Detection of Negative Ions from Individual Ultrafine Particles

Transcription

1 Anal. Chem. 2002, 74, Detection of Negative Ions from Individual Ultrafine Particles David B. Kane, Jinjin Wang, Keith Frost, and Murray V. Johnston* Department of Chemistry and Biochemistry, University of Delaware, Newark, Delaware Aerosol mass spectrometers can be used to classify individual airborne particles on the basis of chemical composition. While positive ion mass spectra are normally used to characterize ultrafine particles (defined here as particles smaller than 200 nm in diameter), negative ion mass spectra can provide complementary information. To effectively utilize the negative ion mass spectra of ultrafine particles, it is important to understand biases in the formation and detection of negative ions. It is found that the intensity of negative ions is generally less than that of positive ions, due to the creation of electrons in the ablation process that must react to form negative ions. The ablation efficiency, defined as the probability that an ablated particle produces a detectable ion signal, exhibits both size and composition dependencies. The ablation efficiency for detection of negative ions follows the same trends as the ablation efficiency for the detection of positive ions: sodium chloride and ammonium nitrate have higher ablation efficiencies than oleic acid, and the ablation efficiency decreases with the particle diameter. The ablation efficiency of negative ions is less than or equal to the ablation efficiency of positive ions, and the relative difference increases as the particle diameter decreases. Pure ammonium sulfate particles exhibit an ablation efficiency too low to be measured in the present experiments. However, trace amounts of sulfate in mixedcomposition particles can be readily detected in the negative ion mass spectra. Aerosol mass spectrometers are becoming an important tool for the chemical characterization of ambient particulate matter. 1,2 With these instruments, aerosol particles are sampled directly from ambient conditions into the source region of a mass spectrometer. The most common version of an aerosol mass spectrometer couples laser ablation with time-of-flight mass spectrometry for ionization and mass analysis. 3 This combination gives the advantages of rapid sampling and particle-by-particle analysis. Recently, the lower size limit of laser ablation aerosol mass spectrometers was extended for composition analysis of ultrafine Current address: Philip Morris USA Research Center, Richmond, VA Current address: Department of Chemistry MS-60, Rice University, P.O. Box 1892, Houston, TX (1) Johnston, M. V. J. Mass Spectrom. 2000, 35, (2) Suess, D. T.; Prather, K. A. Chem. Rev. 1999, 99, (3) Johnston, M. V.; Wexler, A. S. Anal. Chem. 1995, 67, 721A-726A. particles, defined here as particles smaller than 200 nm in diameter In this range, size and composition biases of the hit rate, defined as the rate at which single-particle mass spectra are collected, cause the probability of detection to be different for different size and composition particles. 10 These detection biases must be understood if aerosol mass spectrometers are to be used to infer aerosol compositions. Field studies of particles smaller than 200 nm have been limited to the use of positive ion mass spectra because free electrons dominate the negative ion mass spectra. 6 In principle, positive and negative ion mass spectra can provide complementary information on the composition of particles. 11 Some species, such as sulfate, are only detected in negative ion mass spectra, while in other instances, the negative ion mass spectra can be used to confirm the presence of a particular species. Previously we reported on the biases in the detection of positive ions from ultrafine particles. 10 These biases are found to result from two parameters, both of which exhibit size and composition biases. The first is the sampling rate (SR) of particles into the mass spectrometer. This is the rate at which particles of a given size and composition are sampled from the aerosol into the path of the ablation laser beam when the laser fires. The second is the ablation efficiency (E a ). This is the fraction of particles irradiated by the laser beam that ablate in a manner that produces a detectable ion signal. The relationship between the observed hit rate (HR, particles per unit time that are detected), sampling rate, and ablation efficiency is HR ) n aerosol (SR)E a (1) where n aerosol is the particle number density of the aerosol being sampled. While the sampling rate is independent of the polarity of the detected ions, the ablation efficiency is expected to be dependent on the ion polarity. (4) Reents, W. D.; Downey, S. W.; Emerson, A. B.; Mujsce, A. M.; Muller, A. M.; Siconolfi, D. J.; Sinclair, J. D.; Swanson, A. G. Plasma Sources Sci. Technol. 1994, 3, (5) Reents, W. D.; Ge, Z. Aerosol Sci. Technol. 2000, 33 (1-2), (6) Carson, P. G.; Johnston, M. V.; Wexler, A. S. Rapid Commun. Mass Spectrom. 1997, 11, (7) Ge, Z.; Wexler, A. S.; Johnston, M. V. Environ. Sci. Technol. 1998, 32, (8) Kane, D. B.; Oktem, B.; Johnston, M. Aerosol Sci. Technol. 2001, 35, (9) Kane, D. B.; Oktem, B.; Johnston, M. Aerosol Sci. Technol. 2001, 34, (10) Kane, D. B.; Johnston, M. V. Environ. Sci. Technol. 2000, 34, (11) Galli, M.; Guazzotti, S. A.; Prather, K. A. Aerosol Sci. Technol. 2001, 43 (4), Analytical Chemistry, Vol. 74, No. 9, May 1, /ac011126x CCC: $ American Chemical Society Published on Web 03/21/2002

2 In this study, positive and negative ion detection are compared for single-component particles having compositions representative of ambient particles. Measurements are restricted to particles greater than 40 nm in diameter because the hit rate decreases significantly with decreasing particle size. The detection of sulfate ions in mixed-composition particles is discussed, as are the implications of these measurements for the characterization of ambient aerosol. EXPERIMENTAL DETAILS: The design of the aerosol mass spectrometer is similar to that previously described by Kane et al. 9,10 In the present investigation, the mass analyzer was operated in the linear mode with a postacceleration detector (bipolar time-of-flight detector, Burle Optoelectronics, Inc., Sturbridge, MA). Identical acceleration and postacceleration fields were used in both positive and negative ion detection modes. Because previous investigations showed that free electrons produced by laser ablation of ultrafine particles can saturate the detector, obscuring the negative ion signal, 6 a slight deflection field (100 V) was applied at the end of the ion source to divert electrons from the detector. The deflection field was used in both positive and negative ion modes to maintain equivalent field gradients. An Ar-F excimer laser (PSX-100, MPB Technologies, Pointe Claire, PQ, Canada) was used for ablation. The laser beam was focused with a 20-cm-focal length lens to give a fluence of J/m 2 (193 nm, 8 ns, 2.5 mj pulse) in the source region of the mass spectrometer. The aerosol mass spectrometer was used to collect singleparticle mass spectra and measure normalized hit rates for particles representative of ambient compositions in both positive and negative ion modes. 9,10 The normalized hit rate is the rate at which single-particle mass spectra are collected for a given particle number density of the aerosol. It should be noted that the hit rates reported in this work are not comparable to those reported previously, 10 due to differences in the ionization laser and mass analyzer. The particle beam profile was determined by measuring the hit rate as the laser beam was translated across the ion source region of the mass spectrometer. 9,10 Since the mass spectrometer used in these experiments was not capable of simultaneous positive and negative ion detection, the average mass spectra of a set of particles were compared rather than those of individual particles. In this manner, changes in the particle mass spectra resulting from shot-to-shot variations were minimized. Monodisperse test aerosols used in these experiments were generated with a differential mobility analyzer to size select particles produced by atomization of a solution containing the analyte in either ethanol or an ethanol-water mixture. A second differential mobility analyzer was used in-line with the aerosol mass spectrometer to measure the size and concentration of the particles. RESULTS AND DISCUSSION: Particle Mass Spectra. Negative and positive ion mass spectra of oleic acid, ammonium nitrate, and sodium chloride are compared in Figure 1. Each spectrum is the average of 100 singleparticle mass spectra. The negative ion mass spectrum of each particle composition shows distinctive ions representative of that composition. Oleic acid particles give strong ion signals for H -, C 2-, and C 2 H -. Ammonium nitrate gives strong ion signals for Figure 1. Positive and negative ion mass spectra of (a) sodium chloride, (b) ammonium nitrate, and (c) oleic acid. Each spectrum is the average of 100 single-particle mass spectra from 100-nm mobility diameter particles. O -,OH -, and NO 2-. Sodium chloride particles are characterized by Cl - and small peaks for NaCl 2-. The positive ion mass spectra of these compounds are qualitatively the same as observed previously with a reflectron time-of-flight mass spectrometer. 10 The negative ion mass spectra of these particles have peaks similar to the negative ion mass spectra of micrometer-diameter particles. 3,12,13 However, the mass spectra are dominated by smaller (12) Carson, P. G.; Johnston, M. V.; Wexler, A. S. Aerosol Sci. Technol. 1997, 26, Analytical Chemistry, Vol. 74, No. 9, May 1,

3 Figure 2. (a) Relative intensities of negative ions to positive ions from ammonium nitrate, sodium chloride, and oleic acid as a function of particle mobility diameter. Error bars are 50% confidence limits from at least three measurements, 100 particles each. (b) Integrated intensity of the negative ion signal for the compounds above as a function of particle mobility diameter. Error bars are the standard deviation of at least three measurements, 100 particles each. ions and usually do not contain the larger cluster ions observed from the micrometer-diameter particles. A similar observation was made previously for positive ion mass spectra of ultrafine particles. 6 The greater fragmentation observed for ultrafine particles is most likely a combination of effects: as the optical penetration depth of the laser beam into the particle approaches the particle size, the increased photon flux per unit volume gives a higher probability of multiphoton absorption and fragmentation; and the decreasing plume size and density as the particle size decreases results in fewer collisions and less efficient cooling as the plume expands. Figure 2a is a plot of the integrated negative ion signal relative to the positive ion signal as a function of the particle size. Each point is the average of at least three 100-particle measurements taken over the course of several experiments over several days. (13) Neubauer, K. R.; Johnston, M. V.; Wexler, A. S. Int. J. Mass Spectrom. Ion Processes 1997, 163, While the absolute ion signals exhibited a high degree of variability, the relative intensities in Figure 2a were consistent in all of the experiments. For oleic acid and sodium chloride, the negative ion signal from detected particles is significantly less than the positive ion signal. Since overall charge neutrality must be preserved, this illustrates the dominant role that free electrons can play. Ammonium nitrate produces comparable positive and negative ion signal intensities suggesting that electron capture in the vaporized plume is more efficient than with the other compounds. Sodium chloride exhibits a significant decrease in the relative intensity of negative ions to positive ions as the particle diameter decreases, while oleic acid and ammonium nitrate show little size dependence. The reason for these dependencies is apparent upon examination of Figure 2b, which shows the absolute intensities of negative ions as a function of particle size for the compounds shown in Figure 2a. The dependencies in Figure 2b are similar to those in Figure 2a: the negative ion signal from sodium chloride decreases rapidly with decreasing particle size, while ammonium nitrate and oleic acid show little size dependence. Previously we showed that the positive ion signal from sodium chloride particles decreases only slightly with decreasing particle diameter in this size range. 6,9 Differences in the positive and negative ion signal intensities and their dependencies on the particle diameter suggest that multiple mechanisms may be responsible for the formation of negative ions from ablated particles. It is expected that laser ablation results in the production of free electrons, and negative ions must be formed by the capture of these electrons by the vaporized molecules or fragments. One would expect that the negative ions formed by the ablation process will be biased toward those species with the greatest electron affinities, provided that the energy released by electron attachment is removed by collisional cooling. However, negative ion formation also depends on the electron capture cross sections of the vaporized species and the residence time of the electrons in the plume. The electron capture cross section is dependent on the electron kinetic energy as well as properties of the molecule or fragment. Since the electron kinetic energy distribution is determined by complex and ill-defined processes associated with photon absorption, ionization, and particle disintegration, a quantitative explanation of these dependencies is not possible. However, the size dependence of the negative ion signal from sodium chloride suggests that the electron capture cross section of Cl is low enough that the probability of Cl - formation is proportional to the electron residence time in the plume. Larger particles produce more dense plumes resulting in a larger percentage of negative ions. In contrast, ammonium nitrate and oleic acid do not show significant size dependencies. Evidently, the electron capture cross sections of polyatomic species from these particles must be large enough that the production of negative ions is relatively independent of the plume density and hence the particle diameter. Ablation Efficiency. By measuring the hit rate and determining the diameter of the particle beam from the beam profile, it is possible to determine the ablation efficiency for a given particle composition and size. 10 Measurements of the ablation efficiency for negative ion detection show the same trends that have been observed for positive ion detection. 10 Sodium chloride and ammonium nitrate exhibit higher ablation efficiencies than oleic acid, 2094 Analytical Chemistry, Vol. 74, No. 9, May 1, 2002

4 Figure 4. Negative ion mass spectrum of 77-nm mobility diameter particles composed of ammonium nitrate with 1% ammonium sulfate. The spectrum is an average of 100 single-particle spectra. Figure 3. Size (mobility) and composition dependencies of the hit rate for negative ion detection relative to the hit rate for positive ion detection. The relative hit rates are equivalent to the relative ablation efficiencies. Error bars are the standard deviation of at least three measurements, 100 particles each. and the ablation efficiencies for all three decrease with decreasing particle size. It is not surprising that the trends in the negative ion ablation efficiency are similar to the positive ion ablation efficiency, since the positive and negative ions are created simultaneously. Figure 3 shows the ablation efficiency for negative ion detection relative to that for detection for positive ion detection. Since the sampling rate of particles into the laser beam is independent of the ion polarity, the relative hit rate is equivalent to the relative ablation efficiency (see eq 1). In most cases, the negative ion ablation efficiency is much less than the positive ion ablation efficiency. This behavior indicates that the conversion of free electrons to negative ions varies from particle to particle and does not always yield sufficient ions to exceed the detection threshold. Relative to the positive ion ablation efficiencies, the negative ion ablation efficiencies of all three compounds decrease with decreasing particle diameter. As the particles decrease in diameter, fewer are detected in the negative ion mode, suggesting that free electrons constitute a greater fraction of the negative ion signal. For sodium chloride and oleic acid, the data also suggest that the positive and negative ion ablation efficiencies become equivalent in the limit of a large particle size, although this possibility was not explored for particles above 200 nm in diameter. It should be noted that the size and composition dependencies of the relative hit rates in Figure 3 do not match those of the relative signal intensities in Figure 2a. These differences are not surprising since different quantities are measured. The ablation efficiency is a measure of the fraction of all sampled particles that produce an ion signal exceeding the detection threshold level, while the relative signal intensities are measured only for the subset of particles where the ion signal exceeded the threshold. Detection of Sulfate in Mixed-Composition Particles. Although sulfate is an important component of atmospheric aerosols, the sensitivity of laser ablation mass spectrometry to sulfate is very low in the positive ion mode. Sulfate particles 3 µm in diameter give negative ion signals that correspond to the sulfate species present in the particle, for example, the HSO 4 - ion from ammonium sulfate. 14 However, we were unable to detect pure ammonium sulfate particles of <200 nm in diameter in the negative ion mode, suggesting that these particles have negative ion ablation efficiencies below It is not surprising that the negative ion ablation efficiency is low, since we reported previously that the ablation efficiency of ammonium sulfate in the positive ion mode is less than 0.02 and that the mass spectra of pure ammonium sulfate particles are dominated by trace impurities. 10,15 This suggests that only ammonium sulfate particles containing impurities are detected. Since particles in the atmosphere are likely to consist of a mixture of components, we investigated the detection of sulfate in mixed-composition particles. The average mass spectrum from 77-nm ammonium nitrate particles with 1% ammonium sulfate is shown in Figure 4. In this spectrum, the HSO 4- ion (m/z 97) from sulfate is clearly observed above the noise level. Also present are O - (m/z 16), OH - (m/z 17), and NO 2- (m/z 46). Although, ammonium nitrate is present at a concentration 100 times greater than ammonium sulfate, the signals from NO 2- and HSO 4- are similar, suggesting that sulfate competes favorably for charge in the vaporized plume. While small amounts of sulfate can be detected in mixedcomposition particles, the hit rate decreases as the amount of sulfate in the particle increases. Figure 5 shows the effect of sulfate concentration on the particle hit rate for ammonium nitrateammonium sulfate particles. The hit rate drops significantly as the sulfate mass fraction increases above 20%. Similar behavior was observed for other sizes of ammonium nitrate-ammonium sulfate particles and for oleic acid-ammonium sulfate particles. At high sulfate concentrations, the particles are difficult to detect, since neither positive nor negative ions are efficiently produced by laser ablation. As the ammonium sulfate mass fraction approaches 1, the ablation efficiency becomes dependent on the presence of trace impurities in the particle. It is these impurities that promote ionization and allow sulfate to be detected by electron capture or charge transfer in the vaporized plume. Under the conditions used in this work, the relationship between particle composition and the relative signal intensities (14) Neubauer, K. R.; Sum, S. T.; Johnston, M. V. J. Geophys. Res. 1996, 101 (D13), , 707. (15) Kane, D. B.; Johnston, M. V. Anal. Chem. 2001, 73, Analytical Chemistry, Vol. 74, No. 9, May 1,

5 Figure 6. Dual-polarity mass spectra of a single ambient particle containing sulfate. The particle size is nominally 70 nm in diameter (aerodynamic). Figure 5. Dependence of the normalized particle hit rate (particles detected per second for an aerosol containing 1000 particles/cm 3 ) on the mass fraction of ammonium sulfate for 150-nm mobility diameter ammonium nitrate-ammonium sulfate particles. Error bars are the standard deviation of at least three measurements, 100 particles each. (16) Hinz, K.-P.; Kaufmann, R.; Spengler, B. Aerosol Sci. Technol. 1996, 24, (17) Gard, E.; Mayer, J. E.; Morrical, B. D.; Dienes, T.; Fergenson, D. P.; Prather, K. A. Anal. Chem. 1997, 69, of the corresponding ions is complex and cannot be used to quantitatively determine the sulfate concentration from the mass spectra. This complexity reflects the strong coupling of ionization, particle disintegration, and electron capture processes induced by laser ablation. Implications for Ambient Particle Measurements. The goal of a single-particle experiment is to determine the composition of an ambient aerosol from the distribution of single-particle mass spectra. With this goal in mind, several factors must be taken into consideration when negative ion detection is used for a singleparticle mass spectrometer. First is the need to detect species that do not give strong ion signals in the positive ion mode, such as sulfate. Second is the limitation imposed by lower particle hit rates and lower ion signal intensities. If it is necessary to detect species such as sulfates, which do not give strong ion signals in positive ion mode, then negative ion detection must be used. The poor ablation efficiencies of these particles can potentially be overcome by increasing the sampling rate of particles into the laser beam, either by increasing the laser repetition rate or by decreasing the size of the particle beam. However, if the detection of species in the negative ion mode is not of paramount importance, then positive ion detection is preferred. Aerosol mass spectrometers capable of dual-polarity ion detection have been reported in the literature. 16,17 We recently experimented with a dual-polarity instrument designed for field measurements of ultrafine particles. This instrument is similar in design to a previous field instrument having single-polarity detection. 18 Dual-polarity detection combines the advantage of triggering on the positive ion mass spectrum while simultaneously collecting the negative ion spectrum. An example of a dual-polarity ambient particle spectrum is shown in Figure 6. In the negative ion mode, HSO 4- indicates the presence of sulfate, while the positive ion mode indicates the presence of organic components and iron. While low concentrations of sulfate can be readily detected in particles composed primarily of other materials, particles containing nearly 100% ammonium sulfate are difficult to detect at all. Recently, we demonstrated that by coating particles with a UVabsorbing compound it is possible to increase the ablation efficiency of pure ammonium sulfate particles by a factor of When combined with dual-polarity detection, this provides a reliable method for detecting particles with a high concentration of ammonium sulfate. ACKNOWLEDGMENT This research was supported by Grant CHE from the National Science Foundation. Ambient particle measurements with the dual-polarity mass spectrometer were supported by the Department of Energy National Energy Technology Laboratory through cooperative agreement no. DE-FC26-OINT41017 and were performed in collaboration with Anthony Wexler, Derek Lake, Michael Tolocka, and Yongjing Zhou. Received for review October 29, Accepted February 5, AC011126X (18) Mallina, R. V.; Wexler, A. S.; Rhoads, K. P.; Johnston, M. V. Aerosol Sci. Technol , Analytical Chemistry, Vol. 74, No. 9, May 1, 2002