Sources of Fine Particulate Material along the Wasatch Front - Energy

---

282

Energy & Fuels 2002, 16, 282-293

Sources of Fine Particulate Material along the Wasatch Front Russell W. Long, Rachel Smith, Scott Smith, Norman L. Eatough, Nolan F. Mangelson, and Delbert J. Eatough* Department of Chemistry and Biochemistry, Brigham Young University, Provo, Utah 84602

C. Arden Pope, III Department of Economics, Brigham Young University, Provo, Utah 84602

William E. Wilson U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711 Received July 17, 2001. Revised Manuscript Received October 18, 2001

The concentration and composition of PM2.5 has been measured with a variety of continuous and integrated samplers at the Hawthorne EPA Environmental Monitoring for Public Awareness and Community Tracking (EMPACT) sampling site in Salt Lake City, UT, and at an EPA Science to Achieve Results (STAR) sampling site in Bountiful, UT. Data are considered at both sites during a 10-day winter period with high PM2.5 concentrations due to winter inversions, and at the Hawthorne site only during a fourteen day summer period when the site was impacted by smoke from wildfires in the Wasatch Mountains. The PM2.5 was dominated by organic material and ammonium nitrate in the winter and by organic material in the summer. In both cases, substantial amounts of sem-volatile material, SVM, was present which was not measured by a Tapered Element Oscillating Microbalance (TEOM) monitor but was detected by a Real-Time Ambient Mass Sampler (RAMS). The PM2.5 data have been combined with concentrations of particulate soot and soil corrected potassium, and with gas-phase concentrations of NOx, CO, and SO2 to apportion the PM2.5 to primary emissions emitted by and secondary particulate material formed from emissions from mobile sources, wood smoke combustion (including the forest fire emissions), and oil refineries located near the Bountiful site on an hourly basis. Wood smoke and mobile source emissions dominated the contribution to the non-volatile fraction of the PM2.5, with about equal contributions for each, but very different diurnal patterns for the two contributions. The diurnal patterns for the attribution were consistent with the emission patterns for these two sources. In addition, the oil refineries contributed about 10% of the primary PM2.5 at Bountiful. The SVM could not be directly attributed to any source but appeared to be secondary ammonium nitrate and organic material formed from gaseous emissions from the various sources. The dominant contributors to this SVM were NOx and gas-phase organic compounds in emissions from both mobile sources and wood smoke.

Introduction Human health endpoints associated with exposure to airborne particulate matter (PM) include increased mortality and morbidity from respiratory and cardiopulmonary disease.1-3 These effects are observed with exposure to concentrations substantially below the U.S. PM10 ambient air quality standard. The observed exacerbation of health problems is believed to be associ* Corresponding author. (1) Pope, C. A., III. Epidemiology of fine particulate air pollution and human health; Biological Mechanisms and who’s at risk? Environ. Health Perspect. 2000, 108 (Sup 4), 713-723. (2) Schwartz, J.; Dockery, D. W.; Neas, L. M. Is daily mortality associated specifically with fine particles? J. Air Waste Manage. Assoc. 1996, 46, 927-939. (3) U.S. Environmental Protection Agency. Air Quality Criteria for Particulate Matter, 1996; Environmental Protection Agency, Research Triangle Park, NC, EPA/600/P-95/001aF.

ated more closely with exposure to fine particles than coarse particles. As a result, the U.S. Environmental Protection Agency has promulgated4 revised standards for PM, which establishes new annual and 24-hour fine particulate standards with PM2.5, measured according to the Federal Reference Method (PM2.5 FRM), as the indicator. This recognition of fine and coarse particles as different classes of PM pollutants is an advance in the understanding and control of PM. However, ambient fine particulate matter is not a single pollutant, but a mixture of many chemical species. Major components include the following: sulfate, nitrate, ammonium, and hydrogen ions; trace elements (including toxic and transition metals); organic material; elemental carbon (4) Schaefer, G.; Hamilton, W.; Mathai, C. V. Implementing the revised NAAQS and the FACA subcommittee for ozone, particulate matter and regional haze. Environ. Man. 1997, October, 22-28.

10.1021/ef010168l CCC: $22.00 © 2002 American Chemical Society Published on Web 01/17/2002

Fine Particulate Material along the Wasatch Front

(or soot); and crustal components. Stable species such as trace elements, crustal elements, and sulfate can be accurately measured by single filter samplers such as the PM2.5 FRM (Federal Reference Method).4,5 However, semi-volatile fine particulate species, e.g., these in equilibrium between the gas and particulate phases in the atmosphere with significant concentrations in both phases such as ammonium nitrate (in equilibrium with gas-phase nitric acid and ammonia) and some organic material,6-9 are not collected quantitatively by these techniques. A third semi-volatile component present in ambient particulate matter is water. However, the postsample equilibration proceedures used with the PM2.5 FRM4,5 or the heated filters of the continuous samplers used in this study are intended to give a water-free PM2.5 mass. Past toxicological studies of respirable particulate matter have focused on the evaluation of exposure to genotoxic compounds.10-13 The most extensive of these studies have been those conducted under the umbrella of the Integrated Air Cancer Project of EPA11,13,14 (IACP). Results of the IACP studies have shown that significant concentrations of both gas- and particulatephase mutagens are present in urban atmospheres and that the greatest exposure to these mutagens is associated with the products of photochemistry of NOx and automotive and wood-smoke emissions.13,15,16 Nitrated mutagenic compounds are rapidly formed in wood smoke17,18 and automobile emissions19 in the presence of O3 and NO2 and a large increase in both gas and particulate phase mutagenicity results from this chemistry.20,21 It is reasonable to hypothesize that the semi(5) Musick, D. A summary of the ambient air program for PM2.5. Environ. Manager 2000, February, 17-20. (6) Eatough, D. J.; Obeidi, F.; Pang, Y.; Ding, Y.; Eatough, N. L.; Wilson, W. E. Integrated and real-time diffusion denuder samplers for PM2.5 based on BOSS, PC, and TEOM technology. Atmos. Environ. 1999, 33, 2835-2844. (7) Hering, S.; Cass, G. The magnitude of bias in the measurement of PM2.5 arising from volatilization of particulate nitrate from Teflon filters. J. Air Waste Manage. Assoc. 1995, 49, 725-733. (8) Lewtas, J.; Booth, D.; Pang, Y.; Reimer, S.; Eatough, D. J.; Gundel, L. Comparison of sampling methods for semi-volatile organic carbon (SVOC) associated with PM2.5. Aerosol Sci. Technol. 2001, 34, 9-22. (9) Pang, Y.; Eatough, D. J.; Eatough, N. L. Determination of PM 2.5 Sulfate and Nitrate With a PC-BOSS Designed for Routine Sampling for Semi-Volatile Particulate Matter. J. Air Waste Manage. Assoc. 1999, 49. (10) Hannigan, M. P.; Cass, G. R.; Lafleur, A. L.; Longwell, J. P.; Thilly, W. G. Bacterial mutagenicity of urban organic aerosol sources in comparison to atmospheric samples. Environ. Sci. 1994. (11) Lewtas, J. Emerging methodologies for assessment of complex mixtures. Application of bioassays in the Integrated Air Cancer Project. Adv. Mod. Environ. Toxicol. 1991, 19, 137-146. (12) MacGregor, J. T.; Claxton, L. D.; Lewtas, J.; Jensen, R.; Lower, W. R.; Pesch, G. G. Monitoring environmental genotoxicants. Methods Genet. Risk Assess. 1994, 171-243. (13) Lewis, C. W.; Stevens, R. K.; Zweidinger, R. B.; Claxton, L. D.; Barraclough, D.; Klouda, G. Source apportionment of mutagenic activity of fine particle organics in Boise, Idaho. 1991, Proceedings, 84th Annual Meeting; Air and Waste Management Association, paper 131.3. (14) IACP Integrated Air Cancer Project Study, Session papers in measurement of toxic and related air pollutants. Proc. EPA/APCA Symposium on Measurement of Toxic and Related Pollutants, 1988, Air Pollution Control Association, pp 799-895. (15) Lewis, C. W.; Baumgardner, R. E.; Stevens, R. K.; Claxton, L. D.; Lewtas, J. Contribution of woodsmoke and motor vehicle emissions to ambient aerosol mutagenicity. Environ. Sci. Technol. 1988, 22, 968971. (16) Walsh, D.; Warren, S.; Zweidinger, R.; Claxton, L.; Lewtas, J. Mutagenicity of indoor and outdoor air in Boise, Idaho and Roanoke, Virginia. 1993, Proc. EPA/APCA Symposium on Measurement of Toxic and Related Pollutants, Air Pollution Control Association, pp 190196.

Energy & Fuels, Vol. 16, No. 2, 2002 283

volatile mutagenic material formed in urban atmospheres may also contribute to the exacerbation of respiratory and cardiovascular health effects. However, the accurate assessment of effects associated with exposure to these semi-volatile secondary organic compounds will require monitoring techniques for the correct determination of these semi-volatile particulate components.6,13,22 Results from EPA-funded projects, EMPACT (the Salt Lake City EPA Environmental Monitoring for Public Awareness and Community Tracking) and STAR (Science to Achieve Results) grants, have validated a newly developed Real-time Ambient Mass Sampler, RAMS,6,22 for total (non-volatile plus semi-volatile) PM2.5 mass. These projects also are examining the health relevance of the RAMS method as compared to other measurements of PM2.5 parameters.23 Ammonium nitrate and semi-volatile organic compounds, SVOC, are significant components of fine particles in most western urban atmospheres. However, due to the large loss of semivolatile material from the filter during sampling, these components are not measured correctly with current EPA accepted methods such as the PM2.5 FRM. Data on the carbonaceous components of the fine particles, at sampling sites in Salt Lake City, UT, (Hawthorne site) and Bountiful, UT, have demonstrated that during both summer and winter periods organic material is the major component of PM2.5 and that a substantial fraction of this organic material is composed of semi-volatile compounds which are not retained by a filter during sampling.23 The PM2.5 data from these sites are combined with other atmospheric data indicative of emissions from various types of sources to estimate the probable sources of this fine particulate organic material. Experimental Methods and Procedures Sampling Sites. The EPA EMPACT monitoring site is located immediately adjacent to the State of Utah Air Quality Monitoring site at the Hawthorne Elementary School in Salt Lake City, UT. The site is in a residential area approximately 4 km southeast of the central business district. The population (17) Kamens, R. M.; Bell, D. A.; Dietrich, A.; Perry, J. M.; Goodman, R. G.; Claxton, L. D.; Tejada, S. Mutagenic transformations of dilute wood smoke systems in the presence of ozone and nitrogen dioxide. Analysis of selected high-pressure liquid chromatography fractions from wood smoke particle extracts. Environ. Sci. Technol. 1985, 19, 63-69. (18) Kamens, R. M.; Rives, G. D.; Perry, J. M.; Bell, D. A.; Paylor, R. F., Jr.; Goodman, R. G.; Claxton, L. D. Mutagenic changes in dilute wood smoke as it ages and reacts with ozone and nitrogen dioxide: an outdoor chamber study. Environ. Sci. Technol. 1984, 18, 523-530. (19) Kleindienst, T. E.; Smith, D. F.; Hudgens, E. E.; Snow, R. F.; Perry, E.; Claxton, L. D.; Bufalini, J. J.; Black, F. M.; Cupitt, L. T. The photo-oxidation of automobile emissions: measurements of the transformation products and their mutagenic activity. Atmos. Environ. 1992, 26A, 3039-3053. (20) Cupitt, L. T.; Claxton, L. D.; Shepson, P. B.; Kleindienst, T. E. IACP emissions: transformations and fate. Proc. EPA/APCA Symposium on Measurement of Toxic and Related Air Pollutants, 1987, Air Pollution Control Association, pp 597-604. (21) Cupitt, L. T.; Claxton, L. D.; Kleindienst, T. E.; Smith, D. F.; Shepson, P. B. Transformation of Boise sources: the production and distribution of mutagenic compounds in wood smoke and auto exhaust. Proc. EPA/APCA Symposium on Measurement of Toxic and Related Pollutants, 1988, Air Pollution Control Association, pp 885-889. (22) Eatough, D. J.; Eatough, N. L.; Obeidi, F.; Pang, Y.; Modey, W.; Long, R.W. Continuous determination of PM2.5 mass, Including Semi-Volatile Species. Aerosol Sci. Technol. 2001, 34, 1-8. (23) Long, R. W.; Thompson, W.; Eatough, N. L.; Eatough, D. J.; Pope, C. A.; Wilson, W. E. The Salt Lake City EPA EMPACT Program: I. Sampling plan and first years results, 2001, Proceedings, A&WMA 93rd Annual Conference & Exhibition, in press.

284

Energy & Fuels, Vol. 16, No. 2, 2002

in Salt Lake Valley where the site is located is about 3/4 million. PM2.5 pollution at the site is expected to be dominated by primary emissions and secondary products from mobile source emissions and, during the winter, wood smoke from home combustion. There are no major point sources which are expected to make a significant impact at the Hawthorne site. The EPA STAR monitoring site is located immediately adjacent to the State of Utah Air Quality Monitoring site at the Bountiful Fire Station in downtown Bountiful. The Bountiful site is 20 km north of the Hawthorn site and is sheltered from direct transport of pollutants from the Hawthorne site by a range of mountains which extend from the main Wasatch Mountains out into the valley between the two sites. The population in the Bountiful area is about 100 000. The region has no major business center such as exists in Salt Lake, but is a major commuter traffic feed to the Salt Lake area. Immediately to the west of Bountiful are a series of oil refineries. Bountiful is expected to be impacted by mobile source emissions, wood smoke emissions, and emissions from these refineries during transport from the west. Sampling Periods. Data are being obtained throughout the year at the EMPACT Hawthorne site and during both winter and summer periods at the STAR Bountiful site. Data given in this paper are taken from results obtained at both Hawthorne and Bountiful from 28 December 2000 through 6 January 2001. During this 10-day period, the Wasatch Front was influenced by a generally persistent high pressure system, resulting in valley inversions and generally stagnant air conditions. The high pressure was established at the start of the period. The inversion weakened on 29 December, to be reestablished later that day. A weak front passed through the region on 2 January, followed immediately by a persistent inversion until the end of the period. Skies were clear and there was no precipitation throughout the period. Data are also presented for a 14-day period at the Hawthorne site only from 3 July through 5 August 2000. The region was under the influence of a low pressure system through this time period. Skies were generally clear with no precipitation except for very scattered showers the last 2 days. The west was influenced by wild fires in Idaho and Montana throughout this period. Wildfires started in the Wasatch Mountains to the east of Salt Lake beginning about 31 July, with maximum impact in the Salt Lake Valley occurring August 1-4. These nearby wildfires resulted in a significant increase in PM2.5 at the Hawthorne site. Sampling Methods. A complete description of all samplers used at the EMPACT and STAR sites has been given.23 Sampling methods used to obtain the data given in this paper follow. PM2.5 FRM. The FRM5 contains a 2.5 µm WINS impactor followed by a single filter in a temperature-controlled environment. Collected mass is determined gravimetrically. Protocols for pre- and post- equilibration and weighing are well defined.5 The FRM data were obtained by the State of Utah Division of Air Quality at both Bountiful and Hawthorne. R&P TEOM Monitor. One-hour-averaged, PM2.5 mass concentrations are determined at both sites using an R&P TEOM monitor operating under normal conditions.24 As stated above, semi-volatile PM will evaporate at the standard operating temperature of the instrument, 50 °C, which is required to remove particle-bound water.25 This technique measures only PM which is non-volatile at the TEOM filter temperature. PC-BOSS. The combination of technology used in the HighVolume Brigham Young University Organic Sampling System (BIG BOSS)26 and the Harvard particle concentrator27 has (24) Patashnick, H.; Rupprecht, E. G. Continuous PM10 measurements using the tapered element oscillating microbalance. J. Air Waste Manage. Assoc. 1991, 41, 1079-1083. (25) Mignacca, D.; Stubbs, K. Effects of Equilibration temperature on PM10 Concentrations from the TEOM Method in the Lower Fraser Vally. J. Air Waste Manage. Assoc. 1999, 49, 1250-1254.

Long et al. resulted in the Particle Concentrator-Brigham Young University Organic Sampling System.6,8,28 The configuration and operation of the PC-BOSS as used in the Salt Lake EMPACT program has been previously described.8 The PC-BOSS was used at both sites to determine fine particulate mass, trace elements, crustal material, sulfate, carbonaceous material (elemental and organic), nitrate, semivolatile organic material, and semi-volatile nitrate. Samples for the chemical characterization of PM2.5 in the minor flow following a particle concentrator and a BOSS diffusion denuder are collected in a filter pack containing a pre-fired 47 mm quartz filter (Pallflex) followed by 47 mm charcoal-impregnated glass fiber filter (CIG, Schliecher and Schuell) to determine fine particulate sulfate, and carbonaceous material and nitrate, including semi-volatile species lost from the particles during sampling. A second parallel filter pack containing a 47 mm Teflon (Whatman) filter followed by a 47 mm Nylon (Gelman, Nylasorb) filter is used to determine PM2.5 filter-retained (non-volatile) mass, sulfate and nitrate, plus any nitrate lost from the particles during sample collection. A side flow filter pack, prior to the particle concentrator, containing a 47 mm polycarbonate (Corning, 0.4 µm pore size) filter followed by a 47 mm CIG collects particles (excluding semivolatile species lost during sampling) and gas-phase organic material after the 2.5 µm inlet cut. These data are compared to data from the minor flow filters to determine the particle concentrator efficiency.6,8,28 The filters were also used to determine elemental content by PIXE. Twenty-four-hour PCBOSS samples were collected each day during both winter and summer sampling periods for comparison with equilibrated PM2.5 mass concentrations measured on a 24-hour basis by the PM2.5 FRM. RAMS. The Real-time total Ambient Mass Sampler (RAMS), based on diffusion denuder, Nafion dryer, and TEOM monitor technology, was used at both sites for the real-time determination of total PM2.5 mass, including semi-volatile species.6,22 The RAMS measures total PM2.5 mass with a TEOM monitor using a “sandwich” filter to retain semi-volatile material, SVM, which would be lost from particles in a conventional TEOM monitor. The sandwich filter consists of a Teflon-coated particle collection filter (R&P TX40) followed by a charcoal-impregnated glass fiber filter (CIG, Schliecher and Schuell) to collect any semi-volatile nitrate and organic compounds lost from the particles during sampling. Care must be taken to remove from the sample stream all gas-phase species that can be absorbed by the CIG filter in order to prevent over determination of PM2.5 mass. The configuration and operation of the RAMS, as used in the EMPACT study, have been described previously.22,23 RAMS data were averaged over 1-h periods for comparison with 1-h averaged TEOM data. Twenty-four-hour averaged data were also calculated for comparison with results obtained with the PC-BOSS and PM2.5 FRM samplers. Anderson Aethalometer. An Anderson Instruments Inc. (model RTAA-900) Aethalometer was used for the determination of aerosol elemental carbon (soot) on a continuous basis during the winter at the Bountiful site and during the summer at the Hawthorne site. One-hour average elemental-carbon (soot) concentrations were obtained and compared with the 1-h averaged TEOM and RAMS data. Gas-Phase Data. Hourly averaged concentrations of CO (winter only), NOx, SO2 (Bountiful only), and O3 (summer only) (26) Tang, H.; Lewis, E. A.; Eatough, D. J.; Burton, R. M.; Farber, R. J. Determination of the particle size distribution and chemical composition of semi-volatile organic compounds in atmospheric fine particles. Atmos. Environ. 1994, 28, 939-947. (27) Sioutas, C.; Koutrakis, P.; Burton, R. M. Development and evaluation of a low cutpoint virtual impactor. Aerosol Sci. Technol. 1994, 21, 223-235. (28) Ding, Y.; Pang, Y.; Eatough, D. J. A high volume diffusion denuder sampler for the routine monitoring of fine particulate matter: I. Design and optimization of the PC-BOSS. J. Aerosol Sci. Technol. 2001, in press.

Fine Particulate Material along the Wasatch Front

Energy & Fuels, Vol. 16, No. 2, 2002 285

were obtained from the State of Utah Department of Environmental Quality. Analytical Methods. Microbalance. A Mettler UMT2 Microbalance was used for the determination of collected fine particle mass on a 47 mm Teflon (Whatman) filter sampled from the PC-BOSS minor flow. TPV Analysis. Temperature-Programmed Volatilization26,29 was used in the analysis of collected samples for total carbonaceous material. In this method, the various sampled filters are heated from ambient temperature at a predetermined ramp rate to a predetermined termination temperature. The ramp rate and termination temperatures are dependent on the type of filter being analyzed. Quartz filters are heated to 800 °C in an N2/O2 atmosphere. Charcoal-impregnated filters are heated to 400 °C in an N2 atmosphere. Carbon in compounds desorbed from the filters during the heating process is catalytically converted to CO2 and detected by nondispersive infrared absorption. The particulate carbonaceous material retained on the quartz filters is referred to as non-volatile carbon in this paper. Carbonaceous compounds lost from particles during sampling and collected by the CIG are referred to as semi-volatile organic compounds, SVOC. Dionex Ion Chromatograph. A Dionex model 500 ion chromatograph with a separator column, anion fiber suppressor, and conductivity detector was used for the analysis of collected samples for nitrate and sulfate. Sample filters are ultrasonically extracted and the concentrations of sulfate and nitrate in the various solutions determined by peak area measurement and comparison to standards. PIXE. Proton-Induced X-ray Emission (PIXE) was used for the determination of trace metals in samples collected on Nuclepore filters in the PC-BOSS side-flow.30

Results and Discussion Comparison of RAMS and TEOM Monitor Results. The 1-h average RAMS and TEOM monitor results obtained at both Bountiful and Hawthorne during the 10-day winter study are given in Figures 1A and 1B. These figures illustrate changing PM2.5 concentrations over the 10-day period with the formation and passage and reformation of high-pressure systems (inversions) over the Wasatch Front. The breakdown of the inversion late on December 28 and reformation later that day and the similar change on January 1 and 2 are evident in the data. RAMS PM2.5 mass measurements are generally equal to or greater than the TEOM PM2.5 mass measurements preceding, during and following the inversions for both sites. Peaks in the PM2.5 concentration at Bountiful tend to occur every day between 2 and 4 p.m., with occasional late night (e.g., December 30) or morning (e.g., January 2 and 3) peaks, Figure 1A. The pattern at the Hawthorne site is somewhat different, Figure 1B, with peaks in the PM concentration occurring each morning and each evening. Figure 1C contains the 1-h averaged RAMS and TEOM monitor results obtained during the two-week period of the summer 2000 study. During the first week (July 2329), PM2.5 concentrations were generally low and variable. The peak in PM2.5 the evening of July 24 is associated with emissions from firework displays on Pioneer Day, a state holiday. During the second week (29) Ellis, E. C.; Novakov, T. Application of thermal analysis to the characterization of organic aerosol particles. Sci. Total Environ. 1982, 23, 227-238. (30) Mangelson, N. F.; Hill, Nielson, K. K.; Eatough, D. J.; Christensen, J. J.; Izatt, R. M.; Richards, D. O. Proton-induced X-ray emission analysis of Pima Indian autopsy tissues. Anal. Chem. 1979, 51, 1187-1194.

Figure 1. One-hour average RAMS (total PM2.5) and TEOM monitor (non-volatile PM2.5) results during 28 December 2000 through 5 January 2001 at the Bountiful (A) and Hawthorne (B) sampling sites and at Hawthorne from 23 July through 5 August 2001 at Hawthorne (C). The difference between the two measurements is semi-volatile material (SVM) lost from the heated (50 °C) filter of the TEOM.

(July 30 - August 5), concentrations were higher due to the influence of forest fires in the nearby Wasatch Mountains. The peak in PM concentrations during this time period occurred shortly after noon each day. As in the winter study period, RAMS PM2.5 measurements are equal to or greater than TEOM measurements. Consistent differences in RAMS and TEOM monitor data were seen each afternoon in the winter data, Figures 1A and 1B. The greatest difference between RAMS and TEOM measurements in the summer is observed over the time period of August 01-04, 2000, when the area was most impacted by the smoke from the nearby forest fires, Figure 1C. For the winter sampling period the RAMS and TEOM PM2.5 mass averages were 46 and 29 µg/m3, respectively, at Bountiful and 55 and 29 µg/m3, respectively, at Hawthorne. For the summer sampling period, the RAMS and TEOM averages were 19 and 14 µg/m3,

286

Energy & Fuels, Vol. 16, No. 2, 2002

Long et al.

Table 1. Results of the Statistical Analysis of PM2.5 Samples during the Episodes Studied X vs Y

n

R2 a

slopea

intercepta µg/m3

Bountiful, 28 Dec 2000-5 Jan 2001 Samples RAMS Total vs TEOM Non-Volatile Massb 236 0.82 0.63 ( 0.01 0 ( 6.5 0.82 0.61 ( 0.02 1.4 ( 6.4 Soot vs NOx Calculated Mobile Source PM2.5b 214 0.82 0.85 ( 0.02 0 ( 2.0 0.95 0.99 ( 0.02 0.5 ( 2.3 RAMS Total vs TEOM Non-Volatile Massb

PC-BOSS Total vs PM2.5 FRM Massc PM2.5 FRM vs TEOM Non-Volatile Massc

Hawthorne, 28 Dec 2000-5 Jan 2001 Samples 225 0.09 0.50 ( 0.01 0 ( 10.9 0.24 0.29 ( 0.03 13.2 ( 9.9 Bountiful and Hawthorne Winter Samples 9 0.61 1.01 ( 0.04 0 ( 6.3 0.72 0.73 ( 0.17 16.6 ( 5.7 10 0.13 0.50 ( 0.03 0 ( 5.8 0.82 0.61 ( 0.02 1.4 ( 6.4

Hawthorne, 23 July - 5 August 2001 Samples 335 0.55 0.59 ( 0.01 0 ( 8.0 0.47 0.47 ( 0.02 5.1 ( 7.5 Soot vs NOx Calculated Wood Smoke PM2.5b 295 0.82 1.03 ( 0.01 0 ( 3.3 0.83 0.93 ( 0.02 1.6 ( 2.3 6 0.74 0.54 ( 0.06 0 ( 4.6 PC-BOSS Total vs PM2.5 FRM Massc 0.75 0.60 ( 0.17 -2.1 ( 5.1 PM2.5 FRM vs TEOM Non-Volatile Massc 8 0.94 1.17 ( 0.04 0 ( 1.9 0.82 0.61 ( 0.02 1.4 ( 6.4 RAMS Total vs TEOM Non-Volatile Massb

RAMS Measured vs PC-BOSS Constructed Massc

28

All Samples 0.83 0.97 ( 0.03 0.84 0.89 ( 0.08

0 ( 6.9 3.8 ( 6.9

σd µg/m3

X average µg/m3

X-Y bias

45.6

16.4

NAe

9.1

1.5

1.3

54.5

25.6

NAe

57.5

-1.1

3.3

58.2

29.1

NAe

29.1

10.4

5.2

21.8

12.0

-0.9

2.2

17.6

29.6

13.8

NAe

14.6

-3.0

NAd

42.3

0.7

4.8

σ%

15.5

5.7

11.5

a R2, Slope, and Intercept values are given for (1) zero intercept and (2) calculated intercept. Uncertainties are the standard error of the slope and y estimates. b 1-h data. c 24-h data. d The σ values are biased corrected, see ref 8. e NA, σ could not be calculated because the sampler bias was greater than σ.

respectively, the first week, and 39 and 23 µg/m3, respectively, the second week. The higher RAMS PM2.5 mass, compared to that of the TEOM monitor, can be attributed to the presence of semi-volatile PM which is measured by the RAMS, but not by the TEOM. The linear regression results of RAMS total vs TEOM nonvolatile PM2.5 mass are given in Table 1 for each study location and period. Comparison of PC-BOSS, FRM, and, RAMS Results. Integrated PC-BOSS samples were used to determine the composition and mass of PM2.5 including material retained on a filter and semi-volatile nitrate and organic material lost from the filter during sampling. PC-BOSS sulfate and nitrate were assumed to be present as ammonium sulfate and ammonium nitrate, respectively. It was assumed that fine particulate organic material was 61% carbon, typical of an aged urban aerosol.31 However, during the sampling period impacted by fresh smoke from the forest fires, organic material was assumed to be 71% carbon, more typical of a young aerosol.31 Crustal material was estimated from the Al and Si data, using crustal averages32 for these elements, and constructed mass was obtained for each PC-BOSS sample. Constructed mass was calculated as the sum of estimated crustal material, ammonium sulfate obtained from the Teflon and quartz filters, the ammonium nitrate from the Teflon and quartz filters, the volatile ammonium nitrate from the Nylon and CIG filters, the soot collected on the quartz filters, and the total organic carbonaceous material which is the sum of the organic material on the quartz (31) Turpin, B. J.; Lim, H. J. Species contributions to PM2.5 mass concentrations: Revisiting common assumptions for estimating organic mass. Aerosol Sci. Technol. 2001, 35, 602-610. (32) CRC Handbook of Chemistry and Physics, 102nd ed.; Weast, R. C., Ed.; CRC Press: Boca Baton, Fl, 2001.

filter and the semi-volatile organic material lost from the particles during sampling but collected on the charcoal-impregnated glass fiber filter. Daily PM composition results are given in Figure 2 for each study period. Average PM2.5 composition during the various study periods is given in Table 2. The linear regression results for PC-BOSS constructed PM2.5 mass vs PM2.5 FRM mass for the winter samples are given in Table 1 and Figure 3. During the winter sampling period, PC-BOSS and FRM results were comparable. Linear regression of the PC-BOSS total vs FRM mass results gives R2 ) 0.61 (n ) 9) and a slope of 1.01 ( 0.04 for this time period. This can be attributed to cold, humid meteorological conditions which tend to stabilize semi-volatile species. However, during the summer sampling periods, PC-BOSS mass measurements were much higher than those obtained by the PM2.5 FRM Table 1 and Figure 3. These time periods were dominated by hot and dry weather conditions, resulting in a larger contribution to total PM2.5 mass by semi-volatile species. These semi-volatile species were not retained by the Teflon filter of the PM2.5 FRM during the summer. The TEOM monitor data did not agree with either the RAMS or PM2.5 FRM during the winter sampling periods, Table 1. For both comparisons, the TEOM monitor results averaged only about half of the result obtained by the other samplers. This can be attributed to the loss of semi-volatile material at the elevated temperature of the TEOM monitor filter. In contrast, during the summer time period the TEOM monitor agreed with the PM2.5 FRM sampler but not with the PC-BOSS sampler. Semi-volatile material is lost by both the TEOM monitor and the PM2.5 FRM sampler during the hot and dry summer sampling period.

Fine Particulate Material along the Wasatch Front

Figure 2. Twenty-four-hour average PM2.5 composition determined from PC-BOSS samples during 28 December 2000 through 5 January 2001 at the Bountiful (A) and Hawthorne (B) sampling sites and at Hawthorne from 23 July through 5 August 2001 at Hawthorne (C).

The results obtained by the RAMS for the continuous determination of PM2.5 mass, including semi-volatile particulate material were validated by comparison with results obtained from the PC-BOSS. The linear regression of RAMS-measured vs PC-BOSS-constructed PM2.5 mass results for all samples with a zero intercept gives R2 ) 0.83 (n ) 28) with a slope of 0.97 ( 0.03. The results for all sample locations and seasons were comparable, Figure 4. Results obtained with the RAMS and the PC-BOSS show that PM2.5 mass, including semi-volatile species, can be continuously and accurately monitored (with a PM2.5 precision of (12%) with the RAMS and that the PM2.5 FRM and TEOM monitor can significantly underestimate total PM2.5 mass concentrations. Sources of Fine Particulate Organic Material during PM Episodes. In addition to the mass and composition of PM2.5, species were measured which are expected to be markers of emission from combustion sources in the study area. These included fine particulate soot and potassium, and gas-phase NOx, CO, and SO2. While these species are not expected to be unique to any source, the relative ratio of these species from the various sources will be different. These differences were used in a chemical mass balance source estimation.

Energy & Fuels, Vol. 16, No. 2, 2002 287

As indicated by the data given in Figure 2 and in Table 2, organic material was a major fraction of the PM2.5 for all periods and ammonium nitrate was a major fraction during the winter. An additional quantity given in Table 2 is SVOC. Particulate SVOC is operationally defined here as the semi-volatile organic material that evaporates from the particles collected on the first filter in the PC-BOSS and is collected on the second charcoalimpregnated filter. Gas-phase SVOC in the text refers to the gas-phase semi-volatile organic material collected on the charcoal filter of the side-flow of the PC-BOSS. In addition to the directly determined PM2.5 components, a quantity called SVM (semi-volatile material) is also given in Table 2. SVM is calculated as the difference between the RAMS and TEOM measurement. Comparison of the RAMS, FRM, and TEOM data indicate that more PM2.5 material is lost from the collected particles at the elevated (50 °C) temperature of the TEOM probe than is lost from either the filters of the PC-BOSS or the FRM. SVM includes both organic material and ammonium nitrate lost from the TEOM probe. Possible sources of this SVM can be estimated by comparison of the TEOM and SVM concentrations (RAMS-TEOM) with various potential tracers of emissions sources. Potential markers of emission sources available in the data set include elemental K, soot, NOx, CO, and SO2. Soot will be a primary pollutant emitted mainly from diesel engines and wood smoke, including both home combustion during the winter time periods and the wildfires present during the summer episode. Fine particulate potassium will be a marker of emissions from the combustion of wood. However, some of the PM2.5 K will have originated from crustal sources.13 The measured fine particulate K was corrected for crustal K using the Si and Al concentrations and expected crustal composition.32 This results in Kcor concentrations which are expected to be associated with wood combustion emissions. NOx and CO will be emitted from all potential combustion sources. This will include diesel emission sources, other automotive emission sources, combustion of wood, and emissions from the oil refinery complex (which are expected to impact the Bountiful sampling site). SO2 will be a unique tracer of emissions from the oil refineries at the Bountiful site. A recently completed apportionment study at Lindon, UT,33 in the Wasatch Front region has indicated that the expected ratio of K/soot from wood combustion in this area will be 0.87 ( 0.16 g Kcor/g soot. Other expected ratios for wood smoke emissions are 1.33 ( 0.17 mol NOx/g soot and 18.5 ( 3.4 mol CO/g soot34 and about 100-200 g PM/g Kcor.13 Ratios expected for the mix of diesel and other automotive sources along the Wasatch Front are 0.018 ( 0.001 g Kcor/g soot, 200 ( 30 mol NOx/g soot and 9100 ( 1600 mol CO/g soot.34 These values were used in evaluating potential sources of SVM and total carbonaceous material during the three study periods. Oxygenated fuels were used at Lindon during the apportionment study, but these fuels are not man(33) Ren, Y. Oxygenated fuel use and formation of sulfate in the atmosphere. Ph.D. Dissertation, Brigham Young University, Provo, UT, 1999. (34) Pope, A. C., III; Eatough, D. J.; Gold, D. R.; Yanbo, P.; Nielsen, K. R.; Nath, P.; Verrier, R. L.; Kanner, R. E. Acute exposure to environmental tobacco smoke and heart rate variability. Environ. Health Perspect. 2001, 109 (7), 711-716.

288

Energy & Fuels, Vol. 16, No. 2, 2002

Long et al.

Table 2. Average Composition of PM2.5 during the Various Study Periods, µg/m3 sampling period

ammonium sulfate

non-volatile ammonium nitrate

non-volatile organic

soot

crustal

lost nitrate

SVOCa

SVMb

Bountiful Winter Samples

11.9

15.1

10.5

2.2

0.7

2.2

3.1

17.1

Hawthorne Winter Samples

11.4

13.9

15.8

2.8

1.1

3.9

5.8

25.2

Hawthorne 23 Jul- 29 Jul

1.6

0.6

9.7

1.3

1.6

0.1

3.0

5.1

Hawthorne 30 Jul-5 Aug

3.7

1.6

26.8

1.7

2.4

0.3

5.0

15.6

a SVOC is defined as the organic material collected on the second (carbon-impregnated) filter of the PC-BOSS. b SVM is defined as the RAMS minus the TEOM PM2.5 measurement.

Figure 3. Comparison of PM2.5 FRM measured and PC-BOSS constructed PM2.5 for samples collected during the winter at Bountiful (B) and Hawthorne (H) and during the summer (S) at Hawthorne.

Figure 5. One-hour average RAMS (PM2.5), SVM (RAMS minus TEOM monitor results), Soot (measured with an Anderson Aethalometer), SO2, NOx, and CO at the Bountiful sampling site from 28 December 2000 through 5 January 2001.

Figure 4. Comparison of RAMS measured and PC-BOSS constructed PM2.5 for samples collected during the winter at Bountiful (B) and Hawthorne (H) and during the summer (S) at Hawthorne.

dated at either Bountiful or Salt lake. In addition, the relative importance of diesel vehicles, relative to gasolinepowered vehicles in the region near the sampling sites will be lowest at Bountiful, high at Lindon, and highest at Hawthorne. Thus, the ratios for mobile sources found in the Lindon study will be indicative of ratios expected at the Bountiful and Hawthorne sites, but the actual values may vary somewhat from these ratios. Sources of PM2.5 at Bountiful. The temporal variations in total PM2.5, SVM, soot, SO2, NOx, and CO at

Bountiful during the winter episode are given in Figure 5. The concentrations of NOx and CO were highly correlated at Bountiful for all periods except a few evening and early morning hours on December 28 and 29. The linear regression results for this comparison, excluding these 14 data points, Table 3, gives a CO/NOx mol ratio of 6.7 ( 0.1. However, the ratio of the concentrations of NOx to soot was variable, Figure 6. In addition, the concentrations of soot, CO, and NOx were not correlated with the concentrations of SO2. Sulfur dioxide was elevated each morning, peaking between 6 and 8 a.m., Figure 5. Concentrations were much lower during the rest of the day. This diurnal change in SO2 concentrations is consistent with the oil refineries being the principal source of SO2-associated pollutants. During the early morning, up-canyon winds in the Wasatch Range to the east will carry pollutants

Fine Particulate Material along the Wasatch Front

Figure 6. Comparison of 1-h average Soot and NOx concentrations at the Bountiful sampling site from 28 December 2000 through 5 January 2001. The NOx concentrations have been corrected for NOx associated with SO2 from the nearby refineries, see text. The slopes of the dashed and solid lines are the ratios of NOx/Soot associated with emissions from wood smoke (47 mol NOx/g Soot) and mobile source (113 mol NOx/g soot) emissions, respectively, determined from comparison of the TEOM monitor data with the soot and NOx data, see text. Table 3. Linear Regression Results for the Comparison of NOx and CO Concentrations X vs Y

n

r2

slope

intercept (units)

Bountiful NOx vs CO

217

0.96

6.65 ( 0.09

390 ( 100 (mol CO/mol NOx)

220

0.95

Hawthorne NOx vs CO all data

12.7 ( 0.2

-170 ( 310 (mol CO/mol NOx)

from the refineries to the sampling site. Concentrations will decrease as these winds decrease and vertical mixing increases with the daytime thermal warming. Analysis of the data indicated that concentrations of CO, NOx, and soot associated with the SO2 are all small compared to the average concentrations of these species. Any attempt to associate part of these species to the SO2 emission from the oil refinery significantly deteriorated the correlations among these species. We therefore conclude that the main source of NOx, CO, and soot is associated with wood smoke and mobile emissions and the main source of SO2 is the oil refinery complex.

Energy & Fuels, Vol. 16, No. 2, 2002 289

Possible contributing sources to the non-volatile fraction of PM2.5 as measured by the TEOM monitor were investigated by multi-linear regression analysis of the measured concentrations against the concentrations of SO2, soot, NOx, and CO. The latter two species were highly correlated as shown in Table 3. Thus, inclusion of more than one of these species resulted in collinearity problems in the regression analysis. The correlation including SO2 and any one or two of the other species was strongest for soot and weakest for CO. However, the results were all generally comparable. For all cases, about 7 µg/m3 (of an average of 28 µg/m3) was not explained by the multi-linear regression analysis and an average of 5 µg/m3 was associated with the SO2. Comparison of the TEOM monitor PM2.5 mass and the SO2 data suggests an upper limit to the amount of nonvolatile mass associated with the SO2 of 0.40 µg/mol SO2. This coefficient for SO2 was used to correct the data for non-volatile PM2.5 from the oil refineries. The concentrations of soot and the SO2 source corrected nonvolatile PM2.5, PM2.5cor, are shown in Figure 6. As indicated in Figure 6A, the data are (within the uncertainty in the data) contained within the upper and lower bounds of 20 and 4.9 g non-volatile PM2.5cor/g soot. On the basis of earlier results,13,33 the higher ratio is assumed to be associated with PM2.5 from mobile sources and the lower ratio with PM2.5 from wood smoke. Related coefficients of 9.4 and 0.96 g non-volatile PM2.5/mol NOx for mobile sources and wood smoke, respectively, were obtained by the same method. These two sets of coefficients were then used to estimate the hourly contribution of these two sources to non-volatile PM2.5cor. The results obtained for PM2.5 from mobile sources by the two sets of coefficient were in agreement, and indicate that the precision of the mobile emissions source apportionment is (16%, or (1.3 µg/m3, Table 1. A second check on the validity of the above source apportionment analysis may be obtained from the comparison of NOx and soot concentrations for the Bountiful data set with the upper and lower limit ratios over these two parameters calculated from the PM2.5cor/ soot and NOx/soot parameters. The results of this comparison, Figure 6B, show that the analysis is selfconsistent. The sources of non-volatile PM2.5 obtained from the above-described analysis for the Bountiful winter samples are shown in Figure 7. The average concentration of non-volatile PM2.5 from the three sources was 4.0, 12.5, and 12.9 µg/m3 from the refinery, mobile, and wood smoke sources, respectively. The low concentrations of PM2.5 attributed to these sources on January 2 are due to the passage of a frontal system. However, the low concentrations on January 1 probably reflect reduced traffic on this day. The patterns and concentrations of non-volatile PM2.5 from the refinery source are similar each day of the study. However, the patterns for nonvolatile PM2.5 from the wood smoke and mobile sources are different during the 28-31 December holiday period and during the 3-6 January period after the start of the new year. The concentrations of non-volatile PM2.5 associated with wood smoke emissions averaged 14.5 µg/m3 in December during the holiday period and 10.4 µg/m3 in January. The concentrations of non-volatile PM2.5 from mobile source emissions averaged 11.0 and

290

Energy & Fuels, Vol. 16, No. 2, 2002

Long et al.

Figure 8. One-hour average RAMS (PM2.5), SVM (RAMS minus TEOM monitor results), NOx, and CO at the Hawthorne sampling site from 28 December 2000 through 5 January 2001.

Figure 7. One-hour average contribution of refinery, wood smoke, and mobile emissions to non-volatile PM2.5 and SVM (RAMS minus TEOM monitor results) at the Bountiful sampling site from 28 December 2000 through 5 January 2001.

13.4 µg/m3 respectively, during these two period. Higher peak concentrations were seen in December but the mobile source associated non-volatile PM2.5 was spread out over a longer time period after the holidays, probably due to different commuter traffic patterns after the New Year. The absence of a commuter-associated mobile source peak on 6 January is because that day was a Saturday. The sources of semi-volatile PM2.5 (semi-volatile material, SVM, the difference between the RAMS and TEOM measurements) are more difficult to deduce from the data. Multi-linear regression analyses indicated that the concentrations of SVM were not associated with concentrations of SO2, NOx, CO, or soot. These species were able to account for less than 10% of the average 17 µg/m3 of SVM. The SVM was best correlated with time, Figure 7. Concentrations of SVM were low every morning and peaked every afternoon. These results suggest that the SVM is secondary material formed by photochemical processes from NOx and gas-phase organic material each day. The majority of this SVM is organic, Table 2, and the source emissions which are present when the SVM is formed included both wood smoke and mobile sources with the majority being mobile source emissions. The 24-h PC-BOSS data indicate that an average of only 2% of the gas-phase

SVOC is present as SVOC in particles. The sum of SVOC (average of 5.8 µg/m3) and ammonium nitrate (average of 2.2 µg/m3) lost from particles in the PC-BOSS is less than the average of 17 µg/m3 SVM lost from the TEOM monitor. This could represent additional loss from the TEOM filter of the organic material (average of 16 µg/m3) or ammonium nitrate (average of 14 µg/m3) not lost from the filters of the PC-BOSS. Sources of PM2.5 at Hawthorne. The temporal variations in total PM2.5, SVM, NOx, CO, and (through 31 December) non-volatile carbon at Hawthorne during the winter episode are given in Figure 8. The concentrations of NOx and CO were also highly correlated at Hawthorne. The linear regression results for this comparisons, Table 3, gives a CO/NOx mol ratio of 12.7 ( 0.2, about twice the ratio seen at Bountiful. This increased ratio may reflect the expected increased importance of diesel emissions in Salt Lake City. As was the case in Bountiful, the concentrations of NOx and TEOM measured non-volatile PM2.5 were variable. Using the same approach as was used in Bountiful, coefficients of 9.8 and 1.08 g non-volatile PM2.5/mol NOx were obtained for the mobile source and wood smoke emissions, respectively. The coefficients for the wood smoke and mobile source emissions were essentially the same for the Hawthorne and Bountiful data. The sources of non-volatile PM2.5 obtained using these coefficient for the Hawthorne winter data set are shown in Figure 9. The average concentration of non-volatile PM2.5 from the two sources was calculated to be 6.2 and 18.9 µg/m3 from the mobile and wood smoke sources, respectively, in December and 15.0 and 15.8 µg/m3 from the mobile and wood smoke sources, respectively, in January. Again, the low concentrations of PM attributed to these sources on January 2 are due to the passage of a frontal system. However, the effect of Saturday (January 6) on the PM concentrations was less than that seen in Bountiful. The patterns for non-volatile PM2.5 from the wood smoke and mobile sources are different

Fine Particulate Material along the Wasatch Front

Energy & Fuels, Vol. 16, No. 2, 2002 291

Figure 10. One-hour average RAMS (PM2.5), SVM (RAMS minus TEOM monitor results), Soot and NOx at the Hawthorne sampling site from 23 July through 5 August 2001.

Figure 9. One-hour average contribution of wood smoke and mobile emissions to non-volatile PM2.5 and SVM (RAMS minus TEOM monitor results) at the Hawthorne sampling site from 28 December 2000 through 5 January 2001.

during the 28-31 December holiday period and during the 3-6 January period after the start of the new year. The clear early morning small peak and afternoon larger peak due to mobile source emissions is seen in January, but not in December. The relationship between SVM and the presence of non-volatile PM2.5 from mobile source and wood smoke emissions is not as clear for the Hawthorne results. The concentrations of SVM at Hawthorne averaged 50% higher than those at Bountiful. An average of 3% of the gas-phase SVOC was present as particulate phase SVOC. The highest concentrations of SVM were not associated with the highest concentrations of nonvolatile PM2.5 from either mobile sources or wood smoke. However, the correlation is slightly better for wood smoke emissions. Evidence is seen for the production of SVM after peaks of both mobile source and wood smoke non-volatile PM2.5. Sources of PM2.5 at Hawthorne during the Summer. Both NOx and soot data were available at Hawthorne for assistance in interpreting the data during the summer period when emissions from wildfires were present, Figure 10. The analysis of the summer nonvolatile PM2.5, soot and NOx data lead to the ratios of 1.6 g non-volatile PM2.5/mol NOx and 4.2 g non-volatile PM2.5/g soot for non-volatile PM2.5 associated with wood smoke. These ratios are comparable to those found at both Hawthorne and Bountiful for the winter data, with the ratio for PM2.5/NOx being slightly higher than that found during the winter. The difference may be due to the predominance of emissions from forest fires during the summer. However, the determined ratios of PM2.5 to NOx and soot for mobile source emissions (62 g PM2.5/ mol NOx and 35 g PM2.5/g soot) are about 50% higher for soot, but about six times higher for NOx. The average concentrations of soot at Hawthorne during July (before

the major impact from forest fire emissions) are about half the concentrations seen at Hawthorne or Bountiful during the winter, Table 2. The concentrations of NOx at Hawthorne during the winter varied from about 100 to 500 ppb, with strong diurnal patterns related to morning and evening rush hour periods, Figure 8. While similar diurnal patterns were seen at Hawthorne during the summer, Figure 10, the maximum concentrations were only about 150 ppb. The lower concentrations can be attributed to higher mixing depths during the summer. The lack of a similar decease in the concentrations of soot suggests that there was impact of distant forest fire emissions throughout July and August, leading to increased background concentrations of both NOx and soot from these distant sources. This probably accounts for the increased ratios of PM2.5 to NOx and soot inferred for mobile source emissions. Thus, the unusually high non-volatile PM2.5/NOx ratio can be attributed, in part, to PM2.5 from distant forest fires. The comparison of wood smoke associated non-volatile PM2.5 calculated from the soot and the NOx data were in very good agreement, Table 1, suggesting that valid wood smoke emissions will be estimated from the set of ratios. The precision of this comparison, (18% ((2.2µg/m3) for wood smoke sources at Hawthorne during the summer was comparable to the (16% ((1.3µg/m3) for mobile emission sources at Bountiful during the winter. These two values give the best estimate of the precisim of the source apportionment results given here. The sources of non-volatile PM2.5 obtained from the above-described analysis for the Hawthorne summer samplers are shown in Figure 11. The average concentration of non-volatile PM2.5 from the two sources through July 30 was calculated to be 4.7 and 10.2 µg/ m3 from the mobile and wood smoke sources, respectively. The high concentrations of non-volatile PM2.5 from wood smoke illustrate the importance of distant emissions. The high concentrations of mobile source attributed PM2.5 on the evening of July 24 are due to emissions from firework displays during the celebration of the State Pioneer Day holiday. The average concentration of non-volatile PM2.5 from the two sources after

292

Energy & Fuels, Vol. 16, No. 2, 2002

Long et al.

smoke PM2.5 to Kcor is expected to be about 200 g PM2.5/g Kcor . For the data obtained at the Bountiful and Hawthorne sampling sites during the winter the ratio was (R2 ) 0.48, n ) 14) 240 ( 50 g wood smoke PM2.5/g Kcor, in good agreement with the value expected from the Boise studies.13 The large uncertainty in the ratio determined in this study reflects that fact that the Kcor concentrations were only about 2 to 8 times the limit of quantitation (0.01 µg Kcor/m3). However, the concentrations of Kcor at Hawthorne during the summer varied from 0.1 to 0.5 µg Kcor/m3. For these samples, the ratio of wood smoke to Kcor was (R2 ) 0.73, n ) 8) 40 ( 3 g wood smoke PM2.5/g Kcor , a factor of 6 different from that seen in the winter. The difference is probable due to the difference in emissions from the more controlled combustion of treated wood in homes during the summer and the emissions from uncontrolled combustion of living vegetation in the summer wild fire. Conclusions

Figure 11. One-hour average contribution of wood smoke and mobile emissions to non-volatile PM2.5 and SVM (RAMS minus TEOM monitor results) at the Hawthorne sampling site from 23 July through 5 August 2001. The high indicated concentrations on 24 July are attributed to mobile emissions by the protocols developed in the paper but are probably associated with firework emissions on this Utah state holiday.

July 30 was 10.2 and 16.6 µg/m3 from the mobile and wood smoke sources, respectively. The mobile source emissions will include some wood smoke emissions for the reasons given above. This is probably particularly true for the peaks seen on 2 and 3 August. These are principally due to the nearby forest fires. High concentrations of SVM are either associated with the July 24th fireworks or the August 2-4 impact from nearby forest fires. The concentrations of gas-phase SVOC averaged 115 µg/m3 and were about 100 µg/m3 outside the 2-4 August episode and 140 µg/m3 during this time period. The 80 µg/m3 SVM peak on 2 August will be semi-volatile particulate organic material present in the PM2.5 from the nearby forest fires. This peak and the smaller peaks the day before and after are all slightly later in time than the non-volatile PM2.5 peaks. The appearance of these SVM peaks coincides with the appearance of the ozone peak. The highest ozone concentrations during the entire period occurred along with the August 2 maximum SVM peak. This SVM is secondary organic material formed during the day from gas-phase SVOC associated with these emissions. Comparison of Wood Smoke PM2.5 and Kcor . Concentrations of Kcor could not be used in the 1-h attribution described above because data are only available on a 24-h basis. However, the Kcor concentrations can be used to assess the reasonableness of the attribution. The ratio of Kcor to extractable organic material, EOM, in wood smoke has been previously determined in studies at Boise to be 100 g EOM/g Kcor .13 About half of the wood smoke PM2.5 is EOM, thus the ratio of wood

The combination of continuous gas-phase (NOx, CO, and SO2) and particulate monitor (TEOM, RAMS, and Aethalometer) data has resulted in the 1-h average source apportionment of PM2.5 in Bountiful and Salt Lake City, UT. This is, to the authors’ knowledge, a first attempt to apportion PM2.5 on an hourly basis. Analysis of PM and gas-phase data obtained during winter inversion episodes indicates that both mobile source and wood smoke emissions contribute to PM2.5 present at the STAR Bountiful and EMPACT Hawthorne sampling sites along the Wasatch Front. Contributions of primary emissions from these two sources were similar. A small impact from nearby oil refineries was also seen at Bountiful. In addition to the primary emissions, secondary emissions from the formation of both fine particulate ammonium nitrate and semi-volatile organic material was seen at both sites. Essentially all of this SVM appears to be secondary in nature. Both wood smoke and mobile sources appeared to contribute to the secondary PM2.5. However, formation of secondary semivolatile organic material may have been most associated with the chemistry of emissions from mobile sources during the winter. A summer episode with impact from nearby forest fires also indicated that both mobile sources and wood smoke contributed to primary PM2.5 at the Hawthorne site. An impact of wood smoke fires from distant sources (e.g., the extensive fires in Idaho and Montana) was seen throughout the summer period. However, the dominant formation of secondary semivolatile organic material was clearly associated with high concentrations of wood smoke primary PM2.5 from emissions from nearby forest fires which occurred over about a four-day period during the longer two-week summer episode. These data indicate that the formation of secondary semi-volatile organic material is associated with photochemistry, that the fine particulate SVOC concentrations followed the concentrations of ozone and that essentially all of the semi-volatile fine particulate organic material was secondary. This is the first direct evidence that essentially all fine particulate SVOC is secondary. The 1-h averaged source apportionment technique reported here will be used to investigate possible components or sources of PM2.5 which may be related to cardiovascular morbidity associated with

Fine Particulate Material along the Wasatch Front

exposure to fine particles1,2 by comparison with 1-h heart-rate variability data. Heart-rate variability is an excellent measurement of changes in the autonomic functioning of the heart which has been shown to respond rapidly upon human exposure to fine particulate pollution.34 Acknowledgment. The research reported here was supported by the U.S. Environmental Protection Agency through a cooperative research agreement and a STAR research grant with Brigham Young University. The results of this research have not been subjected to the

Energy & Fuels, Vol. 16, No. 2, 2002 293

Agency’s required peer and policy review and therefore do not necessarily reflect the views of the Agency and no official endorsement should be inferred. The technical assistance of Yanbo Pang, William Modey, Jeff Robison, and Samantha Sizemore and the assistance of Michael Meyer, Jeff Ambs, and Dabrina Dutcher of Rupprecht and Patashnick and of Schleicher and Schuell in the research is gratefully acknowledged. Cooperation by the State of Utah Division of Air Quality in providing the PM2.5 FRM data is also gratefully acknowledged. EF010168L