Direct Identification of Trace Metals in Fine and ... - ACS Publications

Mar 17, 2004 - University of Michigan, 109 Observatory Street, 2518 SPH I,. Ann Arbor ... Exposure to airborne particulates containing low con- centra...
0 downloads 0 Views 1MB Size
Download PDF
Research Direct Identification of Trace Metals in Fine and Ultrafine Particles in the Detroit Urban Atmosphere SATOSHI UTSUNOMIYA,§ KELD A. JENSEN,† GERALD J. KEELER,‡ AND R O D N E Y C . E W I N G * ,§ Department of Geological Sciences, University of Michigan, Ann Arbor, Michigan 48109-1063, National Institute of Occupational Health, Lersø Parkalle´ 105, DK-2100 Copenhagen, Denmark, and Department of Environmental Health Sciences, School of Public Health, University of Michigan, 109 Observatory Street, 2518 SPH I, Ann Arbor, Michigan 48109-2029

Exposure to airborne particulates containing low concentrations of heavy metals, such as Pb, As, and Se, may have serious health effects. However, little is known about the speciation and particle size of these airborne metals. Fine- and ultrafine particles with heavy metals in aerosol samples from the Detroit urban area, Michigan, were examined in detail to investigate metal concentrations and speciation. The characterization of individual particles was completed using high-angle annular dark-field scanning transmission electron microscopy (HAADFSTEM) combined with conventional high-resolution TEM techniques. The trace elements, Pb, As, La, Ce, Sr, Zn, Cr, Se, Sn, Y, Zr, Au, and Ag, were detected, and the elemental distributions were mapped in situ at the nanoscale. The crystal structures of the particles containing Pb, Sr, Zn, and Au were determined from their electron diffraction patterns. Based on the characterization of the representative trace element particles, the potential health effects are discussed. Most of the trace element particles detected in this study were within a range of 0.01-1.0 µm in size, which has the longest atmospheric residence time (∼100 days). Increased chemical reactivity owing to the size of nanoparticles may be expected for most of the trace metal particles observed.

Introduction Epidemiological studies have reported an increase in respiratory symptoms, hospitalization, and cardiovascular diseases in humans with an increase in particulate atmospheric air pollution (1-4). Long-term exposure to certain trace metals (“air toxics”) can also result in pathogenic effects ranging from learning disabilities, development of respiratory inflammation to cancer, and damage of vital organs (5-10). Several particle-related parameters (e.g., particle size, chemical composition, and solubility) have been found to * Corresponding author phone: (734)647-8529; fax: (734)647-8531; e-mail: [email protected]. § Geological Sciences, University of Michigan. † National Institute of Occupational Health. ‡ Department of Environmental Health Sciences, University of Michigan. 10.1021/es035010p CCC: $27.50 Published on Web 03/17/2004

 2004 American Chemical Society

play an important role in the development adverse health effects from air pollution. For example, in vivo experiments showed a ∼10-fold increase in inflammation after inhalation of the same mass-dose of ultrafine (e100 nm) carbon black, TiO2 and latex particles as compared to the effect from particles larger than 200 nm (11, 12). The increase in biological response induced by these insoluble compounds may be linked to specific surface parameters (13). Several studies have also shown that the composition of the particles and their solubility plays an important, but still unclarified, role on the induction of inflammatory proteins and cytoxicity (14-16). Specifically, the effects of soluble polyvalent transition metals have received much attention, because through redox reactions and formation of free radicals, they can induce oxidative stress at the subcellular level (17-20). However, an in vitro study using nontoxic and partially soluble particles (e.g., Mg-, Ca-, and Ba-sulfates) suggests that the presence of particles alone plays an important role, because these phases induce a systematic inflammatory response beginning at concentrations around their solubility-limit in the cell media (16). Owing to their toxicity or carcinogenecity, adverse health effects are also associated with the exposure to certain trace and heavy metals in air pollution (9, 10, 21-23). Currently, 10 metal compounds are listed among the 188 hazardous air pollutant substances (“HAPS” or “air toxics”) defined under the Clean Air Act Amendments of 1990. These metal substances include compounds of As (0.03 µg/m3), Be (0.02 µg/m3), Cd (0.02 µg/m3), Cr6+ (0.008-0.1 µg/m3), Co (0.1 µg/ m3), Pb (1.5 µg/m3), Mn (0.05 µg/m3), Hg (0.09 µg/m3) Ni (0.1-0.2 µg/m3), and Se (0.08-20.0 µg/m3) along with the chemically complex diesel engine emissions (5 µg/m3). The concentrations in parentheses are Inhalation reference Concentrations (RfC) compiled by the U.S. EPA Office of Air Quality Planning and Standards (24). Emission of heavy and trace elements into the atmosphere can also affect the environment; either directly by their ecological toxicity or indirectly through bio- or geochemical accumulation that potentially can result in oral exposure through the food web (25, 26). However, in contrast to human health effects from inhaled particles and air toxics, only soluble compounds can be ecotoxic, because only the soluble fraction readily enters to the hydrobiochemical cycle (9, 25, 27, 28). Hence, to assess their potential impact on human health, as well as their ecotoxicology and dispersion into the environment, it is important to identify the speciation of air toxic compounds and their specific physicochemical parameters. If toxic trace elements are homogeneously dispersed as impurities in insoluble larger-size particles, risks to the health and environment are less than if they occur as major constituents in individual, trace-metal, nanoscale particles. However, if the heavy metals occur in ultrafine particles, the size will influence their reactivity, toxicity, and their fate in the ambient environment (29, 30). Ultrafine particles are known to have increased solubility, as compared to larger-size particles of the same composition because of the increased surface-to-volume ratio for smaller particle sizes (30). Previous analyses have shown that trace elements in airborne particulates by mass are most abundant in the 0.5-1 µm-size fraction (31). However, the bulk composition of particles in urban areas varies depending on the local emission sources. Ultrafine particles collected in Southern VOL. 38, NO. 8, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2289

California (1995-1997) on average contained 50 wt % organic carbon, 8.7 wt % elemental carbon, 8.2 wt % sulfate, 6.8 wt % nitrate, 0.6 wt % sodium, 0.5 wt % chloride, 8.7 wt % ammonium, and 14 wt % trace metal oxides (32, 33). Thus the fraction of total trace metal compounds can be significant in the ultrafine particle fraction. Although the trace metal compounds in the studies mentioned above were classified as trace metal oxides, the metal-bearing phases were actually not identified (32). Bulk chemical analysis of trace elements including heavy toxic elements in atmospheric particulates have been completed in numerous studies using various techniques including ICP-MS, -AES, XRF, and INAA, mainly to evaluate the extent of air pollution and the source of the pollutant (34-39). Aerosols larger than 1 µm have been characterized in several studies using analytical scanning electron microscopy (SEM) (40). In these analyses trace elements have often been detected without the possibility of further identification (4145). Determination of the oxidation state and general speciation of several air toxics have been conducted recently using X-ray Absorption Fine Structure (XAFS) on a PM10 sample, combustion ashes, and NIST Standard Reference Materials for urban air and diesel soot, respectively (46, 47). However, a multiparametric characterization of size, chemical composition, and atomic structure of individual ultrafine particles can be conducted only by transmission electron microscopy (TEM), because this technique has a high spatial resolution at the atomic scale (48). In fact, the major aerosol compounds have already been characterized in samples from several sites using TEM technique (e.g., refs 49-55). Other TEM techniques, electron energy loss spectrometry (EELS), scanning TEM (STEM) combined with an EDX (Energy Dispersive X-ray) spectrometry have been recently applied for characterizing ultrafine aerosols (56-59). However, heavy trace elements have been rarely detected using TEM, EELS, and STEM, because of their low concentration and the dispersed occurrence of the constituent species. Recently nanosize uraninite, UO2, enclosed in soot particles from Detroit was successfully characterized using high-angle annular dark field (HAADF) STEM (60). In a HAADF-STEM, the image is formed by an incoherent scattering process, and the contrast is correlated to the atomic mass and specimen thickness: heavier elements as well as thick areas have a brighter contrast. Thus, HAADF-STEM is an extremely useful tool for detecting ultrafine heavy metal particles (61). In this paper, we have characterized individual fine and ultrafine “trace” heavy metal particles in a sample from the urban atmosphere of Detroit. We focus primarily on identification of ultrafine heavy metal particles and discuss the results with respect to their potential health effects.

Experimental Methods A 24-h Total Suspended Particulate (TSP) sample (LV1-149) was collected from southwest Detroit, MI, U.S.A. on August 29, 1996 at 30 L/min using an open-face Teflon filter pack mounted with 47 mm PTFE membrane filters with a 2 µm pore size (Gelman R2PJ047). The volume of sampled air was m3 at 25 °C, measured directly using calibrated dry gas meters. The PTFE membrane filter were first analyzed at the U.S. EPA National Exposure Research Laboratory using X-ray Fluorescence (XRF). Trace metal concentrations were analyzed using a Perkin-Elmer Elan 5000A Inductively Coupled Plasma Mass Spectrometry (ICP-MS). The samples were introduced into the ICP-MS by pneumatic nebulization after extraction from the filters in nitric acid. Prior to extraction in nitric acid, particle samples were retrieved from the filters for characterization using Field Emission SEM (FE-SEM: Philips XL-30) and high-resolution TEM (HRTEM: JEOL 2010F). Nanoscale elemental mapping on nanocrystals containing heavy metals was conducted 2290

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 8, 2004

FIGURE 1. SE image of typical aerosol from Detroit, MI by FE-SEM.

TABLE 1. Bulk Chemical Composition of the Aerosol from Detroit (LV1-149) (ng/m3)a data analyzed by XRF

Al Si P S Cl K Ca Ti Fe Ni Zr Sn Au Y Cd Pb V

data analyzed by ICP-MS

concentration

EF

861 (29) 3000 (29) 37 (34) 1150 (7) 163 (13) 313 (12) 2630 (13) 86 (23) 3080 (16) 8.2 (13) 2.8 (65) 0.69 (247) ud ud 0.49 (244) 25.0 (10.1) 8.23 (32)

1.0 1.0 3.3 416 119 1.1 6.8 1.8 5.8 10.3 1.6 32.7

Hg Mg V Cr Mn Co Cu Zn As Rb Sr Mo Ag Cd Sb Ba La Ce Nd Sm W Pb

concentration

EF

0.0968 (3.7) 439 (4.9) 8.58 (3.1) 2.73 (3) 45.2 (3) 0.32 (0.9) 13.7 (1.7) 103 (1.8 2.07 (3.6) 0.36 (2.4) 9.59 (0.6) 1.05 (3.8) 0.048 (5.4) 0.45 (0.8) 1.81 (2.6) 30.7 (1.1) 0.53 (4.3) 0.94 (2.8 0.39 (1.7) 0.071 (6.6) 0.24 (9.4) 21.5 (1.9)

114 2.0 6.0 2.6 4.5 1.2 23.6 139 109 0.4 2.4 66.1 64 213 856 6.8 1.7 1.5 1.3 1.1 15.1 156

a The values in parentheses represent analytical standard deviation (%). ud stands for under detection limit.

using HAADF-STEM with an EDX mapping system (Emispec, ES Vision ver. 4.0) (61). TEM specimens were prepared by dispersing the samples from the filters onto holey carbon grids. Before STEM analysis, the TEM specimen holder was cleaned with a plasma (Fischione Model C1020) to minimize contamination. A drift correction system was used for the STEM-EDX mapping.

Results and Discussion The major mineral phases in the aerosol sample consisted of gypsum, sodium chloride, iron oxides, clay minerals, and quartz. Figure 1 shows the typical appearance of the sample using the field emission gun (FE)-SEM. Most of the spherical particles are iron oxides. The angular particles are gypsum. Sheetlike particles are clay minerals. Table 1 gives the bulk chemical composition of the samples as analyzed by XRF and ICP-MS. Three elements, Cd, Pb, and V, are listed in both tables to show the consistency of the results. The data analyzed by ICP-MS have smaller uncertainties for the trace elements. The results show that Mn (45.2 ng Mn/m3 or 58.4 ng MnO/m3) is the only metal that is prone to exceed the current U.S. EPA annual RfCvalues, which is 50 ng/m3 for Mn-compounds (24). Table 1

FIGURE 3. HAADF-STEM image and the elemental map (the red square) of a Pb-particle and REE-bearing particles attached to mica.

FIGURE 2. (a) HAADF-STEM image of Pb-bearing particle on the surface of carbonaceous matter with elemental maps. (b) HRTEM image shows aggregates of nanocrystallites. (c) HRTEM image and the FFT image showing the basis for the identification of the phase as PbO. also lists the Al-normalized enrichment factor (EF ) (CX/ CAl)atmosphere/(CX/CAl)crust) calculated for each element (X) as compared with the average composition of the crust according to the method introduced by Duce et al. (62). The high abundance of the major elements, Si, S, Cl, Ca, Al, K, Mg, and Fe, is consistent with the abundance of the major aerosol phases mentioned above. The EFs of trace metals, Sn, Hg, Zn, As, Mo, Ag, Cd, Sb, and Pb, were significantly higher (30 < EF < 860) than that of the major metals: Si (EF ) 1.0), Ca (EF ) 6.8), Al (reference element), K (EF ) 1.1), Mg (EF ) 2.0), and Fe (EF ) 5.8). The enrichment in Ni (EF ) 10.3) and Cu (EF ) 23.6) was moderate, whereas the enrichment in Cr (EF ) 2.6) was minor. As is normally observed, there was also a high enrichment in S (EF ) 416) and Cl (EF ) 119). High EF values of heavy metals are often reported (36, 37) and are usually interpreted as the result of anthropogenic activities or exceptional geological events such as volcanic eruptions. HAADF-STEM and HRTEM were utilized to obtain further information on the speciation of the trace metal compounds and their potential association with larger-size particles. Fine and ultrafine particles with Pb, As, La, Ce, Sr, Zn, Cr, Se, Sn, Y, Zr, Au, and Ag were observed and characterized. Lead. Figure 2a shows a HAADF-STEM image and elemental maps of the area outlined by the red square. The bright contrast in the image corresponds to the presence of heavy elements, in this case a ∼50 nm-size Pb-rich domain on the surface of carbonaceous matter. HRTEM-imaging showed that the Pb-rich grain consisted of 100 times higher than the value in this study (21.5 ng/m3), primarily due to the use of leaded gasoline (65). The high concentrations of Pb caused serious problems in the central nervous system, peripheral nervous system, and the vascular system of humans (66, 67). The removal of Pb from gasoline dramatically decreased Pb-concentrations by the mid 1980s in the atmosphere and in human blood (68). The speciation of Pb in the atmosphere is crucial with respect to its potential health effects. Generally, soluble and organic Pb-compounds are considered more toxic than insoluble Pb-compounds (69). In the mid 1970s, several traffic-related Pb-compounds in the atmosphere were identified using X-ray diffraction analysis of density separated heavy particles (70). The identified Pb-phases included (NH4)SO4‚PbSO4, PbSO4, PbBrCl, and 2PbBrCl in the sizerange < 2.1 µm. However, Pb-sulfates and Pb-Br-chlorides are known to rapidly degrade photocatalytically to Pb-oxides and Pb-carbonates in the atmosphere (71). The identification of the Pb-sulfates and Br-chlorides occurred before Pbremoval from gasoline was completed. There appears to be a lack of recent data on the speciation of Pb in the “current” atmosphere. Recent attempts to determine the speciation of Pb in atmospheric samples (NIST SRM 1649a and an authentic PM10 filter sample) and a heavy-duty diesel sample (NIST SRM 1648) using XAFS provided systematic data of speciation for each element, but the results are still ambiguous in terms of the phase characterization (46, 47). Both PbO (CAS no. 1317-36-8) and lead arsenate (CAS no. 7784-40-9) are listed in the Hazardous Substance Data Bank (69). PbO is classified as a potential carcinogenic compound, and lead arsenates are confirmed to be carcinogenic. The current RfC-value for Pb-compounds in the United States is 1.5 µg/m3, and the U.S. EPA cancer unit risk ratio is 1.2 × 10-5/(mg/m3). The Pb-content in this sample (21.5 ng/m3) suggests a low risk induced by the Pb-compounds in the air pollution alone in the sampled area of Detroit. However, lead-related problems still appear in industrialized and contaminated regions in the United States and Canada (5, 8). Arsenic. Detailed analysis of the As speciation in combustion emissions/byproducts, diesel soot (NIST SRM 1650), and urban air (NIST SRM 1648a and PM10) suggests that As in particulate air pollution mainly occurs in oxides with ∼10 ( 5% as the more toxic As(III) in urban air and diesel emissions (46, 47, 72). In this study, As was only observed in the Pb-oxygen-arsenic particle in Figure 4. Several anthropogenic sources, such as coal-fired power plants, smelters, and waste incinerators may theoretically emit lead-oxygenarsenic phases. Dispersion of pesticides is another possible source. More than 6.981 × 108 kg Pb-oxygen-arsenical pesticide compounds (e.g., PbHAsO4 and Pb4[PbOH][AsO4]3) were used within the United States between 1900 and 1980 and are still used in some countries (73, 74). The leadarsenates are relatively persistent under surface weathering conditions, and residuals (Pb3As2O8) are still reported in topsoils (75). Hence, aerosolization of pesticide contaminated soil dust is also a possible route for atmospheric lead arsenate. Correlation analysis shows that the As in airborne particulates 2292

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 8, 2004

in Detroit correlates well with Se (R2 g 0.997) and that these elements show good cross-correlations with Pb, REEs, Sb, In, Cd, Cu, S, and soil indicators (e.g., Mg, Si, Al, Ca, K, Ti) (R2 g 0.9). Hence, the identification of the As- or Pb-arsenate source is not possible based on these data. Arsenic broadly affects organs (e.g., peripheral nervous system, renal system, liver, blood system, nasal mucosa) and can also affect the respiratory system and induce lung cancer after long-term exposure (10, 66). Specifically, lead arsenates (CAS no. 7784-40-9) can cause inflammation in the respiratory system, arsenic poisoning, and lung cancer (69). In vivo experiments also show neurotoxic effects with accumulation of Pb and As in the brain (67). The As-concentration in the present sample was 2.07 ng/m3. However, the EF was 110, showing strong enrichment potential for soil and aquatic systems. The current RfC value for chronic exposure to Ascompounds is 30 ng/m3. Assuming that all the As occurs in Pb3As2O8, the concentration of the As-compound is 12.4 ng/ m3. Hence, a notable long-term carcinogenic effect from inhaled As-compounds is not expected. However, based on the U.S. EPA chronic inhalation cancer risk to inhaled Ascompounds (4.3 × 10-3 µg/m3) (24), 12-31 cancer cases are predicted in one million people, as estimated for As2O3 and Pb3As2O8, respectively. Rare Earth Elements (REE). The concentrations of the individual REE were all below 1 ng/m3, and their EF varied between 1.1 (Sm) and 1.7 (La) (Table 1). Despite these low concentrations, REE-particles were observed in the HAADFSTEM mode (Figure 4). The REE-bearing particles were a few tens of nanometers in the size. Elemental mapping showed that La and Ce were associated with P suggesting that the REE phases may be REE-phosphates, such as monazite or rhabdophane. The low enrichment of the REE (EF ) 1.1-1.7) and their speciation with phosphorus (EF ) 3.3) suggests that these REE-phosphates may originate from natural sources. Indeed both of the proposed phosphates are common accessory minerals in plutonic and sedimentary rocks. However, a statistical analysis shows that REEs (La, Ce, Nd, Sm) show a high correlation with both As (R2 g 0.9360.978) and Se (R2 g 0.960-0.987) as well as a cross-correlation with these elements. Hence, a high influence from fossil fuel combustion (primarily coal) is possible, but in the same range as that observed for crustal rock indicators (Mg, Si, Al, Ca, K, Ti). Phosphorus (CAS no. 7723-14-0) is listed as one of the 188 HAPs; however, REE-phosphates are not associated with adverse health effects in humans. Barium and Strontium. Bulk concentrations of Ba and Sr in the sample were 30.7 and 9.59 ng/m3, respectively. Elemental mapping showed that Sr was typically incorporated into Ba-sulfate, barite, as shown in Figure 5a. However, a SrSO4-particle was also observed in a ∼500 nm grain and identified as celestine based on the diffraction pattern (Figure 5b). Both phases are common Sr-bearing minerals. Compared with the average crustal composition, the EF of Sr was 2.4 suggesting a low influence from anthropogenic sources on the concentration of Sr in the Detroit atmosphere. The EF (6.8) of Ba, on the other hand, was enriched somewhat above that of Sr, suggesting that anthropogenic sources may contribute significantly to the Ba-concentrations observed. Using an elemental correlation analysis, R2 values above 0.9 with crustal rock indicator elements are observed for Ba and Sr in Detroit. However, Ba also correlates well with As (R2 ) 0.954), Se (R2 ) 0.974), and Br (R2 ) 0.935). Strontium, on the other hand, only correlated well with As (R2 ) 0.967) and Se (R2 ) 0.964). The correlation factor between Br and Sr was 0.790. Despite the low enrichment factors, the correlation analysis suggests that emissions from both gasoline and coal combustion contributes significantly to the atmospheric concentrations of Ba in Detroit, whereas the Sr-concentra-

FIGURE 5. (a) HAADF-STEM image with the elemental map (red square) of barium sulfate incorporating a small amount of Sr associated with larger particles of Ca-sulfate. (b) HAADF-STEM image with the elemental map of the region in the red square. Sr-particle associated with a larger grain of Ca-Mg-carbonate. The diffraction pattern of Sr-bearing particles is an inset. tions mainly are enriched by emissions from coal combustion. Zinc. The Zn typically occurred as ultrafine particles (e100 nm) attached to larger-size particles (Figure 6a). Reliable HRTEM images of the Zn-particles could not be obtained, because the Zn-rich particles occurred in complex aggregates with other nanoparticles of other phases (Fe-, Si-, and Aloxides and phosphates), as shown in the elemental maps (Figure 6). A few relatively larger Fe-Zn-particles (100-150 nm) were also observed (Figure 6b). The larger Fe-Znparticles were identified as Fe3Zn10 based on the diffraction patterns. The Zn-concentration was relatively high in this sample (103 ng/m3) and was strongly enriched (EF ) 139). The most important anthropogenic Zn-sources are smelters, manufacturing of Zn-compounds, combustion of hydrocarbon fuels and coal, and incineration (76, 77). However, pesticides where ZnO (CAS no. 1314-13-2) is used as an inert ingredient are also possible source (69). The Fe3Zn10 is most likely a high-temperature phase of industrial origin. Fe3Zn10 is produced in the metallurgical industry and can be synthesized by quenching after slowly cooling a Zn-rich ingot (∼92 atom % Zn) from 805 °C to 680 °C (78). The ultrafine size of the unidentified Zn-particles also suggests an anthropogenic origin, such as fossil fuel combustion. The origin of Zn from both metallurgical and fossil fuel combustion sources is further supported from correlation analysis showing that Zn concentration correlates well (R2 g 0.905) with both Br (R2 g 0.968) and the transition/heavy metals (Cr, Mn, Fe, Co, Zr, Mo, Cd, and Tl) (R2 ) 0.905-0.991). Zinc can cause both good and deleterious effects on human health. The addition of Zn in biological systems can alleviate the negative health effects from Pb (66). Currently

FIGURE 6. (a) HAADF-STEM image with the elemental map (red square) of the surface of Ca-sulfate showing that Zn-particles are dispersed with other phases such as Fe oxides, Al-oxides, and quartz. (b) HAADF-STEM image of a particle containing Zn and a larger phase of Ca-sulfate with the elemental map (red square). The inset is the diffraction pattern of the Zn-phase. there is no RfC-value for Zn-compounds. Deleterious health effects resulting in metal fume fever, cough, and pulmonary changes from inhaled ZnO have only been documented in connection with acute high or long-term occupational exposure (e.g., in welders and smelter workers). In vitro bioassays have shown that Zn plays an essential role in controlling programmed cell death and these mechanisms are very sensitive to the Zn-concentration (79). Chromium. The concentration of Cr in the present sample was 2.7 ng/m3, resulting in an EF of 2.6. Such a low EF of Cr is typical in urban aerosols (80). However, Cr is one of the most critical air-toxics with RfC-values ranging from 8 to 100 ng/m3 depending on speciation. Cr(VI) compounds are more harmful than Cr(III) compounds and cause severe air-way irritation and lung cancer (69, 81). In the present sample, Cr was associated with Fe and observed in ∼100-400 nm particles in association with similar-size Fe-silica spheres attached to gypsum (Figure 7a) and Fe-Al-silicates (Figure 7b). Elemental mapping indicated that the Cr-rich particles consist solely of Cr, Fe, and O, indicating that the Cr-phase is a Cr-Fe-oxide such as chromite, Fe(II)Cr(III)2O4. Chromite is commonly found in mafic rocks, but anthropogenic Cr emissions originate mainly from the combustion of coal and oil. Chromium in NIST urban air SRM 1649a and a 1992 PM10 sample from Lexington, KY, was generally present as Cr(III) and occurred in chromite and Cr(III)-sulfate, such as Cr2(SO4)3‚nH2O, respectively (46). However, e10(5% Cr(VI) was present in the sample from Kentucky. Cr(III)-sulfate strongly dominates the Cr-species found in diesel exhaust and residual oil fly ash (46, 47). Silica- and Fe-silica-spheres are typical compounds in coal fly ash, and the association between these particles and chromite in Figure 7a,b suggest the same origin for chromite. However, the potential VOL. 38, NO. 8, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2293

FIGURE 7. (a) HAADF-STEM image of Cr-bearing particle and the elemental map (red square). Larger particles are possibly gypsum and Fe-Al-silicate. (b) HAADF-STEM image of another Cr-phase with the elemental maps (red square). Matrix appears to contain Ca-sulfate and mica.

FIGURE 8. HAADF-STEM image of Se-bearing particles with the elemental maps (red squared region). contribution from natural and anthropogenic sources cannot be eliminated. Correlation analysis shows a high correlation between Cr and Br (R2 g 0.899) and the metals (Fe, Co, Zn, Mo, and Tl) (R2 ) 0.911-0.966). Hence, despite the low EF for Cr, the correlation analysis suggests a significant contribution from traffic (Br) and/or both industrial/metallurgical sources. Selenium. Selenium was not analyzed in the bulk chemical analysis of the sample. Previous analyses showed relatively low concentrations of Se (2.5 ng/m3; EF ) 511) at the same location in Detroit (60). Despite the low Se-concentrations, a few Se-particles were observed (Figure 8). The elemental map did not reveal any association between selenium and oxygen, and electron diffraction pattern analysis of the grains only showed diffuse diffraction patterns, indicating that Se is present as elemental Se in the amorphous state. Selenium is one of the marker elements of fossil fuel combustion (coal 2294

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 8, 2004

FIGURE 9. (a) HAADF-STEM image of Sn-particle embedded in gypsum with the elemental map (red square). (b) HAADF-STEM image of Y-, Zr-particle on a large aluminosilicate grain with the elemental map (red square). and oil) and waste incineration (82, 83). Our previous analysis strongly suggested that Se at the site predominantly originates from coal combustion (60). Emission of SeO2 from the combustion of fossil fuels can be transformed to elemental Se by reaction with SO2 in the atmosphere (84). Based on these observations and the Se reaction path, it is reasonable that the Se-particles observed in this sample consist of elemental Se. Although Se is listed as a hazardous air pollutant (HAP), atmospheric pollution with elemental Se (CAS no. 7782-492) has a low toxicity, as compared with the volatile compounds (e.g., H2Se) also emitted during combustion of fossil fuels (69, 85). Se is also an essential element participating in the formation of antioxidants in most living species (84). Tin. Sn-bearing particles of 300-400 nm-size were observed attached onto a Ca-sulfate particle (Figure 9a). The elemental map showed depletion of O and other elements at the position of Sn, indicating that the Sn-bearing phase consists of elemental Sn (Figure 9a). Elemental tin (CAS no. 7440-31-5) is considered nontoxic to humans and is only found to be an irritant to eyes and airways at occupational exposure levels where it can also induce benign pneumoconiosis (69). The Sn concentration in this sample was 0.69 ng/m3, equivalent to an EF of 33. Correlation analysis shows that Sn generally has a poor correlation with other metals in the Detroit sample. The highest correlation factor for Sn was observed with Ag (R2 ) 0.485). Hence, Sn in Detroit may be derived mainly from noble metal smelters. Zirconium and Yttrium. A ∼200 nm Zr- and Y-bearing particle was observed in association with slightly smaller Fe-Ti-rich particles at the surface of a Ca-Al-silicate (Figure 9b). The concentration of Zr in the sample was 2.8 ng/m3, whereas the concentration of Y was below the detection limit. The low EF of Zr (1.6) suggests a limited contribution from anthropogenic sources. The element mapping indicates that Si is also present in the Zr-Y-phase, suggesting that the Zr-

FIGURE 10. (a) HAADF-STEM image of Au-nanoparticles adhering to the surface of carbonaceous matter. (b) High magnification image of HAADF-STEM showing the occurrence of Au-nanoparticles, even a few nanometers in size. (c) HRTEM image reveals a representative lattice fringe of elemental gold. (d) HAADF-STEM image of a Agparticle on the same carbonaceous matter on which Au-particles are present. Y-phase may be zircon, which is a common heavy mineral in the geosphere originating from igneous rocks. Correlation factor analysis suggests both a natural source and coal combustion as sources of Zr in Detroit. A high correlation for Zr was observed with both As (R2 ) 0.962), Sr (R2 ) 0.978), and Br (R2 ) 0.975) as well as with crustal rock indicators. Gold and Silver. Numerous 1 µm). Probably most of the trace element particles became attached to the surfaces of larger particles during the collection of the sample. This is a well-known problem in particle sampling. However, ongoing analysis of 12-h size-fractioned samples have shown the presence of both individual and agglomerated ultrafine trace metal particles on filters. In this study, we have shown that trace elements were rarely observed as impurities in larger-size particles. The trace metals were observed mainly as individual particles. Hence, the trace metals, including the heavy metals, usually occur as individual e 500 nm-size particles. This has implications for their fate in the environment and impact on health effects, as trace metals occur mainly as very fine and ultrafine sized particles. The trace metal particles will have long residence times in the atmosphere: particles in the size rage from 0.1 to 1 µm have the longest residence time (∼100 days) among all size fractions (88). Such a long residence time allows a wide geographic distribution of the trace metal particles. Longrange transport is often found to dominate the air pollution of certain metals, such as Pb, certain transition metals, and noble metals, in both remote and highly populated urban areas (e.g., refs 37, 62, and 89). The fine submicrometer to ultrafine size of the heavymetal particles enables deep penetration into the respiratory tract and translocation of insoluble particles into tissue and the vascular system. Since, ultrafine particles are much more inflammogenic per mass dose than larger-size particles (11, 12), the inflammatory response of the observed heavy metal particles may be increased significantly above that expected based on bulk elemental concentrations. Moreover, the solubility (S) of ultrafine particles is notably increased as compared to that (S0) of larger-size particles following the equation S/S0 ) exp[(2γV*)/(RTr)], where γ is the surface free energy (mJ/m2); V* is the molar volume (m3/mol); R is the gas constant (mJ/mol*K); T is the temperature (K); and r is the particle radius (m) (90). Accordingly, the chemical toxicity and inflammatory potential as well as the environmental stability of the heavy trace metal particles may be greatly underestimated for nanoscale particles. Based on our in situ analysis of particles in the Detroit aerosol, this phenomenon with ultrafine material may be expected for PbO, Pb-arsenate, REE-phosphates, ZnO, selenium, and the noble metals, gold and silver. Less pronounced size-effects are expected from the 100-500 nm size trace metal particles that consisted mainly of chromite, barite, zircon, and a tin phase.

Acknowledgments M. Kawasaki and S. Johnson are acknowledged for their valuable advice in operating the STEM. J. F. Mansfield and C. J. Wouchope are thanked for their daily maintenance of the TEM facilities. This work was supported by The University of Michigan’s GeoScience and Engineering Initiative and the Environmental Management Science Program of DOE (DEFG07-97ER14816). K.A.J. was additionally supported by the Danish Ministry of the Interior and Health, Research Centre for Environmental Health (J.nr. 383-29-2001).

Literature Cited (1) Pope, C. A., III.; Burnett, R. T.; Thun, M. J.; Galle, E. E.; Krewski, D., Ito, K.; Thurston. G. D. J. Am. Med. Assoc. 2002, 287, 11321142. (2) Oberdo¨rster, G. Philos. Trans. R. Soc. London, Ser. A 2000, 358, 2719-2740. (3) Natusch, D. F. S.; Wallace, J. R. Science 1974, 186, 695-699. (4) Ku ¨ nzli, N.; Kaiser, R.; Medina, S.; Studnicka, M.; Chanel, O.; Filliger, P.; Herry, M.; Horak, F. Jr.; Puybonnieux-Texler, V.; VOL. 38, NO. 8, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2295

(5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24)

(25) (26) (27) (28)

(29) (30)

(31) (32) (33) (34) (35)

Que´nel, P.; Schneider, J.; Seethaler, R.; Vergnaud, J.-C.; Sommer, H. Lancet 2000, 356, 795-801. Margai, F.; Henry, N. Social Sci. Med. 2003, 56 1073-1085. Watkinson, W. P.; Campen, M. J.; Nolan, J. P.; Costa, D. L. Environ. Health Perspect. 2001, 109 Suppl 4, 539-546. Peden, D. B. Environ. Health Perspect. 2002, 110 Suppl 4, 565568. Newhook, R.; Hirtle, H.; Byrne, K.; Meek, M. E. Sci. Total Environ. 2003, 301, 23-41. Denkhaus, E.; Salnikow, K. Crit. Rev. Oncol. Hematol. 2002, 42, 35-56. Mandal, B. K.; Suzuki, K. T. Talanta 2002, 58, 201-235. Donaldson, K.; Stone, V.; Gilmour, P. S.; Brown, D. M.; MacNee, W. Philos. Trans. R. Soc. London, Ser. A 2000, 358, 2741-2749. Brown, D. M.; Wilson, M. R.; MacNee, W.; Stone, V.; Donaldson, K. Toxicol. Appl. Pharmacol. 2001, 175, 191-199. Hetland, R. B.; Schwarze, P. E.; Johansen, B. V.; Myran, T.; Uthus, N.; Refsnes, M. Hum. Exp. Toxicol. 2001, 20, 46-55. Gilmour, P. S.; Brown, D. M.; Lindsay, T. G.; Beswich, P. H.; MacNee, W.; Donaldson, K. Occup. Environ. Med. 1996, 53, 817822. Monn, C.; Becker, S. Toxicol. Appl. Pharmacol. 1999, 155, 245252. Jensen, K. A.; Allermann, L. Abstract of 18th International Mineralogical Association, 2002, p 181. Carter, J. D.; Ghio, A. J.; Samet, J. M.; Devlin, R. B. Toxicol. Appl. Pharmacol. 1997, 146, 180-188. Dreher, K. L.; Jaskot, R. H.; Lehman, J. R.; Richards, J. H.; McGee, J. K.; Ghio, A. J.; Costa, D. L. J. Toxicol. Environ. Health 1997, 50, 285-305. Donaldson K.; Beswick, P. H.; Gilmour, P. S. Toxicol. Lett. 1996, 88, 293-298. Wilson, M. R.; Lightbody, J. H.; Donaldsen, K.; Sales J.; Stone, V. Toxicol. Appl. Pharmacol. 2002, 184, 172-179. Zoroddu, Z. A.; Schinocca, L.; Kowalik-Jankowska, T.; Kozlowski, H.; Salnikow, K.; Costa, M. Environ. Health Perspect. 2002, 110 Suppl 5, 719-723. Gaggelli, E.; Berti, F.; D′Amelio, N.; Gaggelli, N.; Valensin, G.; Bovalini, L.; Paffetti, A.; Trabalzini, L. Environ. Health Perspect. 2002, 110 Suppl 5, 733-738. Rossman, T. G.; Uddin, A. N.; Burns, F. J.; Bosland, M. C. Environ. Health Perspect. 2002, 110 Suppl 5, 749-752. Inhalation Reference Concentration (RfC): The RfC is an estimate (with uncertainty spanning perhaps an order of magnitude) of a continuous inhalation exposure to the human population (including sensitive subgroups which include children, asthmatics, and the elderly) that is likely to be without an appreciable risk of deleterious effects during a lifetime. It can be derived from various types of human or animal data, with uncertainty factors generally applied to reflect limitations of the data used. U.S. EPA Office of Air Qality Planning and Standards. (http://www.epa.gov/ttn/atw/hapindex.html, 2002). Bols, N. C.; Brubacher, J. L.; Ganassin, R. C.; Lee, L. E. J. Dev. Comp. Immunol. 2001, 25, 853-873. Larison, J. R.; Likens, G. E.; Fitzpatrick, J. W.; Crock, J. G. Nature 2000, 406, 181-183. Sures, B.; Zimmermann, S.; Sonntag, C.; Stu ¨ ben, D.; Taraschewski, H. Environ. Pollut. 2003, 122, 401-405. Kjølholt, J.; Stuer-Lauridsen, F.; Mogensen, A. S.; Havelund, S. The Elements in the Second Rank - An Environmental Problem Now or in the Future?; Environmental Project No. 770; Danish Environmental Protection Agency - Danish Ministery of the Environment: Danish Environmental Protection Agency, 2003. Buseck, P. R.; Jacob, D. J.; Po´sfai, M.; Li, J.; Anderson, J. R. Int. Geol. Rev. 2000, 42, 577-593. Navrotsky, A In Nanoparticles and The Environment; Banfield, J. F., Navrotsky, A., Eds.; Reviews in Mineralogy and Geochemistry 44; Mineralogical Society of America, Washington, DC, 2001; pp 73-103. Milford, J. B.; Davidson, C. I. Environ. Sci. Technol. 1985, 35, 1249-1260. Cass, G. R.; Hughes, L. A.; Bhave, P.; Kleeman, M. J.; Allen, J. O.; Salmon, L. G. Philos. Trans. R. Soc. London, Ser. A 2000, 358, 2581-2592. Hughes, L. S.; Cass, G. R.; Gone, J.; Ames, M.; Olmez, I. Environ. Sci. Technol. 1998, 32, 1153-1161. Pakkanen, T. A.; Loukkola, K.; Korhonen, C. H.; Aurela, M.; Ma¨kela¨, T.; Hillamo, R. E.; Aarnio, P.; Koskentalo, T.; Kousa, A.; Maenhaut, W. Atmos. Environ. 2001, 35, 5381-5391. Pakkanen, T. A.; Kerminen, V.-M.; Korhonen, C. H.; Hillamo, R. E.; Aarnio, P.; Koskentalo, T.; Maenhaut, W. Atmos. Environ. 2001, 35, 5537-5551.

2296

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 8, 2004

(36) Bilos, C.; Colombo, J. C.; Skorupka, C. N.; Rodriguez Presa, M. J. Environ. Pollution 2000, 111, 149-158. (37) Veysseyre, A.; Moutard, K.; Ferrari, C.; Van de Velde, K.; Barbante, C.; Cozzi, G.; Capodaglio, G.; Boutron, C. Atmos. Environ. 2001, 35, 415-425. (38) Olmez, I.; Gordon, G. E. Science 1985, 229, 966-968. (39) Yatin, M.; Tuncel, S.; Aras, N. K.; Olmez, I.; Aygun, S.; Tuncel, G. Atmos. Environ. 2000, 34, 1305-1318. (40) Andreae, M. O.; Charlson, R. J.; Bruynseels, F.; Storms, H.; Grieken, R. V.; Maenhaut, W. Science 1986, 232, 1620-1623. (41) Anderson, J. R.; Buseck, P. R.; Saucy, D. A.; Pacyna, J. M. Atmos. Environ. 1992, 26A, 1747-1762. (42) Dzubay, T. G.; Mamane, Y. Atmos. Environ. 1989, 23, 467-476. (43) Katrinak, K. A.; Anderson, J. R.; Buseck. P. R. Environ. Sci. Technol. 1995, 29, 321-329. (44) Post, J. E.; Buseck, P. R. Environ. Sci. Technol. 1985, 19, 682685. (45) Anderson, J.; Buseck, P. R.; Patterson, T. L.; Arimoto, R. Atmos. Environ. 1996, 30, 319-338. (46) Huggins, F. E.; Shah, N.; Huffman, G. P.; Robertson, J. D. Fuel Process. Technol. 2000, 65-66, 203-218. (47) Huggins, F. E.; Huffman, G. P.; Robertson, J. D. J. Hazard. Mater. 2000, 74, 1-23. (48) Buseck, P. R.; Po´sfai, M. Proc. Natl. Acad. Sci. U.S.A 1999, 96, 3372-3379. (49) Parungo, F.; Kopcewicz, B.; Nagamoto, C.; Schnell, R.; Sheridan, P.; Zhu, C.; Harris, J. J. Geophys. Res. 1992, 97, 15, 867-15, 882. (50) Po´sfai, M.; Anderson, J. R.; Buseck, J. R.; Shattuck, T. W.; Tindale. N. W. Atmos. Environ. 1994, 28, 1747-1756. (51) Po´sfai, M.; Anderson, J. R.; Buseck, P. R.; Sievering, H. J. Geophys. Res. 1995, 100, 23063-23074. (52) Po´sfai, M.; Anderson, J. R.; Buseck, P. R.; Sievering, H. J. Geophys. Res. 1999, 104, 21685-21693. (53) Sheridan, P. J. Atmos. Environ. 1989, 23, 2371-2386. (54) Sheridan, P. J.; Schnell, R. C.; Kahl, J. D.; Boatman, J. F.; Garvey, D. M. Atmos. Environ. 1993, 27A, 1169-1183. (55) Po´sfai, M.; Molna´r, AÄ . In Environmental Mineralogy, Vaughan, D. J., Wogelius, R. A., Eds.; EMU Notes in Mineralogy, 2000; pp 197-252. (56) Maynard, A. D. J. Aerosol Sci. 1995, 26, 757-777. (57) Maynard, A. D. Aerosol Sci. Technol. 1995, 23, 521-533. (58) Maynard, A. D. Philos. Trans. R. Soc. London, Ser. A 2000, 358, 2593-2610. (59) Xhoffer, C.; Jacob, W.; Van Grieken, R. J. Aerosol. Sci. 1989, 20, 1617-1619. (60) Utsunomiya, S.; Jensen, K. A.; Keeler, G. J.; Ewing, R. C. Environ. Sci. Technol. 2002, 36, 4943-4947. (61) Utsunomiya, S.; Ewing, R. C. Environ. Sci. Technol. 2003, 37, 786-791. (62) Duce, R. A.; Hoffman, G. L.; Zoller, W. H. Science 1975, 187, 59-61. (63) Harrison, R. M.; Williams, C. R. Sci. Total Environ. 1983, 31, 129-140. (64) Athena Mineralogy: http://un2sg4.unige.ch/athena/mineral/, 2003. (65) Nriagu, J. O. In The Biogeochemistry of lead in the environment part A; Nriagu, J. O., Ed.; Wiley-Interscience Pub.: New York, 1978; pp 137-184. (66) Fergusson, J. E. The heavy elements, chemistry, environmental impact and health effects; Pergamon, NY, 1990. (67) Mejı´a, J. J.; Dı´az-Barriga, F.; Caldero´n, J.; Rı´os, C.; Jime´nezCapdeville, M. E. Neurotoxicol. Teratol. 1997, 19, 489-497. (68) Needleman, H. L. Environ. Res. Sec. A 2000, 84, 20-35. (69) NLM. TOXNET. National Library of Medicine (http:// www.toxnet.nlm.nih.gov). 2003. Ref Type: Electronic Citation. (70) Biggins, P. D. E.; Harrison, R. M. Nature 1978, 272, 531-532. (71) Ter Haar, G. L.; Bayard, M. A. Nature 1971, 216, 353-355. (72) Mukhopadhyay, P. K.; Lajeunesse, G.; Crandlemire, A. L. Int. J. Coal Geol. 1996, 32, 279-312. (73) Murphy, E. A.; Aucott, M. Sci. Total Environ. 1998, 218, 89-101. (74) Reigart, J. R.; Roberts, J. R. R. Recognition and Management of Pesticide Poisonings, 5th ed.; U.S. Environmental Protection Agency: U.S. Environmental Protection Agency, 1999. (75) Kennedy, S. K.; Walker, W.; Forslund, B. Environ. Forensics 2002, 3, 131-143. (76) Johansson, L. S.; Tullin, C.; Leckner, B.; Sjovall, P. Biomass Bioenergy 2003, in press, corrected proof. (77) Stigter, J. B.; de Haan, H. P. M.; Guicherit, R.; Dekkers, C. P. A.; Daane, M. L. Environ. Pollut. 2000, 107, 451-464. (78) Brandon, J. K.; Brizard, R. Y.; Chieh, P. C.; McMillan, R. K.; Pearson, W. B. Acta Crystallogr. B 1974, 30, 1412-1417.

(79) Watjen, W.; Haase, H.; Biagioli, M.; Beyersmann, D. Environ. Health Perspect. 2002, 110 Suppl 5, 865-867. (80) Nriagu, J. O.; Pacyna, J. M.; Milford, J. B.; Davidson, C. I. In Chromium in the natural and human environments; Nriagu, J. O., Nieboer E., Eds.; Wiley-Interscience Pub.: New York, 1988; pp 125-172. (81) Nieboer, E.; Shaw, S. L. In Chromium in the natural and human environments; Nriagu, J. O., Nieboer E., Eds.; Wiley-Interscience Pub.: New York, 1988; pp 399-441. (82) Watson, J. G.; Chow, J. C.; Houck, J. E. Chemosphere 2001, 43, 1141-1151. (83) Chiaradia, M.; Cupelin, F. Atmos. Environ. 2000, 34, 327332. (84) ATSDR. Draft - Toxicological Profile for Selenium; U.S. Department of Health and Human Services: Agency for Toxic Substances and Disease Registry, 2001. (85) Odziemkowski, M.; Koziel, J. A.; Irish, D. E.; Pawliszyn, J. Anal. Chem. 2001, 73, 3131-3139.

(86) Van de Velde, K.; Barbante, C.; Cozzi, G.; Moret, I.; Bellomi, T.; Ferrari, C.; Boutron, C. Atmos. Environ. 2000, 34, 3117-3127. (87) Pirrone, N.; Allegrini, I.; Keeler, G. J.; Nriagu, J. O.; Rossmann, R.; Robbins, J. A. Atmos. Environ. 1998, 32, 929-940. (88) Anastasio C.; Martin, S. T. In Nanoparticles and The Environment; Banfield, J. F., Navrotsky, A., Eds.; Reviews in Mineralogy and Geochemistry 44; Mineralogical Society of America: Washington, DC, 2001; pp 293-349. (89) Barbante, C.; Veysseyre, A.; Ferrari, C.; Morel, C.; Capodaglio, G.; Cescon, P.; Scarponi, G.; Boutron, C. Environ. Sci. Technol. 2001, 35, 835-839. (90) Hochella, M. F., Jr. Earth Planet. Sci. Lett. 2002, 203, 593-605.

Received for review September 12, 2003. Revised manuscript received February 5, 2004. Accepted February 10, 2004. ES035010P

VOL. 38, NO. 8, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2297