Particle-Fluorescence Spectrometer for Real-Time Single-Particle

A particle-fluorescence spectrometer (PFS) for real-time measurements of ... and (4) provide a time stamp for a data block of spectra with time resolu...
0 downloads 4 Views 3MB Size
Download PDF
Environ. Sci. Technol. 2009, 43, 429–434

Particle-Fluorescence Spectrometer for Real-Time Single-Particle Measurements of Atmospheric Organic Carbon and Biological Aerosol Y O N G - L E P A N , * ,† R O N A L D G . P I N N I C K , ‡ STEVEN C. HILL,‡ AND RICHARD K. CHANG† Department of Applied Physics, Yale University, 15 Prospect Street, New Haven, Connecticut 06520, and U.S. Army Research Laboratory, 2800 Powder Mill Road, Adelphi, Maryland 20783

Received June 5, 2008. Revised manuscript received October 8, 2008. Accepted November 14, 2008.

A particle-fluorescence spectrometer (PFS) for real-time measurements of single-particle UV-laser-induced fluorescence (UV-LIF) excited with a pulsed (263-nm) laser is reported. The dispersed UV-LIF spectra are measured by a 32-anode PMT detector with spectral coverage from 280-600 nm. The PFS represents a significant improvement over our previous apparatus [Pinnick et al., Atmos. Environ. 2004, 38, 1657] and can (1) measure fluorescence spectra of bacterial particles having light-scattering sizes as small as 1 µm (previously limited to about 3 µm) and so can measure particles with size in the range of 1-10 µm, (2) measure each particle’s elastic scattering which can be used to estimate particle size (not available previously), (3) measure single-particle fluorescence spectra with a laser and detector that can record spectra as fast as 90,000/s, although the highest rates we have found experimentally in atmospheric measurements is only several hundred per second (previously limited by detectors to only 25/ s), and (4) provide a time stamp for a data block of spectra with time resolution from 10 ms to 10 min. In addition, the PFS has been modified to be more robust, transportable, and smaller. The use of an aerodynamic-focusing sheath inlet nozzle assembly has improved the sample rate. The PFS has been employed to measure UV-LIF spectra from individual atmospheric particles during October-December 2006 and January-May 2008 in New Haven, CT, and during January-May 2007 in Las Cruces, NM.

Introduction The importance of atmospheric aerosol, including anthropogenic aerosols, to earth climate (1), atmospheric chemistry (1, 2), and respiratory health (3) has driven the development of improved sensors to characterize aerosol particles. Part of this effort has been directed toward improving real-time single-particle measurement techniques, which are particularly useful for studying minority aerosol species. Minority * Corresponding author phone: [email protected]. † Yale University. ‡ U.S. Army Research Laboratory. 10.1021/es801544y CCC: $40.75

Published on Web 12/15/2008

203-432-4231;

 2009 American Chemical Society

e-mail:

species of aerosol may be difficult to observe in bulk collections where many particles are analyzed at once. In some cases bulk samples may be amenable to analysis using single-particle techniques, e.g., a Raman microprobe system, but the required time and complexity may be impractical for routine real-time measurements. Single-particle aerosol mass spectrometry, a real-time analysis technique, has been used extensively for atmospheric aerosol monitoring (4-7). This paper focuses on a complementary single-particle technique in which individual atmospheric particles are rapidly interrogated as they flow through an optical cell, using UV-laserinduced fluorescence (UV-LIF) spectroscopy. Even though single aerosol particles contain only a small amount of material (typically a few picograms or less), the LIF of many organic-carbon (OC) compounds in atmospheric aerosol is sufficiently intense for particles near 1-µm diameter or larger that it can be detected in a flow-through system. Techniques for measuring intrinsic LIF of aerosol particles may be divided into two main categories: (a) spectrally undispersed LIF, where the emission is measured only for one or two broadband wavelength channels (8-13); (b) spectrally dispersed LIF, where the emission is measured in many (e.g., 8 to hundreds) of channels (14-18). The PFS system presented here measures dispersed fluorescence spectra. Fluorescence has been used in studies of bulk samples of OC materials that occur in atmospheric aerosol, such as humic and fulvic substances, bacteria, cellulose, lignans, pollens, polycyclic aromatic hydrocarbons (PAHs), and “dissolved organic matter”. Essentially all OC compounds that fluoresce are aromatic. Most PAHs are fluorescent, although many substituted PAHs (e.g., with nitro groups or heavy halogens) and many heterocyclic aromatics (e.g., the bases of DNA or RNA) have very weak fluorescence. Many of the highly fluorescent molecules that appear in atmospheric aerosol comprise a very small fraction of any particle’s mass. Therefore, although OC materials are extremely complex, in some cases only one or a few key fluorescent OC compounds dominate the fluorescence spectra and may serve as markers for a particular type of aerosol. For example, washed bacteria typically have a fluorescence spectrum that is dominated by tryptophan (18), even though there are many different aromatic OC compounds in bacteria (e.g., flavins). The fluorescence spectrum of bacteria, although clearly not unique, can be an indicator for bacteria or other materials that contain the amino acid tryptophan (e.g., most proteins). Similarly, the fluorescence from ferulates may dominate the fluorescence of cellulose. Atmospheric OC particles (at a single site and during a single season) were shown to have fluorescence spectra that can be clustered into a relatively small number of classes (19), using a detection system based on the scheme of Pan et al. (17) but with some modifications facilitating atmospheric aerosol measurements. A virtual impactor was used to concentrate particles in the 2- to 10-µm size range, and the concentrated aerosol entrained in the minority outlet flow was fed through an aerodynamic focusing nozzle. This nozzle formed a narrow (400 µm diameter) laminar jet within an optical cell where individual particles were probed by a 266-nm UV laser, and their fluorescence spectra were measured as they flowed through the cell. However, the aerosol sample rate, sensitivity, and data collection rate of this system were not optimal for several reasons: (a) the inlet aerosol sampling nozzle did not have sheath flow to provides a more highly focused and more laminar aerosol jet; (b) a conventional spectrograph with smaller-than-optimal nuVOL. 43, NO. 2, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

429

FIGURE 1. Schematic of the particle fluorescence spectrometer. Ambient air is drawn into the inlet where particles having sizes between 2 and 10 µm (diameter) are concentrated with a virtual-impactor (not shown). The concentrated aerosol particles entrained in the minority exit flow of the concentrator are drawn into a 2 in. cubic optical chamber through a sheath nozzle (inset), a double aerodynamic-focusing nozzle. Single particles flowing through the cell are illuminated with a pulsed 263-nm UV laser. The fluorescence emission is collected by a reflective objective, focused onto a slit, dispersed by a concave grating, and detected with a 32-anode PMT. merical aperture was used to disperse the fluorescence emission, and so the sensitivity of the detection system was less than optimal; (c) the dispersed emission was detected using an intensified CCD (ICCD) camera (256 × 1024 pixel chip) which limited the measurement rate to about 25/s because of the chip’s read-out time. The present paper describes an improved, more robust, particle-fluorescence spectrometer (PFS). The key improvements are (1) a greater sensitivity to weak fluorescent signals, i.e., fluorescence spectra of bacterial particles as small as about 1 µm in diameter can be detected, (2) an increased sample rate provided by a more tightly focused and more laminar inlet aerosol jet, and on-board electronics which can record up to 90,000 fluorescence spectra/s in a burst mode, although the highest rates we have found experimentally in atmospheric measurements are only several hundred per second, (3) a capability to time-stamp recorded spectra with 10 ms to 10 min resolution, (4) a capability to estimate particle size from the elastic scattering signal, and (5) a reduction in size from 21′′ × 21′′ × 21′′ to 16′′ × 14′′ × 8′′. Some spectral measurements obtained with this PFS were reported previously in a paper that compared spectra obtained at different locations with different instruments (21).

Experimental Section Figure 1 illustrates the present PFS detection system, where the main upgrades are (1) a double inlet nozzle with sheath flow, (2) a miniaturized optical chamber, (3) a light-scattering based particle sizer, (4) a multiple-anode PMT for spectral detection, and (5) a time stamp. For sampling atmospheric aerosol, the PFS employs a virtual impactor concentrator (MSP model 4220) having an inlet sample rate of 330 L/min to concentrate particles in the 2- to 10-µm size range. Particleenriched air in the minority outlet flow of the concentrator (nominally 1 L/min) is fed into the inner-nozzle of a doublenozzle assembly. The double-nozzle is mounted in a small optical chamber (a cubical airtight cell, 2 in. on each side). As air is drawn though this nozzle, a focused, laminar, cylindrical aerosol jet (around 400-µm in diameter) is formed downstream from the nozzle. The tapered inner nozzle aerodynamically focuses the more massive particles (e.g., those with diameter larger than 3-µm) more than it does the 430

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 2, 2009

FIGURE 2. Single-particle UV-laser-induced fluorescence vs elastic scattering measured with the PFS for a fairly monodisperse population of B. subtilis vegetative cells with nominal diameter 1.36 µm. Also shown is the best fit line. The lower histogram is of the single-particle laser-induced elastic scattering (coefficient of variation CV ) 19%), and the upper histogram is of the fluorescence (CV ) 16%). Particles that had fluorescence below about 0.5 were excluded because that is near the lower limit for the PFS to measure spectra with adequate signal-to-noise. smaller particles (e.g., those with diameter of 1-µm). Therefore, the sample rate for the larger particles is further enhanced over that provided by the concentrator. The optical chamber is aspirated through an outlet tube aligned with the inlet nozzle assembly. A piston pump (KNF Neuberger model UN 86) draws air through the chamber. A uniform flow is achieved by a critical orifice between the outlet tube and the piston pump, which reduces pressure fluctuations caused by the pump. Clean air is bled to the outer nozzle inlet through a filter and forms a clean-air sheath for the aerosol jet. This double-nozzle assembly provides for a nearly uniform speed (around 10 m/s) for all particles in the jet. Particles flowing within about 75 µm of the center of the jet are detected by two diode-laser beams (650- and 685-nm wavelength), which are focused (beam diameter around 150

FIGURE 3. PFS fluorescence spectra of 300 B. subtilis particles having nominal 1.36-µm diameter excited by single-shot (263-nm wavelength) laser pulses. The peak at 263 nm is the elastic scattering of the pulsed 263-nm laser beam from the particles leaking through the long-pass filter, which is used for estimating particle size.

FIGURE 4. Effective sample flow rate as a function of particle size for the PFS. The measured particles were made from tryptophan dissolved in water and aerosolized using an IJAG. An interpolation line is also shown.

FIGURE 5. Light-scattering from aerosolized calibration polystyrene latex spheres vs particle diameter. Each data point is an average of scattering from about 500 single particles. The dashed line is an interpolation. µm) to intersect near the center of the jet. Two PMTs (each with its own interference filter that passes the 650-nm or 685-nm scattered light) detect the light scattered from particles passing through the intersection of these two beams. When signals on both PMTs exceed a preset threshold, an AND gate triggers a pulsed UV probe laser with a 1-mm beam diameter, 263-nm wavelength, 0.05 mJ per pulse, and 10-ns pulse length (Nd:YLF, Photonics Industries model DC150-263). Once the probe laser fires, some of the particle

fluorescence is collected by a large-aperture (NA ) 0.4) reflective lens (Newport, model 50105) and focused onto the input slit of a spectrograph (Jobin Yvon, CP-140) for spectral dispersion. The fluorescence spectrum is detected by a 32anode PMT (Hamamatsu H7260) which is gated on by a trigger from the AND gate. The UV-laser energy is used efficiently compared to a CW laser, because the probe laser fires only when a target particle is in the sample volume (near the focal point of the objective lens). The UV-laser spot size is set to about 1 mm to increase the uniformity of illumination of the targeted particles (particle positions vary slightly at the time the laser fires, because of small variations in particle speeds and trajectories). The spectrometer slit width is set sufficiently large (1 mm) to allow fluorescence from particles throughout the sample volume to be collected by the highly (15×) magnifying reflective objective lens. The large slit width reduces the spectral resolution to about 15 nm, and the overall resolution is reduced further because the 32-anode PMT records the spectra from 280 to 600 nm using 20 anodes (about 16 nm per anode). The large slit width is needed for the optics to collect a large fraction of the fluorescence from the particles that trigger the UV laser. The diode lasers cannot be focused too tightly or the sample rate will be reduced (see below). A long-pass liquid filter (dimethylformamide, DMF) diluted with water in a 1-cm thick cell), placed in front of the spectrograph slit, blocks nearly all the elastic scattering from the 263-nm laser but efficiently transmits the fluorescence with wavelength longer than 280 nm. The ratio of DMF to water is adjusted so that the magnitude of the elastic scattering leaking through the filter can be used to estimate particle size (after this scattering is spectrally resolved by the spectrograph) but is not so large as to saturate the detector. The fluorescence spectrum and elastic scattering intensity for each particle, measured by the 32-anode PMT, are captured and analyzed by a custom-designed readout and processing electronics interface (Vertilon, previously Vtech, PhotoniQ OEM). This board is triggered 50 ns earlier than the UV laser. It first reads the background charges immediately and then reads the signal charges 1 µs after the laser fires and the 32-anode PMT has been exposed to the fluorescence emission. The absolute fluorescence intensity at each anode is obtained by subtracting the background charge from the signal charge. The spectral sensitivity vs wavelength was measured using a calibration lamp. To avoid analyzing spectra that appear too noisy, spectra having fluorescence below a threshold are considered to be “nonfluorescent”. The fluorescence threshold is chosen to be twice the average background noise measured without any aerosol particle present but with the experimental conditions otherwise unchanged (including firing the UV laser). The electronic interface can be programmed to (1) read each fluorescence spectrum rapidly from the PMT (less than 1 µs), (2) determine within 11 µs (of the laser pulse) if the spectrum is in a selected category (e.g., if it has a spectrum similar to that of bacteria), and (3) send an electronic pulse to open a fast valve to release a puff of air that aerodynamically deflects the selected particle (if the spectrum is in the category of particles to be collected) so that an enriched sample of targeted particles can be captured for subsequent identification (23). The alignment of the PFS optics (diode lasers, 263-nm laser, collection reflective objective, spectrometer, PMTs, etc.) wasquitestableoveraperiodofthreemonths(January-March, 2008), as evidenced by the nearly constant response (magnitude and spectral profile of fluorescence and elastic scattering) of the PFS to test particles. Only the nozzle required occasional cleaning (with compressed gas) and slight position adjustment to maintain an optimal trigger rate because of a possible tiny movement caused by the switching the inlet nozzle tubing between the concentrator (for atmospheric VOL. 43, NO. 2, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

431

FIGURE 6. Normalized fluorescence spectra of single atmospheric aerosol particles (thin solid lines) measured Jan 8, 2008 at 10:00 to 10:01 a.m. at Yale University, New Haven, CT. The sharp peaks at 263 nm are a measure of the elastically scattered light (attenuated by the DMF/water filter) from each particle and are an indicator of particle size. The particles have sizes (estimated from this elastically scattered light intensity) in the 1-3 µm range. The average spectrum for each cluster is shown by a thick dashed line. The peaks at 527 nm arise from the second-order diffraction (from the grating) of the 263-nm laser line. For the clusters populated with many spectra, the measured spectra look like a thick line with noise on the edges because so many spectra overlap. The percentages of particles falling into each cluster are also shown in the inset. measurements) and the aerosol generators for test particles. The nozzle adjustment has little effect on the fluorescence (magnitude and spectral profile) of the detected particles, because the positions of all optical components are fixed relative to the trigger lasers and only particles passing through the intersection of the two trigger beams are detected.

Results and Discussion To explore the capability of the PFS to measure bacteria we suspended lyophilized Bacillus subtilis vegetative cells (Sigma Chemical) in water and used an Ink Jet Aerosol Generator [IJAG] (20) to create a fairly monodisperse population of test particles composed of agglomerates of cells. The IJAG generates droplets (about 50 µm diameter) and dries the droplets, leaving dry particles. The number of bacteria in the dried particles should have an approximate Poisson distribution. The elastic scattering and fluorescence spectra of this population were measured with the PFS. Figure 2 shows a scatter-plot of the fluorescence vs elastic scattering, as well as histograms of the elastic scattering and fluorescence, and best-fit Gaussian curves for the histograms. Some, and perhaps most, of the dispersion in the fluorescence and elastic scattering signals is a consequence of the particle-size 432

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 2, 2009

dispersion of the test particles (generated with the IJAG), suggesting that the PFS has the capability of measuring uniform particles with good repeatability. Figure 3 shows fluorescence spectra from B. subtilis vegetative cells used to generate the data in Figure 2. It reveals the relatively small variation in fluorescence intensity and similarity of the fluorescence spectral shapes from these bacterial particles. The spectra show a good signal-to-noise ratio even though the particles are small and suggest that the lower limit of measurement of the PFS for typical bacterial particles is a little smaller than about 1.36-µm diameter. The broad fluorescence peak (at 340 nm) is mainly from tryptophan. The shoulder from 400 to 500 nm may be attributable to fluorescent compounds of the growth material in which the bacteria were grown or maybe partially to reduced nicotinamide compounds such as NADH. Because the PFS employs a virtual impactor concentrator and a focusing nozzle (both of which concentrate large particles more) and diode lasers to scatter from a particle within the trigger volume (where larger particles scatter more and consequently result in a larger trigger volume), the sample rate of the PFS is dependent on particle size (aerodynamic and light-scattering). To measure the size-

FIGURE 7. Time dependence of the percentage of fluorescent particles in clusters 2, 5, and 8 at (a) New Haven, CT in October 2006, and (b) Las Cruces, NM in January 2007. The data is binned into 15 min intervals. Particles in the 1-10 µm range are included. dependent sample rate, we used the IJAG to aerosolize tryptophan-water solutions and generate fairly uniformly sized dispersions of dried particles with nominal diameters 2.3, 3, 4.2, 5, 5.7, 8, and 8.7 µm. The IJAG particle production rate and PFS trigger rate were then monitored to obtain the sampling efficiency of the PFS system for particles of different size. The results are summarized in Figure 4. For nominal 1-µm particles, no significant concentration enhancement should result from either the virtual impactor or the focusing nozzle, and the small light-scattering particle size should reduce the trigger volume and cause the sampling flow rate to be even smaller, less than 1 L/min. The size of particles can be approximately determined with the PFS using the elastic scattering signal. The PFS collects elastically scattered light over an approximate 0.27π sr solid angle subtending a 60° cone with axis perpendicular to the laser propagation direction (see Figure 1). For calibrating the particle size deduced from the elastic scattering intensity measurement, we aerosolized polystyrene latex microspheres suspensions (monodisperse PSL spheres with nominal diameters 1, 2, 3, 4.3, 6, and 8 µm from Duke Scientific) using a Royco Aerosol Generator (model 256) and measured the elastic scattering. The results are shown in Figure 5. The PSL calibration particles were also monitored by an aerodynamic particle sizer (TSI, model 3321) to ensure the population of test particles was dominated by single PSL spheres and not multiplets. To demonstrate the new capabilities of the PFS, we used it to continuously monitor the concentration, size, and fluorescence spectra of aerosol on the Yale University Campus in New Haven, CT (during October-December 2006 and January-May 2008), and on the New Mexico State University Campus in Las Cruces, NM (during January-May 2007). Typically there are around ten to a few hundred particles detected per second within the 1-10 µm size range, with around one million spectra measured each day. The threshold for fluorescence was chosen to be as small as possible (2 times the average background noise) and still retain spectra that do not appear too noisy. We consider the January 2008 New Haven data first. During a 48 h sampling period (Jan 8, 10 a.m. to Jan 10,

10 a.m.) we measured 1,611,835 particle spectra, with 1,286,925 (79.8%) particles having size smaller than 3 µm, and 325, 592 (25.3%) of these having fluorescence above threshold. In order to analyze such a large quantity of spectra, a hierarchial cluster analysis was applied to the data (19). Briefly, each spectrum is treated as a vector and the dot product of this vector is taken with each of the template clusters deduced previously (19). When the dot product of a spectrum with a template is greater than 0.9, the spectrum is assigned to that cluster. If a spectrum has a dot product greater than 0.9 with two clusters, it is assigned to the cluster with which it has the higher dot product. Figure 6 shows a small part (1000 in total) of the 325, 592 fluorescence spectra that grouped into different clusters by hierarchical analysis. The eight spectral clusters shown are similar to those found previously at Adelphi, MD (19, 21) and account for more than 91% of all fluorescent particles smaller than 3 µm. The percentages of each cluster within the complete data set that combine with these clusters are also indicated. The most populated clusters are 2, 5, and 8, and the least populated clusters are 1, 7, and 10. Although not shown here, the relative population of particles in each cluster is similar to that for particles larger than 3 µm. As noted previously (21), some materials that have spectra similar to those in the different clusters include polycyclic aromatic hydrocarbons, fulvic, humic acids, or humic-like substances (clusters 8 and 10); bacteria or other protein containing substances (cluster 2, 3 and 4); and cellulose (cluster 7). The PFS provides a time stamp for blocks of spectral data, so that the time dependence of cluster populations can be ascertained, as illustrated in Figure 7. The percentages of particles in the most populated clusters (2, 5, and 8) are shown for (a) New Haven, CT (NH-CT) and (b) Las Cruces, NM (LC-NM). A diurnal variability is evident for most clusters at both locations. We leave the investigation of the possible generality of this finding and likely causes for future study. Some of the questions that can be addressed, at least in part, with the PFS are as follows. (1) What (beyond particle concentration and size, i.e., PM2.5) can be said about the types of particles that affect human health? Epidemiological studies have shown that PM2.5 correlates with mortality and cardiovascular disease (3). Much less is known about the chemical characteristics of the aerosols that most correlate with health, largely because it is expensive to continuously monitor particle composition. Because the PFS could be relatively inexpensive to operate continuously, yet still provide fluorescence spectra which can be an indicator or tracer for certain types of aerosols, it may be feasible to run PFS instruments in cities where epidemiological studies are being conducted and thereby provide another view of particulate matter that can be used to look for correlations between particle numbers in different spectral categories and human health. (2) How common are particles of primary biological origin in the atmosphere? Particles of primary biological origin have been stated to comprise around 20% of particles in the atmosphere, based largely on reactions with Coumassie Blue, which binds to proteins (24). Measurements with the PFS suggest that the fraction of atmospheric particles that have fluorescence spectra similar to bacteria and proteins that are not mixed to other strong fluorophors is around 1% of all particles greater than 1-µm diameter. Spectra of other clusters may be from particles that contain significant protein but also contain other fluorors that are prominent in the spectrum. The biological component of OC-containing atmospheric aerosol is extremely complex, and approaches such as Coumassie Blue staining, mass spectrometry (4-7), the PFS, and mass spectrometry combined with fluorescence (25) can provide different perspectives on atmospheric particles of biological origin. VOL. 43, NO. 2, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

433

(3) What are the sources of the aerosols measured at some location and time? To discover more about the sources of the aerosols (e.g., varying in time as shown in Figure 7), the PFS could be used to measure the spectra at different sources (e.g., downwind from a freeway, a composting plant, a bus station, agricultural facilities). That data could be combined with measurements of urban wind flows and with wind trajectory data. (4) What are the turbulent-mixing and dispersion statistics of atmospheric aerosols? Aerosols are typically assumed to act as passive scalars in atmospheric turbulence. If both particle size and fluorescence is measured, then more detailed questions can be asked in studies of mixing and dispersion of particles in the atmosphere because aerosol particles having different composition from different sources could possibly be differentiated. Because the PFS is relatively simple and potentially inexpensive, multiple PFSs could be used to monitor the mixing of particles from different sources (naturally occurring or anthropogenic).

Acknowledgments We acknowledge support by the Defense Threat Reduction Agency under the Physical Science and Technology Basic Research Program and by the Army Research Laboratory. We thank Professor Menachem Elimelech (Yale) for providing laboratory space with a window that enabled us to make the atmospheric aerosol measurements. John Bowersett (ARL machinist) made the double-nozzle (clean-air sheath) assembly for the PFS.

Literature Cited (1) Seinfeld, J. H.; Pandis, S. N. Atmospheric Chemistry and Physics: from Air Pollution to Climate Change, 2nd ed.; Wiley: New York, 2006. (2) Finlayson-Pitts, B. J.; Pitts, J. N., Jr. Chemistry of the Upper and Lower Atmosphere Theory, Experiments, and Applications; Academic: San Diego, 2004; especially Chs. 9, 10. (3) Pope, C. A.; Burnett, R. T.; Thun, M. J.; Calle, E. E.; Krewski, D.; Ito, K.; Thurston, G. D. Lung Cancer, Cardiopulmonary Mortality, and Long-term Exposure to Fine Particulate Air Pollution. JAMA 2002, 287, 1132–1141. (4) Murphy, D. M.; Middlebrook, A. M.; Warshawsky, M. Cluster analysis of data from the Particle Analysis by Laser Mass Spectrometry (PALMS) instrument. Aerosol Sci. Technol. 2003, 37, 382–391. (5) Kane, D. B.; Johnston, M. V. Size and composition biases on the detection of individual Ultrafine particles by aerosol mass spectrometry. Environ. Sci. Technol. 2000, 34, 4887–4893. (6) Noble, C. A.; Prather, K. A. Real-time single particle mass spectrometry: A historical review of a quarter century of the chemical analysis of aerosols. Mass Spectrom. Rev. 2000, 19, 248–274. (7) Murphy, D. M. The Design of Single Particle Laser Mass Spectrometers. Mass Spectrom. Rev. 2007, 26, 150–165. (8) Pinnick, R. G.; Hill, S. C.; Nachman, P.; Pendleton, J. D.; Fernandez, G. L.; et al. Fluorescence Particle Counter for Detecting Airborne Bacteria and Other Biological Particles. Aerosol Sci. Technol. 1995, 23, 653–664.

434

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 2, 2009

(9) Hairston, P. P.; Ho, J.; Quant, F. R. Design of an instrument for real-time detection of bioaerosols using simultaneous measurement of particle aerodynamic size and intrinsic fluorescence. J. Aerosol Sci. 1997, 28, 471–482. (10) Eversole, J. D.; Hardgrove, J. J.; Cary, W. K.; Choulas, D. P.; Seaver, M. Continuous, rapid biological aerosol detection with the use of UV fluorescence: Outdoor test results. Field Anal. Chem. Technol. 1999, 3, 249–259. (11) Reyes, F. L.; Jeys, T. H.; Newbury, N. R.; Primmerman, C. A.; Rowe, G. S.; et al. Bio-aerosol fluorescence sensor. Field Anal. Chem. Technol. 1999, 3, 240–248. (12) Seaver, M.; Eversole, J. D.; Hardgrove, J. J.; Cary, W. K.; Roselle, D. C. Size and fluorescence measurements for field detection of biological aerosols. Aerosol Sci. Technol. 1999, 30, 174–185. (13) Kaye, P. H.; Barton, J. E.; Hirst, E.; Clark, J. M. Simultaneous light scattering and intrinsic fluorescence measurement for the classification of airborne particles. Appl. Opt. 2000, 39, 3738– 3745. (14) Pinnick, R. G.; Hill, S. C.; Nachman, P.; Videen, G.; Chen, G.; et al. Aerosol fluorescence spectrum analyzer for rapid measurement of single micrometer-sized airborne biological particles. Aerosol Sci. Technol. 1998, 28, 95–104. (15) Pan, Y. L.; Holler, S.; Chang, R. K.; Hill, S. C.; Pinnick, R. G.; et al. Single-shot fluorescence spectra of individual micrometersized bioaerosols illuminated by a 351- or a 266-nm ultraviolet laser. Opt. Lett. 1999, 24, 116–118. (16) Pan, Y. L.; Cobler, P.; Rhodes, S.; Potter, A.; Chou, T.; et al. High-speed, high-sensitivity aerosol fluorescence spectrum detection using a 32-anode photomultiplier tube detector. Rev. Sci. Instrum. 2001, 72, 1831–1836. (17) Pan, Y. L.; Hartings, J.; Pinnick, R. G.; Hill, S. C.; Halverson, J.; et al. Single-particle fluorescence spectrometer for ambient aerosols. Aerosol Sci. Technol. 2003, 37, 628–639. (18) Hill, S. C.; Pinnick, R. G.; Niles, S.; Pan, Y. L.; Holler, S.; et al. Real-time measurement of fluorescence spectra from single airborne biological particles. Field Anal. Chem. Technol. 1999, 3, 221–239. (19) Pinnick, R. G.; Hill, S. C.; Pan, Y. L.; Chang, R. K. Fluorescence spectra of atmospheric aerosol at Adelphi, Maryland, USA: measurement and classification of single particles containing organic carbon. Atmos. Environ. 2004, 38, 1657–1672. (20) Bottiger, J. R.; Deluca, P. J.; Stuebing, E. W.; VanReenen, D. R. An Ink-Jet Aerosol Generator. J. Aerosol Sci. 1998, 29, s965– s966. (21) Pan, Y. L.; Pinnick, R. G.; Hill, S. C.; Rosen, J. M.; Chang, R. K. Single-particle laser-induced-fluorescence spectra of biological and other organic-carbon aerosols in the atmosphere: Measurements at New Haven, Connecticut, and Las Cruces, New Mexico. J. Geophys. Res., Atmos. 2007, 112. (22) Romay, F. J.; Roberts, D. L.; Marple, V. A.; Liu, B. Y. H.; Olson, B. A. A high-performance aerosol concentrator for biological agent detection. Aerosol Sci. Technol. 2002, 36, 217–226. (23) Pan, Y. L.; Boutou, V. E.; Bottiger, J. R.; Zhang, S. S.; Wolf, J. P.; et al. A puff of air sorts bioaerosols for pathogen identification. Aerosol Sci. Technol. 2004, 38, 598–602. (24) Jaenicke, R.; Matthias-Maser, S.; Gruber, S. Omnipresence of biological material in the atmosphere. Environ. Chem. 2007, 4, 217–220. (25) Stowers, M. A.; van Wuijckhuijse, A. L.; Marijnissen, J. C. M.; Kientz, C. E.; Ciach., T. Fluorescence preselection of bioaerosol for single-particle mass spectrometry. Appl. Opt. 2006, 4-5, 8531–8536.

ES801544Y