UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A new linear ion trap time of flight instrument with tandem mass spectrometry capabilities Campbell, Jennifer Mary 1999

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

Item Metadata

Download

Media
831-ubc_1999-463257.pdf [ 10.06MB ]
Metadata
JSON: 831-1.0061530.json
JSON-LD: 831-1.0061530-ld.json
RDF/XML (Pretty): 831-1.0061530-rdf.xml
RDF/JSON: 831-1.0061530-rdf.json
Turtle: 831-1.0061530-turtle.txt
N-Triples: 831-1.0061530-rdf-ntriples.txt
Original Record: 831-1.0061530-source.json
Full Text
831-1.0061530-fulltext.txt
Citation
831-1.0061530.ris

Full Text

A NEW LINEAR ION TRAP TIME OF FLIGHT INSTRUMENT WITH T A N D E M MASS SPECTROMETRY CAPABILITIES By Jennifer Mary Campbell B. Sc. (Chemistry) Queen's, 1991 M . Sc. (Chemistry) Waterloo, 1993 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R O F P H I L O S O P H Y in T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Department of Chemistry) We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A June 1999 © Jennifer Mary Campbell, 1999 in presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of CKe/ftt^-f/"' The University of British Columbia Vancouver, Canada Date X l w \(o DE-6 (2/88) Abstract This thesis summarizes the construction and characterization of a novel hy-brid mass spectrometer with tandem mass spectrometry capabilities, named the linear ion trap time of flight mass spectrometer. In a "linear ion trap" ions are trapped in a 2-dimensional quadrupolar potential by the application of timed stopping potentials on entrance and exit apertures. The perfor-mance characteristics of the linear ion trap as a storage device are initially assessed using a modified triple quadrupole mass spectrometer. On the time scales for tandem mass spectrometry, injection, extraction, and trapping ef-ficiencies are all near 100%. The modified operation of the triple quadrupole mass spectrometer is used to study the kinetics of dissociation of gas phase holomyoglobin in high charge states. The results indicate that the binding of the heme group is relatively unaffected by intramolecular repulsion resultant from excess charge. To construct the linear ion trap time of flight mass spectrometer, a linear ion trap is orthogonally coupled to a linear time of flight mass analyzer. The mass resolutions of the spectrometer could be optimized to attain resolutions near 700. Tandem mass spectrometry in the linear ion trap is enabled by su-perimposing a dipolar excitation voltage on the quadrupolar field by coupling an auxiliary waveform generator to a pair of the quadrupole rods. This volt-age is used to effect precursor isolation via a broadband waveform followed by collision induced dissociation through mass selective resonant excitation. The resulting fragment ions are detected in the time of flight mass spectrom-eter. With 7 mTorr N 2 as the collision gas, the resolutions of ion isolation and excitation are ~ 40 and 70, respectively. Fragmentation efficiency is near 60 %. When a lower pressure is used, the same resolutions increase to 100 and 250, respectively. It is found that the resolution of resonant excitation is strongly dependent upon amplitude of the applied voltage. •ii Contents Abstract ii Table of Contents iii List of Tables vi List of Figures vii List of Symbols xiii List of Abbreviations xvi Acknowledgements xvii Dedication xviii 1 Introduction 1 2 A Review of Instrumentation Similar to the LIT/TOFMS 6 2.1 Quadrupole Mass Spectrometers 6 2.2 Time of Flight Mass Spectrometers 16 2.3 The ESI Source Coupled to T O F M S 20 2.4 Gas Filled RF-only Quadruples in T O F M S 25 2.5 Tandem Mass Spectrometry with ESI-TOFMS Instruments . . 27 2.6 The L I T / T O F M S 29 2.7 Summary and Conclusions 32 3 The Linear Radio Frequency Quadrupole as a Storage Device 34 3.1 The 2-Dimensional Paul Trap 35 i i i 3.2 Experimental Methods 39 3.3 Collisional Energy Loss 46 3.4 Injection and Extraction of Ions with the Linear Ion Trap . . . 50 3.5 Trapping Studies 57 3.6 Summary and Conclusions 60 4 Ion Trapping Studies of the Heme Binding in Highly Charged HoloMyoglobin 61 4.1 Introduction 62 4.2 Experimental Methods 65 4.3 Calculation of the Internal Ion Temperature of Holomyoglobin 69 4.4 Binding Energies 74 4.5 Summary and Conclusions 79 5 The Linear Ion Trap TOF Mass Spectrometer 80 5.1 The Continuous Flow Time of Flight Mass Spectrometer . . . 81 5.2 Deflectors in oa-TOFMS 88 5.3 Performance 93 5.4 Effect of Trapping on Time of Flight Spectra 95 5.5 Summary and Conclusions 99 6 Resonant Excitation for MS/MS in the LIT/TOFMS 101 6.1 Resonant Excitation in the Linear Ion Trap 102 6.2 Experimental Methods 115 6.3 Tandem in Space Mass Spectrometry 124 6.4 Ion Isolation 126 6.5 Ion Fragmentation 130 6.6 Competition between Fragmentation and Ejection 133 6.7 Summary 142 7 Variation of Gas Pressure in the Linear Ion Trap 143 7.1 Introduction 144 7.2 Experimental Methods 146 7.3 Mass Resolution of Resonant Excitation 150 7.4 Isolation at Lower Pressures 164 7.5 M S 3 in an RF-Only Quadrupole 167 iv 7.6 Use of Helium as a Collision Partner 173 7.7 Summary 174 8 Summary, Conclusion, and Outlook 176 Bibliography 181 v List o f Tables 3.1 Operating voltages on the triple quadrupole mass spectrom-eter for the injection and trapping studies. Symbols for the components of the spectrometer are from Fig. 3.1. Voltages marked x/y represent the trapping/transmitting potentials of the pulsed entrance (Q0/Q1) and exit (Q2/Q3) aperture plates of the trap 44 4.1 Observed rates of holomyoglobin dissociation, kT(T), for the +9 to +12 charge states. The Ea and A values are from Gross et al.122 and the ranges of temperatures are calculated from Eq. 4.8 74 4.2 Observed rate constants of hMb loss and calculated activation energies (Ea) for dissociation of hMb ions 78 vi List of Figures 2.1 A cross section of a rod set of a quadrupole mass filter, showing the applied potential, 0 O and the equipotential lines of the resultant quadrupolar potential 8 2.2 The first stability region of the quadrupole mass filter. For a given Vrf and U, ions of different ^ line along the scan line. . 10 2.3 The geometry of the 3-dimensional ion trap 12 2.4 The first stability region for the 3-d ion trap 14 2.5 Schematic of the Wiley-McLaren time of flight mass spectrom-eter 18 2.6 Generalized schematic showing the geometry and components theoa-TOFMS. 22 2.7 The L I T / T O F M S , where 1, is the electrospray ionization source, 2 is an entrance aperture plate, 3 is an RF-only quadrupole, 4, is the exit aperture plate, and 5 and 6 are the source and drift regions respectively of the T O F M S . Pulses to 2 and 4 are used to control the flow of ions into and out of the RF-only quadrupole and to confine ions in the trapping volume. The T O F M S is as shown in Fig. 2.5 31 3.1 The triple quadrupole mass spectrometer system. S, elec-trospray ion source; O, ion sampling orifice; SK, skimmer; Q0, RF-only quadrupole; SRO, prefilter; Q l , mass analyzing quadrupole; Q2, RF-only quadrupole; SR2 prefilter; Q3 mass analyzing quadrupole; D, detector. The linear ion trap is cre-ated from the collision cell 41 vi i 3.2 Time profiles resulting from injection of ions for (a) 100 ms, and (b) 400 ms into a collision cell with 6 mTorr neon. In (b) the decrease in number of ion counts is caused by detector saturation 46 3.3 Time profiles of ions leaving Q2 resulting from injection of ions for 100 (.is, without trapping, into a collision cell of pressure (a) 0.015 mTorr and (b) 1.5 mTorr Ne, and from the trapping of ions in a collision cell with a pressure of 0.015 mTorr for (c) 1 ms and (d) 10 ms 49 3.4 Flight time profiles for injection times of 10, 20 and 40 ms, at collision cell pressures of 0.015 mTorr background, 3.0 and 6.0 mTorr Ne. The vertical axis of each panel represents the mea-sured number of extracted ions per channel normalized by the average ion current measured when no trapping is performed. Only the first 25 ms of the full 100 ms extraction time is shown. 52 3.5 Experimental verification of the high injection and extraction efficiencies of the linear ion trap on a time scale up to 1 s. The solid line labelled "flow" is the product of the ion current with transmitting potentials on the exit and entrance apertures and the injection time. There was 4.5 mTorr neon in the trap for the duration of the experiment 55 3.6 The difference in injection and extraction efficiencies for long injection times with 0.015 mTorr N2 and 4.5 mTorr neon in the linear ion trap 56 3.7 Trapping efficiency on times scales relevant to tandem mass spectrometry in the L I T / T O F M S . 58 3.8 Demonstration of trapping efficiency on a long time scale for the singly charged reserpine ion. Note the change in the ion count axis in the lower panel with the longer trapping time frame 59 4.1 Mass spectrum of myoglobin obtained with a continuous flow mixing apparatus combined with ESI. The orifice skimmer dif-ference was 30 V . Notation: h8 is hmb + 8 a9 is a M b + 9 67 4.2 Change in the relative populations of hMb +9 and aMb +8 as a function of trapping time 72 vii i 4.3 Calculation of kr for the +10 charge state via least squares fitting to the natural logarithm of the relative intensity of hMb. 73 4.4 Variation of the natural logarithm of the total ion counts of hMb +18 with trapping time. The dashed line is the least squares fit giving kr(T)=l\A s _ 1 75 5.1 Schematic of the linear ion trap and the T O F M S . IQ, the interquad aperture, serves as the trap entrance and L I as the trap exit 82 5.2 Effect of flight tube pressure on the resolution of cytochromec (panels (a) and (b)) and reserpine (panels (c) and (d)). The pressure in panels (a) and (c) is 3 .4xl0~ 6 and in (b) and (d) it is 5 . 5 x l O - 5 . Mass resolution are (a) 200 (b) 30 (c) 275 and (d) 195 87 5.3 Effect of altering the voltage on the variable deflector on (a) number of ions detected (b) resolution, for a float voltage of -2.5 kV and the protonated reserpine ion (—=609) 90 5.4 T O F M S spectra of reserpine contrasting instrumental param-eters optimized for maximum resolution and sensitivity. The vertical axis for both plots is the total number of ion counts (in each 5 ns channel) divided by the number of T O F scans, i.e., the number of ions per T O F scan, (a) Flight tube po-tential -2 kV, no deflectors used, lens stack L1=L2=L3=L4=-10 V . Resolution (-£^)FWHM = 740. Peak area/spectral ac-quisition time = 4328 ions/s. (b). Flight tube -3.15 kV, 2-deflector=-3.75 kV, y-deflector=-3.15 kV, L1=L4=-100 V , L2=L3=+50 V . Resolution = 240. Peak area/spectral acqui-sition time = 56 500 ions/s 94 5.5 T O F M S spectrum of C s ^ . ! exhibiting spectral features from the low mass cutoff ( f ~ 750) to f = 5000. The system was optimized for maximum sensitivity for C s ^ I ^ (^=3510)^ shown by (*) 96 5.6 Trapping enhancement of T O F signal. The vertical axis rep-resents the number of ion counts for all observed charge states of cytochrome c in each pulse 98 ix 6.1 Ion trajectories calculated from Eq. 6.24 for fa of (a) 197, (b) 199.5, and (c) 200 kHz. The precursor ion is the +3 charge state of renin substrate, /o = 200 kHz, Aa = 1 V , and the bath gas is 6 mTorr N 2 110 6.2 Schematic of the coupling of the auxiliary drive (aux drive) to the quadrupole rods. The auxiliary drive consists of the output of the arbitrary waveform generator passed through a 2.5x step up transformer. 116 6.3 The timing parameters for a typical 20 ms M S / M S scan. IQ is the entrance aperture of the linear ion trap, and L I , the exit aperture. These apertures control the flow of ions into and out of the linear ion trap. The auxiliary drive controls the coupling of the arbitrary waveform generator to the quadrupole rods and is used for ion isolation and excitation 118 6.4 A typical excitation curve obtained in determining optimal conditions for M S / M S 123 6.5 Results of a tandem in space mass spectrometry study on the +3 charge state of renin substrate with 1 mTorr krypton as the collision gas. The offset between Q0 and the collision cell is shown in the upper right hand corner of each panel 125 6.6 Isolation of the -f-3 charge state of renin substrate = 587) through the use of the notched broadband excitation wave-form, (a) Spectrum of a mixture of renin substrate (3.75 pM) and reserpine (0.75 fiM) trapped for 4 ms without application of the broadband excitation waveform, (b) Spectrum recorded after 4 ms application of the broadband excitation waveform with a notch at ^ 587 (amplitude 30 V 0-peak, notch 217-224 kHz) e 127 6.7 Fragmentation spectra of the isolated -1-3 charge state of renin substrate. Fragmentation was induced by a 4 ms period of resonance excitation with amplitude 1.5 V and qu — 0.623. The horizontal axis is channel number where channel 0 is 30 (JLS and each channel represents 20 ns. The vertical axis shows the number of ion counts in each channel. Spectra are shown for / a=219-223 kHz 132 x 6.8 Fragmentation efficiency of the linear ion trap, (a) Reduction of the spectra shown in Figure 6 showing the precursor ion in-tensity (•) and the sum of the intensities of all fragment ions (o). The F W H M of the excitation profile is 3.0 kHz, mass resolution 73. (b). Similar plot for reserpine with resonant excitation of amplitude 2.4 V applied for 4 ms and qu — 0.51. The nominal mass resolution of excitation is 60 and the frag-mentation efficiency is 60% 134 6.9 The effect of increasing the resonance excitation period for precursor ion reserpine with Aa— 1.5 V . Note in Fig. 6.10 that for an excitation period of 4 ms Aa = 1.5 V is not sufficient to induce full fragmentation 136 6.10 Changes in the fragmentation spectra from the precursor ion reserpine (+1) as a function of amplitude of the resonant ex-citation at qu — 0.51 and an excitation period of 4 ms. The horizontal axis is channel number where channel 0 is 20 /AS and each channel represents 20 ns 139 6.11 Effect of varying the amplitude of the auxiliary excitation for the +3 charge state of renin substrate 140 6.12 Fragmentation spectra with excitation amplitudes of (a) 1.5 V and (b) 3.75 V 141 7.1 The number of ions detected as a function of pressure with the L I T / T O F M S operated with (a) no trapping and (b) a 5 ms injection time. Note that in (a) the number of ion counts increases sixfold and in (b) twofold 148 7.2 Resonant excitation curves for the +3 charge state of renin substrate as a function of both pressure and excitation ampli-tude. Precursor intensity is given by • and the sum of fragment ion intensities by • . . 152 7.3 Variation of resolution as a function of pressure for various Aa. Values shown in legend is the 0 to peak amplitudes of the voltages applied to the rods 154 7.4 Excitation curves with the parameters of excitation voltage op-timized for high resolution. In (a) Aa = 225 mV and t=4 ms, (b) Aa=250 mV and t=3 ms, in (c) Aa=275 mV and t=2 ms, and in (d) 4a=300 mV and t=4 ms 155 xi 7.5 Trajectories for an ion in 3 mTorr N 2 and / 0=200 kHz, irradi-ated with an excitation voltage with / a=197 kHz, and Aa{&) 250 mV, (b) 500 mV and (c) 1000 mV 159 7.6 Trajectories for an ion with /o=200 kHz irradiated with an excitation voltage with J4 a=250mV and / a=197 kHz for pres-sures of (a) 1.5, (b) 3.0 and (c) 6.0 mTorr N 2 160 7.7 Trajectories for an ion with / 0 = 200 kHz in 3 mTorr N 2 irradiated with an excitation voltage with fa = 199.5 kHz and Aa of (a) 250 mV, (b) 500 mV and (c) 1000 mV 162 7.8 Trajectories for an ion with /o=200 kHz irradiated with an ex-citation voltage with 4 a =250mV and / a=199.5 kHz and pres-sures of (a) 1.5, (b) 3.0 and (c) 6 mTorr N 2 . 163 7.9 Isolation with 1.5 mTorr N 2 in the linear ion trap demonstrat-ing a mass selectivity of 100. Panel (b) shows the results of the two step isolation procedure, detailed in the text, applied to the collection of ions shown in (a) 165 7.10 Steps representing M S 3 in the linear ion trap. See text for details 170 xi i List of Symbols Symbol Description x,y,z Spatial coordinates. u General spatial coordinate. m Ion mass. M Neutral gas mass. u0 Field Radius of quadrupole geometry. QUI Qu Mathieu parameters. Pu Function of au and qu. e Total charge on ion. E Electric Field. u>0 Fundamental angular frequency of ion motion Angular frequency of auxiliary voltage. fo Fundamental frequency of ion motion . fa Frequency of auxiliary voltage. va Oscillating auxiliary voltage. Aa Amplitude of auxiliary voltage. V Velocity. E Energy. R Resolution. N The number of particles. P Pressure. T Temperature. a Collision cross section. 6CM Center of mass collision angle. Initial (injection) energy. EN Energy after Nc collisions. a Ratio of energies before and after a collision. xi i i Qdrift Spontaneous drift angle. k Number of charges on an ion. I Length. V Collision frequency. n Neutral gas density. t Time. Ea Arrhenius activation energy. A Arrhenius pre-exponential factor. Arrhenius rate constant. kb Boltzmann constant. du Separation of u deflectors. V Voltage. 7 Collisional damping factor. Vrf Amplitude (0 to peak) of R F voltage on quadrupole rods. u DC voltage on quadrupole rods. ft Angular frequency of R F voltage on quadrupole rods. $(x,y, z) Electric potential in space. 00 Potential applied to an electrode. <l>eff Approximation of $ based on pseudopotential. Du Pseudopotential well depth. A Mean free path of ion. Nc Number of collisions. s Source region in T O F M S . d Middle acceleration region in T O F M S . D Field free drift region in T O F M S . P Charge density. T Velocity relaxation time. V9 Thermal velocity of neutral gas. ki Rate of trapping loss. Eaccel Energy in direction of T O F acceleration. Vaccel Velocity in the direction of T O F acceleration. Fa Force from auxiliary potential. 7 9 Geometrical factor in Fa. Ff Frictional force. aM Polarizability of neutral gas. Reduced mass. x(t) Ion trajectory. xiv Xl(t) Steady state portion of ion trajectory. x2(t) Transient portion of ion trajectory. Frequency of damped oscillations without forcing. Elab Collision energy in laboratory frame. Ecoll Collision energy in center of mass frame. e Permittivity of free space. vr Relative velocity of colliding neutral and ion. a Amplitude of x\(t). 5 Phase shift of Xi(t). b Amplitude of x2(t). Phase shift of x2(t). X V List of Abbreviations Symbol Description aMb Apomyoblobin. BIRD Blackbody Infrared Dissociation. CID Collision Induced Dissociation. ESI Electrospray Ionization. F W H M Full Width at Half Maximum. hMb Holomyoblobin. ICRMS Ion Cyclotron Resonance Mass Spectrometer. IQ Inter-Quadrupole Aperture. IT Ion Trap (referring to the three dimensional Ion Trap). I T / T O F M S Ion Trap-Time of Flight Mass Spectrometer. L I T / T O F M S Linear Ion Trap/Time of Flight Mass Spectrometer. L n The nth Lens. M A L D I Matrix Assisted Laser Desorption Ionization. M C S Mult i Channel Scalar. M S / M S Mass Spectrometry/Mass Spectrometry (also refers to Tan-dem Mass Spectrometry). M S n Multiple or n step Mass Spectrometry. oa-TOFMS orthogonal acceleration-Time of Flight Mass Spectrometer. Qn The nth Quadrupole. Q-TOFMS Quadrupole-Time of Flight Mass Spectrometer. R F Radio Frequency. SWIFT Stored Waveform Inverse Fourier Transform. T O F Time of Flight. T O F M S Time of Flight Mass Spectrometer. xvi Acknowledgements I am tremendously indebted to Don Douglas. Don took me on as a student under unusual circumstances, gave me an exceptional project, and generously allowed me to present and gain recognition for the results. I have been in grad school long enough to appreciate the uniqueness of my situation. In addition, Don's multiple readings of this thesis were thorough, prompt, and completed in exotic locales. I learned much in the exchange. I also thank Don for providing our "real boss", Bruce Collings, who helped me at every level with this project. Bruce has been an excellent teacher, a patient coworker, and a good friend. I thank him for being all three. Thanks also to all my coworkers in the Douglas lab. The technical assistance I received from the U B C electronics (Dave Tonkin) and mechanical shop (Brian Snapkauskas) was greatly appreciated. I gratefully acknowledge financial support received in the form of fellow-ships from NSERC, U B C and the U B C Dept. of Chemistry. Funding for the work was provided by Don's NSERC SCIEX Industrial Chair. It would be dishonest to mention financial support without including my parents. The fact I had the privilege to complete this degree reflects the values I was fortunate enough to gain from them. For love and support I would also like to thank the rest of my family - Christine, Claudio, Noemi, Marcelo, Denise and my special friends - Mark, Mary Ann, Celia, Mark, Pat, Tyler, Todd, Frank. Saving the best for last I would like to thank my husband, Claudio, for more than could be possibly mentioned here. Scientifically, Claudio has always played a dual role as my biggest fan and my most astute critic. His multiple readings of this thesis forced me to understand and clarify this work. Personally, he is my partner in crime, my hero, my drinking buddy, and above all else, my best friend. xvii For Claudio xviii Chapter 1 Introduction Following the ground breaking work of J. J. Thomson, the first "modern" mass spectrometer was constructed in 1919 by Thomson's student, Aston, who devised a method to separate ions based on mass to charge ratios. The earlier work of Thomson, combined with the new "mass spectrograph", made an immediate and fundamental contribution to science as a whole, experi-mentally solidifying the existence of isotopes thereby contributing to a new understanding of atomic structure. Thomson and Aston were separately awarded Nobel prizes for their work. Throughout this century, mass spec-trometers have found numerous, if less grandiose, applications in the scientific enterprise, dominantly in the separation of isotopes - notably in the nuclear industry, routine analysis in the petroleum industry, and structure determi-nation in organic chemistry. In the 1980's the parallel development of two new techniques - matrix as-sisted laser desorption1 (MALDI) and electrospray2 (ESI) ionization - opened a plethora of new applications to mass spectrometry. The novel ability to create gas phase biomolecules has redefined mass spectrometry, creating a multidisciplinary arena in which complicated ion optics techniques adopted from particle accelerators are used in the analysis of biologically derived sam-1 pies of pharmaceutical interest. It would be imprudent not to note that the entry of mass spectrometry in the realm of biomedical research has also re-sulted in a flow of new funding, and commercial interests, providing a strong base for fundamental research. The challenges presented by biological applications to mass spectrome-try have precipitated radical improvements in instrumental capabilities. One primary focus of instrumentation designed for biomolecules has been tandem mass spectrometry, otherwise denoted as M S / M S . 3 Tandem mass spectrome-ters, which were developed prior to the new emphasis on biological molecules, provide the ability to isolate and fragment "precursor" ions with a selected mass to charge ratio. The mass spectrum of the dissociation products pro-vides a window into the structure and energetics of the precursor. In bi-ological mass spectrometry, M S / M S is playing two pivotal roles. First, by examining the fragmentation pattern, the partial or complete 4' 5 primary se-quence of a protein or peptide may be obtained. Second, through performing tandem mass spectrometry on known biomolecules, structural fingerprints or "biomarkers" can be defined from the fragmentation pattern. These patterns are entered into computer databases and used to determine the components of "real" samples, i.e., mixtures of unknown composition derived from biolog-ical materials.6 The combination of tandem mass spectrometry and computer database searching has created a new and powerful methodology to aid in the understanding of protein function and identity, the widespread application of which has emerged as a new discipline - "proteomics" 7 The use of tandem mass spectrometry for the database identification of biological molecules also places new demands on instrumentation. First of all, the molecule of interest has to be isolated from the biological material, a process which often requires some element of liquid phase manipulation. 2 As these processes are typically completed by separation technology, ideally mass spectrometers should be simply coupled to electrophoresis or liquid chromatography. Samples tend to be available in low, that is femtomole, quantities, requiring superior instrumental sensitivity. Unambiguous use of computer databases requires the mass to charge ratio of fragment ions to be provided with high levels of accuracy, which is only possible with high resolution instruments. A brief survey of existing instrumentation highlights the limitations of present technology. At present, no adequate technique for coupling the M A L D I ionization source to effective tandem mass spectrometry has pro-gressed beyond the research phase. Utilization of ESI with triple quadrupole and ion trap mass spectrometers provides only adequate resolution and mass accuracy, particularly when coupled to separation techniques. While com-plex spectrometers, such as the hybrid quadrupole time of flight system,8 and ion cyclotron resonance mass spectrometers9 can address the needs of proteomics, the exclusive price tag does limit widespread use. This thesis discusses the design, construction, and characterization of a new instrument with tandem mass spectrometry capabilities which was designed to, at least in the future, meet the requirements of biological appli-cations - the linear ion trap time of flight mass spectrometer ( L I T / T O F M S ) . The novel centerpiece of the L I T / T O F M S is the "linear ion trap", which is itself a new addition to the field of mass spectrometry. The linear ion trap is created by using electrostatic potentials to confine ions in a radio frequency (RF) only quadrupole rod set. Paralleling the operation of the 3-d ion trap, resonant excitation is used both to isolate precursor ions and to fragment ions via collision induced dissociation with a low pressure neutral bath gas. The fragment ions are mass analyzed and detected in a time of flight mass 3 analyzer, thus completing the M S / M S procedure. This design affords numerous theoretical advantages over existing instru-mentation, including increased sensitivity, high resolution on fragment ions, flexibility of operation, and simplicity. Realizing any of these advantages, however, requires an understanding of the heart of the spectrometer, the lin-ear ion trap. As this work represents the first attempts to complete sequential ion isolation and fragmentation in a linear RF-only quadrupole, the focus of the thesis is proof of the feasibility of storage and resonant excitation in the linear ion trap. A further common theme of this thesis is the orthogonal coupling of both the electrospray ionization source and linear quadrupoles operating in the mTorr pressure regime to the time of flight mass analyzer. This work also discusses the use of trapping in linear quadrupoles prior to time of flight analysis, a topic which has not received significant attention in the literature to date. The thesis is organized around the description of the sequential develop-ment of the linear ion trap. The theories underlying the two technologies which are "hybridized" in the L I T / T O F M S , quadrupole and time of flight mass spectrometers, are outlined in Sees. 2.1 and 2.2 respectively. Chapter 2 also discusses the challenges of coupling the electrospray ionization source to time of flight, mass analyzers and instruments which are similar to, or which influenced, the design of the L I T / T O F M S . The focus of Chapter 3 is the examination of the radio frequency only quadrupole as a storage device, outlining the injection, extraction, and trap-ping efficiencies of the linear ion trap, as well as the total ion capacity of the trap. The procedures developed in Chapter 3 were applied in Chapter 4 to a study of the stability of highly charged gas phase holomyoglobin. Storage in a radio frequency only quadrupole was also used in Sec. 4.3 to determine 4 the temperature of ions confined in a 2-dimensional quadrupolar potential. The remaining three chapters discuss the construction and performance of the L I T / T O F M S . Chapter 5 concerns the experimental details of the in-strument, in particular, factors such as electrostatic deflectors and the use of trapping, which affect the performance of the time of flight mass analyzer. Chapter 6 presents the theory behind, and demonstrates the initial results of, the use of resonant excitation to realize ion isolation and fragmentation in the linear ion trap. As a limitation in the attained performance was poor mass selection, Chapter 7 examines the relationship between bath gas den-sity, amplitude of the applied resonant excitation, and mass resolution. Finally, Chapter 8 provides a summary of the thesis and discusses neces-sary future work for the L I T / T O F M S to realize its potential as a new mass spectrometer. 5 Chapter 2 A Review of Instrumentation Similar to the LIT/TOFMS The first two sections of this chapter introduce the principles of operation of quadrupole and time of flight based mass spectrometers, respectively. The remaining sections of the chapter are dedicated to describing previous instru-mentation similar in design and function to the L I T / T O F M S , in particular, the two "competitive" technologies, quadrupole- and 3-d ion trap- time of flight mass spectrometers. 2.1 Quadrupole Mass Spectrometers The linear ion trap is a further example of the common quadrupole based mass spectrometer and its operation is based on the known principles of ion motion in the quadrupolar field. These concepts, as well as a description of the various types of quadrupole mass spectrometers are reviewed in a text, Quadrupole Storage Mass Spectrometry, by March and Hughes1 0 and discussed in the classic book by Dawson. 1 1 A brief (and simple) summary of quadrupole mass spectrometers focusing on the three quadrupole based instruments referred to throughout the thesis - the mass filter, the RF-only 6 quadrupole, and the 3-d ion trap - is given here. A quadrupole mass filter consists of four circular rods (alternately termed poles or electrodes) whose spacing and applied potential, <f>0, are such that an approximate (ideally hyperbolic) 2-dimensional electric quadrupolar po-tential is created between the rods. The configuration of the rod set, the applied potentials, and the equipotential lines of the quadrupolar potential are shown in Fig. 2.1. The rods are parallel and equally spaced, and the potentials applied to opposite electrodes are in phase, while those applied to adjacent electrodes are of opposite sign. The electric potential, between the rods is given by 9(x,y)= {x2~ V%0 (2.1) where x and y are the displacements from the center to any point in the coordinates shown in Fig. 2.1 and ro is the distance between the centerline and the rod, termed the field radius of the device. The potential is zero in the center of the quadrupole and increases quadratically away from the cen-terline. Creation of a perfect quadrupolar potential requires electrodes with hyperbolic contours. In practice, the same potential is closely approximated by using more simply manufactured circular rods with the radius of the rods, r, and r 0 related 1 0 by r = 1.1487r0. The trajectories of charged particles in any potential are calculated from Newton's law of motion (F = m^f) with the force experienced by an ion of mass m in an electric field, E , given by F = eE where e is the total charge of the ion. The electric field is the negative of the gradient of the potential, E = — V $ ( x , y), and hence, in the x coordinate, the equation of motion is m d ? = - e t o ^ » ) - ( 2 - 2 ) For the quadrupolar potential, the electric field (in the x coordinate) is 7 Figure 2.1: A cross section of a rod set of a quadrupole mass filter, show-ing the applied potential, 0o and the equipotential lines of the resultant quadrupolar potential. given by | _ $ M = _ _ 0 O , ( 2.3) an equality which establishes two important features of the quadrupolar po-tential, both of which are crucial to the operation of the L I T / T O F M S . First, the force on an ion increases linearly from zero at the center line of the rods. Second, motion in x is independent and thus separable from the motion in y, and vice versa. Of course, solving Eq. 2.2 requires a knowledge of <J)Q. For the mass filter, 8 = -e=Z(U-'V^cosfii). (2.5) at2 ** the applied potential is given by (j)0 = U - V r f cos nt (2.4) where £/ is a DC voltage and VTj cos Q.t, is an RF-voltage with a 0 to peak amplitude of Vrj and angular frequency Q. Thus Eq. 2.2 is cPx 2x If Eq. 2.5 is rearranged and a new parameter £ = ^ is introduced, the equation of motion leads to a Mathieu equation d2x — + (ax-2qxcosf)x = 0 (2.6) where the Mathieu parameters for x motion, ax and qx, are given by - 8 U ' e o and 4 ^ (2.8) •*x » 2 . 0 2 r 2 ' e 'o In the mass filter qy — — qx and aj, = — ax. In the following text, u denotes the general coordinate, x or y. The solutions to Eq. 2.6 are complicated and define boundary values of ax and qx between regions of "stable" and "unstable" ion motion. Ions with stable motion will have amplitudes of oscillation less than r 0; ions with unstable motion have trajectories with amplitudes of oscillation that increase and result in the ions striking the rods. Although there are an infinite number of these bounded stability regions, with few exceptions, mass spectrometry involves the first, shown in Fig. 2.2. In the mass filter the potential arises from applying both U and Vrj to the rods. Whether an ion of a jiven ^ is stable in the quadrupole device depends on the ax and qx parameters. By scanning U and K / , but maintaining a 9 Operating line Figure 2.2: The first stability region of the quadrupole mass filter. For a given Vrf and U, ions of different ^ line along the scan line. constant ratio of ax to qx, the mass filter can be used as a mass analyzer. Fixing the values of ax and qx such that only ions of interest have stable trajectories creates a selective mass filter. A n RF-only quadrupole has an identical geometry to the mass filter, but is operated with U = 0, meaning au = 0. Whether or not an ion has a stable trajectory in an RF-only quadrupole is determined exclusively by qu and, as is evident from Fig. 2.2, in the first stability region, the criterion is that 0 < Qu < 0.908. Consequently, if Vrj and ft are fixed, there is a lower limit to the ™ of ions which have stable trajectories in the quadrupole. The RF-only 10 quadrupole thus functions as both an ion beam guide and a low mass filter. One common embodiment of quadrupole instrumentation is the triple quadrupole mass spectrometer. As its name implies, the "triple quad" con-sist of three tandem quadrupoles. Both U and Vrf may be applied to the rod sets of the first and third quadrupoles, and hence these can be used as either mass analyzers or selective mass filters. The second quadrupole is an RF-only "collision cell". This instrumental configuration allows several modes of operation, the most relevant to the work presented here is M S / M S . For recording M S / M S data on a triple quadrupole mass spectrometer, the first quadrupole is used as a mass filter to select the precursor ion, the col-lision cell is filled with an neutral gas in the mTorr regime, and the third quadrupole is used as a mass analyzer. The potential offset between the ion source and the collision cell defines the initial laboratory kinetic energy of the collision between the precursor and the neutral, which, if transferred through sequential collisions into internal energy of the ions will cause fragmentation. The resulting fragment ion population is monitored in the third quadrupole. Another quadrupole mass spectrometer which operates along very similar principles to the mass filter, and also has tandem mass spectrometry capa-bilities, is the 3-d ion trap (IT) also called the Paul trap or the quadrupole ion trap. The 3-d ion trap, shown in Fig. 2.3 is axially symmetric and con-sists of a donut shaped "ring" electrode and two end cap electrodes. To produce a quadrupolar field, the internal radius of the ring electrode, ro, and the distance between the center of the trap and the end cap electrodes, z0, are related by r2, = 2z\. Although several methods are used to create the quadrupolar field,10 the general operation of the 3-d ion trap can be under-stood as being analogous to that of the mass filter - the potential applied to the ring electrodes (opposite rods in the mass filter) is 180° out of phase 11 Figure 2.3: The geometry of the 3-dimensional ion trap. with that applied to the end caps (adjacent rods in the mass filter). There are a few differences between the mass filter and the ion trap which should be highlighted. In the mass filter the quadrupolar potential defines motion in two decoupled dimensions, x and y. Motion coaxial to the rod sets (z) is determined by potential offsets between optical elements of the spectrometer. In the ion trap, the quadrupolar potential defines motion in three dimensions and the trajectories are independent in all three mutually perpendicular directions - x, y, and z. Trajectories in the x and y coordinates are identical and collectively referred to as the r motion. A further property of the ion trap is that ions with stable trajectories are confined ("trapped") in all three dimensions. One early, and still dominant, use of the ion trap is for the storage of charged particles in applications such as spectroscopy.1 2'1 3 Motion and stability in the ion trap are calculated by solving Eq. 2.2 with 12 the electric potential <fr(r, z) given by $(r ,2) = % r 2 - 2 z 2 ) . (2.9) Repeating the calculations of Eq. 2.2 and 2.5, again renders the Mathieu equation (Eq. 2.6). The Mathieu parameters differ from those of the mass filter, namely az = — 2ar and Qz = - 2 g r Correspondingly, the solution to Eq. 2.2, with $(r, z), generates a different stability diagram from that shown in Fig. 2.2. The first stability region for the 3-d ion trap is shown in Fig. 2.4. Aspects of the operation of both the mass filter and the 3-d ion trap are relevant to understanding the linear ion trap. The linear ion trap is created by applying electrostatic stopping potentials in the z direction of an RF-only quadrupole, thus confining ions within the rod set and creating a "linear" or 2-dimensional ion trap (2-dimensional in this instance referring to the quadrupolar potential). The dominant similarity between the mass filter and the linear ion trap is that the equations of motion for re and y, as well as the expressions for qu, are identical. The main parallel between the 3-d and the linear ion traps is that they can both be used for the storage of charged particles. A comparison between 2- and 3-dimensional ion traps as storage devices is included in Chapter 3. Another property shared by both the 2- and 3-dimensional ion traps is that M S / M S is achieved by the use of resonant excitation to manipulate the trajectories of trapped ions. Ions which have stable trajectories in a -16U -8V, rf 2' (2.10) (2.11) 13 Figure 2.4: The first stability region for the 3-d ion trap. quadrupolar potential oscillate with characteristic angular frequencies, cun = (2n + pu)^ (2.12) where n is an integer, —oo < n < oo, and /32 « au + %f where u = x or ?/ in the quadrupole rod set and u = r o r y/2z in the 3-d ion trap. If pu < 0.4, (qx,y < 0.6 in an R F only quadrupole) then the adiabatic approximation 1 4 is valid and ion motion in the quadrupole field is adequately described by the secular motion of a charged particle moving in a harmonic "pseudopotential" of well depth 1 5 D u = ^Vrf. (2.13) The well depth has the units of volts and the maximum energy the ion can have in the rod set is eD u . For a fixed VTj, wo, and ft, each j thus has 14 a unique fundamental resonant frequency u0, (n = 0 in Eq.(2.12)), given approximately by u0 = -^=a (2.14) v8 For higher values of fiu the higher order, frequency terms in Eq. 2.12 become significant and the simple harmonic approximation cannot be used. Since the motion of ions in a quadrupolar field is separable in x and y, the ion can be excited in either coordinate by the application of an auxiliary voltage on one set of pole pairs. This oscillating voltage, Va, has the form Va = AaSwuat (2.15) where Aa and u>a are the amplitude and angular frequency of the auxiliary voltage. Application of the auxiliary voltage at the resonant frequency of an ion (i.e., u>0 = ua) causes the amplitude of its oscillation to increase with time. If the amplitude exceeds r 0 (or equivalently, the energy increase from resonant absorption is greater than D u ) the ion will collide with a rod and be ejected from the quadrupole. In the presence of a background neutral gas, the excited ion motion will result in an increase in the number and energy of collisions. As kinetic energy is transferred to ion internal energy, the ion may reach its critical energy for dissociation and fragment. This method of fragmentation is called resonant excitation, and is used in both the linear, and the 3-d, ion trap to fragment ions through multiple collisions with a neutral bath gas, a process termed collision induced dissociation, CID. Resonant excitation can also be used to eliminate all but the species of interest from the quadrupole device. This is accomplished by creating a broadband waveform from a summation of the excitation waveforms for all u)a from 0 to ^ , excluding a small "notch" for the frequency of and near the fundamental frequency of the precursor. 15 2.2 Time of Flight Mass Spectrometers While the isolation and fragmentation of ions in the L I T / T O F M S are per-formed using quadrupole mass spectrometry, the raw data are recorded as time of flight mass spectra, and thus instrumental performance is dependent upon the operation of the T O F M S . To explain the design of the L I T / T O F M S , in turn, requires a brief discussion of the principles underlying time of flight (TOF) instrumentation. Numerous articles providing detailed reviews of the ideas and mathematics relevant to T O F instruments are available 1 6" 1 9 and a book which focuses on T O F M S and its applications in biological research has appeared recently.20 The general concept of T O F based instruments is that a mixture of ions having different mass to charge ratios (j^j accelerated by a pulse of potential V, to energy, E, will acquire a mass dependent velocity, v, given by and hence exhibit a mass dependent flight time through a field free region of a fixed distance. The ^ of the ions in the pulse can be determined by recording the arrival times, t, at a detector a distance, /, away from the point where the accelerating pulse is applied. The ion flight times are proportional and E — eV. One measure of the quality of a mass spectrometer is its reso-lution, R. Resolution in the mass domain can be understood as the largest ^ for which a neighboring ~ , differing by one, can be detected. If JR=100, then an ion with —=101 could be differentiated from an ion with ^=100. At the same resolution an ion with ^=200 could not be distinguished from an (2.16) as (2.17) 16 ion with ^=201. The mass resolution, R, can calculated by defining A m as the minimum difference in the ^ which a mass spectrometer can distinguish, then R = £ - ( 2- 18) A m In a T O F system, the mass resolution is given by R = A - ( 2 - 1 9 ) 2At v ; where A t is the full width half maximum (FWHM) of the peak arising from the recorded ion arrival times. Hence, any factors which lead ions of the same ^ to have different arrival times, such as variations in the energies and positions of the ions at the time of acceleration or different flight paths to the detector, will decrease the mass resolution. The first instrument based on the T O F idea was constructed in 1948 by Cameron and Eggers and named a "velocitron" . 2 1 This first T O F M S had very low mass resolution of 20. The widespread use of T O F M S began in 1955 when Wiley and McLaren introduced a major instrumental modification to the Cameron and Eggers T O F M S . ' 2 2 The new design of the T O F M S is alternately referred to as the Bendix, 2 3 Wiley-McLaren, linear, or dual stage T O F M S . This seminal paper describes the operation of almost all T O F instruments currently in use and included the concepts of orthogonal acceleration and delayed extraction 2 4 which were re-introduced in the 1990's for coupling ESI and M A L D I , respec-tively, to TOFMSs. The historical importance of this paper is the subject of a recent short article by Karas. 2 5 A schematic of the Wiley-McLaren T O F M S is shown in Fig. 2.5. It consists of a repeller plate, two acceleration grids (labelled middle and final) and a long, field free drift region, followed by a detector. The region between 17 Middle Final Repeller Acceleration Acceleration Plate G r i d G r i d Detector v8 vd VD — s — — d — D Source Middle .C l , r , . * i ,. Drift Tube Region Acceleration Region Figure 2.5: Schematic of the Wiley-McLaren time of flight mass spectrome-ter. the repeller plate and the middle acceleration grid is termed the "source" region (see s in Fig. 2.5) because in the first T O F instruments, the ions were formed within this area. The field free region, denoted D, is termed the flight or drift tube. The key contribution of the Wiley-McLaren T O F M S is space focusing which is achieved through the addition of a second acceleration region, d, between the middle and final acceleration grids. A voltage, Vd, is applied to the middle acceleration grid and the final acceleration grid is maintained at the same voltage as the flight tube, VD, termed the "float" voltage. The repeller plate is connected to a pulsed voltage supply and a pulse, Vs, is applied to it for a short period of time, accelerat-ing the ions from the source region towards the detector. The total energy 18 imparted to an ion is the sum of the energies acquired in the source and ac-celeration regions. Once sufficient time has passed for all ions to be detected, another repeller plate pulse is applied and a second mass scan begins. The need for space focusing relates to problems arising from different initial conditions for ions accelerated from the source region. The source region has a constant electric field, E s , E s = ( V * ~ V d ) , (2.20) s where s is the distance separating the repeller plate and the middle acceler-ation grid. The energy acquired in the source region is given by the product of e, E a , and the distance the ion travels from its position when the pulse is applied to the middle acceleration grid. Hence, an accelerated ion which travels a longer distance in the source region will have a greater total energy. A n ion thus enters the middle acceleration region, and subsequently the flight tube, with a velocity which is a function of its position in space at the time the accelerating pulse was applied. The point at which an ion which travels a longer total distance with a higher energy has an equivalent flight time to an ion which travels a shorter total distance, is termed the space focus plane. In the single stage T O F M S , (d = 0 in Fig. 2.5) space focusing has the geometrical requirement that s = 2D. In the dual stage T O F M S , the location of the space focus plane is determined by the ratio of to Vs, and hence space focusing is achieved by adjusting the repeller plate pulse height. It should be remembered that space focusing only compensates for the initial spatial spread of the ions and does not completely eliminate all resolution degrading effects from the ion source. In the 1955 Wiley McLaren paper, resolutions of 300 were demonstrated, and today, with improved electronics, instruments identical in design to that introduced by Wiley and McLaren achieve resolutions of 1000.24 19 The second major milestone in T O F instrumentation was the introduction of the reflectron in 1973.2 6 As there is no reflectron in the L I T / T O F M S here it will not be discussed other than to mention that the function of a reflectron is to reduce the resolution degrading effects of the ion energy spread in the source region. Numerous modifications in reflectron technology have transpired since its introduction and today the addition of a reflectron to a dual stage T O F M S can be expected to increase resolution by a factor of 5-10. With careful choices of source conditions, ion optics, reflectron voltages, and drift lengths, resolutions of ~ 40 000 have been observed.27 There are numerous advantages of T O F based instrumentation over "tra-ditional" mass spectrometers, many of which derive from the simplicity of its design. The instrument's mass range is theoretically limitless, although in a practical system it is determined by detector technology. As there are no scanning elements, few of the ions from the sample are lost prior to detection, and thus TOFMSs have high sensitivity. The mass scanning time is dictated by the flight time of the heaviest ions and is typically 100 /is. Consequently, the T O F M S can record more spectra in a shorter time and with less sam-ple consumption than other mass spectrometers. The major limitation in T O F M S , as will be discussed below, is the poor duty cycle when coupled with a continuous ionization source. 2.3 The ESI Source Coupled to TOFMS The ionization source in the L I T / T O F M S here is pneumatically assisted electrospray ionization (ESI). 2 8 Briefly, gas phase ions are produced from a solution which is sprayed through a capillary held at a very high potential (~ 5 kV). ESI is a soft ionization technique, hence, ions derived from a broad range of biomolecules, including peptides, proteins, noncovalent complexes, 20 enzymes, oligonucleotides, and D N A have been formed by ESI. Numerous review articles describing the capabilities of ESI have been wr i t t en 2 ' 2 9 - 3 1 and a book on the technique has been recently published.3 2 ESI also provides a simple ionization method for combining mass spectrometric detection with liquid based separation methods, such as high pressure liquid chromatogra-phy and capillary electrophoresis.32 Time of flight mass spectrometry is a particularly attractive technique for the analysis of biomolecules as, with a reflectron, it can deliver a substan-tially higher resolution over a much wider mass range than the quadrupole based mass spectrometers with which the ESI source is generally coupled. The limitation in combining ESI with T O F M S is the incompatability as a source/analyzer pair: ESI creates a continuous stream of ions whereas the T O F M S requires pulsed operation. Consequently, in the direct coupling of ESI sources with TOFMSs a large proportion of the sample which is ion-ized is not detected. The extent of this sample loss is quantified through the instrumental duty cycle, a quantification of the portion of ions formed which are detected. If a pulsed ionization source, such as M A L D I , is used with T O F M S , then a substantial fraction of ions formed can be detected and a near 100% duty cycle is achieved. Conversely, if a continuous ionization source is used, only those ions which are pulsed out of the T O F source region are detected leading to duty cycles which are typically < 1%.33 The challenge for the first instruments using electrospray as an ionization source in TOFMSs was increasing this duty cycle. Two new instruments, the orthogonal extraction- or orthogonal acceleration-TOFMS (oa-TOFMS), and the new hybrid mass spectrometer, the 3-d Ion Trap T O F M S ( IT /TOFMS) emerged from this work. As the L I T / T O F M S incorporates design elements from both these instruments, and one motivation in the development of the 21 o u c w "c3 c o c o 'So u Di u o u. § O H TOF Acceleration Ion Optics Continuous Ion Source Figure 2.6: Generalized schematic showing the geometry and components the oa-TOFMS. L I T / T O F M S was to provide an alternative to these existing technologies, the concepts behind the oa-TOFMS and the I T / T O F M S are discussed here. A schematic of an oa-TOFMS is shown in Fig. 2.6. Ions formed externally are focused into the T O F source region normal to the direction of accelera-tion, hence, the name oa-TOFMS. The oa-TOFMS is operated in a slightly different fashion than the T O F M S . The T O F source region is modified such that both the repeller plate and the middle acceleration grid are connected to pulsed supplies and ions are simultaneously pushed and drawn out of the source region. As the dimension of the source region parallel to the repeller plate, is larger than s, (see Fig. 2.5) a wider pulse of ions is accelerated to-wards the detector. The duty cycle of the oa-TOFMS is given by the ratio of the time required to fill the source to the mass scanning time. Because the time to fill the source region is increased in oa-TOFMS, the duty cycle for the coupling of a continuous ionization source is improved. 22 As is often the case with novel instrumentation, the oa-TOFMS was si-multaneously developed by several researchers. Numerous innovations re-lated to the coupling of continuous ionization sources to TOFMSs have come from the group led by A . F. Dodonov at the Russian Academy of Sciences. Between 1985-87 this group built a series of instruments for high sensitivity mass analysis of ions produced by ESI. There is a Soviet patent dated 1987 concerning the orthogonal extraction configuration,34 but by 1991, when this work was published in Engl i sh 3 4 ' 3 5 similar instruments had been indepen-dently constructed by other groups (outlined below). Moreover, a number of Dodonov's coworkers have worked with K . G . Standing at the University of Manitoba and this group has optimized the performance of Dodonov's oa-T O F M S , 3 6 most notably by the addition of a gas filled RF-only quadrupole before the source region. 3 7 Standing's group has also been instrumental in demonstrating the power of ESI-TOFMS by completing a series of important applications in biochemistry. 3 8" 4 0 The first report in widely available literature detailing the coupling of ESI to a T O F M S was by Boyle et al. in 1991.4 1 The work differed from that of Dodonov and coworkers in that the ion beam enters the T O F source region coaxially, rather than orthogonally, to the direction of acceleration in the T O F M S . In their second generation instrument of 1992, Boyle and Whitehouse 3 3 oriented the ESI source normal to the T O F M S . This new ge-ometry was influenced both by the work of Dodonov, and by similar work on the coupling of other continuous ionization sources to TOFMSs, notably by Grix et al.42 and Dawson and Guilhaus 4 3 with an electron impact source, and Sin et al.44 with a corona discharge. A review discussing recent advances in ESI-TOFMS instrumentation has been published.4 5 A key component in the oa-TOFMSs discussed above, is the placement of 23 the T O F M S perpendicular to the ion source. This orthogonal acceleration ge-ometry is used in the the L I T / T O F M S , and is finding increasing application with a variety of ionization sources, including M A L D I 4 6 and inductively cou-pled plasmas. 4 7 ' 4 8 In addition to the aforementioned improvement in duty cycle, the orthogonal acceleration geometry was designed to minimize the detrimental effects on resolution arising from the spread in energies and po-sitions of ions in the T O F source region. As discussed in Sec. 2.2 resolution is determined by the properties of ions in T O F source region in the direction of acceleration. When ions are sampled from the atmospheric ionization source, into vacuum, there can be a large energy spread (depending on the interface used). Consequently, when ions are transferred from the ionization source to the T O F source region (unless there is an instrumental mechanism of energy normalization prior to the source region) the differences in initial conditions at the point of ionization will translate into a broad distribution of velocities in the ion beam. In the cross beam direction, however, the velocities of the ions will be similar. Thus, if the direction of acceleration in the T O F instru-ment is coaxial to the ion beam, the mass resolution of the T O F M S will be considerably worse than if the direction of acceleration is orthogonal to the ionization source. This idea was discussed and demonstrated in the seminal Wiley-McLaren paper.2 2 Myers and Hieftje have completed a detailed discus-sion of the various differences and the relative advantages of the coaxial and orthogonal coupling geometries for the inductively coupled plasma source.48 Some further advantages of the orthogonal coupling geometry include a re-duction of chemical noise and a mass calibration which is constant over a broad mass range. The second approach to improving the duty cycle was the development of a new hybrid instrument, the I T / T O F M S , which is constructed by combining 24 two mass analyzers: the 3-d ion trap and a T O F M S . This instrument was first used with an ESI source by Lubman and coworkers in 1992.4 9 The ESI source is located external to the ion trap and the ions are transferred into the trapping volume by a series of ion optics. The ions are extracted from the trapping volume as a pulse, which is accelerated in towards the detector. The time of ion extraction triggers the beginning of T O F mass scanning. In this manner, the necessary discretization of the ESI beam is provided by the trap. There is one extraction pulse per T O F mass scan; a near 100% duty cycle is achieved. 4 9 - 5 2 Similar systems have been developed by Purves and L i 5 3 ' 5 4 and, with a M A L D I ionization source, by Doroshenko and Cotter. 5 5 The main attraction of the I T / T O F M S is the high duty cycle. A further advantage is that the resonant excitation can be used for mass selective ma-nipulation of ions, facilitating the isolation and storage of only one species56 and the fragmentation of a mass selected species. 5 7 - 5 9 Resolutions achieved with I T / T O F M S have reached 5 0 00 5 5 whereas resolutions in oa-TOFMSs are routinely double that value. 3 7 2.4 Gas Filled RF-only Quadrupoles in TOFMS A n important innovation in oa-TOFMS has been the addition of a gas filled RF-only quadrupole to transport ions from the ionization source to the T O F source region (see Fig. 2.6). The quadrupole serves as an inlet device for the source region, modifying the beam from the electrospray ionization source such that the resolution, duty cycle, and sensitivity in T O F M S are all im-proved. These benefits derive from the "cooling" (or velocity damping) of an ion beam by collisions in a quadrupolar field. The details of collisional damp-ing will be discussed in Chapter 3. Briefly, ion-molecule collisions in the 25 quadrupolar potential result in translational energies losses that create a slower, lower energy spread, narrower beam entering the source region. The advantages resulting from the lowered velocity of the beam are most ob-vious. The instrumental duty cycle of oa-TOFMS discussed in Sec. 2.3 is mathematically expressed as duty cycle = l s , (2.21) where ls is the length of the source region, vz is the velocity of the ions in the direction perpendicular to the T O F acceleration, and ts is the time between pulses to the repeller plate. Given that ls and ts are fixed, the only way to improve the duty cycle in oa-TOFMS is to lower vz. The reduction in vz results from ion-neutral collisions in the RF-quadrupole. Similarly, the sen-sitivity of T O F M S depends upon the number of ions in the source area, and the slower the ions are moving, the greater the ion density. Consequently, an ion beam with lowered axial energy will result in both an enhanced in-strumental sensitivity and an improved duty cycle. A further advantage of the lower axial velocity, the reduction of the spontaneous drift angle, will be discussed in Sec. 5.2. Douglas and French demonstrated in 1992,60 that the passage of ions through a linear RF-only quadrupole operated at a pressure of 7 mTorr N 2 will spatially focus a broad beam into the narrow low field region in the center of the quadrupolar rods. Hence, if an RF-only quadrupole is placed between the ionization source and the T O F source region, the spatial and energy spreads in the cross beam direction, which can be substantially less than those parallel to the beam to begin with, are minimized. Consequently, the T O F resolution increases. The advantages of using an RF-only quadrupole operating at relatively high pressures have recently been discussed in detail by Tolmachev et al.61 26 and experimentally demonstrated by Krutchinsky and coworkers.37 In the discussed instrumental configuration (TOFMS with a reflectron), the addi-tion of the short RF-only quadrupole operating at 100 mTorr N 2 resulted in a doubling of the resolution (for -f « 1000) from 5000 3 6 to near 10000.37 2.5 Tandem Mass Spectrometry with ESI-TOFMS Instruments In the past two years, literature concerning ESI-TOFMSs has focused on the development of instruments with tandem mass spectrometry capabilities, motivated in part by the potential for the use of such instruments in the lucrative application of biopolymer sequencing.62 These new instruments are hybrid mass spectrometers, resulting from the combination of traditional methods for precursor ion isolation and collision induced dissociation (CID) with T O F based fragment detection. The most widely available hybrid T O F M S , the quadrupole-TOFMS, or "Q-TOF" is created by replacing the second mass filter (i.e., third quadrupole) of a triple quadrupole mass spectrometer with a T O F M S . In the Q-TOFMS, the precursor ion is selected in a quadrupole mass filter, fragmented in a gas filled RF-only multipole (the collision cell), and the re-sulting fragments are analyzed in a T O F M S , oriented orthogonally to the collision cell. As isolation and fragmentation occur in sequential multipoles, the process is termed tandem in space mass spectrometry. In this way, the properties of T O F M S may be exploited in the analyses of fragment ions. Wi th the use of a reflectron, the resolution of the fragment spectrum is typ-ically 5000-10000 compared with a maximum of 4000 expected from triple quadrupole or ion trap mass spectrometers. The mass range for fragment de-tection is increased to — > 10000 from — ~ 4000 on commercially available 27 ion traps and triple quadrupole mass spectrometers. A further advantage of these instruments is that they have a very high mass accuracy, which facili-tates peptide identification via database searching. The Q-TOF system was first introduced by Morris et al. in 19968 and a similar instrument has been developed by Chernushevich et a/. 6 3 As mentioned in Sec. 2.3, an I T / T O F M S can be simply adapted to pro-vide tandem mass spectrometry capabilities. The resonant excitation proce-dures for precursor ion selection and fragmentation in the ion trap are similar to those in the linear ion trap and will be detailed in later chapters. Briefly, the desired species is isolated in the trapping volume by the application of a broadband waveform, or other means, and fragmentation is achieved by using resonant excitation to induce low-energy collisions with a background neutral gas. Since the isolation and fragmentation of the ions occur in the same region, but at different times, the process is referred to as tandem in time mass spectrometry. One advantage of tandem in time over tandem in space mass spectrometry is that it can be used to perform multiple mass spectrometry steps, usually denoted by MS" , where n is one plus the number of times the isolation and fragmentation procedure is invoked prior to the completion of the full mass analysis of the trap's contents. M S n , which can determine the structure of the fragments of fragments, is becoming increasingly desirable as an instrumental capability because of its usefulness in the sequencing of biomolecules and in structural determination. In tandem in space mass spectrometry, n is limited by the number of the mass filter/collision cell combinations, and is fixed at two for the Q-TOF and triple quadrupole mass spectrometers and three for the pentaquadrupole mass spectrometer.64 Tandem in space mass spectrometry, however, does provide access to higher energy fragmentation 28 channels than can be sampled by M S 2 in the ion trap. 6 5 The adaptation of both the oa-TOFMS and the I T / T O F M S for tandem mass spectrometry does alter the performance levels of each as mentioned in Sec. 2.3. The mass selection process in a scanning quadrupole prior to fragmentation lowers the overall sensitivity of the Q-TOFMS relative to that of the oa-TOFMS. The method for calculating the overall duty cycle is unaf-fected by the addition of the collision cell prior to the source region. For the I T / T O F M S , the sensitivity is not altered by the addition of the fragmenta-tion and isolation steps. The 100% duty cycle is applicable when the ion trap is operated for ion storage and beam discretization. However, when the trap is used for M S / M S , the T O F M S cannot scan, and additional ions cannot be injected into the trap, until the isolation and fragmentation steps are com-pleted. The duty cycle for the I T / T O F M S must be separated into the ion trap duty cycle, given by the ratio of the time ions are entering the trap to the total time required for the tandem mass spectrometry cycle to be com-pleted (i.e., the sum of the times required for filling, isolation, fragmentation, and extraction) and the T O F duty cycle, which is 100%. A few other examples of coupling the ESI source with T O F mass analysis incorporating the ability to fragment ions have been introduced. These methods include the manipulation of ion trajectories in gas filled quadrupoles, 6 6 ' 6 7 the use of an electrostatic ion guide,6 8 and hybrid mag-netic sector-oaTOFMS. 6 9 2.6 The LIT/TOFMS A n introductory schematic of the L I T / T O F M S is shown in Fig. 2.7. It con-sists of an ESI source and a linear, RF-only quadrupole which is oriented or-thogonally to a linear T O F M S . The design of the L I T / T O F M S incorporates 29 properties of both the Q-TOFMS and the I T / T O F M S . As in oa-TOFMS, the orthogonal coupling geometry is used in order to improve resolution and duty cycle. Whereas the geometry is borrowed from the oa-TOFMS, the method of precursor ion isolation and fragmentation is analogous to that of the 3-d ion trap and is a tandem in time process. The flow of ions into and out of the RF-only quadrupole is controlled by timed stopping poten-tials on the entrance and exit apertures. When the ions are confined in the quadrupole, an auxiliary voltage is superimposed on the output from the main RF-drive and resonant excitation is used to isolate and fragment the ions. The fragments are then passed to the T O F M S for mass analysis. The use of an RF-only quadrupole for ion storage is an important, and novel, feature of the L I T / T O F M S and is separately described in Chapter 3. While the L I T / T O F M S provides the first demonstration of trapping and dissociating ions in a linear RF-only quadrupole with the fragments ana-lyzed via T O F M S , related work has been both proposed and demonstrated. Ijames70 has described an LIT with ion ejection into a T O F M S through slots in the quadrupole rods. Whitehouse et al.71 have a patent application based on the concept of an L I T / T O F M S and have shown the effects of trapping ions on the T O F spectra (see Sec. 5.4) but no experimental results for tandem mass spectrometry have been given. There are also two recent examples of CID in linear quadrupoles with fragment mass analysis by T O F M S . Dodonov et al.67 introduced a "molecule ion reactor" consisting of a segmented R F -only quadrupole with a longitudinal electric field which was operated at a high pressure. Depending on the mode of operation, CID was accomplished by either increasing the R F or DC voltages along the segments. Loboda et al.66 have modified the RF-drive of the collision cell in a Q-TOFMS to excite ions flowing through the cell, inducing fragmentation. In neither ex-30 Figure 2.7: The L I T / T O F M S , where 1, is the electrospray ionization source, 2 is an entrance aperture plate, 3 is an RF-only quadrupole, 4, is the exit aperture plate, and 5 and 6 are the source and drift regions respectively of the T O F M S . Pulses to 2 and 4 are used to control the flow of ions into and out of the RF-only quadrupole and to confine ions in the trapping volume. The T O F M S is as shown in Fig. 2.5 ample were ions mass selected prior to fragmentation or was trapping used to contain the ions. Whereas the use of an auxiliary oscillating field to manipulate ions for mass analysis, selective ion ejection, and CID is commonplace for the 3-d ion trap, it has found relatively little application in linear RF-only quadrupoles. Exploiting the harmonic nature of the motion of ions in RF-only quadrupoles was the basis of the first mass filter introduced by Paul in 195 8 7 2 and Fis-cher in 1959.7 3 In these instruments an auxiliary voltage, identical to Va (see Sec. 2.1), was connected to the quadrupole rod set. Mass analysis was achieved by measuring changes in the ion current to the rods when ua was scanned. If ua was equal to the secular frequency of an ion in the mass 31 filter and the amplitude of the trajectory increased to greater than r 0 , the current at the rods would increase. This version of the mass filter has re-cently been reimplemented by Welling et al.,74 who have constructed an LIT for ion storage in optical spectroscopy experiments. Forced oscillations of ions in RF-only quadrupoles have been used to achieve selective 7 5 ' 7 6 and broadband 7 6 ' 7 7 ion ejection for ions which were confined75 or flowing 7 6 in the axial direction of the collision cell of triple quadrupole mass spectrometers. A patent on the use of quadrupoles as ion guides discussing the concept of and the potential applications for resonant excitation in linear RF-only quadrupoles was published by Wuerker in 1964.78 There are also a few patents which propose novel mass spectrometers making use of resonant excitation in linear quadrupoles. Two patents held by Syka and coworkers and assigned to the Finnigan Corporation, one on al-ternative (and higher capacity) ion trap geometries79 and one on Fourier transform quadrupole mass spectrometers,80 mention the linear ion trap. Douglas 8 1 suggested using broadband waveforms in the linear ion trap to create a selective "ion bottle" as an ion inlet device to improve the duty cycle of a 3-d ion trap. Senko et al.82 utilize an RF-only octopole as an ion bottle for a Fourier transform ion cyclotron resonance (FTICR) mass spec-trometer and proposed, but did not demonstrate, mass selection of the stored ions. Huang et a/. 8 3 have suggested exploiting the geometry of the linear ion trap for a combined ion trap which makes use of magnetic as well as electric fields. A similar device for laser spectroscopy has been constructed.84 2.7 Summary and Conclusions The instrumentation outlined in this chapter defined the focus for the char-acterization and the desired performance levels of the L I T / T O F M S . The 32 central functionality of the L I T / T O F M S is tandem in time mass spectrom-etry of ions formed by ESI. It was envisioned that the L I T / T O F M S would serve as a simpler (and less expensive) alternative to the Q-TOFMS and the I T / T O F M S . As the ESI, the TOFMS and the orthogonal coupling aspects of the instrument are well known, little work was directed toward novel improve-ments to these techniques. Existing work was used as a guide for expected performance levels which were subsequently achieved (albeit with substantial effort). By contrast, the storage and fragmentation of ions in linear RF-only quadrupoles is new c nd thus is the major contribution of this work. 33 Chapter 3 The Linear Radio Frequency Quadrupole as a Storage Device The linear RF-only quadrupole was assessed as a storage device prior to the construction of the L I T / T O F M S by modifying the collision cell of a triple quadrupole mass spectrometer to operate as a linear ion trap. These experi-ments were motivated by the desire to verify that a sufficient number of ions could be stored in the linear ion trap on a time scale adequate for the pro-cesses of isolation and fragmentation required for tandem mass spectrometry. The linear and 3-d ion traps are compared as storage devices in Sec. 3.1. The experimental details of the modification of the triple quadrupole mass spec-trometer are discussed in Sec. 3.2. Section 3.3 details the effects of multiple collisions between the trapped ions and a background neutral gas, describ-ing the important concepts of the hard sphere collision model and collisional cooling in the quadrupolar potential. Finally the injection and trapping ef-ficiencies of the linear ion trap are demonstrated and discussed in Sees. 3.4 and 3.5 respectively. 34 3.1 The 2-Dimensional Paul Trap Ion storage in linear RF-only quadrupoles is conceptually analogous to the "storage-ring" and "racetrack" traps discussed by Church 8 5 in 1969. In all three, ions are confined in the radial (x, y) direction by the quadrupolar potential. In the linear ion trap, trapping in the axial (z) coordinate is achieved by applying electrostatic stopping potentials at both ends of the quadrupole rod set. Stopping potentials can be maintained on opposite ends of a segmented rod set8 6 or on elements of ion optics, such as aperture plates, located in immediate proximity to the quadrupole rod set.8 7 This simple modification of an RF-only linear quadrupole to facilitate ion storage, was simultaneously developed by research groups with varied interests. In 1989, Prestage et a/.8 8 presented a novel linear ion trap which extended the number of laser cooled atomic ions which could be trapped in a fixed volume with minimal effects from Doppler shifting in observed fluorescence spectra. The success of this apparatus has led to the adoption of similar traps for frequency standards applications, 8 6 ' 8 9 - 9 3 and for recording spectra of trapped ions . 7 4 ' 9 3 - 9 7 This work has necessitated detailed derivations of the trapping potent ia l 9 1 ' 9 8 , 9 9 and the ion capacity 9 0 ' 9 1 of the linear ion trap. In 1988, Dolnikowski et al.87 modified a triple quadrupole mass spectrom-eter by applying pulse sequences to aperture lenses at the entrance and exit of the collision cell, thus enabling the controlled flow of ions; this created a linear ion trap. The motivation behind this experiment was to use the triple quadrupole mass spectrometer to study the time evolution of ion-molecule reactions. This technique has subsequently been used here 1 0 0 (see Chapter 4) and by others 1 0 1 ' 1 0 2 in kinetic studies. The widespread use of the 3-d ion trap and the relative obscurity of the linear ion trap invites comparison of the two trapping geometries for ion 35 storage. In terms of the design of the L I T / T O F M S , an important aspect is a consideration of any differences between the two geometries for the use of each in the respective hybrid T O F M S . The primary distinction between the linear and 3-d ion traps is in the z- coordinate. In the z direction of the 3-d trap there exists a rapidly oscillating quadrupolar potential which must be penetrated to inject ions from an external ionization source to the trap. The injection efficiency for an ESI source into a 3-d ion trap has been quoted to be as low as 0.2%. 1 0 3 By contrast, the processes of injecting and extracting ions from the linear ion trap are very efficient. Potentials on ion optics are lowered from offsets which create z direction trapping conditions, to values which transmit the ions into, and out of, the linear ion trap. As shown below, the injection and extraction efficiencies can be nearly 100%. The differences in the injection and extraction efficiencies of the linear and the 3-d ion traps impact the relative performance of each storage de-vice coupled to a T O F M S . With fewer ions lost in the processes of entering and emptying the trapping volume, less sample is wasted in, and less time is required for, filling the trap. Consequently, if all other parameters are equivalent, the sensitivity of the L I T / T O F M S can be greater than that of the I T / T O F M S . The second major distinction between the linear and 3-d ion traps is the trap capacity, defined as the maximum number of ions, TV, which can be stored without the repulsive forces from ion-ion interaction ("space charge") being greater than the trapping forces of the quadrupolar potential. It should be noted that as the number of trapped ions approaches capacity, the space charge results in a variety of undesirable effects, including spontaneous emp-tying of the trap and shifting of the resonant frequencies of the ions. 1 0 The relative trap capacities of the linear and the 3-d ion traps can be 36 approximated following the work of Douglas 8 1 and Prestage et a/. 9 1 Trap capacity is defined as the product of the trapping volume and the maximum charge density of ions in the pseudopotential well. For the linear ion trap, the volume, V2-d-, can be approximated as a cylinder of length I, V2_d = nr% (3.1) and, for the 3-d ion trap with the geometrical constraint 2z\ — r\, V 3_d is an ellipsoid, i.e., 4 , 4\/2 , V3-d = -7rz2r0 = — T 4 (3-2) Charge density, p, and electric potential, $, are related by Poisson's equation, V 2 $ = -A (3.3) where e0 is the permittivity of free space. In determining the capacity of the 3-d ion trap, 1 0 capacity is approximated by replacing the "true" quadrupo-lar potential with an "effective" potential, 0 e / / , which is a function of the pseudopotential, D T . For the linear ion trap 4><ff = X—^r- (3.4) As Eq. 3.4 is an approximation of the electric potential it does not meet Laplace's condition, thus V2<fieff ^ 0. At capacity, the force from space charge "cancels out" the trapping force of the quadrupolar field, V 2 $ « V2(/>e// « - - . (3.5) It should be appreciated that the force from the pseudopotential and that from the space charge are of opposite sign so that the sum of the two forces at capacity is zero. To calculate the density, one only has to consider the magnitude of the force from space charge. For the linear ion trap 4Dreo „ r v P*-d = • (3-6) ro 37 Correspondingly, for a 3-d ion trap, and hence (r 2 + 2z 2) = 2~? *' ^ ' Ps-d = — — • (3.8) From the product of Eqs. 3.1 and 3.6, N2_d = 4nkoDr, (3.9) and from the product of Eqs. 3.2 and 3.7 7V3_d = 4\/2~7re0zoLV (3.10) The ratio of Eqs. 3.9 and 3.10, gives N2-d I D r (3.11) N3-d V2 z0Bz' If the traps are operated such that pseudopotential well of the linear ion trap, D r , has the same depth as of the 3-d ion trap, N3-d V2z0 ' The linear ion trap with /=20 cm and r0=0.400 cm can theoretically hold 20 times the number of ions of a commercial ion trap with z0=0.707 cm. The above derivation encompasses some critical assumptions. The cal-culations of the trap densities are based upon the pseudopotential approxi-mation and thus valid only when /?u < 0.4. The volumes used for both the linear and 3-d ion traps describe simplified shapes of the trapped ion cloud. Although the absolute numbers resultant from Eqs. 3.9 and 3.10 provide only order of magnitude estimates of the actual trap capacities, the ability for an linear ion trap to store a larger number of ions than a typical 3-d ion trap 38 is well acknowledged. 7 9 ' 8 1 ' 9 0 ' 9 1 The enhanced trap capacity of the linear ion trap translates, with the use of an ESI source, into a possible increase in the concentration linear dynamic range of the L I T / T O F M S as compared to the I T / T O F M S . 3.2 Experimental Methods The initial examination of the linear ion trap as a storage device followed the work of Dolnikowski et al.87 and was accomplished by a simple modification to a custom built triple quadrupole mass spectrometer. This instrument is shown in Fig. 3.1 and has been described in detail elsewhere.104 Briefly, ions from the electrospray source pass through a dry nitrogen 'curtain' gas, a 0.25 mm sampling orifice, a skimmer, and into an RF-only quadrupole, Q0, which acts as an ion guide. Collisions in Q0 (N 2 , 7 x l 0 ~ 3 Torr) cool ions to translational energies of 1-2 e V . 6 0 ' 1 0 4 Ions then pass through an interquad differential pumping aperture, an RF-only prefilter, and into a mass analyzing quadrupole, Q l . Ions enter the collision cell, Q2, an RF-only quadrupole, exit through a second RF-only prefilter, and into a second mass analyzing quadrupole, Q3. A l l quadrupoles had ro = 4.1 mm and Q = 27rl M H Z . A l l of the components are individually connected to DC power supplies, and the axial energies of the ions depend upon the relative D C offsets between quadrupole rod sets. The axial energies of the ions entering the collision cell are determined by the difference between the rod offsets of Q0 and Q2. The overall collision cell length is 20.6 cm and the entrance and exit apertures are 2.5 mm in diameter. Gas could be added to the collision cell and the pressure was measured with a precision capacitance manometer (MKS type 120 high accuracy pressure transducer, manufacturer's stated accuracy 0.12% of reading). If no gas was added to the collision cell it was assumed to have 39 the same pressure as the spectrometer, measured with a vacuum ionization gauge as 1.5 x l O - 5 Torr. For the trapping studies the aperture plates at the entrance and exit of Q2 were connected to an arbitrary waveform generator, which provided the pulse sequences required to operate Q2 as a linear ion trap. The arbitrary waveform generator could only provide voltages in the range of ± 20 V , and hence the D C offsets to the other components of the spectrometer were adapted such that +20 V could trap ions in the collision cell. These voltages are summarized in Table 3.1. Q l was operated in RF-only mode, and hence acted as an ion guide, transferring ions from the ion source to the trapping volume. Q3 was operated in mass selecting mode; thus only ions of a prescribed mass, which were trapped in the collision cell, were detected. The lowering of the exit aperture from a stopping (+20 V) to a transmitting (-20 V) potential triggered the scanning of a multichannel scalar which was connected to a detector operating in ion counting mode. The recorded data are the number of ions versus the sum of the times for the ion to drift from its initial position in the trap to the aperture plate and from the aperture plate to the detector. The distance from the exit aperture of Q2 to the detector was measured as 24 cm and the flight times are dependent upon the axial kinetic energy of the ion. The trap injection energy is calculated by the difference in the potential offsets between Q0 (10 V) and Q2 (0 V) and was typically 10 eV for a singly charged ion. The offsets of the second prefilter and Q3 were -30 V and -15 V respectively. It should be noted that for normal operation of the triple quadrupole mass spectrometer, the second prefilter is generally not used. It was discovered during the course of the trapping studies, however, that adding SR02, and operating it at a low potential offset, increased the extraction efficiency when gas was added to the collision cell. 40 CO o Figure 3.1: The triple quadrupole mass spectrometer system. S, electrospray ion source; O, ion sampling orifice; SK, skimmer; QO, RF-only quadrupole; SRO, prefilter; Q l , mass analyzing quadrupole; Q2, RF-only quadrupole; SR2 prefilter; Q3 mass analyzing quadrupole; D, detector. The linear ion trap is created from the collision cell. 41 For all of the trapping experiments the singly charged reserpine ion (— — 609) electrosprayed from a 1 pM solution of acetonitrile was used. The qu of the ion for all studies was approximately 0.7. There were two objectives for the experiments. The first, as mentioned, was to evaluate the linear ion trap as a storage device, in particular to quan-tify the injection and trapping efficiencies on the time scale necessary for tandem mass spectrometry. The second was to examine the kinetic energy distribution of the extracted ions. This distribution will define the quality of the linear ion trap as an inlet device for the T O F M S , since, as outlined in Sec. 2.3, a number of key performance levels of the L I T / T O F M S , including duty cycle and sensitivity, are determined by the axial energies of the ions in the T O F M S source region. During the preliminary stages of the triple quadrupole mass spectrometer experiments, two problems, the solutions of which (with hindsight) appear obvious, arose. As discussed in Sec. 3.1, one of the putative advantages of the linear ion trap is a high injection efficiency. For an experiment where ions are flowed into the trap for a set period of time, after which the potential on the exit aperture lowers to transmit ions, the injection time and number of detected ions should be linearly related. Initially, however, no matter what the gas pressure in the collision cell, for all injection times greater than ~ 1 ms, only on the order of magnitude of 1000 ions were detected. To understand the origin of this problem a model linear ion trap was constructed using the SIMION software package.105 With SIMION, the user provides the geometry and applied potentials for various electrodes; the pro-gram calculates the electric potential within the prescribed geometry and numerically solves Newton's equation of motion to give the trajectory of ions through the electrodes. The linear ion trap simulation was programmed with 42 oscillating potentials on the quadrupole rod set and timed pulsing sequences to the entrance and exit aperture plates. The voltages on the all compo-nents were identical to those of the triple quadrupole mass spectrometer (Table 3.1). Trajectories of ions in the model linear ion trap provided an explanation for the experimental observations. During the injection period, the exit aperture is held at a stopping po-tential (20 V) . Ions enter the collision cell with an energy of 10 eV, travel to-wards the exit plate, approach the barrier of the stopping potential (+20 V ) , are reflected 180°, and subsequently proceed towards the entrance aperture. Consequently, for the reflected ions, the entrance aperture plate must also provide a potential barrier, otherwise an ion will complete one pass of the linear ion trap and leave the trap. In the initial configuration of the trapping pulses, the potential on the entrance aperture was -20 V , which was enough to literally draw the reflected ions out of the collision cell and towards Q l or to allow the ions to collide with the aperture plate. The plateau for the ob-served ion count was indicative of the number of ions which were completing a single pass of the linear ion trap at the time the exit barrier was low-ered. When the offset of the entrance aperture was increased to a potential which would simultaneously permit the passage of ions into the collision cell and enable trapping of the reflected ions, approximately 7 V , the expected linear relationship between injection time and number of ions detected was observed. Ions which enter the collision cell with 10 eV of energy, will , in traveling through the cell, lose energy via collisions with the gas. The extent of the energy loss is dependent on the pressure of the gas added to the col-lision cell. In these experiments pressures from 1 to 7 mTorr of neon were used. Reflected ions thus approach the entrance aperture with less than 7 eV translational energy and are effectively trapped in the collision cell. 43 Table 3.1: Operating voltages on the triple quadrupole mass spectrome-ter for the injection and trapping studies. Symbols for the components of the spectrometer are from Fig. 3.1. Voltages marked x /y represent the trapping/transmitting potentials of the pulsed entrance (Q0/Q1) and exit (Q2/Q3) aperture plates of the trap. Injection Trapping s 5000 5000 0 150 150 SK 15 15 QO 10 10 SRO 10 10 QI 10 10 Q1/Q2 7 7/10 Q2 0 0 Q2/Q3 -10/10 -10/10 SR2 -30 -30 Q3 -15 -15 Timing Parameters Injection 1-10 s 100 ms Trapping 0 1-40 s Extraction 100 ms 100 ms 44 In a second set of experiments the time profiles of ions leaving Q2 were measured. After injecting ions for a variable period (entrance aperture 7 V) the potential on the exit was lowered to -10 V to transmit ions. The ar-rival time distributions, referred to here a;? time profiles, were measured at the M C S . The second problem which was encountered (once the ability to increase the number of ions injected was realized) was the observation of unusual time profiles when injection times longer than 100 ms were used. The two types of time profiles are shown in Fig. 3.2, with panel (a) showing the behavior at short injection times. As plotted in panel (b), the time pro-files recorded for longer injection periods consisted of a very intense feature at short arrival times, followed by a sharp decrease of intensity, sometimes falling to zero, and then, at longer times, a second rise. After numerous misguided calculations of the energies of each peak feature, it was realized that these unexpected shapes originated from detector saturation. At short arrival times the instantaneous count rate was greater than 5 x 106 counts/s which caused detector saturation. In Fig. 3.2(b), the 82 counts at maximum intensity reflect ions in a 20 fjs channel, and thus an instantaneous count rate of 4 x 10 6/s. As one goal of these studies was to establish the trapping capacity of the linear ion trap, it was desirable to determine the number of ions within the saturated peaks. Consequently, Q3 was operated £ t an inordinately high resolution, to reduce the transmission and thus limit the number of ions reaching the detector. To calculate the actual number of trapped ions from the recorded data, the number of counts at high resolution was multiplied by the ratio of the high to low resolution count rates for conditions in which saturation was not observed. Typically the ratio was in the range of 1000-2000. It should be noted that the data quoted in Sees. 3.4 and 3.5 were 45 TOF/ms Figure 3.2: Time profiles resulting from injection of ions for (a) 100 ms, and (b) 400 ms into a collision cell with 6 mTorr neon. In (b) the decrease in number of ion counts is caused by detector saturation. recorded with Q3 at high resolution and then multiplied by this ratio. 3.3 Collisional Energy Loss Ions lose translational energy in the linear ion trap through multiple collisions with the neutral bath gas. These energy losses can be calculated approxi-mately by a hard sphere collision model, 1 0 6 which has been applied to the initial determination of collision cross sections for biomolecular ions, 1 0 7 and to modeling collisional damping in gas filled RF-only quadrupoles.61 The hard sphere collision model is based on the assumption that particles are rigid spheres which do not interact unless separated by a distance defined 46 by the sum of the radii. If these hard spheres do collide, then the average scattering angle in center of mass coordinate, OCM•> is 90°. Briefly, an ion of mass m, in a neutral bath gas of mass, M , will undergo an average number of collisions, 7VC, calculated by Ne = j , (3.13) where in this case I is the length the ion travels in the collisional environment, and A is the mean free path of the ion, given by A = (no-)'1, (3.14) where a is the collision cross section of the ion and n is the number density of the bath gas. If n is calculated from the ideal gas law, then 7VC becomes Ne = l-f=<r, (3.15) kbT where P and T are the pressure and temperature of the bath gas and kb is Boltzman's constant. The kinetic energy of an ion after 7VC collisions, EN, is calculated from the initial energy of the ion E0 (i.e., the injection energy) via EN = aN<E0, (3.16) where a is the average (i.e., calculated for OCM of 90°) ratio of ion energies before and after a single collision, given by (m 2 + M2) a = (3.17) (m + M ) 2 ' Equations 3.13 and 3.16 allow understanding the interchangeability of the length the ions travel (which is determined by the trapping time) and bath gas pressure. The trapped ion is repetitively reflected between the 47 entrance and exit apertures, increasing / in Eq. 3.13. Hence, trapping ions at a low pressure will result in the same extent of translational energy loss experienced by ions which are not trapped, but travel a shorter distance at a higher pressure. Consequently, trapping can be used in quadrupoles of fixed length and pressure to attain lower final translational energies. The similar effects of increasing trapping time and increasing bath gas pressure are demonstrated in Fig. 3.3. Panel (a) shows the time profiles of reserpine ions (^=609) leaving Q2 after being injected into the collision cell with no gas added. It is assumed that in this instance the number density of neutrals is determined by the 0.015 mTorr background pressure from the N 2 curtain gas. A discrete pulse of ions was created by lowering the stopping po-tential on the entrance aperture for 100 ps and the resultant pulse of ions was not trapped. From Eq. 3.15 a reserpine ion injected with an energy of 10 eV, in traveling 20 cm at 0.015 mTorr N 2 will undergo 0.3 collisions on average, calculated with a = 290 A 2 . 1 0 8 Panel (b) exhibits a time profile recorded with the same pulsing sequence, except that the collision cell pressure was 1.5 mTorr neon (M=20.2 amu), hence Nc = 27. At the low pressure, (a), an asymmetrical distribution of flight times at short times is observed. Wi th a bath gas present (b), a longer flight time and a broadened profile, indicative of loss of kinetic energy, are observed. Panels (c) and (d) show similar plots for trap times of 1 ms in (c) and 10 ms in (d) at a constant trap pressure of 0.015 mTorr and for a longer injection time of 1 ms. The changes in the lineshapes observed between panels (c) and (d) are qualitatively similar to those evident between (a) and (b). Namely, for the longer trapping time (d) a broadened distribution at longer flight times is observed. The average energy of ions leaving Q2 (for Z=24 cm is E = \m ) calculated from the flight time of 250 /is'of (c) is 2.91 eV, and decreases to 0.28 eV, corresponding 48 20 15 10 I 5 O ° 20 15 10 I 1 I 1 I 1 A . . (c) 1 i — 1 n — ~ i 1 i 1 (b) (d) 100 2000 200 300 400 500 1000 1500 TOF/ us TOF / pis Figure 3.3: Time profiles of ions leaving Q2 resulting from injection of ions for 100 fjs, without trapping, into a collision cell of pressure (a) 0.015 mTorr and (b) 1.5 mTorr Ne, and from the trapping of ions in a collision cell with a pressure of 0.015 mTorr for (c) 1 ms and (d) 10 ms. to a flight time of 800 /.is, for the parameters of (d). Figure 3.3 demonstrates the consequences of collisions for motion in the axial, z, direction. As discussed in Sec. 2.1, in the radial coordinate of an RF-only quadrupole, the ion is oscillating in a pseudopotential harmonic well with a depth determined by the Mathieu parameter, qu. A n ion will thus ex-perience collisions resulting in translational energy loss in the radial, as well as the axial direction. In the radial direction, however, the radial transla-tional energy of the ion determines to what extent the ion can overcome the pseudopotential. From Eq. 3.4, the quadrupolar pseudopotential increases 49 from the center of the rod set to the surface of the rods. Consequently, ra-dial energy loss gives a spatial localization of the ion beam. The correlation between the translational energy of ions and position in the quadrupolar po-tential, was observed in the 3-d ion trap in 1959. 1 0 9 If a bath gas is added to the trap, the radial translational energy of an ion will decrease, reducing the amplitude of the ion's motion, and effectively confining the ion to the region of lower potential. This phenomenon has been termed collisional cooling. For the linear quadrupole rod set, the use of a neutral gas to effect colli-sional cooling was first employed by Douglas and French in 1992.6 0 Its most common implementation is in ion guides coupling atmospheric pressure ion-ization sources to mass analyzers operating at low pressures,60 such as QO in the triple quadrupole mass spectrometer shown in Fig. 3.1. As discussed in Sec. 2.4, the properties of a collisionally cooled beam, that is small spread in energy and space in the radial direction, also make the RF-only quadrupole an ideal inlet device for the T O F source region.3 7 3.4 Injection and Extraction of Ions with the Linear Ion Trap One property which was essential to understand and quantify prior to the construction of the L I T / T O F M S was the injection of ions into, and the subsequent extraction of ions out of, the linear ion trap. The injection and extraction properties were analyzed over a range of pressures by constructing pulse sequences designed to inject ions into the trap for varying time periods. This was accomplished by applying a transmitting potential (+7 V) to the entrance aperture plate, and a stopping potential (+20 V) to the exit plate. After the prescribed injection time, a stopping potential was applied to the entrance aperture, and simultaneously, the exit aperture was lowered to a 50 potential (-20 V) which facilitated the extraction of ions from the trap and towards the detector. Time profiles of the extracted ions were recorded. A l l other voltages were as discussed for earlier experiments and shown in Table 3.1. A matrix of data from these experiments is shown in Fig. 3.4, in which the columns represent sequential increases in injection times, from 10 to 20 to 40 ms, and the rows represent a sequential increase in the linear ion trap pressure, from the spectrometer background 0.015 mTorr N 2 , to 3.0, and 6.0 mTorr of neon bath gas. For each of the 9 panels, the recorded number of ions was divided by the average count rate per channel recorded when the triple quadrupole mass spectrometer was operated with the entrance and exit apertures set at offsets for focusing ions into (+7 V) and out (-20 V) of the linear ion trap - subsequently referred to as "flow" rate or steady state conditions. A l l of the panels exhibit a common shape featuring a sharp rise to a maximum, representative of an instantaneous count rate substantially greater than the steady stare ion current, followed at longer flight times, by a rapidly and monotonically decreasing tail. These peaks are similar to those described by Dolnikowski et al.H1 Ions with longer flight times, detected as part of the tail, correspond to lover energy ions. Injection is a sequential process in which the first ions injected are trapped for the remainder of the injection time. Thus an increase in injection time would be expected to parallel a decrease in the average e.oergy of the extracted ions. Several trends in axial ion energy as a function of both injection time and pressure are evident in Fig. 3.4. As both injection time and linear ion trap pressure rise, the fraction of the total ions in the tail, and the effective length of the tail, increase. For a given injection time, as the bath gas pressure is elevated, the ratio of the 51 1 J 1 1 1 1 1 j 1 1 1 L——: T ! 1 1 1 1 ' 1 ' ! ' 1 1 1 I-- —i 1 1 1 1 1 1 1 1 1 1 1 1 1 1,1,1.1, - < — r ' 1 • 1 ' '•••||" 1 1 —1—1—1—1—1—1—1—h— i . i . i -X^t t 1 1 T 1 r f - | . : J i 1 i 1 i —H— . . . . . . , , | i i 1 | i I • i i i i , — i I 1 i 1 i T [ 1 i L - . . . ) 1 | 1 | 1 | , | 1. i . i . i . 4 « = £ : — o C\J i n in E LL O O i n o CM m C/D - E LL 2g i n o CM - E LL o o co o o CM 1 -jjoiuj gi.0"0 o m o CM i - T -JJ01LU O'C i n i n i n CM JJ01LU 0"9 Figure 3.4: Flight time profiles for injection times of 10, 20 and 40 ms, at collision cell pressures of 0.015 mTorr background, 3.0 and 6.0 mTorr Ne. The vertical axis of each panel represents the measured number of extracted ions per channel normalized by the average ion current measured when no trapping is performed. Only the first 25 ms of the full 100 ms extraction time is shown. 52 instantaneous, to the steady state, count rate decreases substantially. The extended injection times and higher pressures also require longer extraction times for all the trapped ions to be detected. Although the total extraction time was 100 ms, Fig. 3.4 only exhibits flight times up to 25 ms, a sufficient time for most of the ions to be extracted with these particular conditions. For the 40 ms injection time at 6 mTorr neon, however, the number of counts did not decrease to near zero until the extraction time was 50 ms. The noticeable increase in the proportion of ion counts measured at long times under conditions for which a larger number of collisions occur, illus-trates a problem in the extraction of ions from the linear ion trap. A well known complication in high pressure ions guides is that collisions can cause translational energy to decrease to the point where an ion has essentially no forward motion. 1 1 0 Unless there is an external applied force, such as an axial electric field, ions with low translational energy will take either an excessively long time to exit (with motion induced by gas diffusion or space charge ef-fects) or be effectively trapped in the collision cell. This trapping is the root of a "cross talk" phenomenon observed in the selective reaction monitoring mode of the triple quadrupole mass spectrometer.111 Briefly, when scanning precursors in rapid succession for fragment ions with identical masses, but stemming from two separate precursors, the fragments from the first precur-sor remain in the collision cell after the fragment ion scan is completed, and hence are erroronously detected as fragments of the second precursor. This problem has motivated the development of new RF-only quadrupole config-urations with axial fields, including dividing the quadrupole into segments and applying increasing DC offsets to each segment,108 and using conically shaped rods with the opposing ends set to different D C offsets. 1 1 1 ' 1 1 2 For the L I T / T O F M S , however, there is no axial field, and thus translational en-53 ergy loss resulting from excessively long injection or trapping times would be expected to degrade the extraction efficiency of the instrument. The lowered extraction efficiency, of course, alters the 100% efficiency hypothesized for the transfer of ions out of the linear ion trap. The time scale for which this assumption is valid is demonstrated in Fig. 3.5. The solid straight line in this plot is the steady state count rate multiplied by the injection time and the dots connected by the dashed line symbolize the total number of ions detected at each injection time (i.e., area under the curve for the peaks in Fig. 3.4). The "flow" curve is calculated from the ion current measured with Q l in RF-only mode and Q3 in mass selective mode as was used in the trapping experiments. For injection times of 1000 ms, background pressures of 4.5 mTorr of neon, and extraction times of 100 ms (more than actually required), the transfer efficiency is near 100 %. As both this time scale and the achieved ion intensities are adequate for the tandem mass spectrometry studies on the L I T / T O F M S , it can be conservatively stated that fewer than 10% of the ions are lost in the process of injection and extraction from the linear ion trap. However, substantial losses could result from the ion optics employed to transfer ions from the linear ion trap to the T O F source region. In order to assess trap capacity, as well as the degradation of the injection and extraction efficiencies at long times, the experiment shown in Fig. 3.5 was repeated with injection times much longer than would be necessary for tandem mass spectrometry in the L I T / T O F M S . The results, for times up to injection 50 s, uniformly recorded with extraction times of 100 ms, are plotted in Fig. 3.6. The vertical axis plots total number of ion counts per scan, and hence the data demonstrate that up to about 2 x 106 ions can be trapped in the linear ion .trap. From Eq. 3.9, for Z=20 cm and 7JU=65.6 V (calculated 54 CD CU c Z3 o O c o CD XJ E Z3 1.5 0.5 200 800 1000 400 600 Injection Time/ms Figure 3.5: Experimental verification of the high injection and extraction efficiencies of the linear ion trap on a time scale up to 1 s. The solid line labelled "flow" is the product of the ion current with transmitting poten-tials on the exit and entrance apertures and the injection time. There was 4.5 mTorr neon in the trap for the duration of the experiment. for c7„=0.708 for reserpine), the theoretical trap capacity is 3 .8xl0 9 . As is evident in Fig. 3.6, however, the number of injected ions detected never approached the calculated trap capacity. The almost unit injection efficiency observed for short time scales is not applicable for longer injection times. Two sources of the lowering of injection efficiency can be reasoned. If the injection time is long enough for substantial trapping losses to occur, then the first ions to enter the linear ion trap, (which are effectively being trapped for the entire injection period) will not be extracted. Factors which result in trapping losses will be discussed in Sec. 3.5. Similarly, ions trapped for longer periods have significantly lower 55 CO cu f 1.5 c o O c o CD .O E 0.5 -* 0.015 mTorr -• 4.5 mTorr Neon 10 40 50 20 30 Injection Time/ s Figure 3.6: The difference in injection and extraction efficiencies for long injection times with 0.015 mTorr N 2 and 4.5 mTorr neon in the linear ion trap. translational energies, and therefore, increases in injection time would be expected to parallel decreases in extraction efficiency. Another feature of Fig. 3.6 is the contrast in the injection process when collision gas is added to the linear ion trap to when the cell is operated at the background pressure of the spectrometer (0.015 mTorr). Wi th 4.5 mTorr neon in the collision cell, there is a near linear increase in the number of ions injected into the linear ion trap up to 10 s - note, almost 10 times the number of ions are observed at 10 s as are at 1 s. When no bath gas is added, however, the injection efficiency decreases substantially. Ion loss occurs when the axial or radial ion energy is enough to surmount the potentials of either the aperture plates, or the trapping potential. The greater the number of collisions that transpire in a given time frame, the higher the probability 56 of these trapping criteria being met. This increased probability of staying within the trapping volume may also contribute to the higher observed ion capacity for the gas filled linear ion trap. It may be that the plateau in the ion capacity at long times results from a steady state, where the rate of ion injection is matched by that of trapping loss. If this were the case, then the trapping losses would be expected to be larger for ions which have higher translational energies in the lower pressure environment. This would result in an enhanced ion capacity when a gas is added to the collision cell. 3.5 Trapping Studies The trapping efficiency of the linear ion trap, like the injection efficiency, has remained relatively unexplored. Beaugrand and coworkers have demon-strated that the trapping efficiency for time scales up to 200 ms is 100%. 1 0 1 This important result was verified here on the triple quadrupole mass spec-trometer by constructing a pulse sequence such that ions were injected into the collision cell for prescribed periods of time, and confined in the linear ion trap by simultaneously applying stopping potentials (+20 V) to both the en-trance and exit apertures plates. After trapping for varying lengths of time, the ions were extracted from the linear ion trap by lowering the exit aperture to a transmitting potential (-20 V) . These experiments were completed on the reserpine ion at qu ~ 0.7 and 3 mTorr neon in the collision cell. As shown in Fig. 3.7, only a 10% decrease in the measured ion counts was observed for trapping times up to 100 ms. This is a significant result as it solidifies that for the time scales which will be utilized for tandem mass spectrometry on the L I T / T O F M S , the linear ion trap has a near 100% trapping efficiency. It was also of interest, however, to determine the trapping efficiency for longer times. This experiment was in part motivated by the desire to use 57 240 CD E 80 -3 40 0 I • 1 • 1 • 1 — — 1 — 0 20 40 60 80 100 Trapping Time/ms Figure 3.7: Trapping efficiency on times scales relevant to tandem mass spec-trometry in the L I T / T O F M S . the linear ion trap for other studies which require maintaining the ions in a fixed volume for longer periods, such as examining ion-molecule reactions and unimolecular dissociation. A high trapping efficiency would allow the L I T / T O F M S to be used in a more extensive range of applications. Fig. 3.8 demonstrates the extent of trapping losses on longer time scales. For this particular plot the injection energy for the singly charged reserpine ion was 10 eV and the gas pressure was 4.5 mTorr of Ne. For these conditions, the trapped ion has a half life of 40 s, i.e., a rate of trapping loss of 0.017 s _ 1 . The trapping efficiency, of course, varies with the ion analyzed a) and with the trapping parameters (qu). Trapping losses will emanate from two main sources. The first is the ion not meeting the simple trapping criterion that ion energy in a given coordinate must be lower than the magnitude 58 in CD c 3 o O c o CD E 3 2 40 60 80 100 Trapping Time/s Figure 3.8: Demonstration of trapping efficiency on a long time scale for the singly charged reserpine ion. Note the change in the ion count axis in the lower panel with the longer trapping time frame. of the trapping potential in that same coordinate. In both the axial and radial directions, the higher the gas pressure the larger the number of energy losing collisions which will occur, and thus, the greater the probability that a trapped ion will meet this energy requirement. In a similar vein, some ions will be lost to "evaporation" as a small percentage of thermalized ions will have sufficient energy to escape the pseudopotential. For long trapping times, it is also possible that the ion can dissociate. The second cause of trapping losses arises from the ions entering a region of the quadrupolar field in which they have an unstable trajectory. One cause of this would be field imperfections. Of more significant concern, however, 59 are the fringing fields at the end of the quadrupolar rod sets. At the ends of the quadrupolar rod sets, the quadrupole electric field diminishes to zero and external fields penetrate into the electrode structure to give a "fringing field". The length of the fringing field is about 1.5 xro from the end of the rods. 1 1 Consequently, as a trapped ion is reflected near either the entrance or exit aperture plate, it may enter the fringing field region, where it could gain energy and have an unstable trajectory. The SIMION simulations of the trapped ions discussed in Sec. 3.2 demonstrated that if the potentials on the aperture plates were greater than the axial ion energy, then all trapping losses occurred in the reflection process. Presumably these losses can be attributed to fringing field effects. 3.6 Summary and Conclusions The characteristics of ions stored in a linear RF-only quadrupole via the application of stopping potentials to entrance and exit aperture plates were examined by modifying the collision cell of a triple quadrupole mass spec-trometer to serve as a linear ion trap. Collisions with a neutral gas result in the ions losing axial energy. The loss of radial energy has the consequence that the majority of ions are confined in the low potential region in the center of the quadrupole rod set. For injection times up to 1 s the efficiency of the transfer of ions from and to the linear ion trap is nearly 100%. Similarly, for trapping times up to 100 ms, the trapping efficiency is almost 100%. For longer injection times, the extraction efficiency decreases; for longer trapping times, significant losses were detectable. The empirical ion capacity of the 20 cm rod set was of the order of magnitude of a million ions. 60 Chapter 4 Ion Trapping Studies of the Heme Binding in Highly Charged HoloMyoglobin The technique for the storage of ions in the linear ion trap, described in Chapter 3, was applied to the investigation of the stability of highly charged gas phase holomyoglobin. This work was part of group collaboration, which was previously published, 1 0 0 in which a new technique to create these ions was developed by Lars Konermann, detailed studies of tandem mass spec-trometry and collision cross sections were completed on the triple quadrupole mass spectrometer by Yu-Luan Chen, and ion temperatures and unimolec-ular dissociation rates were attained with the linear ion trap. The purpose of this chapter is to highlight the contribution of the linear ion trap to the holomyoglobin study. Section 4.1 explains the motivation for the holomyo-globin study and Sec. 4.2 describes the experimental procedures. The inabil-ity to assess accurately the ion temperature limits attempts to derive quan-titative thermodynamic data from observations of unimolecular dissociation and ion-molecule reactions of ions trapped in quadrupolar fields. Section 4.3 details how the temperatures of holomyoglobin ions were estimated. Finally, 61 Sec. 4.4 will discuss the calculated activation energies for the dissociation of holomyoglobin ions of charge states +8 to +20. 4.1 Introduction Contributing to the initial excitement surrounding the electrospray ionization technique was its ability to form gas phase ions of solution complexes bound by noncovalent interactions such as hydrogen bonding and Van der Waals forces. One complex which gained early attention and has served as a model system for the examination of gas phase noncovalent complexes is holomyo-g lob in . 1 1 3 - 1 1 7 In holomyoglobin (hMb) a heme group is noncovalently bound to the protein by Van der Waals interactions, hydrogen bonds, and iron co-ordination. 1 1 8 ' 1 1 9 A fundamental question is whether the specific interactions that bind heme in the solution complex persist in the gas phase. In the gas phase, these complexes are detected as multiply charged ions, the charge on the ion, /c, representing the number of protonated sites plus the charge on the heme. One focus of existing holomyoglobin studies is the strength of the noncovalent binding of the heme group to the protein in the gas phase. This can be probed via the unimolecular dissociation of holomyoglobin into apomyoglobin (aMb) and the heme group: h M b + f c -+ a M b + ( * _ 1 ) + heme +. McLuckey and Ramsey 1 2 0 trapped h M b + f c with A; of 8 and 9 in a 3-d ion trap with a bath gas of 1 mTorr He for times greater than 200 ms, establishing that these holomyoglobin ions could retain the heme group at approximately room temperature for at least this time. In separate experiments, proton transfer reactions were used to determine that the lower bound of the activa-tion energy for the dissociation of the heme group from h M b + 9 was 0.6 e V . 1 2 1 62 A more detailed analysis of the decomposition kinetics of holomyoglobin was completed by Gross et al. who determined the rate parameters for dissoci-ation of the +9 to +12 charge states by blackbody infrared radiative dis-sociation (BIRD) in an ion cyclotron resonance cel l . 1 2 2 In particular, values were obtained for the unimolecular rates of dissociation of holomyoglobin into apomyoglobin and heme (pre-exponential A and Arrhenius activation energies, Ea) with the Ea values ranging from 0.7 to 1.0 eV. Hunter et al. have compared the relative voltages required to dissociate holomyoglobin in the orifice skimmer region of a triple quadrupole mass spectrometer to the known solution phase binding energies for a series of mutants and observed a strong correlation between the solution ar.d gas phase binding strengths.1 2 3 This previous work focuses on relatively low charge states, less than +14, which reflects the difficulties in forming higher charge states. Higher charge states in ESI are normally created by adding a denaturing agent, such as an acid, to facilitate the unfolding of the protein in solution. In the unfolded protein the noncovalent interactions in solution are often destabilized, thus limiting the ability to form highly charged gas phase noncovalent complexes from denaturing solutions. When holomyoglobin is denatured in solution, the heme-protein interactions are disrupted and only apomyoglobin is observed in the mass spectrum. 1 1 6 ' 1 2 4 Konermann et al. have developed a method, "time resolved" ESI, in which the solution containing the native protein is mixed with the denaturing solution for only a short period of time immediately prior to reaching the high voltage region where the solution to gas phase transition occurs. 1 2 4 As the native protein is exposed to the denaturing solution for an insufficient time to allow for the protein to completely unfold, solution transient intermediates can be formed and detected in the gas phase. This technique was used to create the holomyoglobin ions in charge states up to 63 +21, discussed in this chapter. One of the first questions posed in this study was whether the highly charged solution phase intermediates would be stable in the gas phase long enough to be detected. It has been hypothesized that the force of repulsion between charged sites will result in a gas phase unfolding of the complex, ulti-mately creating a "string" like structure in which charges are approximately equally spaced throughout the backbone of the fully unfolded pro te in . 1 2 5 - 1 3 2 For holomyoglobin, it could be reasoned that the geometry changes result-ing from the unfolding due to excess charge parallel a process in which the heme bonding is disrupted. If this were the case, the stability of the heme binding in the holomyoglobin complex would be related to the number of charged sites in the protein and the energy required for dissociation would thus decrease with increasing charge state. A lowered stability in highly charged gas phase holomyoglobin would af-fect the ions' thermochemical stability. Additionally, the lifetimes of the highly charged ions would be substantially lower than those with fewer charged sites. If an ion has a long lifetime in the linear ion trap, then it can also be observed in the 3-d ion trap, and thus the stability of the ion could be assessed using a variety of techniques associated with that instrument, in-cluding collisional activation, ion-molecule reactions, and photodissociation. The extent of unfolding of the highly charged holomyoglobin ions was determined by Yu-Luan Chen through measuring the collision cross sections (tr) for the charge states from +9 to +21. 1 0 0 The +9 charge state has a « 1300 A 2 and this increases monotonically to « 2500 A 2 for the +21 charge state. This highest cross section represents a 90% increase in area, but is still only one half the calculated a of 4944 A 2 for a holomyoglobin ion which is fully unfolded into a string like structure. 1 3 1 64 By measuring the changes in the populations of holomyoglobin ions trapped in the linear ion trap for different times and fitting the result-ing intensities to exponential loss curves, the ther mo chemical stabilities of holomyoglobin ions in charge states from +8 up to +20 were assessed. It was found that highly charged ions can bind heme for 0.5 or more seconds and the rates of loss for the heme are comparable in all charge states. Similarly, the energies calculated for the dissociation of holomyoglobin into heme and apomyoglobin are identical within uncertainty for all charge states. These results suggest that in highly charged gas phase holomyoglobin ions the heme-protein interactions remain relatively unperturbed, even though the protein has almost doubled in cross sectional area. 4.2 Experimental Methods A novel aspect of the experimental work was the use of the continuous flow mixing apparatus, developed for time resolved ESI. This has been described in detail previously. 1 2 4 Briefly, two syringes, one containing 40 fiM horse heart myoglobin in water and the other 0.45% acetic acid (the denaturing agent), were simultaneously advanced by a syringe pump at a rate of 5 y^L/min. The output of each syringe flowed through a capillary and the flow from the two capillaries mixed at a "tee". The time for the two solutions to flow from the "tee" through a 8 cm, 75/zm (inner diameter) capillary to the sprayer is 2.1 s. The mass spectrometer is the triple quadrupole system shown in Fig. 3.1 and operating parameters are discussed below. A typical mass spectrum is shown Fig. 4.1. The relative intensities of the charge states, as well as the ratio of holomyoglobin ions to apomyoglobin ions in the recorded mass spectra were highly dependent on the orifice-skimmer potential difference. Ions are formed 65 at atmospheric pressure and enter a region evacuated to 2 Torr through the sampling orifice. The expansion of gas through the orifice creates a free jet, and a skimmer is placed ~ 0.2 cm behind the orifice such that the centerline flow enters the spectrometer. The energy of collisions in the ion sampling region can be roughly controlled by altering the voltage difference between the orifice and the skimmer, with higher voltages naturally inducing higher energy collisions. The transfer of collisional energy to internal ion energy can result in the undesired (or desired1 2 3) dissociation of ions. Higher orifice skimmer voltage differences (i.e. ~ 100 V) activate and unfold the proteins in the source region and thus the proportion of apomyoglobin ions and lower charge state holomyoglobin ions observed in the resultant mass spectrum is increased. For instance, for holomyoglobin, if the orifice skimmer potential difference was greater than ~ 100 V , then only apomyoglobin would be ob-served in the mass spectrum. For the spectrum shown in Fig. 4.1, exhibiting holomyoglogin in high charge states, the orifice-skimmer voltage difference was a relatively low 30 V . The mass spectrum consists of a bimodal distribution of holomyoglobin ions and a unimodal distribution of apomyoglobin ions. Ions of holomyo-globin in low charge states (+7 to +10) derive from intact heme-protein complexes which have a tightly folded conformation in solution. Ions of apomyoglobin in higher charge states are formed from solution protein which has unfolded and lost the native heme-protein interactions. Highly charged holomyoglobin ions (+11 to +21) are formed from a transient intermediate in solution which has unfolded substantially but still retains the heme group. Ions generated from this transient species could only be detected because the continuous flow mixing technique was used for this study. 1 2 4 The general methodology for trapping ions was described in Sec 3.2, thus 66 001 08 09 Ofr 02 0 Figure 4.1: Mass spectrum of myoglobin obtained with a continuous fl mixing apparatus combined with ESI. The orifice skimmer difference v 30 V . Notation: h8 is hmb+8 a9 is aMb+ 9. 67 only the particulars of the holomyoglobin study will be mentioned here. The injection time was 500 ms and the trapping times were varied from 0 to 500 ms. The trap empty cycle was 100 ms long although the majority of ions were extracted after 40 ms. The time profiles of the extracted ions are similar to those displayed in Fig. 3.4. The background pressure of the collision cell was l x l O - 3 Torr argon. The Q0 to Q2 rod offset difference was fixed at 2 V to minimize energetic collisions which would increase the internal energy of the ions upon entering the linear ion trap (Q2). The orifice skimmer potential difference was 30 V for all charge states to minimize any increase in the ion internal energy resulting from the ESI source. Since the energy of an ion of charge state k traveling through any two potential offsets is k times the difference, it is possible that higher charge states received higher internal energies both in the source and entering the trap. If this were the case, however, the measured binding energy of the higher charge states would be artificially decreased. The temperature studies detailed in Sec. 4.3 required that the populations of both the precursor holomyoglobin ion and the fragment apomyoglobin ion be monitored for various trapping times. For charge states +9 to +12, the holomyoglobin ion with the charge state of interest (k) was mass selected in Q l , and Q3 was set to mass select the apomyoglobin (k — 1) ion. The aMb ion has a charge state of k — 1 as the heme group is lost as a singly charged ion. The same cycle, that is mass selecting for the k charge state and with the identical trapping time, was then repeated and holomyoglobin (k) was mass selected in Q3. In a second experiment, discussed in Sec. 4.4, the holomyoglobin charge states of interest (from charge state +8 to +20) were mass selected in Q l , trapped for varying lengths of time and the total trapping losses were determined by mass selecting the precursor ion in Q3 68 and fitting the recorded populations to exponential decay curves. In this latter experiment the populations of the fragment apomyoglobin ions were not monitored. 4.3 Calculation of the Internal Ion Tempera-ture of Holomyoglobin As discussed in Sec. 4.1, Gross et al. used BIRD to calculate A and the Ea values for the dissociation of the +9 to -1-12 charge states of holomyoglobin into apomyoglobin and heme. 1 2 2 These data were integral to the interpre-tation of the trapping results as they provided a correlation between the measured dissociation rates and the internal temperatures of the trapped ions. An estimate of this internal ion temperature, T, is necessary for the measured rates, k(T), to be used to determine the Ea corresponding to heme loss for the high charge states. This is discussed in Sec. 4.4. The estimation of the ion temperature is, in and of itself, an important experiment. Although ion traps, which can store ions in a fixed volume for relatively long periods of time, provide a convenient tool to monitor ion dis-sociation and ion-molecule reactions, the use of these instruments to derive data, such as Ea or the reaction enthalpy change, is limited by the difficulty in determining the true temperature of an ion trapped in a quadrupolar field. The internal temperature of a trapped ion is a function of the com-peting processes of RF-heating and collisional cooling. 1 5 In the absence of an R F field, the process of collisional cooling would result in the internal ion temperature equilibrating at the bath gas temperature. The micromotion of ions in the RF-field however, induces ion velocities which are greater than the thermal velocity of the bath gas, in effect heating the ions above the bath gas temperature. The resultant competition is dependent on factors 69 which dictate the magnitude of the micromotion, namely the Mathieu pa-rameter, qu, as well as those which determine the rate of collisional cooling, such as bath gas mass and pressure, collision cross sections of trapped ions, a, and the momentum exchange between the bath gas and the precursor, a in Eq. 3.17. Numerous techniques have been used to estimate ion temperatures, including spectroscopy, 1 3 3 - 1 3 6 monitoring of well characterized, temperature dependent chemical reactions, 1 3 7 - 1 4 0 modeling ion mot ion , 1 4 1 ' 1 4 2 and measur-ing the time of flight of ions leaving a t r ap . 5 5 ' 1 4 2 - 1 4 4 In spite of these numer-ous attempts, there is relatively little agreement on the temperature of ions trapped in a quadrupolar field with gas at 295 K, with researchers presenting results which vary from near room temperature 1 3 7 ' 1 3 9 ' 1 4 1 , 1 4 2 to greater than 1 0 0 0 K 55,134,140 This work follows the methodology of groups which measure ion temper-ature by monitoring a reaction for which the rate parameters are already known. For the thermal dissociation of holomyoglobin, Ea and A are known from the BIRD data of Gross et al.122 In order to utilize our rate loss data to predict an ion temperature it was first necessary both to verify that the in-jected ions had ample time to achieve a steady state between RF-heating and collisional cooling and to review the kinetics for the unimolecular dissociation of holomyoglobin ions in the linear ion trap. The time required for ions to reach translational equilibrium in a colli-sional environment has been calculated by Tolmachevet a/. 6 1 Ions injected into a linear trap filled with a neutral gas with an initial velocity, v0 (cal-culated from the injection energy) rapidly lose translational energy. The velocity, v, after a time, t, is given by 70 where, r, the velocity relaxation time is given by AMnaVg and vg is the thermal speed of the collision gas. The lowest temperature at which the ions can equilibrate is that of the bath gas, in this case room tem-perature. Calculating the time required to reach this lower bound for an ion temperature provides an upper bound on the time required for translational equilibrium to be attained. Given that other parameters are constant for all charge states, r will be longest for the holomyoglobin ions with the smallest collision cross section, a. For the +9 state of holomyoglobin, which has the smallest a (<r=1300 A 2 ) , the time, t, to reach room temperature is calculated to be f=s 6.2 ms. This value represents the longest time calculated for any of the observed ions to reach translational equilibrium. The actual time an average ion has to equilibrate is one half the injection time (i.e., 250 ms) added to the full trapping time (i.e., 0 to 500 ms). It is thus more than reasonable to assume that the injected ions have ample time to attain trans-lational thermal equilibrium. However, without fitting the data to existing rate parameters, it is impossible to know the temperature of the equilibrium. The raw data for the dissociation of holomyoglobin in the linear ion trap are represented in Fig. 4.2, which plots the changes in the populations of both the precursor holomyoglobin ion (hMb + f c ) and the fragment apomyoglobin ion (aMb + ( f c - 1 )) as a function of trapping time for the +9 charge state. The number of ion counts have been normalized to the sum of the h M b + f c and aMb + ( f c _ 1 ) populations for each trapping time. As the trapping time increases there is a rise in population of the fragment ion (aMb + ( f c - 1 )) and an equal fall in population of the precursor ion (hMb + f c ) . The measured loss rate constant for the h M b + f c population is a combina-tion of the dissociation loss, fcr, (this is k(T) in the Arrhenius equation) and 71 Trapping Time/s Figure 4.2: Change in the relative populations of hMb +9 and aMb +8 as a function of trapping time. the trapping loss, The experimental decay of holomyoglobin population is given by 4™M = - ( * , + t,)[hMb]. (4.3) Rearranging and integrating from t — 0 to t = t gives [hMb(t)] = [hMb(0)](e-Me- f c '*). (4.4) If one assumes the loss rate constant, of the trap is equal over time and identical for apomyoglobin and holomyoglobin, the trapping losses can be summarized by [hMb(t) + aMb(t)] = [hMb(O) + aMb(0)] e - f c , t . (4.5) Dividing Eq. 4.4 by Eq. 4.5 leads to an expression which is independent of 72 0.05 0.1 Trapping Time/s 0.15 0.2 Figure 4.3: Calculation of kT for the +10 charge state via least squares fitting to the natural logarithm of the relative intensity of hMb. K h [hMbft)] = [hMb(O)] k r t [hMb(i) + aMb(t)] [hMb(O) + aMb(O)] ' 1 ' A plot of In [h Mb^+aMb(t)] v e r s u s t wiU n a v e a slope of -kr. This is illustrated in Fig. 4.3. The values of kr, determined by monitoring the fragment and precursor ion populations as a function of trapping time for the four charge states from +9 to +12 are shown in column 3 of Table 4.1. These values can be substituted for k(T) in the Arrhenius equation kr(T) = Ae-&. (4.7) Rearranging the above equation to T = HT7rr> (4-8) 73 Table 4.1: Observed rates of holomyoglobin dissociation, kr(T), for the +9 to +12 charge states. The Ea and A values are from Gross et a/.x22 and the ranges of temperatures are calculated from Eq. 4.8 Charge State Ea (eV) logio^ M T K s - 1 ) T ( K ) +9 0.8--1.0 9.1- 6.9 4.6 477--807 +10 0.8--1.0 10.1 -7.9 3.3 420--683 +11 0.8--1.0 10.1 -7.9 2.2 411--664 +12 0.7--0.9 9.1--6.9 1.1 396--676 provides an expression for the ion temperature, T, as a function of the experi-mental kr. The resultant ion temperatures are shown in the fourth column of Table 4.1, and range from 398 to 807 K . The average temperature is ~600 K . This range of temperatures is similar to those quoted for other chemical re-actions in the work of Bruce et al.139 and Nourse et a/. 1 3 8 It should also be noted that these temperatures are ~ 150 K above those for which the data of Gross et al.122 were measured, and thus this work assumes that their data can be extrapolated over a temperature range of 33%. 4.4 Binding Energies For the +13 to +20 charge states of the holomyoglobin ion, no rate parame-ters were available to allow for a determination of the ion temperature. For the charge states +8 to +20, holomyoglobin populations were monitored and the total rates of loss of the holomyoglobin were used to derive a rough esti-mate of the Ea such that at least the relative stability of the heme binding in the various charge states could be assessed. Figure 4.4 shows a plot of the natural logarithm of the holomyoglobin ion population against trapping time 74 10 5 1 • ' • 1 • 1 0 0-1 0.2 0.3 Trapping Time/s Figure 4.4: Variation of the natural logarithm of the total ion counts of hMb +18 with trapping time. The dashed line is the least squares fit giving kr(T)=llA s- 1 for the +18 charge state. Similar plots were completed for the remaining 11 charge states. The resultant rate constant for loss, 11.4 s - 1 , is the sum of the rate of trapping loss, hi, and the rate of dissociation, kr. For the +9 to +12 charge states, the rates of loss from the trap were calculated to be 5-8 s - 1 , i.e., about half of the total loss rate, and were found to be independent of charge state. The calculated kr(T) values from the fits of the data for the +8 to +20 charge states to exponential rate loss curves are shown in Table 4.2. The rate constants exhibit no dramatic variations over the observed charge states, indicating that at least the lifetime of the ion is not a function of charge state. These calculated loss rates can be used to estimate Ea if one assumes that T = 600 K and A = 1 0 8 ± 1 1 for all charge states (Table 4.1). The value of Ea 75 is calculated from rearranging Eq. 4.7 and the resultant values are shown in Table 4.2. There is a range of values for each charge state because of the large uncertainty in A from the BIRD experiments. Values for Ea of 0.7 to 1.0 eV are found for all charge states, and there is no systematic change with charge state within the uncertainties of the data. The Ea calculated from the total removal rate will be a lower limit to the actual Ea due to the contribution of ion loss to the observed decay rates. It is important to note that given the large values for A, it is necessary to have at least an order of magnitude change in kT(T) for the change in the calculated Ea to be greater than the quoted uncertainties. If present, such a large difference in dissociation rate could easily be detected in the trapping experiments. The values for Ea quoted in Table 4.2 are obviously dependent upon the validity of a number of assumptions, from the accuracy and relevance of the data of Gross et al.122 to the extrapolation of the thermochemical parameters of the lower charge states to the higher charge states. The aim of the trapping experiments, however, was to assess the relative strength of the heme binding in the low and high charge states. One cannot say with certainty that either A or T is constant for all charge states, but these assumptions are reasonable. As T is a result of the steady state between collisional cooling and R F heating, it would be expected that in similar collisional environments with identical qu, the internal temperatures of the ions will be comparable. The value for the pre-exponential factor can be calculated from 1 4 5 where AS* is the entropy of formation of the activated complex. Thus A will not vary as a function of charge state if the ASt for the dissociation of h M b + A : into a M b + ( f c _ 1 ) and heme is similar for all k. This assumption is (4.9) 76 physically justified if both the structures of h M b + f c and the transition states and the mechanisms of the dissociation pathways are sufficiently similar for the value of AS* to be within the large quoted uncertainties of A for all of the examined charge states. For the +9 to +12 charge the similar A values obtained by Gross et al.122 indicate that is the case. It should also be noted that Gross et al.122 found that Arrhenius pa-rameters depended on solution conditions in their experiment and quoted substantially different A and Ea values for holomyoglobin ions formed from 'pseudonative' (80% water and 20% methanol) and 'denaturing' (50% water, 50% methanol and 0.1% acetic acid) solutions. The trapping data obtained from the continuous flow mixing apparatus is interpreted with the thermo-chemical parameters for the 'pseudonative' solution. M S / M S experiments with a 90% water, 10% methanol solution (i.e., 'pseudonative' conditions) led to dissociation energies which were the same within the experimental un-certainty of those for the identical ions produced with the continuous flow apparatus. 1 0 0 This demonstrates that the most valid Arrhenius parameters to use for calculations involving the trapping data are those of the 'pseudo-native' solution. However, if the Arrhenius parameters quoted by Gross et al.146 for the denaturing solution are used, an average ion temperature from Eq. 4.8 of 1100 K is found. If this higher ion temperature is used in combination with the quoted A values for the holomyoglobin ions from the denaturing solution, the calculated activation energies for holomyoglobin increase by a factor of four but remain similar for all charge states. Thus, use of the rate data from a 'pseudonative' solution gives the lowest estimates of the heme dissociation energies, namely, 0.7-1.0 eV. While the number of assumptions in the quantitative derivation of the Ea 77 Table 4.2: Observed rate constants of hMb loss and calculated activation energies (Ea) for dissociation of hMb ions Charge State Rate (s *) Ea +8 9.6 0.71-0.97 +9 11.3 0.70-0.96 +10 10.8 0.70-0.96 +11 10.9 0.70-0.96 +12 6.6 0.72-0.99 +13 6.9 0.72-0.98 +14 16.2 0.68-0.94 +15 11.1 0.70-0.94 +16 9 0.70-0.96 +17 10.4 0.71-0.97 +18 11.4 0.70-0.96 +19 12.9 0.69-0.95 +20 14.4 0.68-0.95 allow the quoted values to serve only as an order of magnitude approximation to the actual Ea, the trapping work does strongly support the hypothesis that the strength of the heme binding is similar in the low and high charge states. This result is further supported by independent work based on a new method of interpreting tandem mass spectrometry data from the triple quadrupole mass spectrometer, which again resulted in the calculated heme binding energies being between 0.75 and 1.0 e V . 1 0 0 From trapping studies, the strength of the heme noncolvalent binding is apparently unperturbed by the repulsion resultant from additional charged sites in hMb ions up to charge state +20. There was only a small decrease in the binding energies calculated from M S / M S experiments for higher charge states. 1 0 0 This result is particularly interesting in light of the cross section 78 differences between the low and high charge states. From these results it appears that the heme pocket is not disrupted by either the excess Coulom-bic repulsion from the additional charge sites or the unfolding of the pro-tein complex in the higher charge states. This is similar to the behavior of holomyoglobin in the solution phase where it has been demonstrated that hMb can unfold significantly yet still retain a similar structure in the heme binding region. 1 2 4 ' 1 4 7 It is conceivable that the specific interactions of low charge states have been replaced by a myriad of nonspecific interactions that happen to give a similar binding energy. Future experiments with specific residues altered by site directed mutagenesis could give a detailed picture of which individual protein-heme interactions remain in the highly charged ions. 4.5 Summary and Conclusions The modification to the collision cell of the triple quadrupole mass spectrom-eter (described in Chapter 4) was applied to study the stability of highly charged holomyoglobin in the gas phase. Using trapping to monitor the products of thermal dissociation established that the temperature of ions in the quadrupolar field is ~ 600 K . Neither the measured rates of dissociation, nor the calculated Arrhenius activation energies, demonstrated any major changes in values for gas phase holomyoglobin ions with k of +9 to +20. This infers that the noncovalent binding of the heme is not affected by the Coulombic repulsion resultant from the excess charge. 79 Chapter 5 The Linear Ion Trap TOF Mass Spectrometer This chapter describes the design, construction, and characterization of the new spectrometer. The operation of the instrument for recording T O F spec-tra of ions formed by ESI and, briefly, the effect which trapping in the linear ion trap has on T O F spectra are discussed. The adaptations to the instru-ment to achieve tandem mass spectrometry will be described in Chapter 6. Section 5.1 highlights the L I T / T O F M S construction and presents factors which affect the instrumental resolution and sensitivity. Section 5.2 exam-ines the sensitivity resolution tradeoff caused by electrostatic steering deflec-tors in oa-TOF instruments. The resolution, sensitivity, and mass range of the L I T / T O F M S are demonstrated in Sec. 5.3. Section 5.4 shows how trap-ping ions enhances instrumental sensitivity. Some of the results presented in this and the following chapter have been published previously. 1 4 8 80 5.1 The Continuous Flow Time of Flight Mass Spectrometer The L I T / T O F M S is shown schematically in Fig 5.1. Ions are generated by pneumatically assisted electrospray28 and pass through a dry nitrogen curtain gas, a 0.25 mm diameter sampling orifice, a 0.75 mm diameter skimmer orifice, and into the first of two RF-only quadrupoles. The region between the skimmer and the orifice is evacuated by a rotary vane pump to a pressure of 2 Torr (7 L/s , Leybold Trivac D16A, Export, PA, U.S.A.). There is a 2 mm diameter interquad (IQ) aperture between, the RF-only quadrupoles. The first quadrupole, Q0, is 5 cm long and the second, Q l , which acts as the linear ion trap, is 20 cm long. Both quadrupoles have field radii (r 0) of 4.0 mm and are operated by the same main R F drive, which has a maximum Vrf of 5000 V and a drive frequency, f2 == 27r/, where / is 1.0 MHz. The rods of Q0 were custom built at U B C . The rods of Q l were from SCIEX (Concord, ON) and are a prototype of their current commercial rod set. There are 12.5 pF capacitors between the R F drive supply and the rods of Q0; thus the R F voltage on Q0 is approximately one half that of Q l . The D C offsets of the quadrupoles are individually set and typically these are Q0=10 V and Ql=0 V . The pressure in the linear ion trap, 7 mTorr, is determined by the gas (N 2) flow through the ion source region and the pump-ing speed of the attached turbomolecular pump (50 L / s - 1 Leybold Turbovac 50 D, Export, PA, U.S.A.). The use of 7 mTorr of N 2 for the background pressure in the linear ion trap was motivated by considerations of T O F M S performance. This pressure of N 2 was previously shown to give optimal trans-mission with collisional focusing through a 20 cm RF-only quadrupole, and is the pressure commonly used in ion guides connecting ESI sources with the lower pressure regions of triple quadrupole mass spectrometers.60 81 Figure 5 .1: Schematic of the linear ion trap and the T O F M S . IQ, the in-terquad aperture, serves as the trap entrance and L I as the trap exit. 82 The T O F source region is coupled orthogonally to the linear ion trap via four lenses, L1-L4. L I serves as the exit aperture of the linear ion trap and the differential pumping aperture between the linear ion trap and the T O F . As will be discussed below, the diameter of L I was varied from 0.75 to 1.3 mm. The three lenses, L2, L3, and L4, which are separated by 1.27 mm, have apertures of 2 mm diameter and focus the ion beam into the source region of a two stage, 1 m long, Wiley-McLaren T O F M S . The T O F region is evacuated by a turbomolecular pump (345 L/s , Ley-bold). The two turbomolecular pumps are backed by a common rotary vane pump (7 L/s , Leybold Trivac 16A). The chamber of the linear ion trap and lens stack were custom made at U B C and the T O F chamber (custom), flight tube, ion optics (C-677), and 18 mm detector mount (C-701) are from R. M . Jordan Company (Grass Valley, C A , U.S.A.). The detector consists of a dual microchannel plate (Chevron, Galileo Electro-Optics Corp., Sturbridge, M A , U.S.A.). Ion count-ing is used for detection. T O F spectra are acquired using a multichannel scalar (MCS) (Turbo-MCS, E G k G Ortec, Oak Ridge, T N , U.S.A.) with an internal discriminator and a minimum channel width of 5 ns. The M C S is interfaced to a 133 MHz Intel Pentium Pro computer. The T O F repeller plate is pulsed from an offset of 0 V to an amplitude of 200-300 V using a high voltage pulser (rise time < 18 ns, Directed Energy Inc. G R X 1.5K-E, Fort Collins, CO, U.S.A.) . The pulse amplitude on the repeller plate is adjusted to achieve maximum resolution for the ion acceleration energy. This accel-eration energy is determined by the sum of the float voltage, (see Sec 2.2) which was typically between -2 and -4.5 kV, and the energy acquired by the ions in the source region. The ions enter the T O F source region halfway between the repeller plate and the middle acceleration grid (typically set to 83 ground), thus the energy acquired in the source region is one half the repeller plate pulse amplitude (in eV for a singly charged ion). The separations be-tween the repeller plate and the middle acceleration grid (s in Fig. 2.5), and between the middle acceleration grid and the final acceleration grid (d in Fig. 2.5) are both 12.5 mm. The pulse amplitudes which were found to give the best resolutions equaled those calculated to give space focusing for the set acceleration energies.22 The timing of the repeller plate pulsing and the detection electronics are controlled by a four channel pulse generator (Berke-ley Nucleonics Model 501A, San Rafael, C A , U.S.A.) which has a maximum repetition rate of 10 kHz, and hence, electronics dictate a minimum time between spectra of 100 JJS. The width of the pulse applied to the repeller plate is greater than the time required for the highest ^ ions to exit the T O F source region. This width is typically less than the flight time which defines the T O F M S scan-ning rate, typically lOfis and lOOyus respectively. After the repeller plate pulse is complete and during the ion flight time, the repeller plate voltage is set to its offset potential, which allows for ion transmission into the source region. There are 12.5 mm circular holes in stainless steel plates to form the middle and final T O F acceleration grids. The holes are covered with stainless steel mesh. When the repeller plate is pulsed, only those ions contained in the volume circumscribed by this opening enter the flight tube and are acceler-ated towards the detector. Consequently, the ion beam from the linear ion trap is "chopped" into small pulses and the duty cycle of the L I T / T O F M S is given by the ratio of the time for ions to fill the volume defined by 12.5 mm apertures to the time between repeller plate pulses. Duty cycle and sensi-tivity are increased if the ions move more slowly through the source region and consequently it is preferable that the linear ion trap to T O F M S coupling 84 incorporate a method to ensure low energy ions enter the source region. For the linear ion trap, as discussed in Sec. 2.4, axial and radial energy are both lowered by collisional cooling in the gas filled RF-only quadrupole, the linear ion trap. To maintain the low axial energy, the linear ion trap, the offset of the repeller plate, and the middle acceleration grid are all set to 0 V (i.e., the same potential as the Q l offset). Thus ions should not gain energy entering the source region, and the duty cycle will be maximized. For the benefits from collisional cooling to be realized, it is necessary to minimize any beam defocusing in the transfer of ions from the linear ion trap to the T O F source region. In particular, any stray fields from the float potential which penetrate into the source region will cause broadening or deflection of the ion beam, increasing the spatial and/or energy spread and thereby degrading resolution. To reduce the distorting effects of the float voltage a shielding grid was placed 5.1 mm behind, and electrically connected to, the middle acceleration grid. Additional shielding was also installed on the repeller plate and the middle acceleration grid to reduce the effects of stray fields on ions entering the source region. The shielding is shown in Fig. 5.1. In spite of this shielding, there is evidence that stray fields are affecting the recorded spectra. It was found that maximum resolution was achieved using a lens stack with L I to L4 uniformly set to -10 V . Applying lower potentials to the lenses, as well as varying the potentials between the lenses, increased sensitivity but lowered resolution. For example, when the lens stack was operated as an einzel lens (L1,L4=-100 V , L2,L3=-5 V) sensitivity rose by a factor of 4, but resolution fell by ~ 50%. Similarly the number of ions detected was very sensitive to the potential on L4. With L1=L2=L3=-20 V , changing the L4 potential from -20 to -200 V , increased sensitivity by 85 a factor of 4 and lowered mass resolution from 235 to 87. These observed degradations in resolution imply that either the lens voltages are creating stray fields in the T O F M S source region or that these different lens setups result in a divergence in the path of the ion beam. High float voltages also decreased resolution. As the float voltage was increased from 2000 V to 2500 V to 3150 V , the maximum attainable mass resolution on singly charged reserpine ions (with Ll-L4=-10 V) lowered from 680 to 462 to 275. Extensive SIMION simulations of both electric fields in the source region and ion trajectories through the lenses neither predicted nor explained the observed resolution degradation. Another source of resolution degradation for the study of biomolecules, which often have large collision cross sections, is an insufficiently low flight tube pressure. The neutral gas density in the flight tube must be sufficiently low for the distance between the source region and the detector to be greater than the mean free paths (A) of the ions (preferably much greater). Oth-erwise, collisions between ions and residual gas can result in a substantial resolution degradation. 1 4 9 To examine this effect, nitrogen was added to the flight tube to increase the pressure from 3 . 4 x l 0 - 6 Torr to 5 x l 0 - 5 Torr and the spectra of both protonated reserpine (^=609, a — 280 A 2 1 0 8 ) and cy-tochrome c (+13 charge state, f=938 a « 1700 A 2 1 0 4 ) were recorded. The general effects of increased flight tube pressure on T O F resolution are shown in Fig. 5.2. Panels (a) and (b) show T O F spectra for cytochrome c at flight tube pressures of 3 . 4 x l O - 6 Torr and 5 . 5 x l O - 5 Torr respectively. Wi th the number of collisions increasing by a factor of 16, the observed mass resolution degrades from 200 to 30. Panels (c) and (d) show spectra of the reserpine ion at 3 . 4 x l O - 6 Torr (c) and 5 . 5 x l 0 - 5 Torr (d), and demonstrate a fall in resolution from 275 to 195. It should be noted that collisions in the flight 86 650 850 1050 607 609 611 613 615 m/e m/e Figure 5.2: Effect of flight tube pressure on the resolution of cytochroinec (panels (a) and (b)) and reserpine (panels (c) and (d)). The pressure in panels (a) and (c) is 3 . 4 x l 0 - 6 and in (b) and (d) it is 5 . 5 x l O - 5 . Mass resolution are (a) 200 (b) 30 (c) 275 and (d) 195. tube also result in dissociation of ions prior to detection and in the scattering of ions off of the trajectory to the detector, which both can lower the number of the ions of interest detected when high pressures are used. Minimizing the number of collisions an ion undergoes before detection requires lowering the flight tube pressure. From Eq. 3.15 one can show that for an ion to have A > 1 m the product of the pressure in Torr and the collision cross section must be < 3.1 x 1 0 - 3 A 2 Torr. Biomolecules in high charge states often have collision cross sections of approximately 2000 A 2 , thus a minimum desirable pressure would be 1.5 x 1 0 - 6 Torr. Without increasing pumping speed, reducing the flight tube pressure requires either 87 lowering the linear ion trap pressure, which would affect collisional cooling, or making the aperture diameter of L I smaller, which would be expected to result in some ion loss. To improve T O F resolution the aperture diameter of L I was decreased from 1.3 mm to 0.75 mm, decreasing the pressure from 3 . 4 x l O - 6 Torr to 1 .2x lO - 6 Torr. The measured transmission of reserpine ions (^=609) into the T O F decreased by 25%. The resolution of reserpine ions increased, however, to ~ 700, and the resolution of cytochrome c ions increased to 600. For cytochrome c ions, at 1 . 2 x l 0 - 6 Torr, the probability of the ion undergoing zero collisions in the flight path is 52% and of one collision 34%. Extrapolation of a graph of resolution versus j shows that at zero pressure the resolution for cytochrome c is improved ca. 6% over that at 1 .2x lO - 6 Torr. 5.2 Deflectors in oa-TOFMS Prior to beginning the discussion of deflectors, a note on coordinates is nec-essary. The convention adopted in literature discussing RF-only quadrupoles is that the electric fields created by the rods define the x and y coordinates and the z direction is parallel to the rods (thus Ez in Eq. 5.1). Unfortunately, the convention adopted in T O F literature is that the direction of accelera-tion (which is perpendicular to Ez) is also assigned the z coordinate. In order to maintain a consistent set of coordinates for the L I T / T O F M S , the z definition from RF-only quadrupoles is used. It should be appreciated that z-deflectors in the L I T / T O F M S are referred as z-deflectors in much of the T O F literature. 1 5 0 ' 1 5 1 Ions enter the source region of the T O F M S with a residual velocity com-ponent, vz, perpendicular to the velocity in the direction of acceleration, vaccei- From vector addition of the two velocities (v oc \/E), the ion will have 88 a spontaneous drift angle off the axis of the flight tube, 6drift, given by where Eaccei is the T O F acceleration energy and Ez is the axial energy of the ion from the linear ion trap. For example, if Ez, determined by the difference in the offsets between QO, Q l , and the T O F source region, as well as the extent of collisional cooling in the linear ion trap, is 0.2 eV and Eaccei is 2000 eV, then # d r^ t is 0.57° (10~2 radians). An ion with an initial position in the center of the source region will drift 10 mm in the 1 m long flight path to the detector. Considering that a beam 12.5 mm in length (measured along the z coordinate) is being accelerated towards a microchannel plate assembly with an 18 mm diameter, only ~ 40 % of the beam will hit the detector. To minimize the ion loss resultant from the drift angle, electrostatic steer-ing deflectors are used to "push" the ion flight path (post acceleration) to-wards the detector. As shown in Fig. 5.1, there are two pairs of deflector plates in the L I T / T O F M S . Each plate is 22.5 mm long (lz and ly) and the separation between the plates (dz and dy respectively) within each deflector pair is 22.5 mm. The plates located immediately after the final acceleration grid ("2-deflectors") compensate for the spontaneous drift angle. In the z direction of the L I T / T O F M S , the deflector plate farthest from the linear ion trap is electrically connected to, and hence fixed at, the float potential. The opposite plate is electrically isolated from the float potential and connected to an independent voltage supply. The optimal deflector voltage for a given acceleration energy is determined empirically by altering the potential on the variable deflector and maximizing the observed ion counts. Sample data are shown in Fig. 5.3(a), which plots the number of detected reserpine ions as a function of the voltage on the variable z- and y- deflector plates. The float voltage was -2.5 kV. As the voltage on the z-deflector was varied, the y-(5.1) 89 100 o o 80 o a5 60 c O 40 O c o 20 500 c 400 o olut 300 CO CD DC 200 100 : ( a ) • • Z-Deflectors -• — • Y-Deflectors 1 | 1 j ¥ i i • i • i (b) 1 i j 1 1 m j -w j Potential on Variable Deflector / V Figure 5.3: Effect of altering the voltage on the variable deflector on (a) number of ions detected (b) resolution, for a float voltage of-2.5 kV and the protonated reserpine ion (—=609). deflectors were fixed at the float potential and vice-versa. For the z-deflector, the number of ions detected increases when the variable plate has a lower potential than the float voltage. Hence, more ions reach the detector, and consequently overall instrumental sensitivity is enhanced, if there is a net de-celerating force in z. In contrast, variation on the y-deflectors, (there is only a negligible velocity in the y direction in the source region) has a maximum at, and is symmetric about, the float potential. The observed increased sensitivity results from an alteration of the z energy (and hence motion) of the ion beam as it passes through the deflector plates. For simplicity, the following description is based the notion of an "ideal" deflector,151 and assumes that the electric field between the plates is 90 homogeneous. To first order approximation, the angle by which the beam is deflected is identical for all ions, regardless of either their mass or the position at which they enter the deflectors. 1 5 1 ' 1 5 2 A higher Vaccd requires a larger AV^ to achieve an identical deflection angle. In Fig. 5.3(a), the detected ion counts increase to a maximum when the variable deflector plate has a voltage which is 500 V below that of the fixed deflector plate. For this deflector voltage, the number of ions is triple that detected when field free conditions are used. This implies that when no voltage is applied to the deflectors, the beam is 11 mm off axis, and taxiB drift is thus 0.01. Using Eq. 5.1, the z direction energy of ion in the T O F source region is calculated as 0.25 eV. It is also of note that when the variable deflector plate has a voltage 1 kV below the float potential, the same number of ions are detected as when the variable and fixed deflectors are at the float potential. In the two instances, the same proportion of the accelerated beam is detected, but landing on the opposite sides of the detector. The caveat in invoking deflectors is that the alteration of flight path required to maximize sensitivity degrades mass resolution. The kinetic energy of an ion in the deflection field is dependent upon the z coordinate at which it enters the deflectors. An ion which passes closer to a deflector fixed at the float potential will have a lower kinetic energy in the deflection field than one which passes closer to a deflector with a potential which is 500 V below the float voltage. The change in x direction kinetic energy results in an instantaneous acceleration (or deceleration) of ion motion upon entering and exiting the deflection field. The extent of the acceleration is dependent upon the ion's z coordinate. The resultant spread in the arrival time distribution of the ions degrades the mass resolution of the T O F M S . The effect which the z-deflector voltage has on the L I T / T O F M S mass 91 resolution is demonstrated in Fig. 5.3(b). There is a sharp maximum in the achievable resolution when the variable z-deflector plate is set to the float voltage. Obviously, resolution is highest in field free conditions. For the AV — 500 deflector voltage for which maximum sensitivity is achieved the mass resolution is lowered by a factor of two. The problems associated with deflectors in oa-TOFMS have been dis-cussed in de ta i l 4 3 ' 1 5 3 and analyzed both analytical ly 1 5 1 ' 1 5 2 and through de-tailed simulation of ion trajectories.150 A number of instrumental modifica-tions to avoid the detrimental effects of deflectors have been proposed includ-ing shifting the T O F flight tube to match the spontaneous drift angle of the ions , 4 3 ' 1 5 4 and adjusting the location 1 5 5 and angle 1 5 1 of the detector. Park and Koster 1 5 2 have suggested replacing the traditional two plate deflector with a 'multi deflector' consisting of a stack of parallel, electrically isolated plates, such that each deflector pair has a small separation. In the multide-flector, higher deflection angles are possible with smaller AV and thus the effective velocity changes in x in entering and exiting the fields are less pro-nounced. It should be noted that none of these approaches has gained wide acceptance in oa-TOF instruments and none was implemented here. There are, of course, other more simply adaptable methods to minimize the spontaneous drift angle, such as lowering Ez and raising EacCei- When the higher voltages required for increasing EacCei are used, however, due to poor shielding in the L I T / T O F M S any resolution gained by minimizing deflector use is offset by resolution degradation due to stray fields. Therefore the drift angle was lowered by minimizing the ion velocity in the z direction. The potential offsets of the linear ion trap, the repeller plate, and the middle acceleration grid were uniformly set to 0 V . Ion energies in the T O F source are thus dependent on the extent of collisional damping in the axial direction 92 of the quadrupole. 5.3 Performance When the factors outlined in Sees. 5.1 and 5.2 are taken into account, the L I T / T O F M S can be optimized for either sensitivity or resolution. Figure 5.4 demonstrates the resolution and sensitivity achieved on the L I T / T O F M S with the test ion, protonated reserpine. For the sake of comparison, the spectra have been normalized by dividing the raw intensity (number of ions) in each of the 5 ns wide MCS channels by the number of scans coadded in each spectrum. As both spectra were recorded with identical solution composition (1 fjM) and flow rate (1 //L/min), one second of scan time correlates to the consumption of 17 femtomoles of reserpine. Typical scan times in flow-T O F M S were 1-5 s and the raw data of (a) were recorded in 2.5 s, and those of (b) in 5 s. Figure 5.4 (a) shows isotopically resolved reserpine with a resolution (-^)FWHM of 740. Here the system was optimized for resolution. For this spectrum the float potential was -2.0 kV, both sets of deflectors were at the float potential, all lenses (L1-L4) were set to -10 V , and the offsets of the repeller plate and of the middle acceleration grid were 0 V . The peak area divided by the scanning time is 4328 ions/s, or 0.4 ions/scan. The signal intensity was then maximized by increasing the float potential to -3.15 kV and optimizing the potentials on the deflectors to y=-3.15 k V and z=-3.48 kV. The lenses L1-L4 were operated as an einzel lens. The resulting spectrum is shown in Fig. 5.4(b) The mass resolution is decreased to 240 but the ratio of the peak area to the acquisition time is 56,500 ions/s or 5.65 ions/scan, a 13 fold increase in sensitivity relative to the high resolution conditions. The mass range of the instrument was assessed by electrospraying a 93 sueos#/sjunoo suo| jo jaquinN Figure 5.4: T O F M S spectra of reserpine contrasting instrumental parame-ters optimized for maximum resolution and sensitivity. The vertical axis for both plots is the total number of ion counts (in each 5 ns channel) divided by the number of T O F scans, i.e., the number of ions per T O F scan, (a) Flight tube potential -2 kV, no deflectors used, lens stack L1=L2=L3=L4=-10 V . Resolution (-^)FWHM — 740. Peak area/spectral acquisition time = 4328 ions/s. (b). Flight tube -3.15 kV, z-deflector=-3.75 kV, y-deflector=-3.15 kV, L1=L4=-100 V , L2^L3=+50 V . Resolution = 240. Peak area/spectral ac-quisition time = 56 500 ions/s. 94 100 fjM solution of Cs l in acetonitrile. For low mass ions in the RF-only quadrupole, qu defines a sharp low mass cutoff. There is no similar high mass cutoff for the RF-only quadrupole, but transmission is expected to de-crease for the lower qu associated with ions of higher a . 1 5 6 Fig. 5.5 shows Csl clusters with j up to 5000. The peak at ^ ~ 3510 corresponds to the "magic number" ( 2 x 2 x 3 ) cluster, C s ^ I ^ , and has a resolution of 225 and gu=0.194. This spectrum was recorded with parameters maximized for high mass transmission: a high float potential (-3.15 kV), high z-deflector setting (-4.0 kV), y-deflector at the float potential (-3.15 kV), and low potentials on the lens stack (Ll-L4=-20 V ) . In addition, the offsets of the repeller plate and the middle acceleration grid were lowered to -20 V . This change of the lens potentials was necessary to draw the higher mass ions out of the trap. The low mass cutoff of the quadrupole can be seen at ^=750. 5.4 Effect of Trapping on Time of Flight Spectra The purpose of this section is to discuss how the properties of ions in the linear ion trap, as discussed in Chapter 3, can be used to enhance the sensitivity observed on the L I T / T O F M S spectra. This is accomplished by operating the L I T / T O F M S in a trapping mode. The only modifications to the construction outlined in Sec. 5.1 are the addition of timed stopping potentials to L I , the trap exit, and the necessary synchronization between the opening of the trap and T O F M S scanning. For the experiments presented here, there was no controlled flow into the linear ion trap, ions were simply confined in the linear ion trap by the application of a stopping potential to L I and one T O F M S scan per storage cycle was recorded. This is analogous to the injection studies described in Sec. 3.2 as well as to the work of Dresch et al.157 and Senko et 95 o o o o o o o o o o o o CO o o o o CM o o o o A;.isu9iu| Figure 5.5: T O F M S spectrum of C s ^ * . ! exhibiting spectral features from the low mass cutoff (^ ~ 750) to j = 5000. The system was optimized for maximum sensitivity for C s ^ I ^ (^=3510) shown by (*). 96 al}2 for the T O F M S and ICRMS, respectively. The time profiles of the exiting ions are shown in Fig. 3.4 and exhibit points of substantially increased instantaneous ion count rates at short arrival times. If the pulsing to the repeller plate is timed such that the largest portion of trapped ions is pulsed out of the T O F source region, a significant increase in the number of ions detected is realized. This sensitivity enhancement is demonstrated with a storage experiment using ions of cytochrome c, shown in Fig. 5.6. In this experiment, the volt-ages on the T O F M S were identical to those discussed for Fig. 5.4(a) (i.e., resolution optimized) with the exception that Q l was at 5, rather than 10 V . The IQ has an offset to transmit ions, ~ 3.5 V , and LI is held at stopping potential (+12 V) for varying lengths of time. Lowering the L I potential to transmit ions (-10 V) triggers the pulse generator which controls the scan-ning of the T O F M S . When the source region is being filled both the repeller plate and the middle acceleration grid are set to 0 V . The delay between the L I potential drop and the scanning of the T O F M S is varied to determine the time for the densest portion of the trapped beam to reach the accelerating region. There is a single T O F scan for each trapping period. On average, the delay required between lowering L I and the T O F scan-ning was 60 pis and the minimum possible T O F accelerating pulse width was 7 ^s. The delay, however, reflects the time required for the ions to travel from L I to the T O F source region, and is thus highly mass dependent. For these experiments 60 pis was used because it was sufficient time for all charge states to be detected. Low intensities of the high charge states were observed for delay times of 35 pis or less. A delay of 60 pis can be converted, knowing the distance from L I to the center of the T O F source region (31.55 mm), to an ion velocity (for ra=12,200 amu) of 525 m/s and an ion energy of 17.5 eV. 97 » • Trapped Ions Coadded Flow 2 3 Injection Time/ms Figure 5.6: Trapping enhancement of T O F signal. The vertical axis repre-sents the number of ion counts for all observed charge states of cytochrome c in each pulse. Given that the offset between QO and the T O F source region is 5 V , which converts to an ion energy of 50 eV for the +10 charge state of cytochrome c, a significant amount of translational energy loss has occurred. Energy lost per collision for the heavier cytochrome c ion is significantly smaller than that for reserpine ion. Due to its larger cross section, however, it will undergo more collisions. As the z energy calculated by this method is much larger than the 0.3 eV quoted for the reserpine ion in Sec. 5.2, it is evident that cytochrome c ions are less collisionally damped than reserpine. When operated in continuous flow conditions the T O F M S had an inten-sity of 2.2 ion counts per pulse and the intensity which would be realized by the coaddition of these signals for the noted trapping times are represented on Fig. 5.6 by the dashed line. When a stopping potential was applied to 98 L l for the last 40 fis of the 100 /is flight time, the number of ion counts per pulse was found to triple to 6.6. It is important to note that this sensi-tivity enhancement occurs without any sacrifice in T O F M S scanning rate -the system is still operating at 10 kHz. Similarly, if the scanning time was increased to 200 fis (i.e., the repetition rate lowered to 5 kHz) and a stop-ping potential was applied to L l for the last 100 //s of the flight time, the ion intensity increased to 8.5 ions per pulse, which is almost double the 4.4 ions which would be detected in the same time period were no trapping used. Using trapping to create a denser, lower energy ion beam gives overall higher sensitivity than coadding with continuous flow conditions for trapping times of up to 1 ms. In addition, although not completed, this work does demonstrate the possibility for synchronized pulses between the linear ion trap and the T O F source. With careful timing, pulse creation, and high quality ion optics, ion extraction and T O F pulsing could be matched such that the linear ion trap could be operated in a similar mode to the I T / T O F M S , with one trap extraction pulse per T O F pulse, and thus a very high duty cycle could be attained. 5.5 Summary and Conclusions The operation of the L I T / T O F M S can be maximized for high sensitivity or high resolution. When maximized for resolution, ions in the mass range of 22 = 500-1000 have resolutions of « 700. The sensitivity of the instrument is adequate for femtomole level detection in the time scale of one second. Ions with Y of up to 5000 have been detected. Factors independent of the ion-ization source and trapping behavior which affect the intensity distribution in the T O F spectra include the use of electrostatic steering deflectors, the 99 voltages on the lens stack, and collisions with residual gas in the T O F drift tube. Finally, trapping of ions in the linear ion trap can be used to enhance the overall sensitivity. 100 Chapter 6 Resonant Excitation for MS/MS in the LIT/TOFMS The purpose of this chapter is to explain the general methodology of, and present proof of principle for, tandem mass spectrometry in the L I T / T O F M S . As this thesis is organized in chronological order, this chapter discusses the first embodiment of the linear ion trap for M S / M S , with the pressure in the linear ion trap fixed at ~ 7 mTorr N 2 . Results from the second embodi-ment, with the pressure variable from a background of 1.5 mTorr N 2 will be discussed in Chapter 7. Section 6.1 outlines the mathematics behind the damped, forced harmonic oscillator which is used to approximate the trajec-tories of ions when an oscillating auxiliary voltage is superimposed on the quadrupolar potential. The necessary modifications to the linear ion trap and the operation of the L I T / T O F M S for tandem mass spectrometry are described in Sec. 6.2. Section 6.3 describes tandem in space M S / M S for later comparison to the linear ion trap results. Sections 6.4 and 6.5 describe the results for the separate ion isolation and mass selective fragmentation processes. 101 6.1 Resonant Excitation in the Linear Ion Trap As is explained briefly in Sec. 2.1, the motion of an ion in an RF-only quadrupole can be approximated as that of a charged particle in a harmonic potential well. A practical way to understand the motion of an ion in the linear ion trap is to take advantage of available descriptions of harmonic os-cillators to describe the secular ion motion in the pseudopotential well (i.e., macromotion) and consider additional effects on its trajectory (i.e., micro-motion, space charge, scattering from collisions, etc.) as perturbations to the simple harmonic motion. The model used to represent ion trajectories in the linear ion trap (as in the 3-d ion trap and the I C R M S 9 ' 8 2 , 1 5 8 ) is the forced, damped harmonic oscil-lator. If a small oscillating external force is applied to a naturally harmonic system, when the external force frequency approaches the characteristic or resonant frequency of the particle motion, Wo, there is an efficient transfer of energy between the external drive and the harmonic system. The ex-ternal force accelerates the particle, and consequently, the amplitude of the oscillation increases. For the linear ion trap, as discussed in Sec. 2.1 the fundamental angular frequency of ion motion (in the general coordinate u) in the pseudopotential well is and LU0 is related to the frequency / 0 , quoted throughout this chapter, by uo = 2TT/0. (6.2) The key element in Eq. 6.1 is that at a fixed qu u0 is a function of A n external force oscillating at a frequency equal to the resonant frequency of 102 the j of interest provides a mass selective method to increase the amplitude of oscillation of that ion until it strikes an electrode (i.e., eject an ion) and, if this ejection can be monitored, a method of mass analysis (see refs. in Sec. 2.1). Resonant excitation can be explained through the mathematics of the forced harmonic oscillator. In the commonly used "dipolar" excitation mode, the auxiliary voltage source, and therefore the applied force, is attached on opposite poles of the quadrupolar geometry, namely, opposite rods in the linear ion trap, and z end caps in the 3-d ion trap. The connection is such that the auxiliary voltage applied to each pole has identical amplitude and frequency, but is 180° out of phase with that of the opposite pole. The exter-nal force is applied in only one coordinate (u) of the quadrupolar potential. In the description below u = x. In the experimental set up the external force was applied to the x rods. A n ion in a quadrupolar potential with a superimposed dipolar auxiliary voltage, Va = Aa cos uat, where Aa is the zero to peak amplitude in volts, has an equation of motion, -c¥ + W o X = m where, Fa, the force on an ion from V^, is given by 1 5 9 + u^x = — cos(cjat) (6.3) F. = °.%, (0.4) in which ag is a factor which relates to the geometry of the quadrupolar potential, and can be taken as 0.88 for ions near the center of a quadrupole rod set.7 8 The ion trajectory as uja w0 is given by the solution to Eq. 6.3 1 6 0 x(t) = a cos(a;ot -I- S) + a t sin(a;ai), (6.5) 103 where a and S are functions of Fa, u0, and ua. The application oiVa such that c o a - v 0 = 0, will result in the trajectory of an ion with the corresponding ^ having an amplitude which grows linearly in time. The slope of the amplitude increase is F l 2mu>o ' Equation 6.5 describes the principle of resonant excitation in a forced harmonic system. In the linear and the 3-d ion traps, the presence of a neutral bath gas complicates both the trajectories and behavior of the ions. As discussed in Sec. 3.3, non reactive collisions lower the translational kinetic energy of the ion and hence, these collisions act as a frictional, or damping, force in the harmonic system. If the frictional force, Ff, is proportional to the velocity of the ion, (i.e., the faster the ion is moving, the more collisions it undergoes per unit time and hence the greater the Fj) then it can be expressed as „ dx . F, = - c - , (6.6) where the magnitude of the coefficient c determines the extent of the frictional force. In order to add the effects of the frictional force into into Eq. 6.3, a new coefficient 7 = — is introduced, and frictional force and 7 are related by Ff = ~imft. (6.7) The variable 7 , referred to as the damping constant, has the units of inverse time and thus reflects the "rate" at which the frictional force damps the motion of the system. When the frictional force is considered, the equation of motion for an ion moving in a harmonic potential with a bath gas i s i5M6i - i63 d oc doc o F^n / \ , _ _. In the ion trap, the rate at which motion is damped is obviously dependent upon the frequency of ion neutral collisions, which can be calculated by 104 extending the equations introduced in Sec. 3.3. In particular, Eq. 3.15, which expresses the number of collisions which occur when an ion travels a fixed distance, can be used to determine the rate of collisions, v, by replacing I with iv, the relative velocity of the ion (m) and the neutral (M), giving v = nvra. (6.9) To consider the momentum exchanged in each collision, the relative masses of the ion and the neutral must be considered, thus 7 is given by . ^<wL)- <6-io> and termed the reduced collision frequency.158 The use of 7 in Eq. 6.8 requires that the collision frequency be deter-mined such that the rate of collisions be proportional to the velocity of the ion. This assumption becomes problematic, given that the nature of the col-lision, and hence the collision cross section, is also dependent on the velocity of the ion. If the Langevin model of the long range potential between an ion and a spherical neutral is adopted, 1 6 4 ' 1 6 5 then collisions can be catego-rized as spiraling or non-spiraling, referring to the nature of the trajectory of the collision partners. At low translational energies, a spiraling, or orbiting collision occurs when a is given by a = ^ / 2 « . (6.11) VT V A* where CUM is the polarizability of the neutral and n is the reduced mass of the collision partners (j^M)- Substitution of Eq. 6.11 into Eq. 6.10 gives an expression for 7 7 = ^ ) <«•«) M + mJ V A* 105 The above equation has been widely used for describing the effect of ion neutral collisions on the trajectories of ions in both ICR ce l l s 1 5 8 ' 1 6 6 and 3-d ions traps 1 5 9 and is also adopted in this work. It should be noted however, that Eq. 6.12 is valid only when describing orbiting collisions between the ion and the neutral. As vT increases - note from Eq. 6.11 that a very fast ion would have a collision cross section ap-proaching zero - the trajectory is not adequately described as "spiraling". At these collision energies the high angular momentum results in non-spiraling orbits, and the hard sphere model of the cross section 1 6 4 must be used. When substituted into Eq. 6.10, the hard sphere cross section results in a 7 which is dependent upon ion velocity. Since in this case 7 becomes a function of Ff is proportion to the square of the ion velocity. Consequently, for hard sphere collisions, Eq. 6.8 fails to adequately describe the dynamics and nu-merical techniques must be used to calculate trajectories. The description of the trajectories discussed below is valid only if the majority of collisions induced by resonant excitation can be described as orbiting. As noted above, this assumption has been successfully used in simulations of the ion trap and ICRMS. As an aside, the / 0 for the triply charged renin substrate ion used for most studies in the linear ion trap was ~ 200 kHz, and 7 for this ion with 6 mTorr N 2 is ~ 5.3 kHz. As 7 >C UQ, the motion of an ion moving through the linear ion trap when Va = 0 is described by the underdamped harmonic oscillator. In such a system the amplitude of the oscillations decays expo-nentially in time - an alternate explanation of the phenomenon of collisional damping. 1 0 9 , 1 4 2 The literature for quadrupole traps, which provides a full solution to Eq. 6.8, focuses on a unique method of operating the 3-d ion trap in 106 which low amplitude auxiliary excitation is used to slowly scan frequency space . 1 5 9 , 1 6 1 - 1 6 3 When the resonant frequency of an ion is approached, it is ejected from the trapping volume and detected, achieving a mass resolution as high as ~ 1.6 x l O 6 . 1 6 7 The solutions presented in these prior studies in-corporate a frequency scanning rate which is not used in the linear ion trap, and hence, a separate solution to Eq. 6.8 is provided here. Following the text of French, 1 6 8 the solution to Eq. 6.8 is constructed from the sum of two linearly independent solutions, x(t) = x1(t)+x2(t), (6.13) where X\{t) is a solution to the forced, damped harmonic oscillator in the "steady state", and x2(t) is the solution to Eq. 6.8 when no auxiliary force is applied to the system (i.e., Fa — 0). The solution to x\{t) is x\(t) = acos(u>at — 6) (6.14) where Fa/m and tan 6 - 7—5 — . K - ul) The structure of x2{t) is x2(t) = be~ 2 cos(a;it + /3), where u>i is the frequency of damped oscillations without forcing, (6.18) (6.15) (6.16) (6.17) 107 In typical conditions for the linear ion trap, u>0 ^> 7, and thus u>\ ~ UJ0. Substituting Eqs. 6.14 and 6.17 into Eq. 6.13 yields x(t) = acos(u>at — 5) + be~ 2 cos{uit + /?), (6.19) where a and S are given by Eqs. 6.15 and 6.16, and b and B can be calculated for given starting conditions. The requirement that initial conditions be known to solve Eq. 6.19 neces-sitates that a few assumptions be made about ions confined in the linear ion trap. The first assumption is that prior to the application of the auxiliary voltage, the ion is directly in the center of the rods. While not precisely cor-rect, simulations of ion trajectories for RF-only quadrupoles at 7 mTorr N 2 have demonstrated very strong spatial focusing of the beam. 6 1 ' 1 1 1 ' 1 4 2 Solving Eq. 6.19 with the constraint that x(0) = 0 yields, b= -a - . 6.20 cos p The second assumption, while based on the same principles, is more prob-lematic. Collisional damping also results in near thermal ion energies.61 Even though ions are not starting from rest, however, it is reasonable to approx-imate the initial velocity of an ion as zero if the velocity resulting from the excitation is much greater than the starting velocity. If at t=0, 4s. — 0, then the derivative of Eq. 6.19, along with Eq. 6.20, relates B to c5, 7 ui tan/? + a;ataii(5 = ——. (6.21) Further substitution of Eq. 6.16 into Eq. 6.21 gives t a n / ? = JL(4±4V (6.22) From Eqs. 6.22, 6.16, and 6.21, Eq. 6.20 reduces to b=-a~. (6.23) 108 Finally, the full expression for the macromotion of an ion excited by the oscillating external potential, Va, is The ion trajectory is the sum of two cosine waves, the "steady state" solution, Xi(t), which has a constant amplitude, and the "transient" solution, a noticeable effect on the trajectory, and at long times, the motion is fully and exclusively described by X\(t) - a cosine oscillation having the frequency of the applied force, ua. The decay time of the transient solution is determined by the frictional force, through the 7 coefficient. For the typical 7 ~ 5.3 kHz quoted above for the linear ion trap, the contribution of the transient solution to the trajectory vanishes as / approaches 1 ms. Sample trajectories from Eq. 6.24 are shown in Fig. 6.1. The units of x(t) are meters, and it is immediately apparent that the calculated trajectories extend beyond the physical r 0 of the linear ion trap. It should be empha-sized that the desire was to analyze the qualitative behaviour of resonant excitation and not perfectly model the experimental setup. A more complete description of x(t) would require a more thorough consideration of the ion's initial position and velocity, as well as incorporation of micromotion into the model. The trajectories were calculated with the aid of a F O R T R A N pro-gram which provided x(t) for the user entered variables of trap pressure, P , A, /o (and thereby qx), Aa, m, e, and M. In addition to the trajectory, x(t), the output of the program included the values for the coefficients a, 7 , 8, and and verified the equality in Eq. 6.21. For the trajectories displayed in Fig. 6.1 typical excitation parameters for the linear ion trap were entered into the program. The precursor ion was Fa/m cos(uat + 8) - —e~2 cosfwx + /?)]. (6.24) which has an amplitude that decays in time. At short times, the transient has 109 Figure 6.1: Ion trajectories calculated from Eq. 6.24 for fa of (a) 197, (b) 199.5, and (c) 200 kHz. The precursor ion is the +3 charge state of renin substrate, / 0 = 200 kHz, Aa = 1 V , and the bath gas is 6 mTorr N 2 . 110 the 4-3 charge state of renin substrate \Jj = 587J, the bath gas pressure was 7 mTorr N 2 , and / 0 = 200 kHz (g u =0.566). The following description focuses on the effects of varying ua. The amplitude of the auxiliary voltage, Aa, was fixed at 1 V . It should be appreciated, however, from Eq. 6.24, that x(i) is linearly related to Fa and by extension Aa, hence the only consequence of altering Aa is to shrink or expand the vertical axes of the plots. For the quoted bath gas pressure, 7 is 6.2 Hz. The differences in the trajectories plotted in each panel reflect the con-trasting motion of ions excited by fa of (a) 197 kHz, (b) 199.5 kHz, and (c) 200 kHz. For (a) and (b), fa ^ fo, and consequently the short time trajec-tory, when the transient solution is contributing to x(t), is a superposition of two closely moving sinusoidal oscillations of similar frequency, with, as is typical of "beat" motion, alternating periods of constructive and destructive interference. The points of maximum constructive (or destructive) interfer-ence are separated in time by JJ^ZJ^- l n panel (a) fa — fo = 3 kHz, and the time between beats is by 0.333 ms. For this uja: 8 is 0.801 radians, /? is -0.801 radians, and a is 0.0120. Figure 6.1 (b) plots x(t) for fa = 199.5 kHz. The values of 6, (3, and a are 1.39, -1.39, and 0.0192, respectively. There are two salient features of Fig. 6.1 (b). First, the time between points of maximum interference ( 0 5 x k H z = 0.002 s) is longer than the time scale of decay for the transient solution; no beats are observed. Second, while the steady state solution for panel (a) has a lower final amplitude than that of panel (b), at short times, the instantaneous x(t) of panel (a) is actually larger than any x(t) of panel (b). The maximum amplitude in both the steady state and transient solutions is a function of the coefficient a from Eq. 6.15. From Eq. 6.15, it is evident that the coefficient a has a Lorentzian lineshape centered on u>o and thus the steady state amplitude of x(t) reaches 111 a maximum for resonant excitation. The correlation between the value of a and the maximum x(t) in the transient solution, however, is complicated by the fact that the amplitude of constructive interference can reach 2 x a. Consequently, unless the amplitude, a, for excitation at u>a is less than one half that at u>0, the short time amplitude of the off resonance excitation will be greater than any x(t) achieved in the resonant system. Fig. 6.1(c) plots x(t) for resonant excitation (/ a=200 kHz). If ua — UJQ and the relative values of 7 and OJQ are such that u>i « u)0, then Eq. 6.24 reduces to x(t) = —— sin(u;ot)[l - e - £ ] . (6.25) The amplitude of the steady state solution is given by Fajmju;0 and the rate at which the steady state is reached is proportional to 7 . In Fig. 6.1 (c) 8 = 8 = f, and a was 0.017752. The effects of varying 7 and Fa on the trajectories will be discussed in Chapter 7. Some comments on the "perturbations" to the supposedly harmonic sys-tem are necessary. The above equation and the resultant trajectories describe only the macromotion of the ions: the effects of micromotion resultant from the rapidly oscillating RF-drive are ignored. Paul provides a full solution to forced, Mathieu equation without damping. 7 2 In the linear ion trap, however, because of the relatively high neutral gas pressure, it is reasonable to assume that the effects of collisional damping will have a much greater influence on the ion trajectory than those of the micromotion. One plausible effect of micromotion is that a calculated x(t) may be less than x0, and hence the-oretically within the trapping volume, but the addition of the micromotion will result in an ion striking the rods. It is also important to remember that the entire description is based upon the pseudopotential approximation and 112 thus fails if fix > 0.4. The greatest difference between the plotted trajectories and the real sys-tem, however, derives from collisions with the neutral gas. The derivation of Eq. 6.24 assumes the only effect of collisions is unidirectional momentum damping, and that the only consequence of increased amplitude of oscillation is loss from the trap when x(t) > 7*0. Scattering from collisions couples the excitation energy between coordinates. Therefore, even though dipolar exci-tation is applied in only one coordinate, collisions will result in the transfer of translational excitation between all directions of ion motion. More im-portantly, in the real system, the collisions may transfer kinetic energy to internal energy of an ion. For large organic ions this can be quite efficient.169 With multiple sequential collisions, the internal energy of the ion can reach the critical energy of one or more fragmentation channels, resulting in the dis-sociation of the ion. The details of collisional excitation have been reviewed by Cooks. 1 7 0 Some understanding of the relationship between the trajectory, the num-ber of collisions, and the detection of the fragments of the precursor is re-quired. In the presence of the bath gas, an increase in the amplitude of the ion trajectory creates two competing scenarios. First, the application of the excitation voltage can result in a rapid increase in the amplitude of oscil-lation and consequently, an ion is ejected from the trapping volume before the fragmentation occurs. Second, prior to ejection, a sufficient number of collisions to induce fragmentation occurs and the resulting fragments main-tain stable trajectories in the quadrupolar potential. The interplay between these processes is termed the competition between fragmentation and ejec-tion. In general, conditions for ejection will be favored by a rapid increase in the amplitude; fragmentation by a more gradual growth. The fraction of 113 kinetic energy exchanged per collision also affects the time required for an ion to fragment. The parameters which govern this competition are those in Eq. 6.24 which define the rate of increase in x(t) - namely, the frequency of the ion's motion prior to the application of the excitation voltage (u0), the parameters defining the applied external field ( A a , w a , and t), and the mass and density of the bath gas (7 ) . In addition, fragmentation depends on factors which are not easily quantified, or may not be known, such as the structure and fragmentation channels of the precursor ion, the effectiveness of the transfer of collisional energy to internal energy, and the time scale of the unimolecular dissociation of the precursor ion. The competition between ejection and fragmentation in the 3-d ion trap has been studied by Charles et al. 171 and will be discussed for the linear ion trap in Sees. 6.6 and 7.6. Whereas using resonant excitation to eject ions from a trap has a rel-atively long h is tory , 1 0 ' 7 3 ' 1 7 2 ' 1 7 3 the concept of applying the same technique for the fragmentation of ions via collisions with a neutral bath gas was first applied to a 3-d ion trap by Louris et al. 174 in 1989. It was rapidly adopted both as a research and an analytical tool. Several reviews on tandem mass spectrometry in the 3-d ion trap are available. 1 0 ' 1 7 5 ' 1 7 6 In addition, detailed numerical simulations of ion motion in the 3-d ion trap have been com-pleted. 1 7 7 " 1 8 0 As there were literally no references which describe the sequential isola-tion and fragmentation of ions trapped in a linear RF-only quadrupole, the known behavior of ions in 3-d traps was used as a guide for the examination of the effects of resonant excitation in the linear ion trap. It is necessary, however, to highlight the differences between the 3-d and linear ion traps. In the 3-d ion trap, resonant excitation is applied in the z coordinate, and the 114 maximum amplitude of oscillation, without ejection, is typically 7.07 mm. In the linear ion trap, the same parameter is 4.00 mm and consequently, for identical Fa, fewer excitation cycles (and less time), are available for frag-mentation before ejection. In the 3-d ion trap, 2z\ — TQ, and Dz — | D r . For identical qz and qx or qy, the maximum possible energy exchanged in a single collision in the linear ion trap can be twice that in the 3-d ion trap, 6 6 which can result in fragmentation occurring on a faster time scale, and access to higher fragmentation channels. A further difference is in the collision gas. Typically, a 3-d ion trap is operated with 1 mTorr He, as opposed to the 7 mTorr of N 2 in the linear ion trap for the work described in this chapter. The rate of collisional damping is thus significantly lower in the 3-d ion trap. 6.2 Experimental Methods To superimpose the auxiliary drive voltage on the quadrupolar potential, the connection between the main R F drive and the rod sets had to be modified as shown in Fig. 6.2. Normally, opposite rods are electrically connected; the main R F drive has two outputs, corresponding to the x and y rod pairs. When adapting the linear ion trap for tandem mass spectrometry, there is no alteration to the connection between the main drive and the y rods. For the x poles, however, the electrical connection between the two rods is dis-connected, and the output from the main drive is connected to the center tap of the secondary winding of a bipolar transformer (1:1 turns ratio). Each of the secondary leads is connected to an electrically isolated x pole. Dipolar excitation is applied only in the x direction. The primary winding of the transformer is connected to the second channel of the arbitrary waveform generator (PC Instruments AWG-312, Twinsburg, OH, U.S.A.), used to cre-ate the voltage sequences for resonant excitation. As the waveform generator 115 AUX DRIVE A_a(max) =30 V to - 200 kHz X 271 MAIN RF DRIVE V_rf(max) = 5000 V a = 1 MHz x 2TC Figure 6.2: Schematic of the coupling of the auxiliary drive (aux drive) to the quadrupole rods. The auxiliary drive consists of the output of the arbitrary waveform generator passed through a 2.5 x step up transformer. provided a maximum 0 to peak amplitude of 12 V , an additional transformer with a 2.5:1 step up voltage ratio was placed between the output of the gen-erator and the bipolar transformer such that 0-30 V peak amplitudes could be applied to the quadrupole rods. A l l values quoted in the remainder of the text refer to the amplitude of the potential applied to the rods. In the first quadrupole, QO, the x and y poles are connected directly to the main RF-drive through 12.5 pF capacitors; resonant excitation is applied only to QI . With few exceptions the M S / M S capabilities of the linear ion trap were 116 assessed by examining the fragmentation of the +3 charge state of renin sub-strate ( D R V Y I H P F H L L V Y S ) . This is a small peptide with a rich fragmenta-tion pattern which has been used in the characterization of the performance of M S / M S systems.1 8 1 The electrospray solution was 5 /JM renin substrate in a solution which was 50% water, 45% methanol, and 5% glacial acetic acid. The voltages on the components of the L I T / T O F M S were identical to those described in Chapter 5 for high resolution T O F operation. The volt-ages on the sprayer, orifice, and curtain varied depending on electrospray conditions. The voltage on the skimmer was 15 V , the Q0 offset 10 V , and the Q l offset 0 V , and the offsets of the repeller plate, and the middle accel-eration grid were 0 V . Ions thus entered the linear ion trap with energies of 10 eV per charge. The pulse sequences for a typical tandem mass spectrometry scan are shown in Fig. 6.3. The clock controlling the pulsing of the numerous compo-nents of the L I T / T O F M S is provided by the arbitrary waveform generator. As shown in Fig. 6.3, the M S / M S cycle takes 20 ms to complete. It in-volves changing the potentials on the entrance (interquad, IQ, in Fig. 6.2) and exit (Ll) apertures, control of the excitation voltage (Va) output to the quadrupole rods, denoted on the figure by "aux drive" (for auxiliary driver, which includes the arbitrary waveform generator, the stepup transformer, and the bipolar transformer) and the T O F pulsing and detection (TOF). The duty cycle for tandem mass spectrometry in the L I T / T O F M S is identical to that discussed for the I T / T O F M S in Sec. 2.5. The total duty cycle can be separated into the duty cycle for the T O F M S (see Sec. 5.1), and M S / M S respectively. For the timing parameters shown in Fig. 6.3, the ratio of T O F detection time to the total cycle is 7 0 s c a £ 0 s x ^ s 0 1 m s = 3.5%, 0.01 ms being the time to fill the T O F source region. The duty cycle for the linear 117 INJECTION IQ AUX DRIVE L I T O F 5 ms ISOLATION 'A V////A 4 ms EXCITATION [TRAP EMPTY/TOF SCAN 4 ms TRAPPING TrUXTLOrLTL 7 ms 70 TOF SCANS 20 ms MS/MS CYCLE Figure 6.3: The timing parameters for a typical 20 ms M S / M S scan. IQ is the entrance aperture of the linear ion trap, and L I , the exit aperture. These apertures control the flow of ions into and out of the linear ion trap. The auxiliary drive controls the coupling of the arbitrary waveform generator to the quadrupole rods and is used for ion isolation and excitation. ion trap is the ratio of sample fill time, to the total cycle, 2 5 0™ s s = 25%. The overall instrumental duty cycle is the product of these two duty cycles, i.e., 25% x 3.5% = 0.88%. The first and last phases of the M S / M S cycle correspond to the injec-tion and extraction procedures discussed for the collision cell of the triple quadrupole mass spectrometer in Chapter 3. The offset of Q0 was 10 V , of Q l , 0 V . For ion injection, a sync-out pulse from the waveform genera-tor triggers a separate pulse generator (AVTech AV-1002-C, Ottawa, O N , 118 Canada) which controls the potential on IQ, the entrance to the trap. IQ is maintained at a potential to pass ions (~ 7 V) for a set injection time (typically 5 ms as shown in Fig. 6.3) and a stopping potential (12 V) for the remaining 15 ms of the scan. The last phase of the M S / M S cycle is the extraction of ions from the linear ion trap and the subsequent detection of fragment ions. As shown in Fig. 6.3, L l , which is controlled by channel one of the arbitrary waveform generator, is held at a stopping potential (+12 V) for the first 13 ms of the M S / M S scan and at a potential to transmit ions (-10 V) for a set extraction time, typically 7 ms. The other lenses in the coupling stack, L2-L4, are also held at -10 V , and the offsets of the repeller plate and the middle acceleration grid are at 0 V . In addition, channel one of the arbitrary waveform generator gates the pulse generator used to trigger the T O F repeller plate pulsing and detection electronics. Consequently, only when the trap is being emptied are T O F scans acquired. As the T O F scanning rate is 10 kHz, there are 70 T O F scans for each extraction. A few comments on the details of the raw data are required. The majority of the fragmentation spectra from the L I T / T O F M S were collected with the channel width of the multichannel scalar set to 20 ns. A delay, td, was often used between the time the pulse was applied to the repeller plate and the start of MCS scanning. The delay corresponds to the flight times for the ions with ^ corresponding to the low mass cutoff of the linear ion trap. The ion flight time, tj, in / J S , is related to the channel number, p, by tf = 0.02p + td (6.26) and to s by — = .Ci * (tfY + c 2 . (6.27) 119 Typically the float voltage, the z, and the y deflectors were set to -2.0 kV and the calibration constants were Ci=0.345 and C2=0. The choice of injection and extraction periods was aided by the under-standing of the extracted ion energy distribution provided by the prelimi-nary studies on the triple quadrupole mass spectrometer (see Sec. 3.4). To maximize sensitivity, a time short enough to ensure ions are not lost in the processes of injection and extraction (i.e., less than 1 s) must be used. Because of the orthogonal coupling of the T O F acceleration to the linear ion trap, the axial profile of the ion beam (as exhibited for various param-eters in Fig. 3.4) is "chopped" into "pulses" of ions accelerated towards the detector. The effect of this rastering of the beam can be visualized by imag-ining that the time profiles of each panel in Fig. 3.4 is segmented into vertical strips of width 10 fus. In the L I T / T O F M S only the first of every 10 segments wil l be detected. The T O F sensitivity reflects the coaddition of the number of ions in each segment. The pulse corresponding to the initial segment will, of course, contribute more to the total number of ions detected than the last pulse collected. For this reason, short time coaddition of segments with higher count rates can result in a higher sensitivity than long time coaddition of pulses with lower ion count rates. A few additional comments on the injection and extraction times should be made. The injection time serves as a thermalization, or cooling period. For the ions that were studied, fragmentation spectra were independent of orifice skimmer potential difference, suggesting that any ion heating in the ion sampling region equilibrated during the injection period. The. choice of 7 ms for the T O F scanning/trap extraction was influenced by the necessity to record spectra which reflect the true trap population. If shorter exit times are utilized, the total number of ions counted is greater. 120 However, scanning cannot be so short that the extracted ions do not reflect the trap population. Ions of different mass and cross section take different times to leave the trap. Lighter, smaller ions diffuse through the trap and enter the source region of the T O F M S more rapidly than heavier, larger ions. If too short an extraction time is used, the resultant spectra will exhibit an artificially large fragment ion population. For instance, with a 5 ms injection time and a 2 ms extraction time, the ratio of all detected fragment ions to singly charged resperine precursor ions is 4. For exit times from 7 ms to 50 ms the same ratio is ~ 2.5. The additional 3 ms are required to allow full diffusion of the precursor ions from the trap. Were an axial field (mentioned in Sec. 3.4) applied to the linear ion trap, then both a shorter time would be required to empty the trap and the number of ion counts in each T O F pulse would increase. Between injection and extraction, ions are stored in the trapping volume as the ion isolation and fragmentation processes are completed. During this time, both IQ and L l are held at stopping potentials (12 V ) . As discussed and demonstrated in Sec. 3.5, trapping efficiency is near 100% on these time scales and thus there is no need to consider trapping losses for M S / M S in the L I T / T O F M S . The optimal implementation of the isolation and fragmentation processes required preliminary experiments in order to determine the best parameters to use in the final pulse sequences. The procedure is briefly outlined here. The first step in the optimization experiment was the determination of the resonant frequency of the ion of interest. The u>o was calculated from Eq. 2.14. The excitation voltage, Va, applied to the rods was created using the sinusoidal waveform provided with the software of the arbitrary waveform generator. The variable parameters determining Va were frequency, ua = 121 27r/a, amplitude, and time of application. Fig. 6.4 plots the observed precursor and sum of fragment ions as a func-tion of fa for a typical preliminary excitation. The plot is termed an "ex-citation curve", and the half height width of the absorption, A / , is used in calculating the mass resolution of resonant excitation, R, by the relation-ship 1 5 9 R=J± = A ™ . (6.28) Aw A / A m v ' This equality is derived from the pseudopotential approximation, and holds only as long as OJO oc ^ . By scanning fa near the center of the observed curve, and sequentially lowering the amplitude of the excitation voltage, an optimal experimental value for f0 (i.e., that which promoted the most fragmentation) could be attained. The fragmentation process is explained in greater detail in Sec. 6.5. Once the optimal f0 was determined, the tailored waveform used for pre-cursor ion isolation could be constructed. The waveform is based on the "sum of sines" approach to broadband isolation waveforms. 1 8 2 ' 1 8 3 In fre-quency space the waveform spans from 10 kHz to 500 kHz, except for a "notch" on and near / 0 , with a "comb" of 980 equal amplitude sine waves. The corresponding time domain waveform is created by a separate software program, provided by S C I E X . 1 8 4 The parameters under user control included the minimum and maximum frequencies of the notch and the full amplitude of the waveform. A typical notch in the waveform is 2-10 kHz wide and cen-tered on the experimentally determined resonant frequency of the precursor ion. The maximum amplitude of the waveform creates a 30 V (0 to peak) excitation on the rods. The output of the program is a 10,000 line file. Each line in the file represents the voltage output (i.e., amplitude of the waveform) for a discrete time step, determined by the "clock" setting of the generator. 122 198 200 202 204 206 faI kHz Figure 6.4: A typical excitation curve obtained in determining optimal con-ditions for M S / M S . For all experiments the clock was set to 2.5 MHz, thus the separation be-tween consecutive sine waves in the frequency domain is 500 Hz. In the time domain, each line in the file represented 0.4 /is, and 4 ms were required to apply the complete tailored waveform. The ion isolation step illustrated in Fig. 6.3 shows the notch in time, although it will be appreciated that in reality it is in frequency space. The results from the isolation process are discussed in greater detail, in Sees. 6.4 and 7.4. The fragment spectra were transferred from the raw MCS format to an x/y format using software provided by E G & G Ortec. The x/y files were imported to the Galactic G R A M S / 3 2 Version 5.01 software package. Quoted intensities refer to peak heights, rather than peak areas. A n early series of 123 data were reviewed by fitting all of the relevant spectral features twice -once to Gaussian lineshapes using the areas as intensities and once using the peak heights of the raw data for intensities. The results of interest in understanding and optimizing the linear ion trap (i.e., mass selectivity and fragment efficiency), were identical within experimental uncertainty for the two parameterizations. Hence, for simplicity, peak heights were use in the remainder of the work. Peak heights and locations were determined by the peak finding algorithm of the GRAMS/32 software. 6.3 Tandem in Space Mass Spectrometry The energetics of the fragmentation channels for renin substrate (^=587 (M+3H) 3 + ) are not well characterized and hence, the relative energies of the various fragmentation channels were assessed qualitatively by completing tandem in space mass spectrometry studies on the triple quadrupole mass spectrometer (described in Sec. 3.2) with 1 mTorr krypton as the collision gas. The offset of QO relative to the collision cell, Q2, was sequentially increased from 10 V to 35 V in increments of 5 V . The resultant spectra are plotted in Fig. 6.5. At 10 V offset, no discernible fragments are evident in the spectrum, whereas at 20 V all of the precursor has fragmented, and as the axial energy is increased, the observed fragmentation pattern changes. There is a fragment at ^=582 detected with collision cell offsets of 20 and 25 V which should not be confused with the precursor (-^=587). The initial center of mass collision energy, Ecoa, at each potential differ-ence can be calculated from the laboratory axial energy, E^, by M Ecoii = Eiab — — 7 7 , (6.29) m + M where M is 83.8 amu and m is 1761 amu. The fragmentation patterns from 124 'co (a) 10V (b) 15 V L A I . II . . » (c) A A. 20 V . . . . . . J L . (d) j I I l i t 25 V , ! , ! , , , (e) 1 A A J II L All k .. L j 30 V II (f) I J 1 35 V ».. — J A 0 500 1000 1500 m/e Figure 6.5: Results of a tandem in space mass spectrometry study on the +3 charge state of renin substrate with 1 mTorr krypton as the collision gas. The offset between QO and the collision cell is shown in the upper right hand corner of each panel. 125 the tandem in space system are compared with those produced from the tandem in time mass spectrometry to assess which resonant excitation con-ditions produced fragments corresponding to higher collision energies. This technique has been used previously in estimating the relative energies of fragmentation channels observed in a 3-d ion t rap. 6 5 ' 1 8 5 6.4 Ion Isolation Ion isolation in the linear ion trap, with 7 mTorr N 2 , is demonstrated in Fig. 6.6 (a) and (b) which show an electrosprayed solution of 3.75 pM renin substrate and 0.75 pM reserpine before (a), and after (b), the application of the notched broadband waveform with a 30 V amplitude for 4 ms. The timing parameters for this experiment are identical to those shown in Fig. 6.3, except no fragmentation step was used. The notch was designed to isolate the +3 charge state of renin substrate ( j = 587) with qu m 0.623. The notch spanned 217 kHz to 224 kHz and /o for the precursor ion was calculated from Eq. 2.8 to be 220 kHz, which gives (from Eq. 6.28) a resolution of 30. The notch must include all frequencies of the excitation curve for which any decrease in precursor ion intensity is detected. Accordingly, the notch in the broadband waveform will always be wider than the A / of the excitation curve, and thus, the mass selection of precursor isolation is, by definition, worse than the excitation resolution, R. While the broadband waveform does eliminate the greater majority of the targeted ions, leaving the precursor, there are still undesired spectral features discernible in Fig. 6.6 (b), notably at ^ of 786, 885, and 1030. One problem with the ion isolation procedure was the broadband wave-form causing precursor ion fragmentation. The component frequencies of the waveform must eject, and not fragment, the precursor (and other ions). As is 126 4!> A } j S U 3 } U | Figure 6.6: Isolation of the +3 charge state of renin substrate = 587) through the use of the notched broadband excitation waveform, (a) Spec-trum of a mixture of renin substrate (3.75 /iM) and reserpine (0.75 (JM) trapped for 4 ms without application of the broadband excitation waveform, (b) Spectrum recorded after 4 ms application of the broadband excitation waveform with a notch at — 587 (amplitude 30 V 0-peak, notch 217-224 kHz). 127 evident in Fig. 6.4, high amplitude, near resonant, excitation induces signifi-cant precursor dissociation, meaning that if the component frequencies of the broadband waveform are close to the characteristic frequency of the precur-sor ion, fragmentation may result. The peaks at & 786 and 1030 in Fig. 6.6, correspond to fragments observed in Fig. 6.5 and these intensities actually increase after the application of the broadband waveform. In order to min-imize such fragmentation, it was necessary to create a broader notch than would normally be required, subsequently degrading the mass selectivity of ion isolation. Lowering the amplitude of the broadband waveform decreased the extent of fragmentation, as well, but at the expense of efficient ejection at high rj. The converse was observed with increasing application time of the broadband waveform; fragment peak intensities increased. To eliminate all fragmentation an even wider notch, (i.e., ~ 10 kHz) is necessary. It should be appreciated however, that Fig. 6.6(b) demonstrates a successful isolation and the low level of contaminant ions should not affect the M S / M S process. A second problem associated with the isolation process is that the broad-band waveform is constructed with the unreasonable assumption that all ions irradiated for a set time with an identical excitation amplitude will be ejected from the trap with equal efficacy. In Fig. 6.6, the peak at ^=886 lowers in intensity, but does not vanish, after the application of the broadband wave-form. It was found that the elimination of ions with higher ^ exhibited strong dependence on their qu. As the amplitude of the R F - voltage to the rods (Vrf) increased, ions with a high ^ were eliminated. The choice of the relatively high qu in Fig. 6.6 resulted from these experiments. Several causes for this observation were hypothesized, none correct. Any pseudopotential based explanation for the behavior fails. Ions with higher ^ than the precursor have lower qu and, correspondingly, lower uQ. From 128 Eq. 6.24, for a fixed excitation amplitude, the rate of increase in the ampli-tude of oscillation of an ion will be greater for a lower o>0, hence in theory, ions with small qu are ejected with lower amplitudes. 1 7 1 Another idea ex-plored was that since higher mass ions have lower resonant frequencies, they experience a lower number of excitation cycles, which could decrease the ejec-tion efficiency. Increasing the application time of the waveform did somewhat decrease the intensity of the high ^ ions, but the complete elimination of all ions with the same effectiveness as exhibited for low — was never observed. e It could be that the problem is not an inherent feature of the linear ion trap, but rather a result of a particular component of the L I T / T O F M S , such as the step up transformer, which was not as effective in amplifying low frequency components of the broadband waveform. A further possibility is that the ions with a higher ^ have different rates of collisional damping, and correspondingly, diffuse more slowly from the trap. Similarly, higher mass ions have higher 7 values, hence from in Eq. 6.24 will have lower x{t) values for a given Fa and t. Use of longer extraction times did not eliminate the problem. As will be demonstrated in Sec. 7.4, both of the above mentioned prob-lems, (i.e., incomplete ejection and unwanted fragmentation) were eliminated when the pressure in the linear ion trap was lowered. However, better qual-ity isolation at any pressure could be realized with improvements in the construction of the tailored waveform. In particular, a lower voltage could be used for the fa near / 0 which should eliminate near resonant fragmenta-tion. Additionally, higher amplitudes could be used for the lower frequencies, facilitating high ^ ejection. Similarly, only one method of isolation, a broadband waveform, has been used here. For the 3-d ion trap, there are a number of different isolation 129 procedures such as stored waveform inverse Fourier transform (SWIFT) and chirp waveforms, and combined R F and DC isolation which are used, 1 8 2 ' 1 8 6 and all of these could be applied to the linear ion trap. "Colored" waveforms which give the ability to isolate two (or more) precursor ions could also be employed. 1 8 7 6.5 Ion Fragmentation Resonant excitation in gas filled RF-only linear quadrupoles has been demon-strated in two previous studies, 6 6 ' 6 7 neither of which used trapping or precur-sor selection. Maintaining the ions in the rod set as the resonant excitation is applied permits a greater control of the nature and magnitude of the external force. One objective of this work was to determine the nature of the frag-mentation process in an RF-only quadrupole, using the well known behavior of CID in 3-d ion traps 1 7 1 ' 1 7 4 as a guide. The resonant frequency of an ion can be calculated from Eq. 2.14 to an ac-curacy of 1%, provided that qu is less than 0.6.1 0 Any difference between the calculated and experimental resonant frequencies could be indicative of the presence of higher order electric fields or perturbations from space charge effects. In the parameters for CID quoted in this chapter, no significant shifts between calculated and experimental resonant frequencies were ob-served. The success of Eq. 2.14 in predicting resonant frequencies at least somewhat justifies the continued use of the pseudopotential well model for the interpretation of resonant excitation phenomena. Fig. 6.7 shows the raw data for an M S / M S experiment, and demonstrates the variation in fragment spectra of renin substrate ions as the frequency of the auxiliary voltage is changed near u0 (qu « 0.623, thus / 0 ~ 220 kHz). The +3 charge state 587) was isolated via application of a broadband wave-130 form for 4 ms and subsequently irradiated for another 4 ms by an auxiliary voltage with a 1.5 V amplitude. As expected, the intensity of the precursor ion falls as the fundamental frequency of motion is approached and rises again for fa > f0. The fragmentation pattern observed for these excitation param-eters closely resembles the spectrum recorded in the tandem in space system for an axial injection energy of 45 eV (see offset of 15 V in Fig. 6.5(b)) with the notable distinction that precursor elimination is complete in the tandem in time data. This observation is consistent with the general rule that tandem mass spectrometry from resonant excitation typically accesses lower energy fragmentation channels than those available through triple quadrupole mass spectrometers.65 The data of Fig. 6.7 are summarized in Fig. 6.8(a), which plots the in-tensity of the precursor ion and sum of fragment ions against the frequency of the auxiliary voltage. The half height width of the excitation curve in Fig. 6.8(a) is 3.0 kHz, hence the mass resolution, R, is 73. One desired capability of the L I T / T O F M S is a high fragmentation ef-ficiency, indicative of both the ability to dissociate a large fraction of the precursor ions and the capacity to successfully store and detect the resultant fragment ions. As is evident from Fig. 6.7 and Fig. 6.8(a), resonant excitation induces a significant degree of dissociation, but without adequate knowledge of the fragment charge states for +3 renin substrate, the true fragmentation efficiency cannot be calculated. Fig. 6.8(b) plots the precursor and fragment ion intensities as a function of auxiliary voltage frequency for the excitation of the +1 reserpine ion. The isolated precursor at ^ = 609 (gu « 0.51, / 0 ~ 181 kHz) was irradiated with an auxiliary voltage of 2.4 V (0 to peak on the rods) amplitude for 4 ms. The mass resolution of resonant excitation is ~ 60. The fragmentation efficiency can be calculated from the ratio of the total 131 I Figure 6.7: Fragmentation spectra of the isolated +3 charge state of renin substrate. Fragmentation was induced by a 4 ms period of resonance exci-tation with amplitude 1.5 V and qu = 0.623. The horizontal axis is channel number where channel 0 is 30 fis and each channel represents 20 ns. The vertical axis shows the number of ion counts in each channel. Spectra are shown for / a=219-223 kHz. 132 fragment ion intensities to the precursor ion intensity prior to the application of the excitation; for reserpine (Fig. 6.8(b)) this is ~ g|g x 100 = 60%. Due to the relatively high qu setting for the precursor ion (^=609), the calculated low mass cutoffs are at relatively high rj. For instance, for the data plotted in Fig. 6.8(b) with qu = 0.51 for f=609, only ions with ^ > 342 will have a qu of < 0.908, and thus stable trajectories in the linear ion trap. Lower mass fragments will have unstable trajectories and hence, even if formed by the resonance excitation process, cannot be confined for detection in the T O F M S . It should be noted that the triple quadrupole mass spectra exhibit a major fragment peak at ^ = 196. This peak cannot be seen in the T O F spectra and so decreases the apparent fragmentation efficiency. This is an inherent feature of CID in any quadrupole based ion trap. 6.6 Competition between Fragmentation and Ejection Given that the results reported here comprise the first example of M S / M S in a linear ion trap, the nature of the competition between fragmentation and ejection, described in Sec. 6.1, had to be examined. These studies followed similar work completed on the 3-d ion t r a p . 1 7 1 ' 1 7 4 , 1 8 5 ' 1 8 8 - 1 9 0 Studying the nature of the competition between fragmentation and ejec-tion has two motivations. First, analyzing the relative levels of fragmentation and ejection for different operating parameters provides a way to, at least indirectly, assess the nature of ion trajectories (and vice versa). By rough approximation, if fragments are observed, x(t) approached, but did not ex-ceed, r 0 ; if no fragments or precursor are detected, the x(t) of the precursor extended rapidly beyond r 0 without sufficient time for dissociation; and if no alteration in composition of ions in the trap is observed, x(t) failed to reach 133 o o o CO o o o o o o CM 00 1 1 1 1 1 CO 00 / \ CM CO / \ p • 00 "V " o CO — ~ o> I . I . I CO N i t siunoo uo| jo jeqiunN Figure 6.8: Fragmentation efficiency of the linear ion trap, (a) Reduction of the spectra shown in Figure 6 showing the precursor ion intensity (•) and the sum of the intensities of all fragment ions (o). The F W H M of the excitation profile is 3.0 kHz, mass resolution 73. (b). Similar plot for reserpine with resonant excitation of amplitude 2.4 V applied for 4 ms and qu = 0.51. The nominal mass resolution of excitation is 60 and the fragmentation efficiency is 60%. 134 a high potential region of the quadrupole rods. Second, in order to system-atically determine the optimal operating conditions for the L I T / T O F M S , it was necessary to vary the parameters of resonant excitation and examine which provided the desired results. Wi th pressure in the linear ion trap fixed at 7 mTorr N 2 , the parameters which were easily varied were the time of irradiation and the amplitude of the applied potential. The product of these two parameters is termed the "fiuence" of the exci ta t ion. 1 7 8 ' 1 9 0 ' 1 9 1 In the absence of a damping gas, the amplitude of excitation is determined by the magnitude of the fluence, indi-cating that increasing time and decreasing amplitude by the same factor will not affect the behavior of the ions. 1 7 8 This simple relationship is complicated by the presence of a damping gas. The effect of varying the excitation period is shown in Fig. 6.9, which plots the precursor ion and sum of fragment ion populations of the +1 re-serpine ion for various irradiation times with a fixed excitation amplitude of 1.5 V . No fragmentation is observed at short times. As the excitation time increases, fragments appear in the spectrum. For the ions studied, the disso-ciation of the precursor, if observed for a given Aa, occurs at a "threshold". The fragment population gradually rises to a plateau, and further increasing the excitation time effects no additional quantitative or qualitative spectral changes. These observations are consistent with the discussion of the forced damped harmonic oscillator presented in Sec. 6.1. For an ion excited with an auxiliary voltage with ua & u/0, the amplitude of the ion trajectory grows until the steady state is reached and during this period of enhanced motion, it may, or may not, fragment. Once the steady state is achieved, the pre-cursors experience a constant rate of energy exchange between the external 135 6000 0 10 20 30 40 50 Excitation Period/ms Figure 6.9: The effect of increasing the resonance excitation period for pre-cursor ion reserpine with Aa — 1.5 V . Note in Fig. 6.10 that for an excitation period of 4 ms Aa = 1.5 V is not sufficient to induce full fragmentation. force and collisional damping. The steady state behavior of resonantly excited ions is somewhat anal-ogous to the ion temperature in the quadrupolar potential, discussed in Sec. 4.3. Application of the auxiliary voltage "heats" the ions from the initial temperature to a new, higher temperature - now dependent on the steady state between the excitation and collisional damping, rather than on micromotion and collisional damping. As the ion approaches the steady 136 state, a "temperature" sufficient for dissociation may be reached; in this case fragmentation will be observed. The internal energy distribution is not necessarily thermal (Boltzmann) however. It should also be noted that, consistent with behavior in the absence of a damping gas, longer irradiation times can be used to achieve fragmentation with lower amplitudes of excitation. Compare, for example, Fig. 6.9, which demonstrates reserpine ion fragmentation with Aa=l-5 V for varying times, with Fig. 6.10 which shows the effects of increasing Aa for t=A ms. If the irradiation time is 4 ms, the lower voltage is insufficient to fragment the ion; when the application time is increased to 10 ms, the lower voltage facilitates dissociation. The observed interplay between amplitude and time, however, is not consistent with the notion of fluence, i.e., doubling irradiation time did not result in a twofold reduction of the amplitude necessary for fragmen-tation. Rather, within a small time frame (i.e., 1-50 ms) there was some flexibility in the lowest possible amplitude which could be used to induce dissociation. On a practical level, it is worth noting that in Fig. 6.9, after 40 ms the trap population remains essentially constant which serves as further evi-dence of the near 100% trapping efficiency of the linear ion trap on the time scales needed for M S / M S . In terms of choosing operating parameters for the L I T / T O F M S , any increase in the time of irradiation elongates the total time to complete the M S / M S cycle and thus decreases the duty cycle of the in-strument, and little advantage was noted for increasing irradiation time. The choice of optimal irradiation time was thus a combination of the need for an acceptable duty cycle, high fragmentation efficiency, and improved excitation resolution. For the +3 charge state of renin substrate, 4 ms proved to give the optimal conditions for the collection of fragment ions. 137 The second parameter which was varied in order to examine the nature of the competition between ejection and fragmentation was the second com-ponent of the excitation fluence, the amplitude of the auxiliary voltage, Aa. Fig. 6.10 shows the effect on singly charged reserpine ions of increasing the amplitude of the auxiliary voltage (4 ms excitation). While a threshold volt-age is necessary to induce fragmentation on the 4 ms time frame, as Aa increases, ejection dominates and less fragmentation is observed. Again, these observations can be explained by examining the description of the ion trajectory given in Eq. 6.24. The value of the coefficient a, which dictates the amplitude of ion motion, is directly dependent on Fa, and, by inference, Aa. The definitive element in the competition between ejection and fragmentation is the time required for the ion's trajectory to reach a rod (x(t) = r 0 ) . The more rapidly the amplitude of the trajectory increases to r 0 , the more likely it will be ejected before it undergoes enough collisions to fragment. Although this rapid ejection is undesirable in the fragmentation step of the M S / M S process, it is, conversely, the desired phenomenon for the isolation of the precursor (the initial M S / M S step) and is thus the motivation behind the use of high amplitudes in the broadband waveform. A further effect of increasing amplitude results from the corresponding alteration in the nature of ion-neutral collisions. Ions excited with higher voltages are generally moving faster in the trap, and consequently more en-ergy is exchanged per collision. The effects of different collision energetics were noted when the excitation voltage applied to the +3 charge state of renin substrate was increased. The results are summarized in Fig. 6.11. This plot looks like that of Fig 6.10 - there is a threshold for dissociation and fur-ther increasing Aa results in ejection dominating over fragmentation. As the amplitude is increased and precursors are lost, however, there is a change 138 Figure 6.10: Changes in the fragmentation spectra from the precursor ion reserpine (+1) as a function of amplitude of the resonant excitation at qu = 0.51 and an excitation period of 4 ms. The horizontal axis is channel number where channel 0 is 20 fis and each channel represents 20 ns. 139 1500 1000 500 -• Precursor o Fragments o.. "o. o Q. . A. IV ~o.. 8 10 Figure 6.11: Effect of varying the amplitude of the auxiliary excitation for the +3 charge state of renin substrate. in the observed fragmentation pattern. Fig. 6.12 shows the fragmentation patterns recorded with A A of (a) 1.5 V and (b) 3.75 V . In Fig. 6.12(a) the detected fragments of the precursor dissociation are dominated by the peak at 21 = 746, which is observed in the tandem in space system with collision cell offsets of 15 and 20 V . When A A is increased to 3.75 V , (b), the frag-ment at ^=746 is still observed, but the relative intensities of fragments at ^=696 and 582 increase significantly. Comparison with the tandem in space work indicates that the observed fragment spectra correlate to higher energy fragmentation channels, and hence, suggest higher collision or ion internal energies at higher A A . 140 o o o o o o o o o o o CD CM CM T -Aijsue;u| Figure 6.12: Fragmentation spectra with excitation amplitudes of (a) 1.5 V and (b) 3.75 V 141 In terms of choosing operating parameters for the L I T / T O F M S , the ob-servation of fragmentation was far more sensitive to the amplitude than to the irradiation time of the excitation. Typically for maximizing fragmen-tation, a fixed time of 4 ms was used and the amplitude was sequentially lowered, starting from 10 V , until the greatest extent of fragmentation was observed. 6.7 Summary This chapter presents the first demonstration of tandem mass spectrometry in a linear RF-only quadrupole. The isolation resolution was 30, and the mass resolution of ion fragmentation was 73. A reasonably high fragmentation efficiency of 60% was demonstrated, as was the ability to access different fragmentation channels. Fragment spectra were similar to those of a triple quadrupole mass spectrometer operated at a laboratory collision energy of 45 eV with 1 mTorr K r (20 cm cell length). 142 Chapter 7 Variation of Gas Pressure in the Linear Ion Trap The results presented in Chapter 6 demonstrate the feasibility of tandem mass spectrometry in the linear ion trap. The major limitation in the initial performance of the L I T / T O F M S was the relatively poor mass selectivity in the isolation and fragmentation procedures compared to results reported for 3-d ion traps. As outlined in Sec. 7.1, it was hypothesized that one cause of the poor resolution was the relatively high mass and density of the bath gas: 7 mTorr N 2 as opposed to the 1 mTorr He commonly used in a 3-d ion trap. In an attempt to improve the resolution, the vacuum system of the instrument was modified to decrease the pressure to 1.5 mTorr N 2 . This is described in Sec. 7.2. The subsequent studies on mass resolution of resonant excitation as a function of bath gas pressure and excitation amplitude generated the unexpected result outlined in Sec. 7.3 - namely, that it is the amplitude of the excitation, Aa, and not the bath gas density, which has the most significant effect on the mass resolution of resonant excitation. The improvement in the isolation procedure at the lower pressure is outlined in Sec. 7.4. In Sec. 7.5, the enhanced performance of the resonant excitation processes in 143 the linear ion trap is demonstrated through the use of the L I T / T O F M S for M S 3 . Finally, in Sec. 7.6 the use of He as the bath gas is briefly discussed. 7.1 Introduction The mass resolution of-the excitation curves shown in Chapter 6 are compa-rable to that of other groups who quoted values for the same phenomenon in linear RF-only quadrupoles. 6 6 ' 7 5 Watson et al.75 hypothesized that their low resolution was due to the relatively poor mechanical quality of the rod set used. Loboda et a/.6 6 relate the observed mass resolution to the number of ex-citation cycles an ion experiences within the RF-only quadrupole. Neither of these explanations, however, clarifies the low resolution in the L I T / T O F M S . The rods sets are of sufficiently high quality to be used as the mass analyzing quadrupoles on a commercial instrument and, through trapping, the number of irradiation cycles is flexible and proved to be irrelevant. The origins of resolution degradation in the linear ion trap would be expected to be similar to those of the 3-d ion trap. These include space charge and higher order electric fields, as well as ion scattering from collisions with the bath gas. The presence of the two former effects can be detected through a shift in the resonant frequency of an ion. As significant shifts were not observed, attempts to improve resolution focused on the interaction between the ion and the bath gas. The use of 7 mTorr of N 2 for the background pressure in the linear ion trap was motivated by considerations of T O F M S performance, rather than resonant excitation resolution. This pressure of N 2 was previously shown to give optimal transmission with collisional focusing through a 20 cm RF-only quadrupole, and is the pressure commonly used in ion guides connecting ESI sources with the lower pressure regions of triple quadrupole mass spectrom-144 eters.60 There are, however, disadvantages to using high pressure N 2 rather than a lower pressure of a lighter gas, as the collision partner. As is evident from Eq. 6.29, when the precursor ion and the neutral collide, the kinetic energy that can be transferred to internal energy during the collision will be greater for a heavier mass neutral. 1 9 2 Also, in colliding with a heavier neutral, the precursor ion has a higher probability of being scattered through a greater laboratory frame angle. The effects of scatter on the ion trajectory are twofold. First, the scatter could result in the transfer of the precursor trajectory from one which is stable in the quadrupolar potential, to one which is unstable. The subsequent ejection of the ion, degrades the sensitivity and the mass resolution of an ion t rap . 1 7 1 ' 1 9 3 The second impediment to high resolution resulting from a higher mass and higher density gas can be explained by the damped forced harmonic osc i l la tor . 1 4 6 ' 1 5 9 ' 1 6 2 The resonant excitation curves from which the mass res-olution of the linear ion trap are determined are somewhat analogous to the resonant absorption curves discussed for the generalized damped harmonic oscillator, 1 6 8 and used to calculate collision cross sections from ICR data. 9 A general property of the damped harmonic oscillator is that power absorbed by the system plotted against excitation frequency gives an inverted Lorentzian peak shape with an inherent natural width of 7, the rate of momentum damp-ing from collisions (see Sec. 6.1). Additionally, the average power absorbed by the system, Pmax, which determines the "depth" of the inverted peak, is given by — Fa Pmax — o j (7-1) 27m indicating that for a fixed amplitude, 7 limits power absorption. Alternately, to facilitate the same extent of power absorption at a higher 7, a larger 145 excitation amplitude must be used. Unlike the ICRMS, however, ion traps provide no direct observation of power absorbance. If all of the excitation applied to the system were to result in detectable precursor ion depletion then the width of the excitation curve would be 7, and the depth would be determined by Eq. 7.1. From Eq. 6.28, mass resolution for resonant excitation is inversely proportional to the width of the excitation curve; from Eq. 6.10, 7 is proportional to the mass and number density of the neutral bath gas. Consequently a large natural linewidth, resulting from a high mass and/or number of density of the neutral, would be expected to degrade mass resolution. Correspondingly, the F W H M for excitation curves for 7 mTorr N 2 would be expected to be significantly larger than those for 1 mTorr He in the 3-d ion trap and thus, the mass resolution of the linear ion trap operated under these conditions is inherently lower than that of the 3-d ion trap. The modifications to the experimental setup described in the following section were motivated by the desire to lower the natural linewidth of the resonant excitation curves. 7.2 Experimental Methods The pressure in the linear ion trap is determined by the flow rate of curtain gas into the vacuum from the electrospray source and the speed of the tur-bomolecular pump attached to the linear ion trap. A lower pressure in the linear ion trap was attained by replacing the 50 L / s turbomolecular with a 345 L / s pump. When the 50 L/s pump was used, a sufficient backing pres-sure for the two turbomolecular pumps in the L I T / T O F M S (the linear ion trap and 345 L / s pump in the TOFMS) was attained using a common ro-tary vane pump. The second 345 L / s turbomolecular pump required a lower backing pressure than the 50 L / s pump, and hence, a third rotary vane pump 146 was added to the system, such that each turbomolecular pump was backed by a separate 7 L / s pump. With the new pumping system, the background pressure of N 2 in the linear ion trap decreased to ~ 1.5 mTorr. To complete a detailed study of bath gas pressure in the linear ion trap, a few additional modifications to the instrument were required. In the ap-paratus described in Chapter 6, the linear ion trap pressure was measured by a Thermovac TM20 vacuum gauge. To record pressure more accurately, a precision capacitance manometer (MKS type 120 high accuracy pressure transducer, manufacturer's stated accuracy 0.12% of reading) was connected to the linear ion trap. To vary the neutral density, a gas inlet in the trap region of the instrument was equipped with a leak valve. Through the in-let, additional gas was added to increase the pressure above the background pressure in the trap. By adding N 2 to the trap, the L I T / T O F M S could be operated with conditions identical to those described in Chapter 6. Results from the addition of helium will be discussed in Sec. 7.6. As collisional focusing through 7 mTorr N 2 created optimal conditions for the transfer of the ions from the linear ion trap to the T O F source region, one concern in lowering the pressure is degrading T O F M S performance. Fig. 7.1 plots the number of ion counts as a function of pressure in two modes of operating the L I T / T O F M S . The experiment was completed with an elec-trosprayed solution of 100 /xM reserpine in acetonitrile. A l l spectra were recorded in a short time frame on the same day and thus it can be assumed that sensitivity differences are not the result of a change in ion source con-ditions. The voltages of the L I T / T O F M S are as discussed for Fig. 5.4(a). For the data in Fig. 7.1(a), the entrance and exit aperture plates were constantly operated at transmitting potentials. For the data plotted in Fig. 7.1(b), the ions were pulsed into the T O F M S using the 5 ms injection 147 (a) ( b ) 6000 12000 11000 10000 9000 8000 7000 8 6000 2 4 6 8 0 2 4 6 Pressure/mTorr Pressure/mTorr Figure 7.1: The number of ions detected as a function of pressure with the L I T / T O F M S operated with (a) no trapping and (b) a 5 ms injection time. Note that in (a) the number of ion counts increases sixfold and in (b) twofold. and 7 ms extraction times discussed in Sec. 6.2. The result in Fig. 7.1(a) is qualitatively similar to that presented by Douglas and French. 6 0 As pres-sure increases, the number of ions detected rises by a factor of 6, reflecting increases in ion transmission and ion beam density resulting from the colli-sional damping process. At higher pressures however, translational energy losses increase to a point where some ions lose all forward motion, causing a plateau, and then, as ion transmission decreases, a fall, in the number of ions detected. In contrast, Fig. 7.1(b) shows an almost linear increase in the number of ion counts over the same pressure range with no decrease at high 148 pressure observed. As noted on the plot, the vertical axis in (a) encompasses a sixfold increase in the number of ions detected, while (b) shows only a twofold increase. Each data point in Fig. 7.1 represents the coaddition of 50,000 T O F M S scans. The time to record this number of scans for the operating conditions of Fig. 7.1(a) is 5 s (i.e., 50,000 scans divided by the 10 kHz repetition rate of the repeller plate pulsing). For the data plotted in Fig. 7.1(b), the T O F is not pulsing as the linear ion trap is being filled, the total time to record 50,000 scans is 8.6 s. Two distinctions between (a) and (b) should be noted. First, the number of ion counts, relative to both the number of scans and the total time to record the spectrum, is increased when ions are confined in the linear ion trap prior to entering the T O F source. The sensitivity for all pressures in Fig. 7(b) is always greater than that in 7(a). Second, the number of ion counts is far more sensitive to trap pressure when no trapping is used. The number of ion counts per pulse in the L I T / T O F M S is dependent upon the density of the beam, which is defined by its z direction velocity. As discussed and demonstrated in Sec. 3.3 (see Fig. 3.3) trapping ions, through increasing the extent of collisional damping, is an energy, and hence a velocity, normalizing process. Also, as noted previously by others, 6 0 ' 6 1 and further demonstrated in this work, ions traveling 20 cm in the mTorr regime of N 2 , are at near thermal energies. At thermal energies, sensitivity will not change as a function of either trapping time or pressure. When trapping is used, however, ions at a given pressure, are closer to thermal energy, and consequently, considering the Boltzmann distribution, a higher proportion of the ions in the beam is traveling at thermal velocity. Consequently, the number of ions detected is less affected by the pressure in the linear ion trap when trapping is used. 149 Fig. 7.1 serves as further evidence of the advantages of confining ions in a collisional environment prior to T O F scanning. More importantly, it demon-strates that operating the linear ion trap at the lower pressure necessary for improving resonant excitation will not result in a significant adverse effect on overall instrumental sensitivity when trapping is used. 7.3 Mass Resolution of Resonant Excitation Once the pressure in the linear ion trap was lowered to 1.5 mTorr N 2 , excita-tion curves were measured at the new bath gas density. It was immediately apparent that at the lower pressure a lower amplitude of the applied resonant excitation, Aa, had to be used to detect fragment ions. This observation is consistent with the expression for maximum power absorption in a forced damped harmonic oscillator given in Eq. 7.1. The same observation may also be understood in terms of ion trajectories. It is obvious from Eq. 6.25 that the maximum amplitude of an ion excited with resonant excitation is inversely proportional to the magnitude of the frictional damping coefficient, 7, and by inference, the bath gas density. As 7 decreases, the extent of ki-netic energy lost through friction lowers, and consequently the amplitude of the ion trajectory increases. Following the ideas presented in Sec. 6.6, rapid increases in oscillation amplitude result in ejection being favored over frag-mentation. Therefore, in order to observe fragmentation at lower pressures the amplitude of the excitation must be decreased. The relationship between pressure and excitation amplitude is exhibited in Fig. 7.2 which displays 9 excitation curves - a matrix of 3 different ex-citation amplitudes, 450 mV, 900 mV, and 1200 mV, at 3 pressures of N 2 - 2 mTorr, 4 mTorr, and 6 mTorr. The data presented in Fig. 7.2 were recorded on the +3 charge state of renin substrate (qu ~ 0.500) with injec-150 tion for 5 ms, isolation by broadband waveform for 4 ms, resonant excitation for 10 ms, and then, extraction/TOF scanning for 10 ms. The increase in excitation time from the 4 ms discussed in Chapter 6 was motivated by the desire to maximize the fragmentation at low amplitudes. The observed re-lationship between fragmentation and ejection is as expected. For instance, at an excitation amplitude of 450 mV and 2 mTorr, ejection dominates, at 4 mTorr complete depletion of precursor intensity is not observed, but a reasonable number of fragments are detected, and, finally at 6 mTorr, the excitation is not sufficient to induce either ejection or fragmentation. At higher 7 the power absorption profile has a small "depth" and the maximum power absorbed is not enough to result in precursor ion depletion. While the relationship between ejection and fragmentation shown in Fig. 7.2 is as predicted, the width ( A / ) of the curves demonstrates that the experimental mass resolutions = ^j) are contrary to the initial ex-pectation that mass resolution would be defined by 7. The widths of the curves are similar for each amplitude and vary significantly within each pres-sure, indicating that factors other than the natural linewidth are contributing to A / . It is particularly interesting that this observation holds for different extents of power absorption. For a constant pressure of 2 mTorr, all ex-citation amplitudes result in full depletion of precursor population, either through ejection, fragmentation, or a combination of both, but the width of the excitation curve triples. For a constant amplitude of 900 mV full ejection is observed at 2 mTorr and successful fragmentation at 6 mTorr, but A / is similar in each instance. The resolutions measured from the excitation curves shown in Fig. 7.2, along with others recorded at additional pressures and amplitudes, are plot-ted in Fig. 7.3. This figure further emphasizes the observation that at each 151 Figure 7.2: Resonant excitation curves for the +3 charge state of renin sub-strate as a function of both pressure and excitation amplitude. Precursor intensity is given by • and the sum of fragment ion intensities by • . 152 amplitude, the resolution is essentially constant, but at a single pressure, resolution can vary significantly with amplitude. Clearly, in this range of pressures (2 mTorr-7 mTorr N 2 ) the mass resolution of resonant excitation is dependent upon the amplitude of the applied excitation rather than the gas density. It should be noted that an apparent increase in resolution is observed at higher pressure for some amplitudes. The origin of this increase is evident in Fig. 7.2 for Aa=450 mV at 4 mTorr and, to a lesser extent, y4.a=900 mV at 6 mTorr. When the excitation is such that ejection and fragmentation are just barely facilitated, the excitation curves narrow. It should also be noted that the width of these excitation curves, recorded with an irradiation time of 10 ms should not be directly compared to those presented in Chapter 6 for identical excitation amplitudes but shorter (4 ms) irradiation times. As discussed in Sec. 6.6, longer excitation time results in additional depletion of the precursor ion. This would be expected to widen the excitation curve and hence lower resolution. Knowledge of the empirical relationship between the width of the exci-tation curve and the amplitude of excitation permitted an examination of the "best" possible mass selectivity in the L I T / T O F M S . To optimize res-olution, the pressure was maintained at 1.5 mTorr N 2 and the excitation amplitude was sequentially lowered near the calculated resonant frequency until precursor fragmentation, with minimal ejection, was observed. The time of irradiation with low voltage excitation was varied. A selection of the observed excitation curves are shown in Fig. 7.4 and demonstrate mass reso-lutions from 230 to 300. The highest resolutions are achieved by short time application of low voltages as in (b) and (c) which plot data obtained from excitation for 3 ms with 250 mV and for 2 ms with 275 mV, respectively. However, in none of these plots is full precursor depletion achieved. Also, (a) 153 200 Pressure/mTorr Figure 7.3: Variation of resolution as a function of pressure for various Aa. Values shown in legend is the 0 to peak amplitudes of the voltages applied to the rods. demonstrates the excitation curve for incomplete ejection and dissociation on a longer time scale. A broadened, asymmetric curve is observed. It should also be noted the minimum possible width of the excitation curve in the fre-quency domain is dictated by the inverse of the time the pulse is applied. At 4 ms this minimum possible width is 250 Hz, compared to 706 Hz observed, and at 2 ms, 500 Hz, with 922 Hz observed. As the resolutions quoted for the higher pressures do reflect complete precursor depletion, the most valid comparison between the mass resolutions at 7 mTorr and at 1.5 mTorr is to contrast Fig. 6.8(b) with Fig. 7.4(d). In both, the time of irradiation is 4 ms. In Fig. 6.8(b) the amplitude of the excitation voltage is 1500 mV, which is 5 times greater than the 300 mV used in Fig. 7.4(d). The width at half height with the lower amplitude rep-154 1 (a) (b) . 0.8 ^^^A R=230 f \ R=290 / 0.6 \ / 0.4 0.2 / fe ' \ y1 1 0.8 (c) (d) \ R=300 / \ R=230 0.6 \ / 0.4 A * / m\ y J 0.2 / w V *' \ i • \ / m- -m „ > ' . % 211.5 212 212.5 2 1 1 . 5 2 1 2 2 1 2 . 5 f. Figure 7.4: Excitation curves with the parameters of excitation voltage opti-mized for high resolution. In (a) Aa = 225 mV and t=4 ms, (b) Aa=250 mV and t=3 ms, in (c) Aa=275 mV and t=2 ms, and in (d) Aa=300 mV and t—4 ms. resents a slightly more than threefold improvement in resolution. While, of course, further enhancement in mass resolution is possible, this improvement represents a substantial advancement in the development the linear ion trap for practical applications. It should be mentioned, however, that at the lower pressure the fragmentation efficiency decreases. It is also worth mentioning that a low precursor selection resolution is often desired in M S / M S . While separating the isotopic distribution of a mul-tiplied charged high mass ion necessitates a resolution defined by the inverse 155 of its mass to charge ratio, isolation of the whole isotopic cluster, which spans a broader mass range, requires a resolution which can be an order of magni-tude (or more) lower. In high quality hybrid T O F instruments the resolution of the fragment ions is sufficiently high to detect mass differences between isotopic peaks. Therefore isotope peak spacing can be used to assign the charge states of the fragment ions. The observation that low amplitudes result in narrow excitation curves precipitated the desired substantial improvement in the mass resolution of the linear ion trap. The reason the hypothesized correlation between gas pressure and mass resolution failed, however, requires further examination. The most thorough discussion on the relationship between excitation amplitude and mass resolution in a linear RF-only quadrupole is provided by Fischer. 7 3 Fis-cher describes factors which affect mass resolution for resonant excitation in the absence of a bath gas. In this system the "natural" linewidth is infinitely small, corresponding to 7=0. The observed lineshape in this system is de-fined by factors which result in precursor ion depletion for applied excitation frequencies which are near but not equal to the fundamental frequency. Mass resolution can thus be understood by examining the trajectories of excited ions. If the trajectory of an ion irradiated with fa is such that x(t) > ro, then the excitation curve widens to include the non resonant frequency. As discussed in Sec. 6.1, in the absence of a damping gas the amplitudes of os-cillation of ions irradiated with resonant (/ a = /o) dipolar excitation grow linearly with time. Ions which are irradiated with slightly off resonant aux-iliary excitation have trajectories which exhibit beats. If the displacement in the high amplitude portion of beat motion is larger than the field radius of the quadrupole rods or if internal energy gained from collisions induced by beat motion is sufficient to cause fragmentation, the precursor ion will 156 not be detected after irradiation at that fa. The mass resolution is defined by twice the difference between the fa which results in one half depletion in precursor ion population and the fundamental frequency. The amplitude of beat motion is inversely proportional to the factor ( / 2 — / Q ) and directly proportional to Fa. Consequently, beat motion has a higher amplitude for fa ~ /o- However, at a fixed / a , the total amplitude of beat motion is depen-dent upon the magnitude of Fa, and thus A a - Accordingly high excitation amplitudes result in wider excitation curves and lower mass resolutions. In the presence of the neutral gas, two possible causes of resolution degra-dation must be examined: the natural linewidth defined by 7, and the extent to which near resonant irradiation creates trajectories which result in pre-cursor ion depletion. The interplay between these two effects was examined by plotting the trajectories of ions for near resonant excitation for different amplitudes and pressures. The calculations were completed using Eq. 6.24 and the program described in Sec. 6.1. Figures 7.5 and 7.6 exhibit the trajectories for a single ion with / 0=200 kHz, irradiated with non resonant excitation, fa = 197 kHz. If the trajectory of the ion excited at this frequency is such that one half of the precursor ion is depleted, the mass resolution of resonant excitation would be 22°- = 33. Fig. 7.5 exhibits the trajectories of ions irradiated with ex-citation voltages of (a) 250 mV, (b) 500 mV, and (c) 1000 mV and N 2 at 3 mTorr (7=2875 Hz). The fourfold increase in A a from (a) to (c) gives a direct quadrupling of the amplitude of the trajectory. Clearly use of the 1000 mV excitation amplitude would result in a substantial degradation in mass resolution relative to the 250 mV excitation. The system shown in Fig. 7.5 (a) with f0 = 200 kHz, / a=197 kHz, and A a=250 mV is plotted in Fig. 7.6 at pressures of (a) 1.5, (b) 3.0, and (c) 6.0 mTorr N 2 . The variation 157 in pressure results in substantial changes to the character of the trajectory. Most notably, since the time scale of the transient contribution to the tra-jectory decreases with increasing pressure, beats have a discernible effect on the trajectory at a pressure of 1.5 mTorr N 2 (a) for greater than 4 ms while at 6 mTorr (c) the steady state is reached in approximately 1 ms. How-ever, the variation in pressure does not result in the substantial alteration to the amplitude of ion motion - the maximum x(t) at 6.0 mTorr (c) is two thirds the maximum amplitude at 1.5 mTorr (a). As the amplitudes of the off resonant trajectories are not substantially different, the mass resolutions would be expected to be similar at all of the plotted pressures. Hence, for off resonant excitation, mass resolution is determined by the magnitude of the excitation amplitude. Figures 7.7 and 7.8 plot the trajectories varying the same parameters as those discussed for Figs. 7.5 and 7.6, but for fa = 199.5 kHz, and thus | / a — /o|=0.5 kHz. If the trajectories of ions at this frequency resulted in a depletion of one half of the precursor ion intensity, the mass resolution would be 200. At / a=199.5 kHz, the increase in Aa has the same effect as when / a=197 kHz - as is obvious in Eq. 6.24 the amplitude of the excitation and the amplitude of the trajectory are linearly related. Excitation with yla=1000 mV (a), will attain a maximum amplitude of motion which is four times greater than an ion irradiated with Aa=250 mV, (c). In contrast, Fig. 7.8, when compared to Fig. 7.6, demonstrates that the effects of pressure at 199.5 kHz are dramatically more marked than at 197 kHz. A seemingly linear relationship between pressure and maximum amplitude of ion motion is observed. Again, this is evident in Eq. 6.24. As the (u%-a;2)2 term vanishes, the denominator reduces to co0 x 7, and hence x(t) becomes inversely related to 7. Decreasing pressure thus would be expected to have the equivalent 158 CD £ CM O C\J 3 CD O O O O O O O d d d c i c i d C ) I I I CD ^ CM O CM ^ CO o o o o o p p d d d d d d d CO CM O CM ^ CO o o o o o p p d d d d d d d 1 1 1 Figure 7.5: Trajectories for an ion in 3 mTorr N 2 and / 0=200 kHz, irradiated with an excitation voltage with / a=197 kHz, and Aa(a) 250 mV, (b) 500 mV and (c) 1000 mV. 159 CO CO o CO CO CO oo 1— o o o o o o o O o d d d d d d I I O O O C O C O C O O O O C O O t - T - O O T -p o o o o o d d d d d o i i i I Figure 7.6: Trajectories for an ion with / 0=200 kHz irradiated with an exci-tation voltage with ^4a=250mV and / a=197 kHz for pressures of (a) 1.5, (b) 3.0 and (c) 6.0 mTorr N 2 . 160 effect on the trajectory as increasing Aa. The trajectories plotted in Figs. 7.5-7.8 can be used to explain why mass resolution in the linear ion trap is predominantly dependent upon the ampli-tude of excitation and relatively insensitive to bath gas pressure. The natu-ral linewidth defines the extent to which resonant excitation is spread over a range of frequencies about the fundamental frequency at a given pressure. In practice, however, even in the absence of a bath gas, the mass selectiv-ity of resonant excitation is limited by beat motion resulting in precursor ion loss. For resonant excitation in the presence of a bath gas, the natural linewidth defines an effective "zone" of excitation frequencies for which the ion trajectories are a strong function of 7. As fa approaches / 0 , the the ion trajectory is determined by the trap pressure; as fa deviates from f0, the trajectory becomes progressively less sensitive to 7. For these fa, the widths of the excitation curves are determined in an identical fashion to an RF-only quadrupole in the absence of bath gas. Resolution was observed to be a function of the amplitude of beat motion and thus dependent upon the amplitude of auxiliary excitation. In addition, whether or not the system has absorbed enough power to induce fragmentation or ejection is strong dependent on the amplitude of excitation. This result can be related to the expression for average power ab-sorption given in Eq. 7.1, which shows that the maximum power absorption is directly proportional to the square of the amplitude of excitation. Fur-thermore, the extent of power absorption is inversely related to 7, indicating higher pressures require larger excitation amplitudes to achieve threshold energies necessary for fragmentation and ejection. As a consequence of this correlation between power absorption and 7, high excitation amplitudes are required at higher pressures. Completely independent of pressure, resulting 161 o ID ^ W O CO ^ (O (O 5 CM O CM 3 (D CO ^ CM O CM CD o o o o o o o o o o o o o o o o o o o o o d d d d c J d d d d d d d d d d d d d d d d i l l i l l i l l Figure 7.7: Trajectories for an ion with /o = 200 kHz in 3 mTorr N 2 irradiated with an excitation voltage with fa = 199.5 kHz and Aa of (a) 250 mV, (b) 500 mV and (c) 1000 mV. 162 Figure 7.8: Trajectories for an ion with / 0=200 kHz irradiated with an exci-tation voltage with /4.a=250mV and / a=199.5 kHz and pressures of (a) 1.5, (b) 3.0 and (c) 6 mTorr N 2 . 163 from the nature of beat motion as described above, higher excitation ampli-tudes result in lower mass selectively. The net result from these two separate effects is that the minimum observable resolution is, as expected, a function of gas pressure. The important conclusion, however, is that once the threshold power absorption for excitation and fragmentation is reached, the dominant factor affecting resolution is actually the amplitude of the excitation. Consequently, optimizing mass resolution, like optimizing fragmentation efficiency, requires determining the lowest possible amplitude which is suf-ficient for ejection and fragmentation to be observed. The importance of low voltage excitation has been noted for resonant excitation in the 3-d ion trap. 1 9 4 7.4 Isolation at Lower Pressures Of course, the demonstrated improved mass resolution has implications for the quality of isolation of the precursor ion as well. As discussed in Sec. 6.4, the isolation with 7 mTorr N 2 in the linear ion trap was limited by low resolution, poor ejection of high j ions, and the observation of fragmentation resulting from the near resonant frequencies of the broadband waveform. The isolation of the +3 charge state of renin substrate (gu=0.600) with 1.5 mTorr N 2 in the linear ion trap is shown in Fig. 7.9 (compare with 6.6). The details of the waveform are discussed below. Clearly the problems of both high mass ejection and near resonant fragmentation have been eliminated by. lowering the neutral gas density. As discussed in Sec. 7.3, the width of the excitation curve is determined by the amplitude of the auxiliary voltage. Consequently, to achieve an improved isolation resolution, the components of the broadband waveform which are near f0 need to have a low amplitude. To achieve a higher resolution, a 164 CO 4 o o CD O O C O o o ^ CO fc o o LO o o LO CM o o LO o o LO o o LO CM o o LO o o LO o o o o CO Al|SU9}U| Figure 7.9: Isolation with 1.5 mTorr N 2 in the linear ion trap demonstrating a mass selectivity of 100. Panel (b) shows the results of the two step isolation procedure, detailed in the text, applied to the collection of ions shown in (a). 165 two step isolation procedure was implemented. In the first step, a high amplitude (30 V ) , wide notch (5 kHz) broadband waveform was constructed to eject ions across the ^ range. The mass resolution of this first step is 40. In the second step, a low amplitude (1 V) , narrow notch (2 kHz) waveform was used to attain the isolation resolution of 100 demonstrated in Fig. 7.9. The broadband waveforms used in each step have the structure described in Sec. 6.2 and are 4 ms long. It is not unusual to use a two step isolation procedure to improve isolation selectivity. 1 9 5 - 1 9 7 However, the addition of 4 ms to the total time required for a full M S / M S scan does degrade the duty cycle. The same concept for an improved isolation resolution as demonstrated here, i.e., lower amplitudes for near resonant excitation, could be achieved in a one step procedure if SWIFT waveforms were used. 1 8 7 As an aside, the mass selectivity of the isolation process is an important quantity for establishing the capabilities of the linear ion trap as an ion bottle or inlet device for mass analyzers other than the T O F M S . In particular, for mass analyzers whose performance is degraded by space charge effects, as are the 3-d ion trap and ICRMS, it is desirable to allow only the species of interest to enter the mass analyzer. This is the basis of the ion bottle patent of Douglas. 8 1 Whereas quadrupoles and octopoles are already being used as inlet devices to improve the performance of the ICRMS, as noted by Senko et a/., 8 2 adapting such ion guides to facilitate selective ion storage would be an obvious modification. For the ICRMS, external ion isolation in an linear ion trap could also result in a substantial improvement in instrumental duty cycle as present methods for in situ ion isolation require the addition of a neutral gas to the ICRMS cell. In order to attain reasonable resolution, however, the additional gas must be evacuated from the system prior to ICRMS mass 166 analysis, a procedure which requires seconds to complete, and thus the duty cycle is substantially degraded by the isolation process.1 9 8 7.5 MS3 in an RF-Only Quadrupole This section demonstrates the potential of the linear ion trap for multiple step tandem in time mass spectrometry by demonstrating that M S 3 is possible in the L I T / T O F M S . It should be noted that attempts to accomplish M S 3 with the trap pressure .fixed at 7 mTorr N 2 were unsuccessful due to the problems of implementing a second isolation step for a fragment ion with a higher ^ than the precursor. Typically, fragments with higher ^ than the precursor could not be isolated with the existing isolation waveform, as nearby ions could not be ejected from the trap. With the improvement in the performance at low pressures the ability to complete M S " in the linear ion trap was confirmed. One of the advantages of tandem in time mass spectrometry is that mul-tiple steps of isolation/fragmentation can be completed. Through the se-quential application of the isolation and fragmentation procedures outlined in Chapter 6, the fragments of an isolated fragment can be detected. For semantic simplicity, the fragments produced by M S / M S will be termed pri-mary fragments (sometimes referred to as daughter ions) and the fragments of fragments, produced in M S 3 , will be called secondary fragments (some-times referred to as granddaughter ions). M S 3 was first performed with a 3-d ion trap by Cooks and coworkers in 1982 1 9 9 and multiple steps up to n = 8 have been demonstrated.2 0 0 Achieving M S 3 adds substantial capa-bilities to any instrument designed for analyzing biomolecules as assessing the fragments of fragments can be used for determining ion structure, as well as sequencing.2 0 1 In addition, as one of the oft quoted disadvantages of 167 tandem in time CID using resonant excitation is the limitations in the ob-servation of high energy channels, M S n is essentially a method by which to put additional energy into the fragments, thus allowing for the observation of the same fragmentation channels observed with higher collision energies in tandem in space instruments. M S n has been observed in RF-only quadrupoles previously, but only through the use of double or triple resonance techniques, 6 6 ' 2 0 2 that is by ap-plying simultaneously an excitation signal which is the sum of the fundamen-tal frequencies of both the precursor and the primary fragment. Such a sys-tem does not give true M S 3 as there is no in situ isolation step. The analysis relies on spectral subtraction, rather than a direct observation. The following thus comprises the first demonstration of M S 3 in an RF-only quadrupole. The M S 3 experiment was completed with the vacuum system described in Sec. 7.2. Additional N 2 was added to the system to bring the gas pressure to 4 mTorr. This choice of background pressure was based on a compromise between maintaining reasonable isolation and excitation resolutions and pro-ducing a sufficient number of fragments corresponding to a single fragmen-tation channel such that an adequate number of secondary fragments could be detected. A similar argument was used to choose Vrf, and hence the qu values, of the precursor and fragment. More fragments will be formed from the precursor ion for a high qu, however, a higher qu for the precursor corre-sponds to a higher low mass cut off, and hence fewer low ^ fragments will be collected. The qu of the +3 charge state of renin substrate (^=587) was 0.615, and qu for the isolated primary fragment (^=746) was 0.484. The lowest j which had a stable trajectory in the linear ion trap (i.e., c7u=0.908) was ^=402. Spectra reflecting the ion population in the linear ion trap at each step in M S 3 are shown in Fig. 7.10 (a) to (e). Panel (a) shows the 168 composition of the ions produced via ESI from a solution of renin substrate as described in Sec. 6.3 with a 5 ms injection period to the linear ion trap and a 15 ms extraction time. Panels (b) and (c) are identical to the de-scriptions of M S / M S from Chapters 6 and 7. For the isolation shown in (b) /o=217.5 kHz and the broadband waveform was constructed with a notch from 215 to 220 kHz and an amplitude of 25 V , 0 to peak. This corresponds to an isolation mass resolution of ~ 40. The precursor at ^=587 was then irradiated with an excitation voltage having an amplitude 625 mV, and fre-quency 217.5 kHz for 1 ms. The fragmentation of the precursor ion is not complete on this time frame. If the ions at ^=587 were irradiated for the typical t=4 ms, the primary fragment of interest at ^ = 746 did not grow in intensity. Instead, the extra time resulted in complete ejection or dissociation of the precursor and an increase in the intensity of fragments other than the one of interest at ^ = 746. From a starting peak intensity of ~ 8000 counts for the +3 charge state, the intensity of the dominant fragment is ~ 600 (7.5%), and the total fragmentation efficiency (ratio of the sum all fragment intensities to precursor intensity) is 18%. The first step in M S 3 is the isolation of the dominant primary fragment peak at ^ = 746 (c7„=0.484), shown in (d). The broadband waveform had a notch from 162 to 167 kHz, and an amplitude of 25 V , indicating a resolution of 33. The side peaks near the ion of interest result from a poor quality isolation. The isolated primary fragment ion was excited with an excitation voltage having Aa of 1.25 mV and / a=165.3 kHz for 1 ms. The short time was sufficient to attain the rich fragment spectrum shown in (e) and irradiation for longer time scales did not alter the spectrum. The fragmentation efficiency, calculated without the intensities of the adjacent peaks for the process from (d) to (e) is 60%. 169 8000 6000 i ( ~ — i 1 1—H——*-(b) H H (C) (d) Ik. -* h (e) 300 400 500 600 700 m/e 800 900 1000 Figure 7.10: Steps representing M S 3 in the linear ion trap. See text for details. 170 The solution was 5fjM renin substrate and the flow rate at the ESI source was l / /L /min . The T O F M S was operated at conditions for maximum resolu-tion, and in order to detect the low intensity secondary fragments created in the second excitation step, 250,000 T O F scans were recorded. Hence, it took 50 s, and 4.2 picomoles of renin substrate, to record the spectrum shown in (e). As was mentioned, M S 3 , in fragmenting fragments, provides a method to access higher energy fragmentation channels. This is at least partially demonstrated by comparing the spectrum in (e) to those recorded on the tandem in space mass spectrometry experiment shown in Fig. 6.5. Panel (c) and (e) of Fig. 7.10 both exhibit fragments (from different precursors) at ^ of 696, 640, and 582, although the observed relative intensities are substantially different. The fragment distribution in Fig. 7.10(c) is similar to that in Fig. 6.5(b) for a laboratory frame energy of 45 eV on the triple quadrupole mass spectrometer. The fragment distribution in Fig. 7.10(e) more closely resembles the tandem in space spectra for the higher laboratory collision energies of 60 and 75 eV, shown in Fig. 6.5(c) and (d). This exercise proves the feasibility of completing M S " in a linear ion trap. There are however, limitations to the performance levels in the present em-bodiment. The fragmentation efficiency of the full M S 3 (sum of fragments in (e) to precursor ion intensity in (b)) is 4%. Because the total time for completing the M S 3 cycle was 30 ms (primarily due to the long extraction time), the duty cycle for M S 3 is 0.018%. There are some practical (but not fundamental) limitations to completing M S " on renin substrate in the configuration of the L I T / T O F M S . The arbitrary waveform generator has a fixed number of points in time from which the necessary waveforms could be constructed. Completing the six segments required for M S 3 (trap injection 171 and extraction, 2 isolation steps, and 2 excitation steps) used all the avail-able computer memory, rendering implementation of M S 4 impossible. This restriction was also the reason that only the low resolution broadband wave-form could be utilized - there was insufficient memory to apply a second, low amplitude high resolution isolation, which would have made the elimination of the adjacent peaks possible. It should also be noted that renin substrate is not the ideal ion to highlight MS" capabilities as the existence of competing fragmentation pathways limits the intensity in any one channel, and conse-quently, finding successive high yield fragment channels is difficult. On the other hand it is typical of peptide spectra. Part of the interest behind the demonstration of M S " in the RF-only quadrupole is the simple portability of the technology. If the collision cell of either a triple quadrupole mass spectrometer or the Q-TOFMS was modified to operate as a linear ion trap, then fragments formed through tandem in space CID could be isolated and fragmented to effect M S " in instruments which previously had fixed M S / M S capabilities. For example, to adapt the triple quadrupole mass spectrometer described in Sec. 3.2 for M S " would require only the addition of pulsed stopping potentials to existing ion op-tics and the alteration of the connection between the main RF-drive and the quadrupole rod set. These have been discussed in Sees. 3.2 and 6.2. Such an adaptation would open up the possibility of observing new fragmenta-tion channels with higher efficiency than is typically observed in tandem in space based instruments and would increase the capabilities of such instru-ments for the applications of biomolecule sequencing and examination of ion structure. 2 0 3 172 7.6 Use of Helium as a Collision Partner As mentioned in Sec. 7.2, the new pumping system also permitted a gas other than N 2 to be used as the collision partner. To take advantage of this, a few brief studies of resonant excitation on the +3 charge state of renin substrate using helium as a collision gas were completed. Throughout the discussions below it should be remembered that He was added to 1.5 mTorr of residual N 2 in the trap, hence collisions between the precursor and both He and N 2 are possible. As the partial pressure of He increases, the number of collisions with N 2 remains constant and the number of collisions with He increases. Lowering of the gas mass has two expected effects, both of which have been documented in similar experiments in 3-d ion traps. 1 7 1 ' 1 8 5 First, consid-ering the damped forced harmonic oscillator, as the mass of the bath gas M , lowers, the extent of collisional damping, and thus 7, will decrease. Conse-quently, the behavior discussed in Sec. 7.3 for lower neutral gas densities, will also be expected for lighter mass gases - a lower excitation amplitude would be required to achieve ejection and fragmentation. The reason the frictional force is lower for a lighter gas is due to the decrease in the laboratory energy exchanged in each collision. This reduced collision energy is also expected to have an analogous effect on the observed fragmentation pattern as did lower excitation amplitudes discussed in Sec. 6.6 - the fragmentation patterns for resonant excitation with He should exhibit fragments associated with lower energy channels than N 2 . Although only limited experiments were completed, both of these phe-nomena were observed. With a 4 ms irradiation time, (fju=0.565 and fo = 201 kHz) an excitation amplitude of 450 mV was sufficient to induce full precursor depletion with a total pressure (i.e., 1.5 mTorr N 2 + He) of 3.5 and 4.5 mTorr. At the same amplitude/pressure combination, negligible 173 precursor depletion was noted for N 2 . For resonant excitation with He as the collision gas at all pressures, a slightly modified fragmentation pattern was observed. For instance, comparison of a 600 mV amplitude excitation at 2.5 mTorr total pressure demonstrated that the dominant dissociation product when He was added was ^=582, while for N 2 added, the dominant fragment is ^=784. The fragment at & = 582 was observed with lower col-lision energies in the tandem in space study than the fragment at ^=784. Little difference in fragmentation efficiency was noted for the two systems. A comment on the mass resolutions of resonant excitation for both gases is also necessary. In complete agreement with the discussion on the corre-lation between 7 and the amplitude of excitation in Sec. 6.6, no substantial differences in mass resolution for the two gases were observed for identical excitation parameters. This is in itself an important result, as it indicates that a heavier collision gas can be used with little sacrifice in trap perfor-mance. Similar conclusions have been reached for the 3-dimensional ion trap 1 9 3 ' 2 0 4 ' 2 0 5 To understand the effects of collision gas in greater detail it would be necessary to modify the linear ion trap such that the collision gas is added to a higher vacuum region, similar to the collision cell in the triple quadrupole mass spectrometer described in Sec. 3.2. 7.7 Summary Modifying the vacuum system in the L I T / T O F M S such that the pressure in the linear ion trap was reduced to 1.5 mTorr N 2 resulted in a substantial improvement in the mass resolution of resonant excitation, from ~ 70 to ~ 250. The mass selectivity of the ion isolation also increased from 30 to 100. With the improvement in instrumental performance it was possible to 174 observe M S 3 in the linear ion trap. 175 Chapter 8 Summary, Conclusion, and Outlook There are those who believe that to add "Future Work" implies a sense of incompleteness. In many respects, however, this thesis is more of an intro-duction to the possibilities of linear ion traps coupled to T O F mass analyzers than it is a representation of the complete capacities of the L I T / T O F M S . In providing "proof of principle" and a greater depth of understanding for trapping and resonant excitation in R F only linear quadrupoles, this work builds the foundation for a powerful new tandem mass spectrometer. As a brief survey of work completed is given at the end of each chapter; the following summary discusses only the fundamental elements of the linear ion trap and the L I T / T O F M S . As demonstrated in Chapter 3, the linear ion trap has a higher injection efficiency and a theoretically higher ion capacity that the 3-d ion trap. The trapping efficiency on time scales necessary for tandem mass spectrometry is 100%. This high trapping efficiency permits the RF-only quadrupole to be used for kinetic studies. In Chapter 4 the linear ion trap was used to analyze the dissociation dynamics of highly charged holomyoglobin, leading to the conclusion that there was little correlation 176 between the heme binding energy and the charge state of the electrosprayed ion. Also in Chapter 4, the linear ion trap was used to assess the temperature of holomyoglobin ions in a quadrupolar field - ~ 600 K , Chapter 5 outlined the construction of the L I T / T O F M S and the achieved performance levels of the T O F analyzer. The mass resolution was ~ 700 at Y = 609 when optimized for the highest resolution. One limit in the present embodiment of the L I T / T O F M S is the strong trade off between factors which enhance sensitivity and those which improve resolution. The use of resonant excitation to affect the isolation and fragmentation of precursor ions necessary for tandem mass spectrometry was described and demonstrated in Chapter 6. When the linear ion trap was operated at 7 mTorr N 2 the isolation resolution was 30, the mass selectivity of precursor fragmentation was 70, and fragmentation efficiency was 60 %. In Chapter 7, through lowering bath gas pressure the isolation resolution was improved to 100 and the mass selectivity of resonant excitation to 230. The fragmen-tation efficiency was approximately 30%. Chapter 7 also demonstrated and explained the surprising result that the mass resolution of resonant excitation is more dependent upon the excitation amplitude than the bath gas pressure. Numerous additions to the "next generation" of the L I T / T O F M S were alluded to throughout the thesis. Perhaps the best way to summarize sug-gestions for future work is to return to the introduction and the premise that the L I T / T O F M S was designed to present an alternative to, and hence an im-provement upon, existing tandem mass spectrometers. The L I T / T O F M S has a sensitivity advantage over the I T / T O F M S , deriving from the high injection efficiency and storage capacity of the linear ion trap. The L I T / T O F M S also has a sensitivity advantage over the Q-TOFMS, as precursor selection via a broadband waveform ensures 100% ion transmission whereas the selective 177 mass filter of the Q-TOFMS causes some ion loss. A further advantage of the L I T / T O F M S over the Q-TOFMS is the demonstrated capability to complete multiple steps of tandem mass spectrometry. It is also worth noting that the total time scale for M S / M S is, relative to existing instrumentation, "rapid", that is 20 ms for a total scan, compared to ~ 100 ms and up for the more commonly used 3-d ion trap and triple quadrupole mass spectrometers. This fast scanning makes it an ideal M S / M S system for coupling with fast liquid based separation techniques such as capillary electrophoresis. One obvious area in which the present L I T / T O F M S is lagging behind the I T / T O F M S and the Q-TOFMS is mass resolution. The relatively poor resolution limits the mass accuracy which would be desired from a tandem mass spectrometer focused on biomedical applications. The addition of an ion mirror to the T O F mass analyzer would allow for "reflectron" type reso-lution, i.e., 10,000, to be achieved. The improved resolution from a reflectron would also provide a more sensitive probe of the radial spatial spread of the beam exiting the linear ion trap. Consequently, in addition to improving the instrument, high resolution in the T O F M S could possibly aid in a deeper understanding of the effects of resonant excitation on the trajectories of both the precursor and fragment ions. Another instrumental modification discussed in the thesis was the addi-tion of an axial field to the linear ion trap. The axial field would "push" ions out of the quadrupole potential, thus the extraction of ions from the trap would be both more rapid and more effective. If there was an axial field produced by a segmented rod set, fragmentation could be induced by sequentially increasing the potential on the segments (i.e., tandem in space CID), by resonant excitation (i.e., tandem in time CID) or by a combination of the two techniques. The segmented rod set with a method to apply res-178 onant excitation to the rods, and trapping capabilities, would allow greater control and variation of collision energies than either of these techniques in-dividually. The resultant increase in observed fragmentation channels could provide a broader understanding of ion structure as well as a higher prob-ability of producing the fragments necessary to attain a de novo peptide sequence. Furthermore, axial fields combined with trapping could be used to modulate the extraction of the beam from the linear ion trap into the source region of the T O F M S , effecting a higher duty cycle and a lower transmission loss due to slow ions. This thesis also demonstrates the power of the linear ion trap indepen-dent of its coupling to either the electrospray ionization source or a time of flight mass analyzer. The simple modifications to the quadrupole rod set described in Chapter 5 could be implemented in RF-only linear quadrupoles in existing instruments, adding, or improving existing mechanisms of ion isolation and fragmentation. Examples of this application proposed in the literature include acting as a mass selective ion bottle for the 3-d ion trap and the ICRMS, as well as adding MS" capabilities to the triple quadrupole and Q-TOF mass spectrometers. Similarly, one focus of future work in the L I T / T O F M S should be different modes of operations. In theory, all man-ners in which resonant excitation is used in a 3-d ion trap can be applied to the linear ion trap. One example of this mentioned throughout the thesis is using tailored waveforms developed for the 3-d trap for precursor selection in the linear ion trap. Further possibilities include examining the differences between dipolar and quadrupolar resonant excitation, fragmentation by al-ternate techniques, such as square wave forms, and broadband excitation, 2 0 6 and examining the potential effects and advantages of adding non linear field components. 179 In closing, this thesis provides proof of principle for completing ion iso-lation and fragmentation on ions stored in RF-only quadrupoles. It demon-strates the power of the linear ion trap, incorporating it into the design of a new tandem mass spectrometer by coupling the modified rod set with a T O F mass analyzer. This first instrument has proved the potential of the linear ion trap. Hopefully this potential will be realized in the near future. 180 Bibliography [1] M . Karas and F. Hillenkamp. Anal. Chem. 60, 2299-2301 (1988). [2] J . B. Fenn, M . Mann, C. K . Meng, S. F. Wong, and C. M . Whitehouse. Science 246, 64-71 (1989). [3] A . R. Dongre, J . K . Eng, and J . R. Yates III. Trends in Biotech. 15, 418-425 (1997). [4] K . Biemann and H. A . Scoble. Science 237, 992-998 (1987). [5] D. F. Hunt, J . R. I. Yates, J . Shabanowitz, S. Winston, and C. R. Hauer. Proc. Natl. Acad. Sci. U. S. A 83, 6233-6237 (1986). [6] J . R. Yates III, J . K . Eng, and A. L. McCormack. Anal. Chem. 67, 3203-3210 (1995). [7] W. P. Blackstone and M . P. Weir. Trends in Biotech. 17,121-127 (199). [8] H. R. Morris, T. Paxton, A . Dell, J . Langhorne, M . Berg, R. S. Bordoli, J . Hoyes, and R. H. Bateman. Rap. Comm. Mass Spectrom. 10, 889-896 (1996). [9] J . M . S. Henis. Ion Cyclotron Resonance Spectrometry. In Ion-Molecule Collision Vol. II. J . L. Franklin (Ed.) Plenum, New York (1972). [10] R. E . March and R. J . Hughes. Quadrupole Storage Mass Spectrometry, volume 102 of Chemical Analysis. John Wiley, Toronto (1989). [11] P. H . Dawson. Quadrupole Mass Spectrometry and its Applications. Elvesier, Amsterdam (1976). 181 [12] F. Major and H. G. Dehmelt. Phys. Rev 170, 91- (1968). [13] H. G . Dehmelt. Adv. Atom. Mol. Phys. 4, 109-154 (1968). [14] D. Gerlich. Advances in Chemical Physics.LXXlI, 1-176 (1992). [15] H. G. Dehmelt. Adv. Atom. Mol. Phys. 3, 53-72 (1967). [16] R. J . Cotter. Anal. Chem. 64, 1027A-1039A (1992). [17] M . Guilhaus. J. Mass Spectrom. 30, 1519-1532 (1995). [18] C. Weickhardt, F. Moritz, and J. Grotemeyer. Mass Spectrom. Rev. 15, 139-162 (1996). [19] M . L . Vestal and R Juhasz. J. Am. Soc. Mass Spectrom. 9, 892-911 (1998). [20] R. J . Cotter. Time-of-Flight Mass Spectrometry : Instrumentation and Applications in Biological Research. ACS, Washington, D. C. (1998). [21] A. E. Cameron and D. F. Eggers Jr. Rev. Sci. Instrum. 19, 605-607 (1948). [22] W. C. Wiley and I. H . McLaren. Rev. Sci. Instrum. 26, 1150-1157 (1955). [23] W. C. Wiley. Science 124, 817-820 (1956). [24] M . L . Vestal, P. Juhasz, and S. A . Martin. Rap: Comm. Mass Spec-trom. 9, 1044-1050 (1995). [25] M . Karas. J. Mass Spectrom. 32, 1-3 (1997). [26] B. A . Mamyrin, V . I. Karataev, and D. V . Shmikk. Soviet Phys. J. E, T. P. 37, 45-48 (1973). [27] T. Bergmann, T. P. Martin, and H . Shaber. Rev. Sci. In-strum. 63, 2585-2591 (1990). [28] P. Kebarle. Anal. Chem. 65, 972A-980A (1993). 182 [29] R. Bakhtiar, S. A. Hofstadler, and R. D. Smith. J. Chem. Ed. 73, A118-A123 (1996). [30] A . P. Bruins. Mass Spectrom. Rev. 10, 53-77 (1991), [31] S. A . Hofstadler, R. Bakhtiar, and R. D. Smith. J. Chem. Ed. 73, A82-A88 (1996). [32] R. B . Cole. (Ed). Electrospray Ionization Mass Spectrometry, Funda-mentals, Instrumentation, and Applications. John Wiley, New York (1997). [33] J . G . Boyle and C. M . Whitehouse. Anal. Chem. 64, 2084-2089 (1992). [34] A . Dodonov, I. V . Cherunushevich, and V . V . Laiko. In Time of Flight Mass Spectrometry. R. Cotter (Ed.) ACS, Washington D.C. (1991). [35] O. A . Mirgorodskaya, A . A. Shevchenko, I. V . Chernushevich, A . F. Dodonov, and A. I. Miroshnikov. Anal. Chem. 66, 99-107 (1994). [36] A . N . Verentchikov, W. Ens, and K . G . Standing. Anal. Chem. 66, 126-133 (1994). [37] A . Krutchinsky, I. V . Chernushevich, V . L . Spicer, W. Ens, and K . G . Standing. J. Am. Soc. Mass Spectrom. 9, 569-579 (1998). [38] M . C. Fitzgerald, I. V . Chernushevich, K . G. Standing, S. B. H . Kent, and C. P. Whitman. Proc. Natl. Acad. Sci. USA 93, 6851-6856 (1996). [39] X . - J . Tang, C. F. Brewer, S. Saha, I. V . Chernushevich, K . G. Standing, and W. Ens. Rap. Comm. Mass Spectrom. 8, 750-754 (1994). [40] M . C. Fitzgerald, I. V . Chernushevich, K . G. Standing, S. B . H . Kent, and C. P. Whitman. J. Am. Chem. Soc. 117, 11075-11080 (1995). [41] J . G . Boyle, C. M . Whitehouse, and J . B. Fenn. Rap. Comm. Mass Spectrom. 5, 400-405 (1991). [42] R. Grix, U . Gruner, G . L i , H . Stroh, and H. Wollnik. Int. J. Mass Spectrom. and Ion Proc. 93, 323-330 (1989). 183 [43] J . H. J . Dawson and M . Guilhaus. Rap. Comm. Mass Spectrom. 3, 155— 159 (1989). C. H. Sin, E. D. Lee, and M . L . Lee. Anal. Chem. 63, 2897-2900 (1991). I. V . Chernushevich, W. Ens, and K . G. Standing, in Electrospray Ionization Mass Spectrometry. R. Cole (Ed.) John Wiley, New York (1997). A. N . Kruchinsky, A . V . Loboda, V . L. Spicer, R. Dworswak, W. Ens and K . G. Standing. Rap. Comm. Mass Spectrom. 12, 508-518 (1998). D. P. Myers, G . L i , P. Yang, and G. M . Hieftje. J. Am. Soc. Mass Spectrom. 5, 1008-1016 (1994). D. P. Myers and G. M . Hiejftje. Microchemical Journal 48, 259-277 (1993). S. M . Micheal, M . Chien, and D. M . Lubman. Rev. Sci. In-strum. 63, 4277- (1992). S. M . Micheal, B. M . Chien, and D. M . Lubman. Anal. Chem. 65, 2614-2620 (1993). B. M . Chien, S. M . Micheal, and D. M . Lubman. Int. J. Mass Spectrom. and Ion Proc. 131, 149-179 (1994). M . A . Qian and D . M . Lubman. Anal. Chem. 67, 237A-242A (1995). R. W. Purves and L. L i . J. Am. Soc. Mass Spectrom. 8, 1085-1093 (1997). L. L i and R. W. Purves. J. Microcolumn Separations 7, 603-610 (1995). V . M . Doroshenko and R. J . Cotter. Rap. Comm. Mass Spec-trom. 33, 305-308 (1998). S. T. Fountain, H . Lee, and D. M . Lubman. Rap. Comm. Mass Spec-trom. 8, 487-494 (1994). B. M . Chien and D. M . Lubman. Anal. Chem. 66, 1630-1636 (1994). 184 [58] M . G . Qian, Y . Zhang, and D. M . Lubman. Rap. Comm. Mass Spec-trom. 9, 1275-1282 (1995). [59] M . G. Qian and D. M . Lubman. Rap. Comm. Mass Spectrom. 10,1911— 1920 (1996). [60] D. J . Douglas and J . B . French. J. Am. Soc. Mass Spectrom. 3, 398-408 (1992). [61] A . V . Tolmachev, I. V . Chernushevich, A . F. Dodonov, and K . G . Standing. Nucl. Instr. and Meth. in Phys. Res. B. 124, 112-119 (1997). [62] H . R. Morris, T. Paxton, M . Panico, R. McDowell, and A . Dell. J. Prot. Chem. 16, 469-479 (1997). [63] A . Shevchenko, I. Chernushevich, W. Ens, K . G. Standing, B. Thom-son, M . Wilm, and M . Mann. Rap. Comm. Mass Spectrom. 11, 1015— 1024 (1997). [64] M . N . Eberlin. Mass Spectrom. Rev. 16, 113-145 (1997). [65] J . S. Brodbelt, H . I. Kennttamaa, and R. G. Cooks. Org. Mass Spec-trom. 23, 6-9 (1988). [66] A . Loboda, A . Krutchinsky, V . Spicer, W. Ens, and K . Standing. Selec-tive collision induced dissociation by resonance excitation an RF-only quadrupoles at elevated pressures. In Proceedings of the 46th ASMS Conference on Mass Spectrometry and Allied Topics, Orlando, F L . , June 1998, page 37. [67] A . Dodonov, V . Kozlovsky, A . Loboda, V . Raznikov, I. Sulimenkov, A. Tolmachev, A . Kraft, and H . Wollnik. Rap. Comm. Mass Spec-trom. 11, 1649-1656 (1997). [68] P. V . Bondarenko and R. D. Macfarlane. Int. J. Mass Spectrom. and Ion Proc. 160, 241-258 (1997). [69] E. Clayton and R. H. Bateman. Rap. Comm. Mass Spectrom. 6, 719-720 (1992). 185 [70] C. F . Ijames. A proposed two dimensional quadrupole/electrostatic ion trap time-of-flight mass spectrometer. In Proceedings The 44th ASMS Conference on Mass Spectometry and Allied Topics, Portland, Oregon, page 795. American Society for Mass Spectrometry, May 1996. [71] C. M . Whitehouse, T. Dresch, and B. A . Andrien. P C T W O 98/06481 Multipole ion guide ion trap mass spectrometry, 1997. [72] W. Paul, H . P. Reinhard, and U . von Zahn. Z. Phys. 152, 145-182 (1958). [73] E. Fischer. Z. Phys. 156, 1-26 (1959). [74] M . Welling, H . A . Schuessler, R. I. Thompson, and H . Walther. Int. J. Mass Spectrom. and Ion Proc. 172, 95-114 (1998). [75] J . T. Watson, D. Jaouen, H . Mestdagh, and C. Rolando. Int. J. Mass Spectrom. and Ion Proc. 93, 225-235 (1989). [76] R. F. Reinsfelder. Int. J. Mass Spectrom. and Ion Proc. 37, 241-250 (1981). [77] L. Cousins. P C T WO98/52209, RF-only mass spectrometer with aux-iliary excitation, 1998. [78] R. F. Wuerker and R. V . Langmuir. U.S.Patent 3,147,445 Quadrupole focusing means for charged particle containment, 1959. [79] M . E. Bier and J. E . P. Syka. U.S. Patent 5,420,425 Ion trap mass spectrometer system and method, 1994. [80] J . E . P. Syka and W. J . Fies Jr. U.S. Patent 4,755,670. Fourier trans-form quadrupole mass spectrometer and method, 1986. [81] D. J . Douglas. U.S. Patent 5,179,278 Multipole inlet system for ion traps, 1993. [82] M . W. Senko, C. L. Hendrickson, M . R. Emmett, D. -H. Shi, and A . G . Marshall. J. Am: Soc. Mass Spectrom. 8, 970-976 (1997). 186 Y . Huang, G.-Z. L i , S. Guan, and A. G . Marshall. J. Am. Soc. Mass Spectrom. 8, 962-969 (1997). H. A . Schuessler, E . C. Benck, and J. Lassen. Hyp. Int. 81, 263-273 (1993). D. A . Church. J. Appl. Phys. 40, 3127-3134 (1969). J . D. Prestage. U.S. Patent 5,420,549 Extended linear ion trap fre-quency standard apparatus, 1994. G. G. Dolnikowski, M . J. Kristo, C. G . Enke, and J . T. Watson. Int. J. Mass Spectrom. and Ion Proc. 82, 1-15 (1988). J . D. Prestage, G . J. Dick, and L. Maleki. J. Appl. Phys. 6b, 1013-1017 (1989). J . D. Prestage, G. J . Dick, and L . Maleki. IEEE Trans, on Instrum. and Meas. 40, 132-136 (1991). M . Tachikawa, M . Kajita, and T. Shimizu. IEEE Trans, on Instrum. and Meas. 42, 281-286 (1993). J . D. Prestage, G. R. Janik, G. J . Dick, and L. Maleki. IEEE Trans, on Ultrasonics, Ferroelectrics and Frequency Control 37, 535-542 (1990). J . D. Prestage, R. T. Tjoelker, G . J . Dick, and L. Maleki. IEEE Fre-quency Control Symposium 148-149 (1993). P. T. H. Fisk, M . A. Lawn, and C. Coles. Appl. Phys. B 57, 287-291 (1993). M . Welling, R. I. Thompson, and H . Walther. Chem. Phys. Lett. 253, 37-42 (1996). M . G . Raizen, J . M . Gilligan, J . C. Bergquist, W. M . Itano, and D. J . Wineland. Phys. Rev. A. 45, 6493-6501 (1992). M . G. Raizen, J . M . Gilligan, W . . M . Bergquist, W. M . Itano, and D. J . Wineland. J. Mod. Optics 39, 233-242 (1992). 187 [97] M . E. Poitzsch, J . C. Bergquist, W. M . Itano, and D. J . Wineland. Rev. Sci. Instrum. 67, 129-134 (1996). [98] G. R. Janik, J . D. Prestage, and L . Maleki. J. Appl. Phys. 67, 6050-6055 (1990). [99] R. K . Melbourne, J. D. Pretage, and L. Meleki. J. Appl. Phys. 69, 2768-(1991). [100] Y . - L . Chen, J . M . Campbell, B. A . Collings, L . Konermann, and D. J . Douglas. Rap. Comm. Mass Spectrom. 12, 1003-1010 (1998). [101] C. Beaugrand, D. Jaoen, and H. Mesdagh. Anal. Chem. 61, 1447-1453 (1989). [102] J . C. Schwartz, K . L. Schey, and R. G. Cooks. Int. J. Mass Spectrom. and Ion Proc. 101, 1- (1990). [103] W C , M . Nappi, C. D. Cleven, H. Wollnik, and R. G. Cooks. Rap. Comm. Mass Spectrom. 10, 742-746 (1996). [104] Y . - L . Chen, B. Collings, and D. J. Douglas. J. Am. Soc. Mass Spec-trom. 8, 681-687 (1997). [105] D. A. Dahl. The SIMION User's Manual Version 6.0, 1995. [106] M . S. Child. Molecular Collision Theory. Dover, New York (1961). [107] T. Covey and D. J. Douglas. J. Am. Soc. Mass Spectrom. 4, 616-623 (1993). [108] G. Javahery and B . Thomson. J. Am. Soc. Mass Spectrom. 8, 697-702 (1997). [109] R. F. Wuerker, H . Shelton, and R. V . Langmuir. J. Appl. Phys. 30, 342-349 (1959). [110] M . Morris, P. Thibault, and R. K . Boyd. J. Am. Soc. Mass Spec-trom. 5, 1042-1063 (1994). [I l l ] B. A . Thomson. Can. J. Chem. 76, 499-505 (1998). 188 [112] B. A . Mansoori, E . W. Dyer, C. M . Lock, K . Bateman, R. K . Boyd, and B. A . Thomson. J. A m . Soc. Mass Spectrom. 9, 775-788 (1998). [113] Y . - T . L i , Y . - L . Hsih, J . D. Henion, and B . Ganem. J. A m . Soc. Mass Spectrom. 4, 631-637 (1993). [114] R. Feng and Y . Konishi. J. A m . Soc. Mass Spectrom. 4, 638-645 (1993). [115] Y . Konishi and R. Feng. Biochemistry 33, 9706-9711 (1994). [116] V . Katta and B. T. Chait. J. A m . Chem. Soc. 113, 8534-8535 (1991). [117] M . Jaquinod, E. Leize, N . Potier, A . - M . Albrecht, A . Shanzer, and A. V . Dorsselaer. Tetrahedron Lett. 34, 2771-2774 (1993). [118] T. Takano. J. Mol. Biol. 110, 537- (1977). [119] S. V . Evans and G. D. Brayer. J. Mol. Biol. 213, 885- (1990). [120] S. A . McLuckey and R. S. Ramsey. J. A m . Soc. Mass Spectrom. 5, 324-327 (1994). [121] J . L. Stephenson Jr., G . J . Van Berkel, and S. A . McLuckey. J. A m . Soc. Mass Spectrom. 8, 637-644 (1997). [122] D. S. Gross, Y . Zhao, and E. R. Willams. J. A m . Soc. Mass Spec-trom. 8, 519-524 (1996). [123] C. L. Hunter, G. A . Mauk, and D. J . Douglas. Biochemistry 36, 1018-1025 (1997). [124] L. Konermann, F. I. Rosell, A. G. Mauk, and D. J . Douglas. Biochem-istry 36, 6448-6454 (1997). [125] P. D. Schnier, D. S. Gross, and E. R. Williams. J. A m . Chem. Soc. 117, 6747-6757 (1995). [126] P. D. Schnier, D. S. Gross, and E. R. Williams. J. A m . Soc. Mass Spectrom. 6, 1086-1097 (1995). [127] D. E. Clemmer, R. R. Hudgins, and M . F. Jarrold. J. A m . Chem. Soc. 117, 10141-10142 (1995). 189 [128] D. E. Clemmer and M . F. Jarrold. J. Mass Spectrom. 32, 577-592 (1997). [129] K . B . Shelimov, B) E. Clemmer, R. R. Hudgins, and M . F . Jarrold. J. Am. Chem. Soc. 119, 2240-2248 (1997). [130] S. J . Valentine, J . G. Anderson, A . D. Ellington, and D. E. Clemmer. J. Phys. Chem. B. 101, 3891-3900 (1997). [131] K . B . Shelimov and M . F . Jarrold. J. Am. Chem. Soc. 119, 2987-2994 (1997). [132] A. L. Rockwood, M . Busman, and R. D. Smith. Int. J. Mass Spectrom. and Ion Proc. 102, 1991- (1991). [133] Y . Moriwaki, M . Tachikawa, and T. Shimizu. Jpn. J. Appl. Phys. 3 5 , 757-760 (1996). [134] R. D. Knight and M . H. Prior. J. Appl. Phys. 5 0 , 3044-3049 (1979). [135] H . Schaaf, U . Schmeling, and G. Werth. Appl. Phys. 25, 249-251 (1981). [136] Y . Moriwaki, M . Tachikawa, Y . Maeno, and T. Simizu. Jpn. J. Appl. Phys. 31, L1640-L1643 (1992). [137] S. Gronert. J. Am. Soc. Mass Spectrom. 9 845-848 (1998). [138] B. D. Nourse and H. I. Kenttaamaa. J. Phys. Chem. 94, 5809-5812 (1990). [139] J . R. Bruce and J . R. Eyler. J. Am. Soc. Mass Spectrom., 727-733 (1992). [140] C. Basic, J . R. Eyler, and R. A . Yost. J. Am. Soc. Mass Spec-trom. 3, 716-726 (1992). [141] D. E. Goeringer and S. A . McLuckey. J. Chem. Phys. 104, 2214-2221 (1996). [142] K . Taeman. Ph. D. Thesis, Dept. of Physics, McGi l l University, 1997. 190 M . D. N . Lunney, f. Buchinger, and R. B . Moore. J. Mod. Opt. 39, 349-360 (1992). M V , R. J . Champeau, A . Crubellier, J . O. Gaardsted, D. Marescaux, and D. Pavolini. J. Phys. B. 27, 905-912 (1994). S. W. Benson. Thermochemical Kinetics : Methods for the Estimation of Thermochemical Data and Rate Parameters John Wiley, Toronto (1976). J . D. Willams, K . A . Cox, R. G. Cooks, S. A . McLuckey, K . J . Hart, and D. E. Goeringer. Anal. Chem. 66, 725-729 (1994). L. L . Shen and J . Hermans. Biochemistry 11, 1845-1850 (1972). J . M . Campbell, B . A . Collings, and D. J . Douglas. Rap. Comm. Mass Spectrom. 12, 1463-1474 (1998). I. V . Chernushevich, A . N . Verentchikov, W. Ens, and K . G. Standing. J. Am. Soc. Mass Spectrom. 7, 342-349 (1996). M . Guilhaus. J. Am. Soc. Mass Spectrom. 5, 588-595 (1994). T. Dresch. U.S. Patent 5,654,544 Mass resolution by angular align-ment of the ion detector conversion surface in the time-of-flight mass spectrometers with electrostatic steering deflectors, 1996. M . Park and C. Koster. U.S. Patent 5,696,375 Multideflector, 1997. M . Guilhaus, V . Mlynski, and D. Selby. Rap. Comm. Mass Spec-trom. 11, 951-962 (1997). M . Guilhaus and J . H . Dawson. U.S. P 5,117,107 Mass spectrometer, 1988. A. N . Verentchikov. Personal communication, 1998. P. H . Dawson and J . E. Fulford. Int. J. Mass Spectrom. and Ion Proc. 42, 195-211 (1982). [157] T. Dresch, E. E. Gulcicek, and C. Whitehouse. U.S. Patent 5,689,111 Ion storage time-of-flight mass spectrometer, 1997. 191 [158] A. G . Marshall, C. L. Hendrickson, and G. S. Jackson. Mass Spectrom. Rev. 17, 1-35 (1998). [159] D. E . Goeringer, W. B. Whitten, J . M . . Ramsey, S. A . McLuckey, and G. L. Glish. Anal. Chem. 64, 1434-1439 (1992). [160] L. D. Landau and E. M . Lifshitz. Mechanics. Pergamon Press, London (1960). [161] V . M . Dorshenko and R. J . Cotter. Rap. Comm. Mass Spectrom. 8, 766-776 (1994). [162] N . A . Arnold, G . Hars, and H . L. C. Meuzelaar. J. Am. Soc. Mass Spectrom. 5, 676-688 (1994). [163] A. A . Makarov. Anal. Chem. 68, 4257-4263 (1996). [164] E . W. MacDaniel. Collision Phenomena in Ionized Gases. John Wiley, New York (1964). [165] G. Gioumousis and D. P. Stevenson. J. Chem. Phys. 29, 294-299 (1958). [166] J . L. Beauchamp. J. Chem. Phys. 46, 1231-1243 (1967). [167] R. E. March. Rap. Comm. Mass Spectrom. 12, 1543-1560 (1998). [168] A. P. French. Vibrations and Waves. Norton, New York (1971). [169] E. M . Marzluff, S. Campbell, M . T. Rodgers, and J . L . Beauchamp. J . Am. Chem. Soc. 116, 6947-6948 (1994). [170] R. G. Cooks. In Collision Spectropscopy. R. Cooks (Ed.) Plenum Books, New York, 257-450 (1978) [171] J . Charles, S. A . McLuckey, and G. L. Glish. J. Am. Soc. Mass Spec-trom. 5, 1031-1041 (1994). [172] J . E . Fulford, D. -N. Hoa, R. J . Hughes, R. E. March, R. F . Bonner, and G. J . Wong. J. Vac. Sci. Technol. 17, 829-835 (1980). 192 [173] M . A. Armitage, J . E . Fulford, D.-N. Hoa, R. J . Hughes, and R. E . March. Can. J. Chem. 57, 2108-2113 (1979). [174] J . N . Louris, R. G. Cooks, J . E . P. Syka, P. E. Kelley, G. C. Stafford Jr., and J . F. J . Todd. Anal. Chem. 59, 1677-1685 (1987). [175] R. E . March, F. A . Londry, S. Fontana, S. Catinella, and P. Traldi. Rap. Comm. Mass Spectrom. 7, 929-934 (1993). [176] R. G . Cooks, G . L . Glish, S. A . McLuckey, and R. E . Kaiser. Chem. and Eng. News M a r c h 25, 26-41 (1991). [177] M . Splendore, F. A . Londry, R. E. March, R. J . S. Morrison, P. Perrier, and J . Andre. Int. J. Mass Spectrom. and Ion Proc. 156, 11-29 (1996). [178] R E. March, A . W. McMahon, F. A . Londry, R. L . Alfred, J . F . J . Todd, and F. Vedel. Int. J. Mass Spectrom. and Ion Proc. 95, 119-156 (1989). [179] H. A . Bui and R. G. Cooks. Int. J. Mass Spectrom. and Ion Proc. 33, 297-304 (1998). [180] C. Ma, H . Lee, and D. M . Lubman. Appl. Spectrosc. 46, 1769-1779 (1992). [181] B. A . Thomson, D. J . Douglas, J . Carr, J . W. Hager, and C. L . Jolliffe. Anal. Chem. 67, 1696-1704 (1995). [182] M . E. Bier and J . C. Schwartz. In Electrospray Ionization Mass Spec-trometry. R. Cole (Ed.) John Wiley, New York, 235-290 (1997). [183] V . A . Sarurkar and A. G . Menon. Rap. Comm. Mass Spectrom. 13, 469-473 (1999). [184] M . Manolescu. Personal communication, 1996. [185] J . Gronowska, C. Paradisi, P. Traldi, and U . Vettori. Rap. Comm. Mass Spectrom. 4, 306-313 (1990). [186] S. A . McLuckey, D. E . Goeringer, and G . L . Glish. J. Am. Soc. Mass Spectrom. 2, 11-21 (1990). 193 [187] S. Guan and A. G . Marshall. Anal. Chem. 65, 1288-1294 (1993). [188] K . J . Hart and S. A . McLuckey. J. Am. Soc. Mass Spectrom. 4, 250-259 (1994). [189] J . V . Johnson, R. A . Yost, P. E. Kelley, and D. C. Bradford. Anal. Chem. 62, 2162-2172 (1990). [190] G. C. Eiden, C. J . Barinaga, and D. W. Koppenal. J. Am. Soc. Mass. Spectrom. 7, 1161-1171 (1996). [191] P. Liere, T. Blasco, R. E. March, and J.-C. Tabet. Rap. Comm. Mass Spectrom. 8, 953-956 (1994). [192] D. J . Douglas. J. Phys. Chem. 86, 185-191 (1982). [193] S. A . Lammert and J . M . Wells. Rap. Comm. Mass Spectrom. 10, 361-371 (1996). [194] J. F. J . Todd, A . D. Penman, and D. A . Thorner. Rap. Comm. Mass Spectrom. 4, 108-113 (1990). [195] J . C. Schwartz and I. Jardine. Rap. Comm. Mass Spectrom. 6, 313-317 (1992). [196] V . M . Doroshenko and R. J. Cotter. Rap. Comm. Mass Spec-trom. 10, 65-73 (1996). [197] V . M . Doroshenko and R. J . Cotter. US Patent 5696376 Method and apparatus for isolating ions in an ion trap with increased resolving power, 1996. [198] L . Pasa Tolic, J . E . Bruce, Q. P. Lei, G. A . Anderson, and R. D. Smith. Anal. Chem. 70, 405-408 (1998). [199] D. J . Burinsky, R. G . Cooks, E . K . Chess, and M . Gross. Anal. Chem. 54, 295-300 (1982). [200] J . N . Louris, J . S. Brodbelt-Lustig, R. G. Cooks, G. L . Glish, and G . J . Van Berkel. Int. J. Mass Spectrom. and Ion Proc. 96, 117-121 (1990). 194 [201] T. Lin and L . Glish. Anal. Chem. 70, 5162-5265 (1999). [202] L. Cousins. Personal communication, 1998. [203] J . M . Campbell, B . A . Collings, and D. J . Douglas. Canadian Patent Application 2,255,188 Method and Apparatus for Multiple Stages of Mass Spectrometry, 1998. [204] V . M . Doroshenko and R. J . Cotter. Anal. Chem. 34, 2180-2187 (1995). [205] V . M . Doroshenko and R. J . Cotter. Anal. Chem. 68, 463-472 (1996). [206] R. Asam, Michael, K . L. Ray, and G. L. Glish. Anal. Chem. 70, 1831-1837 (1998). 195 

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.831.1-0061530/manifest

Comment

Related Items