Article pubs.acs.org/est
Automated Mineralogical Analysis of PM10: New Parameters for Assessing PM Toxicity Ben. J. Williamson,*,†,‡,§ Gavyn Rollinson,†,§ and Duncan Pirrie#,§ †
University of Exeter, Camborne School of Mines, Penryn, Cornwall TR10 9EZ, United Kingdom Department of Mineralogy, Natural History Museum, Cromwell Road, London, SW7 5BD, United Kingdom # Helford Geoscience LLP, Menallack Farm, Treverva, Penryn, Cornwall TR10 9BP, United Kingdom ‡
ABSTRACT: This work provides the first automated mineralogical/phase assessment of urban airborne PM10 and a new method for determining particle surface mineralogy (PSM), which is a major control on PM toxicity in the lung. PM10 was analyzed on a TEOM filter (Aug.−Sept. 2006 collection) from the London Air Quality Network Bexley, East London, U.K. A cross-section of the filter was analyzed using a QEMSCAN automated mineralogical analysis system which provided 381 981 points of analysis for 14 525 particles over a period of 9 h 54 min. The method had a detection limit for individual mineral components of 0.05 ppm (by area). Particle shape and mineralogical characteristics were determined for particles in the size ranges PM10−4, PM4−2.5, and PM2.5−0.8. The PM2.5−0.8 fraction contained 2 orders of magnitude more mineral particles than the PM10−4 and PM4−2.5 fractions, however the PM10−4 fraction forms 94% and 79% of the mineral mass and surface area, respectively. PSM of the PM10 was dominated by gypsum (36%), plagioclase (16%), Na sulphates (8%), and Fe−S− O phases (8%) in the PM10−2.5, which may be important in explaining the toxicity of the coarse fraction. The wider implications of the study are discussed.
1. INTRODUCTION The World Health Organization has estimated that 2 million deaths worldwide per year can be attributed to breathing airborne particulate matter (PM),1 largely due to increased cardiopulmonary morbidity and mortality.2 Of most concern is particulate matter in the so-called fine fraction, with aerodynamic diameters less than around 2.5 μm (PM2.5), which can reach the alveolar regions of the lung.3 However, there is evidence that the coarse fraction (PM10−2.5), which mainly deposits in the upper airways, may also be impacting on human health.3−9 One of the greatest challenges in understanding the causes of PM toxicity is the difficulty in characterizing the physical and chemical properties of complex mixed dusts.10 PM2.5 is mainly composed of sulfate and nitrate-based compounds, organic and elemental carbon produced from combustion processes, and particles formed by gas condensation and gas to particle conversion in the atmosphere.2 The coarse fraction is largely generated by mechanical processes such as volcanic eruptions,11 industrial and construction activities, resuspension of soil and street particles, and from sea spray, and may often contain pollen, fungal spores, and plant and animal fragments.2,12 Current air quality guidelines in the UK and elsewhere are mainly based on particle concentration (mass) per unit volume of air for certain PM size categories, usually PM10 and also sometimes PM2.5. However, it is clear from numerous studies on the enhanced toxicity of specific dusts (e.g., containing © 2013 American Chemical Society
quartz or certain asbestiform minerals) that there are other important factors such as particle shape, solubility, speciation of contained toxic substances, and surface properties.13 Mineralogical information will undoubtedly help to provide insights into which types and components of PM cause most harm, and by what mechanisms. This may, in the future, provide data for more specific daily air quality information for both the public and risk managers. The most common means of determining PM mineralogy is X-ray diffraction,14 but this is not specific to mineral surfaces. Individual particles can be characterized using techniques such as transmission electron microscopy with electron diffraction; however, this is extremely time-consuming and is not practical for the acquisition of large, statistically significant data sets. The only method able to offer the necessary rapid individual particle analysis, automated mineral/phase identification (including particle surfaces) and size and shape characterization for different size categories (e.g., PM10) is automated scanning electron microscopy (SEM). A range of SEM-based automated mineral analysis systems are currently available. The QEMSCAN instrument used in this study is capable of determining size, shape and mineral/phase (henceforth referred Received: Revised: Accepted: Published: 5570
December 12, 2012 April 25, 2013 April 29, 2013 April 29, 2013 dx.doi.org/10.1021/es305025e | Environ. Sci. Technol. 2013, 47, 5570−5577
Environmental Science & Technology
Article
Figure 1. (a) Backscattered electron image of the surface of the sectioned TEOM filter; (b) QEMSCAN Fieldscan image (mineralogical map) of a portion of the same area showing mineral particles on the surface of the TEOM filter substrate; and (c) images of a range of particles showing gypsum surface coatings or aggregates.
(TEOM, manufactured by Rupprecht & Patashnick) which is used by the London Air Quality Network and many other authorities worldwide for the continuous measurement of ambient airborne particulate matter (PM10 and PM2.5). However, the methodologies tested in this work could reasonably be applied to the analysis of any filter-based PM collection.
to as mineral for brevity) composition, and associations for over 1000 particles per hour, for particles 1 to 10 μm in diameter and with a 1 μm measurement spacing.15 Very few studies of airborne particulate matter using QEMSCAN have been published so far,16−20 although the use of SEM-EDS in the analysis of PM10 and PM2.5 is well documented.21,22 Speirs et al.17 used QEMSCAN to characterize natural aeolian dusts and Shao et al.,20 in their review of the dust cycle, suggested that this automated mineral analysis approach may provide new possibilities for dust analysis. Morrison et al.23 quantified respirable crystalline silica in ambient air around the Hunter Valley open-cast coal mines in NSW, Australia. To our knowledge, ours is the first study to utilize QEMSCAN for the analysis of urban PM10. The aims of the current study are as follows: (1) to demonstrate the potential of QEMSCAN for the mineralogical analysis of PM10, particularly particle surface mineralogy, and (2) to provide the first large (∼15 000 particle) data set on the mineralogical composition of urban PM10. The type of filter analyzed was from a Tapered Element Oscillating Microbalance
2. MATERIALS AND METHODS 2.1. TEOM Air Filter Analyzed. TEOM systems consist of a filter cartridge which contains a circular Pallflex Teflon-coated glass fiber filter mounted on a plastic and aluminum base. The cartridge is positioned at the top of a hollow tapered crystal element, which is fixed at one end and free to vibrate at the other.2 Air is drawn through the filter and then through the element. As particulate matter is deposited on the filter, the frequency of vibration of the crystal element decreases, with this decrease being proportional to the mass deposited on the filter. The filter is heated to 50 °C during operation to drive off moisture.2 5571
dx.doi.org/10.1021/es305025e | Environ. Sci. Technol. 2013, 47, 5570−5577
Environmental Science & Technology
Article
The TEOM filter analyzed was for a PM10 collection from the Bexley-Thames Road South site (BX8) in East London, UK, start date: 10/8/06, 10:30 end date: 26/9/06, 12:15, sourced from the TEOM Filter Archive. During 2006, the BX8 site met the UK Government’s Air Quality Strategy Objectives24 for annual mean levels of PM10, with a figure of 30 μg/m3 compared with the objective of 40 μg/m3, but failed to meet the target for the number of days with a 24 h mean >50 μg/m3 (not to be exceeded more than 35 times a year), this level being exceeded on 49 days at BX8. The average PM10 concentration during the sampling period (i.e., for the filter analyzed) was 23 μg/m3. Meteorological data were available from the Thames Road North site (across the road from the Thames Road South site): average wind direction 232° (i.e., from the southwestlargely from south London), total rainfall 21 mm, average relative humidity 73%, and average temperature 18.3 °C. 2.2. Sample Preparation. The TEOM filter cartridge was removed from its sealed black plastic archive box using ceramic forceps and immediately embedded in EPO-TEK 301-2 epoxy resin in a 30 mm diameter circular mold. When cured, the resin formed a 30 mm diameter cylindrical block. The filter surface was orientated at around 70° to the flat surfaces of the block so that the surface could be ground down to reveal a cross section through the TEOM filter. The block was ground down dry, to prevent dissolution of water-soluble species, and then polished using Kemet water-free 1 μm diamond slurry (1-KDS1429) to limit removal of water-soluble phases. The surface of the polished block was carbon coated to a thickness of ∼25 nm. 2.3. QEMSCAN Analysis. The polished block was loaded into the QEMSCAN 4300 automated mineral analysis system at the Camborne School of Mines, UK. This system is based on a Zeiss Evo 50 SEM with four light element Bruker Xflash Silicon Drift energy dispersive X-ray detectors controlled by iMeasure v. 4.2 software for data acquisition and iDiscover v. 4.2 for spectral interpretation and data processing. The QEMSCAN was set up to run at an accelerating voltage of 25 kV and a beam current of 5 nA. Imaging of a 12 000 × 1000 μm area of the sample surface was first carried out in backscattered electron (BSE) mode where brightness is proportional to average atomic number (Figure 1a). The brightness and contrast of the BSE images were calibrated using quartz, copper and gold standards and then a BSE threshold set to ensure the recognition of all particles with a BSE signal above that of the resin mounting medium. EDS analysis was then carried out in a grid pattern (0.5 μm grid spacing) for those areas identified as containing particles from the BSE image. During the course of the analysis, 14 525 particles were identified, within which there were 381 981 points of analysis. For the area analyzed (12 000 × 1000 μm = 12 000 000 μm2) the data acquisition time was 9 h 54 min, and with a “resolution” (smallest particle identified) of 0.8 μm, the theoretical detection limit for individual minerals within the area analyzed was around 0.05 ppm (by area). A total of 1000 X-ray counts were collected for each point of analysis. In the iDiscover software, the data from each point of analysis is automatically compared with a Species Identification Protocol (SIP) database of mineral and noncrystalline phase spectra to identify the minerals present. The SIP used was modified from the LCU5 SIP provided with the QEMSCAN, which includes a range of oxide, sulfate, and silicate minerals. A mineral map, known as a Fieldscan image, of the sample surface was produced using the coordinates of the individual points of
analysis, each corresponding to an individual pixel (Figure 1b). The precision of QEMSCAN analysis is reported to be ≤1%; accuracy is dependent on the databases used for interpreting mineralogy,25 and a wide range of other factors discussed below. 2.4. Data Processing. From the iDiscover v. 4.2 software, data can either be reported on the basis of mineral name, or on a compositional grouping basis. Minerals with the same (or very similar) chemistries cannot be differentiated on the basis of SEM-EDS analysis, hence the mineral categories used to report the data may include several possible individual minerals. The following are the most likely minerals in each category: Gypsum − may include other Ca sulphates; Quartzquartz or boundary affect (see below); Na sulphatesany mineral with Na, S, O, ± Mg; Plagioclaseany mineral with Na, Al, Si, O to Ca, Al, Si, O (may include trace areas of the TEOM filter); K− Al-silicateany mineral with K, Al, Si, O (e.g., K-feldspar, muscovite); Fe−S−Oany mineral with Fe, S, and possibly O (e.g., pyrite, melanterite, jarosite); Fe(Mg) silicatesany mineral with Si, Mg, Fe such as olivine or talc; Ti minerals any mineral with Ti and O such as rutile, titanite, ilmenite, and paint flakes; FeO/OH/CO3Fe metal, oxides, hydroxides, and carbonates (e.g., hematite, goethite, siderite); Calciteany mineral consisting of Ca with or without C and O; Chlorite may include tourmaline, vermiculite and other Fe, Al, Si, and Fe, Mg, Al, Si minerals; Kaoliniteany mineral with Al and Si with low Fe and Mg; Cu−Fe−S−Oany mineral with Cu−Fe sulphides or sulphates; Dolomiteany mineral with Ca and Mg with or without C and O; Apatiteany Ca phosphates; and Otherany other mineral not included above. To determine the mineralogy and size/shape characteristics of individual particles, areas of the Fieldscan image identified as resin or TEOM collection filter were electronically removed (known as granulation in the processing software). Each discrete area was then defined as a separate particle which could be made up of several different minerals. For each particle, the database contains a mineral map and size/shape (area, perimeter length, axis lengths etc.) and mineral associations data which were output for the following health pertinent particle size ranges: (1) all particles present in the TEOM PM10 collection filter, i.e., PM10, known as the thoracic fraction, particles able to penetrate into the bronchiole regions of the lungs; (2) PM4, known as the respirable fraction, particles able to penetrate into the bronchiolar and alveolar, unciliated regions of the lung; and (3) the 2.5 to 0.8 μm fraction, part of the fine particle fraction (PM2.5).2 The 0.8 μm cutoff was the smallest particle size resolvable using QEMSCAN in this study. As these particle size fractions overlap, it was deemed more informative to compare the characteristics of particles in the size ranges PM10−4, PM4−2.5, and PM2.5−0.8, and from this the coarse fraction (PM10−2.5). The cutoff of 0.8 μm will mainly affect the measured proportion of accumulation mode particles which will be mainly single or aggregates of combustion products or gas condensates (e.g., sulphates and nitrates).12 The parameter used to subdivide the data set into different size categories was “Size” from the iDiscover software. This incorporates a stereological correction, based on the “phase specific surface area” (PSSA),26 necessary when assessing the size of randomly sectioned particles. Such a correction is needed because, for example, a perfectly spherical particle may be sectioned through its center or through the narrowest point(s), which will give maximum and minimum diameters, respectively. Sutherland26 has calculated that for a perfect 5572
dx.doi.org/10.1021/es305025e | Environ. Sci. Technol. 2013, 47, 5570−5577
Environmental Science & Technology
Article
obvious shortcomings as mineral categories can contain a number of different mineral and chemical species. An important additional parameter in assessing the causes of health impacts related to airborne PM is particle surface mineralogy (PSM), which we were able to determine for the first time using QEMSCAN. PSM is important, as it is particle surfaces, rather than material within the particles, which is most immediately available for reactions in the lung. PSM was calculated from the “mineral associations” data in Table 1, specifically the number of pixels of different minerals in direct contact with “background” (either filter substrate, carbonaceous material, or resin mounding medium). The main shortcoming of these data is the small number of pixels which make up the smallest particles, e.g., only one square pixel for a particle of 0.8 μm diameter. Therefore, the surface area data for the smallest particles cannot be accurate. 2.5. Limitations of QEMSCAN Analysis in the Current Study. The main limitations were as follows: (1) only 1000 Xray counts per spectra (point of analysis) were acquired (although this is customizable), which is necessary to match the conditions of the default mineral identification SIP and allows rapid acquisition. This low number of counts gives lower limits of detection for different elements in each spectrum of up to around 10 wt %.27,28 This means that different minerals cannot be distinguished on the basis of differences in their levels of minor constituents. This is not a major issue in the current study because the emphasis is on mineral identification rather than detailed mineral chemistry; (2) N, O, and F give weak signals; (3) it is not routinely possible to detect elements with atomic numbers less than 6 (C); (4) the sample was carbon coated which may contribute to the signal for C, (5) the BSE signal of carbonaceous materials (i.e., C-dominated, e.g., pollen and combustion-derived particles) is typically too close to the resin for them to be designated as separate particles for X-ray analysis, and (6) X-ray interaction volume issues, where for the analysis of particle rims, or particles with volumes smaller than the interaction volume, spectra may include a signal from the mounting resin (especially C and O), nearby particles or adjacent minerals in aggregates or composites. Because of issues 2−6, the mineralogical analysis of carbonaceous materials is usually not possible using QEMSCAN. The current study is therefore restricted to the characterization of mineral and some nonlight element secondary components/phases of urban PM10 (e.g., secondary sulphates). The way the software handles the presence of C and O (from the carbon coat etc.) in the spectra of most mineral categories (including, e.g., calcite and dolomite) is that these categories include a “may have” for C and O. Issues relating to X-ray interaction volume, alluded to in point 6, are one of the main causes of so-called boundary effects. From a Monte Carlo simulation, for the analytical conditions used here (25 kV, spot size 600 nm), the interaction volume will be roughly tear-shaped with an approximate maximum diameter of around 5 μm and a depth of 7 μm. This was calculated for pure carbon as a rough proxy for carbonaceous materials, which will overestimate the interaction volume where particles with higher average atomic mass have been analyzed. Gottlieb et al.29 quoted an interaction volume at 25 kV of 1−2 μm diameter and 1−2 μm depth for an unspecified material. However, the effective resolution of the Fieldscan mineralogical map will be better than this (around 0.8 μm, the smallest particle resolvable), mainly because relatively low counts for X-rays from the peripheries of the interaction
sphere of 1-mm diameter, the mean sectional diameter is 0.67 mm; the “PSSA size” is 1 mm. The situation becomes more complex for nonspherical particles. For a square rod with an aspect ratio of 5:1 (thickness 1 mm), the maximum sectional diameter will be 5.2 mm, mean sectional diameter 0.91 mm, “PSSA size” = 1.36 mm. For a square plate with an aspect ratio of 5:1 (thickness 1 mm), the maximum sectional diameter will be 7.14 mm, the mean sectional diameter 1.42 mm, and the “PSSA size” = 2.14 mm. A wide range of measured and calculated data can be output from the QEMSCAN iDiscover software. In the current study to determine PM10 characteristics most pertinent to human health studies, the following data were output for the three particle size ranges PM10−4, PM4−2.5, and PM2.5−0.8: (1) number of particles of each mineral (Figure 2); (2) mass of each
Figure 2. Number of particles of each mineral in the PM10−4, PM4−2.5, PM2.5−0.8, and PM10−2.5 size fractions.
mineral as a percentage of the bulk mineral mass of PM10 (Figure 3), where the mass was calculated by multiplying the volume % of each mineral (volume calculated using the equivalent circular diameter for the area (number of pixels) of each particle) by its specific gravity (S.G.). The mass calculation, requiring an average S.G. for each category, has
Figure 3. Mass of each mineral in the PM10−4, PM4−2.5, PM2.5−0.8, and PM10−2.5 size fractions, expressed as a percentage of the total mass of PM10, i.e., all points in the graph add up to 100%. 5573
dx.doi.org/10.1021/es305025e | Environ. Sci. Technol. 2013, 47, 5570−5577
Environmental Science & Technology
Article
Table 1. Mineral Associations Matrix for Different Size Classes (PM10‑4, PM4‑2.5, and PM2.5‑0.8) Showing the Number of Pixels of Each Mineral in Direct Contact with Other Mineralsa
a
Mineral names along the top row are abbreviations of those in the first column.
quartz particle. The QEMSCAN operator manually checks each data set in these subclasses to decide whether they should be assigned to quartz (in this example) or to the unknown category “others”; (2) the signal is too mixed to be assigned to any of the main or subclasses and the data set is therefore automatically assigned to the “others” category. The QEMSCAN operator manually checks all entries in the “others” category, with the option to drive the SEM back to
volume (i.e., from nearby particles etc.) are likely to give a signal below the 10 wt % detection limits for each point of analysis. Elements detected from beyond the immediate point of analysis will produce a mixed spectrum which is dealt with by the software in two possible ways: (1) the data are put into a subclass of the main mineral category, for example subclass Si + O, may have Cl and/or Na, where a low level of Cl and S is detected from the substrate in an analysis close to the edge of a 5574
dx.doi.org/10.1021/es305025e | Environ. Sci. Technol. 2013, 47, 5570−5577
Environmental Science & Technology
Article
1b), mainly contained within a 5 μm in length, 3:1, criteria used in standard tests33) indicating a fibrous morphology. However, because a section through the filter was analyzed, this is likely to be an underestimate, as fibrous particles would be expected to lie flat on the filter surface. Another important mineral category is Fe− S−O, which formed 2% of PSM in the PM2.5−0.8 (1020 particles) but shows low particle numbers (10) in the coarse fraction. Its mineral form is unclear, although from its association with gypsum and FeO/OH/CO3 (Table 1), it is thought likely to be mostly present as sulfate, possibly the mineral melanterite (FeSO4·7(H2O)). The remaining minerals in Figure 2 are calcite, dolomite, Ti minerals, and apatite. Ti mineral particle numbers are significant with 134 in the PM2.5−0.8 but only 4 particles were present in the coarse fraction. Ti in road dusts is thought to have mainly a “crustal”34 or road traffic source,35 but can also be derived from paint flakes.36 4.2. Potential Mineralogical Controls on PM Toxicity. As previously stated, the toxicity of PM10 in the lung is generally attributed to the fine (PM2.5) or respirable (PM4) fraction3 due to its relatively large surface area, higher number concentration in ambient air, and small size.2,37 However, given the PSM data presented herein, it is doubtful whether for mineral particles the fine and respirable fractions have a larger bulk surface area than the coarse fraction (PM10−2.5). We propose that the high PSM for gypsum (36%), plagioclase (16%), Na sulphates (8%), and Fe−S−O phases (8%) in the coarse fraction, as a proportion of bulk PM10, may explain why the coarse fraction can be, in certain respects, as toxic as the fine fraction. However, any attempt to prove this is likely to
■
AUTHOR INFORMATION
Corresponding Author
*Phone: +44-1326-371856; e-mail:
[email protected] (B.J.W.). Author Contributions
§ The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. These authors contributed equally to this work.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors wish to thank Steve Pendray for sample preparation, and Frank Kelly and Chrissi Dunster (King’s College London) for providing the TEOM filter, and associated data. QEMSCAN is a registered trademark of FEI Company.
■
REFERENCES
(1) WHO: Tackling the global clean air challenge. New Release, 26 September 2011; http://www.who.int/mediacentre/news/releases/ 2011/air_pollution_20110926/en/index.html. (2) QUARG: Airborne Particulate Matter in the United Kingdom; , Quality of Urban Air Review Group, Department of the Environment: London, 1996; Vol. 3. (3) EPAQS: Airborne Particles: What Is the Appropriate Measurement on Which to Base a Standard? A Discussion Document; Expert Panel on Air Quality Standards; UK Department for Environment, Food & Rural Affairs: London, 2001; http://www.defra.gov.uk/environment/ airquality/aqs/air_measure/index.htm. (4) Diociaiuti, M.; Balduzzi, M.; De Berardis, B.; Cattani, G.; Stacchini, G.; Ziemacki, G.; Marconi, A.; Paoletti, L. The two PM2.5 (fine) and PM2.5−10 (coarse) fractions: Evidence of different biological activity. Environ. Res. Sect. A 2001, 86, 254−262. (5) Gerlofs-Nijland, M. E.; Dormans, J. A. M. A.; Bloemen, H. J. T.; Leseman, D. L. A. C.; Boere, A. J. F.; Kelly, F. J.; Mudway, I. S.; Jimenez, A. A.; Donaldson, K.; Guastadisegni, C.; Janssen, N. A. H.; Brunekreef, B.; Sandstrom, T.; van Bree, L.; Cassee, F. R. Toxicity of coarse and fine particulate matter from sites with contrasting traffic profiles. Inhal. Toxicol. 2007, 19, 1055−1069. (6) Yeatts, K.; Svendsen, E.; Creason, J.; Alexis, N.; Herbst, M.; Scott, J.; Kupper, L.; Williams, R.; Neas, L.; Cascio, W.; Devlin, R. B.; Peden, D. B. Coarse particulate matter (PM2.5−10) affects heart rate variability, blood lipids, and circulating eosinophils in adults with asthma. Environ. Health Persp. 2007, 115, 709−714. (7) Li, R.; Wiedinmyer, C.; Baker, K. R.; Hannigan, M. P. Characterization of coarse particulate matter in the western United 5576
dx.doi.org/10.1021/es305025e | Environ. Sci. Technol. 2013, 47, 5570−5577
Environmental Science & Technology
Article
States: A comparison between observation and modelling. Atmos. Chem. Phys. Discuss. 2012, 12, 11445−11484. (8) Brunekreef, B.; Forsberg, B. Epidemiological evidence of effects of coarse airborne particles on health. Eur. Respir. J. 2005, 26, 309− 318. (9) Meister, K.; Johansson, C.; Forsberg, B. Estimated short-term effects of coarse particles on daily mortality in Stockholm, Sweden. Environ. Health Persp. 2012, 120, 431−436. (10) Kelly, F. J.; Fuller, G. W.; Walton, H. A.; Fussell, J. C. Monitoring air pollution: Use of early warning systems for public health. Respirology 2012, 17, 7−19. (11) Horwell, C. J.; Stannett, G. W.; Andronico, D.; Bertagnini, A.; Fenoglio, I.; Fubini, B.; Le Blond, J. S.; Williamson, B. J. A Physicochemical assessment of the health hazard of Mt. Vesuvius volcanic ash. J. Volcanol. Geoth. Res. 2010, 191, 222−232. (12) Gieré, R.; Querol, X. Solid particulate matter in the atmosphere. Elements 2010, 6, 215−222. (13) Fubini, B. Surface reactivity in the pathogenic response to particulates. Environ. Health Persp. 1997, 105 (Suppl. 5), 1013−20. (14) Williamson, B. J.; Udachin, V.; Purvis, O. W.; Spiro, B.; Cressey, G.; Jones, G. C. Characterisation of airborne particulates in the Cu smelter and former mining town of Karabash, south Ural mountains of Russia. Environ. Monit. Assess. 2004, 98, 235−259. (15) Pirrie, D.; Butcher, A. R.; Power, M. R.; Gottlieb, P.; Miller, G. L. Rapid quantitative mineral and mineral analysis using automated scanning electron microscopy (QemSCAN); Potential applications in forensic geoscience. Geol. Soc. Spec. 2004, 232, 123−136. (16) Martin, R. S.; Mather, T. A.; Pyle, D. M.; Power, M.; Allen, A. G.; Aiuppa, A.; Horwell, C. J.; Ward, E. P. W. Composition-resolved size distributions of volcanic aerosols in the Mt. Etna plumes. J. Geophys. Res. 2008, 113, D17211. (17) Speirs, J. C.; McGowan, H. A.; Neil, D. T. Polar Eolian sand transport: Grain characteristics determined by an automated scanning electron microscope (QEMSCAN). Arct. Antarct. Alp. Res. 2008, 4, 731−743. (18) Radhi, M. Physical and Chemical Properties of Australian Continental Aerosols. Ph.D. Dissertation, The University of New South Wales: Sydney, Australia, 2010; http://scholar.google.co.uk/ scholar?q=Radhi+australian+desert+dust+QEMSCAN&hl=en&as_ sdt=0%2C5. (19) Jia, Q. Dust from Mining Area and Proposal of Dust Emission Factors. Ph.D. Dissertation, Universitetstryckeriet: Luleå, Sweden, 2011. (20) Shao, Y.; Wyrwoll, K. H.; Chappell, A.; Huang, J.; Lin, Z.; McTainsh, G. H.; Mikami, M.; Tanaka, T. Y.; Wang, X.; Yoon, S. Dust cycle: An emerging core theme in Earth system science. Aeolian Res. 2011, 2, 181−204. (21) Casuccio, G. S.; Schlaegle, S. F.; Lersch, T. L.; Huffman, G. P.; Chen, Y.; Shah, N. Measurement of fine particulate matter using electron microscopy techniques. Fuel Process. Technol. 2004, 85, 763− 779. (22) Xie, R. K.; Seip, H. M.; Leinum, J. R.; Winje, T.; Xiao, J. S. Chemical characterization of individual particles (PM10) from ambient air in Guiyang City, China. Sci. Total Environ. 2005, 343, 261−272. (23) Morrison, A.; Nelson, P. F.; Stelcer, E.; Cohen, D.; Haberlah, D. Quantifying respirable crystalline silica in the ambient air of the Hunter Valley, NSWSorting the silica from the silicon. Proceedings of the 20th Clean Air & Environment Conference (2011); http://apo. ansto.gov.au/dspace/bitstream/10238/4236/2/Morrison_ quatifying%20Respirable%20Crystalline_Ambient_2011_07.PDF. (24) DEFRA: The Air Quality Strategy for England, Scotland, Wales and Northern Ireland; Department for Environment, Food and Rural Affairs: United Kingdom, 2007; http://www.defra.gov.uk/ environment/airquality/strategy/index.htm. (25) Pirrie, D.; Power, M. R.; Rollinson, G. K.; Wiltshire, P. E. J.; Newberry, J.; Campbell, H. E. Automated SEM-EDS (QEMSCAN) mineral analysis in forensic soil investigations; testing instrumental variability. In Criminal and Environmental Soil Forensics; Ritz, K., Dawson, L., Miller, D., Eds.; Springer: New York, 2009; pp 411−430.
(26) Sutherland, D. Estimation of mineral grain size using automated mineralogy. Miner. Eng. 2007, 20, 452−460. (27) Andersen, J.C.Ø.; Rollinson, G.; Snook, B.; Herrington, R.; Fairhurst, R. Use of QEMSCAN for the characterization of Ni-rich and Ni-poor goethite in laterite ores. Miner. Eng. 2009, 22, 1119−1129. (28) Rollinson, G. K.; Andersen, J. C. Ø.; Stickland, R. J.; Boni, M.; Fairhurst, R. Characterisation of non-sulphide zinc deposits using QEMSCAN. Miner. Eng. 2011, 24, 778−787. (29) Gottlieb, P.; Wilkie, G.; Sutherland, D.; Ho-Tun, E.; Suthers, S.; Perera, K.; Jenkins, B.; Spencer, S.; Butcher, A.; Rayner, J. Using quantitative electron microscopy for process mineralogy applications. JOM 2000, 24−25. (30) Grobéty, B.; Gieré, R.; Dietze, V.; Stille, P. Airborne particles in the urban environment. Elements 2010, 6, 229−234. (31) Nygaard, U. C.; Samuelsen, M.; Aase, A.; Løvik, M. The capacity of particles to increase allergic sensitization is predicted by particle number and surface area, not by particle mass. Toxicol. Sci. 2004, 82, 515−524. (32) Jones, T. P.; Williamson, B. J.; BéruBé, K. A.; Richards, R. J. Microscopy and chemistry of particles collected on TEOM filters: Swansea, South Wales, 1998−1999. Atmos. Environ. 2001, 35, 3573− 3583. (33) HSE: Methods for the determination of hazardous substances Fibres in air; (MDHS #8), Health and Safety Executive: Sudbury, UK, 1998. (34) Amato, F.; Pandolfi, M.; Viana, M.; Querol, X.; Alastuey, A.; Moreno, T. Spatial and chemical patterns of PM10 in road dust deposited in urban environment. Atmos. Environ. 2009, 43, 1650− 1659. (35) Yeung, Z. L. L.; Kwok, R. C. W.; Yu, K. N. Determination of multi-element profiles of street dust using energy dispersive X-ray fluorescence (EDXRF). Appl. Radiat. Isot. 2008, 58, 339−346. (36) Ondov, J. M.; Wexler, A. S. Where do particulate toxins reside? An improved paradigm for the structure and dynamics of the urban mid-Atlantic aerosol. Environ. Sci. Technol. 1998, 32, 2547−2555. (37) Seaton, A.; Mac Nee, W.; Donaldson, K.; Godden, D. Particulate air pollution and acute health effects. Lancet 1995, 1, 176−178.
5577
dx.doi.org/10.1021/es305025e | Environ. Sci. Technol. 2013, 47, 5570−5577
