Anal. Chem. 1996, 68, 1136-1142
Glow Discharge Mass Spectrometric Analysis of Atmospheric Particulate Matter Wim Schelles, Kris J. R. Maes, Stefan De Gendt, and Rene´ E. Van Grieken*
Department of Chemistry, University of Antwerp (UIA), Universiteitsplein 1, B-2610 Antwerpen, Belgium
A direct current (dc) glow discharge mass spectrometer has been used to analyze atmospheric particulate matter. The sample preparation used is simple and time-saving. The air is sucked by a pump through a single-orifice impactor stage, in which the aerosols are impacted on a metal support, forming a central spot. This metal plate is directly used as a cathode in a dc glow discharge mass spectrometer. Evaluation of the sample loading and of the discharge parameters allowed us to optimize the signal intensity and to minimize its decrease, the latter being a consequence of its consumption by continuous sputtering in the discharge. The available aerosol analysis time could be prolonged to more than 3 h, a time span necessary to perform a multielement analysis using a magnetic sector instrument and long integration times. A NIST reference aerosol was measured to evaluate the quantitative analysis potential. The internal reproducibility was better than 10% RSD, and the limits of detection were estimated to be in the low ppm or sub ppm region. Even without the use of any standards or correction factors, glow discharge mass spectrometry could offer good semiquantitative results, based only on the use of an internal standard. Aerosols (or atmospheric particulate matter) are defined as suspended solid or liquid particles in a gaseous medium.1,2 The particle size ranges from about 0.01 to more than 10 µm. Characterization of these particles has become of great interest in atmospheric chemistry, especially because of health concerns. Quantitative trace analysis of bulk aerosols is usually done by atomic absorption analysis, a single-element technique, or by multielemental inductively coupled plasma emission or mass spectrometry3,4 and voltammetry.5,6 These techniques require a solution sample; hence, they can only be used after (timeconsuming) quantitative destruction of the aerosol sample. Alternatively, there are techniques, like X-ray fluorescence, particleinduced X-ray emission, and neutron activation analysis, that can handle solid samples but have other shortcomings: many elements cannot be measured with satisfactory sensitivity, or not at all, and the quantification may be difficult.4,7 (1) Hinds, W. C. Aerosol Technology, Properties, Behavior, and Measurement of Airborne Particles; John Wiley & Sons: New York, 1982; Chapter 1. (2) Go ¨tz, G.; Me´sza´ros, E.; Vali, G. Atmospheric Particles and Nuclei; Akademiai Kiado´: Budapest, 1991; Chapter 2. (3) Bettinelli, M.; Baroni, U.; Pastorelli, N. J. Anal. At. Spectrom. 1987, 2, 485489. (4) Krivan, V.; Franek, M.; Baumann, H.; Pavel J. Fresnius J. Anal. Chem. 1990, 338, 583-587. (5) Khandekar, R. N.; Dhaneshwar, R. G.; Palrecha, M. M.; Zarapkar, L. R. Fresenius J. Anal. Chem. 1981, 307, 365-368.
1136 Analytical Chemistry, Vol. 68, No. 7, April 1, 1996
In this study, the use of direct current glow discharge mass spectrometry (dc-GDMS) to analyze aerosols is investigated. GDMS has already proven to be a very powerful tool for analysis of conducting and semiconducting materials.8-10 The most important advantages are the low limits of detection (even subppb levels) and the good semiquantitative results obtained even without the use of an external standard. However, the main problem is the requirement that the sample be electrically conducting. Different ways have been investigated to overcome this problem. Substituting the common dc source to a radio frequency (rf) discharge is a very appealing method because nonconducting samples can be measured directly, without the use of an auxiliary conductor.11,12 Promising results have been reported, and rf glow discharges have already been interfaced successfully with commercial mass spectrometers.13,14 For dcGDMS, however, nonconducting powder samples are mostly brought into a conducting host matrix (e.g., copper powder,15,16 gallium,17 tantalum,18 etc.), and electrodes can be formed. Alternatively, flat, solid, nonconducting samples, e.g., glass plates, can be analyzed with dc-GDMS by means of the so-called secondary cathode technique. The sample is exposed to the discharge through an aperture of a conducting diaphragm, the secondary cathode. Due to the continuous in situ sputter-redeposition of the secondary cathode atoms, a thin conducting layer is formed on the nonconducting sample. This allows atomization of the nonconducting material as the bombarding ions can penetrate through the thin conducting layer. This concept has already been applied successfully to the analysis of several sample types with varying chemical and physical properties: glass, iron ore, (6) Nimmo, M.; Fones, G. Anal. Chim. Acta 1994, 291, 321-328. (7) Quisefit, J. P.; de Chateaubourg, P.; Garivait, S.; Steiner, E. X-Ray Spectrom. 1994, 23, 59-64. (8) Harrison, W. W.; Barshick, C. M.; Klinger, J. A.; Ratliff, P. H.; Mei, Y. Anal. Chem. 1990, 62, 943A-949A. (9) Mykytiuk, A. P.; Semeniuk, P.; Berman, S. Spectrochim. Acta Rev. 1990, 13, 1-10. (10) King, F. L.; Harrison, W. W. Glow discharge mass spectrometry. In Glow Discharge Spectroscopies; Marcus, R. K., Ed.; Plenum Press: New York, 1993; Chapter 5. (11) Duckworth, D. C.; Marcus, R. K. Anal. Chem. 1989, 61, 1879-1986. (12) Marcus, R. K.; Harville, T. R.; Mei, Y.; Shick, C. R., Jr. Anal. Chem. 1994, 66, 902A-911A. (13) Duckworth, D. C.; Donohue, D. L.; Smith, D. H.; Lewis T. A.; Marcus, R. K. Anal. Chem. 1993, 65, 2478-2484. (14) Shick, C. R., Jr.; Raith, A.; Marcus, R. K. J. Anal. At. Spectrom. 1993, 8, 1043-1048. (15) Woo, J. C.; Jakubowski, N.; Stuewer, D. J. Anal. At. Spectrom. 1993, 8, 881-889. (16) De Gendt, S.; Schelles, W.; Muller, V.; Van Grieken, R. J. Anal. At. Spectrom. 1995, 10, 681-687. (17) Efimov, A. G.; Pavela, M.; Paulovicheva, E. Presented at 17th Arbeitstreffen Festko ¨rpermassenspektrometrie, KFA, Ju ¨ lich, Germany, 2-4 May, 1994. (18) Tong, S. L.; Harrison, W. W. Spectrochim. Acta 1993, 48B, 12371245. 0003-2700/96/0368-1136$12.00/0
© 1996 American Chemical Society
Figure 1. Aerosol sample spot, impacted on the metal support.
polycarbonate, marble, Teflon, aluminum oxide, and uranium oxide.19-22 Aerosols are, depending on the geographical origin, mainly composed of soil dust, sea salt, sulfates, etc.; therefore, they can generally be considered as “nonconducting”. They have already been analyzed with dc-GDMS, using the above-described method for nonconducting powder analysis.23 For this goal, a large amount of powder had to be collected; after it was pressed into an indium rod, an electrode was formed. We preferred, however, a much less time-consuming and more direct sample preparation method, which has not yet been used for dc-GDMS: the direct impaction technique. A similar concept has already been used for secondary ion mass spectrometry and sputtered neutrals mass spectrometry measurements,24 but those studies aimed at depth profiling of the aerosols rather than bulk trace analysis. In our approach, the sampled atmospheric particles are directly impacted on a metal plate. The nonconducting central aerosol spot is surrounded by the underlying conducting support material (see Figure 1). This situation is comparable with the starting point of the secondary cathode concept, where also a conducting area surrounds a nonconducting one. In this methodological study, the first results concerning this sample preparation method, combined with dc-GDMS analysis, are presented. It is investigated whether the secondary cathode concept also holds for the configuration used, i.e., whether the continuous in situ sputter-redeposition is necessary to create a stable discharge. Different parameters, like the required amount of aerosol, the nature of the support material, and the discharge conditions, are evaluated with regard to the intensity and stability of the aerosol signal. Finally, the quantitative possibilities are estimated by means of a standard reference aerosol analysis. EXPERIMENTAL SECTION Mass Spectrometer. All of the glow discharge work reported in this study was done with a VG9000 (Fisons Instruments (19) Milton, D. M. P.; Hutton, R. C. Spectrochim. Acta 1993, 48B, 39-52. (20) Schelles, W.; De Gendt, S.; Muller, V.; Van Grieken, R. Appl. Spectrosc. 1995, 49, 939-944. (21) Schelles, W.; De Gendt, S.; Maes, K.; Van Grieken, R. Fresenius J. Anal. Chem., in press. (22) Betti, M. European CommissionsInstitute for Transuranium ElementssAnnual Report. EUR 16152 EN 1995, 188-193. (23) Takahashi, T.; Takaku, Y.; Masuda, K.; Shimamura, T. Bunseki Kagaku 1994, 43, 1083-1086. (24) Goschnick, J.; Schuricht, J.; Ache, H. J. Fresenius J. Anal. Chem. 1994, 350, 426-430.
Elemental Analysis, Winsford, Cheshire, UK). This doublefocusing instrument of reversed Nier-Johnson geometry has already been described in detail elsewhere.25 A typical working resolution (at 5% peak height) of 3500 was routinely used for these studies. The detection system consists of a combination of a Faraday cup and a Daly detector, providing a dynamic range of about 10 orders of magnitude (i.e., 1 × 10-19-1 × 10-9 A). For all the measurements, the “new flat cell” was used, already described elsewhere.26 The cell was cryogenically cooled to reduce the background due to residual gases. The glow discharge was supported by high-purity argon (Air Liquide, 99.9997%, 2 ppm H2O as most important impurity). Sample Preparation. Sampling was done at the campus of the University of Antwerp (UIA), Belgium. Air was sucked with an electrical pump at a rate of about 16 L/min into a single-orifice Battelle impactor. This device consists of consecutive funnelshaped stages. The smaller the exit opening, the faster the air goes through and the smaller the size of the particles collected on impactor plates positioned behind the exit orifice. Therefore, impactors are able to separate sampled particles with different sizes on different supports.27 In this study, only one stage of the impactor was used for the aerosol collection. This one stage was completed with two other stages (without impacting plates) to prevent rain from coming directly in the impactor. The collection stage had a 2 mm exit orifice, resulting in an aerosol spot size of about 3 mm on the support, positioned 0.5 mm behind the exit orifice. The theoretical cutoff diameter of the impactor stage used was 0.5 µm; this means that, in principle, all particles with a diameter larger than 0.5 µm are collected. The possible loss of the smallest particles (0.5 ppm) in Different Support Materials
B Na Mg Al Si P S Cl K Ca Ti Cr Ag Cd Sn
Cu
Ta
Al
0.07 1.2 0.04 1.0 1.4 4.7 9.0 0.2 2.2
