Anal. Chem. 1997, 69, 1715-1721
Time-of-Flight Mass Spectrometry with a Pulsed Glow Discharge Ionization Source Robert E. Steiner, Cris L. Lewis, and Fred L. King*
Department of Chemistry, West Virginia University, Morgantown, West Virginia 26506-6045
The pulsed glow discharge (GD) plasma source exhibits several characteristics that make it ideally suited for use with time-of-flight mass spectrometry (TOFMS). TOFMS uniquely affords the ability to monitor a narrow temporal window for a time-varying process such as ion formation in the pulsed glow discharge plasma. Pulsed GD-TOFMS exhibited distinct advantages for the direct determination of trace elements in solid state samples. Initially, the pulse-powered GD-TOFMS system used for these investigations exhibited poor resolution. In an effort to improve resolution, a slit was introduced to narrow the ion beam orthogonally entering the extraction region of the TOFMS. In an effort to determine optimal operating conditions, the influence of slit width on TOF performance was investigated. In the course of this study, the slit width was found to influence isotope ratio accuracy as well as resolution. A slit width of 1.0 mm was determined to provide the best compromise between resolution and isotope ratio accuracy. Pulsed GD-TOFMS affords improved sensitivity and selectivity because Penning ionization is enhanced during the time period immediately following the termination of the discharge power. Ions sampled by an extraction pulse applied after power termination also yield a mass spectrum that is free of contributions arising from electron-ionized interferences. This advantage arises because only ions generated via the Penning ionization mechanism persist after the termination of discharge power. Sampling in the “afterpeak” time regime eliminates the saturation of the detector arising from discharge support gas ion signal. Glow discharge devices are useful as sources of atomization and ionization in the mass spectrometric determination of trace elements.1-5 Glow discharge operation is based in part on cathodic sputtering. The electrical breakdown of the discharge support gas induced by the application of an electrical potential generates ionized discharge gas atoms. These ions accelerate into the sample cathode, where the conversion of their kinetic energy into lattice energy results in the release of sample atoms. This cathodic sputtering effect affords the direct atomization of any solid serving as, or contained in, the near surface region of the cathode.5 Whereas conducting samples can be sputtered and analyzed directly using the dc-powered glow discharge, noncon(1) Harrison, W. W.; Magee, C. W. Anal. Chem. 1974, 46, 461. (2) Colby, B. N.; Evans, C. A. Anal. Chem. 1974, 46, 1236. (3) De Gendt, S.; Van Grieken, R.; Hang, W.; Harrison, W. W. J. Anal. At. Spectrom. 1995, 10, 689. (4) King, F. L.; Harrison, W. W. Mass Spectrom. Rev. 1990, 9, 285. (5) King, F. L.; Teng, J.; Steiner, R. E. J. Mass Spectrom. 1995, 30, 1061. S0003-2700(96)01171-7 CCC: $14.00
© 1997 American Chemical Society
ducting samples must be mixed with a conducting matrix in order to sustain a discharge. Alternatively, nonconducting samples can be analyzed directly using an rf-powered glow discharge source.6 Even small-volume solution samples can be analyzed by evaporating the liquid sample onto a cathode and sputtering the resulting residue.7 Once atomized via the sputter process, sample atoms are subject to collisional excitation and ionization in the surrounding plasma. The independence of the excitation/ionization process from the atomization process in GD sources provides another inherent advantage, freedom from memory of the sample matrix. In techniques such as secondary ion mass spectrometry, where atomization and ionization occur concurrently, the matrix exerts significant influence on the efficiency of ionization for an element. Such matrix effects lead to widely varying elemental sensitivities in these techniques. The decoupling of atomization and excitation/ionization in the glow discharge substantially decreases such matrix effects associated with trace elemental analysis. It has been shown that GDMS relative sensitivity factors for steel, aluminum, and copper matrices do not differ appreciably.8 Researchers have gone to great lengths to facilitate the separation of these two steps for other elemental mass spectrometry techniques. Hieftje et al. recently reported the use of laser ablation inductively coupled plasma time-of-flight mass spectrometry to achieve direct solid sampling with decoupled atomization/ionization.9 Analytical glow discharge plasmas can operate in either a steady-state or pulsed power mode. Although both operation modes yield the desirable characteristics described above, pulsed power operation demonstrates additional advantages not afforded by steady-state operation. The pulsed power mode affords higher instantaneous power with no average current or average voltage increase relative to the steady-state sources. This is of analytical significance because higher instantaneous operation powers increase atomization, excitation, and ionization without inducing thermal degradation of the sample or discharge instability.10 The use of pulsed power plasmas also affords the opportunity to utilize time-gated ion signal acquisition. This improves analytical performance because it permits temporal discrimination against electron-ionized interferences and background species.11,12 These (6) Klingler, J. A.; Harrison, W. W. Proceedings of the 39th ASMS Conference on Mass Spectrometry and Allied Topics; Nashville, TN, May 19-24, 1991; ASMS: Sante Fe, NM, 1991; p 943. (7) Barshick, C. M.; Duckworth, D. C.; Smith, D. H. J. Am. Soc. Mass Spectrom. 1993, 4, 47. (8) Sanderson, N. E.; Hall, E.; Clark, J.; Charalambous, P.; Hall, D. Mikrochim. Acta 1987, 1, 275. (9) Mahoney, P. P.; Li, G.; Hieftje, G. M. J. Anal. At. Spectrom. 1996, 11, 401. (10) King, F. L.; Pan, C. Anal. Chem. 1993, 65, 3187. (11) Klingler, J. A.; Savickas, P. J.; Harrison, W. W. J. Am. Soc. Mass Spectrom. 1990, 1, 138.
Analytical Chemistry, Vol. 69, No. 9, May 1, 1997 1715
unique characteristics have led to the development of pulsed glow discharges as sources for spectrometric determinations.10-21 In general, pulsed glow discharges find use with a variety of spectrometric methods for trace elemental analysis, including atomic absorption,14,15 atomic emission,13,19 atomic fluorescence,21 and mass spectrometries.10,16 The majority of analytical glow discharge mass spectrometry is performed using quadrupole mass filters22,23 or double-focusing mass analyzers.24 Unfortunately, neither of these systems is well suited to the time-gated acquisition of signals. Recently, quadrupole ion trap25,26 and Fourier transform ion cyclotron resonance mass spectrometers27-29 have been adapted for use in GDMS as well. Time-of-flight mass spectrometry has also been used in conjunction with glow discharge ion sources.30 The TOFMS systems uniquely allow the acquisition of an entire mass spectrum with every extraction pulse sequence, enabling rapid, simultaneous spectral acquisition. This characteristic makes TOF especially well suited for isotope ratio measurements, because all analyzed ions are formed at the same time under the same ionization conditions.31 Quadrupoles and sectors may suffer from greater uncertainty in isotope measurements due to the sequential nature of their spectral acquisition. Significantly, time-of-flight mass spectrometry requires either a pulsed ionization source or the gated introduction of ions from a steady-state source. In the past, TOFMS instruments have used a number of pulsed sources, including secondary ion mass spectrometry (SIMS),32 matrix-assisted laser desorption/ionization (MALDI),33 and laser ablation.34 Hieftje et al. have investigated the adaptation of an ICP-TOFMS interface to facilitate the implementation of an rf-powered glow discharge source.30 The GD source used in that study was operated in a steady-state mode with ions gated into the orthogonally positioned flight tube by an applied extraction pulse. Spectra acquired using a steady-state GD source are often dominated by discharge gas and other electron-ionized background species. With a GD source operating (12) Klingler, J. A.; Barshick, C. M.; Harrison, W. W. Anal. Chem. 1991, 63, 2571. (13) Winchester, M. R.; Marcus, R. K. Anal. Chem. 1992, 64, 2067. (14) King, F. L.; Pan, C. Anal. Chem. 1993, 65, 735. (15) King, F. L.; Pan, C. Appl. Spectrosc. 1993, 47, 300. (16) Hang, W.; Walden, W. O.; Harrison, W. W. Anal. Chem. 1996, 68, 1148. (17) Klingler, J. A.; Harrison, W. W. Anal. Chem. 1991, 63, 2984. (18) Parker, M.; Marcus, R. K. Appl. Spectrosc. 1996, 50, 366. (19) Pan, C.; King, F. L. Appl. Spectrosc. 1993, 47, 2096. (20) Pan, C.; King, F. L. J. Am. Soc. Mass Spectrom. 1993, 41, 727. (21) Glick, M.; Smith, B. W.; Winefordner, J. D. Anal. Chem. 1990, 62, 157. (22) Harrison, W. W. J. Anal. At. Spectrom. 1988, 3, 867. (23) Harrison, W. W.; Hess, K. R.; Marcus, R. K.; King, F. L. Anal. Chem. 1986, 58, 341A. (24) Vassamillet, L. F. J. Anal. At. Spectrom. 1989, 4, 451. (25) McLuckey, S. A.; Glish, G. L.; Duckworth, D. C.; Marcus, R. K. Anal. Chem. 1992, 64, 1606. (26) Duckworth, D. C.; Barshick, C. M.; Smith, D. H.; McLuckey, S. A. Anal. Chem. 1994, 66, 92. (27) Shohet, J. L.; Phillips, W. L.; Lefkow, A. R. T.; Taylor, J. W.; Bonham, C.; Brenna, J. T. Plasma Chem. Plasma Process. 1989, 9, 207. (28) Barshick, C. M.; Eyler, J. R. J. Am. Soc. Mass Spectrom. 1992, 3, 122. (29) Watson, C. H.; Wronka, J.; Laukien, F. H.; Barshick, C. M.; Eyler, J. R. Anal. Chem. 1993, 65, 2801. (30) Myers, D. P.; Heintz, M. J.; Mahoney, P. P.; Li, G.; Hieftje, G. M. Appl. Spectrosc. 1994, 48, 1337. (31) Benninghoven, A.; Ru ¨ denauer, F. G.; Werner, H. W. Secondary Ion Mass Spectrometry; John Wiley & Sons: New York, 1987; Vol. 86. (32) Niehuis, T.; Heller, H. F.; Benninghoven, A. J. Vac. Sci. Technol. 1987, A5, 1243. (33) Brown, R. S.; Gilfrich, N. L. Anal. Chim. Acta 1991, 248, 541. (34) Huang, L. Q.; Conzemius, R. J.; Junk, G. A.; Houk, R. S. Anal. Chem. 1988, 60, 1490.
1716
Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
in the pulsed power mode, temporal discrimination can be employed to discriminate against such interferences.14 In this article, we report on investigations performed to ascertain the utility of both rf- and dc-powered pulsed glow discharges as ionization sources for time-of-flight mass spectrometry. Modifications to the instrument were implemented to improve both the resolving power and the isotope ratio measurement accuracy. These modifications were centered around the introduction of a slit positioned between the skimmer cone and the extraction region. Results for a preliminary assessment of the analytical performance of this system are discussed. EXPERIMENTAL SECTION The glow discharge chamber consists of a six-way high-vacuum cross (MDC, Hayward, CA) equipped with 70 mm conflat flanges.14 Gas inlet ports and chamber pressure monitor (Hasting Teledyne-Raydist, Hampton, VA), a probe inlet, and an ion exit orifice are attached to three of the cross ports. The remaining ports are fitted with blank flanges or viewports. Both disk and pin sample geometries are utilized in these studies. The disk sample, 5 mm in diameter and 5 mm in height, is prepared by either machining from a solid ingot reference material or pressing sample powder using a die fabricated in-house. Pin type samples, measuring 2 mm in diameter and 25 mm in length, are fabricated by machining a solid ingot reference material. A direct insertion probe with 5 mm of the sample exposed on its tip facilitates sample introduction into the vacuum chamber without breaking the vacuum.35 Direct current operating power is supplied by a dc power supply (Model OPS-3500, Kepco Inc., Flushing, NY). The pulsed mode of operation is controlled using a square-wave pulse generator fabricated in-house. This pulse generator allows adjustment of power level, pulse width, duty cycle, and gate delay position. Radio frequency power is supplied by a 13.56 MHz rf generator (Model RF 10-S, RF Plasma Products, Inc., Marlton, NJ) and coupled through an automatic matching network (Model AM-10, RF Plasma Products, Inc.). The internal pulsing mode of the rf power supply permits control of frequency, duty cycle, and power level for the applied rf power pulses. The initial trigger output from the pulse generator or rf supply is positioned at the glow discharge power onset. An internal adjustable delay is then set relative to the initial trigger. This delay triggers the application of an adjustable width, positive potential pulse to the repeller, injecting an ion packet into the orthogonally positioned flight tube. Adjustment of the temporal position of this injection pulse permits the mass spectrometric monitoring of ions formed during distinct time regions in the plasma power pulse cycle (Figure 1). The monitoring of power, trigger, and injection pulses is performed using a digital storage oscilloscope (Model 2232, Tektronix, Inc., Beaverton, OR). The time-of-flight mass spectrometer employed is a linear flight instrument (R.M. Jordan Co., Grass Valley, CA) oriented orthogonally to the ion beam from the GD source. This orientation allows pulsed injection of sample ions in the source region of the TOFMS (Figure 2). Operating conditions for the GD-TOFMS are presented in Table 1. The optimal operating conditions were determined experimentally. The signal output from a dualmicrochannel plate detector (Galileo Electro-optic Corp., Stur(35) King, F. L.; Harrison, W. W. Int. J. Mass Spectrom. Ion. Processes 1989, 89, 171.
Table 1. GD-TOFMS Experimental Operating Conditions
Figure 1. Diagram showing glow discharge square-wave drive pulse (5 ms on-time, 25% duty cycle), trigger output for GD drive pulse (positioned at power onset), delay output (adjustable for one pulse cycle), TOF injection pulse (adjustable width from 50 ns to 3.5 µs), and the resulting temporal ion signal profile.
Figure 2. Diagram of GD-TOFMS instrument (refer to Table 1 for component descriptions).
bridge, MA) is preamplified (MIT Combo-100 preamplifier, Boulder, CO) and stored by a PC using a multichannel scaler and control card (Model MCS-702, Comstock, Oak Ridge, TN) operated in an ion counting mode. The data are then imported into a spreadsheet (QuattroPro for Windows, Orem, UT) that generates mass spectral plots. RESULTS AND DISCUSSION The pulsed glow discharge plasma demonstrates great utility as an ionization source for time-of-flight mass spectrometry. Along with the standard advantages associated with the glow discharge ionization source and the TOFMS, operation in the pulsed mode enables temporal discrimination of interferences and spectral acquisition during the time regime in which the analyte signal is maximized. This arrangement is also well suited to the diagnostic investigation of time-dependent processes occurring in pulsed glow discharges.
pressure (Pa) ionization source first stage second stage plasma current (mA) pulse length (ms) duty cycle (%) sampling orifice diameter (mm) skimmer orifice diameter (mm) slit width (mm) TOF repeller pulse time (ns) flight path length (m) instrument potentials (see Figure 2) (V) skimmer, X2 repeller plate, A1 extractor grid, A2 accelerator grid flight tube liner deflection plates, X1 and X2 deflection plates, Y1 and Y2 detector grid plate, G
1.1 × 102 1.3 × 10-2 8.3 × 10-5 0.80 5.0 25 1.0 1.0 0.50-1.0 100 1.0 -345 -175 0 -300 -300 -200 to -300 -200 to -300 -1800
Previous studies in this laboratory examined the origins of differences in the temporal ion signal profiles for an analyte (Cu) and a discharge gas (Ar).11,12 Time-gated injection of sample ion packets into the flight tube facilitates the study of the different temporal regimes associated with these ion signal profiles. Figures 3-5 illustrate both the dynamic processes occurring within the plasma during the associated time regime and the respective spectrum for a 1:1:1 (mass/mass/mass) sample of Fe/ Cu/Ag. In Figures 3-5, the gas phase analyte species are represented by M. At the onset of the pulse power, the argon ion profile exhibits a sharp rise to a maximum referred to as the “prepeak.” During the prepeak time regime, plasma processes are dominated by the electron ionization of the discharge gas, giving rise to the observed maximum for the discharge gas ion signal profile.11,23 Signal contributions from analyte species are delayed approximately 1 ms after power initiation. This delay arises because sample sputtering and analyte diffusion into the negative glow region is not an instantaneous process.10 The spectrum in Figure 3, representative of the prepeak region, shows no signal contributions from the analyte species. As is expected, only the discharge gas species are present as 40Ar+, 40ArH+, and 40Ar + at m/z ) 40, 41, and 80, respectively. As soon as argon 2 ions form, they accelerate toward the sample cathode. Upon collision of the ions with the cathode, energy is transferred to the sample lattice. This energy transfer results in the sputtering of analyte atoms from the sample surface. The gas phase analyte atoms then diffuse through the cathode dark space located one mean free path from the sample cathode and into the negative glow region of the plasma, where they can undergo excitation and ionization via numerous pathways. The corresponding temporal region known as the “plateau” is characterized by steadystate plasma conditions for both the discharge gas and analyte species. The spectrum in Figure 4, representative of the plateau time regime, exhibits contributions from both analyte and discharge gas species. Upon termination of the discharge power, the analyte ion signal profile exhibits a sharp rise toward a maximum. This “afterpeak” maximum occurs approximately 2 ms after the applied power is terminated. The spectrum in Figure 5, representative of the afterpeak region, exhibits contributions Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
1717
Figure 3. Direct current GD-TOFMS spectrum of equal weight Fe/ Cu/Ag sample collected 0.5 ms after pulse power onset and the corresponding dynamic plasma processes occurring during the prepeak time regime.
of only the analyte species. Within this temporal region, the discharge gas ion signal decays rapidly, effectively decreasing potential interferences from discharge gas species. Time-resolved measurements of argon metastable atom populations have indicated that this decrease in the argon ion signal occurs because argon ions recombine with thermal electrons upon power termination, forming argon metastable atoms within the plasma.14 In the present investigations, the plasma was maintained at an argon pressure of 0.9 Torr. This pressure was reported by previous investigators as optimal for argon metastable formation in pulsed dc plasmas.15 An increase in analyte signal intensity is also observed during this time regime. Increased analyte signal arises because the enhanced population of metastable argon atoms reacts with analyte atoms still present in the gas phase to produce ground state argon atoms and analyte ions via Penning ionization.10 These two characteristics make the afterpeak temporal region best suited for analytical data acquisition. Previously, this group has reported the use of pulsed plasmas as ion sources for quadrupole mass spectrometry.10 Time-gated acquisition using a quadrupole system requires the use of a boxcar integrator to collect ion signals through a data gate. This data gate is temporally positioned in the GD pulse sequence using a digital delay generator. The scanning nature of quadrupole instruments permits the collection of only one data point for a 1718 Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
Figure 4. Direct current GD-TOFMS spectrum of equal weight Fe/ Cu/Ag sample collected 4.0 ms after pulse power onset and the corresponding dynamic plasma processes occurring during the pleateau time regime.
given GD pulse, making spectral acquisition a relatively time consuming process. This limitation may also introduce spectral error associated with temporal variations in plasma conditions occurring between the beginning and the end of spectral acquisition. Temporal selectivity is inherent in the operation of any TOFMS. The repeller injection pulse can be adjusted to select ions formed during distinct temporal regions within the GD power pulse cycle. The use of this injection pulse takes the place of the data gate provided by the boxcar integrator in the previous work employing a quadrupole MS system. Another advantageous characteristic of TOFMS is the acquisition of an entire mass spectrum for each ionization event. Unlike scanning instruments, such as the quadrupole, the simultaneous detection of ions decreases the acquisition time required for each mass spectrum at the same time that it decreases uncertainty arising from temporal variation in plasma operation. The orthogonal orientation of the flight tube with respect to the ion beam acts to increase resolution, often degraded by an initial kinetic energy difference between ions with equal m/z, because all ions in a given injection pulse have a kinetic energy of approximately zero along the TOF flight tube axis. However, this orientation does not correct for the spatial distribution of ions
Figure 6. Effect of slit width on resolution of the 63Cu+/65Cu+ isotopes for the NBS 1275 copper-nickel standard (error bars represent 3σ for five replicates).
Figure 5. Direct current GD-TOFMS spectrum of equal weight Fe/ Cu/Ag sample collected 7.0 ms after pulse power onset and the corresponding dynamic plasma processes occurring during the afterpeak time regime.
within the ion beam along the flight tube axis. A slit was added to the GD-TOFMS between the skimmer cone and extraction region to narrow this spatial distribution of ions passing into the TOF extraction region. A similar setup was demonstrated to improve the resolving power of an ICP-TOFMS30 and encouraged this attempt at improving resolution on the current system. The influence of slit width on the resolving power of the instrument was investigated for three slit widths and no slit. Figure 6 illustrates the resolution in the 63Cu+ and 65Cu+ region obtained for these variations in slit width. The numbers of data points for both the resolution and isotope ratio analyses were limited by the availability of suitable slits. It is apparent that the introduction and subsequent decrease in slit width effectively increases the resolving power of the instrument. Unfortunately, the slit was found to influence isotope ratio accuracy as well as resolution. A study was performed to ascertain the effect of slit width on the resulting isotope ratio for the 63Cu+/65Cu+ isotopes. Measurements were taken at four different times within the pulse sequence for four different slit widths. The resulting plot (Figure 7) reveals a number of trends. When a slit is not used, the measured isotope ratio is lower than the natural isotope ratio. This behavior is indicative of detector saturation for the more abundant 63Cu+ isotope. This occurs when the incident ions for a given isotope
Figure 7. Effect of slit width on 63Cu+/65Cu+ isotope ratios at 3 (2), 5 (b), 7 (1), and 8 (9) ms after pulse power onset for the NBS 1275 copper-nickel standard.
saturate the microchannel plate detector and effectively lower the signal intensity. The 65Cu+ isotope does not have as great an abundance as the 63Cu+ isotope and, therefore, does not suffer significantly from this phenomenon. A slit width of 1.0 mm provides isotope ratios that are closer to the natural ratio for 63Cu+/65Cu+. This slit width decreases the 63Cu+ signal sufficiently, alleviating the saturation for 63Cu+, but does not affect the 65Cu+ peak. As the slit width decreases further, the isotope ratio increases. This change arises as decreasing slit widths effectively narrow the ion beam entering the extraction region, decreasing the number of ions injected into the mass spectrometer. Eventually, the number of incident ions striking the detector is not great enough to trigger the detector; therefore, the resulting signal does not accurately represent the isotopic ion population. At a slit width of 0.75 mm, this type of discrimination is observed. The less abundant 65Cu+ isotopes are no longer present in significant numbers to trigger the detector for every ion present. The more abundant 63Cu+ isotope does not suffer from discrimination at this slit width, resulting in characteristic isotope ratio increases. As the slit is narrowed to 0.50 mm, we observe a decrease in the isotope ratio. We believe that this arises because the signal intensity is attenuated to the point where even the more abundant 63Cu+ isotope is being lost. It became clear that a compromise was needed between improved resolution and reduced isotope ratio accuracy. Figure 8 illustrates the dependence of these two measurements on the Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
1719
Figure 9. Direct current GD-TOFMS spectra of NBS 1241 aluminum alloy standard at (a) 4.0 ms after power onset (plateau) and (b) 7.0 ms after power onset (afterpeak).
Figure 8. Direct current GD-TOFMS spectra of NBS 1275 coppernickel standard utilizing slit widths of (a) 0.75 and (b) 1.0 mm and (c) no slit.
slit width. At a narrow slit width, 0.75 mm (a), baseline resolution is observed. However, the isotope ratio accuracy suffers from an insignificant number of 65Cu+ ions. When no slit is present (c), the spectrum suffers from degraded resolution and detector saturation, negatively affecting isotope ratio accuracy. When a slit width of 1.0 mm is utilized (b), the negative aspects of both extremes are minimized. However, it is not significant enough to have a negative effect on either the resolution or the isotope ratio. Currently, the implementation of Einzel lenses positioned between the ionization source and the extraction region is being considered to increase overall signal strength. A similar setup has been described previously.36 This lens system would act to focus the ion beam onto the slit, effectively narrowing the spatial distribution of the ion beam further. The ability of pulsed GD-TOFMS to discriminate against discharge gas and other spectral contaminants is of analytical significance and was the driving force behind this work. The spectrum of NBS 1241b aluminum alloy shown in Figure 9a was collected during the plateau time regime of the discharge pulse. The spectrum is characterized by the 27Al+ peak as well as the much more intense 40Ar+ signal. As the pulse gate is moved into the afterpeak region of the pulse sequence, a considerable increase in the analyte signal occurs at the same time that signal contributions from the discharge gas and other interferences (36) Myers, D. P.; Li, G.; Yang, P.; Hieftje, G. M. J. Am. Soc. Mass Spectrom. 1994, 5, 1008.
1720 Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
Figure 10. (a) Full-scale and (b) low-intensity region blowup of direct current GD-TOFMS spectra of NBS 1275 copper-nickel standard 8.0 ms after power onset (afterpeak), utilizing a slit width of 1.0 ms (100 000 sweeps).
approach zero (Figure 9b). Small contributions from oxygen are observed at m/z ) 16 and 32. Again, these interferent contributions decreased as the injection gate was moved farther into the
resulting from differences in mobilities between electrons and ions in the oscillating electric field. This bias allows the direct sputtering of nonconductive samples such as glasses and ceramics.37 Figure 11a represents a spectrum for stock silica glass procured from the university glass shop. This spectrum corresponds to data acquired through a temporal acquisition window located within the plateau region of the pulse cycle. Again, this spectrum is dominated by signal contributions from the discharge gas species 40Ar+ at m/z ) 40. A large signal contribution is also observed at m/z ) 32, arising from 16O2+, a large component of silica glass. Signals arising from the silicon isotopes 28Si+, 29Si+, and 30Si+ are observed; however, they exhibit poor isotope ratio accuracy and are partially overlapped by the peaks at m/z ) 28, 29, and 32 corresponding to 14N2+, 14N2H+, and 16O2+, respectively. Figure 11b represents a glass spectrum acquired during the afterpeak time regime. The signal resulting from the analysis of the glass sample was of sufficient intensity to allow total discrimination of electron-ionized species by positioning the temporal acquisition window farther into the afterpeak region. This discrimination was possible only with a substantial decrease in signal intensity. Peaks arising from the silicon isotopes 28Si+, 29Si+, and 30Si+ are clearly observed without contributions from 14N + at m/z ) 28, 14N H+ at m/z ) 29, and 16O + at m/z ) 32. 2 2 2
Figure 11. Radio frequency GD-TOFMS spectra of silica glass at (a) 4.0 ms after power onset (plateau) and (b) 7.0 ms after power onset (afterpeak).
afterpeak. Spectra such as those shown in Figure 10 demonstrate the discrimination and consequent analytical utility that is possible using the pulsed GD-TOFMS system. Part a illustrates the fullscale spectrum of the NBS 1275 copper-nickel alloy, while spectrum b is an enlargement of the low-concentration constituent signals. Again, this pulsed spectrum exhibits the characteristic absence of interfering discharge gas species, giving the spectrum a much cleaner appearance as well as increased signal intensity resulting from a higher contribution from the Penning ionization process. Detection limits for the NBS 1275 copper-nickel alloy are characteristically in the low part-per-million range for this system at the present time. For example, the detection limit of the 56Fe+ isotope with a concentration of 1.46% is 19 ppm for a spectral acquisition time of 25 min. Future efforts will be directed at improving the sensitivity of this system. Pulsed ratio frequency GD-TOFMS is an alternative technique for the analysis of nonconducting samples. The rf-powered plasma operates by establishing a dc bias voltage on the sample surface, (37) Duckworth, D. C.; Donohue, D. L.; Smith, D. H.; Lewis, T. A.; Marcus, R. K. Anal. Chem. 1993, 65, 2478.
CONCLUSIONS The data presented illustrate the analytical utility of time-offlight mass spectrometry monitoring of glow discharge ion sources operating in a pulsed power mode. Spectra acquired during the afterpeak time regime exhibit increased analyte sensitivity with only small contributions from interfering species. A narrow slit increases the resolving power of the instrument but also influences isotope signals measured for a given element. A compromise slit width was arrived at to optimize the resolution and isotope ratio accuracy of the system. The use of an rf-powered GD source facilitated the direct analysis of solid nonconducting samples such as glasses and ceramics. Future studies will involve the implementation and use of an Einzel lens system to improve both the resolution and the sensitivity of the current instrument. ACKNOWLEDGMENT The technical assistance of Mr. Donald Feathers in the design and fabrication of the rf probe and modifications to the TOFMS system is greatly appreciated. Dr. Charles Pan is recognized for pioneering GD-TOFMS in our laboratory. We gratefully acknowledge the support of this research by the Office of Naval Research and the U.S. Department of Energy. Received for review November 19, 1996. February 13, 1997.X
Accepted
AC961171I X
Abstract published in Advance ACS Abstracts, April 1, 1997.
Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
1721
