Comprehensive On-Line Characterization of Complex Gas Mixtures

Furthermore, several electronic devices for instrument control and data acquisition are mounted in 19-in. racks (vacuum control units, voltage supply ...
0 downloads 15 Views 497KB Size
Download PDF
Anal. Chem. 2004, 76, 6753-6764

Comprehensive On-Line Characterization of Complex Gas Mixtures by Quasi-Simultaneous Resonance-Enhanced Multiphoton Ionization, Vacuum-UV Single-Photon Ionization, and Electron Impact Ionization in a Time-of-Flight Mass Spectrometer: Setup and Instrument Characterization F. Mu 1 hlberger,† K. Hafner,† S. Kaesdorf,‡ T. Ferge,†,§ and R. Zimmermann*,†,§,|

Institut fu¨r O ¨ kologische Chemie, GSF-Forschungszentrum fu¨r Umwelt und Gesundheit, D-85764 Neuherberg, Germany, Analytische Chemie, Lehrstuhl fu¨r Festko¨rperchemie, Universita¨t Augsburg, Universita¨tsstrasse 1, D-86159 Augsburg, Germany, Stefan Kaesdorf-Gera¨te fu¨r Forschung und Industrie, Gabelsberger Strasse 59, D-80333 Mu¨nchen, Germany, and Abteilung Umwelt- und Prozesschemie, BIfA-Bayerisches Institut fu¨r Umweltforschung und -technik, Am Mittleren Moos 46, D-86167 Augsburg, Germany

This paper reports on a newly developed mobile mass spectrometer for comprehensive on-line analysis of complex gas mixtures such as ambient air or industrial process gases. Three ionization methods, namely, the resonance-enhanced multiphoton ionization (REMPI), vacuum-ultraviolet single-photon ionization (SPI), and electron impact ionization (EI) are implemented in this instrument and can be operated (quasi-) simultaneously. By means of this setup, a wide range of compounds can be analyzed due to the unique ionization selectivitiy and sensitivity profiles provided by the different ionization techniques. The mass spectrometer is designed for field application even under severe conditions. The REMPI technique is suitable for the selective and soft ionization (without fragmentation) of aromatic compounds at trace level (ppbv/pptv). The also soft but less selective SPI technique with 118-nm vacuum-ultraviolet laser pulses is used as a second laser-based ionization method. Mass spectra obtained by this technique show profiles of most organic compounds (aliphatic and aromatic species) and of some low IP inorganic substances (e.g., ammonia, nitrogen oxide) down to ppbv concentrations. In addition to the laser-based ionization techniques, EI ionization can be used for analysis of the bulk components such as water, oxygen, nitrogen, and carbon dioxide as well as for detection of inorganic minor components such as HCN or HCl from combustion flue gases at ppmv concentration levels. Each method yields specific mass spectrometric information of the sample composition. Special tech* Corresponding author. E-mail: [email protected]. † GSF-Forschungszentrum fu ¨ r Umwelt und Gesundheit. ‡ Stefan Kaesdorf-Gera¨te fu ¨ r Forschung und Industrie. § Universita¨t Augsburg. | BIfA-Bayerisches Institut fu ¨ r Umweltforschung und -technik. 10.1021/ac049535r CCC: $27.50 Published on Web 10/06/2004

© 2004 American Chemical Society

niques have been developed to combine the three ionization methods in a single mass spectrometer and to allow the quasi-parallel application of all three ionization techniques. Mass spectrometry (MS) is a key technique in analytical chemistry. One current field of application is the development of MS-based on-line monitoring instrumentation, e.g., for industrial process analysis and environmental monitoring. Laser-based resonance-enhanced multiphoton ionization timeof-flight mass spectrometry (REMPI-TOFMS) is known as a highly selective and sensitive analytical method,1-7 which is well suited for the on-line analysis of trace compounds in complex gas mixtures. The REMPI technique uses two or more photons for photoionization, which takes place via an optical resonance absorption step. Due to this resonant absorption, the selectivity of UV gas-phase laser spectroscopy is included in the ionization process. Depending on the molecular system to be analyzed and the sample introduction method (e.g., supersonic jet inlet, providing cooling of the sample molecules, or effusive inlet, generating a warm molecular beam), REMPI selectivity ranges from substance class-specific up to isomer-specific ionization. Particularly, the REMPI mass spectrometry technique is well suited for the (1) Hurst, G. S.; Payne, M. G.; Kramer, S. D.; Young, J. P. Rev. Mod. Phys. 1979, 51, 767-819. (2) Lubman, D. M., Ed. Lasers and Mass Spectrometry; Oxford University Press: New York, 1990. (3) Boesl, U.; Zimmermann, R.; Weickhardt, C.; Lenoir, D.; Schramm, K.-W.; Kettrup, A.; Schlag, E. W. Chemosphere 1994, 29, 1429-1440. (4) Zimmermann, R.; Lenoir, D.; Kettrup, A.; Nagel, H.; Boesl, U. In 26th Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, 1996; pp 2859-2868. (5) Cool, T. A.; Williams, B. A. Combust. Sci. Technol. 1992, 82, 67-85. (6) Thanner, R.; Oser, H.; Grotheer, H.-H. Eur. Mass Spectrom. 1998, 4, 215222. (7) Oudejans, L.; Touati, A.; Gullett, B. K. Anal. Chem. 2004, 76, 2517-2524.

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004 6753

on-line analysis of aromatic compounds8-12 because of its selectivity and softness (i.e., nearly fragmentation-free ionization).13 This includes the application of REMPI-TOFMS instruments for online analysis of the flue gas of industrial incineration plants,6,10,14-16 coffee-roasting off-gas,12,17 and monitoring of internal combustion.7,18,19 Different REMPI ionization schemes are possible. Aromatic species, for example, are easily ionizable with (1+1) REMPI using wavelengths in the range from 200 to 320 nm. Aliphatic compounds can be ionized by REMPI with wavelengths below 220 nm or by more complex REMPI processes as, for example, (2+1) REMPI, where two nonresonantly absorbed photons are needed to reach the resonant absorption step. Many inorganic species can only be ionized using (2+1) or even more complex and unfavorable REMPI processes.2,20 Another laser-based ionization technique, the single-photon ionization (SPI), uses vacuum-ultraviolet (VUV) photons for ionization.21 Selectivity is provided via the ionization potential (IP), as only compounds with an IP lower than the photon energy can be ionized. A typical wavelength for VUV-SPI is 118 nm (generated by frequency tripling of intensive 355-nm Nd:YAG UV laser pulses in a rare gas cell22-24), which is equivalent to an energy of 10.49 eV. This method has gained some attention as an ionization method for mass spectrometry in recent years.3,17,19,25-37 A number compounds such as aliphatic and aromatic hydrocarbons can be (8) Oser, H.; Thanner, R.; Grotheer, H.-H. Combust. Sci. Technol. 1996, 116117, 567-582. (9) Gittins, C. M.; Castaldi, M. J.; Senkan, S. M.; Rohlfing, E. A. Anal. Chem. 1997, 69, 286-293. (10) Heger, H. J.; Zimmermann, R.; Dorfner, R.; Beckmann, M.; Griebel, H.; Kettrup, A.; Boesl, U. Anal. Chem. 1999, 71, 46-57. (11) Franzen, J.; Frey, R.; Holle, A.; Betzold, H.; Ulke, W.; Boesl, U. SAE Tech. Pap. 1993, 930082, 55. (12) Zimmermann, R.; Heger, H. J.; Yeretzian, C.; Nagel, H.; Boesl, U. Rapid Commun. Mass Spectrom. 1996, 10, 1975-1979. (13) Boesl, U.; Neusser, H. J.; Schlag, E. W. Chem. Phys. 1981, 55, 193-204. (14) Zimmermann, R.; Heger, H. J.; Kettrup, A.; Boesl, U. Rapid Commun. Mass Spectrom. 1997, 11, 1095-1102. (15) Zimmermann, R.; Heger, H. J.; Blumenstock, M.; Dorfner, R.; Schramm, K.-W.; Boesl, U.; Kettrup, A. Rapid Commun. Mass Spectrom. 1999, 13, 307-314. (16) Nomayo, M.; Thanner, R.; Pokorny, H.; Grotheer, H.-H.; Stu ¨ tzle, R. Chemosphere 2001, 43, 461-467. (17) Dorfner, R.; Ferge, T.; Yeretzian, C.; Kettrup, A.; Zimmermann, R. Anal. Chem. 2004, 76, 1368-1402. (18) Weickhardt, C.; Boesl, U.; Schlag, E. W. Anal. Chem. 1994, 66, 10621069. (19) Mu ¨ hlberger, F.; Zimmermann, R.; Kettrup, A. Anal. Chem. 2001, 73, 35903604. (20) Letokhov, S. Laser Photoionization Spectroscopy; Academic Press: Orlando, FL, 1987. (21) Pallix, J. B.; Schu ¨ hle, U.; Becker, C. H.; Huestis, D. L. Anal. Chem. 1989, 61, 805-811. (22) Bjorklund, G. C. IEEE J. Quantum Electron. 1975, QE-11, No. 6, 287296. (23) Maker, P. D.; Terhune, R. W. Phys. Rev. 1965, 137 (3A), 801-818. (24) Vidal, C. R. In Tunable Lasers; Mollenauer, L. F., White, J. C., Eds.; SpringerVerlag: Berlin, 1987; Vol. 59, pp 56-113. (25) Becker, C. H. Fresenius J. Anal. Chem. 1991, 341, 3-6. (26) Butcher, D. J. Microchem. J. 1999, 62, 354-362. (27) Kornienko, O.; Ada, E. T.; Tinka, J.; Wijesundara, M. B. J.; Hanley, L. Anal. Chem. 1998, 70, 1208-1213. (28) Materer, N.; Goodman, R. S.; Leone, S. R. J. Vac. Sci. Technol. 1997, A 15 (4), 2134-2142. (29) McEnally, C. S.; Pfefferle, L. D.; Mohammed, R. K.; Smooke, M. D.; Colket, M. B. Anal. Chem. 1999, 71, 364-372. (30) Steenvoorden, R. J. J. M.; Kistemaker, P. G.; Vries, A. E.; Michalak, L.; Nibbering, N. M. M. Int. J. Mass Spectrom. Ion Processes 1991, 107, 475489.

6754

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

readily detected by VUV-SPI-TOFMS while most background gases such as nitrogen (IP ) 15.58 eV), oxygen (IP ) 12.06 eV), carbon dioxide (IP ) 13.77 eV), and water (IP ) 12.62 eV) are excluded from ionization, which otherwise might saturate the TOFMS channel plate detectors. Electron-impact (EI) ionization is one of the standard ionization techniques for mass spectrometry. Due to its nonselective ionization profile and the high fragmentation rate, however, EI is not suited for direct analysis of complex mixtures of organic trace compounds. Overlapping fragmentation patterns cannot be deconvoluted, making the identification of single compounds impossible. Nevertheless, EI represents a valuable method for the ionization of bulk components, such as N2, O2, CO2, and H2O as well as small inorganic molecules such as SO2, HCN, or H2S for direct-inlet mass spectrometry. Furthermore, relatively stable organic compounds such as methane and acetylene, which are present in rather high concentrations, for example, in combustion flue gases, can be analyzed via EI-MS. These compounds show low or even no fragmentation and are not accessible via the laserbased SPI photoionization due to their high IP. In summary, the three ionization methods SPI, REMPI, and EI are complementary regarding their ionization selectivity. If the three methods are applied in parallel or in rapid alternation in a mass spectrometer, the information obtained from the different mass spectra gives a rather comprehensive overview of the sample composition in real time. In this work, a newly designed instrument was developed, combining the three ionization methods in one mobile time-of-flight mass spectrometer. The rugged design allows the quasi-parallel application of all three ionization methods even under severe field measurement conditions. The instrument uses the third harmonic of an Nd:YAG pumped OPO laser system with tunable wavelength for REMPI. The same laser radiation (third harmonic) is used to generate 118-nm pulses by frequency tripling in a rare gas cell. In addition, an electron gun is used for EI ionization. Instrument Setup. A scheme of the mobile laser mass spectrometer (155 cm × 170 cm × 70 cm, length × height × width) is shown in Figure 1A. The general setup consists of a laser cabinet containing the Nd:YAG laser and the OPO laser as well as a nonlinear optical setup, a second cabinet for generation of alternating UV and VUV laser pulses for REMPI and SPI, optical devices for incoupling of the different laser beams into the ion source, and the mass spectrometer with the ion source. Furthermore, several electronic devices for instrument control and data acquisition are mounted in 19-in. racks (vacuum control units, voltage supply for the electron gun used for EI ionization, DAQ computer and OPO laser control computer, high-voltage supplies for the mass spectrometer, pulse generator to establish the trigger setup needed for parallel use of the different ionization methods, drawer containing an external gas standard generator used for quantification and calibration purposes, and a power supply for (31) Shi, Y. J.; Hu, X. K.; Mao, D. M.; Dimov, S. S.; Lipson, R. H. Anal. Chem. 1998, 70, 4534-4539. (32) Trevor, J. L.; Hanly, L.; Lykke, K. R. Rapid Commun. Mass Spectrom. 1997, 11, 587-589. (33) Vries, M. S.; Hunziker, H. E. J. Photochem. Photobiol.. A 1997, 106, 31-36. (34) Werner, J. H.; Cool, T. A. Chem. Phys. Lett. 1997, 275, 278-282. (35) Werner, J. H.; Cool, T. A. Chem. Phys. Lett. 1998, 290, 81-87. (36) Zoller, D. L.; Sum, S. T.; Johnston, M. V. Anal. Chem. 1999, 71, 866-872. (37) Tonokura, K.; Nakamura, T.; Koshi, M. Anal. Sci. 2003, 19, 1109-1113.

Figure 1. (A) Schematic representation of the alternating beam pathways for parallel use of REMPI and SPI. (B) Schematic diagram of the ion source with indicated electron gun for EI and the beam paths for REMPI and SPI as well as the position of the inlet needle.

the OPO laser system). The laser mass spectrometer requires a 220 V/16 A single-phase power supply and no external gas or water cooling in order to allow field applications. In the following, the relevant and newly designed parts of the system are described in more detail. The TOFMS (custom-made by Stefan Kaesdorf, Munich, Germany) can be used in linear and reflectron modes, respectively. When used in reflectron mode, the field-free drift region is 801 mm long and a mass resolution R50% of 1800 m/z measured at 92 m/z can be achieved. The ion source and the flight tube are differentially pumped by a 520 (N2) and a 210 L/s (N2) turbomolecular pump (TMU 521, TMU 261, Pfeiffer Vacuum, Aslar, Germany). The Wiley-McLaren type ion source38 can be operated with static or pulsed acceleration fields. The repeller electrode and the first extraction electrode are on the same potential with opposite polarity to maintain ground potential for the central ionization region, thus minimizing the influence of the grounded inlet needle.10,39 The field-free drift region lies on the potential of the second acceleration electrode (linear). Ions are detected using a two-stage multichannel plate detector (MCP) in Chevron assembly with 40-mm active diameter. The inlet system is based on a continuous molecular beam inlet technique.10,14,19,39 It consists of a heated, hollow, stainless steel needle, pointing into the center of the ion source. Inside this needle, a deactivated fused-silica capillary of 320-µm i.d. is placed. The outlet of the capillary is (38) Wiley, W. C.; McLaren, I. H. Rev. Sci. Instrum. 1955, 26, 1150-1157. (39) Hafner, K.; Zimmermann, R.; Rohwer, E. R.; Dorfner, R.; Kettrup, A. Anal. Chem. 2001, 73, 4171-4180.

aligned with the tip of the needle and is located ∼2 mm above the center of the ion source. The capillary runs through a heated vacuum seal and a heated transfer line and is connected to the on-line sampling path. The gas flow through the capillary is ∼7 mL/min (capillary length 2 m, i.d. ) 320 µm, 250 °C). The capillary acts as restrictor between vacuum (MS ion source) and ambient pressure (sampling system). Behind the orifice of the capillary, an effusive molecular beam is formed. This molecular beam is directly hit underneath the capillary tip19,39 by laser pulses, for photoionization (REMPI, SPI), as well as by the electron beam, for EI ionization. By this effusive molecular beam inlet, only a medium optical selectivity is achieved when REMPI is applied because the sampling gas is not expanded through a restriction nozzle and therefore no adiabatic cooling of the analyte molecules occurs.40 However, with the same needle inlet, a continuous supersonic molecular beam at low flow rates (2-10 mL/min) can be realized by using a fused-silica capillary, which has a conically tapered restriction nozzle at its tip.39,41 Due to the jet cooling of the sample molecules, an enhanced optical selectivity for the REMPI process is achieved and residual fragmentation of, for example, alkanes (SPI) caused by elevated internal energy can be suppressed.42 The TOFMS is equipped with a mass gate, preventing ions of preselected specific m/z value from reaching the detector. MCPs need some time for recharging when a very large number of ions are received in a short time (i.e., peaks with very high intensity are detected). This saturation effect results in a decreased sensitivity for ions reaching the detector shortly after such a highintensity peak. In the present setup, peaks with a signal height exceeding 1 V result in a suppression of peaks from the subsequent higher mass ions in the same TOFMS transient. However, the saturation of the MCP can be avoided by use of the mass gate, which is realized as an array of parallel wires (50-µm wire diameter with distances of 500 µm) aligned orthogonal to the TOF ion path. When the ions of unwanted m/z value pass the mass gate, the wires are charged to (500 V within 10 ns with neighboring wires exhibiting the opposite polarity. Due to the high electric fields, the ions that are close to the mass gate are efficiently deflected. However, as the distance of the positively and negatively charged wires is low, the influence of the mass gate is restricted to ions close to the mass gate. For more distant ions, the fields compensate. Thus, for best time resolution, the mass gate should be located in the Wiley-McLaren spatial focal point of the dual-stage ion source. At the Wiley-McLaren focal point, single masses (i.e., m/z values) can be eliminated from the TOF mass spectrum without influencing neighboring masses. Laser mass spectrometric methods require reliable quantification and calibration methods. In particular, with REMPI, different analytes often exhibit largely different responses. Standard gases, containing relevant analytes in ppmv or ppbv quantities, usually are used for external calibration and quantification of REMPI-MS results.9,10,14,19,43 In this work, the standard gas is generated according to the “defined leak” principle, using diffusion and (40) Heger, H. J.; Boesl, U.; Zimmermann, R.; Dorfner, R.; Kettrup, A. Eur. Mass Spectrom. 1999, 5, 51-57. (41) Zimmermann, R.; Rohwer, E. R.; Dorfner, R.; Boesl, U.; Kettrup, A. German Patent Application, Gemany, 1999. (42) Dagan, S.; Amirav, A. J. Am. Soc. Mass Spectrom. 1996, 7, 737-752. (43) Grotheer, H.-H.; Nomayo, M.; Pokorny, H.; Thanner, R.; Gullett, B. K. Trends Appl. Spectrosc. 2001, 3, 181-206.

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6755

permeation tubes.44,45 A compact calibration gas generation unit, which is mounted in a 19-in. drawer, was built. The core of the unit is a temperature-controlled (26 ( 0.1 °C) aluminum block containing a glass tubing system through which a constant flow of 10 mL/min precleaned and prethermalized (26 °C) air is drawn by a small membrane pump with a flow control system. The temperature control is performed by a Peltier element-based heating/cooling system (type PRG 400, Peltron, Fu¨rth, Germany). The diffusion and permeation tubes, which are filled with the calibration standard compounds, are placed inside the thermalized glass tubing system. The standard compounds are either diffusing through the inserted capillaries at the top of the diffusion tubes with a constant rate or permeating through the Teflon membranes of the permeation tubes at a defined and constant rate into the air flow. The concentration of the different components in the calibration standard gas can be determined by the gravimetrically determined weight loss rate of the respective diffusion/permeation tube (∼µg/h range) and the applied flow rate.10 The main novelty of the here presented instrument is the integration of three ionization methods for parallel use. In the following, the instrumental sections of the system relevant for the ionization processes are considered. Resonance-Enhanced Multiphoton Ionization and SinglePhoton Ionization. Both, the OPO system (custom-made by GWU Lasertechnik, Erftstadt, Germany)46,47 and the VUV rare gas cell are pumped by 355-nm pulses (3-5 ns, 225 mJ, 10 Hz) generated by a Nd:YAG laser (Surelite-III, Continuum, Santa Clara, CA). The OPO technique is based on a nonlinear optical process by which a blue photon of energy E1 is split by means of nonlinear refractive optical crystals (β-barium borate, BBO) into two photons of lower energies E2 (signal) and E3 (idler) with E1 ) E2 + E3. The signal radiation (here, 412-710 nm) is then again frequency doubled resulting in UV laser radiation, which is used for the REMPI ionization process (here, 220-355 nm, 2-4 mJ/pulse, and ∼0.07-nm bandwidth). The field use of OPO laser systems in mobile instruments is challenging. The whole laser setup is thermally stabilized by conducting the temperature-controlled cooling water of the Nd:YAG power supply through a tubing system inside the OPO chamber. For the generation of VUV laser pulses, a rare gas third harmonic generation (THG) cell is used.19 About 10% of the 355nm beam (22 mJ/pules), generated by the pump laser (Nd:YAG), is focused by a quartz lens (f ) 100 mm) through a quartz window into a 180-mm-long stainless steel tube. Special care was taken for minimization of contamination sources for the THG cell rare gas filling. The dimensions of the beam parameters and the paths used were calculated according to the literature.22 The 118-nm radiation was separated from the 355-nm beam (off-axis irradiation of the 355/118-nm beams onto a MgF2 lens according to the literature21,30,48,49) to avoid fragmentation of the analyte molecules (i.e., due to the high intensity of the fundamental beam) and is focused into the center of the ion source. (44) Namiesnik, J. Chromatographia 1983, 17, 47-48. (45) Namiesnik, J. J. Chromatogr. 1984, 300, 79-108. (46) Harris, S. Proc. IEEE 1969, 57, 2096-2113. (47) Fan, Y. X.; Eckardt, R. L.; Beyer, J.; Nolting, J.; Wallenstein, R. Appl. Phys. Lett. 1988, 53, 2014-2016. (48) Steenvoorden, R. J. J. M.; Hage, E. R. E.; Boon, J. J.; Kistemaker, P. G.; Weeding, T. L. Org. Mass Spectrom. 1994, 29, 78-84. (49) Nir, E.; Hunziker, H. E.; Vries, M. S. Anal. Chem. 1999, 71, 1674-1678.

6756

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

The OPO laser as well as the tripling cell for generation of 118-nm laser radiation is pumped by 355-nm pulses from the Nd: YAG pump laser (Figure 1A). The 355-nm beam is split into two beams by a beam splitter. The main part (90%) of the pumping energy is directed to the OPO and the remaining energy into the THG cell for frequency tripling. The 355-nm beam used for the THG process and the resulting OPO UV laser beam are blocked alternately by shutters in order to avoid simultaneous irradiation of UV (i.e., for REMPI) and VUV (i.e., for SPI) pulses. Usually five spectra recorded by each ionization method are averaged, so that the beam shutters have to be changed every 0.5 s. The beam shutters consist of Teflon plates mounted on magnetic switches. The data acquisition program controls the magnetic shutters along with recording of the mass spectra, resulting in separate SPI- and REMPI-mass spectra. However, it is possible to move the shutters with 10 Hz, allowing an alternate REMPI and SPI measurement with 5 Hz. The acquisition of the laser ionization TOF mass spectra is performed by a 250 MHz/1 GS transient recorder PC card (model DP 110, Acqiris, Geneva, Switzerland) at a repetition rate of 10 Hz. The spectra are stored in real time on the hard disk by a home-written software package (developed with LabView, National Instruments, Austin, TX) and can be displayed on the monitor. For more detailed information on data acquisition, see for example refs 10 and 19. Electron Impact Ionization. EI is considerably less efficient when compared with the REMPI or SPI processes. Therefore, to increase the sensitivity of the EI-TOFMS, it is useful to accumulate a large number of EI-TOFMS transients. The typical duration of a time-of-flight spectrum is ∼20 µs; thus, employing a laser repetition frequency of 10 Hz, the spectrometer is “used” only for 200 µs/s for laser ionization-based mass spectrometry. The remaining 999 800 µs/s are available for recording of EI-TOFMS spectra. The implemented electron gun and the gated ion extraction can be operated at repetition rates up to 20 kHz. Therefore, in the interval between the laser shots, the molecular beam can be analyzed by electron impact ionization. The data acquisition for EI-TOFMS is based on a fast multichannel analyzer (see above). Figure 2B shows the timing sequence for laser and electron impact ionization. A schematic drawing of the ion source with indicated electron gun and the respective laser beams as well as the position of the inlet needle is shown in Figure 1B. The extraction field of the ion source is switched off during operation of the electron gun, as the electrons otherwise would be defelcted immediately after entering the ion source. Using two high-voltage pulse generators, the high voltages of the repeller and the first extraction plate can be switched on and off within some nanoseconds. For recording an EI-TOFMS spectrum, the electron gun emits a 1-µs electron pulse. After a break of 50 ns, the ion extraction field is switched on and the ions formed are subsequently accelerated into the spectrometer. The heating current of the electron gun defines the intensity of the electron beam. It is possible to record EI spectra with electrons of different kinetic energy by adjusting the filament potential. In contrast to laser ionization TOF mass spectrometry, the acquisition of the TOFMS data obtained by electron impact ionization was performed using ion-counting techniques based on

Figure 2. (A) Timing sequence for quasi-parallel use of REMPI, VUV-SPI, and EI for mass spectrometric analysis. The laser is operated at 10 Hz, the first five mass spectra per second are recorded with SPI; then the beam dump is switched so that REMPI spectra can be recorded. Between the laser ionization mass spectra, 9980 single-sweep spectra are recorded with electron impact ionization. Resulting EI spectra are sum spectra of 99 800 single-sweep spectra. (B) Trigger setup for the (quasi) simultaneous use of all three ionization methods. (A) Master trigger (from Nd:YAG), (B) electron extraction (electron gun), (C) acceleration plates, (D) signal switch. Channel A has an operation frequency of 10 Hz; channels B-D have an operation frequency of 10 kHz. The first two sequences of channels B and D are suppressed by a function generator.

a fast multichannel analyzer (Fast P7886, Fast ComTec, Munich, Germany). The P7886 analyzer has a minimum channel width of 500 ps; therefore, a typical time-of-flight spectrum with 20-µs duration is composed of 40 000 time channels. The analyzer interprets a detector signal larger than the (adjustable) discriminator threshold as a true ion signal and a signal below the threshold as noise; i.e., a time channel contains either the value 0 (no ion) or 1 (one or more ions). Subsequently, these spectra are added up in real time at a very high repetition rate (up to 20 kHz). The discriminator level is adjusted such that a single-ion signal is just above the threshold. The relative high gas density in front of the inlet needle in the ion source as well as the nonselective ionization process results in difficult quantification of EI mass spectra. If the concentration of molecules is high, several ions of one species can be detected within one spectrometer sweep. However, the multiple event time-to-digital converters will interpret this as a single-ion signal. For quantitative measurements, this saturation effect has to be avoided. The heating current of the electron gun needs to be reduced, thus lowering the electron current to a level where statistically not more than one ion of a specific mass is

detected per cycle. Dealing with highly dynamic on-line measurements, this must be maintained even for the highest possible concentration of any relevant compound that has to be quantified. In practice, often higher electron currents are used, taking into account the saturation for species in high concentrations but enhancing the sensitivity for low-level compounds at acceptable measuring time. Integration of the Three Ionization Methods for QuasiParallel Use. The time sequence of the ionization events in the case of quasi-parallel operation of the three ionization methods REMPI, SPI, and EI is depicted in Figure 2A. For the purpose of quasi-simultaneous use of the two laser-based and the electron impact ionization, a complex trigger setup is required (see Figure 2B). This includes the triggering of the pump laser, the beam block switch, the electron gun, and the extraction HV pulse for the TOF analysis as well as the two data-acquisition systems. Depending on the ionization method (laser-based or EI) the signal from the detector needs to be supplied alternately to the transient recorder (REMPI/SPI) or to the multichannel analyzer (EI). For selection of the appropriate acquisition mode, the detector signal Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6757

Figure 3. Left: Two REMPI spectra of the same external gas standard recorded with two different wavelengths (259 and 290 nm). Right: Wavelength spectra of compounds present in the standard mixture. Due to specific absorption bands, compounds can be selectively ionized. Determined detection limits depend both on laser energy and laser wavelength (REMPI cross section).

preamplifier contains a triggerable signal switch delivering the received signal either to the multichannel analyzer (high trigger input level) or to the transient recorder (low trigger input level). The internal clock of the Nd:YAG laser (i.e., the laser sync out signal) is used as master trigger. This master trigger starts the transient recorder card for the data acquisition of the laser ionization mass spectra. Furthermore, it starts a function generator (type PDG204, Scientific Instruments, Gilching, Germany) that controls the extraction plates of the ion source, the electron gun for EI ionization, and the signal switch for data acquisition. In Figure 2B, details of the trigger scheme are depicted. After the laser trigger A (master trigger-negative edge), ions are formed due to laser ionization (either REMPI or VUV, depending on the position of the beam shutter, which is controlled separately by the data acquisition program). During this time, the electron gun is turned off (no trigger B) and the gain for the signal splitter (D) is set to low, directing the data from the detector to the transient recorder. The extraction field is pulsed (C), and the ions formed due to laser ionization are accelerated into the flight tube. The function generator is capable of suppressing two or more sequences of a distinct channel. Here, channel A has an operating frequency of 10 Hz (corresponding to the frequency of the laser system). Channels B-D have an operating frequency of 10.000 Hz, exhibiting 1000 sequences during the time between two separate laser shots. Channels B and D are suppressed for the first two sequences. Only during the first sequence (t ) 0-0.1 ms) are the laser mass spectra recorded; during the second 6758

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

sequence (t ) 0.1-0.2 ms, channels B and D still suppressed), no recording takes place. From the third sequence forth, trigger B fires the electron gun, channel C triggers the ion extraction field, and channel D is set to high gain, so that signals from the detector are processed by the multichannel analyzer. Then 998 EI-TOFMS transients are recorded (t ) 0.2-99.9 ms) until the next laser ionization sequence is started at t ) 100 ms. By means of this setup, the analysis of a sample gas is possible, recording 10 times/s laser ionization mass spectra (5 REMPI-MS and 5 SPIMS spectra, respectively) and 9980 times/s electron impact ionization mass spectra as shown in Figure 2A. RESULTS AND DISCUSSION Analytical Performance of the Mass Spectrometer with the Individual Ionization Methods. The performance of the instrument using REMPI and SPI using VUV photons was evaluated by recording mass spectra of the external calibration standard. Based on the known concentration of the standard compounds, detection limits were determined. As the REMPI process is extremely dependent on wavelength and laser pulse energy, it is important to be aware that the obtained detection limits are valid only for the given ionization conditions. Additionally, one has to keep in mind that the OPO output energy varies with wavelength. Because two photons are necessary for the (1+1) REMPI process, the ion yield depends on the energy fluence by the power of two. Thus, for determining the detection limit of a distinct compound, one has to find the

Figure 4. (A) SPI mass spectrum of the external BTX calibration gas standard (10 ppmv benzene, toluene, and m-xylene in nitrogen). (B) SPI mass spectrum of the calibration gas mixture as delivered from the standard gas supply unit.

wavelength with an optimum both in laser power and in ionization cross section, which has not necessarily to be the wavelength, where the REMPI cross section would be best on its own. The left part of Figure 3 shows two mass spectra of the same standard gas mixture recorded at two different wavelengths. The right part shows wavelength spectra of the single compounds present in the standard gas mixture in the wavelength range from 255 to 295 nm. The dotted vertical lines represent the wavelengths used for the recording of the mass spectra of the standard gas. The mass peaks below 70 m/z are fragment ions of aniline with a maximum signal height of 1.2% of the molecular ion peak. By lowering the laser fluence at the center of the ion source, these peaks vanish. The mass spectrum obtained with 259-nm UV radiation shows all five analytes present in the standard gas mixture (benzene, toluene, aniline, indole, naphthalene) as they all exhibit absorption bands at 259 nm in the wavelength spectra. As only aniline exhibits an absorption band at 290 nm, aniline is solely detectable by REMPI-TOFMS at this wavelength. The detection limits are calculated according to the literature.10,50 Furthermore, it is assumed that a substance can be detected if its signal peak is two times larger than the noise level of the baseline (S/N >2). In Table 1, the REMPI detection limits of the standard compounds are listed together with the respective wavelength used for measurement. Note that for other wavelengths the detection limits could be even lower. Under optimal conditions, sensitivities in the low pptv region can be achieved with REMPITOFMS.10,43 Similar to the REMPI technique, single-photon ionization was characterized regarding the achievable detection limits. Different standard gas mixtures (premixed standard gas, benzene, toluene, (50) Williams, B. A.; Tanada, T. N.; Cool, T. A. In 24th Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, 1992; pp 1587-1596.

Table 1. REMPI Detection Limits of Different Standard Compoundsa

compound

mass (m/z)

wavelength (nm)

std concn

av spectra

detection limit (S/N >2)

benzene toluene aniline p-xylene indole naphthalene

78 92 93 106 117 128

259 259 290 266 280 290

1.4 ppmv 300 ppbv 800 ppbv 10 ppmv 78 m/z). The SPI spectrum (118 nm) on the other hand shows a profile of the most abundant organic constituents, which predominately are low-mass species such as ammonia, acetaldehyde, acetone. and small hydrocarbons. The very intense peak at 162 m/z is due to nicotine. For many mass peaks, several isomeric or isobaric compounds are present in tobacco smoke and the assigned species must be considered as exemplary. The EI spectrum shows predominantly water, argon, and carbon dioxide as well as some minor peaks originated from the manifold of organic species.

mass ions gain intensity in the spectrum (Figure 6B) and higher methylated phenols, alkyl styrenes, and alkanes and alkenes of higher molecular mass become detectable. Hence, with use of the mass gate, it is possible to avoid detector saturation effects and therefore to increase the dynamic range for detection of target compounds. This is necessary if, despite the use of a selective ionization mechanism, some substances are present in very high concentration causing overwhelming ion signals.

Demonstration of Parallel Operation of the Three Ionization Methods REMPI, SPI, and EI. In Figure 7, a measurement is depicted with all three ionization techniques operated in quasisimultaneous manner. The subject of thedemonstration is tobacco mainstream smoke. Part A shows the resulting REMPI at 266nm mass spectrum, part B the SPI at 118-nm mass spectrum, and part C, the corresponding EI mass spectrum. The different selectivities of the ionization methods are evident. Whereas the REMPI mass spectrum only shows aromatic compounds (mass range above 78 m/z), with SPI aliphatic hydrocarbons can be detected in addition. In both spectra, homologous series of alkylated compounds are visible. Such homologous series build up on phenols, benzene, indole, and styrene in the case of REMPI as well as on alkenes and unsaturated aldehydes, dienes, ketones, benzene, and pyridine in the case of SPI. A tentative assignment of the peaks was performed by considering REMPI selectivities of possible compounds known to be present in tobacco smoke.54,55 In the case of VUV-SPI, the relative abundances of isobaric compounds are also taken into account. If more than one compound is possible, the one with the highest relative abundance is assigned.55 Nicotine is present only in the SPI spectrum as REMPI is not efficient enough at the employed wavelength. This, in addition to accessible mass ranges and compound classes, underlines the potential analytical impact of a comprehensive characterization by parallel use of the two laser-based ionization techniques. Part C of Figure 7 shows the corresponding EI spectrum (23 eV) with suppressed nitrogen (mass gate). The bulk components such as water and carbon dioxide dominate the spectrum. In addition, argon is visible (atmospheric trace gas). The peaks with m/z values above 50 are due to a mixture of fragments of larger molecules and eventually some molecular peaks. The degree of fragmentation is already reduced by using an electron energy of 23 eV in comparison to standard 70-eV EI spectra. The bottom spectrum (marked with an asterisk) in Figure 7 depicts an EI spectrum of air. When compared with the REMPI and VUV spectra of tobacco smoke, this emphasizes the applicability of the instrument for analysis of compounds present in a wide range of concentrations. Obviously a simultaneous measurement of substances is possible from the pptv/ppbv level (see Tables 1 and 2) up to compounds present in percent concentrations (like nitrogen in ambient airsFigure 7). In Table 4, a tentative assignment of peaks detected with all three ionization methods is given. CONCLUSION On-line analysis of complex samples on one hand is desirable for several research fields (e.g., catalysis or investigation of reaction mechanisms) and on the other hand may be applied for monitoring of industrial processes such as combustion processes or food processing. The on-line signals of specific target compounds or compound classes can be used for process control measures, product quality control, or production process optimization. However, in many application fields, the concentrations of the relevant compounds may differ by several orders of magnitude. (54) Zimmermann, R.; Heger, H. J.; Kettrup, A. Fresenius J. Anal. Chem. 1999, 363, 720-730. (55) Stedman, R. L. Chem. Rev. 1968, 68, 153-207.

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6763

By means of the new developed time-of-flight mass spectrometer using the three different ionization techniques (i) REMPI, (ii) vacuum-UV SPI, and (iii) EI in parallel, it is possible to comprehensively analyze compounds on-line in concentration ranges from the percent (EI) down to the ppt/ppbv level (REMPI/ SPI). As mentioned above, a wide variety of applications is possible. In a forthcoming publication,56 results from applications such as on-line analysis of waste incineration flue gases will be presented in more detail. (56) Mu ¨ hlberger, F.; Hafner, K.; Adam, T.; Mitschke, S.; Ferge, T.; Zimmermann, R., in preparation.

6764

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

ACKNOWLEDGMENT The authors thank the members of the laser mass spectrometry group at GSF and University of Augsburg, J. Maguhn, T. Streibel, T. Adam, S. Gallavardin, S. Mitschke, W. Welthagen, M. Grieshaber, and T. Hauler, for contributions during the field measurement campaign. Funding from the German Ministry of Research, BMBF (HGF-Strategiefondsprojekt “Stickoxidminderung”, Grant 01SF9920/7) is gratefully acknowledged. T.F. thanks the Deutsche Bundesstiftung Umwelt, Osnabru¨ck, Germany, for a Ph.D. scholarship. Received for review March 24, 2004. Accepted August 26, 2004. AC049535R