Motor Vehicle Contributions to Ambient PM10 and PM2.5 at Selected ...

3 downloads 36 Views 236KB Size Report
monitoring sites across the USA: Westbury, NY; ... Westbury, NY, 3, 11- to 12-h samples; (2) Albany, ...... K. Willeke & P. Baron (Eds.), Aerosol measurement:.
Environ Monit Assess (2007) 132:155–163 DOI 10.1007/s10661-006-9511-3

Motor Vehicle Contributions to Ambient PM10 and PM2.5 at Selected Urban Areas in the USA Mahmoud Abu-Allaban & John A. Gillies & Alan W. Gertler & Russ Clayton & David Proffitt

Received: 1 May 2006 / Accepted: 15 September 2006 / Published online: 14 December 2006 # Springer Science + Business Media B.V. 2006

Abstract A source apportionment study was carried out to estimate the contribution of motor vehicles to ambient particulate matter (PM) in selected urban areas in the USA. Measurements were performed at seven locations during the period September 7, 2000 through March 9, 2001. Measurements included integrated PM2.5 and PM10 concentrations and polycyclic aromatic hydrocarbons (PAHs). Ambient PM2.5 and PM10 were apportioned to their local sources using the chemical mass balance (CMB) receptor model and compared with results obtained using scanning electron microscopy (SEM). Results indicate that PM2.5 components were mainly from combustion sources, including motor vehicles, and secondary species (nitrates and sulfates). PM10 consisted mainly of geological material, in addition to emissions from combustion sources. The

M. Abu-Allaban Department of Water Management and Environment, Heshemite University, P.O. Box 150459, Zarga 13115, Jordan J. A. Gillies (*) : A. W. Gertler Division of Atmospheric Sciences, Desert Research Institute, University and Community College System of Nevada, 2215 Raggio Parkway, Reno, NV 89512, USA e-mail: [email protected] R. Clayton : D. Proffitt ARCADIS Geraghty and Miller, 4915 Prospectus Drive, Durham, NC 27713, USA

fractional contributions of motor vehicles to ambient PM were estimated to be in the range from 20 to 76% and from 35 to 92% for PM2.5 and PM10, respectively. Keywords Air pollution . Particulate matter . Road dust . Source attribution . Chemical mass balance . Scanning electron microscopy

1 Introduction With the adoption of the newly revised National Ambient Air Quality Standard (NAAQS) for particulate matter, all industrial sectors as well as US states and local municipalities have needs to better understand their contributions to ambient fine particulate matter. Most fine particles, those with less than 2.5 μm aerodynamic diameter, are produced by combustion sources such as automobiles and trucks, while many of the particles between 2.5 and 10 μm are fugitive dusts such as those from roadways. Therefore development of effective pollution control strategies for the control of airborne particulate matter derived from motor vehicles is an important issue in air quality management. Identifying pollution sources (particularly motor vehicles) and determining their contributions to the observed levels in the atmosphere is necessary for resolving air pollution problems. Source apportionment techniques are applied in air quality studies in order to attribute the relative amounts of measured

156

atmospheric contaminants in a selected region to their local sources. Source apportionment modeling attempts to attribute atmospheric contaminants to local sources by relating chemical and physical properties of the source material to the properties observed at a receptor site. In this paper two approaches, chemical mass balance (CMB) and scanning electron microscopy (SEM), are utilized in order to estimate the fractional contribution of motor vehicles to the ambient PM10 and PM2.5 levels within urban areas at different US cities.

2 Ambient Measurements 2.1 Sampling sites Ambient data were collected during the period September 7, 2000 through March 9, 2001 at seven monitoring sites across the USA: Westbury, NY; Albany, NY; Birmingham, AL; Houston, TX; El Paso, TX; Las Vegas, NV; and Long Beach, CA. Sites were chosen to represent a range of climatic and air quality conditions including the Los Angeles Basin (high PM2.5, nitrates, and semi-volatiles), the western USA (high elevation, wind blown dust, and carbon aerosols), the southern USA (high temperatures and humidity), and the northeastern USA (high sulfates and low temperatures). The two primary considerations for site selection were to use existing national or state local air monitoring sites (NAMS/ SLAMS) that are near roadways and downwind with respect to prevailing winds. Other considerations for selecting sites within a geographical area included roadway grade, curvature and surface quality, nearby sources of non-roadway airborne PM, typical traffic composition (trucks versus cars), and density and variability of traffic (rush-hour time periods).

Environ Monit Assess (2007) 132:155–163

(6) El Paso, TX, 9, 10- to 16-h samples; (7) Houston, TX, 8, 8- to 14-h samples. Samples were collected during both daytime and nighttime hours. In each location, two medium-volume samplers designed to collect samples for chemical analyses were utilized. This type of sampler employs a SierraAndersen 254 PM10 inlet or Bendix PM2.5 cyclone (Chan & Lippmann, 1997) to determine the size fractions collected. The ambient air is transmitted through the size-selective inlet into a plenum, then drawn simultaneously through two parallel filter packs, one with a ringed 47 mm Teflon-membrane filter (Gelman Scientific, Ann Arbor, MI) and one with a 47 mm quartz-fiber filter (Pallflex, Putnam, CT). In addition to the quartz and Teflon samples, polycyclic aromatic hydrocarbons (PAHs) were also collected. The PAHs were critical to apportion the carbon components of the PM based on the uniqueness of PAH compounds associated with motor vehicles and other combustion sources (Fujita et al., 1998). A sequential fine particulate/semi-volatile organic compounds sampler (PSVOC sampler) was used for PAH sampling. Samples were collected on a Teflon-impregnated glass fiber filter (TIGF) followed by a backup solid adsorbent to account for the total ambient concentrations of semi-volatile organic species that occur in both the particle and gas phases. The solid adsorbent was a polyurethane foam (PUF)–XAD-4 resin sandwich cartridge (PUF/XAD/PUF) that is used to collect semi-volatile compounds (SVOC). Samples for SEM analysis were collected on polycarbonate (PC) filters using an Andersen dichotomous sampler, which separates each PM10 sample into fine (particles with aerodynamic diameter ≤2.5 μm) and coarse fractions (particles with aerodynamic diameter >2.5 and ≤10 μm).

3 Analytical Methods 2.2 Measurement methods Ambient PM2.5, and PM10, samples were collected on a 10- to 16-h basis at the seven locations using the sampling protocol described by Watson et al. (1994). The breakdown of the samples by location is: (1) Westbury, NY, 3, 11- to 12-h samples; (2) Albany, NY, 4, 12-h samples; (3) Birmingham, AL, 6, 6- to 12-h samples; (4) Long Beach, CA, 5, 11- to 13-h samples; (5) Las Vegas, NV, 7, 10- to 12-h samples;

The Teflon-membrane filters were used for gravimetric mass and chemical analyses by x-ray fluorescence where a basic suite of elemental concentrations (Mg, Al, Si, P, S, Cl, K, Ca, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Br, Rb, Sr, Ba, U and Pb) is determined. The quartz-fiber filters were used for ions following the methodology described by Watson and Chow (1993). Organic and elemental carbon components were measured by thermal/optical reflectance (TOR) on

Environ Monit Assess (2007) 132:155–163

0.5 cm2 punches taken from one half of the quartzfiber filter (Chow et al., 1993). The PAH samples were first extracted with deuterated internal standards added to confirm extraction efficiencies (Zielinska et al., 1998). The extracts are combined and reduced in volume then analyzed by gas chromatograph/ mass spectrometry (GC/MS) in the selected ion monitoring mode. The SEM analysis was conducted by RJ Lee Group (Monroeville, PA) using the secondary and backscattered electron modes at magnifications typically ranging from ×20 to ×20,000. The secondary electron signal produces an image with a threedimensional perspective, high depth-of field, and the appearance of overhead illumination. The backscattered electron signal yields an image containing compositional information because the signal is dependent on the atomic number of the feature being examined. Higher atomic numbered elements generate more backscattered electrons and therefore, appear brighter than features consisting of lighter elements. The backscattered electron signal permits light element species (e.g., carbon particles) to be easily distinguished from particles having a higher atomic number (e.g., geological, metals, etc.). 3.1 CMB method The CMB consists of a set of linear equations that express the ambient concentrations of a chemical species as the sum of products of source contributions and source composition profiles. The forms of these equations are: Ci ¼ Fi1 S1 þ Fi2 þ ::: þ Fij Sj þ ::: þ FiJ SJ þ ei þ eij; i ¼ 1; 2; ::I j ¼ 1; 2; :::; J where Ci is the concentration of species i measured at a receptor site, Fij is the fraction of species i in emissions from source j, Sj is the estimate of the contribution of source j, ei and eij are the random measurement errors of the ith species in the ambient sample and in the jth source, respectively, I is the number of chemical species, and J is the number of source types. The source contributions are the unknowns in these equations. The current CMB software (EPA/DRI versions 7 and 8) applies the effective variance least-squares solution developed and tested by Watson, Cooper, and

157

Huntzicker (1984) to solve these equations. The CMB procedure requires several steps. First, the contributing sources must be identified and their chemical profiles selected. Then the chemical species to be included in the model must be selected. The next step is the estimation of the fractions of each chemical species contained in each source type and the estimation of the uncertainties in both the ambient concentrations and source contributions. The final step is the solution of the set of chemical mass balance equations. These procedures are described in detail in an application and validation protocol by Pace and Watson (1987). The CMB output contains statistics that can be used to evaluate how well the model’s calculated species concentrations match the ambient measurements for these species. The most important of these are: (1) the source contribution estimates (SCEs) and their uncertainties; (2) “chi-square,” the weighted sum of the squares of the differences between calculated and measured species concentrations. Values between one and two indicate acceptable fits; values less than one indicate very good fits to the data; (3) “R-square,” the fraction of the variance in the measured concentrations accounted for by the variance in the calculated species concentrations. Values of “R-square” greater than 0.9 indicate a good fit to the measured data; and (4) “percent mass,” expresses the percent of total mass accounted for by the source contribution estimates. Values between 80 and 120% are considered to be acceptable. 3.2 SEM method Another method for estimating the contribution of different sources to the observed ambient PM concentrations is to apply SEM analytical techniques to characterize the collected particulates. The SEM is useful in distinguishing particles originating from different sources based on their composition and morphological characteristics (Figure 1). In this study the SEM was used to provide this information for individual fine and coarse particles. One time-integrated sample was examined for each of the test locations. The average number of particles examined per sample was 1,200. The size range of the particles examined was 0.2 to 10 μm. According to Mamane, Gillies, Connor, and Gertler (1998), individual particle characteristics can be used to attribute the material to a source. For example, the presence of fly ash

158

Environ Monit Assess (2007) 132:155–163

Figure 1 Secondary electron image and elemental spectrum of: a carbon chain agglomerate, b fine carbon particles, c geological material, d sodium/sulfur-rich particles, e sodium/chlorinerich particles, and f biological material.

(spherical silicon/aluminum-rich [SAS] particles) or particles of biological origin can be distinguished from mineral grains by appearance alone.

4 Results and Discussion 4.1 Mass and major chemical species Table I summarizes the average mass concentrations of major species detected at the seven sampling locations. The minimum detection limits (Table I) are calculated as three times the standard deviation of multiple analyses of known zero concentrations from laboratory filter blanks. The measurement precision of a species is based on propagating the precision of the volumetric measurements, the chemical composition measurements, and the blank variability. Not all species are shown in Table I because their relative concentrations are small, nevertheless they were used

in the CMB modeling, provided they were greater than their detection limit, for their important role in differentiating between sources that produce similar emissions. The primary species were found to be organic carbon (OC), elemental carbon (EC), sodium (Na), aluminum (Al), silicon (Si), calcium (Ca), and iron (Fe). PM2.5 mainly consisted of EC, OC, nitrates, and sulfates, with a small component of elements that suggest a geological source. In addition to OC and EC, PM10 samples were rich in geological related elements including Al, Si, and Ca. The significant presence of the OC and EC in the PM2.5 and PM10 samples reflects the important contribution from combustion sources (mainly motor vehicles) to the ambient PM levels at the sampling locations. 4.2 CMB modeling results Profiles representing the major sources assembled from previous studies (Fujita et al., 1998; Hildemann,

Environ Monit Assess (2007) 132:155–163

159

Table I Detection limits and average PM2.5 and PM10 concentrations for major species (μg/m3) Species

MDL1

Size

Albany

Birmingham

El Paso

Houston

Las Vegas

Long Beach

Westbury

Chloride

0.03 0.03

Sulfate

0.03

Ammonium

0.03

Soluble Sodium

0.004

Soluble Potassium

0.003

Organic Carbon

0.05

Elemental Carbon

0.05

Sodium

0.02

Magnesium

0.006

Aluminum

0.002

Silicon

0.002

Sulfur

0.001

Chlorine

0.002

Potassium

0.002

Calcium

0.001

Iron

0.0004

0.4±0.1 4.7±0.5 3.6±0.2 4.0±0.2 1.9±0.1 2.1±0.1 1.7±0.1 1.7±0.1 0.2±0.0 4.8±0.3 0.1±0.0 0.1±0.0 5.6±0.7 9.4±0.8 2.6±0.2 2.2±0.2 0.6±0.5 5.0±0.6 0.1±0.1 0.1±0.1 0.0±0.0 0.3±0.1 0.1±0.1 1.3±0.5 0.6±0.0 0.7±0.0 0.2±0.0 5.0±2.0 0.1±0.0 0.3±0.1 0.1±0.0 0.7±0.1 0.1±0.0 0.4±0.0 16.2±0.8 32.5±2.2 20.9±1.2 34.9±2.0

0.1±0.1 0.1±0.1 1.8±0.1 2.0±0.1 2.8±0.2 2.5±0.2 1.5±0.1 1.3±0.1 0.0±0.0 0.1±0.0 0.0±0.0 0.1±0.0 5.7±0.8 5.7±0.8 3.0±0.2 2.6±0.2 0.1±0.8 0.6±0.8 0.0±0.1 0.0±0.1 0.0±0.0 0.4±0.1 0.2±0.1 1.8±0.6 0.9±0.0 0.8±0.0 0.0±0.0 0.0±0.0 0.0±0.0 0.2±0.0 0.1±0.0 0.7±0.1 0.1±0.1 0.5±0.1 15.4±0.9 17.8±1.1 16.2±1.1 24.1±1.4

0.0±0.1 0.1±0.1 0.2±0.1 0.5±0.1 0.6±0.1 0.8±0.1 0.2±0.1 0.2±0.1 0.0±0.0 0.2±0.0 0.0±0.0 0.1±0.0 1.5±1.1 3.7±1.1 1.1±0.3 1.7±0.5 0.2±0.6 0.3±0.8 0.1±0.1 0.1±0.1 0.2±0.0 1.8±0.7 0.4±0.0 7.7±3.1 0.2±0.0 0.3±0.0 0.0±0.0 0.2±0.1 0.1±0.0 0.6±0.2 0.3±0.0 2.6±0.5 0.2±0.0 1.1±0.1 4.9±1.1 21.4±3.4 5.5±0.6 31.5±1.9

0.1±0.1 0.2±0.1 1.7±0.1 2.0±0.1 1.2±0.1 1.3±0.1 0.9±0.1 0.8±0.1 0.0±0.0 0.1±0.0 0.1±0.0 0.1±0.0 3.6±0.7 5.0±0.8 2.0±0.2 2.7±0.2 0.5±0.6 0.3±0.7 0.0±0.1 0.0±0.1 0.1±0.0 0.2±0.1 0.3±0.1 1.3±0.5 0.4±0.0 0.5±0.0 0.0±0.0 0.1±0.0 0.1±0.0 0.1±0.0 0.3±0.0 1.0±0.2 0.2±0.1 0.5±0.1 10.4±0.8 15.3±1.0 12.4±0.9 17.6±1.2

0.0±0.0 0.1±0.0 0.3±0.0 0.6±0.1 0.4±0.1 0.8±0.1 0.2±0.0 0.2±0.0 0.0±0.0 0.2±0.0 0.0±0.0 0.0±0.0 1.1±0.8 3.2±0.8 0.9±0.2 1.3±0.3 0.3±0.6 0.0±0.8 0.0±0.1 0.2±0.1 0.0±0.0 0.4±0.1 0.1±0.1 2.4±0.8 0.2±0.0 0.3±0.0 0.0±0.0 0.1±0.0 0.0±0.0 0.2±0.0 0.2±0.0 2.4±0.4 0.1±0.0 0.6±0.0 3.4±0.9 12.4±1.3 2.5±0.6 18.1±1.1

0.4±0.1 0.6±0.1 12.6±0.7 15.0±0.8 1.5±0.1 1.7±0.1 4.3±0.3 4.7±0.3 0.1±0.0 0.4±0.0 0.1±0.0 0.2±0.0 4.8±0.7 7.2±0.8 3.2±0.2 3.6±0.3 0.3±0.7 0.3±0.8 0.0±0.1 0.1±0.1 0.0±0.0 0.5±0.2 0.1±0.1 1.7±0.6 0.6±0.0 0.7±0.0 0.2±0.0 0.5±0.2 0.1±0.0 0.3±0.1 0.0±0.0 0.5±0.1 0.2±0.0 0.9±0.1 27.2±1.1 37.4±1.4 30.0±1.7 46.8±2.6

0.1±0.1

Nitrate

PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10 PM2.5 PM10

Sum of Species Mass a

0.069

1.7±0.1 3.6±0.3 1.9±0.1 0.1±0.0 0.0±0.0 3.6±0.7 2.5±0.2 0.2±0.7 0.0±0.1 0.0±0.0 0.1±0.1 1.4±0.1 0.0±0.0 0.0±0.0 0.0±0.0 0.1±0.0 13.8±0.8 11.5±0.9

MDL Minimum detection limit (μg/m3 )

No PM10 samples were collected at the Westbury site.

Markowski, & Cass, 1991; and Rogge, Hildemann, Mazurek, & Cass, 1993) were used in the CMB analysis. Initial tests with different combinations of PM2.5 source profiles were done to determine which profiles best explain the ambient data and the robustness of the results with respect to choice of source profiles. CMB performance measurements were examined to determine how well the ambient

concentrations were explained by the CMB source contribution estimates. Examples of the PM2.5 test apportionments are presented as a series of trials representing different combinations of source profiles in Table II. The last highlighted column represents the best fit of the data. Determining the profile combination that represents the acceptable apportionment was based on R2, chi2, percent calculated mass, and

160

Environ Monit Assess (2007) 132:155–163

Table II Source profiles sensitivity test results for CMB for PM2.5 from four selected sites Source profile

1st Group 2nd Group 3rd Group 4th Group 1st Group 2nd Group 3rd Group 4th Group

Long Beach 2/16/01 2.00 2.04 13.07 13.04 – – 0.52 0.54 3.88 4.13 1.66 1.28 – – 1.40 – 0.75 – – 1.89 24.59 24.59 23.30 22.92 2.99 2.99 2.56 2.28 2.76 2.76 3.32 3.46 0.85 0.84 1.93 1.86 Albany 12/8/00 Ammonium sulfate 2.05 2.05 Ammonium nitrate 5.42 5.43 Light-duty, gasoline no. 1 – 0.21 Light-duty, gasoline, no. 2 – – Heavy-duty diesel 1.40 – Woodstove burning hardwood – – Fireplace burning softwood 2.02 1.57 Brake-wear – – Road dust – – Composite road dust 1.53 1.58 Measured mass 15.13 15.13 Calculated mass 12.42 12.31 Measured organic carbon 1.67 1.67 Calculated organic carbon 2.08 1.89 Measured elemental carbon 1.30 1.30 Calculated elemental carbon 1.31 1.31 R-square 0.82 0.81 Chi-square 1.7 1.7

Ammonium sulfate Ammonium nitrate Light-duty, gasoline no. 1 Light-duty, gasoline, no. 2 Heavy-duty diesel Woodstove burning hardwood Fireplace burning softwood Brake-wear Road dust Composite road dust Measured mass Calculated mass Measured organic carbon Calculated organic carbon Measured elemental carbon Calculated elemental carbon R-square Chi-square

2.04 13.04 0.66 – 3.50 – 1.32 – 1.54 – 24.59 22.11 2.99 2.12 2.76 3.00 0.91 0.81

2.00 13.07 0.75 – 3.14 2.25 – 1.40 0.70 – 24.59 23.32 2.99 3.10 2.76 2.80 0.87 1.42

2.05 5.42 0.21 – – 1.49 – – – 1.65 15.13 12.06 1.67 1.42 1.30 1.49 0.80 1.7

2.05 5.43 0.21 – 1.36 – 1.65 – – 1.57 15.13 12.27 1.67 1.87 1.30 1.31 0.82 1.74

Birmingham 12/17/00 1.85 1.85 1.06 1.06 3.40 2.86 – – – 0.27< 1.25 1.40 – – – – – – 0.13< 0.16< 3.51 3.51 7.69 7.60 2.66 2.66 2.84 2.74 1.16 1.16 1.10 1.16 0.98 0.98 0.3 0.3 Westbury 9/7/00 3.92 3.94 2.88 2.88 – – 0.89 0.86 3.02 2.91 – 3.16 3.53 – – – – – 1.85 1.96 17.77 17.77 16.09 15.71 4.91 4.91 4.37 3.26 2.54 2.54 2.67 3.36 0.86 0.83 1.35 1.63

1.84 1.06 3.89 – – – 0.7< – – −0.01 3.51 7.49 2.66 2.52 1.16 1.22 0.99 0.2

1.84 1.05 – 2.28 – 2.05 – – – – 3.51 7.23 2.66 2.69 1.16 1.20 0.97 0.49

– – – – – – – – – – – – – – – – – –

3.90 2.89 2.83 – 2.05 – 2.83 – – 1.68 17.77 16.18 4.91 4.43 2.54 2.61 0.87 1.38

The < sign means that the value is uncertain. The predicted mass at Birmingham is apparently over predicted (approximately twice as much), but when the ambient data were checked, it appears that the ambient mass is under measured.

percent calculated organic and elemental carbon. The same exercise was carried out for the PM10 samples, but the results are not shown. The average CMB apportionment results for each site are shown in Table III. The results indicate that PM2.5 is primarily from tailpipe emissions, with some secondary particulate (sulfate and nitrate) from background sources. There was a small contribution from

road dust. Cooking and vegetative burning sources from nearby activities were also observed. Major PM10 sources included tailpipe emissions, road dust, and background particulate (sulfate and nitrate). A brake-wear contribution was observed mainly in the PM10 fraction. Due to the fact that road dust profiles were, to some extent, contaminated by brake-dust, accounting for brake-wear depended on both the

Environ Monit Assess (2007) 132:155–163

161

Table III Average PM2.5 and PM10 source attribution results for the seven sampling locations (average±standard deviation, μg/m3) City

Size

Ammonium sulfate

Ammonium nitrate

Sodium Chloride

Vegetative burning

Tailpipe

Brakewear

Road Dust

Primary motor vehicles (%)

Albany Birmingham Houston Long Beach Las Vegas El Paso Westbury Albany Birmingham Houston Long Beach

PM2.5 PM2.5 PM2.5 PM2.5 PM2.5 PM2.5 PM2.5 PM10 PM10 PM10 PM10

2.4±0.2 3.7±0.3 1.6±0.1 2.1±0.2 0.5±0.1 0.7±0.1 5.2±0.4 2.7±0.2 3.7±0.3 1.6±0.1 2.1±0.2

4.6±0.4 2.4±0.3 2.2±0.3 16.3±1.5 0.3±0.1 0.3±0.1 2.2±0.3 4.9±0.4 2.7±0.2 2.5±0.2 18.9±1.2

0.0±0.0 0.0±0.0 0.0±0.0 0.0±0.0 0.0±0.0 0.0±0.0 0.0±0.0 11.7±1.7 0.1±0.0 0.1±0.0 0.8±0.1

5.5±0.7 3.3±0.8 3.1±0.7 4.6±1.1 1.0±0.4 0.7±0.3 1.7±0.9 7.6±0.7 3.9±0.9 3.1±0.7 4.6±1.1

2.9±0.5 5.7±0.7 2.6±0.5 4.1±0.7 1.5±0.3 2.0±0.7 5.3±0.8 2.9±0.5 5.7±0.7 3.8±0.6 6.5±0.9

0.0±0.0 0.0±0.0 0.0±0.0 0.4±0.2 0.0±0.0 0.3±0.2 0.0±0.0 0.0±0.0 0.0±0.0 0.7±0.5 3.5±0.7

1.9±0.6 1.4±0.5 2.6±0.9 1.3±0.2 2.0±0.5 2.8±0.5 0.7±0.3 11.7±1.5 8.4±1.3 6.9±1.2 9.6±0.6

27 43 43 20 65 76 40 35 58 61 43

brake-wear and road dust profiles used in the model. Thus, the road dust profile with the lowest brake-wear contamination would yield the highest estimate for the brake-wear contribution. The last column in Table III shows the percentage contribution of motor vehicles to the observed ambient PM at sampling locations based on the sum of the tailpipe, brake-wear, and road dust contributions. Since this assumes that all the geological material is due to resuspended road dust, this is likely to overestimate the primary mobile source contribution. It is, however, important to note that the contribution of mobile sources to the observed secondary particulate levels has not been included. Since motor vehicles are significant sources of the precursors of these species, if this contribution were included, the mobile source contribution to the observed PM would be significantly higher. 4.3 Scanning electron microscopy The SEM results are presented in Table IV. The values presented in Table IV represent the percent, by mass, of the fraction of particles of a given size class that are associated with a chemical composition class (e.g., Carbon-, or Na/Cl-rich). Particle mass was estimated based on the assumption that the identified particles are spherical, which provides an estimate of particle volume and that particle density is a function of its identified chemical composition. The results indicate that the majority of the fine fraction samples consisted primarily of sub-micron carbonaceous particles. Moderate to major concen-

trations of sulfur species were observed from the majority of the eastern monitoring sites and several of the western sites. Other particles consistently observed on the fine samples were moderate to minor amounts of geological material, and minor to trace concentrations of iron-rich particle types, primarily iron/barium species. Particles in the coarse fraction samples ranged from sub-micron to 10 μm, although the sample volume was primarily attributed to particles greater than 2 μm. The coarse material associated with most samples consisted primarily of Si/Al-rich and Si-rich particles indicative of geological material. Carbonaceous particles were observed at moderate to minor concentrations. However, unlike the carbonaceous particles observed on the fine samples, the majority of the carbon mass observed on the coarse samples was often biological, consisting mainly of pollen and spores. Similar to the fine samples, Fe-rich particles often containing a trace amount of Si, S, and Ba were observed on most of the coarse samples at minor to trace concentrations. In some cases, Fe-rich particles had a spherical morphology, indicative of a high temperature combustion source. The El Paso, TX sample revealed a moderate contribution of Ca species and the Albany, NY sample had a significant amount of Na/Cl-rich (salt) particulate matter. High Ca in El Paso, TX could be soil derived. The salt particulate detected on the Albany sample collected in December was presumably related to road salt. Although sub-micron carbonaceous particles are not necessarily unique to vehicular emissions, there is a

162

Environ Monit Assess (2007) 132:155–163

Table IV Summary of the SEM results Location

Size

Albany

Fine

Albany

Major (>40%)

C-rich (submicron and a trace of biological) Coarse Na/Cl-rich

Birmingham Fine Birmingham Coarse Houston Fine C-rich (submicron) Houston

Moderate (20–40%)

Coarse

C-rich (submicron and a trace of biological) Long Beach Coarse Crustal

Minor (5–20%)

Trace (