76
Anal. Chem. 1866, 58,76-80
(2) Lawrence, J. F.; Laver, G. W. J . Assoc. Off. Anal. Chem. 1974, 5 7 , 1022. (3) Glatt, V.; Meler, W. Mift. Geb. Lebensmmelunfers. Hyg. 1980, 71(4), 526. (4) Alcock. N. J.; Corbelli, L.; Games, D. E.; Lant, M. S.; Westwood, S. A. Biomed. Mass Spectrom. 1982, 9( 1 I ) , 499. (5) Ha, Y-D.; Bergner, K. G. Dfsch Lebensm. Rundsch 1980, 76( 1 I ) , 390. (6) Udenfriend, S.; Stein, S.; Bohlen, P.; Dairman, W.; Leimgruber, W.; Weigele, M. Sclence (Wphingfon, D . C . , 1883-) 1972, 178, 871. (7) Cruces Blanco, C.; Garcia SBncher, F. Anal. Chem. 1984, 5 6 , 2035. (8) Garcia Sanchez, F.; Navas, A.; Santlago, M. Anal. Chlm. Acta 1985, 167, 217.
.
.
(9) Cruces Blanco, C.; Garcia SBnchez, F., J . Assoc. Off. Anal. Chem. 1988, 69, 50. (IO) Miller, J. N.; Ahmad, T. A,; Fell, A. F. R o c . Anal. Div. Chem. SOC. 1982. 19. 37. (11) Rublo, S.:Gomez-Hens, A.; ValcBrcel, M. Anal. Chem. 1985, 5 7 , 1101. (12) Silva de, J. A. F.; Stronjny, N. Anal. Chem. 1975, 47, 714. (13) Lloyd, J. B. F.; Ewett, I. W. Anal. Chem. 1977, 49, 1710. (14) Guidelines for Data Adquisition and Data Quality Evaluatlon In Environmental Chemistry Anal. Chem. 1980, 52,2242.
RECEIVED for review May 28,1985. Accepted August 9,1985.
Negative Ion Mobility Spectrometry for Selected Inorganic Pollutant Gases and Gas Mixtures in Air G. A. Eiceman,* C. S. Leasure, and V. J. Vandiver Department of Chemistry, New Mexico State University, Las Cruces, New Mexico 88003
Spectra for inorganic gases common to combustion emisdons were obtained in air by using negative ion mobility spectrometry (NIMS). Gases were characterized individuaiiy from 10 to over 4000 ppb in alr and in two binary mixtures of S02/N02 and HCI/SOP. Spectra for HCI and SO2 were distinguished by single Intense peaks at reduced mobilities ( K O )of 2.98 cm2/(V 8 ) and 2.30 cm2/(V s), respectively. The product ion for NO, had a K O of 2.80 cm2/(V 8 ) as determined using binary mixture studies and was unresolved otherwise from reactant Ions. I n contrast, NIMS spectra for H2S showed three peaks that were concentration dependent and had mobility values of 2.38, 2.28, and 1.70 cm2/(V 8 ) . Peak currents for H2S ranged from 10 to 50 pA and the estimated limit of detection (LOD) was 10 ppb. Peak currents for SO2 ranged from 15 to 65 pA and the estlmated LOD was 2 ppb. The range of peak currents for HCI was comparatively large at 140-215 pA and estimated LOD was NO2 > SOP
An environmental aspect in the operation of coal-fired power plants, municipal incinerators, and other combustionbased units is production of large amounts of inorganic waste gases, including SOz, NOz, HC1, CO, and COz. In addition to corrosion of equipment from acids, release of these gases into atmospheric environments has also been associated with other deleterious effects, such as acid deposition and respiratory irritation in humans. Subsequent to the Clean Air Act of 1968, monitoring of gaseous emissions for inorganic gases from stacks as point sources has been practiced nationally. Presently, at least 10 methods have been developed for sensing such pollutants and include wet chemical ( I ) , electrochemical ( 2 , 3 ) ,spectroscopic ( 4 , 5 ) ,and chemiluminescence (6) methods. When evaluated in terms of speed, sensitivity, selectivity, continuous operation, automation, and low cost, these approaches all have certain limitations. For example, chemiluminescence detection of SOz may have interference from other sulfur-containing compounds (6),while peroxyacetyl nitrate (a common photochemical pollutant) can greatly in-
terfere in the detection for NOz. Electrochemical sensors for these gases have been reviewed in detail (2, 3) but, despite promise, are limited by specificity, the number of sensors needed, and maintenance demands. Instrumentation for optical spectroscopictechniques, including Fourier transform infrared spectrometry, with better specificity and speed can be expensive, and inexpensive wet chemical and adsorbent methods are slow and labor intensive. Additionally, a more serious limitation for wet chemical techniques is the timeaveraged or integrating nature of such sensors. Ion mobility spectrometry (IMS) is an atmosphericpressure based technique that has been used almost exclusively in prior applications for separation and detection of organic compounds in the vapor phase. Reactant ions which are produced from /3 emission into a supporting gas (Nz, air) react with analyte through ion/molecule collisions and product ions are produced. Although formation of positive product ions through proton (or NH4+and NO+) exchanges has been more generally used, negative product ions can also be formed through electron attachment and other mechanisms (7,8). Thus, negative ion mobility spectrometry (NIMS) will result in a mobility spectrum for separation and detection of only negative product ions. These reactions are especially useful for compounds containing atoms with high electron affinity such as halogens, oxygen, and nitrogen. Since compounds that do not attract or hold protons may show some sensitivity as negative ions, NIMS has advantages of selectivity and possible extension to applications with inorganic gases not usually associated with IMS. Although inorganic gases have been characterized by use of atmospheric pressure ionization mass spectrometry (9, IO), and NIMS spectra for a few of these gases have been reported ( I I ) , response of NIMS toward such compounds has been unexplored. Work described below is directed at two aspects of use of IMS in sensing of atmospheric environmental pollutants: (1) possible interferences from inorganic gases in NIMS detection of organic compounds, and (2) direct multiple analysis of emissions for inorganic gases. Should IMS be used for in-stack monitoring of organic compounds such as polycyclic aromatic hydrocarbons or polychlorinated biphenyls, inorganic gases may certainly be major chemical interferences. Thus, knowledge of IMS behavior for HC1, SO,, NO,, and others will provide background information on feasibility for using NIMS
0003-2700/86/0358-0076$01.50/00 I985 American Chemlcal Society
ANALYTICAL CHEMISTRY, VOL. 58, NO. 1, JANUARY 1986
77
(or positive IMS) as an in-stack monitor. Negative IMS has properties including simple instrumentation, continuous operation, high sensitivity, and fast response, which are attractive for contiquous atmospheric monitoring. Although promising, NIMS has not been demonstrated as suitable for inorganic gases and the objective for this work is to document quantitatively NIMS response to select stack gases, including NO2, SO2, HCl, and H2S.
EXPERIMENTAL SECTION Instrumentation. An ion mobility spectrometer, constructed in-house, with closed-segmented design and unidirectional gas flow (12),was equipped with an 11.2 mCi 63Niionization source, computer-based digital signal averager (13),0-10 kV +/-power supply (ModelRR-10-6R Gamma High Voltage, Mt. Vernon,NY), and picoammeter (Model 427, Keithley, Cleveland, OH). The IMS was housed in an oven from a Hewlett-Packard Model 5720 gas chromatograph and has been described in detail (12). Conditions for operation were as follows: tube temperature, 211 "C; drift gas, 99.5% grade air at 500 mL/min; carrier gas, Nz; atmospheric pressure, 660.0-665.0 torr; gate time, 0.1 ms; amplification, 1 to 10 X 1OI1 A; total voltage on IMS tube, -3600 V; length of drift region, 9 cm; drift field strength, 355 V/cm, and frequency of signal acquisition, 28 Hz. All gas flows were controlled with Nupro fine metering valves (Willoughby, OH). Procedures. Negative ion mobility spectra were obtained by using inorganic gases (Matheson Co., Rutherford, NJ) at concentrations from 10 to 4000 ppb in air. Concentrations were determined from mass flow rates for pure reagent gas and for nitrogen as diluent gas. Gases were directed throughout the gas manifold by using uncoated 0.25 mm i.d. fused silica capillary tubing. Spectra were acquired only after signal became stable (usually less than 30 s after flows were fixed) and lo00 scans were averaged for each spectrum. At a frequency of 28 Hz, signal averaging lasted about 40 s while sample was introduced continuously into the IMS tube in the ring adjacent to and upflow from the reactor ring. Of the compounds sampled, only NOz had a mobility indistinguishable from reactant ions and collection of response curve and limits of detection directly was not possible without a mass spectrometer detector. However, for other compounds,response curves and limits of detection were prepared in plots of product ion peak current vs. concentration. These compounds included SOz, HCl, and HzS. Response of NIMS for binary mixtures was characterized by use of two combinations of gases, SOZ/NOzand HC1/SOZ. A constant and continuous flow of SO2at 512 ppb in air was mixed in the IMS with concentration of NO2 from 6 to 250 ppb and spectra were collected at regular intervals. Similarly, 60 ppb of HCl in air was introduced continuously with a variable concentration of SOz from 98 to 833 ppb.
RESULTS AND DISCUSSION Spectra for Inorganic Gases. A spectrum for reactant ions in negative ion mobility spectrometry (NIMS) is shown in Figure 1A as a plot of detector current vs. time (milliseconds). As found in prior studies, three negative reactant ions were observed by use of air as the drift gas. Reduced mobilities for these three peaks were 3.04,2.82, and 2.54 cm2/(V s), two of which (2.82 and 2.54) matched favorably with earlier NIMS/MS reports where identities were assigned based on mass analysis. The peak with reduced mobility of 3.04 cm2/(V s) was assigned as C1- impurity. Reported values for reduced mobilities of mass identified reactant ions in NIMS were (in cm2/(V s)) 2.76 for NOz- and CNO- and 2.55 for COS- (7). Results for SOz, HzS, HCl, and NO2 are shown in NIMS spectra in Figure 1B-E. In NIMS, product ions are created through three possible reactions, including charge (-) transfer, proton abstraction, and dissociative attachment reactions (8). Therefore, spectra for the inorganic gases should be expected to be M-, (M - H+)-, and (M - F)-, where F is a neutral fragment from M. The spectrum for HC1 in Figure 1D was a single intense peak with reduced mobility of 2.98 cm2/(V
z +
w K
U 3
z
P
0
I2
16
DRIFT T I M E C M 5 )
. Ion mobility spectra in air for (A) reactant Ions, (6)SO2, HCI,
and (E) NOz.
Since HC1 is likely to undergo dissociative attachment, the intense peak in NIMS for HCl is probably C1-. This is also consistent with IMS results from other C1-containing compounds such as chlorobenzenes, chlorinated phenols, and some pesticides where dissociative capture leads to loss of the C-C1 bond (14-16). Thus, use of /3 ionization from 63Niwill not lead to distinguishable spectra for C12 and HC1, since neither species shows a molecular ion in NIMS. The spectrum shown was acquired with 60 ppb of HC1 in air, but band broadening (most likely due to clustering) became severe at concentrations above 250 ppb. For example, peak widths were 150 ps at 100 ppb but were 400 ps a t 300 ppb. The limit of detection was estimated as
