Articles Anal. Chem. 1997, 69, 286-293
Real-Time Quantitative Analysis of Combustion-Generated Polycyclic Aromatic Hydrocarbons by Resonance-Enhanced Multiphoton Ionization Time-of-Flight Mass Spectrometry Christopher M. Gittins,† Marco J. Castaldi,‡ Selim M. Senkan,‡ and Eric A. Rohlfing*,†
Combustion Research Facility, Sandia National Laboratories, Mail Stop 9055, Livermore, California 94551, and Department of Chemical Engineering, University of California, Los Angeles, California 92004
We have combined resonance-enhanced multiphoton ionization (REMPI) time-of-flight mass spectrometry with on-line flame sampling to determine the centerline concentrations of naphthalene, fluorene, and anthracene in a pure methane + oxygen/argon (1:5) diffusion flame. Naphthalene concentrations between 100 parts per billion by volume (ppbV) and 6 parts per million by volume (ppmV) and fluorene concentrations below 50 ppbV are determined using one-color REMPI on jet-cooled samples extracted from the flame; anthracene concentrations in the 5-40 ppbV range are determined using two-color REMPI. The REMPI ion signals are converted to absolute concentrations in real time by performing gas-phase standard additions to the flame sample. Isomer-selective detection of larger polycyclic aromatic hydrocarbons, such as perylene and benzo[a]pyrene, is possible using the twocolor REMPI approach. Risk assessment studies consistently point to the need to regulate the emissions of polycyclic aromatic hydrocarbons (PAHs) and their derivatives formed at trace levels in a variety of combustion devices. Recent legislative actions on air toxics, for example as stipulated in the 1990 Clean Air Act Amendments, target the emission of such trace combustion byproducts with increasing stringency. The average concentration of PAH in the exhaust from a typical 250 000 barrel/day petroleum refinery or 300 MW fossil-fuel-fired power generation plant under normal operating conditions is in the parts per billion by volume (ppbV) or lower range.1,2 However, controlled experiments using a research-grade, industrial style diffusion flame burner show that emissions of PAH precursors increase by orders of magnitude †
Sandia National Laboratories. UCLA. (1) Stockdale, R.; Lev-On, M.; Meeks, N.; Reheis, C. H. Western States Petroleum Association (WSPA) Pooled Source Test Program. Industry Specific Air Toxics Emission Factors. Paper presented at the Air and Waste Management Association Meeting, Pittsburgh, PA, March, 1991. (2) Bjorseth, A., Ramdahl, T., Eds. Handbook of Polycyclic Aromatic Hydrocarbons, Vol. 2; Marcel Dekker Inc.: New York, 1985; Chapter 2. ‡
286
Analytical Chemistry, Vol. 69, No. 3, February 1, 1997
during substoichiometric operation.3 (PAH concentrations were not measured in ref 3.) This observation implies that PAH emissions could be dramatically reduced through active control of the combustion process that avoids conditions favorable to their formation. To implement effective control strategies for environmentally benign combustion processes, a highly sensitive technique that can provide emissions data in real time and at a moderate cost is needed. The detection sensitivities of commercial analytical equipment dictate that all current PAH determinations be made via the “preconcentration” method.4-6 In this approach, gases bearing trace PAHs are withdrawn from combustion processes using sampling probes and concentrated on a variety of organic absorbents. The adsorbed PAHs are recovered by solvent extraction and concentrated. Their concentrations are quantified by addition of internal standards and subsequent analysis using gas chromatography/mass spectrometry and liquid chromatography/fluorescence. The conventional analytical approach only provides evidence for PAH emission long after the sampling process and is clearly unsuited for use as a continuous emissions monitor. We address the need for real-time methods for the speciation, characterization, and quantification of gas-phase PAHs at the ppbV level by coupling two well-known analytical techniques: standard additions, commonly used to compensate for sample matrix effects when making absorbance-based concentration measurements, and resonance-enhanced multiphoton ionization time-of-flight mass spectrometry (REMPI-TOFMS). We demonstrate the effectiveness of our approach on an atmospheric pressure, coflowing diffusion flame of methane and oxygen, which provides a prototypical source of various PAHs in the ppbV to ppmV range. (3) Edwards, C. F.; Goix, P. J. Combust. Sci. Technol., in press. (4) Bjorseth, A., Ed. Handbook of Polycyclic Aromatic Hydrocarbons; Marcel Dekker Inc.: New York, 1983. (5) Otson, R.; Leach, J. M.; Chung, L. T. K. Anal. Chem. 1987, 59, 17011705. (6) CARB, California Air Resources Board. Determination of Polycyclic Aromatic Hydrocarbon Emissions from Stationary Sources. Method 429; Sept 12, 1989. S0003-2700(96)00969-9 CCC: $14.00
© 1997 American Chemical Society
REMPI-TOFMS is a highly sensitive, isomerically selective speciation method and has been used for qualitative assessment of PAH levels in laboratory flames7-10 and quantitative analysis of PAH mixtures preseparated by capillary column gas chromatography;11-14 however, it has not been used for real-time quantitative analysis of PAHs in the gas phase. The primary difficulty in using REMPI as a real-time analytical technique is establishing the relationship between ion signal (coulombs or ion counts per laser pulse) and the analyte concentration (ng/cm3) in the original sample. In principle, the relationship between ion signal and analyte concentration can be determined from a detailed knowledge of the frequency, spatial and temporal profiles of the REMPI laser pulse(s), the molecular excitation and ionization cross sections, and the collection efficiency and responsivity of the mass spectrometer. However, the plethora of parameters required for an accurate calculation generally makes this approach impractical. Postacquisition processing of REMPI-TOFMS data has been used to quantify abundance and establish detection limits for numerous environmental contaminants.15-17 In those experiments, the REMPI signal from a sample with unknown analyte concentration is quantified by off-line comparison to a calibration standard. Off-line analysis yields an accurate, precise determination of analyte concentration but is not best suited for real-time monitoring of target species. The standard addition method discussed here is similar in principle to the off-line method of Cool and coworkers.15-17 A major difference between this work and that of refs 15-17 is that sequential standard additions are made to our continuously drawn samples, and the additions are performed rapidly enough to obtain analyte concentrations in real time. Real-time quantitative analysis of trace gases using REMPITOFMS is not unprecedented. Syage18 demonstrated real-time detection of chemical warfare agent surrogates at sub-ppbV levels by on-line comparison of sample signal with that from a calibrated reference standard. Weickhardt et al.19 have employed one-color REMPI-TOFMS for real-time monitoring of several volatile organic compounds (VOCs) and aldehydes at ppmV levels in engine exhaust and propose to extend measurements to dibenzodioxin and dibenzofuran in the exhaust gases of waste incineration (7) Hepp, H.; Siegmann, K.; Sattler, K. Chem. Phys. Lett. 1995, 233, 16-22. (8) Siegmann, K.; Hepp, H.; Sattler, K. Combust. Sci. Technol. 1995, 109, 165181. (9) Siegmann, K.; Burtscher, H.; Hepp, H. J. Aerosol Sci. 1993, 24, S373-S374. (10) Siegmann, K.; Hepp, H.; Sattler, K.; Burtscher, H.; Siegmann, H.-C. Symp. At. Cluster, Surf. Phys. 1994, 260-263. (11) Weeks, S.; D’Silva, A. P.; Dobson, R. L. M. In Lasers and Mass Spectrometry; Lubman, D. M., Ed.; Oxford University Press: New York, 1990; pp 510529. (12) Dobson, R. L. M.; D’Silva, A. P.; Weeks, S. J.; Fassel, V. A. Anal. Chem. 1986, 58, 2129-2137. (13) Rhodes, G.; Opsal, R. B.; Meek, J. T.; Reilly, J. P. Anal. Chem. 1983, 55, 280-286. (14) Klimcak, C. M.; Wessel, J. E. Anal. Chem. 1980, 52, 1233-1239. (15) Williams, B. A.; Tanada, T. N.; Cool, T. A. Resonance Ionization Detection Limits for Hazardous Emissions. Twenty-fourth symposium (International) on Combustion; The Combustion Institute: Pittsburgh, 1992; pp 1587-1594. (16) Tanada, T. N.; Velazquez, J.; Hemmi, N.; Cool, T. A. Combust. Sci. Technol. 1994, 101, 333-348. (17) Cool, T. A.; Williams, B. A. Combust. Sci. Technol. 1992, 82, 67-85. (18) Syage, J. A. Anal. Chem. 1990, 62, A505-A509 and references therein. (19) Weickhardt, C.; Boesl, U.; Schlag, E. W. Anal. Chem. 1994, 66, 10621069.
plants.20 The analyses of Weickhardt et al. are facilitated by a single, continuous standard addition of calibration gases to their exhaust sample. Analyte concentrations are determined in a single laser shot from the calibrated ratio of analyte ion signal to that from the internal standard(s) coincidentally ionized by the REMPI laser pulse. Their calibration procedure is designed for one-color REMPI and is ideally suited for continuous monitoring of VOC emissions, where most target species are efficiently ionized by 1 + 1 or 2 + 1 REMPI. An analogous calibration procedure for two-color REMPI is far more complex as the energy, overlap, spatial, and temporal profiles of the excitation and ionization laser pulses must be determined. In addition, appropriate calibration gases for two-color REMPI have not been identified. Because isomerically selective detection of larger PAHs will likely require two-color REMPI schemes, the internal calibration method is less well suited for on-line monitoring. Liquid standard additions are commonly used to determine the relation between absorbance and analyte concentration when interferant species in the sample matrix preclude a direct determination from a single absorbance measurement. We use gasphase standard additions equivalently to quantify the relation between the REMPI signal of jet-cooled molecules and their partial pressure in the original vapor sample. The standard addition method circumvents the problem of nonanalyte calibration gases and two-color REMPI normalization schemes. Because the analyte concentration is determined from the relative increase in REMPI signal resulting from the standard additions, it is not subject to the difficulties typically associated with quantitative REMPI-TOFMS: absolute analyte density and Boltzmann temperature in the jet, degree of complexation, or even long-term stability of laser fluence and ion detection efficiency. Field data1 suggest that naphthalene, fluorene, and anthracene are among the most abundant PAH products of incomplete hydrocarbon production. Although naphthalene, C10H8, and fluorene, C13H10, are amenable to single-shot analysis using the internal calibration method of Weickhardt et al., they are also ideal test molecules for the standard addition method. In addition to being particularly abundant products of incomplete hydrocarbon combustion, both species are sufficiently volatile that the creation of low-concentration standards is not difficult. Although anthracene, C14H10, is present at only ppbV levels in our flame, it is the most abundant of the combustion-generated PAHs requiring two-color REMPI for selective detection. This makes it an excellent molecule on which to test the sensitivity and isomer selectivity of two-color REMPI for practical, real-time combustion measurements. EXPERIMENTAL METHOD Figure 1 displays a schematic diagram of the experimental apparatus. The diffusion flame burner consists of three concentric stainless steel tubes, 16, 32, and 48 mm in diameter. Pure methane flows through the center tube at 0.06 mmol s-1 (0.09 slm [slm ) L(STP) min-1]), and the 16% O2/balance Ar mix flows through the middle tube at 2.3 mmol s-1 (3.40 slm). Pure Ar flows through the outer tube at 2.7 mmol s-1 (4.0 slm) to serve as a shield gas, and the burner is isolated from laboratory drafts by a piece of acrylic tubing. The methane flow rate is set by a (20) Zimmerman, R.; Lenior, D.; Kettrup, A.; Nagel, H.; Boesl, U. On-line Emission Control of Combustion Processes by Laser-Induced Resonance-Enhanced Multi-Photon Ionization/Mass Spectrometry. Twenty-sixth symposium (International) on Combustion; The Combustion Institute: Pittsburgh, in press.
Analytical Chemistry, Vol. 69, No. 3, February 1, 1997
287
Figure 1. Schematic diagram of the REMPI-TOFMS flame sampling system configured for two-color REMPI detection.
calibrated mass flow controller (MKS Instruments Inc., Andover, MA), while the argon and oxygen are supplied through calibrated needle valves. The overall composition of the unburned gases is 2.5% CH4, 15.5% O2, and balance Ar. The flame length is 16 mm from the burner surface to the tip of the luminous zone. We use tapered quartz microprobes to perform the on-line sampling of the flame. Two different probes are used for the naphthalene concentration measurements; each has a sampling orifice diameter of ∼0.3 mm but different taper angles, 30° and 60°. The flow rate through each probe is ∼0.14 mmol s-1 (0.2 slm) but depends on its position in the flame. The difference in taper angle did not affect the measured naphthalene concentrations to within the 90% confidence limits of our measurements, and only the 60° probe is used for the anthracene and fluorene concentration measurements. The vertical position of the sample probe relative to the burner surface is controlled by adjusting the burner height with a motorized translation stage (Newport Corp., Irvine, CA). Micrometer-driven translation stages provide adjustment of the probe in the horizontal plane. The sampling probe is oriented along the flame axis to preserve the cylindrical symmetry of the flame and to minimize the perturbation of the visible flame profile. Because the gas flow rate through the probe exceeds the methane flow rate, the probe severely distorts the flame profile near the burner surface (burnerto-probe separation
