the linear quadrupoles were applied to all four rods. This should not be confused with Chapter 2. Mass Spectrometry 51 the DC potential U in Equation 2.39. Ions stored in Q0 are pulsed into the ion trap by briefly lowering the Gate potential. The linear quadrupoles are the key elements of the ion optics. The col-lisional focusing in the high pressure Qo narrows the beam of ions collected from the skimmer. Collisional focusing allows more ions to pass through the small hole in the entrance endcap. The RF-potential applied to the linear quadrupoles was generating using a single SCIEX (Toronto, Canada) power supply. The RF-supply operated at 1 MHz and was capacitively coupled to Qi which was, in turn, capacitively coupled to Qo- The impedance of the capacitors creates a voltage divider. The amplitude of the RF-potential on Qo is approximately 75% of the amplitude on Q\. Externally generated ions must lose kinetic energy after they are injected into the ion trap through collisions with the buffer gas.3 7 Ions injected with too much kinetic energy will not be trapped. Therefore, the voltage drop between the Gate and the grounded entrance endcap must be kept low. The RF-voltage applied to the ring electrode during ion injection is also critical. To optimize injection of a particular ion, the RF-voltage applied to the ion trap ring electrode should correspond to a qz of between 0.1 and 0.2 for that ion. 3 7 The phase of the potential is obviously important but a pulse of Chapter 2. Mass Spectrometry 52 ions from Q0 will persist for many RF-cycles so no synchronization is possible or necessary. The ions should also be shielded from the entrance endcap as much as possible before injection so that voltages applied to the ion trap do not disrupt the ion beam. Early versions of the ESI front-end included an aperture between LI and L2 that was connected to the ring electrode. The ion trap could not be filled to capacity until this aperture was insulated from the endcap and connected to L2 as shown in Figure 2.11. This modification of L2 and the cleaning described in Section 2.5.1 increased the ion transmission by at least two orders of magnitude. This increase in ion transmission allowed the ion trap to be filled to capacity. As ions are scanned out of the ion trap during a mass scan they are accelerated towards the detector with an exit lens. The exit lens is biased to -300 V. The exit lens also shields ions from the exit endcap as they leave the trap. The ions pass through the mesh electrode before they hit the detector. The mesh electrode is biased to +300 V when a mass scan is not in progress to prevent unwanted ion from hitting the detector. Chapter 2. Mass Spectrometry 53 2.5.3 Ion Trap Control The commercial control software of the Teledyne 3DQ Discovery is versatile. The Teledyne 3DQ uses a mass selective instability mode enhanced with early resonant ejection (Section 2.2.1). The frequency and amplitude of resonant excitation, scan speed, injection voltage, trapping voltage, trap temperature, and multiplier voltage can all be controlled with the software. As previously discussed, the main limitation is the 15 s time limit for a single mass spectro-metric sequence. The data processing package supplied by Teledyne also has limitations so a Lab VIEW program (National Instruments, Austin, USA) was developed to process mass spectra. The Teledyne 3DQ was delivered with a 0 — 650 Th mass range. The mass range was extended to approximately 2000 Th by changing the gain of the two amplifiers which supply the resonant excitation to the endcaps. When operated with an extended mass range, the mass scan speed was reduced from the default 12000 Th/s to 1000 or 2000 Th/s. The buffer gas pressure must also be lowered to extend the mass range. Unfortunately, lowering the buffer gas pressure also lowers the trapping efficiency and fewer ions can be trapped. Chapter 3 54 Spectroscopy and Optics It was hypothesized that combining the sensitivity of fluorescence spectroscopy with the selectivity of ion trap mass spectrometry would produce a mass spec-trometer with unique detection capabilities. The fluorescence from a single barium ion had been detected in an ion trap 8 and fluorescence emission from molecular cations insolated in a frozen noble gas matrix had been character-ized. If one atomic ion could be detected then surely a million fluorescent molecular ions could also be detected. The ion trap's rarified environment and long trapping times seemed to be ideal for ultra-sensitive fluorescence spectroscopy. Fluorescence spectroscopy is a very sensitive technique when used with condensed phase experiments. Generally, the limit of detection for fluores-cence experiments is « 1000 times lower than absorption spectroscopy.38 It is intrinsically easier to measure a small fluorescence signal against a zero Chapter 3. Spectroscopy and Optics 55 background than to detect small differences in two large signals in an ab-sorption experiment. The intensity of the fluorescence signal (If) is given by the following equation: If = 2.303(j) fI0ecd (3.1) where 0/ is the fluorescence quantum efficiency, IQ is the excitation intensity, e is the wavelength dependent molar absorptivity, c is concentration and d is path length. Another advantage of fluorescence spectroscopy is that the signal If can be increased by using the high intensity output of a laser for excitation (I0). Furthermore, since trapped ions can be contained almost indefinitely, the signal from many absorption and emission events can be integrated to increase the total amount of fluorescence detected . While If is obviously important, the limit of detection is really dependent on the signal to noise ratio (S/N) and the magnitude of the background signal. Fluorescence is normally significantly red shifted from the excitation so the excitation does not directly contribute to the background. In condensed phase experiments, the background signal is usually dominated by Rayleigh and Raman scattering from the solvent, and fluorescence from the solvent and impurities.1 7 Since the ion trap eliminates any solvent or impurities from the probe volume, it was theorized that detection limits may be even better than Chapter 3. Spectroscopy and Optics 56 those obtained in solution. Modifying an ion trap to allow optical access for both excitation and emission, while preserving the functionality of the mass spectrometer, was expected to be the main instrumental challenge of this project. Unfortu-nately, several aspects of gas phase spectroscopy were not anticipated and the experiment proved impossible without major instrumental modifications. This chapter describes the development of the optical system used to de-tect fluorescence emission from large gas phase molecular ions trapped in a quadrupole ion trap mass spectrometer. While initial experiments failed to detect fluorescence, they led to an improved understanding of the funda-mental processes involved. This chapter describes how the knowledge gained from these early experiments was used to develop a successful instrument. 3.1 Optical Access to the Trapping Volume In order to gain optical access to the trapping volume, modifications to the vacuum chamber and ion trap electrodes were necessary. A new vacuum chamber was machined with three optical ports to allow the laser radiation to enter and exit the chamber. It also allowed any fluorescence emission to Chapter 3. Spectroscopy and Optics 57 exit the chamber. Two slots in the entrance endcap and one slot in the exit endcap allowed the laser beam to enter and exit the trapping volume, and allowed emission from inside the trap to be collected. Brian Snapkauskas (UBC Chemistry Mechanical Engineering Shop) ma-chined a new vacuum chamber from a solid block of aluminium. The chamber was then black anodized (Altech Anodizing Ltd., Delta, Canada) to reduce light scatter. Custom flanges were built to hold glass windows. These flanges were then welded at angles that corresponded to the asymptotes of the hy-perbolic electrodes. To avoid problems with reflections, the laser entrance window was replaced with a quartz Brewster window on the end of a long black anodized tube. Laser light from the Spectra Physics (Mountain View, USA) Stablite 2017 argon ion laser was spatially filtered before being passed into the ion trap. A long focal length (500 mm fl) plano-convex lens was used to focus the laser. Two cone shaped 3 mm apertures, located behind the Brewster window, were used for the spatial filtering. The beam waist is « 1 mm in diameter at the centre of the ion trap. It would be easier to pass a more tightly focused laser beam through the ion trap without hitting an electrode, but fewer ions would be excited because the ion cloud radius is at least 2 mm in diameter. Chapter 3. Spectroscopy and Optics 58 A C W laser was chosen over a pulsed laser because the CW laser can deliver more photons to the probe volume without causing unwanted multiphoton absorption. Figure 2.2 shows the Teledyne electrodes without optical access slots. In order to create a line of sight through the ion trap, 5 mm wide slots were cut from the outside edge of the endcap to the edge of the channel surrounding the hyperbolic surface. The modifications made to the electrodes and vacuum chamber did not degrade the performance of the instrument as a mass spectrometer. 3.2 Optical Detection The original optical collection system used several singlets for confocal imag-ing and to correct for chromatic abberations. The first singlet (i.e. lens) was positioned inside the vacuum system close to the slot cut into the entrance endcap. The aperture used for confocal imaging was positioned at the back focal point the first lens (usually inside the vacuum system). Outside the vacuum system, two plano-convex lenses separated by the sum of their fo-cal lengths were used to correct for chromatic abberations and to focus the Chapter 3. Spectroscopy and Optics 59 collected light onto the entrance slit of the spectrograph. The optical con-figuration was aligned by irradiating a small amount of dye solution, which was held in the centre of the ion trap with the vacuum system open and the trap electrodes removed. A software program was written to calculate the exact front and back focal lengths of the lenses as a function of wavelength. The 0.150 m spectrograph used for this work has a 600 groove/mm grat-ing blazed at 500nm (Acton SpectraPro-150, Acton, USA). The spectro-graph is a Czerny-Turner type with a flat focal plane, an f-number of 4.0, and a dispersion of 9nm/mm. The optical system was designed with the last lens having an f-number of 4.0 so that the grating would be fully illumi-nated. The charged coupled device (CCD) detector used for all preliminary experiments with the spectrograph was a 1024 x 512 CCD (TE/CCD 1024 E E V UV1, Princeton Instruments, Tucson, USA). The optical setup used with the thermal desorption probe (Section 2.4.1) is shown in Figure 3.1. 3.3 First Experiments with PAH Cations PAH cations were the first analytes investigated. PAH cations were chosen because they are toxic in the environment even at low levels and sensitive Chapter 3. Spectroscopy and Optics 60 Thermal Desorption Probe Figure 3.1: Instrumentation used to study PAH cations. (N.B. only part of the laser is shown) Chapter 3. Spectroscopy and Optics 61 methods to detect them are needed.39 Their mass spectrometric identification can be complicated by the fact that there is often more than one geometric isomer corresponding to the same m/z in a mass spectrum. It was predicted that fluorescence detection of trapped PAH cations would be a more sensitive and selective analytical technique than ion trap mass spectrometry alone. 3.3.1 Methods Two PAH molecules were chosen based on the matrix isolated absorption spectra of their cations40 (see Figure 3.2). As shown in Equation 3.1, flu-orescence intensity (If) is directly proportional to molar absorptivity (e). The absorption of tetrabenzo[acfh]napthalene and benzo[ghi]perylene corre-sponds to the wavelengths available with the argon ion laser (i.e. 488 ram and 514nm). Obviously, in order to detect fluorescence it must be possible to excite the cations. Tetrabenzo[acfh]napthalene was synthesized in the UBC Chemistry department by Tim Daynard who followed a procedure found in the literature.41 Benzo[ghi]perylene was purchased from Aldrich Chemical Company (St. Louis, USA). Benzene solutions of the PAH molecules were applied to the tungsten filament of the thermal desorption probe (Section 2.4.1). The probe was Chapter 3. Spectroscopy and Optics 62 400 500 600 700 40(5 ' 500 600 700 wavelength (nm) wavelength (nm) Figure 3.2: The matrix isolated cation absorption spectra of Benzo[ghi]perylene (left) and Tetrabenzo[acfh]napthalene (right).(from Electronic absorption spectra of radical ions., volume 34 of Physical sciences data. Elsevier, New York, 1988.) then inserted into the vacuum chamber with the filament positioned between the ring and exit endcap. A current was passed through the filament and neutral PAH molecules were thermally desorbed into the trapping volume. Radical cations were then created using a beam of 70 eV electrons from an electron gun (Equation 2.40). The original Teledyne instrument was equipped with an electron gun positioned behind what is now referred to as the entrance endcap. Electrons were accelerated into the ion trap by pulsing an electrostatic lens at the end of the electron gun. The duration of the electron pulse is software controlled Chapter 3. Spectroscopy and Optics 63 which allows one to control the number of ions created. It was very easy to fill the ion trap to maximum capacity by increasing the duration of the ionization pulse. The trapping level during the ionization step was normally kept close to qz = 0.3 for these first optical experiments. The light given off by both filaments, which was a major limitation, was initially overlooked. The tungsten filaments of the electron gun and thermal desorption probe give off a significant amount of light. The filament supply for the Teledyne's electron gun was modified to include a high current me-chanical relay. A similar relay was used with the thermal desorption probe. These relays were gated on and off so that light from the filaments would never reach the CCD. As well, the laser light was gated on and off using a crude but effective shutter constructed in the lab from a business card and a solenoid. Figure 3.3 shows the usual timing for the experiment. Only dark current and read noise were initially considered to be limiting sources of noise in this experiment. It was expected that the Stokes shift in the gas phase fluorescence might be different than the Stokes shift observed in the matrix experiments, however the fluorescence was still expected to be red shifted significantly from the laser line. If this were true, the fluorescence measurement could be made against a zero background, as detector noise Chapter 3. Spectroscopy and Optics 64 Trap RF Start Pulse Electron Gun _l Filament Probe Filament J l (desorption) Electron Gate T (ionization) CCD Shutter Laser Shutter Figure 3.3: Timing diagram for early optical experiments on PAH cations using the thermal desorption probe. The start pulse is TTL, the next three are either on (high) or off (low), and the shutters are open (high) or closed (low). Chapter 3. Spectroscopy and Optics 65 was the only background noise expected. The CCD was cooled to — 50°C to reduce its dark current and full vertical hardware binning was used to reduce the read noise. Since noise in a CCD follows Poisson statistics, the rms noise is the square root of the background. In the low-light regime, the significant noise sources are read noise and dark current noise. The total background noise for a CCD is given by the following equations: dark noise — Vdark current (3-2) total background noise — \J(read noise)2 + (dark noise)2 (3.3) 3.3.2 Photodissociation The first optical experiments were conducted before the CCD detector was available. Absorption was detected by observing photodissociation of the PAH cations. Photodissociation was used to align the laser beam with the centre of the ion cloud as well as to determine the dimensions of the ion cloud. Figures 3.4 and 3.5 show typical photodissociation spectra. Initially, photodissociation was a welcomed indication that the ions were absorbing, but unfortunately, photodissociation also limited the laser power that could be used. In solution experiments, much higher laser powers can Chapter 3. Spectroscopy and Optics 66 • •• - -— i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i 70 90 110 130 150 170 190 210 230 250 270 290 310 330 LJji ^g.i,,...th,. X JL~J~J - I — i — i — i — i — i — i — n — i — i 1 — i — i — i — i — i — i — i — i — i — i 1 1 — i — i — i 1 — i — i — r 70 90 110 130 150 170 190 210 230 250 270 290 310 330 m/z Figure 3.4: Tetrabenzo[acfh]napthalene. The upper mass spectrum without photodissociation and the lower mass spectrum with photodisso-ciation (488 nm, 0.25 W) Chapter 3. Spectroscopy and Optics 67 i i i 1 i i i i i I I I i i i i i i n — i — i — i — i — r 50 70 90 110 130 150 170 190 210 230 250 270 290 310 330 ~ i—i—i"" I i i — i — i — i — i — i — T — i — i — i — i — i — i — i — i — i — r n — i — i — r 50 70 90 110 130 150 170 190 210 230 250 270 290 310 330 m/z Figure 3.5: Benzo[ghi]perylene. The upper mass spectrum without photodis-sociation and the lower mass spectrum with photodissociation (514 nm, 0.6 W) Chapter 3. Spectroscopy and Optics 68 be used before photodissociation becomes a problem. Laser powers as high as 70,000W/cm2 are used in droplet experiments42 but only 1 to 5W/cm2 could be used in the ion trap without quickly fragmenting all of the ions. A typical ion cloud has a diameter of at least 2 mm so the argon ion laser would only be able to deliver 25 to 35 W/cm2 to the entire ion cloud. In solution experiments, photodissociation is thought to be a single pho-ton event where one photon can excite the molecule to a dissociative excited state.17 For many fluorescent molecules it is routinely estimated that a fluo-rescent molecule can withstand, on average, close to one million absorption and emission cycles before dissociating. This mechanism is not consistent with the increased susceptibility of trapped ions to photodissociation. 3.3.3 Background Scatter When optical detection was first attempted, the laser power was kept just below the level that would cause photodissociation. The only emission de-tected away from the laser line turned out to be plasma lines from the laser tube. The plasma emission was scattered off of the ion trap electrodes and collected as unwanted background. This problem was easily solved by plac-ing a filter in the laser beam which was designed to pass only the laser line. Chapter 3. Spectroscopy and Optics 69 Filters for 488 nm and 514 nm were obtained for this purpose (03-FIL-202 and 03-FIL-204, Melles Griot, Rochester, USA). The scatter from the laser line was overwehelming. No matter how well the laser beam was spatially filtered (or otherwise cleaned up), unwanted Rayleigh and Raman scattering from the last window into the vacuum system (entrance window) caused problems. The integration time for the optical experiment was limited to 1-3 s because the CCD would saturate if the spectrograph's grating was positioned with laser wavelength hitting the CCD. A solution to this problem was to position the grating in the spectrograph so that the laser wavelength was not incident upon the CCD. By doing this the integration time can be extended to 10 s. As a result, the total integration time was limited by the amount of light collected, the well capacity of the CCD and the stray light from the laser excitation wavelength. The Raman scattering and fluorescence generated by the laser beam as it passed through the entrance and exit windows was significant. This back-ground can reflect off the outside and inside ion trap electrode surfaces. Even with confocal imaging of the centre of the ion trap, it is not possible to elim-inate all of this unwanted scatter. Since the fluorescence is expected to be Stokes shifted from the laser line this background scatter was not expected Chapter 3. Spectroscopy and Optics 70 to interfere with the fluorescence experiment. To further reduce unwanted background generated by the laser, the exit window was replaced with a specially designed beam dump. The dimensions of the beam dump were chosen so that the laser beam would be completely contained as it reflected around the interior of the device. The exit window was only necessary for photodissociation experiments used to determine ion cloud size. 3.3.4 Unwanted Neutrals To this point, no fluorescence from the ions was detected. To address the problem of background scatter from the electrodes, efforts were taken to improve the collection optics. The first collection optic outside the trap was doubled in diameter to 1 inch. The purpose of this change was to collect more light exiting the trap and to reduce the depth of field of the image created by this lens. Decreasing the depth of field improves the confocal imaging at the expense of losing light from the fringes of the ion cloud. Several optical configurations were used to image the ion cloud. A minimum number of optics is desirable because approximately 5% of the light will be lost by reflections at each lens surface. More optics are needed if confocal imaging Chapter 3. Spectroscopy and Optics 71 is to be efficient and chromatic abberations are to be minimized. Optimizing an optical system without a detectable signal was very dif-ficult. Especially problematic was adjusting the lens and aperture inside the vacuum system. Before the adjustment could be evaluated, the vacuum system would need to pump down for a few hours. After many months of optical changes and adjustments, a signal other than unwanted laser scatter was finally detected. Unfortunately, this signal was extremely chaotic and was present even without ions in the trap. Figure 3.6 is an example of the background collected in an experiment with a total integration time of 100 s. The varying component of the background was attributed to neutral PAH molecules that remained in the probe volume after they were desorbed. The improved ion optics was partly responsible for making it possible to detect emission from neutrals. After discussing the high neutral background with a Teledyne scientist it was revealed that a recent software change could be responsible. In their last software update Teledyne decided to reduce the heating of the ion trap without informing their customers. For all previous PAH experiments, the ion trap was kept hot enough to evaporate any PAH molecules that adsorbed to the electrodes. With this latest software change the neutral background would remain a problem because it was no longer Chapter 3. Spectroscopy and Optics 72 Figure 3.6: Chaotic background from laser scatter and neutral emission. Tetrabenzo[acfh]napthalene excited at 488 nm. Upper spectrum is magnification of lower spectrum. Chapter 3. Spectroscopy and Optics 73 possible to keep the electrodes hot enough. Reverting to old versions of the software may have been possible, but the instrument was already being converted to incorporate the ESI source. Although not ideal for fluorescence experiments the thermal desorption in-strument could have been used to collect absorption spectra of PAH cations. Photodissociation would be used to indicate absorption. The astrophysics community was, and still is, interested in identifying diffuse interstellar band (DIB) carriers and PAH cations are prime candidates. The absorption spec-tra of the PAH cations would be useful for evaluating potential candidates. A tunable laser source would be needed for such experiments. 3.3.5 Results from PAH Experiments Although fluorescence from PAH cations was not detected, these experiments were still very useful. The susceptibility of gas phase ions to photodissoci-ation was a problem that had not been anticipated. The emission of neu-tral molecules suggested that ion emission may be significantly closer to the laser line than originally predicted. If the fluorescence is not Stokes shifted away from the laser generated background, then the fluorescent measurement would no longer be against a zero background. Without a zero background, Chapter 3. Spectroscopy and Optics 74 the main sensitivity advantage of fluorescence spectroscopy is lost. Another group reported that fluorescence from similar PAH cations could not be detected in a molecular beam experiment.4 Lower temperatures, less photodissociation, and longer integration times should make molecular beam experiments easier than ion trap experiments. Maier reported experiments where two thirds of ions that were fluorescent in matrix experiments did not fluoresce in the gas phase.43 Furthermore, it has been predicted that fluores-cence experiments of large gas phase molecular ions may not be possible in an ion trap. 1 4 There were many aspects of the fluorescence experiment that remained unknown and further investigation was clearly necessary. The problem of emission from neutral molecules was solved by changing to an electrospray ion source. However, there were many other issues which were not addressed by this major instrumental change: photodissociation would still limit the laser intensity, absorption spectra have not been recorded, the fluorescence quantum efficiency ((/>/) for gas phase ions remained unknown, and ion fluorescence (if present) may be totally obscured by background scatter. Chapter 3. Spectroscopy and Optics 75 3.4 Preliminary Experiments Using ESI The decision to add an electrospray source to the ion trap mass spectrometer represented a significant shift in both experimental techniques and goals. It was clear that fluorescence spectroscopy of trapped molecular ions may not be a particularly sensitive technique. However, fluorescence spectroscopy could still be combined with ion trap mass spectrometry to probe the structure of molecular ions. The structure of gas phase biomolecules is of particular interest.44 Using an ESI source allowed these ions to be investigated. It would have been impossible to trap biomolecules using the thermal desorption probe since thermal desorption would destroy most biomolecules. Initially, the ESI source could not efficiently deliver ions to the ion trap. The benefits of the ESI source were difficult to appreciate when the total number of ions that could be trapped was reduced by ay least two orders of magnitude. While successful fluorescence experiments were not possible until the ESI front-end was properly optimized (Section 2.5.2), further pho-todissociation experiments led to a better understanding of the photophysics of trapped ions. Chapter 3. Spectroscopy and Optics 76 Figure 3.7: Structures of 5-CR 6G and R6G. These molecules have very sim-ilar structures. 3.4.1 Ionic Dyes The first molecules investigated were the ionic dyes Rhodamine 6G (R6G) and 5-carboxyrhodamine 6G (5-CR 6G). These dyes have very similar struc-tures (Figure 3.7). The dyes were chosen because they are ions in solution and they have a high fluorescence quantum efficiency in solution ((f)f > 0.9). R6G was purchased from the Eastman Kodak Company (Rochester, USA) and 5-CR 6G was purchased from Molecular Probes Inc.(Eugene, USA). For all experiments using these dyes, 1 pM, (in 50:50 water methanol solutions) were used for spraying (Section 2.4.2). The gas phase absorption spectra and quantum efficiencies were unknown. Chapter 3. Spectroscopy and Optics 77 Gas Phase Absorption of 5 -CR 6G Ions A photodissociation experiment was devised to measure the absorption of gas phase 5-CR 6G ions. Ideally, a tunable laser should be used for this experiment but the argon ion laser was the only one available. Trapped ions were irradiated at five different wavelengths. The following conditions were kept constant for each wavelength: trapping level(g2), buffer gas pressure, irradiation time, and laser power. The percentage decrease in the molecular ion peak was then plotted against wavelength. The gas phase absorption spectrum was approximated by connecting these points with a smooth line. Figure 3.8 shows the gas phase absorption compared to the solution absorp-tion, measured with a Hewlett Packard (Palo Alto, USA) 8452A diode array UV-Vis spectrophotometer. The gas phase absorption is blue shifted 22 nm from the solution spectrum. The gas phase 5-CR 6G ions are clearly absorbing at the argon ion wave-lengths (476.5, 488.0, 496.5, 501.7, and 514.5 nm). Differences in fluores-cence quantum efficiencies at these wavelengths could affect the amount of photodissociation observed. Measuring fluorescence emission and photodis-sociation at each wavelength could be used to generate a more accurate absorption spectrum. Chapter 3. Spectroscopy and Optics 78 0.9 0.9 0.8 § 0 7 .2 0.8 0.7 o g 0.6 | 0.5 « — g a s phase 0.2 0.2 as 0 1 •*- solution 0.1 0 0 0 100 2 0 0 3 0 0 4 0 0 500 6 0 0 7 0 0 8 0 0 wavelength (nm) Figure 3.8: The solution and gas phase absorbance of 5-CR 6G. The gas phase absorbance was approximated by measuring the percentage of the molecular ion which was photodissociated (right vertical axis). The same laser power was used at each wavelength. The gas phase absorbance is blue shifted « 22 nm relative to the solution absorbance. Adduct Ions The first mass spectra obtained using the ESI source usually included adduct ions. It was assumed that adduct ion formation was the unavoidable conse-quence of the electrospray process. Adduct ions are non-covalent complexes of the molecular ion of interest and an impurity or solvent molecule. The unwelcome molecule was likely water or methanol. Problems with the ion optics delayed the development of an accurate mass calibration method so Chapter 3. Spectroscopy and Optics 79 a definitive identification of the adduct was not made. The intensity of the adduct peak varied from 1-5% of the molecular ion peak. Photodissociation of the molecular ion required a certain amount of laser power depending on the irradiation time, buffer gas pressure and trapping po-tential. Photodissociation of the molecular ion could be completely avoided by using low enough laser power or short enough trapping times, but the adducts were much more fragile. In contrast, the adduct ion could not sur-vive even the lowest possible laser power setting. The fragile nature of the adduct ions suggests that the trapped 5-CR 6G ions gain vibrational energy very quickly upon irradiation. Since the adduct ions did not survive any amount of irradiation, they could not affect the fluorescence experiment. It was later determined that leaks in the Qo vacuum chamber were causing the adduct ions to form. After eliminating these leaks the adducts were no longer observed. 3.4.2 Photodissociation Studies The fluorescence signal was still undetectable; fluorescence (if any) was buried in the noise. To detect fluorescence, the signal would have to be increased or the noise would have to be decreased. Increasing the collection efficiency of Chapter 3. Spectroscopy and Optics 80 the optical system could increase the fluorescence signal. Only 0.16% of the ion cloud can be imaged through the 5 mm slot cut into the entrance endcap. Increasing the size of the slots in the ion trap electrodes would allow for larger collection optics, but significant changes to the electrodes would degrade the performance of the instrument as a mass spectrometer. Larger collection optics may also collect a disproportionately greater amount of background scatter. It was therefore believed that larger collection optics would not help separate the fluorescence from the excitation background. The fluorescence signal could be enhanced by increasing excitation inten-sity (see Equation 3.1). However, increasing the laser intensity will cause photodissociation of the ions. At this point, a better understanding of the photodissociation mechanism was needed to optimize the fluorescence exper-iment. Photodissociation vs. Laser Power Photodissociation was obviously dependent on laser power and irradiation times. When ions were trapped at lower RF-voltages (i.e. lower qz) they could withstand greater laser powers without dissociating. The ion cloud is more diffuse at lower qz values, so the benefit of using higher laser pow-Chapter 3. Spectroscopy and Optics 81 ers would be mitigated by a lower ion concentration. Simply changing the trapping potential and laser powers was not enough to produce a detectable fluorescence signal. The degree of photodissociation varies with laser power. Lower laser pow-ers can be used to generate modest fragmentation, while higher laser powers can be used to increase the fragmentation. Figure 3.9 shows the photodis-sociation of 5-CR 6G at two different laser powers. A more detailed picture of the photodissociation mechanism can be seen by monitoring the molecu-lar ion (precursor) and the first fragment ion (product) as the laser power is slowly increased. Figure 3.10 shows that the photodissociation does not start until a critical threshold is reached. The precursor ion then decreases nearly linearly with laser power. The first fragment increases in intensity before it is photodissociated at higher laser powers. If photodissociation was caused by a single photon absorption, there would not be a threshold as seen in Figure 3.10. This minimum energy required for photodissociation indicates that there must be a cooling mech-anism to counteract the heating caused by the laser irradiation. Radiative emission would certainly help dissipate energy gained from absorption, but the quantum efficiency would have to change with laser power to produce Chapter 3. Spectroscopy and Optics 82 1 oo % • 50 • -i 1 r*——h \ ft l 100 1 Jo 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 570 m/z I O O %H 50 160 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 570 m/z Figure 3.9: Mass spectra showing the photodissociation of 5-CR 6G at two different laser powers. Laser power is 0.15 W for upper trace and 0.5 W for lower trace. the threshold. Trapped ions undergo collisions with neutral helium atoms with a mean time between collisions on the order of 1 [is if the buffer gas pressure is 1 mtorr. Typically, photon absorption occurs in 10~15s and fluorescence emission happens within 1 — 3 ns after absorption. Collisions with neutral buffer gas would not affect the fluorescence spectroscopy but could help cool the ions between absorptionevents. If photodissociation is the result of a Chapter 3. Spectroscopy and Optics 83 5-CR 6G 0 0.1 0.2 0.3 0.4 0.5 0.6 Laser Power (W) Figure 3.10: Peak height of 5-CR 6G's molecular ion and first fragment vs. laser power. Photodissociation of the molecular ion produces a first fragment which itself is photodissociated at higher laser powers. slow heating process requiring the absorption of many photons, collisions with buffer gas molecules could help cool the ions. Photodissociation vs. Pressure The effect of buffer gas pressure was determined by measuring the laser power required to dissociate the precursor ion by 50% at different pressures. Chapter 3. Spectroscopy and Optics 84 0.6 Buffer Gas Pressure (mtorr) Figure 3.11: The laser power required to photodissociate 50% of the 5-CR 6G molecular ion is plotted vs. helium buffer gas pressure. Higher laser powers are required to photodissociate ions trapped at higher buffer gas pressures. Higher laser powers were required to fragment ions held at higher buffer gas pressures (see Figure 3.11). The pressure dependence suggests that the photodissociation of trapped ions is a slow process that can be affected by the buffer gas. The purpose of studying the photodissociation of trapped ions is to find ways to prevent it. Increasing the buffer gas pressure allows a higher laser power to be used before ions are destroyed by photodissociation. Typical ion Chapter 3. Spectroscopy and Optics 85 trap mass spectrometers are operated with buffer gas pressures from 1 x 10~5 to 1 x 10~3 torr. During fluorescence experiments the vacuum system and mass spectrometer should be operated close to their high pressure limits. This will allow a significantly higher laser power to be used and therefore produce a much higher fluorescence signal. Helium is the buffer gas most commonly used in ion traps. As discussed in Section 2.2.2, the buffer gas improves the performance of the ion trap. Helium is used because a heavier collision partner would cause greater deviations in the trajectory of the ion with which it collides. To kinetically cool a trapped ion it is better to have many collisions with a light buffer gas, than to have fewer collisions with a heavier one. On the other hand, if a heavier buffer gas could decrease the internal energy of the ions more effectively it ought to be used in fluorescence experiments. To investigate the effect of heavier buffer gases on photodissociation some minor instrumental changes were necessary. The plumbing on the high pres-sure side of the leak valve was changed to accommodate several different gas lines. Each gas cylinder had its own regulator and could be isolated from the leak valve. A roughing pump was also included so that the high pressure side of the leak valve could be evacuated to ensure the buffer gas could be Chapter 3. Spectroscopy and Optics 86 changed cleanly. This configuration permitted quick changes between the different gases tested. To accurately monitor the pressure in the ion trap a model 120AA Bara-tron gauge (model 120AA: MKS Instrumentation Inc., Andover, USA) was used. The pressure gauge normally used with the ion trap is an ionization gauge (model 342: Granville-Phillips, Boulder, USA). Ionization gauges de-pend on the ionization potential of the gas which is being measured and are not as accurate as Baratron gauges. A Baratron gauge will also accurately measure the pressure of a mixture of gases while the output of the ionization gauge would depend on the exact ratio of gases present. Figure 3.12 shows the results for three gases: argon, nitrogen, and helium. Heavier gases are more effective at cooling the ions and could therefore be used to improve the fluorescence experiment. The internal energy of trapped ions is increased by the absorption of photons. These ions can be cooled by collisions with the buffer gas and by emitting radiation. If the heating rate exceeds the cooling rate, the ions will eventually fragment. When the ion trap is filled to capacity or the ions are held at a low RF-potential, the ion cloud will be large and ions will oscillate in and out of the laser beam. When an ion is in the beam it will be heated Chapter 3. Spectroscopy and Optics 87 • A • Ar • N2 A He .6 .8 1.0 1.2 1.4 1.6 1.8 2.0 Buffer Gas Pressure (mtorr) Figure 3.12: The laser power required to photodissociate 50% of the 5-CR 6G molecular ion is plotted vs. buffer gas pressure for argon, nitrogen and helium gasses. Using a heavier buffer gas can de-crease the amount of photodissociation and allow a higher laser power to be used. by absorption and cooled by collisions with the buffer gas. Outside the laser beam the ion will only cool. The more time the ion spends outside the laser beam the more difficult it will be to photodissociate. Chopping the Laser Beam Another experiment was designed to prevent photodissociation involved chop-ping the laser beam. A mechanical chopper (SR540, Stanford Research Sys-Chapter 3. Spectroscopy and Optics 88 tem Inc., Sunnyvale, USA) was used to give ions more time to cool. Even though the background noise increases with laser power, the S/N. should still increase as more fluorescence and background are collected. Figure 3.13 demonstrates the effect of chopping the laser beam at two different laser powers. A chopping frequency of zero corresponds to continuous radiation. The molecular ion signal increased with chopping frequency until a steady value was reached. The chopping allowed the ions more time to cool through collisions with the buffer gas. Chopping the laser beam should increase the S/N of the fluorescence measurement by allowing a higher laser power to be used. Unfortunately, increasing the trap pressure and chopping the laser beam was still not enough of an improvement to detect fluorescence. Although the fluorescence experiment was not successful, the chopping experiments were still interesting. Dunbar had already completed similar ex-periments in the mid 1980's.45 Dunbar's experiments were performed in an ion cyclotron resonance (ICR) trap. 4 5" 4 7 An ICR trap uses magnetic fields to confine ions 4 8 at low operating pressures. The main advantage to using an ICR trap to conduct photodissociation experiments is the low operat-ing pressures (0.5 — 3 x 10~8 torr) where heating rates can be measured without interference from collisional cooling. For fluorescence experiments Chapter 3. Spectroscopy and Optics 89 100 chopping freq (Hz) Figure 3.13: The 5-CR 6G molecular ion peak height is plotted vs. the fre-quency at which the 488.0 nm laser excitation was chopped for a 3 s experiment. Results for two laser powers are shown (0.08PF upper trace and 0.12W lower trace). Chopping the laser beam provides more time for the ions to cool thus reducing photodis-sociation. collisional cooling is a welcomed benefit to working with the higher pressure of a quadrupole ion trap. 3.4.3 Fluorescently Tagged Peptides Photodissociation limits the intensity of the laser light that can be used to excite gas phase molecular ions. Also, any fluorescence emission was likely Chapter 3. Spectroscopy and Optics 90 not significantly Stokes shifted from the excitation wavelength. It was hy-pothesized that attaching a fluorescent molecule to a small peptide might address one or both of these problems. If the fluorescent molecule could dis-sipate excess vibrational energy to the peptide, it may be able to withstand more excitation and emission cycles without dissociating. Unlike collisions with the buffer gas, the peptide tag may help dissipate excess vibrational energy when the molecule is in the excited state and induce a greater Stokes shift in the emission. In condensed phase experiments the solvent or matrix removes excess vibrational energy of the excited state. Two very similar tagged peptides (see Figure 3.14 for structures), fluorescein-glycine-histidine (FITC-Gly-His) and fluorescein-glycine-glycine-histidine(FITC-Gly-Gly-His), were purchased from Molecular Probes Inc.(Eugene, USA). For these experiments, 1 fiM in 50:50 water methanol solutions of the pep-tides with 0.1% acetic acid were used. Adding acetic acid ensures that the molecules are positively charged in solution and easily sprayed. The peptides seemed to withstand a greater amount of laser power than the dye molecules used in previous experiments. Of course, this decrease in photodissociation may be caused by a lower molar absorptivity and com-paring molecules is difficult. There was some evidence that the peptide Chapter 3. Spectroscopy and Optics 91 FITC-Gly-Gly-His FITC-Gly-His Figure 3.14: Structures of fluorescently tagged peptides FITC-Gly-Gly-His (ion is 659.6 Th) and FITC-Gly-His (ion is 602.6 Th). tag was acting as a heat sink for the dye, but fluorescence emission could not be detected. As the laser power was increased the tagged peptides ap-peared to be losing histidine and then glycine. To break the peptide bonds the molecule must have transferred energy from the dye tag to the peptide. Only detectable fluorescence emission could prove that this energy transfer was happening fast enough to affect the excite state. Differences between the CID fragmentation and the photodissociation were also observed (see Figure 3.15). Although a costly addition to a mass spectrometer, a photodissociation capability could be used to increase the selectivity of an ion trap. An important qualification is that one would need to find a chemical system where MSn was not already selective enough. PAH analysis would be one case where photodissociation would certainly be more Chapter 3. Spectroscopy and Optics 92 fluorescein-Gly-His+ Photodissociation fluorescein 40 *0 80 1O0 120 HO 160 180 2<X) 210 240 1GO 2SO JOG .1JO MO ,10O 3S0 400 410 440 460 4SO 500 J20 540 560 380 ' fluorescein-Gly-His+ CID fluorescein-Gly+ fluorescein' ••iWl.lillfa.JJ..UJJI,L liml l.iu.1. Ii 40 00 90 I0O 110 MO 160 180 200 110 HO 160 280 JOO J20 340 560 380 40O (lO 440 460 4SO SOO 320 340 360 380 600 620 640 Figure 3.15: Photodissociation spectrum of FITC-Gly-His (top) and it's CID spectrum (bottom). selective than CID. 3.5 Requirements for a Successful Fluorescence Experiment The methods, technology and instrumentation used to this point were not adequate to allow the detection of fluorescence from molecular ions confined Chapter 3. Spectroscopy and Optics 93 in a quadrupole ion trap. The original instrumentation and methods were not adequate because critical aspects of this experiment were not anticipated when the project began. Photodissociation limits the laser intensity and this alone would make gas phase fluorescence measurements at least three or four orders of magnitude less sensitive than solution measurements. The problem of separating the fluorescence from the excitation was also not expected. While it was expected that the Stokes shift of the emission in the gas phase would be different than in solution, the fluorescence was still expected to be wavelength separated from the laser generated background. The laser background, composed of Raman scattering and fluorescence from the win-dows, was very significant. At this point, it was thought that the fluorescence emission was not Stokes shifted away from the laser generated background and the gas phase experiment was no longer a zero background measurement. If the fluorescence experiment was to be successful, major improvements to the instrumentation were needed. If fluorescence emission was to be de-tected then the S/N would have to be increased dramatically. The total number of ions trapped must also be increased. The trapping efficiency was improved by optimizing the ion optics of the ESI front-end. The main prob-lem yet to be solved was reducing the amount of background light collected. Chapter 3. Spectroscopy and Optics 94 A new optical system and new detectors were necessary. Chapter 4 95 Pulse Counting Experiments It was not possible to detect fluorescence emission from trapped molecular ions with the methods and instrumentation described in the previous chap-ter. It did not appear that wavelength dispersion with a spectrograph could separate the fluorescence from the background generated by the excitation. The background signal was much too high and any fluorescence would likely be too close in wavelength to the excitation. For fluorescence to be detected, new instrumentation and a new approach would be necessary. In an attempt to solve this problem methods used to detect small signals in other fields of analytical chemistry were considered. Lock-in amplification is one technique that can dramatically improve the S/N when measuring a small signal in the presence of a large, noisy background. Basically, the analytical signal of interest is modulated and the entire signal (background + analytical signal) is multiplied by a wave, which has been generated at Chapter 4. Pulse Counting Experiments 96 the same frequency as the modulation and with a fixed phase. The result of this multiplication (demodulation) is that the signal is recovered with only a small fraction of the original noise. Only noise that happens to fall within the narrow bandwidth of the lock-in amplifier will be detected. This technique would also be able to detect any emission that was resonant with the excitation. For example, lock-in amplification is used to detect fluorescence emission from atoms in flames. The excitation source is modulated at a certain fre-quency so that the emission from the atoms become modulated at that same frequency. Most of the background generated by the flame stays constant (DC), or only varies at low frequencies, while the analytical signal is mod-ulated to a much higher frequency and can be detected against an almost zero background. Unfortunately, in the ion trap fluorescence experiments it is the excitation source (laser) that is creating the unwanted background and modulating the laser would also modulate the background and nothing would be gained. Modulating the fluorescence without modulating the background signal may be achieved by using a technique known as sample modulation. For the most part, sample modulation has never been a very effective or pop-Chapter 4. Pulse Counting Experiments 97 ular technique. In molecular absorption and luminescence spectroscopy the sample and reference cells can be moved in and out of the the light path to modulate the signal. In atomic emission, fluorescence, and absorption ex-periments the sample introduction system can be turned on and off. These techniques are not very effective because the signal modulation is usually too slow (low frequency) to separate the analytical signal from the DC back-ground. If a nebulizer is used to inject a sample into a plasma (or flame) the act of modulating the sample can also modulate the plasma background emission at the same frequency. However, in an ion trap the ions are already moving at well defined secular frequencies. In an ion trap experiment, sample modulation could be achieved by mov-ing the ions in and out of the laser beam. In most of the experiments de-scribed so far, the ions are already moving in and out of the laser beam. To enhance the modulation the laser beam would be tightly focused and a reso-nance excitation voltage applied across the endcap electrodes (Section 2.1.3). A higher voltage for the resonance excitation would result in a greater per-centage of the ions moving in and out of the beam on every cycle but it would also heat the ions. Hence the resonant excitation would have to be kept modest to avoid CID (Section 2.2.4). Chapter 4. Pulse Counting Experiments 98 The use of sample modulation and lock-in amplification requires a detec-tor that will generate an analog signal if amplitude modulation, as described above, is to be used. An obvious choice is a sensitive photomultiplier tube (PMT). Experiments with P M T detection had already been attempted but were unsuccessful because of the high level of background. With the existing optical system, the P M T operating voltage would have to be kept so low that its detection efficiency would be severely compromised. To obtain a sufficient analog fluorescence signal the P M T gain should be as high as possible. How-ever, the maximum achievable gain was also limited by the large background. Before lock-in amplification and sample modulation can be applied to sepa-rate the fluorescence signal from the background, the background must first be reduced by other means. To reduce the background, a new optical collection system was designed. Confocal imaging was abandoned, and all of the collection lenses were re-placed with a fiber optic probe. The fiber optic probe was designed to be inserted between the entrance endcap and ring electrodes of the ion trap. The fiber optic probe blocked any stray light reflected off of the outside of the trap electrodes. Unfortunately, laser generated background continued to scatter off the inside of the electrode surfaces. Chapter 4. Pulse Counting Experiments 99 An additional approach to reducing the background was to incorporate in-strumentation used for Raman spectroscopy into the fluorescence experiment. Small signals are detected very close to laser lines in Raman spectroscopy. A relatively new type of holographic notch filter has been developed for Ra-man experiments that can block a very narrow bandwidth. To attenuate the scatter from the inside electrode surfaces such a filter was obtained (HSPF-488.0-1.0, Kiaser Optical Systems, Inc.). This filter was designed to have an optical density (O.D.) greater than 6.0 at 488.0 nm and a bandwidth of less than 10 nm. To achieve the rated performance the light to be filtered must be perfectly collimated and hit the filter at a specific angle. The fiber optic probe was designed to accommodate a holographic notch filter. Experiments using the notch filter are described in Chapter 5. 4.1 Fiber Optic Probe The switch to a fiber optic probe for light collection eliminates serious weak-nesses of the previous optical system. Most importantly, the background scatter from the outside of the electrodes no longer reached the detector. Also, the old system was difficult to align and vulnerable to inadvertent mis-Chapter 4. Pulse Counting Experiments 100 alignment when other parts of the instrument were being manipulated. The fiber optic probe was very easy to align and its correct positioning could be quickly verified. Using optical fibers to collect and guide light is simple and effective. The fiber optic probe was designed with the following criteria: it must fit through the slot cut into the endcap electrode, collect light efficiently, be compatible with a holographic notch filter, and be vacuum tight so that an exit window is not necessary. An exit window would mean two more inter-faces where signal would be lost to reflections. To make the probe durable and vacuum tight the fibers were contained in a stainless steel tube. The fibers must protrude several centimeters from the stainless steel tube to prevent arcing from the high voltage on the ring electrode. Making a vacuum seal on the outside of the tube (1/4 inch outside diameter) is easy and commercially available tube connectors (SS-4-UT-1-4, Swagelok, Solon, USA) were used for this purpose. To make the inside of the probe vacuum tight an epoxy (Torrseal, Varian, Inc., Walnut Creek, USA) was used to fill the spaces between the fibers. A prototype was constructed. After these fibers were glued, they were cut and polished using lapping paper attached to a rotating disk. Unfortunately, the cutting and polishing heated the fiber Chapter 4. Pulse Counting Experiments 101 optic bundle excessively and the epoxy melted and contaminated the ends of the fibers. This meant a window had to be used to make the vacuum seal after all. The new fiber optic probe was designed so any one inch diameter optic could be used to make the vacuum seal. The fibers used for the bundle were chosen so the probe would collect light efficiently. Superguide fused silica fiber optic (SFS400/440T, Fiberguide Industries, Stirling, USA) was purchased with the following specifications: 400 pm core diameter, 40 \im thick cladding, and a 40 am thick Thermacoat jacket. This fiber has a numerical aperture of 0.22 which corresponds to an acceptance angle of 12.7°. The Thermacoat jacketing material was chosen because it can withstand —190°C to 300°C and is thin. The thinnest possible jacketing was chosen because only light hitting the core will be transmitted through the fiber. The temperature stability of the jacketing material will allow the probe to be used in experiments where the ion trap electrodes are heated or cooled. To keep the probe durable, the jacketing material was not removed. Seventy individual lengths of fiber were cut and polished using both silicon carbide (9//) and aluminum oxide (0.3//) lapping paper with a commercial polishing device (Ultrapol 1200, Ultra Tech Mfg., Inc., Santa Ana, USA). The fiber optic probe is shown in Figure 4.1. The input end of the probe Chapter 4. Pulse Counting Experiments 102 is easily aligned with the ion cloud because the fiber bundle just fills the access slot cut into the entrance endcap. At the other end of the probe, the filter holder attaches to a one inch tube lens system (Thorlabs, Inc., Newton, USA) and can be coupled directly to any detector equipped with a compatible flange. Any one inch diameter filter, or window, can be used to make the vacuum seal. The optic is simply held against an o-ring positioned at the bottom of the filter holder. The probe is attached to the vacuum system using a KF-25 flange. To estimate the collection efficiency of this probe the total cross-sectional area of the bundle must be compared to the area represented by the fiber cores. Assuming the fibers are efficiently packed, the empty spaces between the fibers will represent 9.31% of the cross-sectional area. The core is 400 \im in diameter and the total diameter of the fiber, including the cladding and jacketing, is 480 pm. Therefore, the area of the core is 69.4% of the total fiber cross-section. Assuming a 5% loss to reflections and negligible attenuation within the fibers, the entire probe will transmit 60% of the light striking it with an angle of incidence below 12.7°. The fiber optic probe is positioned approximately 2 cm from the centre of the ion trap. The total effective cross-sectional area of the probe is simply Figure 4.1: Fiber optic probe. Chapter 4. Pulse Counting Experiments 104 70 times the cross-sectional area of one core (0.088 cm 2). The emission orig-inating from the centre of the trap can be approximated by the surface area of a sphere with a 2 cm radius. The probe's effective area is approximately 0.18% of this sphere's surface area. Adding losses from reflections at both ends of the probe, this probe transmits « 0.16% of the light originating at the centre of the trap. 4.1.1 PMT Detector for the Fiber Optic Probe A P M T housing (model SWH/RFI, Electron Tubes Limited, Rockaway, USA) and a P M T (R1477-06, Hamamatsu, Hamamatsu City, Japan) were used for these experiments. The P M T housing has an RF-shielded socket assembly which helps reduce noise pick-up from the ion trap electronics. The R1477-06 has a high sensitivity multialkali photocathode with a spectral response from 180 nm-900 nm (peak at 450 nm). A special flange was built in the UBC Chemistry Mechanical Engineering Shop to couple the P M T housing to the fiber optic probe. The high voltage for the P M T was supplied by a Stanford Research Systems, Inc.( Sunnyvale, USA) power supply (model PS350/5000 V - 35 W). The background noise was reduced considerably by using the fiber optic Chapter 4. Pulse Counting Experiments 105 probe. The exact level of reduction cannot be measured because different PMTs and housings were used before the probe was added. In addition, improved spatial filtering of the laser beam and other adjustments to the excitation optics had also reduced the amount of unwanted background. 4.2 Coating the Ion Trap Electrodes While assembling the instrumentation for lock-in amplification detection Dr. John Hepburn was consulted. This invaluable discussion lead to the idea of coating the ion trap electrodes with carbon. The hemispheres of electron momentum spectrometers are coated with carbon to adsorb any traces of pump oil. Coatings on ion trap electrodes could potentially degrade the performance of the instrument as a mass spectrometer. If the surfaces in an electron momentum spectrometer can be coated with carbon, it was thought that ion trap electrodes could also be coated. A commercial product, Aquadag, was used to coat the inside surfaces of the ion trap electrodes. Aquadag is an aqueous-based colloidal dispersion of 5 pm diameter carbon particles produced by Acheson Colloids Company (Port Huron, USA). Aquadag is used as a commercial lubricant because it Chapter 4. Pulse Counting Experiments 106 adheres well to most metal surfaces and has a long wear life. The fine particle size produces a very stable suspension for commercial applications. The fine particle size also produces a very black coating; particles with diameters exceeding 10 u.m would produce a shiny grey coating. Aquadag is commonly applied to metal surfaces using a brush or by immersing the metal into a diluted suspension of the particles. Changing the shape or roughening the surface of the ion trap electrodes would degrade the performance of the ion trap as a mass spectrometer. To produce an even coating, the Aquadag was diluted in ethanol (50:1 by volume) and sprayed onto the electrodes using a nebulizer. The metal surface was kept hot during the coating process to avoid creating rings of dried carbon. A hot plate was used to heat the electrodes to between 80°C and 100°C The electrode temperature was monitored using a thermocouple during the spraying process. A spare set of ion trap electrodes was coated to test the impact of the coating on mass spectrometer performance. The coated set of electrodes was evaluated for a week and they did not negatively impact on the ion trap performance as a mass spectrometer. The coating was then applied to the endcaps which had been modified for optical experiments. Flash photography was used to demonstrate the effectiveness Chapter 4. Pulse Counting Experiments 107 Figure 4.2: Endcap electrodes with (left) and without (right) carbon coat-ings. Carbon coating reduces scatter off the inside electrode sur-faces. of the coating (see Figure 4.2). The black coating on the electrodes reduced the background by « 3000 times. Figure 2.2 shows the coated spare end-cap electrodes and an uncoated ring electrode; it is easy to appreciate how uniform the coating is in this figure. To prepare for the lock-in amplification experiment R6G was sprayed into the trap and irradiated with 488.0 nm light from the argon ion laser. The output of the P M T was monitored directly with an oscilloscope ( TDS 340A Tektronix, Beaverton, USA). Individual pulses were detected indicating that Chapter 4. Pulse Counting Experiments 108 the P M T was now operating in a pulse counting regime. The output of the P M T was no longer a DC current as was the case in past experiments. Since the analytical signal was not sufficient to create a DC current output, amplitude modulation and lock-in amplification cannot not be used. While it was difficult to see changes in the P M T output by visually mon-itoring the oscilloscope, there appeared to be a higher frequency of pulses with ions in the trap. Pulse counting hardware was attached to the P M T to get a precise measure of its output. A F-100E pre-amplifier and discriminator was connected to the PMT, and the output of the F-100E was connected to a PRM-100 precision ratemeter. Both devices were manufactured by Advanced Research Instrumentation, Corp (Wheat Ridge, USA). This preliminary experiment resulted in the first successful detection of emission from trapped molecular ions using the instrument described in this thesis. The first signal to background ratio was approximately 2:1. At this point it was clear that emission was being detected from trapped R6G ions. However, it was possible that the emission was Rayleigh scattering and not fluorescence. A long pass filter (CG-OG-530, CVI Laser Corporation) was placed in front of the P M T and the laser background dropped dramatically and the Chapter 4. Pulse Counting Experiments 109 S/N ratio improved to greater than 20 for 10 s experiments. The signal enhancement from using a long pass filter is evidence that the emission de-tected is indeed fluorescence. Without a fluorescence emission spectrum, it is difficult to predict how much fluorescence is also being blocked by the filter. A problem was encountered during these preliminary experiments: the probe was poorly aligned with the centre of the trap because the KF-25 flange and the filter holder were not parallel. This alignment problem was quickly fixed by Des Lovrity of the UBC Chemisrty Mechanical Engineering Shop. 4.3 Pulse Counting Instrumentation At this point the instrument was adjusted to accommodate pulse counting experiments. The pulse counting experiments had to be synchronized with the ion trap control software. Software was developed to control the pulse counting experiment and to record data. Figure 4.3 shows the configuration of the instrument used for pulse counting experiments. Proper spatial fil-tering of the laser beam and the use of a Brewster window are important aspects of the instrumental design. However, it is the fiber optic probe and carbon coated electrodes, and to a lesser extent the internal beam dump, that Figure 4.3: Instrumental configuration for pulse counting experiment. resulted in exceptional performance. Figure 4.4 is a block diagram showing the more important peripherals used in the experiment. Chapter 4. Pulse Counting Experiments 111 Pulse Counting Data Acquisition Coumputer Digital Delay Generator Pulse Counter Pre-amp PMT 4 Ion Trap ESI Teledyne 3DQ Control Computer Figure 4.4: Instrumental block diagram for pulse counting experiment. Chapter 4. Pulse Counting Experiments 112 4.3.1 Data Collection for Pulse Counting A data acquisition board (AT-MIO-16, National Instruments, Austin, USA) was borrowed so that Lab VIEW could be used to develop the software nec-essary for the pulse counting experiment. The PRM-100 ratemeter is an excellent counter with a pulse pair resolution of less than 10 ns. Unfortu-nately, it does not have a convenient way of outputting count rates. To get around this problem, two counters on the AT-MIO-16 board were used to count T T L pulses generated by the PRM-100 ratemeter. For every current pulse from the P M T that is above the threshold set by the discriminator, the PRM-100 receives an ECL pulse from the F-100E pre-amplifier. The PRM-100 converts the ECL pulses into T T L pulses. Counters in the PRM-100 then count the T T L pulses and the rates are displayed on the front panel of the ratemeter. The PRM-100 outputs rates based on a time scale set on the front panel. There was no way to set the counting pe-riod exactly without writing a complete instrument driver for the PRM-100. Instead, the T T L pulses generated by the PRM-100 were counted directly using a counter on the AT-MIO-16 data acquisition board. The counting period was easily and accurately synchronized with the ion trap experiment. The Teledyne electronics can generate a pulse at the Chapter 4. Pulse Counting Experiments 113 beginning of every trapping sequence. This "start" pulse triggers a digital delay generator. The digital delay generator then sends a pulse to the gate lens between Q0 and Q\ allowing ions to enter the ion trap. The gate pulse drops the gate voltage by 19 V and lasts for 1 ms for all experiments. At a chosen delay after this, a pulse is sent to the data acquisition board to define the pulse counting period. The counters of the AT-MIO-16 can be gated on and off themselves. The pulse from the digital delay generator is sent to the gate input of one counter and at the same time sent to the input of a second counter. The input of the "gated" counter is, of course, the T T L output of the ratemeter. The software program can record how many pulses are detected in a given counting period. The software also controls how many counting periods are measured in each experiment. The software saves the data from each counting period in an array and calculates the mean and standard deviation. Since the time between ion injection into the trap and pulse counting begins is small (usually < 10ms), the laser is kept on all the time. The timing diagram of the experiment is shown in Figure 4.5. Chapter 4. Pulse Counting Experiments 114 Trap RF Start Pulse _ l I Gate for ion injection L Pulse Counting I Figure 4.5: Timing diagram for pulse counting experiments. For a 1 s optical experiment the following timings were used: ion injection for 10 ms, ion storage for 1 s, mass scan of 50 ms, QQ injection gate of 1 ms, and pulse counting for 1 s. 4 . 3 . 2 P M T C o o l i n g It is important to note that the R1477-06 PMT was not designed for pulse counting experiments. The R1477-06 was designed for maximum sensitivity while a pulse counting PMT would be designed to have very low dark counts. Most of the dark counts arise from thermionic emission from the PMT's photocathode. At room temperature, the dark counts from the R1477-06 were much higher than the laser background when a filter was being used. To improve the S/N of the measurement, the dark counts were reduced by Chapter 4. Pulse Counting Experiments 115 lowering the temperature of the photocathode. To cool the photocathode the entire P M T housing was cooled using solid CO2. With 1000 V applied to the PMT, cooling to — 1°C reduced the dark noise to « 700 counts per second (CPS) compared to « 2700 CPS at room temperature. Cooling the P M T can cause condensation problems. To prevent conden-sation on the inside of the P M T housing, the base was filled with molecular sieves and the entire housing was sealed in a glove bag filled with nitro-gen. To prevent condensation on the outside of the P M T window, the lens tube system between the probe and the P M T was also filled with nitrogen. A pulse counting P M T which would not require such cooling was ordered. While waiting for the new PMT, several pulse counting experiments were conducted. 4.4 Fluorescence vs. Laser Power The first pulse counting experiments measured the laser power dependence of the fluorescence signal. The R6G ions were held at the same trapping voltage for each laser power (i.e. qz = 0.5). The counting period was 1 s and ten measurements were made at each laser power. Mass spectra were Chapter 4. Pulse Counting Experiments 116 also recorded for each laser power. Background measurements were made without ions in the trap. To ensure nothing interfered with the background subtraction, only the pulse sent to the gate lens was stopped; every other instrumental setting remained constant. R6G ions were excited with 488.0 nm light from the argon ion laser for all pulse counting experiments described in this thesis. The long pass filter (CG-OG-530, CVI Laser Corporation) was used to attenuate the laser generated background in all pulse counting experiments. 4.4.1 Instrumental Parameters The various instrumental parameters used in this experiment are noted in Table 4.1. These settings remain the same for all of the pulse counting experiments except were indicated. Helium was used as a buffer gas in all of the fluorescence experiments. 4.4.2 Results and Discussion As expected, the fluorescence signal initially increased with increasing laser power. At higher laser powers, photodissociation begins to reduce the number of ions and the fluorescence signal decreases. The amount of fragmentation Chapter 4. Pulse Counting Experiments 117 ESI & Ion Guide Component Voltage (V) Sprayer 3750 Curtain Plate 195 Sampling Orifice 80 Skimmer 46 Q0 Offset 35 Gate 38.5 Qi Offset 30 LI 30 L2 30 N2 flow 1.4 L/min Syringe Pump IpL/min MS Operating Pressures Zone Pressure Orifice / Skimmer 950 mtorr Qo 7.0 mtorr Ion trap 1.4 mtorr Teledyne Software Settings Injection qz 0.16 Detector Voltage 1000 V Scan Speed 12000 Th/s Resonant Excitation 549 kHz offset 5.0 V(48pV/DAC) Linear Quads 300 V Table 4.1: Instrument settings. was determined in the mass spectrum acquired at the end of every optical experiment. When evaluating the effect of photodissociation, it is important to know that the amount of fragmentation only indicates what has happened to the ions by the end of the irradiation period. Figure 4.6 shows the mass spectrum of R6G and fragments recorded for each laser power. The fluorescence signal as a function of laser power is plotted in the insert of Figure 4.6. Curiously, the maximum fluorescence signal corresponded to some photodissociation. The fluorescence signal is directly proportional to both ion concentration and laser power. Increasing the laser power initially increased the signal, but after a certain threshold (Figure 3.10) the signal decreased as the ion concentration was reduced by Chapter 4. Pulse Counting Experiments 118 300 400 m/z Figure 4.6: Background subtracted fluorescence signal of R6G us. 488.0 nm laser power (insert) and corresponding mass spectra. Maximum fluorescence signal corresponds to some photofragmentation of R6G. photodissociation. The maximum fluorescence signal corresponded to some photodissociation, but the dissociation occurred gradually over the whole irradiation period. Further experiments were needed to explain why the maximum signal corresponded to some photodissociation. Chapter 4. Pulse Counting Experiments 119 The mean S/N ratio of the fluorescence signal for these experiments is 33. The standard deviation calculated for the ten measurements at each laser power was used as the noise in the measurement. The fluorescence signal is the total signal minus the background signal that was collected with no ions in the trap. Fluorescence signals reported in this thesis always have the background subtracted. The background signal is a combination of detector dark counts and back-ground light from the laser scatter and room lights. S/N ratios are not to be confused with signal to background ratios. The standard benchmark for a signal to be detected is a S/N= 3. The signal does not have to be three times the background, but rather three times the total noise in the measurement. Figure 4.7 shows the fluorescence signal, background signal, and measured S/N ratio plotted against laser power. The S/N ratio is still very good even when the background is larger than the signal. There are many possible sources of error in most analytical measure-ments. The laser intensity may fluctuate, but the total photon flux for a 1 s experiment should be very constant if the laser is operating properly. When operated in an analog mode, all pulse height fluctuations will end up on the P M T output as noise. In a pulse counting mode, pulse height fluctuations, Chapter 4. Pulse Counting Experiments 120 T ' 1 — - > 1 1 1 1 1 • 1- 100 1 1 1 ' 1 1 1 1 1 1 r 1-0 0 50 100 150 200 250 300 Laser Power (mW) Figure 4.7: Fluorescence signal of R6G and background signal vs. laser power (left axis). S/N ratio vs. laser power (right axis). unless very large, will not affect the count rate. The P M T gain is set high enough so that the discriminator level of the pulse counting system can be easily set higher than the amplifier noise. For this reason, pulse counting experiments have a significant S/N ratio advantage over analog detection. In photon counting experiments, where signal pulses are detected for a certain period of time, there will always be fluctuations in the count rate. Chapter 4. Pulse Counting Experiments 121 These fluctuations, called shot noise, can be expressed as a Poisson distribu-tion. The shot noise is simply the square root of the number of counts and represents the lowest achievable noise level in a photon counting experiment. Shot noise contributions from the fluorescence (S), background light (Nb), and dark counts (Nj) are given by the following relationships: Since this fluorescence measurement is background subtracted, the total noise (ntot) hi the measurement is given by the following equation: The factor of 2 in Equation 4.1 arises because background noise (n\ + n2d)\s present in the signal and in the subtracted "blank" measurement. The back-ground shown in Figure 4.7 is dominated by dark counts of « 740 CPS. The S/N would be improved by using a pulse-counting P M T that had fewer dark counts. However, the S/N measured for these fluorescence experiments were very close to the best that could be achieved in a pulse counting experiment (i.e. the most of the noise is from the shot noise of the signal itself and not noise from signal = ns = noise from background light = rib = noise from dark counts = rid = (4.1) Chapter 4. Pulse Counting Experiments 122 from the background signal). 4.5 Fluorescence vs. Trapping Level (qz) In this experiment the effect of the trapping level was examined. The in-strumental parameters used in the previous experiment were not changed (Table 4.1). As with all of the pulse counting experiments discussed in this thesis, R6G was the molecule interrogated. For this experiment the laser power was held constant at 120mW while the trapping level (qz) was varied. 4.5.1 Results and Discussion The results of the experiment are shown in Figure 4.8. The fluorescence signal initially increases with qz and then reaches a broad maximum before decreasing. As the trapping voltage, and therefore qz, was increased, the ion cloud becomes smaller. A smaller ion cloud means the ions are more concentrated. As expected, the increased concentration leads to a higher signal since fluorescence is directly proportional to concentration. However, as the ion cloud shrinks the ions spend more time in the laser beam. An ion spending a greater percentage of time inside the laser beam experiences Chapter 4. Pulse Counting Experiments 123 00 CL o c D) C/D <D o c 0 o o 3 2200-| 2000-1800-1600-1400-1200-1000-800-600-400-0.0 ~~' 01 ' 02 ' 03 ' 04^  ' 05 ' olo" Figure q z 4.8: Background subtracted fluorescence signal from 5-CR 6 G vs. qz for a fixed laser power. greater heating rates. Increased heating rates will eventually cause enough photodissociation to decrease the total fluorescence signal detected at high laser powers. Another factor that could play a role in this measurement is the collec-tion efficiency of the fiber optic probe. As the ion cloud size decreases, the collection efficiency may be improved because the ions will be emitting closer Chapter 4. Pulse Counting Experiments 124 to the centre of the trap. Changes in collection efficiency are not expected to have a large effect. While a lens outside of the ion trap would not be able to collect more light through the optical access slot, a lens mounted on the end of the probe may improve collection. One important consideration is that a lens on the end of the probe may also collect an even greater amount of scattered light. It is clear that the laser power or trapping level can be optimized to give a maximum fluorescence signal. For a global optimization, the signal depen-dance on the laser power had to be monitored at several different trapping levels. 4.6 Fluorescence vs. Laser Power and qz To optimize the fluorescence experiment, the signal dependence on the laser power was measured at six different trapping levels. The instrumental pa-rameters have not changed and are given in Table 4.1. As with the previous experiments, each data point is an average of ten 1 s measurements. Chapter 4. Pulse Counting Experiments 125 3000-^ 2800 -CL 2600 -O 2400 -CD & 2 2 0 0 : ^ 2000 -CD c 1800-d) oo 1600-cu O 1400-u- 1200-1000-—t— 50 100 150 200 Laser Power (mW) - • - q = .1 - • - q = .2 - A - q = .3 - • - q = .4 —•—q = .45 - ^ - q = .5 250 Figure 4.9: Background subtracted fluorescence signal from 5-CR 6G vs. laser power for six trapping levels (qz). Higher count rates are obtained for smaller ion clouds (larger qz) and lower laser power. 4.6.1 Results and Discussion The results of these experiments are shown in Figure 4.9. For low trapping levels the fluorescence signal increases fairly linearly and then appears to level off. At higher trapping levels (i.e. higher qz), the fluorescence decreases after reaching a maximum. For low qz values, the ion cloud is so diffuse that very little photodis-Chapter 4. Pulse Counting Experiments 126 sociation occurs at the laser powers used. As the trapping level increases photodissociation begins to limit the maximum fluorescence signal which can be collected. As qz increases, lower laser powers are needed to achieve the maximum signal. A more significant observation is that greater fluorescence signals can be achieved by using smaller, more concentrated ion clouds (i.e. higher qz) and lower laser powers. Using smaller ion clouds and lower laser power is superior to using higher laser power with diffuse ion clouds for several reasons. Increasing the laser power may increase the fluorescence but it also increases the heating rate and causes photodissociation. As well, more background scatter is generated at higher laser powers thus decreasing the S/N ratio of the measurement. As previously discussed, light collection may also be more efficient for smaller ion cloud sizes. Finally, it is likely that photodissociation has a greater than linear dependance on laser power because many photons are required to dissociate an ion. Chapter 4. Pulse Counting Experiments 127 4.7 Irradiation Time All pulse counting experiments described thus far have used 1 s storage times. In this experiment, the fluorescence signal was measured for each 1 s interval of a 10 s total storage time. Since the laser was on for the entire experiment the storage time was equivalent to the irradiation time. The experiment was repeated for two different laser powers and ten measurements were averaged for each interval. For both experiments, R6G ions were held at a trapping level corresponding to qz = 0.45. After 10 s of irradiation at the lower laser power (31 mW), almost no photodissociation was observed in the mass spectrum. The second laser power (174 mW) was chosen so that 80% of the molecular ions would be photodissociated at the end of the 10 s irradiation time. 4.7.1 Results and Discussion The results are displayed in Figure 4.10. For the higher laser power, the fluorescence signal is initially quite high but quickly decreases. The drop in signal is clearly the result of photodissociation. For the lower laser power, the fluorescence signal is very constant for each interval. The signal is constant because the concentration of ions is not decreasing from photodissociation. Chapter 4. Pulse Counting Experiments 128 Figure 4.10: Fluorescence measured in 1 s intervals of a 10 s experiment with high laser powers (causing fragmentation) and lower laser power (no fragmentation). Chapter 4. Pulse Counting Experiments 129 The total amount of fluorescence collected for the entire 10 s is slightly higher for the lower laser power. Also, the lower laser power created less background scatter. For the full 10 s, the lower laser power measurement has « 10% higher S/N ratio. However, using a lower laser power for a 10 s experiment is obviously not the way to improve the S/N ratio of the fluorescence experiment. To make best use of 10 s of measurement time, it would be better to use ten separate ion storage cycles and make ten 1 s measurements at the higher laser power. Shorter irradiation times and higher laser powers will give better S/N ratios until the trapping time becomes so short that the ion trap is no longer completely filled between experiments. 4.8 Final Optimization For complete optimization of the fluorescence experiment the full range of trapping potentials was investigated. In Section 4.6, only trapping levels up to qz = 0.5 were considered. Pressure effects were also evaluated. A new P M T specifically designed for pulse counting was used for the remaining pulse counting experiments. The Hamamatsu R4632 has a much lower dark count than the R1477-06. The P M T housing was cooled to « 15°C Chapter 4. Pulse Counting Experiments 130 using a Peltier cooler. To avoid condensation problems, the P M T was not cooled further. This modest cooling reduces the dark counts from 50 CPS at room temperature to 30 CPS at 15°C. The reduction of dark counts from 700 CPS, with the R1477-06, to 30 CPS with the R4632 significantly improved the S/N of the measurements. Using the R4632 pulse counting P M T the total background was reduced to 35 CPS consisting of: 30 CPS dark counts, 4 CPS laser generated background, and 1 CPS from room lights for most measurements. Al l other instrumental parameters (Table 4.1) remain unchanged except for pressure. 4.8.1 Complete Optimization at Two Pressures The photodissociation studies reported in Section 3.4.2 show that increas-ing the buffer gas pressure increases the cooling rate of absorbing ions and decreases photodissociation. Thus, higher pressures allow the use of higher laser powers. For the same concentration of ions, higher laser powers must increase the fluorescence signal. Unfortunately, the photodissociation results provide the only evidence that increased pressure will lead to improved flu-orescence detection. Increasing the pressure in the current instrument affects more than just Chapter 4. Pulse Counting Experiments 131 the cooling rates of trapped ions. It also affects the focusing properties of the ESI front-end, especially Q i , and injection efficiency into the ion trap (Section 2.5.2). Increasing the pressure in the trap will cause more ions to be trapped. Attempts to trap equal numbers of ions at each pressure would be complicated by the fact that ion ejection and detection is also pressure dependent. For this experiment, the laser power was optimized for each trapping level (qz) tested. As before, the maximum signal reported for each qz was the average of ten 1 s experiments. The optimization was conducted at 1 mtorr and 0.55 mtorr of helium buffer gas. 4.8.2 Results and Discussion The results of the optimization experiment are shown in Figure 4.11. A trap-ping level corresponding to qz = 0.65 gives the maximum fluorescence signal at both pressures. The higher pressure experiment gives greater fluorescence signals at every qz value. Increased cooling rates, from more frequent colli-sions with helium atoms, must contribute to this effect. However, the main reason for superior signals at the higher pressure is the increase in the total number of trapped ions. The mass spectra clearly show more ions being Chapter 4. Pulse Counting Experiments 132 trapped at 1 mtorr compared to 0.55 mtorr. Although this experiment does not prove that higher fluorescence signals can be obtained at higher buffer gas pressures, subtle differences in the results do illustrate some interesting complexities of the fluorescence experiment. At both pressures, the fluorescence signal increases linearly with qz at low qz values. At 0.55 mtorr, the linear portion of the graph has a lower slope and begins to roll off at lower qz values. Two related effects are responsible for this difference. First, the ion cloud size is much larger for the higher pressure; there are simply more ions in the ion cloud. At the lower pressure, the smaller ion cloud will requires less compression (lower qz) to become fully illuminated by the laser beam. As the ion cloud becomes fully illuminated there is less concentration advantage to further increasing qz (compressing the ion cloud) and the slope of the graph decreases. Furthermore, space charge effects are more serious for larger numbers of trapped ions. Space charge effects are caused by the electric fields generated by the ions themselves. At low ion numbers, the motions of trapped ions are primarily affected by only the quadrupole field. When the ion trap is filled to capac-ity the ion-ion repulsions become much more significant. The mass spectra shown in this experiment are severely "space charged". The ion-ion repul-Chapter 4. Pulse Counting Experiments 133 — • — 1 mtorr 0 100 200 300 ' 400 ' 500 600 (m/z)' 0 100 200 300 400 500 600 (mil) Figure 4.11: Optimization of qz and corresponding mass spectra. The upper graph shows 5-CR 6G fluorescence signal vs. qz for two differ-ent buffer gas pressures ( 1 mtorr and 0.55 mtorr). The mass spectra for both pressures are plotted on the same vertical scale. Chapter 4. Pulse Counting Experiments 134 sions interfere with ion ejection during the mass scan and ions appear in the mass spectrum as broad peaks. Especially sensitive to space charge effects are low concentration ions which have a slightly lower mass to charge ratio than an abundant ion. Space charge effects are also partially responsible for the longer linear portion of the high pressure curve. Larger ion clouds will require stronger trapping fields (higher qz) to overcome the space charge effects and fully compress the ion cloud. At both pressures, there is a noticeable suppression of the fluorescence signal for qz values of 0.7 and 0.75. This decrease in signal is the result of a decrease in ion Concentration. The ion concentration is lowered at these qz values because trapped ions are resonantly excited at f3z = 0.5. This resonance is attributed to higher order fields caused by truncating and/or stretching electrodes. Normally the f3z = 0.5 resonance does not affect ion trap performance because ion numbers are kept low and trapping times used for mass spectrometric analysis are much less than 1 s. At high ion concen-trations the cloud size increases and the ions are subjected to stronger fields farther away from the centre of the trap. This "dip" in the fluorescence signal is more pronounced at higher pressures because the ion cloud is larger and a greater number of ions is affected by the fjz = 0.5 resonance. Chapter 4. Pulse Counting Experiments 135 At very high qz the fluorescence signal drops quickly. It turns out that the first few fragment ions created by photodissociation are also fluorescent. Since they are of lower mass, these fragments have higher qz values than the precursor ion. As the qz value of the intact R6G ions approaches the stability limit defined by (3Z = 1 and qz = 0.908, the trajectories of the lower mass fragment ions are no longer stable. The sharp decrease of the fluorescence signal occurs because the fragment ions are no longer contributing to the fluorescence signal since they are ejected from the trap as soon as they are formed. 4.9 Fluorescence From Fragments To prove that a particular photodissociation product is fluorescent, it must be isolated in the ion trap. Ion isolation is achieved using a filtered noise field (FNF, Section 2.2.3). After all other ions have been removed from the ion trap, a standard pulse counting experiment can be performed. First, the precursor R6G ions are irradiated until enough of the desired fragment is formed. A FNF is then applied to eject all other ions. Finally, the counter on the data acquisition board is gated on and fluorescence data is recorded. CID Chapter 4. Pulse Counting Experiments 136 was not used to generate fragments because some photodissociation products are not observed in the CID spectrum. The mass scan for this experiment is shown in Figure 4.12. After ion injection, the RF-voltage is increased so that the R6G ions are held at a trapping level corresponding to qz = 0.65. The laser power is held constant for the entire experiment and depends on which fragment is being analyzed. After 1 s of irradiation a sufficient amount of the desired fragment is pro-duced. Since the application of the FNF requires 100 ms, photodissociation must be stopped while the FNF is applied. Gating the laser off for the ion isolation is one method to stop photodissociation. In this experiment photodissociation is stopped by lowering the trapping potential. The FNF was applied at this lower potential so that ion isolated could be achieved without photodissociation also occurring. After the ion isolation step is complete, the trapping level is increased so that the fragment of interest is held at qz = 0.65 for 1 s. The photon counting electronics are gated on for this time period and any fluorescence counts can be attributed to the fragment. The first few fragments of R6G corresponding to loss of CH2, 2 x CH2, 3 x CH2, and OC2H4 are fluorescent. Chapter 4. Pulse Counting Experiments 137 Photon Counting 1 > Time Figure 4.12: Timing diagram to test fragment ions for fluorescence. 4.10 Chopping Experiments The idea of chopping the excitation laser was first presented in Section 3.4.2. Chopping the laser beam allows the ions more time to cool and higher laser powers can therefore be used. While chopping certainly permitted the use of higher laser powers, the experiment was unsuccessful when first attempted. The chopping experiment was revisited and reevaluated after fluorescence was detected using pulse counting experiments. The total fluorescence emission Chapter 4. Pulse Counting Experiments 138 collected for a given counting period was reduced by 66% by chopping the laser. The chopping wheel used for these experiments had a 50% duty cycle. Chopping the laser beam would only be useful for increasing the S/N ratio during the "on" phase of the chopping. If the fluorescence signal is already sufficiently above the background noise then a chopping experiment should not be used. 4.11 Conclusions The main objective of building an instrument capable of detecting gas phase fluorescence emission from molecular ions has been achieved. An instrument has been developed to detect fluorescence emission from gas phase molecular ions confined in an ion trap. This is the first such instrument capable of mak-ing high S/N measurements of the fluorescence emission from large molecular ions. A S/N of 260:1 was measured for a 15 s experiment and larger S/N measurements could easily be achieved by increasing the counting period. The fluorescence experiment is optimized by using compact ion clouds and lower laser powers. A trapping potential corresponding to qz = 0.65 gives the optimal fluorescence signal for R6G ions (2000-3000 CPS). Many significant Chapter 4. Pulse Counting Experiments 139 instrumental innovations were required to achieve this goal. The ESI front-end allowed ions to be trapped without also filling the trap-ping volume with neutrals. A large, noisy neutral background and emission from filaments were the main disadvantages of using a thermal desorption probe with the first version of the instrument. The most significant instrumental challenge was to reduce the amount of laser generated background. This background problem was not anticipated at the beginning of this work because the fluorescence emission was expected to be significantly Stokes shifted. The sensitivity of gas phase molecules to photodissociation also required the use of lower laser powers, thus decreas-ing the total amount of fluorescence that could be expected. The only way to make this fluorescence measurement was to significantly reduce the total amount of background light detected. The use of a fiber optic probe and the coating of the electrodes with carbon reduced the amount of background scatter collected by at least six orders of magnitude. This background reduc-tion led to the successful detection of fluorescence from trapped gas phase molecular ions. At the beginning of this chapter it was stated that separating the fluores-cence emission of trapped ions from the laser generated background might not Chapter 4. Pulse Counting Experiments 140 be possible with a spectrograph. Wavelength dispersion was thought to be inadequate for separating the small fluorescence signal from the overwhelm-ing laser background. Fluorescence was expected to be almost resonant with the excitation and wavelength dispersion would therefore not help. The dra-matic reduction of background noise and the fact that a filter could be used to separate some of the fluorescence from the laser background prompted a re-evaluation of the wavelength dispersion experiment. The next experimental goal and supplementary objective of this thesis was to collect a fluorescence emission spectrum of a gas phase molecular ion. An intensified CCD (ICCD) was provided for this purpose. The ICCD is an ideal detector for this ion trap fluorescence experiment. 4.12 Perspective While many laboratories around the world have abandoned their attempts to detect fluorescence emission from trapped molecular ions, other groups are making progress. To put the instrumental accomplishments described in this chapter into perspective it is important to consider the progress made by these groups. Chapter 4. Pulse Counting Experiments 141 Joel Parks' group at the Rowland Institute (Harvard) has detected emis-sion from trapped ionic dyes and has now moved on to gas phase F R E T ex-periments.16 Parks was able to detect an average of 1 photon for each pulse of a Nd:YAG laser. Two dielectric high reflectors, designed to attenuate the laser wavelength by « 99%, and two long pass filters were used to reduce the background detected. Relying on several filters to reduce the background limits the excitation wavelengths that can be used and attenuates the fluo-rescence. Collecting a full emission spectrum is likely impossible when long pass filters are used. Also, a fluorescence excitation spectrum could only be collected if filters were purchased for every excitation wavelength measured. Another group actively developing instrumentation in this field is Alan Marshall's group at the National High Magnetic Field Laboratory in Florida. Marshall's group has been able to detect fluorescence emission of CeF6+ con-fined in an ICR trap. 1 5 , 4 9 Pulsed excitation (10 Hz) was also used for these experiments. C 6 F 6 + has an excited state lifetime that is long enough to allow a detector to be turned on after the excitation from the pulsed laser has passed. They were able to collect close to 2 photons per pulse. Also, since no filters were required to block the excitation wavelength, an excitation spectrum was recorded. For ions with sufficiently long fluorescent lifetimes a Chapter 4. Pulse Counting Experiments 142 pulse laser is a reasonable excitation source. 143 Chapter 5 Fluorescence Emission Spectroscopy This chapter describes the first instrument capable of recording the gas phase emission spectrum of large molecular ions. The objective of the work described in this thesis was to develop an instrument capable of detect-ing fluorescence emission from large gas phase molecular ions confined in a quadrupole ion trap. This was demonstrated in Chapter 4. The possi-ble future applications of the instrument included: probing the gas phase structure of biomolecules, identifying potential diffuse interstellar band car-riers, and complementing other forms of molecular spectroscopy. Experi-ments based on the pulse counting detection methods and instrumentation presented in Chapter 4 demonstrate that the instrument can potentially be used for these applications. However, these pulse counting experiments only Chapter 5. Fluorescence Emission Spectroscopy 144 record the quantity of fluorescence emitted by the trapped ions. The wave-length component of the emission is only roughly estimated by considering the transmittance of the filter used in the experiment. The ability to record a complete emission spectrum of trapped ions would significantly enhance the versatility and effectiveness of the instrument. A fluorescence emission spectrum is recorded by spatial dispersion of the emis-sion and recording of the fluorescence intensity as a function of wavelength. The spectroscopic information contained in an emission spectrum could be used help elucidate the structure of gas phase ions. Also, the design of pulse counting experiments could be improved with knowledge of the emission spectrum of the ion under investigation. As well, it is likely that fluorescence experiments would be easier to opti-mize with prior knowledge of the emission spectrum of the ion to be studied. For example, a F R E T experiment (Section 1.1) requires that the emission of the donor molecule correspond in energy to the absorption of the acceptor chromophore. In Chapter 3 it was shown that the gas phase absorption of ions differs from their solution absorbance (Section 3.4.1). Also, if emission of the donor is significantly different in the gas phase it may not excite the acceptor at all. FRET pairs that work well in solution experiments may Chapter 5. Fluorescence Emission Spectroscopy 145 not be appropriate for gas phase investigation. An instrument capable of recording complete emission spectra of gas phase ions could be used to select suitable F R E T pairs and provide a proof of principle for all gas phase F R E T experiments. The results from the pulse counting experiments suggest that the in-strument might be capable of recording emission spectra because the laser background was significantly reduced. The fluorescence signal was two or-ders of magnitude higher than the laser generated background. However, the pulse counting experiments relied on a long pass filter to reduce the total amount of laser generated background detected. Earlier in this thesis it was hypothesized that fluorescence emission is very close in wavelength to the laser excitation, so a long pass filter may not be the best choice for fluo-rescence emission spectroscopy. A long pass filter would likely attenuate a significant proportion of the emission. In order to evaluate the possibility of recording an emission spectrum of trapped ions, modifications to the instrument were necessary. The P M T used to detect photons in pulse counting experiments was replaced with a spectrograph. The same spectrograph used in previous experiments (Section 3.2) was coupled to the fiber optic probe. The CCD used in earlier work Chapter 5. Fluorescence Emission Spectroscopy 146 (Section 3.2) was replaced with an ICCD (Pi-Max 1024 x 256, Roper Sci-entific, Inc., Tucson, USA). Initially, a holographic notch filter was used to reduce the amount of laser scatter reaching the spectrograph. 5.1 Instrumentation Most of the instrumentation used for the pulse counting experiments was also used to collect emission spectra. The excitation optics and fiber optic probe were not changed. The methods and instrumentation used to confine the ions also remained the same. The only change from the pulse count-ing experiments was how the fluorescence was separated from the excitation background and detected. The pulse counting P M T and long pass filter were exchanged for an ICCD mounted on a spectrograph. Other filters, including the one used for pulse counting, could easily be added if required. Figure 5.1 is a diagram showing the final version of the instrument. 5.1.1 Spectrograph and ICCD The 0.150 m spectrograph has been previously described (Section 3.2). The ICCD uses a microchannel plate (MCP) to amplify light prior to CCD detec-Chapter 5. Fluorescence Emission Spectroscopy 147 Syringe Pump Figure 5.1: Final version of the instrument. Chapter 5. Fluorescence Emission Spectroscopy 148 tion. The ICCD has five main components: photocathode, MCP, phosphor screen, fiber optic bundle, and CCD. Light hitting the photocathode will gen-erate photoelectrons, which are accelerated towards the front of the MCP. A photoelectron then enters one channel of the MCP and becomes amplified via a cascade process. Amplified electron packets exiting the M C P strike the phosphor screen producing visible photons that are then coupled through the fiber optic bundle to the CCD where they are detected. By gating both the photocathode and the MCP this ICCD can achieve off/on ratios of 106 : 1. An ICCD can detect extremely low light levels that would otherwise be far below the read noise of a CCD. A gain setting of 200 was used for all emission experiments; according to manufacturer specifications this corresponds to an w 7800 times amplification of the signal. The CCD was cooled to — 20°C for all experiments. The ICCD was controlled with commercial software supplied by the manufacturer (WinSpec V2.4.7, Roper Scientific, Inc.). Spectra shown in this thesis are plotted with a vertical axis in counts. These counts correspond to the intensified signal and have not been corrected to represent the number of photon incident on the detector. The spectrograph was mounted on a kinematic platform which sits on a translation stage. After fluorescence emission was confirmed with a pulse Chapter 5. Fluorescence Emission Spectroscopy 149 Figure 5.2: Photograph of the instrument with the P M T attached to the fiber optic probe and spectrograph ready to be connected. counting experiment the P M T was removed from the fiber optic probe and the spectrograph was moved forward into place. The entrance flange of the spectrograph has been modified to couple directly to the fiber optic probe. The fine thread used in the tube lens coupling system aided in the alignment of the spectrograph. The translation stage allows the spectrograph to be moved on and off the probe without altering the alignment. Figure 5.2 is a photograph of the instrument with the P M T attached to the fiber optic probe and the pre-aligned spectrograph ready to be connected. The entrance slit width used on the spectrograph was 760 pm giving a Chapter 5. Fluorescence Emission Spectroscopy 150 resolution of « 14 nm at the detector. Room temperature fluorescence emis-sion from R6G and 5-CR 6G ions was expected to be diffuse and greater resolution unnecessary. Filters and Optics The 488.0 nm holographic notch filter (Chapter 4) was used as the output window of the fiber optic probe to reduce the scattered 488.0 nm excitation. The probe was designed so that only light exiting the fibers at small angles (< 10°) would reach the filter. Light exiting the fibers at greater angles would hit a beam stop built into the probe. To achieve optimum performance the light to be filtered must strike the notch filter with an angle of incidence of less than 10°. After the collected light passes through the notch filter it was focused onto the entrance slit of the spectrograph with a 25 mm diameter plano-convex lens with a 100 mm focal length. This lens has an f-number of 4 to match the spectrograph. Since the output of the fiber optic probe does not illuminate the entire lens the grating was underfilled. Chapter 5. Fluorescence Emission Spectroscopy 151 5.1.2 Methods The instrumental parameters shown in Table 4.4.1 were used to trap ions of the ionic dyes R6G and 5-CR 6G. The only instrumental parameter changed was the buffer gas pressure. In these experiments the buffer gas pressure was 1 mtorr. Each dye was investigated separately and held at a trapping level corresponding to qz = 0.65 for every experiment. Laser wavelength and laser power are specified for each experiment. The room lights were turned off for these experiments. The ion storage time was 1 s and the total time for the complete mass scan was 1.065 s. Since the duty cycle for the storage time was 94%, the ICCD was operated asynchronously with the mass spectrometer and the total integration time was dictated by the light collected. For the PAH experiments described in Chapter 3, the CCD was usually gated on and off with every mass scan to avoid collecting light emitted by the filaments. Gating the detector on and off for each mass scan adds read noise to the measurement and should always be avoided if possible. Full vertical binning of the ICCD was used to minimize read noise. The capacity of the readout pixel below every column of pixels on the ICCD is 1.2 x 106 electrons. To obtain the best possible S/N the integration time of the Chapter 5. Fluorescence Emission Spectroscopy 152 ICCD was set so that the highest signal on the detector would be just below the maximum capacity of the read pixel. The relationship between total counts and S/N is basically the same as for the pulse counting experiments. The S/N will increase with the square root of the total number of counts. It is important not to saturate the pixels or charge will bloom into adjacent pixels. 5.1.3 Solution Spectra R6G and 5-CR 6G are very similar dye molecules which have high fluores-cent quantum efficiencies in solution (4>f > 0.9). The structures of the dyes are shown in Figure 3.7 and their solution absorption spectra are shown in Figure 5.3. The solution absorption spectra were recorded using a Hewlett Packard (Palo Alto, USA) 8452A diode array UV-Vis spectrophotometer. Emission spectra for R6G and 5-CR 6G were recorded using the same fiber optic probe, spectrograph, and ICCD used with the ion trap experiments. A quartz cuvette filled with a 1/iM solution of the dye was positioned approxi-mately 25 cm from the end of the fiber optic probe. The 488.0 nm output of the argon ion laser was directed through the cuvette and the emission spec-trum was recorded. The solution emission spectra are shown in Figure 5.4. Chapter 5. Fluorescence Emission Spectroscopy 153 1.8 -1.6 -1.4 -o 1.2 -c ro 1.0 -.a i _ OS 0.8 --Q < 0.6 -0.4 -0.2 -0.0 -100 200 300 400 500 600 700 Wavelength (nm) 800 900 Figure 5.3: Solution absorbance spectra of R6G and 5-CR 6G. 36000 H C/D -»—' c O 32000 O c 3 c — 28000 max 561 nm R6G 5-CR 6G 400 450 500 550 600 Wavelength (nm) Figure 5.4: Solution emission spectra of R6G and 5-CR 6G. Fluorescence maximum is Stokes shifted 73 nm. Chapter 5. Fluorescence Emission Spectroscopy 154 The emission spectra of R6G and 5-CR 6G are very similar. The emission maximum for both dyes is Stokes shifted 73 nm from the 488.0 nm excita-tion. The long wavelength side of the solution emission shows a sharp cutoff which corresponds to the edge of the ICCD. R6G and 5-CR 6G are expected to also have very similar gas phase emission spectra. 5.2 Emission Spectra with Notch Filter The first emission spectrum collected was for R6G. R6G ions were excited with 66 mW of 488.0 nm laser light. The emission spectrum of R6G was recorded in three 60 minute integrations. One 60 minute background spec-trum was recorded with no ions in the trap. The only change made in the experiment to collect the background spectrum was to disconnect the gate pulse and therefore prevent ions from reaching the ion trap. The full 180 minute fluorescence emission spectrum was obtained by adding the three 60 minute integrations and subtracting the background spectrum three times. Since the laser power was chosen to maximize the fluorescence emission, some photodissociation was occurring during this experiment. The recorded emis-sion spectrum shown in Figure 5.5 must contain emission from the fragments Chapter 5. Fluorescence Emission Spectroscopy 155 in addition to emission from the R6G ion. Lower laser powers should be used when collecting emission spectra if any of the photodissociation fragments have a different emission spectrum from the precursor ion. The emission spectrum of R6G was collected again at a lower laser power. The laser was set to 32 mW so that almost no photodissociation would occur during the experiment. As before, three 60 minute integrations and one 60 minute background spectrum were recorded. The background subtracted sum of the three integration times is shown in Figure 5.5. The emission spectrum of 5-CR 6G ions was also recorded. The total integration time was 105 minutes for 5-CR 6G. The background subtracted emission spectrum is shown in Figure 5.5. A laser power of 66 mW was used to excite the 5-CR 6G and some photodissociation occurred during this experiment. The emission spectra shown in Figure 5.5 were smoothed using a five point moving average. 5.2.1 Discussion The two most obvious characteristics of the emission spectra shown in Fig-ure 5.5 are that the fluorescence emission is very close to the excitation Chapter 5. Fluorescence Emission Spectroscopy 156 I ' 1 ' 1 1 1 ' 1 ' 1 > 1 4 0 , 0 0 0 - i 350 400 450 500 550 600 650 Wavelength (nm) Figure 5.5: Emission spectra collected with the notch filter. Chapter 5. Fluorescence Emission Spectroscopy 157 wavelength and that there is a gap in the spectrum between the excitation wavelength and the fluorescence. The holographic notch filter was supposed to provide a 10 nm notch centered at 488.0 nm. The actual transmission of the notch filter was measured with a Hewlett Packard 8452A diode array UV-Vis spectrophotometer. The transmission of the notch filter was compared with the R6G emission spectrum. The results of this comparison are shown in Figure 5.6. The filter is not a 488.0 nm notch filter but a 496.5 nm notch filter which does little to attenuate the 488.0 nm excitation used for these experiments. The filter manufacturer (Kaiser Op-tical Systems, Inc., Ann Arbor, USA) had inadvertently shipped the wrong filter. While disappointing there was a positive aspect to discovering the incorrect filter had been used. The use of an expensive notch filter was un-necessary. A fluorescence excitation spectrum, for example, could now be measured without requiring a separate notch filter for every excitation wave-length measured. Experiments without the notch filter will be 20% more sensitive because the transmission of a notch filter is less than 80% outside the notch. Without the need for notch filters the instrument is more versatile, sensitive, and less expensive. The use of an inappropriate notch filter interfered with the collection of Chapter 5. Fluorescence Emission Spectroscopy 158 Figure 5.6: R6G emission (left axis) and notch filter transmittance (right axis). Chapter 5. Fluorescence Emission Spectroscopy 159 the fluorescence emission spectra of R6G and 5-CR 6G. The edge of the notch filter was almost certainly defining the shape of the emission spectrum and experiments without this notch filter were necessary. Despite this, some use-ful information was extracted from these first emission spectra. Comparing the emission spectra collected for R6G with and without photodissociation and comparing the emission of R6G and 5-CR 6G is still possible. The elastically scattered laser light appears in most of the emission spec-tra shown in this thesis. This collected laser scatter is « 200 times larger than the maximum fluorescence signal. This laser scatter appears in the background subtracted emission spectra because there is more noise associ-ated with subtracting two large signals. Drifting of the laser power during long experiments also contributed to this effect. The normalized emission spectra of R6G with and without photodissoci-ation are plotted together to aid in their comparison (Figure 5.7). At this resolution they appear to be identical and photodissociation will not have to be avoided in future experiments unless the information being sought is lost by fragmentation (e.g. FRET experiments). The fluorescence emission spectra of R6G and 5-CR 6G are plotted to-gether in Figure 5.8. The 5-CR 6G emission spectrum is broader and its Chapter 5. Fluorescence Emission Spectroscopy 160 120,000 100,000 80,000 0 60,000 "§ 40,000-| j= 20,000 -| R6G • R 6 G (no frag) 350 ' 400 ' 450 ' 500 ' 550 ' 600 ' 650 Wavelength (nm) Figure 5.7: R6G emission spectra with and without pho-todissociation. 120,000 100,000 80,000 §5 60,000 •D 40,000 (0 N ^ 20,000 0 4 350 400 450 500 550 600 650 Wavelength (nm) Figure 5.8: R6G and 5-CR 6G emission spectra. Chapter 5. Fluorescence Emission Spectroscopy 161 emission maximum has a greater Stokes shift. One of the intended appli-cations of the instrument is to investigate the conformations of gas phase biomolecular ions trapped in an ion trap. It was encouraging that differences in the emission spectra of two very similar ions can be observed even at such low resolution. 5.3 Emission Spectra without Filter The inappropriate (and, in fact, unnecessary) notch filter was removed from the fiber optic probe and replaced with a quartz window. The emission spectra of trapped R6G and 5-CR 6G ions were collected using a laser power of 47 mW for both dyes. An 80 minute R6G ion emission spectrum was collected in eight 10 minute integrations with four 10 minute background integrations. A 90 minute 5-CR 6G ion emission spectrum was collected in six 15 minute integrations with two 15 minute backgrounds integrations. The emission spectra of R6G and 5-CR 6G are plotted together in Figure 5.9. Chapter 5. Fluorescence Emission Spectroscopy 162 350 400 450 500 550 600 Wavelength (nm) Figure 5.9: Fluorescence emission spectra for R6G (80 min integration time) and 5-CR 6G (90 min integration time). Chapter 5. Fluorescence Emission Spectroscopy 163 5.3.1 Discussion R6G has an emission maximum at 503 nm and 5-CR 6G has an emission maximum at 509 nm. As with the previous spectra of these two ions (Figure 5.8), the 5-CR 6G emission spectrum is more diffuse and has a greater Stokes shift than the R6G emission. The fluorescence emission from gas phase R6G ions is Stokes shifted 15 nm from the 488.0 nm excitation. The fluorescence emission from gas phase 5-CR 6G ions is Stokes shifted 21 nm from the 488.0 nm excitation. These Stokes shifts are small compared to the 73 nm Stokes shifts measured in solution (Figure 5.4). In condensed phase experiments molecules excited to a high vibrational level of the excited state quickly lose excess vibrational energy to the sur-rounding solvent or matrix. Any fluorescence will therefore originate from the lowest vibrational energy level of the excited state. Franck-Condon fac-tors are non-zero for a wide range of transition energies and the resulting emission is a diffuse band. After emission, some molecules will be in a high vibrational level of the ground state. The excess vibrational energy is again absorbed by the solvent and the molecules end up in the lowest vibrational energy level of the ground state. The Stokes shift observed in solution ex-periments is a consequence of vibrational relaxation of the excited state and Chapter 5. Fluorescence Emission Spectroscopy 164 solvent rearrangements. In the gas phase, no such relaxation of the excited state can occur be-tween absorption and emission events. As a result, smaller Stokes shifts are observed in gas phase emission spectra. These small Stokes shifts make detecting gas phase fluorescence from trapped molecular ions difficult. In solution measurements the fluorescence is Stokes shifted away from the ex-citation and emission is detected against an almost zero background. In gas phase experiments the fluorescence emission is very close in energy to the excitation. Separating the weak fluorescence signal of trapped ions from the large excitation generated background is therefore challenging. Figure 5.10 shows the fluorescence signal dominated by the background laser scatter. Figure 5.11 is the same data with an expanded vertical scale so the small fluorescence signal can be observed. The data shown in Figure 5.10 was used to generate the R6G background-subtracted fluorescence emission spectrum shown in Figure 5.9. Since the fluorescence emission spectrum of R6G ions has been collected it is possible to evaluate the effect of the long pass filter (CG-OG-530) used in the pulse counting experiments. The transmittance of the filter was mea-sured using a Hewlett Packard 8452A diode array UV-Vis spectrophotometer. Chapter 5. Fluorescence Emission Spectroscopy 165 4000000-, 35000001 j2 300000CM c 3 g 250000CH . ~ 20000004 CO c . £ 150000O-I 1000000-1 500000-0 --signal + background b a c k g r o u n d only 350 400 450 500 550 Wavelength (nm) 600 650 Fig ure 5.10: R6G fluorescence emission spectrum without background subtraction (black) and the cor-responding background spectrum (red). 30000Ch 28000O, c o o 26000CH >> 2400004 c CD 2200001 2000001 1800001 -signal + background| background only 350 400 450 500 550 Wavelength (nm) 600 650 Figure 5.11: R6G fluorescence emission spectrum without background subtraction (black) and the cor-responding background spectrum (red) with expanded vertical scale. Chapter 5. Fluorescence Emission Spectroscopy 166 Figure 5.12 shows the filter transmittance plotted with the R6G ion emis-sion spectrum. To calculate the attenuation the R6G emission spectrum was multiplied by the measured transmittance of the long pass filter. The result of this calculation is also shown in Figure 5.12. Using the long pass filter was necessary for pulse counting experiments to reduce the amount of background light detected, but the fluorescence emis-sion was attenuated as well. The long pass filter reduced the total amount of fluorescence by « 83%. If this long pass filter was used for the collection of the emission spectrum, the peak fluorescence would have been erroneously measured at 547 nm. Now that the emission spectrum has been recorded, a more effective long pass filter could be selected for pulse counting experi-ments. 5.4 Anti-Stokes Fluorescence Although the holographic notch filter was not required for the detection optics of the instrument, it was reinstalled into the fiber optic probe and evaluated using 496.5 nm excitation. Two 20 minute integrations and one 20 minute background were collected for R6G ions. The trapping level remained Chapter 5. Fluorescence Emission Spectroscopy 167 60000 40000 co C 30000 o > , 20000 CO c cu • R6G emission • Calculated Attenuation • Filter Transmittance 350 ' 400 ' 450 ' 500 ' 55T Wavelength (nm) 660 co o 'co c (D -t—* c 650 500 510 Wavelength (nm) Figure 5.12: R6G emission (left axis), transmittance of the filter used for pulse counting experiments (right axis) and the calculated at-tenuation of the emission spectrum by the filter. The long pass filter used for pulse counting experiments attenuates the fluo-rescence significantly. Chapter 5. Fluorescence Emission Spectroscopy 168 at qz = 0.65, the laser power was 47 mW, and the buffer gas pressure re-mained at 1 mtorr. The background subtracted R6G emission was smoothed using a five point moving average and the resulting emission spectrum is shown in Figure 5.13. Another instrumental change was required before 496.5 nm excitation could be used. The filter used to pass only the laser line and block all other argon ion plasma lines was removed. This filter was necessary in older versions of the instrument. Since the black carbon coating of the electrodes and the fiber optic probe reduced the total amount of scattered light collected this filter was no longer necessary. 5.4.1 Discussion The emission spectrum of R6G ions is shown in figure 5.13. Excitation at 496.5 nm produced a smaller Stokes shift of 9.5 nm compared to the 15 nm Stokes shift measured for 488.0 nm excitation. The emission maximum is now 506 nm. This maximum is directly adjacent to the long wavelength side of the notch filter's attenuation and the notch filter is likely obscuring part of the fluorescence emission. The fluorescence emission of molecules is usually easy to differentiate Chapter 5. Fluorescence Emission Spectroscopy 169 0 1 ' 1 1 1 ' 1 > 1 > 1 1 r 350 400 450 500 550 600 650 Wavelength (nm) Figure 5.13: R6G emission from 496.5 nm excitation (left axis) and the notch filter transmission (right axis). from Raman scattering. For sufficiently intense emission or long fluorescence lifetimes fluorescence can be temporally resolved from Raman scattering. In condensed phase experiments the fluorescence is often significantly Stokes shifted from the excitation and cannot be confused with Raman scattering (e.g. Figure 5.4). In these gas phase experiments the fluorescence emission is very close to Chapter 5. Fluorescence Emission Spectroscopy 170 the laser line and therefore involves transitions similar in energy to those ex-pected in a resonance Raman spectrum. The emission observed for R6G ions excited at 488.0 nm and 496.5 nm is consistent with fluorescence emission. Resonance Raman scattering would have produced an emission spectrum with identical Stokes shifts (in energy units, e.g. cm'1) at both excitation wavelengths. The observed Stokes shift in nanometers between excitation and emission was smaller for 496.5 nm excitation and therefore even smaller in energy units. The difference in the Stokes shifts observed for the two excitations is indicative of fluorescence. For comparison, the emission from 488.0 nm and 496.5 nm excitation were plotted on the same graph (Figure 5.14). Within the noise and resolu-tion limits of this measurement, it appears that the R6G emission is identical for excitation at both wavelengths . The R6G ion is excited from a low vi-brational energy level of the ground state into a higher vibration level of the excited state. The excited ion will keep this vibrational energy when it returns to the ground state by fluorescence emission. The fluorescence emis-sion wavelength therefore corresponds to the difference in energy of ground and excited states. The diffuseness of the fluorescence emission depends on Franck-Condon factors. Excitation of a gas phase molecule with 488.0 nm Chapter 5. Fluorescence Emission Spectroscopy 171 460 470 480 490 500 510 520 530 540 550 560 570 580 590 600 Wavelength (nm) Figure 5.14: R6G ion emission from 488.0 nm and 496.6 nm excitation (left axis). The notch filter transmittance shown for reference (right axis). and 496.5 nm radiation is expected to produce the same emission spectrum at this resolution. Anti-Stokes emission can be observed in the R6G emission shown in Fig-ure 5.13. Anti-Stokes emission can occur in the gas phase fluorescence spec-trum because the ions are not vibrationally cooled effectively in the excited or ground state. Non-radiative relaxation of the excited state will obviously Chapter 5. Fluorescence Emission Spectroscopy 172 lead to heating of the ion. However, a radiative transition will also cause heating of the ion. The Stokes shift observed in the gas phase emission is related to the amount of energy retained by the ion after one excitation and radiative emission. Heating of the ions will lead to photodissociation and anti-Stokes emission. Anti-Stokes emission occurs when a hot ground state ion absorbs a photon and then relaxes to a lower energy level of the ground state. 5.5 Conclusions This chapter described the first instrument capable of recording gas phase emission spectra of large molecular ions. The instrument developed was not only able to collect emission spectra of large gas phase molecular ions, but the background light rejection of the instrument was even better than expected. Emission spectra were collected without filtering the excitation or collected emission. Without the need for a different pair of filters for every excitation wavelength used, fluorescence excitation spectra of ions can be recorded using this instrument. Simultaneously recording absorption spectra (by monitoring photodissociation) and fluorescence emission spectra can provide a wealth of Chapter 5. Fluorescence Emission Spectroscopy 173 spectroscopic information. Detecting differences in the fluorescence emission spectra of R6G and 5-CR 6G ions was very encouraging. The ability to detect such small struc-tural differences supports the hypothesis that fluorescence emission spec-troscopy can be a sensitive probe of gas phase molecular ion structure. Fi-nally, knowledge of an ion's fluorescence emission spectrum would be very useful for selecting filters for pulse counting experiments. 5.6 Perspective The results described in Chapter 4 and Chapter 5 were first reported at the 51st Annual ASMS Conference on Mass Spectrometry and Allied Topics (June 8-12, 2003, Montreal, Canada). At the same conference Alan Mar-shall's group at the National High Magnetic Field Laboratory in Florida reported the collection of the emission spectrum of C§F§ confined in an ICR trap. CeFo~ has an excited state lifetime that is long enough to allow a detec-tor to be turned on after the excitation from a pulsed laser has passed. Their emission experiment required 80 hours of integration time. However, tempo-rally resolving the fluorescence from the background may limit this technique Chapter 5. Fluorescence Emission Spectroscopy 174 to small or simple molecules. Since larger molecules have shorter fluorescence lifetimes their emission spectra are very difficult to measure in this way. Since most of the intended applications for the instrument described in this thesis involve large molecules, temporal separation methods are not desirable. In-terestingly, this group also revealed plans for an instrument based on a linear quadrupole ion trap. 175 Chapter 6 Summary and Conclusion 6.1 Summary This thesis described the first instrument capable of high S/N gas phase flu-orescence measurements of large molecular ions. The instrument is also the first capable of recording the gas phase emission spectra of large gas phase molecular ions. This instrument was used with filters and a pulse counting P M T to detect fluorescence, or the fluorescence was dispersed with a spec-trograph and an emission spectrum recorded with an ICCD. The instrument is versatile, sensitive and robust. The two most significant experimental challenges associated with detect-ing the gas phase fluorescence of ions were discovered rather than anticipated. First, the susceptibility of gas phase ions to photodissociation was an unex-pected problem. Photodissociation limited the laser power which could be used. Second, the fluorescence emission of gas phase ions was not significantly Chapter 6. Summary and Conclusion 176 Stokes shifted from the excitation. The lack of Stokes shift meant that small fluorescence signals must be detected in the presence of background scatter generated by the excitation. The technical difficulties associated with such a measurement are similar to those encountered in condensed phase Raman experiments. The background scatter recorded with the initial version of the instrument 1.7 x 107 times larger than the fluorescence signal eventually detected. Initial experiments were carried out on PAH cations. Although fluo-rescence from these cations was not detected, the experiments outlined in Chapter 3 were still very useful. The most obvious result from the PAH ex-periments was that molecules are more vulnerable to photodissociation in the gas phase. The emission of neutral molecules also suggested that ion emis-sion may be significantly closer to the laser line than originally predicted. The lack of detectable emission indicated that the fluorescence was below the noise level, or not present at all. Changing the ion source to electrospray eliminated the neutral background problem, but the real challenge was to extract the fluorescence signal from the much larger laser background. The ESI source permitted the investigation of molecules that were not accessible with the thermal desorption probe (Section 2.4.1). Initial exper-Chapter 6. Summary and Conclusion 177 iments were carried out on R6G and 5-CR 6G cations. Unfortunately, the fluorescence emission from these ions was still buried in the background noise and was undetectable. Photodissociation experiments with R6G and 5-CR 6G did however lead to a improved understanding of the gas phase pho-tophysics of trapped ions. For fluorescence to be detected, the ion optics of the ESI front-end needed improvement and major changes to the optical collection system were necessary. The innovative changes to the instrument reduced the background by six orders of magnitude. The internal beam dump was an addition that reduced some background noise; however, the carbon coating of the electrodes and the development of the fiber optic probe were the key elements introduced to reduce the background. As a result, background noise was reduced without limiting the versatility of the instrument. For fluorescence experiments, the 3D ion trap is an excellent device for confining ions. Optically, the ion volume is well shielded from the rest of the instrument. As a result, much of the unwanted light scattered off of the vacuum manifold does not reach the inside of the trap. W i t h inside surfaces of the electrodes coated black the ion trap makes a good "ion cuvette". The pulse counting experiments described in Chapter 4 were used to Chapter 6. Summary and Conclusion 178 optimize fluorescence detection. The greatest fluorescence signals were mea-sured for compact ion clouds and lower laser powers. A trapping potential corresponding to qz = 0.65 gave the best fluorescence signal for R6G ions. With the fluorescence experiment optimized, the emission spectra of R6G and 5-CR 6G ions were collected. It was possible to collect the emission spectra discussed in Chapter 5 for two reasons. First, the background was reduced 1.3 x 106 times with the instrumental innovations detailed in Chap-ter 4. Second, the signal was amplified « 7800 times using an ICCD. Finally, anti-Stokes fluorescence was detected from R6G excited at 496.5 nm. 6.2 Conclusion In condensed phase experiments molecules excited to a high vibrational level of the excited state quickly lose excess vibrational energy to the surrounding solvent or matrix. Any fluorescence will therefore originate from the lowest vibrational energy level of the excited state. Franck-Condon factors are non-zero for a wide range of transition energies and the resulting emission is a diffuse band. After emission, some molecules will be in a high vibrational level of the ground state. The excess vibrational energy is again absorbed by Chapter 6. Summary and Conclusion 179 the solvent and the molecules end up in the lowest vibrational energy level of the ground state. Since all fluorescence emission in the condensed phase originates from the lowest vibrational level of the excited state, excitation at different wave-lengths produces the same fluorescence emission spectrum. The Stokes shift observed in solution experiments is a consequence of vibrational relaxation of the excited state and solvent rearrangements. In the gas phase, no such relaxation of the excited state can occur between absorption and emission events. As a result, smaller Stokes shifts are observed in gas phase emission spectra. Furthermore, without a solvent or matrix to remove excess vibra-tional energy, gas phase ions remain vibrationally hot after emission occurs. To detect fluorescence from gas phase ions confined in an ion trap an instrument must be capable of separating the emission from the excitation. For atomic ions this is very easy. Figure 6.1 shows two different excitation schemes that are used to detect atomic fluorescence. Multiphoton absorption will produce an emission that is at least half the wavelength of the excita-tion. Alternatively, emission from a lower energy transition of the atomic ion can be monitored. In both cases the excitation and emission transitions are well separated in wavelength; thus fluorescence is detected against a zero Chapter 6. Summary and Conclusion 180 Molecules o CO 00 "3-L O Solution 3 - * 2 1 0 o C O 00 8 2 L O 92 s S L O CO CD V Gas Phase Figure 6.1: Energy level diagrams for fluorescence experiments. Chapter 6. Summary and Conclusion 181 background. Furthermore, the excitation laser power is not limited by pho-todissociation. Fluorescence from atomic and small molecular ions can also be detected using a pulsed laser source. These ions typically have long fluorescence life-times and their emission can be detected after an excitation pulse. For larger molecules excited state lifetimes are usually 1 to 3 ns and their fluorescence must be detected in the presence of the excitation background. The limita-tion in this case is not the length of the laser pulse, but the time required to gate a detector. The best intensifiers currently require several nanoseconds to turn on. Therefore, in the time it takes to gate on the detector the signal will have almost completely decayed. For a large gas phase molecular ion, excitation places the ion in a high vibrational level of the excited state. The ion cannot lose energy before emis-sion occurs so it retains this vibrational energy when it emits a photon. A 7T-7T* transition is not expected to induce a significant distortion in a large dye molecule and the Franck-Condon factors will therefore be greatest for transitions to the same vibrational level of the ground state. Figure 6.1 illus-trates why excitation of R6G at 488.0 nm and 496.5 nm produces the same emission spectrum. The emission is mainly dependent on the difference in Chapter 6. Summary and Conclusion 182 energy between the ground state and the excited state. Anti-Stokes fluo-rescence can be detected when a vibrationally hot ion in the ground state absorbs a photon and then relaxes to a lower vibrational level of the ground state. Since the fluorescence is not Stokes shifted far from the excitation wave-length the fluorescent measurement is no longer made against a zero back-ground. Without a zero background, the main sensitivity advantage of fluo-rescence spectroscopy is lost. However, difficulties associated with measuring the fluorescence emission of trapped ions do not detract from the desired ben-efits of such a measurement. 6.3 Future Directions 6.3.1 Improvements to the Instrument A tunable laser for excitation would be very useful for molecular spectroscopic investigations and for optimizing the fluorescence experiment. It would allow the fluorescence excitation spectrum of an ion to be collected. A tunable laser would also allow the optimization of the fluorescence experiment with respect to excitation energy. The optimal excitation wavelength will likely be a trade Chapter 6. Summary and Conclusion 183 off between increasing the molar absorptivity and decreasing the Stokes shift. A spectrograph with improved resolving power would also enhance this instrument. The spectrograph initially acquired for this instrument has a 600 grooves/mm grating. Since the fluorescence emission is not significantly Stokes shifted from the excitation a grating with a higher groove density (e.g. 1200 grooves/mm) would be more effective in separating the fluores-cence emission from the excitation at the detector. A greater physical sepa-ration at the detector will decrease the amount of stray light contaminating the emission signal. The original grating would still be useful for solution measurements where greater spectral coverage is required at the focal plane. More modest gains in instrument performance could be realized by in-creasing the size of the fiber optic probe and using a buffer gas that would more efficiently cool the ions. Increasing the number of fibers will improve collection efficiency. However, this must be accomplished without degrading the capabilities of the ion trap as a mass spectrometer. Sulphur hexafluoride (SFQ) should be evaluated as a collision gas during a fluorescence experiment. SF$ will likely be more effective at cooling ions during collisions and allow the use of higher laser powers. For an improved understanding of the experiment pulse counting tech-Chapter 6. Summary and Conclusion 184 niques should be applied to the ion trap detector. Currently, the ion trap uses analog detection but pulse counting could help estimate the total num-ber of ions in the trap. The maximum capacity for an ion trap similar to the one used in this work is thought to be approximately one million ions. Pulse counting the output of the ion trap during a mass scan would help refine this approximation. This would be a first step towards quantifying gas phase molar absorptivities and fluorescent quantum efficiencies. 6 . 3 . 2 A p p l i c a t i o n s Future work will involve applications of the instrument. Gas phase structure of proteins could be investigated using fluorescent tags or by detecting native fluorescence from tryptophan or green fluorescent protein (GFP). GFP has been cloned and purified for this purpose. Fluorescence emission detection could also be used to complement other molecular spectroscopies. PAH cations are potential DIB carriers that could be evaluated using fluorescence spectroscopy. Photodissociation is commonly used to measure the gas phase absorption of PAH ions. However, if the fluorescence quantum efficiency of particular transition is very high it may appear to have a lower absorption in a photodissociation experiment. Fluorescence experiments may Chapter 6. Summary and Conclusion 185 help refine absorption measurements based solely on photodissociation. In addition, F R E T experiments may help determine the structures of gas phase proteins. FRET experiments rely on the emission of one molecule to be resonant with the absorption of the second molecule. F R E T pairs that work well in solution may not be appropriate for gas phase experiments. As demonstrated in this work, the absorption and fluorescence emission of a molecule is very different in the gas phase. Methods and instrumentation developed in this thesis could be used to select suitable molecules for gas phase F R E T experiments. 6.3.3 Concluding Remarks The instrumental innovations described in this thesis permitted high S/N fluorescence detection and fluorescence emission spectroscopy of large gas phase molecular ions. Others wishing to investigate the fluorescence emis-sion of trapped molecular ions will be able to duplicate these innovations in a straightforward manner. Detecting fluorescence emission from large gas phase ions was an experimental problem that had frustrated many groups over the past twenty or thirty years. With the knowledge gained during the development of the instrument described this problem has now been solved. 186 Bibliography [1] L. Stryer and R.P. Haugland. Energy transfer: A spectroscopic ruler. Proc. Natl. Acad. Sci. U.S.A., 58(2):719-726, 1967. [2] Paul R. Selvin. The renaissance of fluorescence energy transfer. Nature Struct. Biol., 7(9):730-734, 2000. [3] Guo-Zhong Li , Bryan A. Vining, Shenheng Guan, and Alan G. Marshall. Laser-induced fluorescence of B a + ions trapped and mass-selected in a fourier transform ion cyclotron resonance mass spectrometer. Rapid Communications in Mass Spectrometry, 10:1850-1854, 1996. [4] C. Moutou, L. Verstraete, P. Brechignac, S. Piccirillo, and A. Leger. Laser induced fluorescence spectroscopy of jet-cooled molecular species:a tool to identify diffuse interstellar band carriers. Astronomy and Astro-physics, 319:331-339, 1997. Bibliography 187 [5] D. Romanini, L. Biennier, F. Salama, A. Kachanov, L. J. Allamandola, and F. Stoeckel. Jet-discharge cavity ring-down spectroscopy of ionized polycyclic aromatic hydrocarbons: progress in testing the pah hypothe-sis for the diffuse interstellar band problem. Chemical Physical Letters, 303(1-2):165-170, 1999. [6] J.J. Thomson. Rays of Positive Electricity. Longmans Green and Co., London, 1913. [7] W Paul and H Steinwedel. Apparatus for separating charged particles of different specific charges. German Patent, 944,900, 1956. [8] W. Neuhauser, M . Hohenstatt, P.E. Toschek, and H. Dehmelt. Localized visible B a + mono-ion oscillator. Phys. Rev. A, 22(3):1137, 1980. [9] Fred J. Grieman, Bruce H. Mahan, and Anthony O'Keefe. The laser induced fluorescence spectrum of trapped BrCN cations. J. Chem. Phys., 74(2):857-861, 1981. [10] Cecilia C. Martner, Jurgen Pfaff, Neil H. Rosenbaum, Anthony O'Keefe, and Richard J. Saykally. Radiative lifetimes of trapped molecular ions:HCl+ and H B r + . J. Chem. Phys., 78(12):7073-7076, 1983. Bibliography 188 [11] Alexander Baldwin Young, Raymond Evans March, and Richard James Hughes. Studies of infrared multiphoton dissociation rates of protonated aliphatic alcohol dimers. Can. J. Chem., 63:2324-2331, 1985. [12] F. Diedrick, J.C. Bergquist, Wayne M . Itano, and D.J. Wineland. Laser cooling to the zero-point energy of motion. Physical Review Letters, 62 (4):403-406, 1989. [13] R.C. Thompson. Spectroscopy and quantum optics with trapped ions. Comments At. Mol. Phys., 27(6):349-364, 1992. [14] D.J. Wineland, Wayne M . Itano, and R.S. Jr. Van Dyck. High-resolution spectroscopy of stored ions. In Advances in Atomic and molecular Physics, volume 19, pages 135-186. Academic Press, New York, 1983. [15] Yang Wang, Christopher L. Hendrickson, and Alan G. Marshall. Di-rect optical spectroscopy of gas-phase molecular ions trapped and mass-selected by ion cyclotron resonance: laser -induced fluorescence excita-tion spectrum of hexafluorobenzene (CQF^ ). Chemical Physical Letters, 334:69-75, 2001. Bibliography 189 [16] Joseph T. Khoury, Sandra E. Rodriguez-Cruz, and Joel H. Parks. Pulsed fluorescence measurements of trapped molecular ions with zero back-ground detection. J. Am. Soc. Mass Spectrom., 13:696-708, 2002. [17] Michael D. Barnes, William B. Whitten, and Micheal J. Ramsey. De-tecting single molecules in liquids. Analytical Chemistry, 67(13):418A, 1995. [18] Raymond E. March and John F.J. Todd, editors. Practical Aspects of Ion Trap Mass Spectrometry: Fundamentals of Ion Trap Mass Spectrometry, volume I. CRC Press, New York, 1995. [19] Raymond E. March and John F.J . Todd, editors. Practical Aspects of Ion Trap Mass Spectrometry: Ion Trap Instrumentation, volume II. CRC Press, New York, 1995. [20] Raymond E. March. Quadrupole ion trap mass spectrometry: a view at the turn of the century. International Journal of Mass Spectrometry, 200:285-312, 2000. [21] M . Nappi, C. Weil, C D . Cleven, L.A. Horn, H. Wollnik, and R.G. Cooks. Visual representations of simulated three-dimensional ion trajectories in Bibliography 190 an ion trap mass spectrometer. International Journal of Mass Spectrom-etry and Ion Processes, 161:77-85, 1997. [22] E. Mathieu. Memoire sur le mouvement vibratoire d'une membrane de forme elliptique. J. Math. Pures Appl, 13:137, 1868. [23] Raymond E. March. An introduction to quadrupole ion trap mass spec-trometry. Journal of Mass Spectrometry, 32:351-369, 1997. [24] J.E.R Syka, J.N. Louris, R E . Kelly, G.C. Jr. Stafford, and W. Reynolds. Method of operating an ion trap. U.S. Patent, 738,018, 1985. [25] Raymond E. Kaiser, John N. Jr. Louris, Jonathan W. Amy, and R. Gra-ham Cooks. Extending the mass range of the quadrupole ion trap using axial modulation. Rapid Communications in Mass Spectrometry, 3(7): 225-229, 1989. [26] J. Louris. Presentation WOB 10:45. In 40th ASMS, Washington, D.C., 1992. [27] P.E. Kelley, G.C. Jr. Stafford, and D.R. Stephens. Method of mass ana-lyzing a sample by use of a quadrupole ion trap. U.S. Patent, 4,540,884, 1985. Bibliography 191 [28] Scott Gronert. Estimation of effective ion temperatures in a quadrupole ion trap. J. Am. Soc. Mass Spectrom., 9:845-848, 1998. [29] D.J. Douglas and J.B. French. Collisional focusing effects in radio fre-quency quadrupoles. J. Am. Soc. Mass Spectrom., 3:398-408, 1992. [30] Paul Kebarle and Liang Tang. From ions in solution to ions in the gas phase, the mechanism of electrospray mass spectrometry. Analytical Chemistry, 65(22):972A-986A, 1993. [31] Terri L. Constantopoulos, George S. Jackson, and Christie G. Enke. Challenges in achieving a fundamental model for ESI. Analytica Chimica Acta, 406:37-52, 2000. [32] Scott T. Quarmby and Richard A. Yost. Fundamental studies of ion injection and trapping of elctrosprayed ions on a quadrupole ion trap. International Journal of Mass Spectrometry, 187:1-23, 1999. [33] Richard B. Cole, editor. Electrospray Ionization Mass Spectrometry. Fundamentals, Instrumentation & Applications. John Wiley & Sons, Inc., Toronto, 1997. Bibliography 192 [34] R.P.A. Hartman, D.J. Brunner, D.M.A. Camelot, J .C.M. Marijnissen, and B. Scarlett. Electrohydrodynamic atomization in the cone-jet mode physical modeling of the liquid cone and jet. J. Aerosol Sci., 30(7): 823-849, 1999. [35] A. Jaworek and A. Krupa. Classification of the modes of EHD spraying. J. Aerosol Sci., 30(7):873-893, 1999. [36] Th Dulcks and R. Juraschek. Electrospray as an ionization method for mass spectrometry. J. Aerosol Sci., 30(7):927-943, 1999. [37] Scott A. McLuckey, Gary J. Van Berkel, Douglas E. Goeringer, and Gary L. Glish. Ion trap mass spectrometry of externally generated ions. Analytical Chemistry, 66(13):689A-696A, 1994. [38] David Rendell. Fluorescence and Phosphorescence. Analytical Chem-istry by Open Learning. John Wiley & Sons, Toronto, 1987. [39] L.H. Kieth and W.A. Telliard. ES&T special report: Priority pollutants: I-a perspective view. Environ. Sci. TechnoL, 13(4):416-423, 1979. [40] Tadamasa Shida, editor. Electronic absorption spectra of radical ions, volume 34 of Physical sciences data. Elsevier, New York, 1988. Bibliography 193 [41] Robert E. Buckles, Arthur Serianz, and David Naffziger. Dibenzo(g.p)chrysene: A challenging experiment in organic syn-thesis. Proc. Iowa Acad. Sci, 80:45-49, 1973. [42] Steven C. Hill, Michael D. Barnes, Noah Lermer, William B. Whit-ten, and J. Michael Ramsey. Simulation of single-molecule photocount statistics in microdroplets. Analytical Chemistry, 70:2964-2971, 1998. [43] John P. Maier. Spectroscopic studies of open-shell organic cations in the gas phase. Chimia, 34(5):219-231, 1980. [44] C. L. Hunter, A. G. Mauk, and D. J. Douglas. Dissociation of heme from myoglobin and cytochrome b5: Comparison of behavior in solution and in the gas phase. Biochemistry, 36:1018, 1997. [45] Robert C. Dunbar, Jyh Horung Chen, Hun Young So, and Bruce Asamoto. Infrared fluorescence relaxation of photoexcited gas-phase ions by chopped-laser two-photon dissociation. J. Chem. Phys., 86(4): 2081-2086, 1987. [46] James D. Faulk and Robert C. Dunbar. Photodissociation spectroscopy of gas-phase ferrocene cation. J. Am. Soc. Mass Spectrom., 2:97-102, 1991. Bibliography 194 [47] Robert C. Dunbar. Photodissociation of trapped ions. International Journal of Mass Spectrometry, 200:571-589, 2000. [48] Melvin B. Comisarow and Alan G. Marshall. Fouier transform ion cy-clotron resonance spectroscopy. Chemical Physical Letters, 25(2):282-283, 1974. [49] B. Cage, M.A. McFarland, C L . Hendrickson, N.S. Dalai, and A.G. Mar-shall. Resolution of individual component fluorescence lifetimes from a mixture of trapped ions by laser-induced fluorescence/ion cyclotron res-onance. J. Chem. Phys. A., 106:10033-10036, 2002. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061168/manifest

Comment

Related Items