«14‘:ng LIBRARY Michigan State University This is to certify that the thesis entitled SELECTED ION FRAGMENTATION WITH A TRIPLE QUADRUPOLE MASS SPECTROMETER presented by Richard Alan Yost has been accepted towards fulfillment of the requirements for Ph 0 D 0 degree in ChemiStry ML\ Major professor Date July 23, 1979 0-7639 OVERDUE FINES ARE 25¢ PER DAY PER ITEM Return to book drop to remove this checkout from your record. SELECTED ION FRAGMENTATION WITH A TRIPLE QUADRUPOLE MASS SPECTROMETER By Richard Alan Yost A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1979 ABSTRACT SELECTED ION FRAGMENTATION WITH A TRIPLE QUADRUPOLE MASS SPECTROMETER By Richard Alan Yost A triple quadrupole mass spectrometer has been designed and developed for the direct analysis of mixtures and the elucidation of molecular structures. The triple quadru— pole system is a simple and efficient implementation of the selected ion fragmentation technique, whereby several ionic species are generated from a sample, ions of a particular mass are selected for fragmentation, and the resulting fragment ions are mass analyzed. The instrument consists of, in series, a dual chemical ionization/electron impact (CI/El) ionization source, a quadrupole mass filter, an RF-only quadrupole that can be pressurized with a collision gas, a second quadrupole mass filter, and an electron multiplier. The ion fragmentation process is performed by collision-induced dissociation (CID), in which the ion acquires internal energy by collision with a neutral molecule. The RF—only quadrupole collision chamber provides focusing of scattered ions. In this Richard Alan Yost instrument the selected ion fragmentation process can provide enhanced selectivity and discrimination over normal mass spectrometry without significant loss of sensitivity. The instrument is described and its per- formance evaluated. There are several applications of mass spectrometry which benefit greatly from the added information contained in the fragmentation spectrum of each ionic species pro- duced in the source, including the elucidation of molecular structures and the analysis of mixtures. For structure elucidation, the formation and fragmentation of every fragment ion from a compound can be directly determined. A number of examples of the use of selected ion fragmenta- tion for structure elucidation are presented, including a detailed study of the fragmentation of nonan-H—one. Over MOO distinct fragmentations of the molecule are observed, providing valuable information about the fragmentation path- ways of the molecule and the structure of the various frag- ment ions. For mixture analysis, the first mass analyzer can separate the mixture components as their molecular ions, produced by soft ionization (945;, CI) of the sample. The fragmentation spectra of the individual components can then be obtained by fragmenting the selected molecular ions and scanning the second analyzer. The mixture analysis technique is discussed and specific application examples Richard Alan Yost are presented. The identification of isomeric and isobaric compounds in a mixture is demonstrated. Detection limits 15 mole for methane and nitrobenzene are shown. of 10' Chemical noise is virtually eliminated by the selection of both parent and fragment ion. The low-energy CID process which occurs in the center quadrupole is characterized. It is shown to be a highly efficient process for the fragmentation of organic ions. The low-energy process (10-20 eV ion kinetic energy) prob— ably occurs by direct vibrational excitation through momentum transfer, and is a very different process from the electronic excitation process which occurs at high kinetic energy (3-20 keV). The collision of CCl+ with several collision gases produces not only fragment ions but also charge ex— change ions and addition products. The ion products with +, Ar+, CAr+, and ClAr+. Ar as collision gas include 0+, Cl The effect of collision gas identity and pressure, and ion kinetic energy on the collision process is also dis- cussed. The characterization of the CID process provides information of fundamental interest as well as making pos- sible the efficient use of the process for chemical analysis. Triple quadrupole mass spectrometry is shown to be a powerful new technique for mixture analysis and structure elucidation. The simplicity and versatility of the tech- nique, combined with its sensitivity and potential for chemical analysis, all point to a bright future. ACKNOWLEDGMENTS I would like to express my deepest gratitude to Dr. Christie G. Enke, under whose guidance this research was performed. I will always value his friendship and counsel. I acknowledge my guidance committee at Michigan State University, including Drs. Stanley R. Crouch, George E. Leroi, William H. Reusch, and Charles C. Sweeley. I would like to express special thanks to George Leroi for his assistance as second reader and to Stan Crouch for many valuable discussions. I would also like to acknowledge Chuck Sweeley, and Graham Cooks of Purdue for their en- couragement early in the project when it was most needed. This research would have been considerably more dif- ficult without the benefit of preliminary experiments per- formed in the Department of Physical Chemistry at La Trobe University in Bundoora, Victoria, Australia. Special thanks go to "Prof" James D. Morrison for his assistance and advice. Don McGilvery and Dianne McLoughlin are acknowledged for their unselfish help at La Trobe. I am grateful to the Office of Naval Research for fund- ing the triple quadrupole project. I also acknowledge three years of fellowship support from the National Science Foundation, and an American Chemical Society Division of Analytical Chemistry Fellowship sponsored by the Upjohn Company. The support of MSU, including teaching assistant- ships and an L. L. Quill Memorial Fellowship, is gratefully 11 recognized. This research would have been impossible without the help of the departmental shops at MSU. Russ Geyer and Len Eisele of the machine shop deserve special recognition for converting my sketches into a working instrument. I also want to thank for their help our electrical designer Marty Rabb, and Ron Haas, head of the electronics shop. The members of the C. G. Enke & Company research group, past and present, have been the best of colleagues and friends. A special word of thanks to Drs. Tom Last, Erik Carlson, Jim Hornshuh, and Spyros Hourdakis. I thank Ed Darland for sharing his wisdom about quadrupole mass spectrometry, and Tom Atkinson for his assistance in the computer programming game. I extend my best wishes and gratitude to the "mass spec subgroup": Kaz Latven, John Chakel, Phil Hoffman, and J. W. Chai. Each of them has played an important role in the research described here, and will continue to influence the future of triple quadrupole mass spectrometry. Finally, I acknowledge the loving concern and support of my family. My parents, my sister Kathe, and my brother David have all helped make this research worthwhile. Most of all, I thank my wife Katie for her love and inspiration. She has helped me keep my sanity these past few years, and put up with me at even the toughest moments. In addition, she proofread the rough and final drafts of this thesis, and provided invaluable editorial assistance. iii TABLE OF CONTENTS Chapter LIST OF TABLES O O O O O O O O O O O 0 0 LIST OF FIGURES . . . . 1. INTRODUCTION. . . . . . . . . . Organization of the Thesis . . . . Selected Ion Fragmentation Concept. . . . . . . . . . . . Perspective on Mass Spectrometry Overview of Triple Quadrupole Mass Spectrometry. . . . . . . Modes of Operation and Applica- tions I O O I O O O O O O O O O O 2. LOW-ENERGY COLLISION PROCESSES. . . Collision Processes in Mass Spectrometry. . . . . . . . . Effect of Experimental Parameters on Low-Energy CID. . . . . . . . Collision Gas Species. . . . . Collision Gas Pressure . . . . Ion Kinetic Energy . . . . 3. MIXTURE ANALYSIS. . . . . . . . Mixture Analysis Technique . . . Application Examples . . . . . Conclusions. . . . . . . . . . . . A. STRUCTURE ELUCIDATION . . . . . . . Data Available Through Selected Ion Fragmentation . . . . Selected Ion Fragmentation Instrumentation. . . . . . . . . . The Fragmentation of Nonan-u-one . Conclusions. . . . . . . . . . . iv Page 12 12 l7 17 21 22 26 27 29 32 33 3A 37 39 5A Chapter 5. INSTRUMENTATION . . . . . . . . . . . . . . Description of Instrument. . . . . . . Performance of Instrument. Resolution . . . . . . . . . . . CID Efficiency . . . . . . . . . . . . . Sensitivity. . . . . . . . . . . . Structure Elucidation. . . . . . . . . . Conclusions. . . . . . . . . 6. FUTURE WORK . . . . . . . . . . REFERENCES. . APPENDIX A. APPENDIX B. Selected Ion Fragmentation with a Tandem Quadrupole Mass Spectrometer. High Efficiency Collision-Induced Dissociation in an RF—only Quadrupole . . . . . . . . Page 57 63 67 67 68 71 73 714 77 79 82 818 Table B-3 LIST OF TABLES Page (Table l). Expressions describing simulated ion trajectories in RF-only quadrupoles. . . . . . . . . . . . . . . . 87 (Table 2). Effect of ion axial voltage on fragmentation efficiency. . . . . . . . 89 (Table 3). Comparison of CID spectra and ion appearance potentials. . . . . . . 92 vi Figure 2-3 2-H 14-1 LIST OF FIGURES Ion-molecule reactions of 35ClC+ with various collision gases . . . . Effect of collision gas identity on the fragmentation of CH“+ . . . . Relative intensity of CHu+(P) and its fragments (F1) as a function of the square of the collision gas diameter . . Effect of ion axial energy on the relative intensity of the fragments of CHu+. . . . . . Comparison of CID spectra of selected ions in the BI spectrum of a five component mixture with reference CID spectra from pure components . . . . . . . . . . . . . . Collision-induced fragmentations of nonan-U-one from this study (below the diagonal), metastable fragmentations from previous IKES study (38) (above the diagonal), and EI mass spectrum (along the diagonal). . . . . . . . . . . . . vii Page l6 19 20 23 31 A0 Figure 5-3 5-1; Page Major fragmentations of nonan- u—one observed in this CID study. Starred transitions have also been observed in previous metastable studies (38) . . . . . . . . . . . . . . . N2 Three-dimensional fragmentation map for cyclohexane. . . . . . . . . . . . 58 Conceptual diagram of the triple quadrupole mass spectrometer showing each component and its function . . . . . . . . . . . . . . . . . 60 Scale drawing (top view) of triple quadrupole mass spectrometer . . . . . . . . . . . . . . . 6“ Effect of collision gas (N2) pressure on CID efficiency for methane CH“? at ion axial energy of 10 eV. . . . . . . . . . . . . . 70 Structure elucidation of h3+ functional moiety by inter- pretation of CID spectrum and comparison with reference CID spectra. . . . . . . . . . . . . . . . . . 75 viii Figure A-l B-ll Page (Chart I). . . . . . . . . . . . . . . . . 82 (Figure l). CID spectrum of the parent ion (m/c98) of cyclohexanone present as 5% of a mixture . . . . . . . . 82 (Figure 1). Diagram of the tandem quadrupole mass spectrometer system showing the center RF-only quadru— pole CID region. . . . . . . . . . . . . . 86 (Figure 2). Simulated ion trajectory. Ion m/z 36, 25 V peak RF, 0.35 MHz, 1.9 cm diameter rods, 1 eV off— axis energy, time=h0 us. . . . . . . . . . 87 (Figure 3). Relative intensity of the parent ion (P/PO) and individual fragment ions (Pi/PO) as a function of CID pressure (argon) for cyclohexane 8h+. Ion axial voltage=5 V. Peak RF voltage=72 V. . . . . . . . . . . . . . 89 (Figure A). CID spectra of (a) benzene 78+ and (b) n-hexane 86+. Argon CID gas at 2 x 10'“ torr. Ion axial voltage=5 V. Peak RF voltage=250 V. . . . . . . . . . . . . . . 91 ix CHAPTER 1 INTRODUCTION This thesis is a description of the design and charac- terization of a new triple quadrupole mass spectrometer and its application to problems of mixture analysis and structure elucidation. We first introduced the concept of selected ion fragmentation with a triple quadrupole mass spectrometer based on preliminary experiments per- formed in 1977 at La Trobe University in Australia (1). A second publication based on the Australia experiments describes the characterization of the highly efficient low-energy collision process (2). The reader is referred to these two publications (reprinted in the Appendices) for information regarding the preliminary experiments on the La Trobe instrument. The experiments described in the following chapters were all performed on the instrument deve10ped at Michigan State University, which is described in Chapter 5. Organization of the Thesis The thesis is divided into six chapters. This intro- ductory chapter provides background information to help the reader understand the arrangement and significance of the research described in succeeding chapters. This includes a perspective on the field of mass spectrometry, a review of triple quadrupole mass spectrometry, and a description of the Operational modes and applications of the instrument. Chapter 2 describes the characterization of the low- energy collision process which occurs in the center quad- rupole. A brief overview of collision processes in mass spectrometry is followed by a discussion of the effect of experimental parameters on the low-energy collision process. This characterization not only sheds some light on the funda- mental nature of the collisional process, but also provides the necessary background for the analytical application of the technique. The third chapter is a brief description of the mix- ture analysis capabilities of the triple quadrupole system. This chapter, in combination with Chapter 5, comprises a paper submitted for publication. A description of an earlier mixture analysis application appears in the Journal of the American Chemical Society (1). The technique of structure elucidation with the triple quadrupole system is described in Chapter A. This chapter has been prepared for publication in two parts, an overview of the technique, and a description of the nonan-h-one experiments. Another example of a structure elucidation problem appears in Chapter 5. The fifth chapter is a description of the triple quadrupole instrument constructed at Michigan State Uni- versity, and a detailed study of its performance. This chapter has been submitted for publication, and was also presented at a poster session at the 27th Annual Conference on Mass Spectrometry and Allied Topics in Seattle, June 1979. The final chapter includes a statement of the conclu- sions of this work and an outlook on the future of triple quadrupole mass spectrometry. Selected Ion Fragmentation Concept One of the most active research areas in mass spec- trometry today is selected ion fragmentation, in which several ionic species are generated from a sample, ions of a particular mass-to-charge (m/z) are selected for fragmentation, and the resulting fragment ions are mass— analyzed. The fragmentation can occur unimolecularly or as a result of collision-induced dissociation (CID) with residual gas molecules. Because the technique requires two stages of mass separation in tandem (one to select the parent ion and one to analyze the daughter ions result- ing from fragmentation), it has also been dubbed MS/MS. The process of determining the fragmentation of mass- selected ions can provide an added dimension of informa- tion about a sample compared to conventional mass spec- trometric techniques. Perspective on Mass Spectrometry Mass spectrometry was introduced in 1913 (3), but did not come into common usage until the 1950's when interest in the quantitative analysis of petroleum frac- tions spurred its commercial development. Early mass spectrometers were all magnetic sector instruments. Double focusing instruments, in which an electric sector is added in order to focus ions of varying kinetic energy into the magnetic sector, were developed to increase the avail- able mass resolution. It was not until the late 1950's that the quadrupole mass filter was proposed (A), and not until the 1960's were quadrupole instruments commer- cially available. The mass filter, as its name implies, allows ions of a chosen m/z to pass, while rejecting all others. It is constructed of four parallel rods with diagonally opposed pairs coupled together and an RF po- tential applied between the pairs. Under these conditions all ions experience stable trajectories and are trans- mitted. If DC potentials of opposite sign are applied to the pairs of rods as well, only those ions falling in a window of m/z valves are transmitted, and lighter or heavier ions experience unstable trajectories and are lost. A quadrupole mass spectrometer includes an ion source and an ion detector positioned on either end of the mass filter. Despite the fact that the mass spectrometer was com- mercially developed for the quantitation of relatively complex mixtures, in qualitative studies even low levels of impurities can make identification of a sample impos- sible. For this reason, in the late 1950's the gas chromato- graph was coupled to the mass spectrometer (CC/MS) to effect prior separation of multicomponent mixtures (5). Coupling of liquid chromatography to mass spectrometry (LC/MS) has generated considerable interest recently (6) because it minimizes the problems with thermally labile or involatile samples associated with GC. Chromatography provides an extra dimension of information when mass spectra are taken at each retention time interval. However, when mixture separation is performed chromatographically, the time required to separate a component of interest can be need- lessly long. In structure elucidation applications, an extra dimen- sion of information can be provided by the detection of "metastable" ions (ions which are stable only long enough to leave the ion source, and then decompose unimolecularly before analysis). The use of metastable peaks to determine specific fragmentation paths was first suggested in l9h5 (7). The use of metastable data for structure elucidation was extended dramatically with the development of Mass-Analyzed Ion Kinetic Energy Spectrometry, or MIKES (8). MIKES makes use of double-focusing mass spectrometers in which the ions are mass-analyzed by a magnetic sector before frag- -mentation, and the kinetic energy of the resulting ions is analyzed by a succeeding electric sector. This kinetic energy analysis of the ions can provide information on their m/z and the energy released on fragmentation. The added dimension of information that results not only provides valuable assistance in the elucidation of molecu- lar structure, but can also make possible the separation and analysis of mixture components. The most straightforward approach to providing an added dimension of mass spectral information is tandem mass spectrometry (MS/MS), in which two mass analyzers are connected in series. Lindholm in 195“ introduced the concept of using a tandem mass spectrometer with a col- lision chamber between the analyzers to study ion-molecule reactions (9). A dozen or more researchers are now using tandem instruments for this purpose (10). Most of these are tandem sector instruments, although a few tandem quadrupole instruments have been built and applied to ion- molecule reactions (11) and to photodissociation of ions (12,13). Tandem mass spectrometers designed for analytical applications of selected ion fragmentation are currently under development or are operating in the laboratories of Cooks, Hunt, and McLafferty. The work described in this thesis, however,constitutes the first application of tandem quadrupole mass spectrometry to problems in chemical anal- ysis, and in particular, the first use of collision-induced dissociation within a quadrupole field. Overview of Triple Quadrupole Mass Spectrometry The triple quadrupole mass spectrometer developed at Michigan State University uses tandem quadrupole mass filters to provide two stages of mass analysis. The center RF-only quadrupole, positioned between the mass filters, has no mass filtering action, but rather provides strong focusing of the ions during fragmentation. The combina- tion of three quadrupoles produces a simple and efficient MS/MS instrument for selected ion fragmentation. Triple quadrupole instruments were independently designed and constructed for photodissociation studies by two groups in the early 1970's (12,13). It was on one of these instruments, in the laboratory of J. D. Morrison at La Trobe University, that I performed the preliminary experiments described in our earliest papers (1,2). In the photodissociation experiments, the mass-selected ions in the center quadrupole are excited by interaction with photons from a dye laser or xenon lamp. The light ab— sorbance is too low to measure, but the resulting frag- ment ions can be detected. In this way it is possible to study the optical spectroscopy of ions in the gas phase, and to study the fragmentation process using a monoener- getic source of excitation. The photofragmentation ef- ficiency is extremely low, however, and even at low residual 8 torr), the photodissociation signal is pressures (10' swamped by the collision-induced dissociation products. By increasing the pressure in the center quadrupole, this troublesome interference becomes a highly efficient frag- mentation technique for analytical applications. The instrument developed at Michigan State University consists of, in series, an ionization source, a quadrupole mass filter, an RF-only quadrupole collision chamber, another mass filter, and an ion detector. The selected ions are fragmented between the mass filters by collision— induced dissociation (CID) at low kinetic energies of 10-20 eV. The parent ions may be produced by ionization with or without significant fragmentation from gaseous, liquid, and solid samples. Mass analysis over the range 1 to 1000 amu is possible, with resolution of 1 part in 1500. The entire instrument is designed for ease of computer control. Only under complete control of micro- and mini-computers will the ultimate potential of the system for chemical analysis be realized. Modes ofggperation and Application There are several possible modes of Operation for the triple quadrupole mass spectrometer, as described in Chapter 5. Each of these modes finds application in the solution of specific analytical problems. The triple quadrupole instrument may be used as a single stage mass spectrometer by scanning the first mass filter with the second and third mass filters in RF-only (total ion) mode. In this way the instrument may be applied to problems which do not require the added dimension of selected ion fragmentation. This mode can also be used to tune the first quadrupole to the parent peak which is to be fragmented. For analysis of mixtures, the molecular ions for all components can be produced by a soft ionization tech- nique which produces very little fragmentation (345;, chemical ionization), and can then be separated by the first mass analyzer. For one component at a time, the molecular ion can be selected with the first quadrupole, fragmented in the second, and the mass spectrum obtained by scanning the third. If the application requires only the detection of a single component of interest, single reaction monitoring may be employed, in which the first and third quadrupoles are set to select a specific parent ion/daughter ion pair. If several compounds are to be lO detected, then several parent/daughter ion pairs may be selected in sequence using multiple reaction monitoring. A potentially powerful technique for analysis of mixture components is to scan the first and third quadrupoles with a fixed difference in mass. In this way a selected neutral loss which is characteristic of a specific functional group may be monitored, and mixture components which con- tain that group identified. In a similar way, the molecu- lar ions may be monitored which contain a functional group that forms a specific addition product with a selected reactive collision gas. A special case of mixture analysis is the identifica- tion of isotope labelling in samples with less than complete isotopic substitution. The first quadrupole can be set to mass select only those ions which do contain the isotope 1abel(s), and can therefore eliminate the interference from the unlabelled ions. All of the operational modes described for mixture analysis may also be used in solving problems of structure elucidation. In one mode, any fragment ion in a compound's normal mass spectrum may be mass selected by the first quadrupole. The selected fragment can be further frag- mented by collision, and the spectrum of resulting ions obtained by scanning the third quadrupole. In this way the mass spectrum of a particular portion or functional moiety of a molecule can be determined. Alternatively, 11 all fragments which may further fragment to yield a specific ion may be determined by scanning the first quadrupole with the third set for that ion. Neutral loss spectrometry is realized by scanning the two mass filters with a fixed difference in mass, and thus those ions which show loss of a specific neutral may be determined. The availability of several modes of operation makes the triple quadrupole mass spectrometer a versatile instrument for both struc- ture elucidation and mixture analysis. These same modes may be implemented with MIKES instruments; however, they are much simpler in the triple quadrupole due to the in- dependence of the two stages of mass separation. There are many indications that triple quadrupole mass spectrometry will prove to be a significant advance in the field of mass spectrometry due to its simplicity, its efficiency and sensitivity, and its potential for chemical analysis. These indications include the wide- spread interest in the technique at recent scientific meetings and the evaluations of major figures in the mass spectrometry field. The commercial development of triple quadrupole instruments is very likely, and the resulting widespread availability of the technique would further increase the impact of triple quadrupole mass spectrometry on chemical analysis. 12 CHAPTER 2 LOW-ENERGY COLLISION PROCESSES Characterization of the low-energy collision processes which occur in the center quadrupole is an important step in the development of the triple quadrupole mass spec- trometer. This characterization is important both as a fundamental study of the processes and as a method for optimization of the sensitivity and selectivity of the analytical technique. Collision Processes in Mass Spectrometry The role of collision processes has been of concern to mass spectrometrists for a number of years. Indeed, signals corresponding to collision-induced dissociation (the so-called "Aston bands") were observed in the very first mass spectra (3). Since that time there has been a great deal of interest in the collisions between ions and molecules, especially in the fundamental nature of the ion—molecule reaction. Major treatises have been written on the nature of ion-molecule reactions for simple mole- cules (14). From the study of ion-molecule reactions, the important analytical technique of chemical ionization was 13 developed in the mid-1960's (15). Another important analytical technique which evolved from these studies was the use of collision-induced dis- sociation (CID) or collisional activation (CA) for frag- mentation of selected ions as a complementary technique to unimolecular decompositions (metastable studies). The widespread use of the CID technique was catalyzed by a paper of Jennings in 1968 (16). Most of this CID work has been performed by Mass-Analyzed Ion Kinetic Energy Spectrometry (MIKES) with reversed-geometry double focusing mass spectrometers. The present work extends the use of CID to the triple quadrupole system. The ion-molecule collisions in MIKES instruments occur at relatively high ion kinetic energies of 3-10 keV. At these energies the conversion of kinetic into internal energy occurs primarily by a vertical Franck-Condon type of electronic excitation. Fragmentation may occur after the ion and molecule are again separated, through relaxa- tion of the electronic energy into vibrational levels (17). Because the electronic excitation occurs over a time inter- val much shorter than vibrational periods, the ion and molecule each retain their identity afterwards, and col- lision products due to attachment and displacement re- actions are not formed. Ions due to changes in the electric charge of the ion and molecule can be produced, however, including products of charge exchange and charge 1A inversion. Although charge inversion products have been detected in MIKES experiments (18), the collision gas ions which result from charge exchange would not be analyzed even if they managed to enter the second sector. Never- theless, charge exchange can cause a loss of ion signal. Indeed, the cross section for charge exchange and scatter- ing can be relatively large compared to the cross section for fragmentation, and the resultant loss of ion signal can limit the sensitivity of the MIKES technique (19). The collision processes observed in the triple quadru- pole system are very different from those seen in MIKES instruments. The kinetic energy of the ions in the center quadrupole (typically 10-20 eV) is two to three orders of magnitude less than that in MIKES experiments. At these low energies the cross section for electronic excitation is quite low. Fragmentation probably occurs through direct vibrational excitation as a result of momentum transfer (17). Because the ion and molecule may "stick" together for a reasonably long time (several vibrational periods), product ions may result from addition or displacement re- actions. Charge exchange product ions from the collision gas may also result. At high pressures, the product ion from a collision may undergo a second collision and be further changed. In those experiments where it is neces- sary that the observed product ions be the result of only single collisions, it is important to maintain a low l5 collision gas pressure, typically below 10-5 torr. Examples of the types of collision products that are observed with the triple quadrupole mass spectrometer are depicted in Figure 2-1. The collision spectra are those observed when 35ClC+ ions from 001“ are collided with various gases at approximately 10"3 torr and 10 eV axial kinetic energy. In addition to the CCl+ reactant ion, three types of product ions are observed. Both possible fragment ions (C+ and Cl+) are formed with all five collision gases. In addition, charge exchange products are seen with all the collision gases except helium. Examples are Ar+, H2O+, and N+ and N2+. Helium has the highest ionization potential and would be the least likely to undergo charge exchange. A third class of ions, addition products, are also observed. Collision with argon, for example, produces ArC+ and ArCl+; with nitrogen, CN+, CN2+, ClN+, and CClN+ are formed. These ion complexes are stable long enough to reach the detector approximately 50 us after formation. These results, and those which follow, reinforce the conclusion that the low- energy collision process is a very different one from that observed in MIKES. l6 ION-MOLECULE REACTIONS OF CCI+ IOOP CCl’ ‘5 '"” no: CAI' 9‘ :51” I xlO‘ Ar 0 1 Carl C!“ ”63 '00!- "2* K? KP '2N1 ”Nedaflt 3 hi IO gr? I [b l éhm+ 2 ,_ o . II I . l t: (I) IOO- 3 El 7:3 ”03 m3 ‘2 Z " CiO’ 0 -' M63 (3 Lu "0’, ccaot 2 5 g o we I 22 ‘4 .1 .4 Ei C) (I U He I U j I] 0 2'0 4'0 610 som/z Figure 2-1. Ion-molecule reactions of 3501C+ with various collision gases. 17 Effect of Experimental Parameters on Low-Energy CID A major part of the characterization of the low- energy CID process is the study of the effects of varying the experimental parameters. We have already reported the results of a preliminary study of these effects on the triple quadrupole system at La Trobe University (2). This publication is reprinted in Appendix B. The effects of collision gas identity, collision gas pressure, and ion kinetic energy (axial and transverse) will be discussed in the sections which follow. Collision Gas Species An example of the effect that collision gas identity can have on the fragmentation process is shown by the 001+ study in Figure 2-1. The intensities of the 0* and Cl+ fragment ions relative to each other, and to the unfragmented CCl+ peak, vary with different collision gases. The collision-gas peaks which result from charge exchange vary from relatively intense for N2, 02, and H20, to weak for Ar, to undetectable for He. Addition product ions are observed for all the collision gases except helium, and vary with the chemical nature of the collision gas. As 18 indicated above, association product ions are probably not formed in MIKES instruments, and collision gas ions due to charge exchange that may be formed are undetected. Furthermore, the relative intensities of fragment ions in MIKES spectra do not vary when the collision gas is changed (20). This result would be expected for the high energy electronic excitation process described earlier. The effect of changing the collision gas on the low— energy CID of CH“? is illustrated in Figure 2-2. In these experiments the noble gases from helium to xenon, plus nitrogen, were all studied, each at the same pres- sure, approximately 10"3 torr, with an ion axial energy of 10 eV. In Figure 2-2 the intensity of each fragment ion (F1) of CH“? (P) relative to the total ion signal (2F1+P) is plotted against the energy lost in each frag- mentation. The energy loss values are taken from an earlier MIKES study (21). The fragment ion intensities increase for collision with the noble gases from He to Xe, with N2 falling between Ar and Kr. The relative intensity of the C? and CH+ fragments compared to CH2? and CH3+ also increases from He to Xe. The degree of fragmentation appears to be related to either collision gas mass or size. Figure 2—3 shows the relative intensity of each ion plotted against the square of the collision gas diameter. Experi— mental conditions are the same as for Figure 2-2. The data points for each fragment ion are reasonably smooth l9 CH * c t CH" C“ Lo—rz—F‘L 1 l I03— “341114 111 o 2 4 s 3 IO l2 l4 l6(eV) FRAGMENT ENERGY LOSS Figure 2-2. Effect of collision gas identity on the fragmentation of CHut. 20 He Ne Ar N2 Kr Xe 1 PI P+EH I i I I I l I I I l L xii O 4 8' I2 IS 20 24 ( ) (MOLECULAR DIAMETER)2 OF COLLISION GAS Figure 2-3. Relative intensity Of CHu?(P) and its frag- ments (F1) as a function Of the square Of the collision gas diameter. 21 and linear except for the points corresponding to the N2 collision gas, which lie above the line for C? , CH+, and Cfizf. The diatomic N2 molecule is more efficient for fragmentation than would be a noble gas molecule Of the same diameter. As mentioned above, in most MIKES studies it has been found that the collision gas identity has little effect on the appearance of the CID spectrum, although helium is Often cited as more efficient for fragmentation than heavier collision gases (20). Recently, however, COOks has Obtained data which indicate that heavier collision gases are more efficient at both fragmenting and scatter- ing ions (22). The fact that, in some experimental con- figurations, the lighter collision gases show higher over— all CID efficiencies indicates that the fragmentation ef- ficiency does not increase as rapidly as does the scatter- ing loss as the collision gas size is increased. Collision Gas Pressure The effect Of collision gas pressure on the fragmenta- tion process is discussed in Chapter 5. In that chapter, Figure 5-“ shows the effect Of increasing N2 collision gas pressure on the fragmentation Of CHut, Several effects are Observed. As the pressure is increased, the number 22 of ions undergoing collision (and therefore the frag- mentation efficiency) is increased; at higher pressures multiple collisions become important, increasing further the fragmentation efficiency. As the pressure is increased, however, the collection efficiency drops due tO increased scattering. The overall CID efficiency exhibits a maximum at some intermediate pressure dependent on the collision gas and ions involved. Ion Kinetic Energy Finally the effect of ion kinetic energy on the frag— mentation process has been studied. The ion kinetic energy is composed Of two components, the axial energy from the accelerating potential between the quadrupole common and the ion source, and the transverse energy from the quadru— pole RF field. The effect Of ion axial energy on the fragmentation Of CH”? is shown in Figure 2-N. The argon collision gas is at a pressure high enough (10"3 torr) for multiple col- lisions to occur. The formation Of CH3+ is nearly inde- pendent Of axial energy, but loss Of 2,3, and A hydrogens shows increasing axial energy dependence, with a maximum Ct~ production (which requires approximately 15 eV (21)) at about 15 eV. At lower axial energies, the path followed 23 IO O '70 I07.— 4. f U o ff. - I— ' 4. C “>J N 'a d (I l".— ). E: U) 2 DJ p. 2 CH’ OJ9$- l I l OINVG O '20 '40 -60 —80 -'IOO “I20 (6V) ION AXIAL ENERGY Figure 2-U. Effect Of ion axial energy on the relative intensity Of the fragments Of CHut. 2A by the ions is longer due to more orbits in the RF field, and hence more collisions (Of lower energy) occur. As the axial energy is increased, the fragmentation efficiency increases to a maximum, and then decreases as the cross section for momentum transfer drops, as Observed in other momentum transfer reactions (17). Initially, an increase in axial energy increases the amount Of energy available for vibrational excitation. It can be visualized that at higher kinetic energies (greater relative velocities), the ion and gas molecule do not "stick" together long enough for significant vibrational excitation to take place. The ions may pick up transverse kinetic energy from the RF field in the center quadrupole. From digital simulations published earlier Of the ion trajectories in the RF field, an average transverse energy Of a few eV is indicated (2). Furthermore, the simulations indicate that the transverse energy is independent Of both the peak RF voltage and the RF frequency. Experimental data confirm that the RF voltage over a range from 0 to 720 VpRF has no effect on the fragmentation process (2). The total ion kinetic energy is the sum Of the trans- verse and axial energies, and is typically 10-15 eV. The fragmentation process occurs, therefore, by a very ef— ficient conversion Of a portion Of the ion's kinetic energy into internal energy. From these experiments it appears that the low—energy collision process in the center quadrupole 25 can involve a significant transfer Of momentum. Further- more, the ion-molecule interaction can last long enough for a collision complex to form, and a variety Of cOl- lision products, including association products, tO result. The characterization Of the collision process is a con- tinuing project, with special emphasis on the energy transfer process. Included are experiments to determine the effect Of collision gas parameters such as polarity and chemical reactivity, and studies Of isotopic scrambling to elucidate the nature Of the collision complex. CHAPTER 3 MIXTURE ANALYSIS Commercial mass spectrometers were developed in the 1950's to analyze the complex mixtures encountered in petroleum production. The difficulties involved in the analysis Of the resultant spectra led to the increased use Of chromatographic methods, at the expense Of mass spectral techniques. The eventual combination Of mass spectral analysis with gas chromatography (CC/MS) has produced one Of the most powerful and widely accepted analytical methods (5). Interest in alternate techniques has remained active, however, due to the limitations on the thermal stability and volatility Of samples, and the time-consuming nature Of GC. Although difficult to inter- face tO MS, liquid chromatography (LC) has generated considerable interest in recent years as a prior separa- tion technique (6). While LC reduces the problems as- sociated with involatile and thermally labile samples, it does not significantly increase the speed Of separation. This chapter describes the use Of mass spectrometry for both separation and identification Of mixture com- ponents (MS/MS). While MS/MS has been implemented using MIKES instruments (23,2h), this is the first application Of tandem quadrupoles tO mixture analysis. 26 27 Mixture Analysis Technique The molecular ions for all the components of a mixture can be produced by a soft ionization technique (ELEL’ chemical ionization) which results in very little frag- mentation, and can then be separated by the first mass filter. The molecular ion for one component at a time can be selected with the first quadrupole, fragmented in the second, and the third scanned to Obtain the result— ing mass spectrum. The selectivity and discrimination which result from the use Of two stages Of mass separation can Often improve the detection limit Of the mass spec- trometric technique due to the elimination of chemical noise (23). The use Of mass spectral separation Of the components Of a mixture in MS/MS eliminates the time delays as— sociated with the chromatographic separation used in GC/MS and LC/MS. Because all the components are available at any time and in any order, only those components Of par- ticular interest need be analyzed. Each component may be selected for as much or as little time as is required to determine its identity and its quantity. Continuous analysis Of a sample is possible in such applications as atmospheric analysis without the time delays associated with batch sampling for chromatographic analysis. The limitations Of GC/MS regarding involatile and thermally labile samples 28 are largely eliminated by the MS/MS technique. Direct insertion probes and field desorption or chemical desorp- tion could be used to introduce samples which are not amenable to CC separation. The triple quadrupole mass spectrometer is especially versatile for mixture analysis due tO the availability Of the different Operational modes described in Chapter 5. A spectrum Of all the ions produced by soft ionization Of the sample may be Obtained by scanning the first quadru- pole with the second and third in RF-Only (total ion) mode. The ions may then be selected by the first quadru- pole, fragmented, and their mass spectra Obtained by scan- ning the third quadrupole. If specific compounds which produce known fragment ions are to be detected, then mul- tiple reaction monitoring may be employed. In this tech- nique, selected parent ion/daughter ion pairs are sequen- tially selected by the first and third quadrupoles. Those components which contain a specific functional group may be rapidly identified if that functionality produces a characteristic fragment ion, and the third quadrupole is set to that ion while the first quadrupole is scanned. If instead, a characteristic neutral loss is being sought, the first and third quadrupoles may be scanned with a fixed difference in mass. 29 Application Examples Our first publication (1) contained an example Of mix- ture analysis performed on the triple quadrupole mass spectrometer at La Trobe University (see Appendix A). In this study a mixture Of cyclohexane and three minor components (benzene, n-hexane, and cyclohexanone, each present as 5% Of the mixture), was analyzed. The spec- trum Of each mixture component was reproducible and showed good agreement with the spectrum Of the pure compound. An example Of the improvement in detection limit that can result from the elimination Of chemical noise appears in Chapter 5. In this case a sample Of 10 ppm CH“ in N2 was analyzed by monitoring the CH“? -* CH3+ reactfbn. A detection limit Of 20 femtograms was determined, a hundred- fOld improvement over the detection limit Obtained by simply I‘ 2 improvement results from the elimination Of chemical noise monitoring the CH“? , CH3+, or CH ions directly. This caused by the interference of 1“N": , lSNf and residual Of ions. As a third example Of mixture analysis, the following sample containing five components at equal concentration was chosen: two isomers Of molecular weight 11“, 3-hepta- none and n-heptanal; a compound isobaric with the other two at 11”, n-octane; plus cyclohexane (8H) and 2-pentanone (86). The mixture was ionized by 70 eV EI, although CI 30 would have simplified the mass spectrum. The CID spectra Of the 11H+ ion from the individual components show unique fragmentation: 99+ (M-l5) for 3-heptanone, 96+ (M-18) for n-heptanal, and 70+ (CSHIO) for n-octane. In the CID spectrum Of the llI-I+ ion in the mixture, it is possible to detect the individual components despite their being iso— meric or isobaric. Figure 3-1 shows the CID spectrum Of the lllI+ ion in the mixture as well as the reference CID spectra Of lllI+ from the three pure compounds. The two isomers in the mixture may be identified with even greater confidence by Obtaining the CID spectra Of fragment ions in the El spectrum Of the mixture that are unique tO the specific components. The CID spectra Of 99+ and 81+ from the mixture exactly match the spectra Of the frag- ments 99+ from pure 3-heptanone and 81+ from n-heptanal, respectively. The other two components are easier to identify in the mixture; the CID spectra Of pure cyclo- hexane (8A+) and 2-pentanone (86+) are nearly perfect matches with the CID spectra Of these ions from the mix- ture. The CID spectra Obtained for 8A+ and 81+ from the mixture are compared with reference spectra in Figure 3-1. A sixth and unexpected component was discovered in the mixture as well. The electron impact spectrum showed peaks at 120+ and 105+ which could not be attributed to any Of the five known components. The CID spectra showed 31 .mpcocoa IEOO mesa Eopm mpuocam 0H0 mocmpmmcp Sufi: meduxas ucmcanOO O>Hm a no Esppomam Hm on» :H mcofi cmuooacm no mnpomam 9H0 no comfipmaEOo .HIm mpswfim ON. 8. om om ov r P ha I b u b ”A we $21.WJ .8231-.. 2.5 . .82: r h P E a... _l _. 2.9.2.226 co co 3 k n h J — h 8828:.» 2. no «A «s. In; & p *4 . _ @Hfio pox mn 2.1 USE-I...» 6 a. me 8. NE om. Om ow 9. OW Cw. 0-0 firm 0% J1 d O I j am momma .fl Dc 59.. c Eat ..< .6 +HHMU .me If h n b P — mo OF Mul— unexocsgu 3 541 .+flHHU hVEw L \ mecca—col In ON. N\E 00. Ram «figOW—M Q I _ee_r...f.. I... i E. T1. + . 2.9.281- n 9560 I: Eat :\/\/\/\L $.5me wozwmmhmm mZO_ oEbmumm no $.5me ZOFflrZMEOdE 20. 32 fragments from 120+ at 105+, N3+, and 77+ and fragments from 105+ at 77+, 51+ and 26+. Interpretation of these data suggested acetophenone as the impurity, and comparison with the reference CID spectra Of acetophenone confirmed this. Careful study Of the E1 spectra Of the pure com- ponents showed that the acetophenone was present as an impurity in the n—heptanal. A single standard addition experiment indicated an acetophenone concentration Of ap- proximately 5 parts per thousand. This gives an impurity concentration Of 1 part per thousand in the mixture. These results demonstrate the ability Of the triple quadrupole system to identify mixture components, including both isomers and isobaric compounds. Conclusions The technique Of direct analysis of mixtures by triple quadrupole MS/MS has been developed, and it may eventually join GC/MS in widespread usage. The simplicity and versatility Of the technique, combined with its ability to eliminate chemical noise, makes it a powerful alterna- tive to current GC/MS and LC/MS methodology. CHAPTER U STRUCTURE ELUCIDATION An added dimension for structure elucidation is pro- vided by selected ion fragmentation, in which several ion species are generated from a sample, ions Of a particular mass are selected for fragmentation, and the resulting fragment ions are mass analyzed. The use Of two sequen— tial stages Of mass analysis with fragmentation occurring between them makes it possible to determine the routes Of formation and fragmentation Of ions in the mass spectrum. A familiar example is the analysis Of the metastable peaks Observed in the mass spectrum tO show parent/daughter relationships for some Of the ions. A more systematic technique for determining fragmentation pathways is the combination Of collision-induced dissocia- tion (CID) with Mass-Analyzed Ion Kinetic Energy Spec- trometry (MIKES) (25). We recently introduced the concept Of performing selected ion fragmentation with a triple quadrupole mass spectrometer (l). The use Of tandem quadrupole mass filters for mass analysis and the high efficiency Of the low-energy CID process when performed in a center quadru— pole provides an extremely sensitive technique for the 33 34 direct and systematic analysis Of fragmentation pathways. The kinds Of additional analytical information available and its interpretation for elucidation Of structure are discussed in the sections that follow, and the chapter concludes with a thorough analysis Of the fragmentation Of nonan-A-one. Data Available Through Selected Ion Fragmentation The use Of two stages Of mass analysis in tandem makes possible the identification Of the formation and frag- mentation pathways for every ion in a compound's mass spectrum. If no collision gas is present, the unimolecu— lar fragmentations Observed correspond to the metastable ions Often Observed in normal mass spectra. These transi~ tions can provide some information on the genetic relation— ship Of some few ions in the spectrum. Introduction Of a collision gas adds energy to the ions and increases both the number and intensity Of fragmentations that can be Observed. The ions produced by low-energy CID are less likely to show isomerization or rearrangements than are metastable ions, which have higher internal energies from the source (25). Most Of the ions Observed in the low- energy CID spectrum arise from simple cleavages and thus provide useful ’data on the structures Of the original molecule. 35 As an example Of the amount Of data provided about a compound by selected ion fragmentation, consider the results for isopropanol. The electron impact mass spectrum Of i- propanol indicates 33 fragment ions, but gives no informa- tion about their formation or fragmentation. McLafferty catalogs 9 metastable ions Observed for i-propanol (26). Calculation Of possible parent/daughter ion pairs from the metastable data indicates 8 confirmed fragmentations and one set Of two possible transitions which could correspond to the ninth peak. We have measured the CID spectra Of all 3“ ions in the E1 spectrum Of i-propanol and have Observed 201 different fragmentations. The added dimension Of information provided by knowing the formation and frag- mentation Of all the fragment ions Of the compound makes this low-energy CID data extremely valuable for elucida- tion Of structure. The interpretation Of the CID data can provide a number Of different types Of structural information about the sample. Consider the modes Of Operation Of the triple quadrupole system: 1) By scanning the third quadrupole with the first set to pass a specific m/z ion, a spectrum Of all the fragments which arise from the selected parent ion is Obtained. 2) By scanning the first quadrupole with the third set for a specific ion, all the parent ions which can fragment tO form that daughter ion are Observed. 3) All the ions which lose a selected neutral 36 mass (§;g;, H20 or CH3°) are monitored by scanning both mass filters with a fixed mass difference between them. The Object Of the interpretation effort is the assign- ment Of structure tO the major fragment ions and ultimately tO the molecular ion. The interpretation techniques normally applied to EI mass spectra are quite useful for CID spec- tra since CID produces spectra qualitatively similar to those provided by EI. Most Of the CID spectra Obtained, however, are for even-electron fragment ions, whereas mass spectroscopists are most familiar with the fragmentation Of Odd-electron (g;g;, molecular) ions. Another very power- ful approach tO determining the structure Of specific fragment ions is comparison Of their CID spectra with a library Of reference CID spectra Of fragment ions Of known structure. In contrast tO library matching of E1 spectra, which requires a library Of enormous numbers Of reference spectra, the library Of reference spectra for identifica- tion of fragment ions requires relatively few entries. As an example Of the interpretation Of CID data to answer a specific structural question, consider again the structure elucidation Of i-prOpanOl, (CH3)2CHOH. Another possible compound that could give the Observed EI spectrum is acetic acid, CH3COOH. In order tO eliminate acetic acid as a possibility, consider the CID spectrum Of (M-CH3)+ at m/z A5. The CID spectrum Of "5 includes peaks at AH+, ”3+, “2+, and Al+ which indicate the presence Of at least 37 A hydrogens in the fragment. This eliminates COOH+ as the structure Of A5+ and eliminates acetic acid as the compound. The selected ion fragmentation technique is carried to its ultimate when all the fragmentation steps for a compound are completely elucidated. The 201 fragmentations Observed for i—prOpanOl provide a dramatic indication Of the amount Of structural information made available by selected ion fragmentation. Selected Ion Fragmentation Instrumentation The tandem mass spectrometer makes possible the mass analysis Of the fragment ions Of mass-selected parent ions. Two quadrupoles in tandem have been employed to study ion- molecule reactions (11), and triple quadrupole systems have been applied to study the photodissociation Of ions (12,13). The triple quadrupole mass spectrometer used in this study has been described (27). It consists Of, in series, a dual chemical ionization/electron impact (CI/EI) ioniza- tion source, a quadrupole mass filter, an RF-Only quadrupole that can be pressurized for CID, a second quadrupole mass filter, and an electron multiplier. Mass resolution Of one part in 1500 is possible over the entire mass range Of 1-1000 amu. The high efficiency Of each component in the system makes detection limits in the femtomole range 38 possible by monitoring specific fragmentations (27). Prior to the development Of the triple quadrupole mass spectrometer, all selected ion fragmentation for analytical purposes had involved the use Of double-focusing mass spectrometers to determine fragmentations (either metastable on collision-induced dissociations) which occur either between the sectors or before the first sector. For a double-focusing spectrometer with the electric sector pre- ceding the magnetic sector, the technique is called Ion Kinetic Energy Spectrometry (IKES) if the electronic sector voltage is scanned, and high voltage (HV) scanning if the accelerating voltage is scanned. For a spectrometer in which the magnetic sector precedes the electric sector, the technique is called MIKES. All three techniques display peaks that are broadened by kinetic energy loss upon frag- mentation,which makes it difficult tO achieve unit mass resolution. The RV technique has the additional dis— advantage Of a limited range Of fragment mass (neutral loss less than half the parent mass). Nevertheless, both IKES and MIKES have been proven to be powerful techniques for structure elucidation (25,28-31). 39 The Fragmentation Of Nonan-A-One The electron-impact induced fragmentation Of aliphatic ketones has been widely investigated (32). A number Of specialized techniques have been utilized tO help in these studies, including metastable decomposition studies (33-35), ion cyclotron resonance (ICR) (36), and the use Of deuterium labelled compounds (35-38). Nonan-A-One has received par— ticular attention, including a study Of the metastable ions Observed from this compound withtnilon Kinetic Energy (IKE) Spectrometer (38). We have performed a detailed study Of the fragmentation Of this compound using low-energy colli— sion-induced dissociation (CID) Of selected ions in a triple quadrupole mass spectrometer. Figure ”-1 displays all the collisional fragmentations Observed in this study (below the diagonal) and all the metastable transitions seen in the IKES study (above the diagonal) (38). The EI mass spectrum Of nonan-u-One shows A7 fragment ions, and the metastable study indicates 37 confirmed and 7 unconfirmed metastable transitions. The triple quadrupole system shows over 400 CID peaks, which include the “A fragmentations Observed in the metastable data as well as a host Of others. Not only does the triple quadrupole system with CID produce a significant increase in the number Of fragmentations Observed, it also provides direct mass analysis Of both the parent and A0 AMEUMSUMEJE FRAGNEJIT KIN NDQ ¢I40I30I20IIOIoosoeovososo4oso I40 ”0 I30 I30 5 N .20 I20 3 1 ~§ no no :2 5i £2 E§mo I00 F- _. Z: 90 so “J E E g 80 80 (.9 70 70 B 3 9 LL. 60 60 a.) C3 <1 “'50 so +- U LIJ 40 40 :5 30- 30 20-- 20 " ' '9: ‘I I40I30I20IIOI009080706050403020 CHD F¥VNHVT KIN NM? Figure “-1. Collision-induced fragmentations Of nonan-h- one from this study (below the diagonal), metastable fragmentations from previous IKES study (38) (above the diagonal), and EI mass spectrum (along the diagonal). ”1 daughter ions at unit mass resolution. A combination Of complementary methods was used in the IKES study Of metastable transitions (38). The kinetic energy loss upon metastable fragmentation results in broad peaks which make difficult unambiguous mass assignment Of the parent and fragment ions (.e_._g_._, 1142+ + 99+ or 1142+ + 98+) in IKES scans. (In these scans it is the ratio Of parent ion mass/charge squared tO daughter ion mass/charge that is actually measured, so that several pairs Of parent/daughter ions (g;gL, 1A2+ + 86+ and 71+ + 43+) may contribute tO a single peak). Comparison Of IKE spectra of isotopically labelled and unlabelled samples can help in the assign- ment Of specific transitions for each peak. In the high voltage scanning technique that was also employed, mass losses Of greater than 50% cannot be Observed. The major fragmentations Observed in the CID studies are depicted in Figure “-2. The stars indicate transitions that are also Observed in the metastable studies (38). Redundant pathways are not shown (e.g., 1&2+ + 85+ is not shown since the transitions 132+ + lllI+ and lllI+ + 85+ do appear). Note that the CID data do not indicate any ions Of m/z not seen in the EI mass spectrum; rather, they provide interconnections between these ions, specific fragmentation pathways for the production and further fragmentation Of each ion. The fragmentations normally attributed to alkanones .Ava moacsuw OHpmpmMch mSOH>mnq :H om>hcmno coon Oman m>mn meowufimcma» Oceanum .zp59m 9H0 was» ca cm>nomno OCOIzlcmcoc no wCOHpmpccEwwhm nOnmE .mla madman “3 are all Observed, including a-cleavage (1H2+ + 99+ and lh2+ + 71+), loss Of CO or 02H“ from the a—cleavage ions (99+ + 71+ and 71+ + h3+), McLafferty rearrangement in the long chain (llI2+ + 86+), and the resultant double Mc- Lafferty rearrangement (86+ + 58+). These transitions, also seen in the IKES study Of metastables, account for the genesis Of the five major fragment ions in the electron impact mass spectrum. These fragmentations also shed some light on the preferential a-cleavage Of the short or long A torr), chain. At high collision gas pressures (l x 10- the long chain is preferentially eliminated, as is the case in 70 eV EI spectra (32). At lower collision gas pressure (2 x 10"5 torr), however, the short chain is preferentially lost, a feature that is also Observed in 10 eV EI spectra (39) and metastable studies (33). It has been postulated (39) that further decomposition Of the ion resulting from a-cleavage Of the short chain is the cause Of its reduced intensity in the 70 eV EI spectrum. Further fragmentations are minimized at 10 eV, and therefore the short chain a-cleavage ion remains the more intense. The CID data are in agreement with this rationaliza- tion, since multiple collisions and therefore further fragmentations are more likely at higher collision gas pressure. This rationale requires faster rates Of de— composition and/or more pathways for the further decomposi- tion Of the ion which results from a—cleavage Of the short HA chain compared tO that from the long chain loss. That reasoning can be checked by comparing the degree Of frag- mentation in the CID spectra Of the two a-cleavage ions. The ion arising from d-cleavage Of the short chain (99+) shows 97% fragmentation (only 3% of the ion current is due tO 99+) at l x 10'“ torr compared to only 88% fragmentation for 71+, the ion arising from the a-cleavage Of the long chain. This higher probability for further fragmentation Of 99+ compared tO 71+ explains why 71+ predominates in cases where enough energy is available for further frag- mentation (70 eV EI or high pressure CID) and why 99+ is larger when further fragmentation is unlikely (low energy EI or low pressure CID). In addition to these fragmentations already Observed, a number Of new transitions not previously reported are Observed, such as the McLafferty rearrangement in the short chain and its further fragmentation (ll-I2+ + 11A+, lllI+ + 86+, llII+ + 70+, 1114+ + 58+), eleven fragmentations Of the methyl-loss ion (127+) which confirm the methyl loss in the short chain, the formation and fragmentation Of the hydrogen-loss ion (1H2+ + 1A1+ and 1A1+ + 99*, 1A1+ + 86+), the loss Of water from the a-cleavage ions (99+ + 81*, 71+ + 53*, 57+ + 39*) and the fragmentation of the resultant hydrocarbon ions. Over 350 fragmentations not previously reported are shown in Figure A-l. A number Of these are discussed in the sections which follow, together A5 with fragmentations which confirm or contrast with those reported by the IKES study. Discussion Of the formation and fragmentation Of specific ions appearsiJIthe following sections, arranged in order Of decreasing m/z. All the transitions discussed can be seen in Figure ”-2. Group A; 1A2+, llIl+ The CID spectrum Of 1A2+ is very similar in appearance tO the BI spectrum Of the molecule. The same fragments are Observed, with the exception Of the isotopic peaks, which are eliminated by the selection Of 142+. By control- ling the collision gas pressure and ion kinetic energy, the relative intensity Of the fragment peaks in the CID spectrum can be made to agree almost exactly with that in the normal EI spectrum. The 1H2+ + lill+ transition, not seen in the metastable study, is Observed in the CID data. Group B;4127+ The CID data confirm that the 1A2+ + 127+ transition is due tO CH3~loss from the short chain, as previously indicated by deuterium labelling in the metastable study. Major CID fragments Of 127+ are due tO further loss in the short chain to produce 11h+ and 99*, indicating that the long chain in 127+ is still intact. Another A6 fragmentation is the loss Of the elements Of propanol to form 67+, CSH7+ + ou c- 11 +--11 The fragmentation Of 113+ indicates that it is formed by loss Of C2H53 probably from the long chain Of the molecular ion, since both 99+ and 86+ are missing. Note also the loss Of CZHAO tO form C5H9+. The authors Of the metastable study assigned the weak 1142+ + lllI+ metastable to the fragmentation process in Equation (1), based on deuterium labelling. -02Hu .+ 1u2+ 1114+ All fragmentations Of the 11A+ ion, however, were attribut- ed tO the ion which arises from the McLafferty rearrange- ment in the short chain as shown in Equation (2). H ,+ +' O \_I -02Hu OH I ——->)\/\/\ (2) 1u2+ 1114+ A7 Based on the intense fragment ions Of 11A+ observed at 57+ and 58+ in this study, the structure Of the 11A+ ion prO- duced by electron impact is that shown in Equation (2). It is quite possible that the McLafferty rearrangement which is Observed in El also occurs by CID. The ion at 115+ is not an isotopic counterpart of 11A+ since its fragmentation bears little resemblance tO that Of 11A+. The 55+ fragment ion is approximately three times as intense as any other ion which could result from the following formation Of 115+ from a two-hydrogen rearrange- ment, loss Of C3H6 to form 73+,and loss Of H20 tO form 55+ [Equation (3)]. ) H (I + H+ @WA 02*; 3"“ °H .- 0 .- H 4§;‘\/’~\~/"\~ 4¢4\\//' (3 1A2+ 115+ 73+ I [::]+ 55+ Group D; 98+-100+ In the study Of metastable fragmentation, 99+ was Observed to be formed from 1A2+ by a-cleavage Of the short chain as well as from 11A+. The CID data show formation A8 of 99* from 11:1+ and 127+ as well. The fragmentation of 99+ shows the expected losses, plus the unexpected loss Of water to form 81+, as shown in Equation (A). + + OH O H ’2'. Haz; + -H 0 99+ 81+ The mechanism Of H20 loss in branched-chain ketones has been elucidated (37), but the very low intensity Of the ion due to this loss in straight-chain ketones has pre- cluded its study by deuterium labelling or from experiments involving metastables. Despite the low intensity of the water-loss ion, the CID data make it possible to Observe not only the loss Of water, but also the further fragmenta- tion Of the product ion (81+). Analogous fragmentations which indicate the loss Of water are seen for the other a-cleavage ion (71+), as well as for 73+, 57+, A5+, AA+, and 31*. The 98+ ion is seen as a fragment Of 1A2+ in both the CID and metastable studies, with an intensity Of approxi— mately 5% Of 99+. The CID spectrum shows a moderately F! A9 intense 99+ + 98+ fragmentation, which suggests that the 98+ may be formed by a two-step fragmentation from 1A2+ through 99*. Group E; 81+,g85+-87+ The 81+ ion does not appear as part Of any metastable transition with sufficient intensity tO be Observed in previous studies. In addition to its formation, as described in Group D above, its fragmentation is clearly Observed. The ion series 53+, Al+/39+, 29+, 1A+ from the even-electron ion suggests a 6-membered ring. The formation Of 85+ from 1A2+ and 11A+ was noted in metastable studies. We Observe that it may also arise from the fragmentation Of 100+, 99+ and 86+. The 85+ shows the ion series 57*, u3+/u1t 29+/27+, 15+/1u+; the absence Of a 71+ or 70+ peak is in agreement with the cyclo- pentanone structure. The ion at m/z 86+ is formed as expected by the Mc- Lafferty rearrangement in the long chain Of the molecular ion. Another route for its formation, however, is by loss Of CZHA from the other McLafferty rearrangement ion, 11A+. The 87+ ion is formed by the "McLafferty+l" rearrangement from the molecular ion. Its fragmentation is quite similar tO that Of 86+ plus a hydrogen, with the exception Of the very large 87+ + A5+ transition (loss Of C3H6). 50 Group F; 67+-73+ The 57+ ion has not been Observed in earlier metastable studies. Its formation was discussed in Group B above. Its fragmentation pattern follows that expected for the even-electron C5H7+ ion, and is analogous to the fragmenta- tion Of 81+ and 53+. The 69+ and 70+ ions also show frag- mentation patterns indicative Of cyclic alkane ions. The CID spectrum Of the even-electron 69+ shows very good agree- ment with that Of 69+ from cyclohexane; the Odd-electron 70+ ion shows very similar fragmentation to the 70+ ion Of n—octane. Two ions are isobaric at 71+, 05H11+ and CuH7O+. By careful deuterium labelling in the metastable studies, the 71+ ion from 1A2+, 1A1+, 113+, and 86+ has been shown to be CuH7O+; the 71+ peak from 99+ has been found tO correspond to both CSH + and CuH7o+. In this CID study, 11 the formation of 71+ from all Of these ions is Observed, as well as the production Of 71+ from 127+, 115+, llA+, 98+ and 87+. By consideration Of other fragmentations and the neutral losses involved, it is possible to assign the CuH7o+ structure for the 71+ fragment Of 87+ (by 87+ + 86+ + 71+) and 98+ (loss Of 27). For the other three parent ions (127+, 115+, llA+) it is not possible to confirm the structure Of the 71+ fragment based on the available data. Deuterium labelling would aid in these studies. 51 The fragmentation spectrum Of 71+ from the El spectrum Of nonan-A-One shows A3+ as the most intense fragment, with other fragments at 27*, 29+, 31+ (with intensities 0.A0, 0.25 and 0.015, respectively, relative to A3+), 15+, 39+, Al+, 53+, and 55*. For comparison, the CID spectra Of CuH7O+ from 2—pentanone and CSHll+ from n-octane were measured under similar experimental conditions. The CAH70+ ion shows the most intense fragment at A3+, with other fragments at 27+, 29+, 31+ (intensities o.uo, 0.08, and 0.075 relative to u3+), 15+, 39+, Al+, A5+ and 53*. The CSHll+ ion produces a CID spectrum with A3+ also the most intense fragment, and additional fragments at 27+ and 29+ (intensities 0.05 and 0.36 relative tO A3+), 15+, 39+, A1+, 55+, and 56+. The 71+ ion from nonan-A-one shows fragments corresponding to both the possible ions, and is therefore undoubtedly composed Of both CuH7O+ and CSHll+ Based on the relative intensities Of the 27+ and 29+ peaks, the contributions to the 71+ ion signal are esti- mated to be 30% CSHn+ and 70% CuH7O+o The 72+ ion yields the fragmentations expected for the addition of H to 71+. The major pathways for formation of 72+ are "y-cleavage + 1" from 11A+ and 87+, and McLafferty rearrangement from 100+. The formation of 73+ from 115+ and its fragmentation by loss Of H20 to form 55+ was dis- cussed in Group C. Other major fragmentation pathways for 73+ are loss Of CZHA and C2H6 tO form A5+ and A3+ respectively. ‘7 52 Group G;i50+-50+ The ions at 50+, 51+, 52+, 53+, and 5u+ all show frag- mentation spectra that are quite similar tO those for the same m/z ions from hydrocarbons such as cyclohexane and n- octane. The 55+ ion, however, shows a peak in the CID spectrum at 27+ that is 1/3 greater in relative intensity than the 27+ ion seen from 55+ (CuH7+) in reference alkane CID spectra. This may indicate that a small portion Of the 55+ peak is due to the c3H30+ ion. Although the CID spectrum Of C3H3O+ has not been measured for reference, the major fragmentation would likely be 55+ » 27+, with loss Of CO. In the EI mass spectrum Of cyclopentanone, the C3H3o+ ion comprises 9A% Of the 55+ base peak, and its formation has been carefully studied (A0). It is improbable or impossible that 55+ formed from 1A2+, 115+, 113+, 86+, 73+, 72+, 71+ +) but it is possible that fragmentation Of 127+, 99*, 98+, + + + , (CSH11 , 70 or 69 would be C3H3O , and 71+ (CuH7O+) could produce C H3O+ as well as CuH + 3 7 ° The peak at m/z 56+ can also correspond to two dif- + ferent ions, CAHBT and C3HuO-. The fragmentation matches that Of CAH8t with the exception Of the 28+ ion which is more intense than in the reference spectrum. The extra intensity at 28+ is undoubtedly due to the C2Hu? frag- 1- + + ment Of C3Hu0 . Both Cqu and C3H50 are possible structures for the 57+ ion. Only the 57+ fragment from 85+ has been assigned (as C H50+) based on metastable 3 53 results. The CID data shows that the 86+ ion also frag- ments tO form C3H50+, but it is impossible to positively assign without further study either the Cqu+ or C3H3O+ structure to the other 11 fragmentations Observed that lead to 57+. The CID spectrum of 57+ from nonan-A-one shows a peak at 31+ that is not Observed in the otherwise similar spectrum Of CuH9+ from n-octane. The 31+ ion probably corresponds to the loss of C2H2 from C3H50+- The CID spectrum Of the 58+ ion shows twO major frag- ments, A3+, and 15+. The interpretation Of the CID spectrum indicates the acetone ion (CH3)2CO?; comparison with reference spectra shows it tO be identical with those 4. for (CH3)2CO- from acetone and 2—pentanone. Group H;38+-A5+ The ions at 38+, 39+, A0+, and Al+ all show very good agreement with alkane reference CID spectra. The ions at A2+, A3+, and AA+, however, are mixtures Of alkane and ketone ions. The A2+ ion shows an additional intense fragment at 1A+ undoubtedly due tO the formation Of CH2? from CH2COt. The reference CID spectra Of the two possible A3+ ions, CH3CO+ and C3H7+, include all the same ions; the major difference is the ratio Of the intensities Of the 27+ and + 15+ peaks: approximately 5:1 for C3H7+, and 1:20 for CH3CO The A3+ peak from nonan-A-one shows fragments at 27+ and 5A 15+ in the ratio l:A. Based on these data, the A3+ peak corresponds to 75% CH3CO+, the remainder being C3H7+. The A3+ peak is quite common in the CID spectra; all but 3 ions above m/z 57+ show a A3+ peak. The CID spectrum Of AA+ shows the fragmentations that would be expected for CH3CO+ With one C13 or an extra H. The A5+ ion includes H O+ (loss Of acetylene) as one Of 3 the most intense fragment ions in its CID spectrum. Group I;_26+-31+ +, 28+, 29+ and 30+ all give CID The ions at 26+, 27 spectra which agree well with the reference CID spectra for hydrocarbon ions. The 31+ ion, however, must contain 0, and indeed fragment ions are Observed at 18+ and (31-18)+. Conclusions The triple quadrupole mass spectrometer provides a wealth Of data that can be interpreted tO provide a new dimension of information for structure elucidation. The formation and fragmentation pathways Of specific ions can be obtained in order to determine ion structure. Start- ing with every fragment of a compound, every fragmentation pathway for the entire molecule can be determined. 55 The application Of the triple quadrupole system to the collision-induced dissociation Of the ions Of nonan-A- one has identified over A00 fragmentation paths. This enormous increase in the amount Of information available for the structure elucidation Of the compound compared to that Obtained by earlier IKES metastable studies is a result Of several features Of the triple quadrupole system: (1) The use Of CID increases the number Of fragmentations that occur, and their intensity. (2) The high sensitivity Of the system makes it possible to study transitions Of very low intensity. (3) The use Of two stages Of direct mass analysis eliminates any ambiguity in the identifica- tion Of parent and daughter ion m/z. (A) There is no restriction on the neutral loss. (5) Unit mass resolution is achieved and is unaffected by the kinetic energy loss upon fragmentation. Interpretation Of the CID spectra has made it pos- sible tO assign specific structures tO a number Of the fragment ions of nonan-A-one. Confirmation of these structures can be achieved by comparison Of the spectra with reference CID spectra of ions Of known structure. In the case Of fragment peaks which correspond tO two isobaric ions, it is possible to estimate the contribution Of each ion by comparison with the reference spectra, without resorting to the high resolution necessary tO separate the isobaric ions. The ability to measure the 56 probability Of further fragmentation for the fragment ions which arise from u-cleavage has made it possible tO explain the apparent preference for a-cleavage in the long chain in 70 eV EI spectra and for a-cleavage in the short chain at 10 eV. A potentially powerful system for structure elucida- tion Of organic compounds should be realized by the com- bination Of complete CID fragmentation data with computer programs which use heuristic or pattern recognition tech- niques for structure identification. The completeness Of the CID fragmentation data and their ability tO show rela- tionships among the ions in the mass spectrum should make computer interpretation extremely powerful. CHAPTER 5 INSTRUMENTATION An added dimension Of mass spectral information is provided by selected ion fragmentation whereby several ionic species are generated from a sample, ions Of a particular mass are selected for fragmentation, and the resulting fragment ions are mass analyzed. We recently introduced the concept Of performing selected ion frag— mentation with a triple quadrupole mass spectrometer (1). In this system, the selected ion fragmentation process can provide enhanced selectivity and discrimination over normal mass spectrometry without loss Of sensitivity. There are several applications Of mass spectrometry which benefit greatly from the added information con- tained in the fragmentation spectrum Of each source ion. These include the elucidation Of organic structures and the analysis Of mixtures. In structure elucidation applica- tions, any fragment ion in a compound's normal mass spec- trum can be selected with the first mass analyzer. This "parent" ion is further fragmented, and the mass spectrum Of the resulting daughter ions is determined by scanning the second mass analyzer. A complete fragmentation map may be Obtained by recording the mass spectrum Of each fragment ion Of a compound. Figure 5—1 is an example Of such 57 58 “/2937: // \ El SPECTRUM ;/ Z 9 + IIIIII 02 *- a “VQA wwwv‘j I553 "“::: .W {333 .......... \fi Its? wwvurh A“ I'I"\"-E 1:15;: ..... 3' stEE“ '2- . ”IWEEE. IONS 0F 69 + Three—dimensional fragmentation map for cyclohexane. Figure 5-1. 59 a map, for cyclohexane. Note that the normal electron im- pact mass spectrum, displayed along the diagonal (fragment ion m/zsparent ion m/z), is the only information available without the added dimension Of selected ion fragmentation. For analysis Of mixtures, the molecular ions for each component can be produced by soft ionization and then sepa— _ rated by the first mass analyzer. Thus, one component at a time, the molecular ion species can be selected, fragmented and the individual mass spectrum Obtained by scanning the second analyzer. The added dimension in this case is the selectivity and discrimination achieved through mass separa- tion Of the molecular ions of the several components. The elimination Of chemical noise that results can Often improve the detection limit Of the mass spectral technique (23). I The triple quadrupole mass spectrometer is a particu- larly simple and efficient approach tO selected ion frag- mentation. The ion fragmentation process is performed by collision-induced dissociation (CID) in an RF-Only quadrupole i which provides ion focusing and is highly efficient (2). The instrument consists Of, in series, a dual chemical ionization/ electron impact (CI/EI) ionization source, a quadrupole mass filter, an RF-Only quadrupole that can be pressurized for CID, a second quadrupole mass filter and an electron multi- plier, as shown in Figure 5-2. There are several possible modes Of operation for the instrument: (1) In order tO Obtain a normal mass spectrum (only 60 .COHuocsm npfi ncm ucmcoaeoo comm weasonm LmumEOpuOcam mmmE maoasaomsv mamas» one no Empmmfip annuaoocoo ZOFUMI—mm 20:0; so}. 20. PUDOOE ZO_._.<._.Zw—2.U aOpv wcfizmsp carom .mIm mhzwam wH16WE33 .1921 ‘98 HE. D I: fie L: Li L L ”v0.3 «.3? .3 Kg; R53 is KKSQiimwfiI Hfifil IIM%WI MM 6:9 :2 I. <.om~II_IH\ SE J 583 I O? _ i is. #5:. 65 first except that the small (0.2 cm diameter) entrance aperture has been removed. The axial ion energy in the three quadrupoles may be varied independently over the range ~100 tO +100 V, with -10 V being typical for positive ions. The positive ions are detected with a Galileo A770 high-current Channeltron with a measured gain Of 3 x 106 at -3 kV. The multiplier is mounted Off axis to eliminate noise from stray neutrals and photons, with a deflection electrode at —65 V which helps tO focus the ions into the Channeltron. The multiplier is interchangeable with a A870 Channeltron which has a gain that is sufficient (3 x 108) tO allow ion counting for positive and negative ions. The current from the multiplier is detected with a Keithley 18000-20 picoammeter with programmable ranges Of 10 V/lO"10 A to 10 W10"3 A. The components Of the mass spectrometer are enclosed in three differentially pumped stainless steel chambers, each pumped by an Oil diffusion pump with water-cooled baffle (see Figure 5-3). Normal Operating pressures in the three chambers, as indicated by ion gauges, are source chamber: EI 2 x 10"7 torr, CI 6 x 10"l'I torr; center chamber: 2 x 10 torr (no collision gas); detector chamber: 2 x 10'”6 torr. Source pressures up to 10 torr are possible but 1 torr is typical in CI mode. Source pressure is measured by a Gran- ville—Phillips Thermogauge on the CI gas line. Collision -6 if] 66 6 gas pressures from 2 x 10' torr to l x 10'2 torr are pos- sible. The ion gauge on the center chamber is calibrated against a thermogauge mounted on the center quadrupole in order to provide routine measurement Of the collision gas pressure. The entire vacuum system is interlocked to provide fail- safe operation and permit computerized cycling Of the vacuum system. Foreline pressures, coolant water flow, electric current through the pumps, pneumatic pressure, and valve position are all monitored and used to control the proper sequencing Of all pumps, electropneumatic valves, and electronics. Data acquisition from the system is currently either manual (strip chart recorder or oscilloscope), or auto— mated with an Intel SDK-85 microprocessor, CRT terminal, and floppy disc. A multiple microprocessor system employ- ing the Intel 8085 is currently under development for data acquisition and intelligent control of the instrument. The entire system has been designed for computer control, including the mass selection in both mass filters, the selection Of mass filter or total—ion mode, the RF vol- tage on all three quadrupoles, the control Of collision and CI gases, and the vacuum interlock. The capabilities Of the triple quadrupole system for selected ion frag- mentation will be significantly enhanced under complete computer control. 67 Performance Of Instrument The use Of quadrupoles as mass filters and as the CID chamber has provided the anticipated excellent selectivity and sensitivity. Selectivity is achieved by tandem mass separation up to mass 1000 with resolution as high as one part in 1500. The high sensitivity (detection limit of 10'15 mole) results from the very efficient low-energy CID process and the high transmission Of each component along the ion path. Resolution The two mass filters have been Operating with a mass range Of 1—500 amu (2.8 MHz RF), but the maximum mass is being increased to 1000 amu (1.9 MHz RF). The ultimate resolution Of the quadrupoles is approximately 1 part in 1500, as determined by measuring peak width at half height. As an example Of the resolving power, the Nib/CH1-2 doublet at 1A amu can be resolved (50% valley), which requires a resolution Of 1 part in 1100. The mass filters can be operated in two modes: constant resolution and resolu- tion proportional to mass. At constant resolution (g;g;, 1 part in 1000 over entire mass range) the relative abun- dance Of the ions in a mass spectrum closely resembles that Observed with double-focusing mass spectrometers. With resolution proportional to mass (e.g., 1 part in 100 68 at mass 50, 1 part in 1000 at mass 500), the intensity Of the low mass peaks is enhanced. The quadrupoles are typically operated with unit mass resolution (1% valley) over the entire mass range. The peak broadening due to kinetic energy loss on fragmentation that is Observed with MIKES instruments does not occur with the quadrupole mass filter. CID Efficiency The high sensitivity Of the triple quadrupole system is a result Of the high efficiency Of each component. The ion source and lenses produce approximately 1 ion for every 2 x 105 molecules, as determined by measuring the ion current entering the first quadrupole at a known sample flux. The mass filters have a transmission efficiency Of approxi- mately 60% in RF-Only mode, and 10% in mass filter mode with resolution Of 1 part in 200. The center quadrupole has virtually 100% transmission. Even more important is the efficiency Of the CID process which occurs in the center quadrupole. In an earlier demonstration Of the high efficiency Of the low-energy CID process in a triple quad- rupole system (2), three expressions were developed to describe the efficiency. The collection efficiency is the ratio Of ions exiting the quadrupole tO those entering. With no collision gas present, there is 100% collection. At 2 x 10-“ torr collision gas pressure, the collection 69 efficiency ranges from 50% for light ions like CH“? up to 75% for heavier ions which are less prone to scatter. The strong focusing Of the quadrupole field minimizes scattering losses. The fragmentation efficiency is the fraction Of the ions exiting the center quad that are frag- A ment ions. At 2 x 10“ torr, fragmentation efficiencies range from 15% to 65% for various compounds (2). As the collision gas pressure is increased, the fragmentation ef- ficiency for all compounds approaches 100% due tO multiple collisions, but the collection efficiency decreases due to scattering. The overall CID efficiency, which is the product Of the collection and fragmentation efficiencies, exhibits a maximum at some intermediate pressure. The collection efficiency as a function Of collision gas pres- sure for the dissociation Of CH“? from methane is shown in Figure 5-A. The fragmentation efficiencies for the production Of the CH3+ and CH2? ions are also shown. Several other factors besides collision gas pressure can affect the efficiency Of the CID process (2). The larger the molecular diameter Of the collision gas, the more efficient the CID. Ion axial energy and ion internal energy also affect the CID process. A more detailed study Of these effects is in progress. 70 Lot 00 (‘3 F—\ Collection Efficiency E.) CH4+-’CH3+ O Fragmentation Efficiencies G erg—ma; ll. IO"I F- C) >- . L) 2! 9.1 2 u. n. “J I0'2 I J o -5 4 3 IO’ IO IO’ (torr) IO' COLLISION GAS PRESSURE (N2) Figure 5-A. Effect Of collision gas (N3) pressure on CID efficiency for methane CHu- at ion axial energy Of 10 eV. 71 Sensitivity The overall sensitivity Of the instrument can be esti- mated from the product Of the efficiencies Of the individual processes. The source efficiency and the transmission through the three quads (without collision) are 2 x 10'5 2 and 10' , respectively. The fragmentation efficiency is + a function Of the ion and fragment selected, but for CH“- .1. + H C3 overall efficiency is 2 x 10' from Figure 5-A it is seen to be about 0.1. The 8, i;g;, two CH3+ ions reach the detector per 108 CH” molecules passing through the source. The detection system can measure the current due tO one ion per second which is an average current Of 5 10 A x 10'13 A. For methane, then, a current Of A x 10' will be produced from a sample flux Of 1 pg per second. The ultimate detection limit depends on the system sensitivity, the chemical and electrical noise levels, and the lowest measurable signal level. The electrical noise '13 A. The two stages in the system is typically 3 x 10 Of mass separation Often make it possible tO reduce chemi- cal noise (ions detected from other than the desired reaction) to well below the 10'13 A level. In such a case, at the extreme sensitivity limit, the peak height is quantized depending upon the integer number Of ions reaching the detector during the scanning time Of the peak. The detection Of A ions would give an S/N Of 2 and, for methane, would require, 6 fg of sample. The detection 72 limit for methane has been determined experimentally to be 16 femtograms by measuring a sample Of 10 ppm CH“ in N2 at a sample flux Of 20 femtograms per sec Of CH“. The large excess Of N2 is required to increase the ion source pressure so that it can be accurately measured and the sample flux calculated. The selectivity Of monitoring the CH“? + CH3+ reaction has effected a two order Of magnitude im- provement in detection limit compared to that Obtained by simply monitoring the CHuf, CH3+, or CH2? ions from elec- tron impact. This improvement is due to the elimination Of the chemical noise introduced by background 1”Nut, 15N? and 160? peaks. The selected ion fragmentation capability Of the instrument produces only a small loss in transmission and sensitivity, and produces a significant gain in selectivity. Indeed, in the normal case where the system is limited by chemical rather than electrical noise (23), a substantial improvement in detection limit can be achieved as was Observed for methane. The detection limit Of approxi- mately 10'15 mole is made possible by the high efficiency Of every component Of the system. The detection limit for a higher mass organic compound, nitrobenzene, has also been determined experimentally. + was monitored at unit mass .4. The transition N02C6H5- + C6H5 resolution in both mass filters, with the parent ion produced by electron impact on a sample Of 10 ppm nitro- benzene in N2. A detection limit of 120 femtograms (S/N=2) 73 was Obtained with a sample flux Of 200 femtograms/second. In contrast, for the MIKES technique, Cooks has estimated a detection limit Of 10 pg for the 1A0+ + 123+ transition Of protonated nitrophenol ions produced by CI (A5). Al- though consumption Of only femtograms is required for re— cording the single reaction, actual sample size is signifi- cantly larger than this and is dependent on the sample inlet used. Total consumption Of a sample introduced by direct probe should make it possible to detect quantities Of individual components at close to these detection limits. Structure Elucidation As an example Of a structure elucidation application, consider unknown 5.13 in McLafferty's classic text (A6) on mass spectral interpretation. The EI spectrum shows molecular weight 120 with major peaks at 105 and 77. Con- sidering the 77 peak (phenyl moiety) and 105 (loss of methyl), there are two possible compounds, C6H5COCH3 and C6H5CH(CH3)2. Relying on the 121/120 ratio and the unusually small hydrogen-loss ions from 120 and 105, Mc- Lafferty concludes that the unknown is acetophenone, C6H5COCH3. The structure of this unknown can be Obtained in a direct manner using selected ion fragmentation on the triple quadrupole mass spectrometer. The large A3+ peak in the El spectrum corresponds tO the remainder Of the molecule [CH3CO+ or (CH3)20H+] after loss Of 77 (0635). 7A The CID spectrum Of A3+ is shown in Figure 5-5. Interpre— tation Of the data indicates the fragmentations shown for CH3CO+. Comparison with reference CID spectra Of known ions, as shown in Figure 5-5, shows a nearly perfect match with the CID spectrum Of CH3CO+ from acetone. Thus, Mc— Lafferty's conclusion is confirmed. The process Of struc- ture elucidation Of this unknown has been made simpler and more reliable by the added information available by selected ion fragmentation. Conclusions The CID process performed in the strong focusing field Of an RF-Only quadrupole is efficient and effective in producing characteristic spectra Of selected ions. The independence Of the mass filtering Operation on kinetic energy is ideal for collision product analysis. Inde- pendent control Of the two mass filters provides easy implementation Of the various analytical modes. The system can be used tO detect species present in a mixture including isomers without prior separation at the 10'15 mole level. The capability tO select a desired initial mass and specific collision product mass reduces chemical noise dramatically. For structure elucidation applications, the spectra Of selected functional moieties in the molecule may be Obtained. A number Of other 75 CID SPECTRUM OF 43+ FRAGMENTATION OF 43" Wm LT W—Jfl‘} 0+ 40 Liz CHag REFERENCEo CID SPECTRA CH3CO’ l .5 from Acetone 1 27,, f /—o a, h {I ‘Confrrnoncn ' '43 043/ chof (CH3IZCH* I I5 a:,c* CHO" from C Figure 5-5. Structure elucidation Of A3+ functional moiety by interpretation Of CID spectrum and comparison with reference CID spectra. 76 applications Of the selected ion fragmentation technique are promising, including the analysis Of isotopically labelled samples, in which the first mass filter can eliminate interferences from molecules that are not com- pletely labelled. CHAPTER 6 FUTURE WORK The triple quadrupole mass spectrometer has been demonstrated to be a simple and efficient instrument for mixture analysis and structure elucidation. A number Of projects remain to be undertaken in the further refinement Of the instrument and the technique, the extension Of chemical applications, and the quest to better understand the low-energy CID process. A number of instrumental improvements and additions are advisable. The mass range Of the quadrupoles needs to be extended to 1000 amu. The addition Of ion counting should increase the precision and convenience Of measuring low—level ion signals. The addition Of negative ion detec- tion will Open up important new areas Of study. Improve- ment in the efficiency Of the ion source is an important step in improving the instrumental sensitivity. The direct insertion probe for the introduction Of solid samples should be completed and tested. The development Of the multiple microprocessor system and the interfacing Of the triple quadrupole instrument to it is undoubtedly the most im- portant project yet tO be accomplished. The analytical capabilities Of the system will be fully realized only 77 78 under complete computer control. The low-energy CID process needs to be further characterized in order to better understand and apply it to analytical applications. In particular, the collis- ion gas pressure in the center quadrupole must be more ac- curately measured. Studies Of reactions Of known cross section and simulation studies of the ion collision process should help in this characterization process. Finally, the applications Of the triple quadrupole system should be extended into new areas. More complex mixtures and real-life samples must be analyzed. Com- puterized mass spectral interpretation and pattern recogni- tion need to be implemented for real-time structure elucidation Of unknowns. The selected ion fragmentation technique must also be applied to new types Of chemical problems such as the identification Of isotope substitu- tion. The research described in this thesis has demonstrated dramatically the potential Of the triple quadrupole mass spectrometer for chemical analysis. The future research proposed here should help realize this potential and further expand the horizons Of triple quadrupole mass spectrometry. REFERENCES 9. 10. 11. 12. 13. 1A. 15. 16. §§, 2621 (1966). REFERENCES R. . Yost, C. G. Enke, J. Amer. Chem. Soc., 100, 227a- 2275 (1978) R. A. Yost, C. G. Enke, D. C. McGilvery, D. Smith, J. D. Morrison, Int. J. Mass Spectrom. Ion Phys., 0, 127-136 (1979)- J. J. Thomson, "Rays Of Positive Electricity and Their Application to Chemical Analysis", Longmans, Green, London, 1913. W. Paul, H. P. Reinhard, U. von Zahn, Z. Phys., 152, 1A3 (1958). J. C. Holmes, F. A. Morrell, Appl. Spgc., ll, 86 (1957). P. J. ArpinO, G. Guiochon, Anal. Chem., 5;, 682A- 701A (1979)- J. A. Hipple, E. U. Condon, Phys. Rev., 68, 5A (19A5). J. H. Beynon, R. M. Caprioli, T. Ast, Org. Mass Spectrom., 5, 229 (1971). E. Lindholm, Z. Naturforsch., 2;, 535 (195A). J. H. Futrell, T. O. Tiernan in "Ion-Molecule Reac- tions", J. L. Franklin, Ed., Plenum, New York, 1972, Chapter 11. T. Y. Yu, M. H. Chang, V. Kempter, F. W. Lampe, J. Phys. Chem., 6, 3321-3330 (1972). M. L. Vestal 559-561 (1975). D. C. McGilvery, J. D. Morrison, Int. J. Mass Spectrom. Ion Phys., 8, 81-92 (1978). J. H. Futrell, Chem. Phys. Lett., g8, J. L. Franklin, Ed., "Ion Molecule Reactions", Plenum, New York, 1972. M. S. B. Munson, F. H. Field, J. Amer. Chem. Soc., K. R. Jennings, Int. J. Mass Spectrom. IonP hys 1,227 (1968). 79 17. 18. 19. 20. 21. 22. 23. 2A. 25. 26. 27. 28. 29. 30. 31. 32. 33. 80 J. Durrup, in "Recent Developments in Mass Spectros- copy", K. Ogata and T. Hayakawa, Eds., University Park Press, Baltimore, 1970, p. 921. R. W. Kondrat, G. A. McClusky, R. G. Cooks, Anal. Chem., 12g, 1222-1223 (1978). C. J. Porter, R. P. Morgan, J. H. Beynon, Int. J. Mass Spectrom. Ion Phys., gg, 321-333 (1978). F. W. McLafferty, P. F. Bente III, R. Kornfeld, S.-C. Tsai, I. Howe, J. Amer. Chem. Soc.,‘2§, 2120-2129 (1973). T. Wachs, F. W. McLafferty, Int. J. Mass Spectrom. Ion PhX§-. 2;. 2A3-2A7 (19777. R. G. Cooks, personal communication. R. W. Kondrat, R. G. Cooks, Anal. Chem., 2g, 81A-92A (1978). F. W. McLafferty, F. M. Bockhoff, Anal. Chem., 53, 69-76 (1978). M. H. Bozorgzodeh, R. P. Morgan, J. H. Beynon, Analyst, 103, 613-622 (1978). F. W. McLafferty, "Interpretation Of Mass Spectra", 2nd Ed., W. A. Benjamin: Reading, MA, 1973, D. 190. R. A. Yost, C. G. Enke, Anal. Chem., in press. R. G. Cooks, J. H. Beynon, MTP Int. Sci., Ser. II, 2, 159 (1975). F. W. McLafferty, P. F. Bente, R. Kornfeld S.-C. Tsai, I. Howe, J. Amer. Chem. Soc., 95, 3886 (1973). U. P. Schlunegger, Angew. Chem. Int. Edn., gi, 679-688 (1975). K. Levsen, H. Schwarz, Angew. Chem. Int. Edn., 1;, 509-519 (1976). H. Budzikiewisz, C. Djerassi, P. H. Williams, "Mass Spectrometry Of Organic Compounds", Holden-Day, Inc., San Francisco, 1969, Chapter 3. R. G. Cooks, A. N. H. YeO, D. H. Williams, Org. Mass Spectrom., ;, 985-995 (1969). 81 3A. J. H. Beynon, R. M. Caprioli, R. G. Cooks, Org. Mass Spectrom., 3, 1-11 (197A). 35. D. J. McAdOO, P. F. Bente III, M. L. Gross, F. W. McLafferty, Org. Mass Spectrom., 9, 525-535 (197A). 36., G. Eadon, J. Dickman, C. Djerassi, J. Amer. Chem. Soc., 2;, 6205-6212 (1970). 37. G. Eadon, C. Djerassi, J. Amer. Chem. Soc., 2;, 308A- 3089 (1970). 38. G. Eadon, C. Djerassi, J. H. Beynon, R. M. Caprioli, Org. Mass Spectrom., 5, 917-933 (1971). 39. W. Carpenter A. M. Duffield, C. Djerassi, J. Amer. Chem. Soc., , 6167-6170 (1967). A0. P. Natalis, Bull. Soc. Chim. Belges., pl, 599 (1958)- Al. M. J. Lacey, C. G. MacDonald, Anal. Chem., 5;, 691- 695 (1979). A2. R. W. Kondrat, G. A. McClusky, R. G. Cooks, Anal. Chem., 2Q, 2017-2021 (1978). A3. R. J. Stradling, K. R. Jennings, S. Evans, Org. Mass Spectrom., 13, A29-A30 (1978). AA. J. R. Olson, g£_§l., Rev. Sci. Instrum., 52” 6A3- 6A9 (1978). A5. T. L. Kruger, J. F. Litton, R. W. Kondrat, R. G. Cooks, Anal. Chem., 3;, 2113-2119 (1976). A6. F. W. McLafferty, "Interpretation Of Mass Spectra", 2nd Ed., W. A. Benjamin, Reading, MA, 1973, pp. 95,96. APPENDICES APPENDIX A 2274 Selected loo Fragmentation with a Tandem Quadrupole Mass Spectrometer Sir: An added dimension of mass spectral information is pro- vided by a tandem mass spectrometer when it is used to create ion species from a sample. select one individual ion specres. fragment it. and obtain the mass spectrum of the fragments. Metastable ion peaks and collision-induced dissociation (CID) have been used to relate daughter ions and their precursors.‘ Specialized "NHKES"2 instruments have been developed to allow systematic acquisition of data on metastable and CID fragmentation spectra. The potential of such instruments for mixture analysis and structure elucidation is currently being explored. These processes are illustrated in Chan I. The second Chartl : : Mass : cm I \tass I 55m": “THU“! 5W": Frownt'n: Sea'n : Detection i : ,-H‘ : : -r : " °‘ ”30‘" «W ~4- A‘.a‘.C'*°Ies’ .. “iii?” 5 3 Ln’ 5 i -c’ i : “'5. i : to“. i m.» -' K'm’m' *4 I-M‘ - n’.o’.n’..'....o' .. "xiii“ I I . l , I . I I '-N I : -a mass separation in M IKES instruments is actually an ion ki- netic energy separation interpreted to provide the fragmen- tation mass spectrum. Impressive sensitivity has been achieved with a MIKES instrument’ despite the substantial ion losses and the ion energy spread produced by the CID process.” Data presented here demonstrate that a selected ion frag- mentation mass spectrometer based on tandem quadrupole mass filters is completely practical and that the C10 process in the quadrupole instrument is effective and extremely cf fi- cient. The system consists of. in series. an electron impact (El) ionization source. a quadrupole mass filter. an “RF-only" quadrupole C lD region. a second quadrupole mass filter. and an electron multiplier. The use of quadrupoles for mass sepa- ration provides higher transmission efficiency than magnetic sectors operated as the same resolution and mass range.7 and unit mass resolution in the CID spectra is easily obtained. Tandem quadrupole mass spectrometers have been devel- oped for the study of ion-molecule reactions."'° A center RF-only quadrupole has been added for phatodissociation studics' “'2 and the investigation of long-lived metastable ions. ' 3 Prior to this study. however. all of the reported selected ion fragmentation work has been performed on reversed-sector MIKES instruments.'~3‘°-" Experiments demonstrating the practicability of selected ion fragmentation in a quadrupole system have been performed on a tandem quadrupole mass spectrometer'2 in the laboratory of J. D. Morrison at LaTrobe University. Bundoora. Victoria. Australia. The central quadrupole of three was used as the C [D chamber by the admission of the collision gas and operation in the RF-only mode. The El source sensitivity was 3 x lo” A/Torr of cyclohexane (1 ion / I ()5 molecules). Transmission through the ion optics into the first quadrupole was 3096. In RF-only mode. quadrupole transmission was 30%. la mass filter mode. transmission dropped to 2.5%. Neglecting CID. this gives an overall sensitivity of 1.5 X 10‘9 A/Torr of cy- clohexane (5 ions detected /' lO'° sample molecules). The se- lected ions emerged from the first quadrupole and entered the CID quadrupole with a translational energy of ID V. The RF oscillations of the ions in the CID quadrupole increase the ion energy by a few volts. However. the kinetic energy of ions in these experiments is very small compared with that of MIKES Journal of the American Chemical Society / 100:7 / March 29. 1978 55 69 8| (mo) M /L L l l l l l J! I I'D/O GO 50 GO YO OO 90 Fig.0 l. CID spectrum of the parent ion (ni/e 98) of cyclohexanone prment as 596 of a mixture. instruments where ions enter the CID chamber with an energy of 3- l0 kV. The CID spectra obtained in the quadrupole in. strument resemble the l4- l6oeV El spectra of the pure com- pounds. Clearly the low translational energy combined with the relatively long (~5 X IO‘5 5) residence times of the ions in the CID region is sufficient to provide a characteristic and relatively rich fragmentation spectrum. The efficiency of the CID process is determined by two factors. the fragmentation cf ficrency and the collection effi- ciency. We can let Po and P symbolize the selected ion beam current at the entrance and exit of the CID region. respectively. and SF. the total current of all fragment ions at the exit of the rcgion. The fragmentation efficiency is E p I SEX/I P + SE). the collection efficiency is Ec - (P + SKI/Po. and the overall CID efficiency is EC“) I SF;/Pq I EFEC. The overall CID efficiency of the quadrupole system ranges from l5% for benzene to 65% for n-hexane. The CID efficiency of a M IKES instrument has recently been reported as < I096.‘ The collec- tion efficiency in the tandem quadrupole system is nearly IOO'Ib: there is virtually no scattering loss in the CID process. even at the 2 X l0“ Torr C ID pressure used in these studies. (Larger pressures could not be obtained with this instrument.) In this system. CID occurs in a strong-focusing quadrupole field: the field-free drift region used for CID in MIKES in- struments produces scattering losses of 90% (collection effi- ciency of IO%) at similar C 10 pressures." The scattering losses in the MIKES CID chamber increase as the mass of the collision gas increases. This has led to a preference for hydrogen or helium as the collision gas.°-" Because the collection efficiency is nearly 100% in the qua- drupole CID region. heavier collision gases can be used to in- crease the fragmentation efficiency. Argon shows CID cf- ficiencies three to four times higher than hydrogen in this system. To demonstrate the mixture analysis capabilities of the quadrupole system. a mixture of cyclohexane and three minor components (benzene. n-hexane. and cyclohexanone. each present as 5% of the mixture). was analyzed. The mixture components were selected to minimize interference between fragment ions and molecular ions in the El spectrum of the mixture. This. of course. would not be necessary if a low energy ion source such as chemical ionization (C l) were used. The CID spectrum of the parent ion of each of the four components was obtained: that for cyclohexanone is shown in Figure I. The C ID spectra of all of the mixture components are reproducible and show good agreement with the CID spectra of the pure compounds. On the basis of these results. the probability of being able to achieve a highly effective. yet fundamentally simple. quadrupole-based selected ion fragmentation mass spectrometer seems virtually certain. A tandem quadrupole system for selected ion fragmentation is currently under construction in this laboratory. It will have a dual C l/El source and more efficient quadrupoles with a larger mass range and be able to tolerate higher CID pressure than the instrument on which these experiments were per- formed. The ability to vary the translational energy of ions entering the CID region will also be incorporated. This in- strument will be used to further characterize the low-energy AReprinted from the Journal of the American Chemical Society. too. 2274 (1978).] Copyright 1978 by the American Chemical Society and reprinted by permission of the copyright owner 82 83 Book Reviews C ID process and to explore selected ion fragmentation appli- cations in mixture analysis and structural elucidation. Acknowledgments. We gratefully acknowledge fruitful discussions with .I. D. Morrison and the use of facilities in his laboratory in this work. Special thanks are extended to Don McGilvery and Dianne Smith for their assistance. This work was supported in part by the Office of Naval Research. One of us (R.A.Y.) gratefully acknowledges a National Science Foundation Graduate Fellowship. an American Chemical Society DiVision of Analytical Chemistry Fellowship sponsored by the L‘pjohn Co.. and an L. L. Quill Memorial Fellowship from MSU. References and Notes (1) R. w Koiuetandn G. Coolie. Anal. Chem. 30. autism. (2) MKES (nines-analyzed ion kinetic energy may). DAD! (dtect miysis ot unmet ions). and CAMS (eotIisionei activation mean spec- nometryiaiiretettotheteemiquenmxhencetieruyMe-nhe mimnimotarmmmmm 2275 WitmeMMmmaCOm. (3iT.an-.J.F,Lmn.w.KmmR.GCoan.-t.m.4|. 2113(1976i. (a) H H Tuitnot. Int. J MW looms 23147I1977i (5) UP CM"- Int Ed.En¢..1d.679(1975). (GIRLeveenaiidHSchwarLAngerhenLME Enpt...16500 (1976). ITIGLaweonandJ F. J Todd.Chem...8r O. 373(t972) (I)J.FFwellandTOT'iert-iin Ion-Molecue LFruIdh. Ed. PIenumPreea NewYont NY t972..CMptettt (Olcwnlaenfluamonmndw SKMIHJM...” 5105109721 (tOITWYYuMHMVKmmFWWJanem.n 3321(1972) (11) H LVeetatandJ. N. Forte. WWMM 20. 559(1974) (tzionflorrieonaridDCMcleererltflJmJAheW press (13) U manuals Tnterczyit.Phys.Let1..11.190(1m). (14) F w. W P. F meaws-c Tatum. None. J.Am. cum Soc., es. 2120(1973). ILA. Yost. C. G. Enlie‘ Department of Chemistry. Michigan State University East Lansing. Michigan 48824 Received November i' O. ’977 APPENDIX B International Journal of Mass Spectrometry and [on Physics. 30 (1979) 1 27-136 127 © Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands HIGH EFFICIENCY COLLISION-INDUCED DISSOCIATION IN AN RF -ONLY QUADRUPOLE R.A. YOST and C.G. ENKE * Department of Chemistry, Michigan State University, East Lansing, Michigan 48824 (U.S.A.) D.C. McGILVERY ", D. SMITH and JD. MORRISON Department of Physical Chemistry, La Trobe University. Bundoora, Victoria. 3083 (Australia) (First received 5 July 1978; in final form 2 October 1978) ABSTRACT Collision-induced dissociation (CID), when performed in an RF-only quadrupole is a highly efficient method of fragmenting ions. The low-energy (5—10 eV) CID process may involve direct vibrational excitation by momentum transfer but, in any case, is a very different process from the high-energy (3—10 keV) electronic excitation CID process observed in MIKES and CAMS. Experimental results are presented which demonstrate the efficiency of CID fragmentation (up to 65%), the elimination of scattering losses, and the effects of varying such experimental parameters as choice and pressure of collision gas, ion velocity in the quadrupole, and quadrupole RF voltage and frequency. The appearance of the CID fragmentation spectra is roughly similar to 14-eV El spectra. The results of digital simulation of ion trajectories in an RF-only quadrupole are presented. The correspondence between simulated and experimental results aids in the understand- ing of the quadrupole CID process. The high efficiency of the CID process in an RF-only quadrupole is significant in the development of a tandem quadrupole mass spectrometer for selected ion fragmentation studies. INTRODUCTION One of the most exciting areas in mass spectrometry is selected ion frag- mentation, in which an ion is mass-selected, fragmented, and the resulting fragment ions are mass-analyzed. Collision-induced dissociation (CID) has received considerable interest recently as a fragmentation technique for selected ion fragmentation. These studies have been performed on reversed- geometry double-focussing mass spectrometers at ion kinetic energies of 3— 10 keV by several groups, see for example Kondrat and Cooks [1] and Mc- * To whom correspondence should be addressed. “ Present address: Department of Chemistry, Cornell University, Ithaca, New York 14853, U.S.A. 814 85 128 Lafferty and Bockhoff [2], under labels such as MIKES, CAMS, and DADI '. We have investigated the selected ion fragmentation technique using a very different implementation, i.e. a triple tandem quadrupole system [3]. In this system, CID occurs in the center quadrupole operated in an RF-only (total ion) mode, and at ion kinetic energies of 5—20 eV. At these low ener- gies, the CID process may involve vibrational excitation by momentum trans- fer, but in any case is quite different from the high-energy C1D process ob« served in MIKES, which involves electronic excitation [4]. Experiments show that the low-energy quadrupole CID process is cha- racterized by very efficient fragmentation (conversion of up to 65% of the ions into fragments) and the virtual elimination of ion losses even at collision gas pressures of 2 X 10" torr. The CID spectra obtained show some similar- ity to the low-energy (12—20 eV) electron impact (El) spectra of the corre- sponding molecules. The efficiency of the induced fragmentation and the degree of fragmentation (relative abundance of fragments which have higher appearance potentials) in the CID spectra are affected similarly by changes in the experimental parameters. Increasing the collision gas pressure or molecular weight increases the induced fragmentation. With the instru- ment used in these studies, the C1D efficiency exhibits a minimum at an ion axial energy of 10 eV. Neither the RF voltage nor frequency on the CID quadrupole has any significant effect on the CID process. These effects can be explained by theoretical and simulation studies of the low-energy CID process in a quadrupole. EXPERIMENTAL These experiments were performed on a triple quadrupole system origi- nally designed for studies of laser-induced photodissociation which has been described [5]. A diagram of the system appears in Fig. 1. For these experi- ments, a collision gas (hydrogen or argon) was introduced into the center quadrupole (1.9 cm diameter rods, 15 cm long). The quadrupole was oper- ated at 2.8 MHz over a range 0-1 kV peak RF. Collision gas pressures up to 2 X 10" torr were used; this limit is imposed by the lack of differential pumping between the CID region and the electron multiplier. Collision gas pressures were measured with an ionization gauge and corrected for the specific CID gas. The first and third quadrupoles (0.79 cm diameter rods, 20 cm long) were tuned to provide unit mass resolution over the entire mass range 0-100 u. Samples were introduced through an unheated batch inlet system and ionized typically in an electron impact source with 7 O-eV elec- trons. " MIKES-Mass-analyzed Ion Kinetic Energy Spectrometer. CAMS- Collisional Acti- ' vation Mass Spectrometer. DADI - Direct Analysis of Daughter Ions. 86 1 29 common sauna 61:5 L H i E, 10., MD 1 :3 r 10" sounce ouao 2 QUAD 3 $9532, GAUGE (nr ONLY) 1.... m F—lwl Fig. 1. Diagram of the tandem quadrupole mass spectrometer system showing the center RF-only quadrupole C1D region. SIMULATIONS The quadrupole CID process was studied using digital simulation of ion trajectories in an RF-only quadrupole. The simulation programs already described [6] operate on a PDP 11/40 minicomputer with 28K words mem- ory, cartridge disk, Tektronix 4010 graphics terminal, and an RT-ll operat- ing system. The simulations assume round rods and a point-by-point descrip- tion of the field within the quadrupole. Fringing fields are ignored. An ion may be “injected” into the field and its motion traced. The trajectories are observed to resemble ellipses, centered around the quadrupole axis. An example appears in Fig. 2. Ion trajectories for the parameters used in these experiments exhibit much smaller orbits. Approximate empirical expressions which describe the ion motion have been obtained by observing the effects of each of the variables on the characteristics of the trajectory. Simulations covered the range 100 kHz to 20 MHz, 10-1000 V peak RF, 1-2 cm diameter rods, and ions of mass 5— 600. The expressions thus obtained for the period and path length of the generally elliptical orbits (and hence average velocity), as well as the upper and lower mass limits, appear in Table 1. Note especially that the ion’s average transverse velocity (and hence the path length through the quadru- pole and transverse kinetic energy) is independent of the RF voltage and fre- quency. Neither RF parameter should then affect the CID process, as is experimentally observed. Other experiments have been performed to test the validity of the simula- tion results in Table 1. Measuring the transmission of ions with m/z 15, 28, and 41 from cyclohexane as a function of peak RF voltage gives a value of 0.4 for the constant in the expression for Mm“, the lower mass limit. In comparison, the stability diagram for the mass filter [7] predicts Mm", = (0.3 V)/(f’ d’). The upper mass limit is experimentally too indistinct to com- pare with the simulated expression. The period of the ion orbit may be 130 Fig. 2. Simulated ion trajectory. Ion m/z 8 36. Peak RF voltage = 25 V. RF frequency = 0.35 MHz. Diameter of rods = 1.9 cm. Off-axis energy = 1 eV. Time = 40 us. experimentally determined. Under certain Operating conditions, the maxi- mum excursion of the ions from the axis of the quadrupole is large com- pared to the exit aperture. In this case the signal exhibits a maximum when- ever the ions have executed an integral number of half-orbits (that is, they TABLE 1 Expressions describing simulated ion trajectories in RF-only quadrupoles ' . 6 X 10'6 mfd2 Orbit period T V (s) Orbit length l - 5 X 10':an M, (m) Average velocity F- '3 I 9 X 103\/E/—rn (m s”) Lower mass limit Mm", . 9f—fd—r (u) Upper mass limit Mm“ . W (u) ‘ V - Peak RF voltage (V); I - RF frequency (MI-12); E - off-axis energy (eV); m . ion mass (u); d - pole diameter (cm). 88 131 are near the axis of the quadrupole) as they reach the end of the quadrupole. By changing the RF voltage, frequency, or ion mass and observing the resul- tant beat patterns in the ion signal, the dependence of the orbit period on these experimental parameters can be determined. Experimentally, the orbit period appears to obey the relationship 7 = (7 X 10“ mfd3)/V. a result which is quite close to the expression for the simulated behavior. The period of the fundamental ion motion can be calculated from the theory [7], and can be approximated for an RF-only quadrupole by ‘r = (7.5 X 10"5 mfd2)/V. The validity of the trajectory simulations is demonstrated by the agreement between the simulated results and both the experimental and theoretical values. The usefulness of the simulations is proven by the assistance they provide in understanding the quadrupole CID process. CID EFFICIENCY The efficiency of the CID process is determined by the efficiency of frag- mentation and the efficiency of collection of the ions. We can let P“ and P represent the parent ion beam current at the entrance and exit of the CID region, respectively, and 21", the total current of all fragment ions at the exit of the region. The fragmentation efficiency is E,- = SFi/(P + 51“,), the collec- tion efficiency is EC= (P+ SFQ/Po, and the overall CID efficiency is the product 5cm = E; - EC = DIR/Po. The collection efficiency in the quadrupole CID region is virtually 100%. There is no detectable loss of ions owing to scattering, neutralization, or similar mechanisms at pressures up to 2 X 10" torr. The strong focussing nature of the RF-only quadrupole stabilizes the ion after collision and frag- mentation. Preliminary simulation studies show that the ion trajectories after collision and fragmentation may contract, but do not get larger. Similar results have been seen for simulated collisions in a three-dimensional quadru- pole [8]. Fragmentation efficiency in the quadrupole CID system for the com- pounds studied ranges from 15% for benzene to 65% for n-hexane. This cor- responds to fragmentation cross-sections of 9X 10"‘-—1 X 10'“ cm’. At pressures of 2 X 10" torr, the mean free path (assuming an ion collision dia- meter of 5 A) is approximately equal to the length of the CID quadrupole, 15 cm. The actual ion path length through the CID quadrupole is approxi- mately one-and-a-half times the length of the quadrupole owing to the ion’s oscillations in the RF field. While transversing 1.5 mean free paths, approxi- mately 87% of the ions experience one or more collision, 47% two or more, and 26% three or more. In the case of the n-hexane, in which 65% of the parent ions are fragmented, 75% of the collisions must lead to fragmentation. This very high efficiency of the low-energy CID process is a very clear indi- cation that the low-energy collisions do not follow the Massey “adiabatic criterion” [9] which gives the limit for the maximum energy transfer in high. energy collisions. 89 132 TABLE 2 Effect of ion axial voltage on fragmentation efficiency ‘ Ion axial voltage :Fi/(P + 217,) (V) 0 0.22 2.5 0.19 5 0.17 10 0.1 5 15 0.45 ' Cyclohexane 84’ parent ion. Argon CID gas at 2 X 10" torr. Peak RF voltage = 72 V. Ion axial voltage in first and third quadrupoles I 11 V. Although the collection efficiency remains nearly 100% under all experi- mental conditions, the fragmentation efficiency is strongly influenced by changes in experimental variables such as collision gas species and pressure and ion axial energy (see Table 2). Experiments on the triple quadrupole system show that argon produces fragmentation efficiencies three to four times greater than those measured when hydrogen is the collision gas. At the low energies used in the quadrupole system (5—20 eV), CID may occur through vibrational excitation by momentum transfer [4]. Hence the more massive argon atoms would be expected to be more efficient for low-energy C1D than hydrogen. At the high energies used in MIKES systems (3-10 keV), CID occurs by vertical electronic excitation followed by relaxation ’0 84O-084. J ’ - '\ “‘7' \\ / I. 03 ” CYCLOHEXANE . Vz . 5v ‘ VpRF - 72V 06 ‘- c463 5 04 >- O4°'°56.(15) ,J -10 02- f4“? sausage-50)“ 0 IO" IO' 5 IO" IO'3iorr CID GAS PRESSURE (ARGON) Fig. 3. Relative intensity of the parent ion (PIPO) and individual fragment ions (Pi/Po) as a function of CID pressure (argon) for cyclohexane 84’. Ion axial voltage - 5 V. Peak RF voltage - 72 V. 90 133 into vibrational excitation [4]. In this light it is not surprising that some workers find hydrogen or helium to be most efficient for CID in MIKES systems [10,11]. The use of a light collision gas undoubtedly reduces scat tering in the field-free drift region used for CID in MIKES systems. ' Fragmentation efficiency increases as the CID pressure is increased. The effect of the argon collision gas pressure on the intensity of the 84° parent peak and the individual fragment ions of cyclohexane is shown in Fig. 3. The intensity of all the fragment ions increases dramatically above 10" torr. Future experiments will extend these studies above the 2 X 10"-torr limit of the system at La Trobe. For a first-order dependence of the CID process on collision gas, we would expect plots of IMP/Po) vs. pressure or ESE/P vs. pressure to be linear. The quantity and precision of the current data are not sufficient to confirm or refute the hypothesis of a first-order fragmenta- tion process. The axial energy of the ions as they pass through the CID quadrupole has a marked influence on the efficiency of fragmentation. The fragmentation efficiency exhibits a minimum at an ion axial voltage of 10 V. At higher axial voltages, the increased kinetic energy of the ions leads to more ener- getic collisions and therefore more efficient fragmentation. At ion axial energies below 10 eV, the kinetic energy of the transverse motion of the ions as they oscillate in the RF field (about 5 eV according to the simula- tions) is the main source of collision energy. Although the collision energy is lower, the lower ion axial velocity increases the number of orbital oscilla- tions the ion undergoes through the quadrupole, and this increased path length results in more collisions. We did not initially expect that the frag- mentation efficiency would be nearly independent of the RF voltage and fre- quency. However, this experimental observation is supported by the simula- tions, in which the ion’s transverse velocity (and hence the total path length and ion kinetic energy) is independent of the RF voltage and frequency. In summary, the overall CID efficiency may be'increased by (1) increasing the mass of the collision gas molecules, (2) increasing the CID pressure, or (3) increasing or decreasing the ion axial voltage from 10 V. The efficiency of the CID process is independent, however, of both the frequency and peak RF voltage on the CID quadrupole. APPEARANCE OF CID SPECTRA A very important factor in assessing the usefulness of CID as a technique for selected ion fragmentation is the appearance of the resulting mass spectra. Examples of the CID spectra obtained are presented in Fig. 4. The spectra of n-hexane and benzene demonstrate the characteristic and rela- tively rich spectra seen. An estimate of the amount of energy involved in fragmentation may be obtained by comparing the CID spectra with electron impact spectra. High- energy MIKES CID shows similarities between CID spectra and 50-eV El 91 134 as D 52 2-IO“ to" Argon a 2-IO" '0" ARGON 56 V1 _ 5v v: '5)! VpRF - 250v VpRF ”25°" 70 -. HO) 4:: “ 39 _ . ‘a j in ’7 i," f‘. E [3 U I? 1' ’ I; .:‘ 1 J5, l 26 l 5 (I 29 s II ...__._.__/'_.___.’I' ___.§\3_. __J I . _J~____J‘-.L_ - .... lg - --.—L -. ._J _.'___ .__-.-_ : _ i l 1 » . - m/l 3O 40 50 60 7O 80 m/z 310 40 50 610 7O 80 CD SPECTRUM 0F PURE BENZENE cno SPECTRUM 0F PURE n- HEXANE Fig. 4. CID spectra of (a) benzene 78’ and (b) n-hexane 86°. Argon CID gas at 2 x 10" torr. Ion axial voltage = 5 V. Peak RF voltage - 250 V. spectra [12]. In the quadrupole system, however, we are observing a very different CID process. The CID spectra obtained resemble 12—20-eV EI spectra, with 14 eV being typical, although differences in the relative abundance of some fragments are apparent. The abundance of the CID fragment ions can also be compared with the differences between the appearance potentials of the normal ions and the ionization potential of the parent ion. The CID spectra and appearance potential of the corresponding ions for benzene and n-hexane are presented in Table 3. Note that the intensity F , of the fragment ions generally shows a correlation with the appearance potentials; the greater the appearance poten- tial, the lower the intensity. The effects of the experimental parameters of the quadrupole CID system on the appearance of the spectra follow closely the effects seen on CID effi- ciency. Increasing the mass of the collision gas or increasing the ion axial vol- tage above 10 V increases the intensity of the ions of higher appearance potentials (that is, increases the electron energy in the corresponding EI spectrum). Increasing either of these variables produces more energetic colli- sions. Increasing the collision gas pressure, or decreasing the ion axial voltage below 10 V also increases the intensity of ions of higher appearance potentials, since either change increases the number of collisions and the probability of multiple collisions. It would not be expected that the change due to more multiple collisions would be the same as that due to more energetic colli- 92 135 TABLE 3 Comparison of CID spectra ‘ and ion appearance potentials m/z AP—IP (eV) b CID c Benzene 77 4.6 18 76 5.0 3 52 5.3 45 39 5.5 22 63 6.5 11 50 8.3 4 51 8.4 6 26 9.8 0.3 27 9.9 0.1 n-Hexane 71 0.8 0.4 70 0.8 0 57 0.8 30 56 0.8 32 42 0.8 8 43 1.1 19 ' 2 X 10“ torr argon CID gas. Ion axial voltage 5 V. Peak RF voltage 250 V. b Appearance potential (AP) for normal ion minus ionization potential (IP) for molecular (parent) ion. Benzene data from ref. 13, n-hexane data from ref. 14. c Total fragment ions - 100. sions, but more complete data will be necessary to make that comparison. Neither RF voltage nor frequency have a noticeable effect on the appearance of the CID spectra. CONCLUSIONS It has been demonstrated that the triple quadrupole system using CID at low energies in a center RF-only quadrupole is a simple, efficient system for selected ion fragmentation. Quadrupole CID shows a high fragmentation effi- ciency (up to 65%) and essentially 100% collection efficiency. The low energy CID process is different in several significant aspects from the high- energy process employed in MIKES. Digital simulation of ion trajectories in an RF-only quadrupole have been shown to be in good agreement with quadrupole theory and with experimen- tal data. Results of the simulation are a valuable aid in understanding the quadrupole CLD process. A triple quadrupole system for selected ion fragmentation has recently been completed at Michigan State University. It will enable us to extend these studies to higher masses, tolerate higher CID gas pressures, control the ion axial voltage over a wide range, and fragment ions that are generated by 93 136 chemical ionization as well as electron impact. The new system will allow us to more completely characterize the low-energy CID process and investigate its application to mixture analysis and structure elucidation. ACKNOWLEDGEMENTS This work was supported in part by the Office of Naval Research. One of us (R.A.Y.) gratefully acknowledges a National Science Foundation Grad- uate Fellowship, an American Chemical Society Division of Analytical Chemistry Fellowship sponsored by the Upjohn Company, and an L.L. Quill Memorial Fellowship from Michigan State University. REFERENCES R.W. Kondrat and R.G. Cooks, Anal. Chem., 50 (1978) 81A. F.W. McLafferty and F.M. Bockhoff, Anal. Chem., 50 (1978) 69. RA. Yost and CG. Enke, J. Am. Chem. Soc., 100 (1978) 2274. H. Yamaoka, D. Pham and J. Durup, J. Chem. Phys., 51 (1969) 3465. D.C. McGilvery and J.D. Morrison, Int. J. Mass Spectrom. Ion Phys., 28 (1978) 81. D.C. McGilvery and J.D. Morrison, in preparation. Peter H. Dawson, Quadrupole Mass Spectrometry and Its Applications, Elsevier, New York, 1976, p. 2. RF. Bonner and RE. March, Int. J. Mass Spectrom. Ion Phys., 25 (1977) 411. H.S.W. Massey and E.H.S. Burhop, Electronic and Ionic Impact Phenomenon, Oxford University Press, London, 1952, p. 513. 10 E“). McLafferty, P.F. Bente, III, R. Kornfeld, S.-C. Tsai and I. Howe, J. Am. Chem. Soc., 95 (1973) 2120. 11 K. Levsen and H. Schwarz, Angew. Chem., Int. Ed. Engl., 15 (1976) 509. 12 KR. Jennings, Int. J. Mass Spectrom. Ion Phys., 1 (1968) 227. 13 C. Lifshitz and B.G. Reuben, J. Chem. Phys., 50 (1969) 951. 14 B. Steiner, C.F. Giese and MC. Inghram, J. Chem. Phys., 34 (1961) 189. \ld}(fl&wNH coon