u...“ ,E'L.‘ an ., ‘ 2X7 l ll'llllllllllillllfllllll 1— 1293 00852 7255 ‘ i Illljl E ~-k‘.i:-'¢-7 ! uwmnr'mqu a we.” . . .3 .1 -\'A.PA—Jr I. L-II v' ‘- Ui- :u This is to certify that the thesis entitled MODIFICATION OF A TIME—OF—FLIGHT MASS SPECTROMETER FOR THE ANALYSIS OF GAS CHROMATOGRAPHIC EFFLUENTS presented by David Michael Guido has been accepted towards fulfillment of the requirements for M. S . degree in Chemistry Major professor Date 23 AWIVCESI. 0-7639 MS U is an Affirmative Action/Equai Opportunity Institution MSU RETURNING MATERIALS: Place in book drop to LIBRARIES remove this checkout from .—_——. your record. ‘FINES will be charged if book is returned after the date stamped below. 89v 132:. ». . . v ‘ ’@32401 uq-_ MODIFICATION OF A TIME-OF-FLIGHT MASS SPECTROMETER FOR THE ANALYSIS OF GAS CHROMATOGRAPHIC EFFLUENTS By David Michael Guido A THESIS submitted to Michigan State University in partial fulfillment of the requirement for the degree of MASTER OF SCIENCE Department of Chemistry 1985 ABSTRACT MODIFICATION OF A TIME-OF-FLIGHT MASS SPECTROMETER FOR THE ANALYSIS OF GAS CHROMATOGRAPHIC EFFLUENTS BY David Michael Guido The technique of gas chromatography/mass spectrometry (CC/MS) employing time-of-flight (TOP) mass spectrometers has not been as popular as the use of more conventional magnetic sector and quadrupole mass spectrometers. However, the advent of an integrating transient recorder which will permit summing of sequential transients should stimulate increased utilization of TOP mass spectrometry. A conventional TOP mass spectrometer was configured for GC/MS by modifying the pumping system to include differential pumping. The GC/MS system utilized an open-split interface and, more successfully, a direct column to source connection method of interface. Chromatographic separation and detection of hydrocarbons up to C20 were possible before the unheated source became a problem. A study was also conducted to compare the fragmentation patterns or mass spectra acquired on the time-of-flight and magnetic sector mass spectrometers. Results indicate that the fragmentation pattern of a given compound is independent of instrument type but changes as a function of source temperature. Additionally, compounds of different type manifest different changes in their patterns as a function of temperature. ACKNOWLEDGEMENTS I would like to thank Dr. J.T. Watson for his support and guidance throughout my graduate career. I want to thank all the members of the Watson group, the Mass Spectrometry Facility, and the Biochemistry Shop for their friendship and advice during the course of this work. Finally, I wish to thank my wife, Jane, and my parents whose love and encouragement were vital to the completion of this degree. ii II. III. TABLE OF CONTENTS List of Tables List of Figures Introduction The Technique of Gas Chromatography/Mass Spectrometry Methods of Data Acquisition TOF Ideal for ITR Goal of this Work The Time-of-Flight Mass Spectrometer Principles of Operation CVC M82000 Time-of-Flight Mass Spectrometer Ion Source Magnetic Electron Multiplier Pumping System Modification to Pumping System Interface of GC to MS Introduction Separators Open-Split Interface Direct Connection Interface Open Source vs. Tight Source Experimental iii vi 10 ll 11 14 18 22 24 25 32 32 36 39 41 42 43 IV. iv Results and Discussion Comparison of Fragmentation Patterns Introduction Experimental Results and Discussion List of References 48 59 59 62 63 81 LIST OF TABLES Table 1: System pressure before and after vacuum system modifications. The Penning gauge measures the pressure in the analyzer region and the ion gauge monitors it in the source. Table 2: Compatibility of gas chhromatography with other instrumental analytical methods [1]. Table 3: Comparison of percent total ion intensity (T11) of some peaks in mass spectra of selected compounds acquired on the time-of-flight and magnetic sector mass spectrometers. Table 4: Variation in relative intensities of major peaks in five consecutively recorded mass spectra of acetophenone from the time-of-flight mass spectrometer. Table 5: Variation in relative intensities of major peaks in five consecutively acquired mass spectra of acetOphenone from the magnetic sectror mass spectrometer. 31 ,33 66 72 73 LIST OF FIGURES Figure 1: Comparison of true gas chromatographic profile (dashed line) with mass chromatograms (solid lines). (A) Mass chromatogram prepared from mass spectra acquired at a rate of l scan/s. (B) Mass spectra acquired at a rate of l scan/s, but synchrony of chromatogram and scan cycle shifted by one-third second. (C) Mass spectra acquired at a rate of 3 scans/s [7]. Figure 2: Spectrum of n-butane for comparison of time slice detection and time array detection. (A) In time slice detection, only one time bin is measured for each pulse of the ion source. Thus, many pulses are needed to acquire the entire spectrum. (B) In time array detection, the entire spectrum is acquired from each pulse of the ion source [7]. Figure 3: Simplified diagram of a time-of-flight mass spectrometer. (Adapted from ref. 25.) Figure 4: (A) The principles of time-lag focusing. Part I shows the position of the ions at the time of ion formation. The electron beam is then turned off and the ions have motion due to their kinetic energy. The time delay between the instant the electron pulse is turned off and the instant the ion draw-out pulse is turned on is called the time lag. In Part II the draw-out pulse is applied and ions are drawn out from their new spatial positions. As the ions drift toward the detector, in Part III, the ions furthest away from the detector recieve additional energy and catch up to the ion bunch as it reaches the detector, this is shown in Part IV [29]. (B) Plot of flight time versus ion initial position. Figure 5: (A) Partial spectrum of toluene with and without time-lag focusing. Resolution at m/z 91 is 800 using FWHM definition when no time-lag is used. With a time-lag of 1.1 microsecond, the resolution is 1138. (B) Partial spectrum of the xenon isotopes. Resolution at m/z 132 is 858 with no time-lag and 1353 with a time-lag of 1.3 microsecond. Figure 6: Ion source of the CVC M82000 time-of-flight mass spectrometer. Figure 7: Magnetic electron multiplier used in the CVC M32000 time-of-flight mass spectrometer. (adapted from ref. 25.) vi 12 16 19 20 23 vii Figure 8: Vacuum system of the CVC M82000 time-of-flight mass spectrometer. The shaded areas indicate the modifications that were made to accomodate the GC inlet system. The vacuum valves indicated are (1) valve from analyzer tube to high vacuum pump; (2) backing valve for high vacuum pump; (3) valve for roughing analyzer tube; (4) valve for evacuating inlet system; (5) needle valve from batch inlet to ion source; (6) valve from ion source to high vacuum; (7) two position backing/roughing valve; (8) GC/MS interface probe vacuum lock; (9) valve for roughing vacuum lock; (10) valve for venting system. Figure 9: Diagram of circuit that was built to Operate the Diffstak diffusion pump and the backing/roughing valve. Figure 10: Sketch of pneumatically operated valve system showing solenoid valves in their de-energized positions. Figure 11: Molecular separators. (A) watson-Biemann effusion separator [1]; (B) Ryhage two-stage jet separator [30]; (C) Llewellyn membrane separator [30]. Figure 12: Open-split interface. Figure 13: Schematic showing open-split interface in CC floor. Figure 14: GC/MS interface probe. Figure 15: Schematic of the direct-connection interface showing the capillary line running through the bore of the open-split interface. Figure 16: Sensitivity comparison of (A) open-split interface and (B) direct connection interface. In each case 2 micrograms of acetophenone were injected. The chromatographic conditions used were: injection port, 280°C; isothermal, 130°C; interface, 250°C. A 30 meter, 0V-101 capillary column was used. Ion currents at m/z 105 were monitored. Figure 17: Selected ion current profile of an equal weight mixture of selected n-alkanes. C was injected but no peak was detected. The chromatographic conditions used were: injection port, 250°C; temperature program, 55°C to 265°C at 3°C/min; interface, 300°C. A 30 meter, 0V-101 capillary column was used. Ion currents at m/z 57 were monitored. Figure 18: Selected ion current profile of standard methyl ester of fatty acid mixture. The chromatographic conditions used were: injection port, 250°C; temperature program, 70°C to 265°C at 3°C/min; interface, 300°C. A 30 meter, 0V-101 capillary column was used. Ion currents at m/z 74 were monitored. 26 29 30 37 40 44 45 49 51 52 54 viii Figure 19: Selected ion current profile of standard methyl ester of fatty acid mixture. The chromatographic conditions used were: injection port, 250°C; temperature program, 200°C to 265°C at 3°C/min; interface, 300°C. A 30 meter, 0V-101 capillary column was used. Ion currents at m/z 74 were monitored. Figure 20: Selected ion current profiles showing detection limit of methyl stearate using on-column injection. (A) 45ng and (B) 9ng methyl stearate. GC conditions used were: injection port, 250°C; temperature program, 200°C to 265°C at 3°C/min; interface, 300°C. A 30 meter, 0V-101 capillary column was used. Ion currents at m/z 74 were monitored. Figure 21: Possible effect of the magnetic field of the magnetic elctron multiplier. (A) Absence of the field. (B) Deflection of ion bunch in the presence of the field. Figure 22: Mass spectra of acetophenone. (A) Spectrum produced by a time-of-flight mass spectrometer. (B) Spectrum produced by a magnetic sector mass spectrometer. Figure 23: Mass spectra of 1,5-cyclooctadiene. (A) Spectrum produced by a time-of-flight mass spectrometer. (B) Spectrum produced by a magnetic sector instrument. Figure 24: Mass spectra of hexachloro-l,3-butadiene. (A) Spectrum produced by a time-of-flight mass spectrometer. (B) Spectrum produced by a magnetic sector instrument. Figure 25: Mass spectra of 1,5-cylcooctadiene acquired on the magnetic sedtor mass spectrometer. (A) Source temperature was 180°C and separator temperature was 55°C. (B) Source temperature was 180°C and separator temperature was 300°C. Figure 26: Mass spectra of acetophenone acquired on the magnetic sector mass spectrometer. (A) Source temperature was 180°C and separator temperature was 55°C. (B) Source temperature was 180°C and separator temperature was 300°C. 55 57 61 64 67 70 76 78 I. INTRODUCTION The Technique of Gas Chromatography/Mass Spectrometry The coupling of gas chromatography with mass spectrometry has provided the chemist with one of the most powerful analytical techniques. Before combined gas chromatography/mass spectrometry (CC/MS) was utilized, it was impractical to do a complete qualitative analysis on a complex organic mixture of 20 or more components [1]. Today GC/MS is used for the analysis of such diverse substances as biological fluids [2], air pollutants [3], and pesticides [4] which may be composed of more than 400 components. Furthermore, the amount of any component may range from the nanogram level to more than 992 of the mixture [5]. When confronted with the requirements of this type of analysis, simple methods fail. Gas chromatographic data alone cannot yield reliable evidence of structure in the absence of prior information. The retention time of the unknown must be used as a means of identification. This method has a serious drawback: many compounds can be assigned to any given retention time. Only when general structural features of the sample are already known, can GC provide information through correlations with reference compounds. Thus, for qualitative analyses of complex mixtures, gas chromatography must be coupled with other analytical systems. In mass spectrometry, structural evidence can be obtained from fragment ions produced in the source including, quite often, the molecular ion from which molecular weight information can be inferred. When bombarded by electrons, every substance ionizes and fragments uniquely. Some groups of compounds give very similar mass spectra, but as a rule, the process provides a distinctive fingerprint. Structure can be inferred by applying general rules governing fragmentation, for example, cleavage being favored at branched carbon atom sites or at bonds beta to a hetero atom and so forth. Also, a basic requirement for compound identification by mass spectrometry is that the sample be pure and in the gas phase. The great capacity of gas chromatography to accomplish the separation of complex mixtures and supply the eluting components in the gas phase combined with the sensitivity and the specificity of mass spectrometry to provide identification and quantitative determinations of the components make GC/MS a highly desirable method of analysis. Methods of Data Acquisition In a GC/MS system, the most common mode of operation is repetitive scanning in which complete or partial mass spectra (i.e., ion abundance versus mass/charge ratio) are continuously recorded [6]. A computer is used to continuously and automatically record mass spectra of the gas chromatographic effluent, whether or not a sample component is present in the chromatographic fraction, and stores all spectra on disk. It is desirable to obtain a minimum of 6 or more mass spectra across a chromatographic peak. From the consecutively recorded spectra that are stored on disk, the computer can plot the intensity of the total ion current versus scan number. These reconstructed total ionization plots are very similar to gas chromatograms as recorded conventionally. The computer can also plot the intensity of one specific mass versus spectrum number. Such a plot can be thought of as a gas chromatogram recorded with a detector that is specific for a given_mass spectral fragment and thus has increased specificity for a given compound type. Since each plot delineates the history of a single mass throughout the chromatogram, these graphs are called "mass chromatograms". Because it is impractical to generate such plots for each of several hundred masses, these plots are generated only for those masses selected on the basis of knowledge of the sample, mass spectrum of a compound, or a group of compounds suspected of being present rather than for every mass ° in the range of the spectrum. The resolution of gas chromatography has been improving due to the need for better separation, but mass spectrometer scanning rates are not keeping up. The improvement in GO resolution makes it difficult to obtain a mass spectrum of the emerging sample component from the GC column because the mass spectrometer scanning speed is not fast enough to take the complete spectrum before the concentration of the GC fraction in the source changes. Mass spectrometers' scanning speeds are being pushed to their limits [7]. This problem is illustrated in Figure 1, which compares a true gas chromatographic profile with reconstructed mass chromatograms. The greater the frequency of spectrum acquisition, the greater the number of points available to define the chromatographic -" J'- o" ( ‘ salelJA l gh—A—l—h - O 2 4 6 8 K) O 2 SEUONDS $803 4 6‘s o 05 SECONDS Figure 1: Comparison of true gas chromatographic profile (dashed line) with mass chromatograms (solid lines). (A) Mass chromatogram prepared from mass spectra acquired at a rate of 1 scan/s. (B) Mass spectra acquired at a rate of 1 scan/s, but synchrony of chromatogram and scan cycle shifted by one-third second. (C) Mass spectra acquired at a rate of 3 scans/s [7]. profile. To get around this problem, a smaller mass range can be used to increase the scanning rate. However, this method fails when a complete spectrum is needed as in the case of an unknown. Another way to bypass the problem is to use a technique originally designed for increased sensitivity and detection of co-eluting components from the GC, called selected ion monitoring [8]. Selected ion monitoring (SIM) can be used when the time required to scan the full spectrum is too long. It is capable of much greater sensitivity than rapid, repetitive scanning primarily due to the fact that in SIM, the output from the electron multiplier can be integrated over a longer time interval, thus enhancing the signal to noise ratio. In this technique, the detection and recording system of the mass spectrometer are dedicated to acquiring the ion current profiles at only certain selected m/z ratios. However, SIM can only be used if the m/z values of interest are known at the start of the run. Therefore SIM requires some pre-knowledge about the sample [9]. SIM precludes the capacity for repetitive scanning since the selected ion currents are continuously being monitored. The selected ion current profile, as the output from this method of data collection is called, provides the data necessary for good chromatographic profiles. However, the shortcomings of this method are that complete mass spectra are not provided, one must know which m/z values to monitor, and it is a highly specific rather than a universal method of detection. This last disadvantage is actually an advantage when SIM is used for the detection of certain compound types or for selectivity in the presence of co—eluting components which is the reason SIM was initially develOped. Thus, when little or nothing is known about the sample, the method of repetitive scanning is necessary. There is a limit to how fast the spectrum can be scanned in the magnetic sector and the quadrupole instruments due to inherent design. In the magnetic sector instrument, the mass axis is usually scanned by scanning the magnetic field. However, the magnetic field must not change significantly during the transit time of an ion; otherwise the resolution and the sensitivity will be reduced. This loss in resolution and sensitivity occurs because the exit slit is no longer at the focal point for that particular m/z value. In the quadrupole mass spectrometer, ions are focused by the use of alternating fields on the quadrupole rods. Rf and dc voltages are applied between opposite pairs of rods and the ions are sorted according to their path stability as they pass through the analyzing region. For maximum transmission, the rf and dc voltages should not change during the transit time of an ion through the filter. Otherwise, the result would be loss of intensity because the ion of a particular m/z value will not have the optimum trajectory through the analyzer region. Thus, in the magnetic sector and quadrupole mass spectrometers, the mass scanning rate is limited by design characteristics of the instrument. During repetitive scanning with conventional data collection, the instrument detects ion currents of only one mass at a time while ions of all other masses strike the walls of the vacuum chamber and are lost. Time-of—flight mass spectrometry (TOFMS), on the other hand, operates in a cyclic mode in which 10,000 or more spectra are produced each second. However, the full data output capability of TOFMS is rarely used [10]. Thus, it is not the design of the instrument that limits the scanning rate in TOFMS to several spectra per second, but rather the detection method. Most commercial time-of-flight mass spectrometers contain an analog output system that effectively samples only one point in the spectrum during each cycle of the instrument. This is referred to as time slice detection (TSD) as in Figure 2(a). To obtain a spectrum using TSD in TOFMS, the output device integrates the current in individual time slices or time bins over a relatively large number of instrument cycles. With a magnetic electron multiplier (discussed in chapter 2), it is possible to selectively gate a particular mass peak from each cycle of the instrument and collect it by one of the multiplier anodes. The current can then be measured by an electrometer circuit. Scanning is accomplished by changing the time relative to the extraction pulse at which the gate is pulsed to collect the current during a given time bin for each cycle of the instrument. The gate literally "jumps" along but the increments are so small that a continuous scan is the apparent result [11]. Thus, scan times are long. Little work has been done to collect all of the information produced in TOFMS. However, using state-of-the-art electronics, a method of data collection is being prepared at Michigan State University by Holland et a1 [7], that will use all of the information produced by TOFMS. This system is called an integrating transient recorder. To get the maximum utilization of the information, all of the ion currents striking the detector must be collected, stored, and summed in real time. This may be referred to as time array detection (TAD). TAD is TIME SLICE DETECTION Arrival Tums (p8) r-N Lor -s as -m -N to: IHENSHY :-—---——--o in- — I--- I.----- P O. to--. to. J Klfl Hmiii TIME ARRAY DETECTION mh43 IQIHRSE ’ 1 g? Auhkmunlxmxnn ”29 mass an5 ! Figure 2: Spectrum of n-butane for comparison of time slice detection and time array detection. (A) In time slice detection, only one time bin is measured for each pulse of the ion source. Thus, many pulses are needed to acquire the entire spectrum. (B) In time array detection, the entire spectrum is acquired from each pulse of the ion source [7]. compared with TSD in Figure 2. Consecutive scans will be summed and stored as a single spectrum to give increased sensitivity and dynamic range over conventional TSD. TAD will use all available information by acquiring the entire spectrum from each pulse of the ion source and is expected to provide 10 to 1000 complete, averaged spectra per second. TOF Ideal for ITR As stated above, a unique feature of the time-of-flight mass spectrometer is that ions of all masses are accelerated from the source simultaneously with each cycle of the instrument. All of the ions reach the detector in less than 100 microseconds. However, using TSD with TOFMS, spectral acquisition rates take about the same amount of time as with conventional mass spectrometers, which is on the order of seconds. If TAD were used on TOFMS, all of the ions would be detected from each pulse of the ion source. Because of poor ion statistics for each spectrum resulting from a single source pulse, it would be desirable to sum, say 10 or more instrument cycles to obtain a spectrum. Thus, a complete mass spectrum could be acquired in 1 msec with TOFMS as Opposed to several seconds with conventional instruments. It is necessary to acquire spectra rapidly because any fluctuations in composition or pressure of the gas stream during the mass scan of the instrument will cause the output spectrum to be skewed in peak relative intensities. 10 Goal of this WOrk In the past, time-of-flight mass spectrometry has not been very popular in combination with gas chromatography. Despite this fact, it is expected that the use of this type of mass spectrometer will increase for gas chromatography and other uses with the advent of the integrating transient recorder. To carry out the ITR research, a working capillary gas chromatography/mass spectrometry system is needed. Thus, the principal objective of this project is to provide a working gas chromatography/mass spectrometry system utilizing a time-of-flight mass spectrometer on which the ITR can be implemented. A secondary goal is to compare spectra acquired with a time-of-flight mass spectrometer to those from a more popular type of mass spectrometer, a magnetic sector instrument. This study will show whether there are any significant differences in fragmentation patterns (mass spectra) obtained from the two types of instruments. II. The Time-of-Flight Mass Spectrometer Principles of Operation The Operating principle of the time-of-flight mass spectrometer is based on the time necessary for ions of different masses but identical energy to travel down a field-free drift tube. The simplest description of the time-of-flight mass spectrometer is that it is made up of an ion source and a detector situated at opposite ends of an evacuated flight tube as shown in Figure 3. The ions are formed in the ionization region of the source, usually by electron impact. The ions are then accelerated out of the source toward the detector by a series of electric fields produced by the acceleration grids that are pulsed on at the end of the ion formation period. All of the ions receive essentially the same energy according to the equation: KE - 1/2 mv2 where RE is the kinetic energy, m is the mass of the ion, and v is its velocity. Thus ions of different masses will have different velocities. The velocity of the ions in the field-free flight path is a function of the ratio of their mass, m, to the number of charges, 2. The time-of-flight of the ions follow the equation: t-dW. Here, t is the time-of-flight, d is the distance the ion travels, e is electronic charge, and V is the accelerating voltage. By the time the ions reach the detector they have separated into bunches corresponding to their m/z ratio, the lightest ions reaching the detector first and being followed by ions of successively heavier mass. Each source pulse, 11 12 Field-Free Ion Source Drift Region Detector 9 I l | E I : -F + I I . I I *" l i + +1 : I ' + + 0+ I | + I ’: u '\ \\& Acceleration Grids Scope \\ Anode Electron Beam JLL Figure 3: Simplified diagram of a time-of-flight mass spectrometer. (Adapted from ref. 25.) 13 therefore, results in a complete mass spectrum. This occurs approximately 10,000 times each second. The first working non-magnetic TOF mass spectrometer was reported by Cameron and Eggers [12] in 1948. Their mass spectrometer consisted of a 317 cm flight tube and a single grid ion source which accelerated the ions through 300 volts. The resolution of this first device was poor, but the work encouraged others to investigate TOFMS more thoroughly. Another TOF mass spectrometer similar to the one built by Cameron and Eggers was reported by Wolff and Stephens [13] in 1953. Their single field ion source accelerated ions to 300 volts and the ion flight path between the source and the detector was 100 cm long. Discernible adjacent mass separation extended to about m/z 20. Katzenstein and Friedland [14], in 1955, reported their TOF mass spectrometer in which several electric fields were used in the source region and the ions received an energy of 250 volts. The flight path of the ions was 100 cm. A grid near the collector was pulsed negative to select a single mass bunch. Only ions gaining extra energy by this pulse could pass a repeller grid in front of the collector. The mass spectrum was scanned by slowly changing the time delay between the acceleration pulse and the selector pulse. Masses were separated up to m/z 75. 14 A real landmark in the development of time-of-flight mass spectrometry came in 1955, when Wiley and McLaren [15] reported their TOF mass spectrometer. They showed that the experimental resolution agrees closely with a more rigorous mathematical theory. The higher resolution was due to a new two-grid ion source based on careful mathematical study of the focusing preperties of TOFMS. They did not disregard the fact that ions have an initial spatial and energy spread. The double field source introduced new parameters not available in the single field source which greatly improved the resolution. The drift space was approximately 40 cm long and the accelerating potential 1600 volts. With the new ion source, peaks representing adjacent mass units were completely separated well beyond m/z 100 and useful resolution was obtained to at least m/z 300. CVC M82000 Time-of-Flight Mass Spectrometer The instrument used for this project was a CVC M82000 time-of-flight mass spectrometer. The M82000 mass spectrometer is equipped with an electron impact (EI) source of a modified Wiley‘McLaren type. It is capable of producing mass spectra at a frequency of 11 kHz with an approximate mass range of l to 900 amu. It operates with positive ion production and detection in which the ion energy is 2.7 kV. Ions are detected by a magnetic electron multiplier which has four gated outputs and predynode gating capability. The distance between the ion source and the magnetic electron multiplier is two meters. The inlet system is a batch, molecular-leak type. A four-inch oil diffusion pump with a refrigerated baffle provides a pressure of 3x10”7 torr in the 15 analyzer. The drift region between the ion source and the ion detector is maintained at high voltage by an insulated liner. The instrument is equipped for time-lag focusing, a method for improving the resolution of ions that have an initial energy by the insertion of a time delay between the ion formation pulse and the ion drawout pulse. Figure 4a illustrates that by delaying the ion drawout, the ions are allowed to move in the direction of their initial velocities. This causes the ions nearest to the drawout grid to fall through the least potential and to be accelerated to a lower velocity than ions farther from the grid. Ions that had an initial velocity away from the detector fall through more of the electric field, giving them more energy and a slightly higher velocity. The ions of higher velocity catch up to the bunch just ;s it arrives at the detector. Time-lag focusing is mass dependent. Thus, a certain delay time only focuses a single mass or a very small mass range so it is not useful for the entire spectrum for a given pulse of the ion source. The time-lag can range from 0 to about 8 microseconds. The plot of flight time versus ion initial position in Figure 4b shows that without the time lag, the three ions initially at the same position, but with velocities 0 and +vO or - v0, would have different flight times. But, during the lag period each ion moves a distance vot prior to extraction, where upon all three ions have the same flight time. If the initial energy spread is to be eliminated, the change in flight time produced must compensate for the flight time difference with no lag between ions of velocity v and 0. 16 Figure 4: (A) The principles of time-lag focusing. Part I shows the position of the ions at the time of ion formation. The electron beam is then turned off and the ions have motion due to their kinetic energy. The time delay between the instant the electron pulse is turned off and the instant the ion draw-out pulse is turned on is called the time lag. In Part II the draw-out pulse is applied and ions are drawn out from their new spatial positions. As the ions drift toward the detector, in Part III, the ions furthest away from the detector recieve additional energy and catch up to the ion bunch as it reaches the detector, this is shown in Part IV [29]. (B) Plot of flight time versus ion initial position. II III IV -270V I I. O-‘I Z --2 s I I 0V -3kV -270V 17 t = time-lag ____. .—p V‘AV O . — O Detector FLIGHT TIME I I I I I I ° I l I l | I I I I I I I I I T r f -v°r vor ION mm“. Posmou B Figure 4 18 Figure 5 shows a partial mass spectrum of toluene containing m/z 91 and 92 to illustrate the difference in resolution with and without time-lag focusing. The toluene pressure was 2 x 10”6 torr (background pressure of 8.6 x 10.8 torr) as measured in the source region. The trap current during the acquisition of the spectrum was 0.68 microamps. Using the full width at half maximum intensity (FWHM) definition of resolution, the resolution was 800 with no time-lag and 1138 with a time-lag of 1.1 microsecond. The optimum time-lag was obtained by viewing the spectrum on the oscilloscope display and adjusting the time-lag until the two peaks were in focus. Similarly, the resolution was obtained for an ion originating from an inorganic substance, xenon. The resolution with no time-lag was 858, while a time-lag of 1.2 microseconds gave a value of 1353. Ion Source The ion source, illustrated in Figure 6, has a four-grid electron beam system in which the electron beam is collimated by a series of narrow slits at the center of the grids. The operational cycle is started at 88 microsecond intervals. The electron beam is normally kept from entering the ionization region by a negative bias on the first electron control grid in front of the filament. This prevents the electrons given off by the tungsten filament from entering the ionization region by repelling them. At the beginning of a cycle, a very short, 1 microsecond, positive pulse is applied to the first electron grid, thus permitting the electron beam to pass through this grid. The energy with which the electrons pass through the source 19 91 .92 no time-lag JOLL. 129 132 no time-lag DU L. 91 92 time—lag 1.1 usec 129 132 time-lag 1.2 usec UL Figure 5: (A) Partial spectrum of toluene with and without time-lag focusing. Resolution at m/z 91 is 800 using FWHM definition when no time-lag is used. microsecond, the resolution is 1138. the xenon isotOpes. Resolution at m/z 858 time-lag and 1353 with a time-lag of 1.3 microsecond. With a time-lag of 1.1 (B) Partial spectrum of is with no 20 0V 150V max +15V 30V min 4us l ( 250V 4us _ I Trap IG-I IG-Z IG-3 IG-4 Backing Plate Deflection Plates 2 ¢ Ion Beam : :l I Electron 3 Beam 1500V 2700V Electron Control r_-1 Grid Filament Figure 6: Ion source of the CVC M82000 time-of—flight mass spectrometer. 21 region is controlled by adjusting the bias on the filament; this bias is 0-100 V with respect to the ionizing region which is at ground potential at the time of ionization. A magnetic field created by small permanent magnets, on the exterior of the manifold, collimates the electron beam so that it impinges upon the trap anode on the opposite side of the source. A sample of gaseous molecules, introduced by an inlet system is ionized in the electron beam. During ionization the components surrounding the ion chamber are at ground potential. Immediately after the electron grid pulse is turned off, the ion-drawout grid (IG 1) is pulsed to a variable voltage between 30V and 150V and the ion acceleration grid (IG 2) is pulsed to 250V simultaneously. (If time lag focusing were being used, there would be a delay time between the electron grid pulse and the ion-drawout pulse.) The duration of this pulse, 4 microseconds, is sufficient for all ions to be drawn out of the ionization region and to pass into the drift region beyond the ion energy grid (IG 4). A 2.7kV potential on the ion energy grid accelerates the ionized sample along the length of the field-free flight tube. Inside the flight tube there are pairs of horizontal and vertical deflection plates that position the most intense part of the ion beam on the cathode of the detector by way of electric fields. Ions are separated according to their various m/z ratios as they drift from the source to the detector. The lighter ions arrive at the detector first. As successively heavier ions reach the detector, the time between bunches becomes shorter because their velocity is inversely proportional to the square root of their mass. 22 If all of the ions were at rest and in a plane parallel to the ion acceleration grid before the pulse was applied, the resolution would be limited only by the extraction pulse shape and the detecting equipment [15]. In practice, however, the ions vary in initial position and velocity so that the resolving power of the time-of-flight mass spectrometer is dependent on its ability to reduce the time spread caused by the ever-present initial space and initial kinetic energy distributions. Magnetic Electron Multiplier At the end of the flight path, the separated bunches, or sheets, of ions strike the cathode of the magnetic electron multiplier with 2.7kV energy and dislodge electrons. See Figure 7. Crossed electric and magnetic fields cause the electrons to follow cycloidal paths. The cycloiding electrons enter the magnetic electron multiplier where the multiplication process is continued by successive impacts on the dynode strip. These resistive glass strip dynodes are unique and made by a process which yields a resistive surface that is very durable. They can be exposed to moisture and other atmospheric contaminants yet suffer no adverse effects on the gain characteristics of the glass. Since each primary electron dislodges about 5 electrons at the first multiplication 5 step, after 40-60 impacts, gains of 10+ to 10+7 are achieved. 23 G) B secondary gates MEI "“1 JUUU J .... “ ‘ ‘ anode 3 3 3%, 2"“ am... «mu. m... ‘ .0. ~ ion ”K17 fl resistigflgggductive o p c s s —l (21" cathode @ CRT Figure 7: Magnetic electron multiplier used in the CVC M82000 time-of-flight mass spectrometer. (adapted from ref. 25.) 24 After leaving the multiplier region, the electrons move along in cycloidal motion returning to their original energy level just above the rail. By prOperly adjusting the equipotential lines of the electric field, the electrons can be directed to any output channel desired because the cusp of the cycloid will try to return to the starting potential. The final channel, the scope anode, is always on. Any electron pulses not gated to preceding channels are picked up by the scope anode and form a series of voltage pulses across a resistor. The voltages are amplified and applied to the vertical plates of the oscilloscope. Pumping System The vacuum system of the CVC M82000, when purchased, consisted of a single 4-inch oil diffusion pump located near the middle of the flight tube; this pump was backed by an Alcatel rotary pump which also provided a rough vacuum. With the addition of a gas chromatograph, this vacuum system was not adequate to maintain a sufficiently high vacuum. A high vacuum is essential in the mass spectrometer for several reasons as outlined by G.M. Message [16]: (1) High voltage breakdown may occur in the multiplier, source or analyzer if the pressure becomes too great. (2) Oxygen from residual air and leaks will cause filament burnout in the ion source and ion gauges. (3) The mean free path of molecules in the system must be longer than the ion path through the analyzer, or ion-molecule collisions will 25. occur. Ion-molecule reactions start to occur as the pressure rises, producing changes in fragmentation patterns. This effect is used to advantage in the ion source during chemical ionization (CI), but the pumping requirements for the analyzer relative to the ion source are even more stringent in the CI mode. Also, for good chemical ionization, the reagent gas pressure must be stable and controllable. Ion molecule collisions in the analyzer degrade resolution and ion transmission. (4) As the pressure rises, regulation of the electron current through the ion source becomes more difficult. (5) High background pressures are due to the presence of compounds in the mass spectrometer. These give interfering mass spectra, making interpretation difficult in some cases. (6) Contamination of ion source components, slits or rods, and multipliers increases with increasing pressure. This necessitates down time for cleaning. Modification to Pumping System Since the connection of a gas chromatograph to the source region of the mass spectrometer places an increased burden on the mass spectrometer pumping system, it was necessary to modify the existing vacuum system. See Figure 8. To counteract the high pressure of helium gas constantly being introduced into the source, differential pumping was incorporated into the system. Differential pumping is the independent pumping of two regions of a vacuum system that are separated by a restriction. A restriction was already present between the source 26 SEPTUM n I:én'I'HERMOCOUPLE INTERFACE PENNING ION SOURCE ANALYZER TUBE ..................... 10 6 INLET FOR _ VENTINC ,3 V 7 _ 4 ’3 av THERMOCOUPLES I PUMP ..... 2 THERMOCOUPLE FV- P UMP Figure 8: Vacuum system of the CVC M82000 time-of—flight mass spectrometer. The shaded areas indicate the modifications that were made to accomodate the GC inlet system. The vacuum valves indicated are (1) valve from analyzer tube to high vacuum pump; (2) backing valve for high vacuum pump; (3) valve for roughing analyzer tube; (4) valve for evacuating inlet system; (5) needle valve from batch inlet to ion source; (6) valve from ion source to high vacuum; (7) two position backing/roughing valve; (8) GC/MS interface probe vacuum lock; (9) valve for roughing vacuum lock; (10) valve for venting system. 27 and the drift region of the flight tube due to the design of the flight tube liner. Thus, differential pumping could be incorporated between the source and analyzer regions by installing a second high vacuum pump. The new pump, a large diffusion pump, was suspended beneath the source. The pump used is an Edward's Cryo-cooled Diffstak MK2 diffusion pump, model 160. The Diffstak system comprises a 3-stage oil diffusion pump, a high vacuum isolation butterfly valve that is pneumatically actuated and a cryo-cooled baffle. The cooling baffle incorporates a disc, positioned below the butterfly valve, that is cooled by cold ethanol. The cryo-cooled surfaces are attached to a capper probe and cooled by conduction from a reservoir of cold ethanol. A vacuum around the ethanol reservoir is common with the vacuum at the inlet to the diffusion pump to provide thermal insulation for the reservoir. A Neslab cold bath is used to refrigerate and recirculate the cold ethanol to the baffle. The weight of the diffusion pump was too great to allow it to hang freely without causing damage to the instrument. Thus, a support was constructed from an used computer frame. An Edward's backing/roughing valve, model BRVP, is utilized on the Diffstak diffusion pump which combines the functions of separate backing and roughing valves in a single, integral 3-port unit. The valve is pneumatically actuated and is double acting so that either the backing port or the roughing port can be connected to the third common port that connects to the rotary pump without an overlapping connection, i.e., all three ports are isolated from each other momentarily during the change-over period. The mechanical pump which backs this diffusion pump is Edward's Model 18. 28 TO control the heater and pneumatically operated high vacuum butterfly valve Of the Diffstak diffusion pump and the pneumatically Operated backing/roughing valve, a circuit was built as shown in Figure 9. Two relays located in a Granville-Phillips gauge controller are employed as safety features for the butterfly valve and the backing/roughing valve. If the pressure in the source Of the mass spectrometer, as measured by a Granville-Phillips ionization gauge added near the mouth Of the Diffstak, were tO rise above a certain preset value, the high vacuum butterfly valve would automatically change to the closed position and the backing/roughing valve, if not already in the backing position, would change to the backing position. These pneumatically Operated valves are driven by a compressed gas at a pressure between 60-90 psi as represented in the schematic in Figure 10. Three-way and two-way solenoid valves were used in the system. The two-way valves are either normally Opened or normally closed; when the solenoid valves are energized, the position changes. Similarly, in the three-way valves, two ports are normally connected. Then, as the solenoid valve is energized, the position changes sO that one Of the ports is connected tO the third. The added pumping Of the Diffstak diffusion pump produced a suitable vacuum in the system even with helium carrier gas flowing through the GC capillary column into the source. Table 1 shows the typical pressure Of the system at several stages Of development. The Open-split interface and the direct connection will be described in the following chapter. 29 exam zonnameaa — «.mfi O>Ha> coco oku . Iuo>o m ma non Sfifisnzfo .92 o easemeeam emcee .O.z sacs; H . zsnfi A.nn AA A season assuage Awauxummv Anusomv .o.z .u.z mm nosom can: - IIIOII 0 sous: UZH¥U (3 ’lcas Lines E: Fused Silica Capillary Restrictor Line Heated Probe \ Figure 13: Schematic showing Open-split interface in CC floor. 45 I Electrical Connection for Heater Assembly Stainless Steel Heat Insulating Stainless Steel Shaft\. Handle Stainless Steel Capillary Tube / Heater Assembly \ F \ Glass Insulating Tube Probe Tip Union \ ‘-\ Kalrez Ferrule Fused Silica / Capillary Tube Figure 14: GC/MS interface probe. Quartz Insulating Tube 46 lock seal. A Vacumetrics temperature controller number 50753 is used for controlling and measuring the temperature Of the GC/MS interface probe. TO assemble the entire system, a 0.05 mm ID restrictor tube was fished through the interface probe. With approximately 3/4 inch Of the restrictor line extending past the probe tip, the nut was tightened down over a 0.4 mm ID Kalrez ferrule tO provide a vacuum seal at the probe tip. In order tO avoid crushing the vitreous tube, the nut was tightened down just enough so that the restrictor line did not move when gently tugged. Next, the probe was inserted through the vacuum lock until the tip was at the edge Of the ionizing region and then the outside end Of the restrictor line was plugged by squeezing a piece Of Parafilm over its end so that the system could be checked for vacuum leaks. The gas chromatograph was elevated over the platform and the restrictor line cut to the prOper length so that its end would be inside the interface when the G0 was lowered into place. A nut and ferrule at the bottom Of the interface was tightened tO eliminate leaks. The GC capillary column was inserted through the Opposite end Of the interface and fed in until it touched the restrictor line. It was pulled back 1 mm and a nut and ferrule tightened. See Figure 14. 47 At this point, it was necessary tO find a way tO heat the section Of the interface that extended out Of the bottom Of the GC oven tO the heated probe. Several methods were tried before a satisfactory one was found. First, the entire area was wrapped with aluminum foil tO provide a gOOd heat conductive mass around the interface. This was all wrapped with a heating tape in order to keep the area hot. However, with this method, even fairly volatile compounds did not make it through to the mass spectrometer. This method probably allowed cold spots. The next method Of heating this area that was attempted was tO make a cylindrical wire mesh housing for the interface area that could be wrapped with a heating tape. On the outside Of the heating tape, layers Of insulative glass wool were applied and then covered with aluminum foil. Initially, the volume inside the wire cage heated nicely, but within hours, the fiberglass insulation Of the heating tape had become damaged because the heat could not dissipate evenly. A method that finally did succeed consisted Of a cylindrically shaped BriskHeat heating sleeve. Glass wool was loosely applied to the Open ends tO minimize convective heat loss. However a disadvantage to the heating sleeve was that, being a continuous tube, it was difficult tO tighten the nuts at the end Of the Open-split interface. The sleeve was the exact length Of the separation between instrument modules, 30 in order tO work in the area inside Of the sleeve, it had tO be compressed like a bellow. With both ends Of the restrictor line fastened taut, working in this area Often caused the restrictor line tO snap. 48. In order to eliminate this difficulty with the Open-split interface, a direct connection to the ion source was attempted. TO accomplish this, a vitreous capillary tube was connected tO the heating probe and the Opposite end fished through the hole in the oven floor directly into the oven. The GC column was connected tO the vitreous capillary by means Of a zero dead volume connector as seen in Figure 15. The same heating sleeve was used in this case. Since the end going to the oven was not fastened rigidly, the capillary did not easily snap when disturbed. Results and Discussion A data system is currently being prepared for the CVC time-Of-flight mass spectrometer. However, because Of several electronic and software problems, the work has faced numerous set-backs. Due tO time restraints, it was necessary to use an alternative data collection device for this work. Thus, it must be pointed out that the data were collected for this part Of the project with a strip chart recorder. In order to monitor the effluent Of the gas chromatograph, selected ion monitoring was used. By using a mass selector potentiometer on the mass spectrometer, a time delay is set sO that only the ions Of a certain m/z value (or more correctly, a certain flight time) are recorded. 49 : CC Column Zero Deada/llllllr Volume Connecter (3 CC Oven Floor Fused Silica Capillary Line Heated Probe .__‘~.1 I Figure 15: Schematic of the direct-connection interface showing the capillary line running through the bore of the Open-split interface. 50 The Open-split method Of interface initially used, was discarded because Of the physical assembly problems that caused the capillary restrictor line tO continually snap. By directly connecting the GC column tO this line, the problem was bypassed and, as an added advantage, the sensitivity was increased. Figure 16 represents some data that demonstrate the increased sensitivity. Although one might expect that the direct connection Of the column tO the ion source would give shorter retention times because the high vacuum Of the mass spectrometer is acting on the GC effluent, the Opposite was found to be true in this case. TO explain this one must consider the method Of connection. The GC column with a 0.32 mm ID is joined tO a vitreous capillary column whose diameter is only 0.11 mm ID. This restriction causes the flow tO be reduced. Thus, even though the flow measured at the GC column exit was the same in both cases, the flow tO the source was reduced in the direct connection interface. A hydrocarbon mixture was separated and detected with the GC/TOFMS system using a 30 meter, OV-101 capillary column. The following chromatographic conditions were used: injection port, 250°C; interface probe, 300°C; heating sleeve on interface 300°C; helium carrier gas flow (measured at column exit), 3.3ml/min; temperature programming: initial temperature, 55°C; program, 3°C/min tO 265°C. Figure 17 shows the selected ion current profile for the hydrocarbon mixture containing decane ( ), undecane ( ), dodecane ( ), tetradecane (C14), C10 C11 C12 hexadecane (C16), octadecane (C18), eicosane (020), tetracosane (C24), and hexacosane (026). 51 I I 1.2 1.7 2.2 Time (min) A 4.3 4.7 5.1 Time (min) B Figure 16: Sensitivity comparison Of (A) Open-split interface and (B) direct connection interface. In each case 2 micrograms of acetophenone were injected. The chromatographic conditions used were: injection port, 280°C; isothermal, 130°C; interface, 250°C. A 30 meter, 0V-101 capillary column was used. Ion currents at m/z 105 were monitored. 52 .vouOumeos one: hm ex! on oueouusu new .vOns an: oasaoo stanza: 878 £32. 8 < 8.8..” .3882: “350% us come“ on comm .Eauuoua ousumuanOu "voomu .uuoo :OmuOOmam "one: no»: n=OmumchO Oasmoquusaouno one .vOuOOuov one anon on non vaguenum an: emu .noeaxuaia vauoofion mo ousuxma Osman: fiasco an no summon; ucouuso :Om vouOOHom us~ ouswmh . A:_EV cE_& om ca . om ON 53 As can be seen from Figure 17, as the molecular weight becomes higher, the peaks are less intense and less resolved. This indicates that a cold spot is present. In the time-Of-flight mass spectrometer, the source is not heated; thus, one explanation for these results is that the cold spot is after the heated portion Of the interface probe. There is about a 3 or 4 cm extension Of the fused silica capillary line that extends past the heated area Of the probe. This consists Of a quartz insulating tube and about 1 cm extension Of the capillary line. This same problem is evident in a series Of methyl esters Of fatty acids. A standard mixture containing 1 microgram per microliter Of caprylic acid methyl ester (C8), capric acid methyl ester (Clo), lauric acid methyl ester (clz), mysteric acid methyl ester (014), and palmitic acid methyl ester (016). Figure 18 shows the selected ion current profile. for this mixture. The m/z value that was monitored here was mass 74, which is an abundant fragment due tO a McLafferty rearrangement present in all methyl esters. The conditions for this analysis were: injection port 250°C; interface probe, 300°C; heating sleeve on interface, 300°C; temperature programming: initial temperature, 70°C; program 3°C/min tO 265°C. The problem becomes even more severe when higher molecular weight methyl esters Of fatty acids are analyzed. Figure 19 contains the selected ion current profile Of another mixture containing palmitic acid methyl ester(016), stearic acid methyl ester (018), arachidic acid methyl ester (020), and behenic acid methyl ester (C22). 54 ----.—-- ow. .c..——. .vouOumaoa one: on a\a an nuaouuso com .mon: ans aesfioo hunaamono “cause .noooa on < .o.oom .ouuunuuen "cae\o.m no o.ne~ on ooos .amuwoum ousuouomaou moocmu .uuoo acauOOMaw "ouo3.poms n:OmuvaOO Omnonuwouoaouso use .Ousuxma omen house no nouns Hugues chansons «O mammoua ucouuoo aOm vouooaom "mg ouswmm bewEV OEPE .——. -———.—-_ : I. W_ : : .a———. .- l' . —. .4... __.___ 55 .vouOumeoa one: on w\a um nuaouuso OOH .voos no: casfioo munaamono moHI>c .uouoa am < .oocom .oommuou=M mama\0on an Damon Ou cocoa .amuwouo unaunMOABOu “Deana .uuoo aOfiuOoflOm "one: puns maOmuwocoo Owsomquumsouno one .ousuxma Oman haunw mo nouns amnuoa composua mo sawmoum uaouuso aOm nauseaom "mfi ouswwm Acwev oeflh mm ON mH OH 56 An on-cOlumn injector was used in order tO find the limit Of detection. Figure 20 shows results Obtained for methyl stearate. Nanogram quantities can be detected with a signal tO noise ratio better than two. Similar quantities Of acetophenone and naphthalene gave similar results. This work can be expanded upon to improve the interface Of the GC/TOFMS system. Viewports can be installed in a flange near the source sO that it would be possible tO visually line up the exit end Of the capillary restrictor line with the ionization region. This could increase the sensitivity by having more analyte ions interact with the ionizing electron beam. Also, the insulating probe tip could be removed tO reduce the unheated distance Of restrictor line and allow molecules Of higher molecular weight to reach the electron beam. Replacing the heating sleeve around the interface system with one that can easily be installed or removed while the interface is in place would eliminate many difficulties. The clyindrical-shaped heating sleeve that is presently being utilized must be installed before the GC is lowered into place. Using one that is wrapped, then laced, zippered, or sealed with Velcro, would be an improvement. 57 4 6 8 10 12 Time (min) A 4 6 8 10 12 Time (min) B Figure 20: Selected ion current profiles showing detection limit of methyl stearate using on-column injection. (A) 45ng and (B) 9mg methyl stearate. GC conditions used were: injection port, 250°C; temperature program, 200°C to 265°C at 3°C/min; interface, 300°C. A 30 meter, OV-lOl capillary column was used. Ion currents at m/z 74 were monitored. 58 There are advantages tO the Open-split interface that were mentioned earlier; thus, this method should not be abandoned. Arrendale, et a1, [24] have had success with this interface method. They reported that when the GC column flow was equal to or less than the interface tube flow, the sensitivity was equivalent tO that obtainable with direct connection Of the GC column tO the mass spectrometer. Their Open-split interface was placed inside the oven and was constructed Of glass in order tO allow precise visual adjustments Of the GC column and the restrictor line. For maximum chromatographic efficiency, without distortion Of the vacuum system, the restrictor line was inserted inside the GC column. The Open-split interface Of the GC/TOFMS system can be installed inside the oven. Further experiments may show improvements in sensitivity. IV. Comparison Of Fragmentation Patterns Introduction In this study, both the time~Of-f1ight and the magnetic sector mass spectrometers produce ions by the same method, namely electron impact. Thus, differences in fragmentation should not exist due tO the ionization process. The 70eV electron beam used for ionization will cause the sample molecules to decompose in the same manner in either instrument. If differences do occur in the spectra produced by the two different types Of mass spectrometers, they will be due tO the inherent differences in the twO instruments. A primary factor that could affect the fragmentation patterns is source temperature. Typically, the source Of a time-Of-flight mass spectrometer is not heated while that Of the magnetic sector instrument is. Adding heat would increase the internal energy Of the ions resulting in a larger initial energy spread. In TOFMS this causes a decrease in resolution. Heating the source, therefore, results in poorer resolution. On the other hand, ion sources in magnetic sector instruments are typically heated tO 300°C in order tO prevent the sample from condensing on the source components. The thermal energy available from the heated source could contribute to the total internal energy Of the ions causing more fragmentation to take place and give rise tO decreased intensity Of the molecular ion peaks and enhanced intensity Of some fragment ion peaks. 59 60 Another, probably less important, factor that may affect the ion intensities is the use Of a magnetic field on the multiplier Of the TOF mass spectrometer. The Operation Of the magnetic electron multiplier requires a magnetic field provided by an array Of bar magnets mounted externally on the vacuum housing. The field might extend over part Of the drift tube as well as over the entire multiplier structure and act throughout a short portion Of the Otherwise field-free drift tube in addition tO the multiplier itself. Thus, the magnetic electron multiplier might introduce a geometric mass discrimination [31] caused by the deflection Of ions by the magnetic field as shown in Figure 21. If ions are deflected, the multiplier entrance aperture intercepts a mass-dependent fraction Of the incoming ions and prevents their detection. If the field is not present, ions striking the detector will cover a circular area. However, because some ions might be intercepted by the aperture, the shape, location, and area or the geometry, Of the transmitted bunch could be changed by the magnetic field Of the multiplier. Generally, geometric mass discrimination is negligible in most commercial TOF instruments when they are Operated at the normal accelerating potential Of approximately 3000V. Hunt, et a1 [31] reported that during normal Operation, this geometric mass discrimination is negligible for m/z values greater than or equal tO eight. However, for low mass ions at small accelerating voltages, this discrimination can be quite significant; it can even prevent detection, because the amount that an ion bunch is deflected increases as its kinetic energy, as well as its mass tO charge ratio, decreases. 61 *— Multiplier Cathode I Ion Bunch _I L_ Multiplier Entrance Aperture \ ‘7 I Figure 21: Possible effect Of the magnetic field Of the magnetic elctron multiplier. (A) Absence Of the field. (B) Deflection Of ion bunch in the presence Of the field. 62 Experimental Fragmentation patterns for selected compounds were compared by Observing the spectra produced by electron impact in time-Of-flight mass spectra and magnetic sector- mass spectra. Seventy electron-volt ionizing electrons were used. The CVC 2000 and the LKB 2091 mass spectrometers were used. The LKB 2091 is a gas chromatograph/mass spectrometer system. The system consists Of three main parts: a gas chromatograph, a two stage jet separator, and a single focusing magnetic sector mass spectrometer. It is equipped with a 90° sector 140 mm radius magnetic analyzer and has a mass range Of 680 daltons at an accelerating voltage Of 3500V. The samples were introduced via the GC inlet. Due tO the absence Of a working data system on the CVC 2000 time-Of-flight mass spectrometer, spectra were collected in the analog form on a strip chart recorder. Mass peak intensity pairs were manually acquired and used tO produce bar graph mass spectra. 63 Results and Discussion Figure 22 contains a comparison Of mass spectra Of acetophenone Obtained on the TOF mess spectrometer and the magnetic sector mass spectrometer. Both spectra contain the same major peaks due tO fragmentation Of the molecule, but upon examination, differences in peak intensities are Observed. The percentages Of the total ion intensity (TII) for selected peaks are shown in Table 3. The peak at m/z 120 exemplifies a significant difference in the percent TII between the two instruments. This difference illustrates that some feature between the two mass spectrometers is affecting the fragmentation pattern. An examination Of the relative intensities Of the peaks at m/z 51, 77, and 120, indicate that the ratio Of these peaks is approximately the same in both spectra. Thus, it appears that the intensity Of the peak at m/z 105 changes. The fact that the ratio Of peaks at m/z 51 and 77 tO the base peak at m/z 105 is decreased in the magnetic instrument suggest that the process in which m/z 120 goes tO m/z 105 is enhanced in relation to the Other fragmentation processes. A comparison Of spectra Of 1,5-cyclooctadiene taken on both instruments is shown in Figure 23. As in the case Of acetOphenone, the same basic spectral pattern is given by either instrument. The molecular ion peak is twice as intense in the TOF spectrum. The remaining peaks in the mass spectrum are virtually superimposable. Again, it is evident that the fragmentation pattern is altered. 64 Figure 22: Mass spectra Of acetophenone. (A) Spectrun produced by a time-Of-fl ight mass spectrometer. (B) Spectrum produced by a magnetic sector mass spectrometer. 65 Tosaseeoa.nar acute banana 30-Nov-IB loan 0 S I! for 9108- 2230. TS..- 0:00 1003- 0.9. 30.9 ca 103 a I3 C30 77 >351 120 51 J an d l. I L 114 I" l IUIUIUIUT II II TIII IIUI IIIUIIITUJ; IIIIIIIIUIIU III! III! III. TUIT Ill! lltl III. IITT 40 43 80 BB 00 BB 70 .0 05 .0 98 100105 110 113130 123 180 A M3? 0. “A AGITW sue/UL 633333335scun O 147 T1: for plot- 0.5.0. Tamo- 4:94 :OOI- 38944. 33 103 ‘ Zn 5‘ aao II I] ILII llh II I. L La II "II I"! I"! I II "ITI’IIIIIIIHIIIII IITI 11W "1! T111 II" "I ITIU I'U' T 40 48 80 55 00 05 70 75 00 IS 90 95 100 105 110 115 120 125 130 B Figure 22 66 Table 3: Comparison of percent total ion intensity (TII) of some peaks in mass spectra of selected compounds acquired on the time-of—flight and magnetic sector mass spectrometers. Z TII Compound gig ‘_I‘_(_)_F_ M_a_gnetic Sector Acetophenone 120 14.1 9.8 105 30.9 34.1 77 26.8 25.1 51 13.3 10.5 1,5-Cyclooctadiene 108 3.4 1.6 93 3.0 3.2 80 10.2 9.6 67 21.9 22.6 54 28.8 29.1 Hexachloro-1,3-butadiene 260 6.0 5.2 225 13.3 13.2 190 5.4 5.1 153 1.7 1.6 141 3.5 3.1 118 5.0 3.1 94 2.2 1.2 83 3.4 1.6 71 2.0 1.0 67 Figure 23: Mass spectra of 1,5-cyclooctadiene. (A) Spectrum produced by a time-of-fl ight mass spectrometer. (B) Spectrun produced by a magnetic sector instrunent. 68 TOIaBISOB.M8F 1.5-6 closet-atone 84-Nov-05 Seln O 1 II for 910%. 2302. Tan.- 0:00 100‘- 000. 28.8 54 d 67 Z 52 J I .i J 80 ‘ 93 100 L IIJ II I J A A J IrIIIIIIIIIITilIIIIrIjII IIIIr TIIIIIIIrITII IjTjjl'r'LtII 40 45 ,BO 55 .0 65 7O 75 BO 85 .0 .5 100 105 110 102100391.N8F .IUL 1.5-CYCL000TADIENE 0-00%-.5 ‘CIfl O 1. 71! 90P 010%- 39.300. TQICI 2:31 100‘. .8400. 29.2 5‘ 57 q q 2:52 J fi .0 'I 93 I Tilt I!5 LA ‘ TIJrII Ilfi‘All Irt ‘ IiI? rifi IIIILI—III Tn TrrT I r 111! I I I I I' I I I 40 43 30 55 00 05 70 75 00 BS 90 05 100 105 110 B Figure 23 69 Figure 24 illustrates the same type of comparison for hexachloro-l,3-butadiene. In this case, the differences in the relative intensities are more evident in lower mass ions. The percent total ion intensities of some of the major peaks are also given in Table 3 for these compounds. In order to see how much variation in relative ion abundances is obtained on an instrument due to factors such as noise and ion statistics, several spectra of the same compound were taken. Table 4 contains the results of an experiment in which five spectra of acetophenone were acquired on the time-of-flight mass spectrometer. It shows the variation that occurs in peak intensities in the group of spectra. Examining the value of the percent total ion intensity for m/z 120, it can be seen that 952 of the time the percent total ion intensity will have a value between 21.9 and 23.1 (22.5 1. 2s). The results in Table 5 are from the same experiment performed on the magnetic sector instrument. Here, 952 of the time, the intensity for m/z 120 will be between 11.12 and 12.71 of the total ion intensity. Note that each of the tables compares values for intensity from data acquired on a single instrument. Next, having compared the variations within each group, one can compare the differences between the two groups. Still referring to m/z 120, it is obvious that, even with error taken into account, these values will not overlap. In other words, the variation between these groups is larger than the variation within the groups. Thus the spectra are truly different. The differences between spectra acquired on the TOP and magnetic sector mass spectrometers that were discussed earlier are not merely random, but distinct differences. .'70 Figure 24: Mass spectra of hexachloro-l,3—butadiene. (A) Spectrun produced by a time-of—fl ight mass spectrometer. (B) Spectrun produced by a magnetic sector instrunent. T:1315501.M8F Hoxachloro-1.3-butadionc 71 Figure 24 85-Nov-55 Scan s 1 T1! for plot- 4505. 73.0- 0:00 :003- 595. 13.3 T c1 c1 ‘ . . .4 c12c-c-c-cc12 ' 115 I I: 141 - 7: 94 J l 10. I l ‘53 .1 I I 1 1 ll .. III JL 1 l -n nu 1 "‘T"*‘lfi""rl"fi“'l""lf""’ ""I""l"'ffI""1‘""' ""15 I: 40 5O 50 7O 50 90 100 110 120 130 140 150 150 51 13.3 J 225 J .4 250 q 190 q ..-: 4.--- .-- -‘ .... r--.. --_,, -- ‘1‘-..,- --..-I-..." 1., --.- 150 170 150 190 200 210 220 250 240 250 250 370 250 A 115110557.NIF HIXACHLOfl0-1.8-OUTADIINI ID-Nov-IB loan 0 154 TI: for 919:— 1208530. 7100- 5:50 8005- 1.0785. Cockoround Icahn: 1.00 I 145 , 16 d ‘ 115 141 d 1 71 .3 3‘ J [I 3" -vv‘ Vi, _“r' 7' l;‘; v‘ slglr' 'Ame- ll; Ill, Al.-- I. lj‘l ‘ l I I l l""l""l""l""I""I"" 40 50 70 so :00 :10 120 150 :40 150 :50 2:11 la a 225 . ' 190 250 .,:.::: l - - ..- . - - 'II .‘5-. - .7- . . - .7. "l : I I 57“. . . I‘ I ' 150 £70 150 150 300 810 220 sec ' 840 850 8‘0 870 150 B 72 Table 4: Variation in relative intensities of major peaks in five consecutively recorded mass spectra of acetophenone from the time-of-flight mass spectrometer. m/z Relative Intensity (8) Percent Total Ion Intensity (s) n = 5 n I 5 77 82 (2) 31.5 (0.4) 105 100 - 38.3 (0.5) 120 59 (2) 22.5 (0.3) . 73 Table 5: Variation in relative intensities of major peaks in five consecutively acquired mass spectra of acetOphenone from the magnetic sector mass spectrometer. m/z Relative Intensity (8) Percent Total Ion Intensity (a) n - 5 n I 5 77 76 (2) 31.7 (0.4) 105 100 - 41.9 (0.8) 120 29 (1) 11.9 (0.4) 74 Stated below, are three possible reasons for the differences in the fragmentation patterns acquired from the two instruments. The first possibility is that the heated source in the magnetic sector instrument provides additional internal energy to the molecular ions. This would lead to a higher percentage of the molecular ions having adequate energy to fragment. The amount of energy actually being added to the molecular ions in a 300°C source was calculated to be 0.07eV. Considering that the bombardment of analyte molecules with 70eV ionizing electrons gives rise to molecular ions having a internal energy in the range of zero to twenty electron volts, with the average being approximately a few electron volts, it seems unlikely that the addition of 0.07eV would have a significant effect. A second possible explanation is that any metastable decompositions that take place in the field-free drift region of the TOP mass spectrometer, will have the same arrival time as the parent ions. However, this is not true for a magnetic sector instrument. Ions decomposing after leaving the source, but before entering the magnetic field, will be detected at m/z values dependent on the masses of the precursor ions and on the masses of the product ions. This behavior generally gives rise to a very broad peak that may not show up in the bar graph spectrum. Thus, if many metastable decompositions are occurring, this could explain why certain peaks will be more intense in TOF than in magnetic sector instruments. However, metastable peaks in TOF usually have broad bases since the decomposition will cause an energy spread. This would be evident in the analog TOF spectra; however, no such peaks were observed. ‘75 Finally, another possibility could be that the hot metal of the heated source is catalyzing thermal degradation of the analyte molecules and molecular ions. To investigate this possibility, a simple experiment was performed. Mass spectra of selected compounds were run in the magnetic sector instrument with the source cooled down to 180°C and the molecular separator and the injector of the GC cooled down to SS’C. Under these conditions, the mass spectra obtained were much the same as those acquired on the TOF mass spectrometer. Compare Figure 25A with Figure 23A and Figure 26A with Figure 22A. This was an interesting result. But, it does not answer the question whether this occurs because there is an addition to the internal energy of the molecular ion in the hot source that is sufficient to assist fragmentation, or because thermal decompositions are taking place in the hot source. To determine which of these mechanisms may be occuring, the separator was heated to 300°C. This would provide the additional thermal energy without providing a hot metal surface in the source. Under these conditions, the spectra obtained were much like those acquired on the TOF mass spectrometer as shown in Figure 258 and 268. Apparently, the fragmentation pattern is a function of a hot catalytic metal surface rather than a purely thermal effect, and is independent of the type of 10888 spectrometer . 76 Figure 25: Mass spectra of 1,5-cyclooctadiene the magnetic sector mass spectrometer. separator were cool. hot. acquired on (A) Both source and (B) Source was cool and separator was 77 111312IH591HJ43F’ 1.13-{rVCLJNDTIlIIIfliI SO-Doc-OS Scon O 75 TX! for 915%- 020732. 71-.— 4:01 1003- 180344. £54 d Cool Source .. .7 Cool Separator 00' d I I I I I I I .J a [JJ _La_. I J L l I I j A _ l A L IITI TIITITI IIII rfI I IrIIIII I ITIIIIIIIIIIIIIII 40 45 50 BB 00 08 70 75 00 08 100 $03 110 A 110120554.N5F 1.8-CYOL0007ADIENE ta-Dec-ls Scan 0 50 T1! for plot- 441550. Tana— 3:08 3005- 10514.. 54 4 Cool Source 57 Hot Separator q 50 d I d - .3 105 d I I.I1 IIIIIIIIIIIW {IIII[II!I%I IIIIIIIiITI ITTII [IIITrIIT IIIIII'IIIIIIII I 40 43 30 85 00 08 70 75 00 05 20 J5 100 105 3110 Figure 25 78 Figure 26: Mass spectra of acetophenone acquired on the magnetic sector mass spectrometer. (A) Both source and separator were cool. (B) Source was cool and separator was hot. 79 112129'9&.M8P ACETOPHENONE SO-Do;-B? Scan 0_813 711 f0? 0100- 13.00. 7:..- 10:47 100'. 37.4. 10!! Cunl Source Cool Separator 130 I IIIII IIIIIIIIIITIIIIIIIIIIII I II "IIIIIIIIIIIIIIIIIIIITI IIII III IIII IIII IIII 40 45 30 35 00 03 70 73 00 05 .0 08 100 105 110 $38 130 I” 180 A 110120599.MSP ACETOPHENONE 11-006-03 Icon 0 83. TX! for Plat- 48.220. 71..— 7:05 800’. 1308.0. . Cool Source :05 Hot Separator 880 51. .II! I; 44 1- I l k ‘1 A. A ; A III IIII I TIII III IIII IIII IIIIIIIIIIIIII IIII IIII “I "II IIII IIII III 40 A! 30 35 00 B! 70 73 00. 05 00 95 100 105 110 115 180 135 £30 B Figure 26 80 In summary, the data show that some variation occurs in relative peak intensities in mass spectra acquired at different source temperatures. The temperature of the ion source affects the distribution of the total ion intensities in mass spectra. This variation appears to be independent of instrument type, but dependent upon the nature of the analyte. LIST OF REFERENCES 100 11. 12. l3. 14. 15. 16. 17. 18. 19. 20. List of References W. H. McFadden, Techniques of Combined Gas Chromatography/Mass Spectrometry, John Wiley and Sons, New York, 1973. E.C. Morning and M.G. Morning, Clin. Chem. 17, 802 (1971). J.K. Majer and R. Perry, Pure Appl. Chem. 24, 685 (1970). G. Vander Velde and J.F. Ryan, J. Chromatogr. Sci. 13, 322 (1957). C. Merritt, Jr., Appl. Spectrosc. Rev. 3, 263 (1970). R.A. Bites and K. Biemann, Anal. Chem. 42, 855 (1970). J.F. Holland et al., Anal. Chem. 55, 997A (1983). C.C. Sweeley, W.H. Elliott, I. Fries, R. Ryhage, Anal. Chem. 38, 1549 (1966). 8.8. Middleditch and D.M. Desiderio, Anal. Chem. 45, 806 (1973). K.A. Lincoln, Dyn. Mass Spectrom. 6, (1981). Instruction Manual for the Time-of-Flight Mass Spectrometer Type M82000, CVC Products, Inc., Rochester, N.Y. A.E. Cameron and D.F. Eggers, Jr., Rev. Sci. Instr. 19, 605 (1948). um. Wolff and v.3. Stephens, Rev. Sci. Instr. 24, 616 (1953). H.S. Katzenstein and 8.8. Friedland, Rev. Sci. Instr. 26, 324 (1955). W.C. Wiley and 1.8. McLaren, Rev. Sci. Instr. 26, 1150 (1955). G.M. Message, Practical Aspects 2: Gas Chromatography/Mass Spectrometry, John Wiley & Sons, New York, 1984. A.N. Feedman, Anal. Chim. Acta 59, 19 (1972). A.A. Ebert, Jr., Anal. Chem. 33, 1865 (1961). J.C. Holmes and F.A. Morrell, Appl. Spectrosc. 11, 86 (1957). R.S. Gohlke, Anal. Chem. 31, 535 (1959). 81 21. 22. 23. 24. 25. 26. 270 28. 29. 30. 31. 82 R.R. Freeman, High Resolution Gas Chromatography, Hewlett- Packard, 1981. D. Henneberg, U. Henrichs, G. Shomberg, Chromatographia 8, 449 (1975). J.J. Manura, The Mass Spec Handbook 2: Service, Scientific Instrument Services, Inc., Pennington, N.J., 1983. R.F. Arrendale, R.F. Severson, O.T. Chortyk, Anal. Chem. 56, 1533 (1984). J.T. Watson and K. Biemann, Anal. Chem. 36, 1135 (1964). A. Copet and J. Evans, Org. Mass Spectrom. 3, 1457 (1970). R. Ryhage, Anal. Chem. 36, 759 (1964). J.T. Watson, Introduction £2_Mass Spectrometry, Raven Press, New York, 1985. D.C. Damoth, Dyn. Mass Spectrom. 1, 199 (1970). D.R. Black, R.A. Plath, R. Teranishi, J. of Chromatogr. Sci. 7, 284 (1969). W.W. Hunt, et a1., Rev. Sci. Instr. 39, 1793 (1968).