Article pubs.acs.org/est
Characterization of Selenium in Ambient Aerosols and Primary Emission Sources Arlette De Santiago,† Amelia F. Longo,† Ellery D. Ingall,*,† Julia M. Diaz,‡ Laura E. King,† Barry Lai,§ Rodney J. Weber,† Armistead G. Russell,∥ and Michelle Oakes†,⊥ †
School of Earth and Atmospheric Sciences, Georgia Institute of Technology, 311 Ferst Drive, Atlanta, Georgia 30332-0340, United States ‡ Biology Department, Woods Hole Oceanographic Institution, 266 Woods Hole Road, Woods Hole, Massachusetts 02543, United States § Advanced Photon Source, Argonne National Laboratory, 9700 South Cass Avenue, Argonne, Illinois 60439, United States ∥ School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332, United States S Supporting Information *
ABSTRACT: Atmospheric selenium (Se) in aerosols was investigated using X-ray absorption near-edge structure (XANES) spectroscopy and X-ray fluorescence (XRF) microscopy. These techniques were used to determine the oxidation state and elemental associations of Se in common primary emission sources and ambient aerosols collected from the greater Atlanta area. In the majority of ambient aerosol and primary emission source samples, the spectroscopic patterns as well as the absence of elemental correlations suggest Se is in an elemental, organic, or oxide form. XRF microscopy revealed numerous Se-rich particles, or hotspots, accounting on average for ∼16% of the total Se in ambient aerosols. Hotspots contained primarily Se0/Se(−II). However, larger, bulk spectroscopic characterizations revealed Se(IV) as the dominant oxidation state in ambient aerosol, followed by Se0/Se(−II) and Se(VI). Se(IV) was the only observed oxidation state in gasoline, diesel, and coal fly ash, while biomass burning contained a combination of Se0/Se(−II) and Se(IV). Although the majority of Se in aerosols was in the most toxic form, the Se concentration is well below the California Environmental Protection Agency chronic exposure limit (∼20000 ng/m3).
■
INTRODUCTION Selenium (Se) is known to have deleterious effects on human health, ranging from neurological conditions to pulmonary edema.1 The toxicity of Se is a function of both concentration1−4 and oxidation state.5 Approximately 16000 tons of total Se are cycled through the troposphere on a yearly basis.6 Fossil fuels are a major source of Se to the atmosphere.7,8 With projected increases in fossil fuel use,1 Se emissions are expected to substantially increase.5,9 Thus, Se associated with aerosols could have regional and global health repercussions through direct inhalation or indirect exposure via deposition. The concentration of Se in the atmosphere has been most commonly studied using inductively coupled plasma-mass spectroscopy (ICPMS).1,10,11 These studies have indicated that coal and biogenic emissions are important potential Se sources to the atmosphere.12−15 To date, aerosol Se studies have primarily focused on using Se as a tracer for coal fired power plants.5,9,16 Although Se concentration in the atmosphere has been previously studied,7,11,16−20 Se oxidation state has been largely unexplored. Traditionally, studies of Se oxidation state have used leaching procedures21−24 and have focused on aquatic and sedimentary environments.15,23−28 Study of Se oxidation state in atmospheric environments has © 2014 American Chemical Society
likely been hampered by the large sample mass required for the application of leaching techniques. The high sensitivity and uncomplicated sample preparation have made synchrotron-based spectroscopy and microscopy (hereafter termed “spectromicroscopy”) a powerful tool for the characterization of Se and other elements in atmospheric particulate matter.5,9,25−28 X-ray spectromicroscopy allows for the determination of oxidation state and speciation at micron spatial scales and provides information on elemental associations with Se. For atmospheric studies, X-ray spectromicroscopy has only been applied to the characterization of Se in coal and coal fly ash emissions.5,9,21,29 To the best of our knowledge, this is the first study characterizing Se oxidation state in ambient atmospheric aerosols as well as other noncoal associated primary emission sources. Received: Revised: Accepted: Published: 8988
January 22, 2014 July 21, 2014 July 30, 2014 July 30, 2014 dx.doi.org/10.1021/es500379y | Environ. Sci. Technol. 2014, 48, 8988−8994
Environmental Science & Technology
■
Article
METHODS Sample Collection. Aerosols with an aerodynamic diameter of less than 2.5 μm, or PM2.5, were collected from urban sites in Atlanta, GA, and from surrounding rural sites (Table 1). The three sampling sites, South DeKalb, Fire Station
Samples of selenocysteine (CAS 863394-07-4), sodium selenite (CAS 10102-18-8), and sodium selenate (CAS 13410-01-0) represent Se(−II), Se(IV), and Se(VI), respectively, and were analyzed as standard reference material. Se foil, run as an energy calibration (see below), provided spectroscopic data for elemental Se in the Se0 oxidation state. These standards contain the most common oxidation states of Se (Figure S1, Supporting Information). Standard materials were mounted on cellulose acetate filters, and approximate 0.5 cm × 0.5 cm sections of filters were mounted on aluminum supports. New Zefluor and cellulose acetate filters run as blanks did not produce a Se signal. Synchrotron-Based X-ray Spectromicroscopy. The ambient aerosols, primary emission sources, and Se standards were analyzed on the 2-ID-D beamline at the Advanced Photon Source at Argonne National Laboratory. Ambient aerosol and primary emission source samples were characterized using both X-ray fluorescence (XRF) microscopy and X-ray absorption near edge structure (XANES) spectroscopy. For XRF microscopy, the incident energy was 12.67 keV. Elemental content (atomic numbers from aluminum to Se) was mapped for selected samples. Filter areas averaging 117 × 117 μm were characterized for elemental content using a focused X-ray beam with a spot size of approximately 200 × 200 nm. Samples were mounted perpendicular to the beam. In order to maximize the number of samples analyzed in the allotted time, XRF maps were constructed by rastering the beam in 1 μm steps using a 1 s dwell time. This created a lower spatial resolution map, but relevant features were clearly discernible. Full fluorescence spectra were fit with modified Gaussian curves to obtain counts relevant for the mapping and spectral analyses. Elemental concentrations were calibrated with thin-film standards NBS1832 and NBS-1833 from the National Bureau of Standards (Gaithersburg, MD). XANES spectroscopy data were collected in two modes for ambient aerosols. First, XRF maps were used to identify hotspots. Hotspots are particles with a diameter of greater than approximately 1 μm with Se content that makes them readily identifiable in Se XRF maps. Hotspots were analyzed with a focused beam (200 nm spot size). While hotspots were an apparent source of Se on the filters, unresolvable particles also contributed to the Se loading on the filters. Large filter areas were examined with an unfocused beam (0.4 mm2 spot size) to obtain bulk spectra, capturing an average Se spectra for the scanned area. Primary emission sources and Se standards were characterized using bulk spectroscopy. XANES measurements were collected by scanning from 12.645 to 12.705 keV in 0.5 eV steps. The zone plate used for XANES measurements was 160 μm in diameter. The sample was in focus at 12.675 keV. Variation in the size of the focused beam that occurs during XANES scans was smaller than the diameter of hotspot particles analyzed. Thus, the focused beam stayed within the diameter of hotspot particles, eliminating a potential source of measurement artifacts.31 A range of dwell times were used, within practical limits, to maximize the number of fluorescence counts collected relative to the background. For regions with high Se concentration, several thousand Se counts could be obtained with a dwell time as low as 1 s. For areas with lower Se content, dwell times of up to 10 s were used to obtain XANES spectra. Dwell time and maximum count rates of spectroscopy are not expected to affect the peak position of the spectra.31 An energy-dispersive Si-drift detector (Vortex EM, with a 50 mm2 sensitive area, and a 12.5 um Be window; SII NanoTechnology,
Table 1. Ambient Aerosol Sample Collection Dates site
type
dates collected
Fort Yargo
rural
Fire Station 8
urban
South DeKalb
urban
9/17/2008 5/19/2009 6/10/2011 6/13/2011 6/16/2011 8/18/2011 6/14/2011 6/17/2011 8/11/2011 8/22/2011 8/24/2011 8/19/2011 8/24/2011
8, and Fort Yargo State Park, are part of the ongoing study Assessment of Spatial Aerosol and Composition in Atlanta.30 Samples were collected on Zefluor filters over 24-h periods at a flow rate of 16.7 L min−1 using cyclone inlet samplers (URG, Chapel Hill, NC). The multichannel particle samplers were mounted approximately 2.5 m off the ground. Six samples were collected from Fort Yargo State Park (33°58′4.18″N, 83°43′29.16″W), a rural setting that is frequently impacted by biomass burning plumes and power plant emissions. Two samples were collected in South DeKalb, a mixed commercial− residential area, approximately 8 km from a major interstate (33°41′16.48″N, 84°17′25.26″W). Five additional samples were collected from Fire Station 8, located in an industrial area within close proximity to a rail yard, a fire station, and an intersection with heavy diesel truck traffic (33°48′6.01″N, 84°26′8.75″W). PM2.5 samples were collected from gasoline and diesel emissions, biomass burning, and coal fly ash.26 Emissions from ultralow sulfur diesel fuel running a 10.8 L engine and conventional gasoline fuel running a 3.3 L engine were collected according to US Environmental Protection Agency protocols under typical urban driving conditions.26 Polydisperse coal fly ash, provided by The Southern Co., from an electrostatic precipitator of a midsized power plant was aerosolized and collected with a cyclone inlet sampler to separate the PM2.5 fraction.26 In an effort to produce fewer emissions that potentially react in the atmosphere, power plants sometimes use low-iron coal. Thus, low-iron and standard coal fly ash samples were analyzed in this study. Sampling was also conducted during a controlled biomass-burning experiment. Fuel consisted of materials from local coniferous and deciduous trees. A filter sampler with a PM2.5 cyclone inlet was placed 1 m above the burn area at a flow rate of 16.7 L min−1 for approximately 30 min.26 All primary emission source samples were collected on polytetrafluoroethylene (Zefluor) filters. All samples were stored in clean Petri dishes at −20 °C until analysis. Preparation for synchrotron analyses consisted of mounting an approximate 0.5 cm × 0.5 cm section of ambient aerosol and primary emission source filters on an aluminum support. 8989
dx.doi.org/10.1021/es500379y | Environ. Sci. Technol. 2014, 48, 8988−8994
Environmental Science & Technology
Article
Northridge, CA) was used to measure X-ray fluorescence of the samples under a helium environment. Absorbance spectra from the elemental Se metal foil were collected at the beginning and end of an experimental session, which confirmed that the energy calibration remained constant. Data Analysis. The concentration of Se in hotspots and the total Se concentration of each sample were determined using XRF maps and MAPS.32 The area densities from each pixel of the XRF maps were multiplied by pixel area and summed to determine Se content over a user-defined region of interest. This value was then normalized by the volume of air sampled using the average flow rate of the sampler; however, there is up to 20% variation in the flow rate due to uneven distribution of mass collected on each sample. Colocation of Se was investigated by plotting area densities of elements with atomic numbers from aluminum to arsenic against Se. Ten XRF maps of ambient aerosols and six XRF maps for the primary emission sources were analyzed. Following procedures typical of XANES studies, data was normalized to account for small variations in beam intensity that can occur during data collection; such variations in beam intensity were monitored with an ion chamber mounted upstream of the zone plate. After normalizing to the upstream ion chamber, the data was also processed using ten iterations of a three-point smoothing algorithm built into the software Athena.33 Finally, the data for an individual XANES spectrum were normalized to create a relative intensity value of approximately 1 for post edge area of the spectra. The postedge features of the reduced Se standards made determining relative contributions of each oxidation state in ambient samples difficult. For this reason typical XANES spectral processing methods such as linear combination fitting and Gaussian peak fitting and peak integrations were not used. Instead, the simplest approach of directly comparing the peak position of ambient aerosols and primary emission sources to pure Se standards was used.5,9,21,33 With this approach, we were able to identify the presence of a particular oxidation state within a sample but were unable to quantify the exact proportion of Se in different oxidation states.
Figure 1. Bulk ambient aerosol spectra. All of the XANES bulk spectra are shown and separated between sites: Fire Station 8 (blue), South DeKalb (orange), Fort Yargo (green). Black vertical solid lines indicate the spectral peak positions for different Se oxidation states.
■
RESULTS Ambient Aerosol. Bulk spectra from Fire Station 8 and Fort Yargo predominately exhibited peaks at 12.664 keV (Se(IV)) with both sites containing one spectrum exhibiting a peak at 12.662 keV (Se0/Se(−II)) (Figure 1). Some of the observed peaks in these samples were slightly below (0.5 eV) the energy typically observed in Se(IV) standards, indicating that a minor portion of Se within these samples is reduced (Se0/Se(−II)) (Figure 1). Half of the Fire Station 8 and Fort Yargo samples also exhibited a secondary peak at higher energy (12.668 keV), indicating the presence of Se(VI) (Figure 1). For two Fire Station 8 samples, the beam was moved to a different area of the same filter, examining a total of 0.8 mm2 of a sample. Spectra from different filter areas from the same sample exhibited similar spectral features. The South DeKalb spectra indicated the presence of Se0/Se(−II) with a secondary peak indicating the presence of Se(VI) (Figure 1). Spectra were also collected on a total of 48 hotspots from six Fire Station 8 samples, two South DeKalb samples, and two Fort Yargo samples (Figure 2). Hotspots accounted for ∼16% of total Se in the samples on average. Fire Station 8 had 17 hotspot spectra indicating Se0/Se(−II), one hotspot spectrum exhibiting a double peak at both Se(VI) and Se0/Se(−II), and
Figure 2. Hotspot ambient aerosol spectra. XANES spectra for each hotspot are presented and separated between Fire Station 8 (blue), South DeKalb (orange), and Fort Yargo (green). Black vertical solid lines indicate the spectral peak positions for different Se oxidation states.
five hotspots spectra indicating Se(IV) (Figure 2). Fort Yargo had 12 hotspots with peak positions indicating Se0/Se(−II). Of these 12, one hotspot also had a secondary peak indicating the presence of Se(VI). A total of 13 hotspot spectra were interrogated for South DeKalb, eight with peak positions 8990
dx.doi.org/10.1021/es500379y | Environ. Sci. Technol. 2014, 48, 8988−8994
Environmental Science & Technology
Article
indicating reduced Se0/Se(−II) and five with peaks indicating the prevalence of oxidized Se(IV) (Figure 2). Using XRF maps, which include Se in both hotspots and small unresolvable particles, elemental associations between Se and other elements (aluminum (Al), silicon (Si), phosphorus (P), sulfur (S), chlorine (Cl), potassium (K), calcium (Ca), titanium (Ti), vanadium(V), chromium (Cr), manganese (Mn), iron (Fe), cobalt (Co), nickel (Ni), copper (Cu), zinc (Zn), and arsenic (As)) were assessed using linear regressions and the corresponding coefficients of determination (R2) (Table S1, Supporting Information). Of the six XRF maps collected from Fire Station 8 samples, four showed moderate associations between Se and As (0.5 ≤ R2 ≤ 0.61, 95% confidence levels range from 0.022 to 0.0025). The remaining XRF maps from Fire Station 8 did not reveal any elemental associations for Se (R2 < 0.5). XRF maps of samples collected at Fort Yargo and South DeKalb also showed little association between Se and other elements (R2 < 0.5). Examination of elemental relations in only the hotspots revealed no significant correlations (0 < R 2 < 0.43) (Table S2, Supporting Information). Primary Emission Sources. Bulk Se in primary emission sources associated with fossil fuel combustion had average peak positions of 12.664 keV, indicating an oxidation state of Se(IV). Standard deviations on this average were ±0.0003 for coal fly ash, ± 0.0004 for gasoline, and ±0.0006 for diesel (Figure 3).
As (R2 = 0.75, 95% confidence level 0.021) and Co (R2 = 0.6, 95% confidence level 0.027) but less so with other elements ( R2 < 0.5). Se concentration in one of the standard coal fly ash samples was weakly correlated with Ca ( R2 = 0.52, 95% confidence level 0.021), while Se in the other sample was not correlated strongly with any element (R2 < 0.5). Se in biomass burning, diesel, and gasoline did not have any prevalent associations (R2 < 0.5). In analysis of hotspots, there was a strong correlation between Zn and Se (R2 = 0.85, 95% confidence level 0.0006) in the diesel exhaust. There were no other strong correlations observed for any other elements in hotspots examined in the primary emission source samples or ambient aerosols (R2 < 0.5). Aerosol Se Concentrations. Aerosol Se concentrations are presented in Table 2 and Figure 4. Fire Station 8 had a wide Table 2. Total and Hotspot Concentration (ng/m3) for Each Site and Emission Source. Median Values Are Interpolated for Each Data Set site Fire Station 8
Fort Yargo South DeKalb biomass coal fly ash
diesel gasoline
total conc (ng/m3)
median total conc (ng/m3)
2.29 29.6 3.94 3.54 4.32 4.32 8.92 1.95 2.72 1.52 12.7 10.5 17.4 15.4 6.18 1.51
4.13
5.43 2.12 12.7 15.4
6.18 1.51
hotspot conc (ng/m3) 0.02 1.33 0.10 0.61 0.36 0.21 6.21 0.03 0.22 0.02 6.64 3.96 3.42 0.54 0.22 0.74
median hotspot conc (ng/m3) 0.29
3.12 0.12 6.64 3.42
0.22 0.74
range of Se concentrations, ranging from 2.29 to 29.6 ng/m3. The Fire Station 8 sampler is located next to a rail yard and intersection with high traffic, which may result in high filter loadings at times. Interpolated median values for each site are used in the discussion because of the wide range seen in Se concentration. Of the sites, Fort Yargo had the highest median total concentration of Se at 5.43 ng/m3. Fire Station 8 and South DeKalb had total median Se concentrations of 4.13 and 2.12 ng/m3, respectively. Se concentration associated with hotspots was the highest in Fort Yargo, 3.12 ng/m3. Hotspots from Fire Station 8 and South DeKalb had lower median Se concentrations of 0.29 and 0.12 ng/m3, respectively.
Figure 3. Oxidation state of Se in primary emission sources. Average XANES spectra for each primary emission source are shown. Positions for Se oxidation states are indicated by vertical lines. Coal fly ash and gasoline produce the most oxidized Se, with a peak position reflective of Se(IV). In contrast, biomass burning and diesel both show peak positions between Se0/Se(−II) and Se(IV). None of the primary emission sources showed Se(VI).
Se(VI) has been seen in other XANES studies of coal fly ash;5,9 however, coal fly ash analyzed here only shows Se(IV). Biomass burning had a peak position of 12.663 keV (±0.0002) indicating a combination of Se0/Se(−II) and Se(IV). Three XRF maps were collected for coal fly ash. One XRF map was collected for each of the remaining primary emission sources. On the basis of previous studies, associations between Se and As, Fe, or S were anticipated.5−7,9,11,17,18,20,21,29,34−38 Associations between Se and S are based on proposed reactions between the two elements as well as expected similarities in chemical behavior of these two group VI elements.6,7,35−37 The average correlation between S and Se for all samples was R2 = 0.10. In the low Fe coal fly ash sample, Se was correlated with
■
DISCUSSION In the majority of bulk ambient aerosol samples, the main Se XANES peak position (average ∼12.6634 keV) is shifted slightly below the Se(IV) peak. This suggests that gasoline exhaust, diesel exhaust, and coal fly ash, which all have peaks at higher energies (i.e., predominantly Se(IV)), are not the sole sources of Se to ambient aerosols (Figure 5). Biomass burning has a similar Se XANES peak position to bulk ambient aerosols and could, thus, account for the oxidation state observed in bulk ambient aerosols. However, it is unlikely that biomass8991
dx.doi.org/10.1021/es500379y | Environ. Sci. Technol. 2014, 48, 8988−8994
Environmental Science & Technology
Article
Figure 4. Concentration bar graph. Median bulk and hotspot Se concentrations (ng/m3) are shown for each site and emission source.
locations, especially in areas impacted by mobile sources and biomass burning. The differences in the elemental associations and Se oxidation state between potential primary emission sources and ambient aerosols make it challenging to apportion Se in ambient aerosols to specific sources and suggests that atmospheric processing plays a key role in the final composition of primary Se emissions. Additionally, the featureless nature of Se XANES spectra, indicative of oxides, makes it challenging to use spectral features, and therefore composition, for source apportionment. Two bulk ambient aerosol samples, out of 15 analyzed, were dominated by the Se0/Se(−II) oxidation state, which cannot be explained by contributions from any of the primary emission sources analyzed (Figure 5). The presence of Se0/Se(−II) in these samples may be explained by atmospheric processing. The reduction of selenium dioxide coupled to the oxidation of sulfur dioxide has been proposed as a mechanism to produce Se0 in aerosols.36,42,43 In a previous study this mechanism was dismissed due to lack of evidence for Se(IV) in aerosols.42 Our results indicate the presence of Se(IV) in both primary emission sources and ambient aerosols; thus, this mechanism is plausible. A correlation in the XRF data between Se and S might be expected given the proximity (8−35 km) of coal-fired power plants to the sampling sites.6,7,18,34−37 The lack of associations observed between Se and S in hotspots or bulk ambient aerosols may reflect differences in the chemical behavior of Se and S during transport. When SO2 and Se are emitted as gas in the combustion process, Se quickly condenses as particles, within kilometers of the source.1,6 Whereas SO2 is very slowly oxidized in the gas phase, which typically results in an increase in sulfate concentrations with distance from the source.19,44 Such processes can confound correlations making them difficult to attribute to specific sources. Another explanation for Se0/Se(−II) observed in bulk samples and hotspots could be a contribution from an unknown source such as volatile organic Se compounds. Microorganisms and plants are thought to emit volatile organic Se compounds such as DMSe (dimethyl selenide) and DMDSe (dimethyl diselenide).1,3,6,17 Volatile organic Se compounds can react with hydroxyl and nitrate radicals in the atmosphere to form aerosol Se0/Se(−II).1,6,45 The heavy vegetation in the Atlanta region could be a potential source of such volatile organic Se compounds. In summary, the main peak position of ambient aerosols indicating an oxidation state slightly more
Figure 5. Oxidation state of Se in literature, primary emission sources, and ambient aerosol. Box and whisker plot depicts the range of oxidation states in primary emission sources and ambient aerosols. The solid black line in each box represents the mean value, boxes represent data contained within one standard deviation, whiskers represent data contained within two standard deviations, and open circles represent data outside two standard deviations of the mean. Dark gray, light gray, and dotted boxes represent primary emission sources, ambient aerosols, and literature values for coal fly ash,5,9,42,46,47 respectively. Several ambient aerosol samples contained multiple oxidation states, which are both represented in this figure.
burning emissions are a major contributor to ambient aerosol Se because PM2.5 samples collected in Atlanta contain on average only a 7% contribution from biomass burning emissions.39 Additionally, biomass burning is typically more prevalent in the winter and spring in the Atlanta region, and the samples were collected in the summer months.40 Se is commonly linked to coal-fired power plant emissions.7,12−16,41 In a few ambient samples (Fire Station 8), Se was associated with As, another power plant tracer, providing evidence that power plants could be a source of Se to ambient aerosol in this study. However, the detection of Se in other primary emission sources including gasoline, biomass burning, and diesel emissions (Figure 4) demonstrate Se may not be a good tracer alone, for coal burning emissions in some 8992
dx.doi.org/10.1021/es500379y | Environ. Sci. Technol. 2014, 48, 8988−8994
Environmental Science & Technology
Article
1357375. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. Use of the Advanced Photon Source is supported by the U.S. Department of Energy, Office of Basic Energy Sciences, under contract No. DE-AC02-06CH11357. We thank John Jansen at Southern Co. and Bill Preston at the EPA for providing primary emission source samples.
reduced than that of Se(IV) as well as the presence of Se0/ Se(−II) in some samples could be attributed to (1) an unlikely dominance of aerosol supply from biomass burning; (2) reactions in the atmosphere leading to the reduction of Se(IV); or (3) a mixture of fossil fuel derived aerosol with a source of reduced selenium Se0/Se(−II) such as aerosol particles derived from volatile organic Se. Several bulk spectra also had a second peak at 12.668 keV indicating the presence of Se(VI) (Figure 1). The peak for Se(VI) is not observed in any of the primary emission source spectra (Figure 3). Oxidative atmospheric processes are a likely mechanism to transform Se(IV) in ambient aerosols to Se(VI).42 Furthermore, it is reasonable to assume that oxidative processes could also eventually transform Se0/Se(−II) observed in both bulk and hotspot spectra to Se(IV) and Se(VI). The observation of Se(VI) in some ambient aerosol samples suggests that oxidative processes in the atmosphere may work to reduce Se toxicity. In the majority of samples hotspots were comprised of Se0/Se(−II) (Figure 2). The prevalence of Se0/Se(−II) in the larger hotspot particles may be a function of their low surface area to volume ratio making them less prone to oxidation relative to smaller submicron particles. The most toxic oxidation state of Se, +IV, was the most prevalent oxidation state in bulk ambient aerosol. Typical values range from 1 to 10 ng/m3 of Se in the atmosphere,6 and Se concentrations for the ambient aerosols analyzed in this study fall within this range (Table 2). The Environmental Protection Agency has not released atmospheric limits for Se, but the California Environmental Protection Agency has set chronic exposure limits as low as 20000 ng/m3 for Se-containing compounds. Even though the majority of Se in bulk ambient aerosols was in the most toxic oxidation state, fortunately total Se concentrations in all samples were 1000 times under recommended exposure limits.
■
■
(1) Amouroux, D.; Liss, P. S.; Tessier, E.; Hamren-Larsson, M.; Donard, O. F. X. Role of oceans as biogenic sources of selenium. Earth Planet. Sci. Lett. 2001, 189 (3−4), 277−283. (2) Davis, W. C.; Jin, F. X.; Dempster, M. A.; Robichaud, J. L.; Marcus, R. K. Development of a new liquid chromatography method for the separation and speciation of organic and inorganic selenium compounds via particle beam-hollow cathode glow discharge-optical emission spectroscopy. J. Anal. At. Spectrom. 2002, 17 (2), 99−103. (3) Ham, Y. S.; Tamiya, S. Selenium behavior in open bulk precipitation, soil solution and groundwater in alluvial fan area in Tsukui, Central Japan. Water Air Soil Pollut. 2006, 177 (1−4), 45−57. (4) Hamilton, S. J. Review of selenium toxicity in the aquatic food chain. Sci. Total Environ. 2004, 326 (1−3), 1−31. (5) Shah, P.; Strezov, V.; Prince, K.; Nelson, P. F. Speciation of As, Cr, Se and Hg under coal fired power station conditions. Fuel 2008, 87 (10−11), 1859−1869. (6) Wen, H. J.; Carignan, J. Reviews on atmospheric selenium: Emissions, speciation and fate. Atmos. Environ. 2007, 41 (34), 7151− 7165. (7) Ellis, W. G.; Arimoto, R.; Savoie, D. L.; Merrill, J. T.; Duce, R. A.; Prospero, J. M. Aerosol Selenium at Bermuda and Barbados. J. Geophys. Res.-Atmos. 1993, 98 (D7), 12673−12685. (8) Mosher, B. W.; Duce, R. A.; Prospero, J. M.; Savoie, D. L. Atmospheric selenium - Geographical-distribution and ocean to atmosphere flux in the Pacific. J. Geophys. Res.-Atmos. 1987, 92 (D11), 13277−13287. (9) Shah, P.; Strezov, V.; Stevanov, C.; Nelson, P. F. Speciation of arsenic and selenium in coal combustion products. Energy Fuels 2007, 21 (2), 506−512. (10) Wallschlager, D.; London, J. Determination of inorganic selenium species in rain and sea waters by anion exchange chromatography-hydride generation-inductively-coupled plasma-dynamic reaction cell-mass spectrometry (AEC-HG-ICP-DRC-MS). J. Anal. At. Spectrom. 2004, 19 (9), 1119−1127. (11) Chiaradia, M.; Cupelin, F. Gas-to-particle conversion of mercury, arsenic and selenium through reactions with traffic-related compounds (Geneva)? Indications from lead isotopes. Atmos. Environ. 2000, 34 (2), 327−332. (12) Husain, L.; Parekh, P. P.; Dutkiewicz, V. A.; Khan, A. R.; Yang, K.; Swami, K. Long-term trends in atmospheric concentrations of sulfate, total sulfur, and trace elements in the northeastern United States. J. Geophys. Res.-Atmos. 2004, 109, D18. (13) Kowalczyk, G. S.; Choquette, C. E.; Gordon, G. E. Chemical element balances and identification of air-pollution sources in Washington, Dc. Atmos. Environ. 1978, 12 (5), 1143−1153. (14) Vossler, T. L.; Lewis, C. W.; Stevens, R. K.; Dzubay, T. G.; Gordon, G. E.; Tuncel, S. G.; Russwurm, G. M.; Keeler, G. J. Composition and origin of summertime air-pollutants at Deep-Creek Lake, Maryland. Atmos. Environ. 1989, 23 (7), 1535−1547. (15) Dutkiewicz, V. A.; Husain, L. Spatial pattern of non-urban Se concentrations in the Northeastern United States and its pollution source implications. Atmos. Environ. 1988, 22 (10), 2223−2228. (16) Chow, J. C.; Watson, J. G.; Kuhns, H.; Etyemezian, V.; Lowenthal, D. H.; Crow, D.; Kohl, S. D.; Engelbrecht, J. P.; Green, M. C. Source profiles for industrial, mobile, and area sources in the Big Bend Regional Aerosol Visibility and Observational study. Chemosphere 2004, 54 (2), 185−208.
ASSOCIATED CONTENT
S Supporting Information *
Elemental associations between Se and other elements for bulk aerosols and hotspots in both primary emission sources and studied sites as well as spectral data for standards. Table S1 shows the corresponding coefficients of determination (R2) for each element in the primary emission sources and ambient aerosols for both the bulk spectra and hotspots. Figure S1 depicts the spectral data of the Se standards. Figure S2 shows scatter plots generated for the high coefficients of determination (R2). This material is available free of charge via the Internet at http://pubs.acs.org.
■
REFERENCES
AUTHOR INFORMATION
Corresponding Author
*Tel: 404-894-3383. Fax: 404-894-5638. E-mail: ingall@eas. gatech.edu. Present Address ⊥
Environmental Protection Agency, National Center of Environmental Assessment, Research Triangle Park, NC 27711.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This material is based upon work supported by the National Science Foundation under Grant Nos. OCE 1060884 and OCE 8993
dx.doi.org/10.1021/es500379y | Environ. Sci. Technol. 2014, 48, 8988−8994
Environmental Science & Technology
Article
(17) Arimoto, R.; Duce, R. A.; Savoie, D. L.; Prospero, J. M. Traceelements in aerosol-particles from Bermuda and Barbados Concentrations, sources and relationships to aerosol sulfate. J. Atmos. Chem. 1992, 14 (1−4), 439−457. (18) Beceiro-Gonzalez, E.; Gonzalez-Soto, E.; Lopez-Mahia, P.; Prada-Rodriguez, D. Total arsenic and selenium levels in atmospheric particulate matter of La Coruna (Spain). Sci. Total Environ. 1997, 208 (3), 207−211. (19) Ross, H. B. An atmospheric selenium budget for the region 30 degrees N to 90 degrees N. Tellus, Ser. B 1985, 37 (2), 78−90. (20) Waldman, J. M.; Lioy, P. J.; Zelenka, M.; Jing, L.; Lin, Y. N.; He, Q. C.; Qian, Z. M.; Chapman, R.; Wilson, W. E. Wintertime measurements of aerosol acidity and trace-elements in Wuhan, a city in central China. Atmos. Environ., Part B 1991, 25 (1), 113−120. (21) Seames, W. S.; Sooroshian, J.; Wendt, J. O. L. Assessing the solubility of inorganic compounds from size-segregated coal fly ash aerosol impactor samples. J. Aerosol Sci. 2002, 33 (1), 77−90. (22) Hutchinson, T. J.; Basappa, L.; Dikshit, A.; Luo, Y.; Catalano, J. G.; Giammar, D. E. Fate of metals in fly ash during aging in laboratoryscale ash impoundments. Environmental Engineering Science 2012, 29 (12), 1085−1091. (23) Saitanis, C. J.; Frontasyeva, M. V.; Steinnes, E.; Palmer, M. W.; Ostrovnaya, T. M.; Gundorina, S. F. Spatiotemporal distribution of airborne elements monitored with the moss bags technique in the Greater Thriasion Plain, Attica, Greece. Environ. Monit. Assess. 2013, 185 (1), 955−968. (24) Steinnes, E.; Allen, R. O.; Petersen, H. M.; Rambaek, J. P.; Varskog, P. Evidence of large scale heavy-metal contamination of natural surface soils in Norway from long-range atmospheric transport. Sci. Total Environ. 1997, 205 (2−3), 255−266. (25) Longo, A. F.; Ingall, E. D.; Diaz, J. M.; Oakes, M.; King, L. E.; Nenes, A.; Mihalopoulos, N.; Violaki, K.; Avila, A.; Benitez-Nelson, C. R.; Brandes, J.; McNulty, I.; Vine, D. J. P-NEXFS analysis of aerosol phosphorus delivered to the Mediterranean Sea. Geophys. Res. Lett. 2014, 41 (11), 2014GL060555. (26) Oakes, M.; Ingall, E. D.; Lai, B.; Shafer, M. M.; Hays, M. D.; Liu, Z. G.; Russell, A. G.; Weber, R. J. Iron Solubility Related to Particle Sulfur Content in Source Emission and Ambient Fine Particles. Environ. Sci. Technol. 2012, 46 (12), 6637−6644. (27) Oakes, M.; Weber, R. J.; Lai, B.; Russell, A.; Ingall, E. D. Characterization of iron speciation in urban and rural single particles using XANES spectroscopy and micro X-ray fluorescence measurements: investigating the relationship between speciation and fractional iron solubility. Atmos. Chem. Phys. 2012, 12 (2), 745−756. (28) Schroth, A. W.; Crusius, J.; Sholkovitz, E. R.; Bostick, B. C. Iron solubility driven by speciation in dust sources to the ocean. Nat. Geosci. 2009, 2 (5), 337−340. (29) Seames, W. S.; Wendt, J. O. L. Partitioning of arsenic, selenium, and cadmium during the combustion of Pittsburgh and Illinois #6 coals in a self-sustained combustor. Fuel Process. Technol. 2000, 63 (2− 3), 179−196. (30) Butler, A. J.; Andrew, M. S.; Russell, A. G., Daily sampling of PM2.5 in Atlanta: results of the first year of the assessment of spatial aerosol composition in Atlanta study. J. Geophys. Res.-Atmos. 2003, 108, (D1). (31) Ingall, E.; Diaz, J.; Longo, A.; Oakes, M.; Finney, L.; Vogt, S.; Lai, B.; Yager, P. L.; Twining, B. S.; Brandes, J. Role of biogenic silica in the removal of iron from Antarctic Seas. Nature Commun. 2013, 4, 1981. (32) Vogt, S. MAPS: A set of software tools for analysis and visualization of 3D X-ray fluorescence data sets. J. Phys. IV 2003, 104, 635−638. (33) Ravel, B.; Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT. J. Synchrotron Radiat. 2005, 12, 537−541. (34) Atkinson, R.; Aschmann, S. M.; Hasegawa, D.; Thompsoneagle, E. T.; Frankenberger, W. T. Kinetics of the atmospherically important reactions of dimethyl selenide. Environ. Sci. Technol. 1990, 24 (9), 1326−1332.
(35) Ball, S.; Milne, J. Studies on the interaction of selenite and selenium with sulfur donors 0.3. Sulfite. Can. J. Chem. 1995, 73 (5), 716−724. (36) Ross, H. B. An atmospheric selenium budget for the region 30degrees-N to 90-degrees-N. Tellus Ser. B 1985, 37 (2), 78−90. (37) Bronikowski, T.; Pasiuk-Bronikowska, W.; Ulejczyk, M.; Nowakowski, R. Interactions between environmental selenium and sulphoxy radicals. J. Atmos Chem. 2000, 35 (1), 19−31. (38) Abad-Valle, P.; Lopez-Anton, M. A.; Diaz-Somoano, M.; Juan, R.; Rubio, B.; Garcia, J. R.; Khainakov, S. A.; Martinez-Tarazona, M. R. Influence of iron species present in fly ashes on mercury retention and oxidation. Fuel 2011, 90 (8), 2808−2811. (39) Kim, E.; Hopke, P. K.; Edgerton, E. S. Improving source identification of Atlanta aerosol using temperature resolved carbon fractions in positive matrix factorization. Atmos. Environ. 2004, 38 (20), 3349−3362. (40) Lee, D.; Balachandran, S.; Pachon, J.; Shankaran, R.; Lee, S.; Mulholland, J. A.; Russell, A. G. Ensemble-trained PM2.5 source apportionment approach for health studies. Environ. Sci. Technol. 2009, 43 (18), 7023−7031. (41) Rahn, K. A.; Lowenthal, D. H. Pollution aerosol in the Northeast-Northeastern-Midwestern contributions. Science 1985, 228 (4697), 275−284. (42) Andren, A. W.; Klein, D. H.; Talmi, Y. Selenium in coal-fired steam plant emissions. Environ. Sci. Technol. 1975, 9 (9), 856−858. (43) Husain, L. A technique for determining in-cloud formation of So4. Geophys. Res. Lett. 1989, 16 (1), 57−60. (44) Glarborg, P.; Marshall, P. Mechanism and modeling of the formation of gaseous alkali sulfates. Combust. Flame 2005, 141 (1−2), 22−39. (45) Haygarth, P. M.; Jones, K. C.; Harrison, A. F. Selenium cycling through agricultural grasslands in the UK - Budgeting the role of the atmosphere. Sci. Total Environ. 1991, 103 (1), 89−111. (46) Huggins, F. E.; Senior, C. L.; Chu, P.; Ladwig, K.; Huffman, G. P. Selenium and arsenic speciation in fly ash from full-scale coalburning utility plants. Environ. Sci. Technol. 2007, 41 (9), 3284−3289. (47) Niss, N. D.; Schabron, J. F.; Brown, T. H. Determination of selenium species in coal fly-ash extracts. Environ. Sci. Technol. 1993, 27 (5), 827−829.
8994
dx.doi.org/10.1021/es500379y | Environ. Sci. Technol. 2014, 48, 8988−8994
