Characterization of Fine Particle and Gaseous Emissions during School Bus Idling

Transcription

1 Environ. Sci. Technol. 2007, 41, Characterization of Fine Particle and Gaseous Emissions during School Bus Idling J. S. KINSEY* United States Environmental Protection Agency, Office of Research and Development, National Risk Management Research Laboratory, MD E343-02, Research Triangle Park, North Carolina D. C. WILLIAMS, Y. DONG, AND R. LOGAN ARCADIS U.S., 4915 Prospectus Drive, Suite F, Durham, North Carolina The particulate matter (PM) and gaseous emissions from six diesel school buses were determined over a simulated waiting period typical of schools in the northeastern U.S. Testing was conducted for both continuous idle and hot restart conditions using a suite of on-line particle and gas analyzers installed in the U.S. Environmental Protection Agency s Diesel Emissions Aerosol Laboratory. The specific pollutants measured encompassed total PM-2.5 mass (PM e2.5 µm in aerodynamic diameter), PM-2.5 number concentration, particle size distribution, particle-surface polycyclic aromatic hydrocarbons (PAHs), and a tracer gas (1,1,1,2,3,3,3-heptafluoropropane) in the diluted sample stream. Carbon monoxide (CO), carbon dioxide, nitrogen oxides (NO x ), total hydrocarbons (THC), oxygen, formaldehyde, and the tracer gas were also measured in the raw exhaust. Results of the study showed little difference in the measured emissions between a 10 min post-restart idle and a 10 min continuous idle with the exception of THC and formaldehyde. However, an emissions pulse was observed during engine restart. A predictive equation was developed from the experimental data, which allows a comparison between continuous idle and hot restart for NO x, CO, PM- 2.5, and PAHs and which considers factors such as the restart emissions pulse and periods when the engine is not running. This equation indicates that restart is the preferred operating scenario as long as there is no extended idling after the engine is restarted. Introduction The U.S. Environmental Protection Agency (EPA) has determined that diesel exhaust is a likely human carcinogen that can also contribute to other acute and chronic health effects (1). In addition, children are generally more susceptible to air pollutants such as diesel particulate matter (PM) because their respiratory systems are still developing and they have a faster breathing rate (1). For these reasons, concern has been raised about the exposure of children to PM exhaust pollutants associated with diesel school buses during the commute to and from school. Of particular * Corresponding author phone: (919) ; fax: (919) ; kinsey.john@epa.gov. importance is the exposure of children to idling buses during loading and unloading operations. In these circumstances, the engine tends to run at less than optimum efficiency with limited dispersion of the exhaust pollutants (2). A number of studies has been conducted to assess children s exposure to diesel pollutants during school bus commutes, some of which address potential exposures during loading/unloading (3-11). In 2002, both Wargo and Brown (10) and Sabin et al. (4) found significantly higher concentrations of black carbon (BC) inside idling buses as compared to those measured in buses while in motion. Gilliam and Reeves (7) also determined similar increases in particle concentration for idling buses after the door was initially opened and also found significant spikes in the PM-2.5 (particles e2.5 µm in aerodynamic diameter) concentration when the buses were first started in the morning. In addition, Behrentz et al. (3) determined the concentrations of BC, particle surface polycyclic aromatic hydrocarbons (PAHs), and NO 2 to be times higher (depending on pollutant) at bus stops as compared to a school loading/unloading zone where idling was limited. Also, in addition to tailpipe pollutants, Hill et al. (6) found the crankcase vent tube to be a major source of PM-2.5 observed at bus stops. On the basis of these and similar data, many regulatory agencies and school districts, including the EPA s Clean School Bus USA initiative, have issued guidance or regulations limiting the idling of school buses during the loading/unloading of school children (12). A question frequently posed to the EPA and anti-idling advocates is whether restarting school buses will result in higher emissions of diesel pollutants than those attributable to periods of continuous idle. This paper addresses this question by measuring the idle emissions from a limited number of diesel school buses under wintertime conditions. The objective of the study was to test the hypothesis that the benefit of anti-idling, including restart, results in less net emissions than continuous idling. Experimental Procedures Testing was performed during early March 2005 at the bus yard of the Katonah-Lewisboro School District located in Cross River, NY. The District provided the test site, buses, and fuel used in the study. The District also allowed each test bus to be taken out of service so that it could be evaluated in a more cost-effective manner. A total of six District buses with model years ranging from 1997 (odometer ) km) to 2004 (odometer ) 1191 km) were evaluated. The buses were equipped with one of three different models of Caterpillar diesel engine along with a Donaldson Diesel Oxidation Catalyst (DOC) muffler and a Spiracle Crankcase Ventilation Filtration System. Standard pump grade diesel fuel with a sulfur content of 226 ppm (weight) and a cetane index of 44.6 was used during testing. A further description of the test vehicles and fuel used can be found in the Supporting Information. For this study, the Diesel Emissions Aerosol Laboratory (DEAL) was used as the basic sampling platform (13). A special test fixture (Figure 1) was needed, however, to connect the tailpipe of the bus to the DEAL sampling system while providing adequate dilution of the exhaust sample with clean, pollutant-free air. As shown in Figure 1, a tapered fitting was inserted into the bus exhaust pipe and connected to a short-radius 90 elbow and straight section of heated stainless steel pipe containingvariousprobesandsensors. Tracergas(1,1,1,2,3,3,3- heptafluoropropane or FM-200) was injected and mixed with ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 14, /es CCC: $ American Chemical Society Published on Web 06/08/2007

2 FIGURE 1. Exhaust sampling system. the exhaust flow in the elbow after which the mixture entered the straight working section of the apparatus. In the working section, the volumetric flow rate was determined using a calibrated Annubar and differential pressure cell followed by PM and gas sample extraction with the gas sample directed to the continuous emission monitor (CEM) bench of the DEAL through a heated line. The exhaust temperature was also measured at two locations with the remaining raw exhaust vented to the atmosphere. In the diluter portion of the sampling system (Figure 1), the raw sample flow was mixed with ambient air (test average temperature of 4-14 C) filtered through a combination high efficiency particulate air (HEPA) and activated carbon filter. A final Teflon filter was also used to remove any carbon particles that might contaminate the diluent air stream. The ratio of dilution air to sample air was generally <10:1 to simulate real world conditions near the bus. After introduction of the diluent stream, mixing of the sample with the dilution air occurred by passing the combined flow through a short-radius 90 elbow before entering the PM-2.5 preseparator in the DEAL. After entering the DEAL sampling tunnel, particles greater than PM-2.5 were removed in a high volume virtual impactor. Samples were then provided to the various analyzers using a series of staggered probes and flow splitters. The preseparator major and minor flows were controlled by variable frequency drives and calibrated mass flow meters. A series of particle analyzers was connected to the flow splitters in the DEAL sampling tunnel (see Supporting Information). These analyzers measured the PM mass concentration (tapered element oscillating microbalance [TEOM] and quartz crystal microbalance [QCM]), PM number concentration (condensation particle counter [CPC]), particle size distribution (electrical low pressure impactor [ELPI] and scanning mobility particle sizer [SMPS]), black and blue carbon (Aethalometer), and PAHs (EcoChem 2000) in the diluted sample stream. In addition, the particle number (CPC) and mass (TEOM) were also determined downstream of a Dekati Model EKA-111 thermal denuder operated at 250 C to determine the percent volatile fraction of the PM. Both TEOMs were operated at 30 C to reduce volatilization losses (13). All appropriate calibrations were performed prior to the start of the study. A description of the various PM analyzers and their operation is provided elsewhere (13, 14). The fixed combustion gases in the raw exhaust were also monitored for CO, CO 2,NO x, total hydrocarbons (THC), and O 2 as described by Kinsey et al. (13). The sample flow was provided to the CEM bench through the long heated sample line described previously. All gas analyzers were calibrated prior to testing and checked on a daily basis using certified standards. To determine the dilution ratio, the concentrations of tracer gas in the raw and diluted exhaust were monitored using two INNOVA Model 1314 photoacoustic analyzers. In the case of the raw exhaust, an aliquot of sample gas obtained from the flow stream to the CEM bench was provided for analysis. For measurement of the diluted exhaust, the Model 1314 was connected to Splitter 1 of the DEAL sampling tunnel (see Supporting Information). In addition, the raw gas analyzer was also equipped with an optical filter to measure formaldehyde, which was used as a general indicator of gasphase air toxics. Both analyzers were calibrated with certified gas standards prior to testing and were checked daily. Finally, the data for all instruments were recorded and stored using the automated data acquisition system (DAS) described by Kinsey et al. (13). Also, except for bus 203, a separate laptop computer was used to record engine operating data (e.g., fuel flow) using the Caterpillar software. Bus VOL. 41, NO. 14, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

3 FIGURE 2. Total gas-phase emissions over an equivalent 10 min engine idling period for (a) nitrogen oxides (NO x), carbon monoxide (CO), and total hydrocarbons (THC) and (b) formaldehyde. Note that the restart pulse emissions shown in panel (a) are only a small fraction of those measured at idle. 203 was a 1997 model equipped with a mechanical fuel injection system and thus did not have an onboard computer to monitor engine operation during testing. To meet study objectives, both continuous idle and hot restart emissions were measured using a simulated 10 min waiting period specified by EPA Region 2. For the continuous idle tests, the bus was started and allowed to warm up for a period of 5 min per District policy, after which the vehicle was driven for about 17 min to represent travel from the bus yard to a local school. The bus was then parked, the tunnel was attached to the tailpipe, and the emissions were continuously monitored for up to 20 min using the equipment described previously. The bus was then shut off and allowed to cool down, and the procedure was repeated. During the hot restart tests, the same basic protocol was used except that the engine was shut off for a period of 10 min after the bus was parked, the bus was restarted, and the emissions were measured during the post-restart period. Triplicate runs for each operating mode were conducted for all buses except bus 259, where only one continuous idle test was conducted. Data Analyses. Fuel-specific emission factors (mass/mass of fuel burned) were calculated from the sampling data. These emission factors were calculated from a carbon balance using the fuel analysis, CO 2 concentration measured in the raw exhaust, and dilution ratio determined from the FM-200 concentration measured in the raw and diluted exhaust streams. Applicable emission rates (mass/unit time) of the various pollutants were also calculated using the diesel fuel feed rate recorded by the engine computer. In addition, the total mass of each pollutant was calculated for an equivalent 10 min operating period for both continuous idle as well as the emissions occurring after engine restart. Detailed calculation procedures are shown in the Supporting Information. To produce the particle size distributions (PSDs) for the different buses using the ELPI data, the data inversion method ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 14, 2007

4 FIGURE 3. Total PM-2.5 emissions and percent volatile content determined by the TEOMs over an equivalent 10 min engine operating period for continuous idle and post-restart idle. Note the lower emissions for the restart pulse. FIGURE 4. Total particle-bound PAH emissions over an equivalent 10 min engine idling period. Note much lower restart pulse emissions indicated by dashed lines. discussed by Dong et al. (15) was used. A composite PSD was calculated for all three tests of continuous idle and hot restart for each bus. Finally, it should be noted that the Aethalometer (black and blue carbon) and the SMPS (particle size distribution) generated no useful data. The PM-2.5 mass emission rates and particle size information provided next were derived from the TEOM and ELPI data, respectively. Results Gas-Phase Emissions. In general, the gas-phase emissions generated during continuous idle were found to be either relatively constant or to slowly rise over time depending on the particular species being measured. Upon restarting the engine, however, a short (i.e., s) emissions pulse was observed followed by a quick stabilization period and then constant emissions. Appropriate emission factors, emission rates, and the total emissions determined over an equivalent 10 min engine operating period were calculated for NO x, CO, THC, and formaldehyde for both continuous idle and post-restart idle using the procedures described previously. The emission factors and rates for all four gaseous pollutants are provided in the Supporting Information with the 10 min total emissions shown graphically in Figure 2 as compared to the restart pulse and the results of Toback et al. (16). As shown in Figure 2a, there is little difference in the emissions between post-restart and continuous idle for CO and NO x with the exception of buses 260 and 288, respectively. For bus 260, the relative percent difference (RPD) in the CO VOL. 41, NO. 14, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

5 FIGURE 5. Average mass-based particle size distributions determined from ELPI data for (a) continuous idle and (b) post-restart idle. emissions was 15% for the two idle conditions, and for bus 288, the RPD of the NO x emissions was 19%. Note, however, that bus 288 has a newer engine design with very low mileage and probably had not been adequately broken in prior to testing. With regard to the emissions of THC and formaldehyde, except for bus 203, these emissions were generally higher for post-restart as compared to continuous idle, which is most apparent for buses 260 and 288, which had the highest ( km) and lowest (1191 km) odometer reading, respectively. For buses 260 and 288, the THC emissions were 3-5 times higher, and the formaldehyde emissions were 2-4 times higher for restart. A noticeable odor was also noted during the testing of bus 288, which was not present for the other vehicles tested. Since these tests were conducted in the winter, the catalyst bed may have dropped below its activation temperature as a result of the 10 min cool off period coupled with the low exhaust gas temperature typical of idle. When a catalyst is not active, minimal reduction in organic gases are achieved (17). Comparing the previous data with the restart pulse showed much lower overall emissions for the pulse. Also, reasonably good agreement was found between the current study and that of Toback et al. (16) for some pollutant/bus combinations (e.g., NO x for buses 255 and 260, CO for buses 255 and 256, and THC for buses 256 and 259), whereas substantial differences were found in other cases. There are several reasons for different results being obtained in the two studies, including a lower ambient temperature in the current work as well as the fact that different engines were evaluated ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 14, 2007

6 TABLE 1. Average Idle Emission Rates and Restart Emissions a pollutant continuous idle emission rate (ER i(p)) (mg/min) total restart emissions pulse (E r(p)) (mg) post-restart idle emission rate (ER ri(p)) (mg/min) nitrogen oxides (NO x) carbon monoxide (CO) total PM PM surface-bound PAH a Arithmetic average for all buses tested and rounded to two significant figures. Does not include bus 203, which was not equipped with an engine computer or emissions data for formaldehyde and total hydrocarbons which were inadequate to calculate the restart pulse. TABLE 2. Emissions for Hypothetical 10 Min Bus Waiting Period a operating scenario total NO x emissions (mg) total CO emissions (mg) total PM-2.5 emissions (mg) PM surface PAH emissions (mg) continuous idle restart + 0 min idle restart + 2 min idle restart + 4 min idle restart + 6 min idle restart + 8 min idle continuous idle ) restart pulse (s) a Rounded to two significant figures. Particle-Phase Emissions. As was the case for the gaseous emissions, the PM emissions occurring during the continuous idle period were also relatively constant over time. However, far more variation in the emissions was observed than was the case for the various gases discussed previously. Restart of the engine again resulted in a s pulse of emissions followed by a leveling off period. This pulse was generally more variable and longer in duration than was observed for the gaseous species. Total PM-2.5 emission factors, emission rates, and total emissions over an equivalent 10 min engine idling period were also calculated in the same manner as the gaseous emissions. The emission factors and rates are also provided in the Supporting Information with the 10 min total PM-2.5 mass emissions shown graphically in Figure 3. Also shown in Figure 3 is the volatile fraction of the PM as determined by the TEOMs with and without the thermal denuder, the restart emissions pulse, and the experimental results obtained by Toback et al. (16). As shown in Figure 3, the total PM-2.5 emissions and volatile fractions are both about the same or lower for postrestart as compared to continuous idle with the exception of bus 288. In the case of bus 288, both the total PM-2.5 emissions and volatile fraction are slightly (i.e., 40% RPD for total PM-2.5) higher for restart. Again, these observations could be attributable to the lack of an adequate break-in period as compared to the older buses as discussed previously. The volatile content of the PM emissions for all buses tested was generally consistent with other published idle data generated using a similar measurement system (18). Also, according to Kittleson et al. (19), the volatile components removed in the thermal denuder were generally comprised of water, sulfates, and organic carbon. It should also be noted in Figure 3 that generally good agreement was found between the PM emissions data obtained in the current study and that collected by Toback et al. (16) with the exception of bus 260. For bus 260, the observed PM emissions were a factor of 2 higher. Again, variations in ambient temperature and engine design could explain the difference in the PM emissions between the two studies. With respect to PM surface-bound PAHs, the total emissions generated over the 10 min engine operating period are provided in Figure 4 as compared to the restart pulse. As shown in Figure 4, the PAH emissions are generally lower for post-restart as compared to continuous idle. Another interesting observation from Figure 4 is that the PAHs generated by the older, mechanically injected bus (203) were substantially (i.e., a factor of 5-10) lower than the newer buses equipped with electronic fuel injection. A similar trend was also observed for NO x in Figure 2. The lower PAH and NO x emissions could be attributed to differences in injection timing and a lower combustion temperature for the older engine. Prior studies have found that a higher combustion temperature tends to favor both the production of particlephase PAHs (20, 21) and NO x (22). Further investigation would be necessary, however, to determine whether these lower emissions are typical of all mechanically injected engines or just this example. PSDs. The differential mass PSDs determined from the ELPI data are shown in Figure 5. For 10 min of continuous idle (Figure 5a), the PM emissions from all buses exhibited a single accumulation mode with bus 260 producing smaller particles. The geometric mass mean particle diameter (GMD) for all buses except bus 260 averaged 360 nm, whereas the PSD for bus 260 showed a GMD of 200 nm. Also, although bus 260 had the highest total PM-2.5 mass emissions (Figure 3), it produced lower number concentrations than bus 203. The higher mass emissions for bus 260 are a result of the greater fuel consumption and exhaust flow for this particular engine as compared to the others tested. For the post-restart condition (Figure 5b), all buses again produced a single accumulation mode with two of the buses having somewhat smaller particles as compared to the others. For buses 203 and 255, the GMD averaged 390 nm, whereas for buses 260 and 288, the GMD was 280 nm. It should also be noted that the sizes of the particles produced after restart were qualitatively the same as those occurring during continuous idle except for bus 260. For bus 260, the PSD produced after restart was approximately the same as for the other buses, whereas somewhat smaller particles were generated during continuous idle as mentioned previously. VOL. 41, NO. 14, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

7 There is no apparent reason for this observation except inherent differences in the engines tested. Again, bus 260 exhibited slightly lower number concentrations as compared to bus 203 but higher PM-2.5 mass emissions for the reasons mentioned previously. Discussion The analysis described previously was made using simulated loitering periods of 10 min. As such, no consideration was given to the time during which the bus was turned off and no pollutants were generated. In addition, a number of sitespecific conditions such as meteorology (e.g., need for air conditioning or heat), driver preference, local school district policy, and whether the buses are parked overnight at the school instead of at a remote bus yard will also influence how the engine is actually operated and thus a true comparison between continuous idle and post-restart. Therefore, an additional examination of the data was performed to further assess differences in the emissions between operating modes that more closely reflect the real world. As the first step, vehicle-averaged emission rates were calculated from the experimental results for both idling conditions. The average emissions generated during the short restart pulse, separate from the post-restart idling, were also determined for those pollutants where sufficient data were available. The results of these calculations are shown in Table 1. It must be noted, however, that the emissions data provided in Table 1 are very limited and only applicable to the specific engines, emission controls, diesel fuel, ambient conditions, and operating procedures evaluated in the study. Therefore, caution is advised when applying these data to other engine types and emission controls. Using the data shown in Table 1, the difference in total emissions for a particular operating scenario can be calculated as follows: E p ) (ER i(p) t i ) - [E r(p) + (ER ri(p) t ri )] (1) where E p is the difference in emissions of pollutant p between continuous idle and restart (mg); ER i(p) is the emission rate of pollutant p for continuous idle (mg/min); t i is the continuous idle time (min); E r(p) is the total emission of pollutant p for the engine restart pulse (mg); ER ri(p) is the emission rate of pollutant p for idling after engine restart (mg/min); and t ri is the idle time after restart (min). Eq 1 allows the input of different waiting times and engine operating modes to determine the overall difference between continuous idling and restart for the pollutants listed in Table 1. To illustrate the significance of these findings, Table 2 shows the results of using eq 1 to model the emissions for a hypothetical 10 min waiting period. In the first scenario, the bus idles continuously for the entire 10 min. In the other scenarios, the engine is turned off, restarted, and allowed to idle for periods ranging from 0 to 8 min. Also shown in Table 2 is the amount of continuous idle time that would be equivalent to the restart emissions pulse (i.e., ER i(p)t i ) E r(p) in eq 1). As can be seen from Table 2, restart and immediate departure have the lowest emissions for all pollutants. Also, the time period when ER i(p)t i ) E r(p) varied from 2 s for NO x to 200 s for PM-2.5. Thus, even for PM-2.5, it takes very little continuous idle time to produce the same emissions as the short pulse associated with restart. Table 2 indicates, therefore, that restarting the engine is generally preferred as long as there is no extended idling after the restart. Acknowledgments The U.S. Environmental Protection Agency through its Office of Research and Development funded and managed the research described here under Contract EP-C to ARCADIS U.S. It has been subjected to Agency review and has been approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. Supporting Information Available Test vehicle description, DEAL equipment layout, emission calculation procedures, fuel composition, and idle emission factors/rates. This material is available free of charge via the Internet at Literature Cited (1) Health Assessment Document for Diesel Engine Exhaust; PB , U.S. Environmental Protection Agency, National Technical Information Service: Springfield, VA, (2) Ning, Z.; Cheung, C. S.; Lu, Y.; Liu, M. A.; Hung, W. T. Experimental and numerical study of the dispersion of motor vehicle pollutants under idle conditions. Atmos. Environ. 2005, 39, (3) Behrentz, E.; Sabin, L. D.; Winer, A. M.; Fitz, D. R.; Pankratz, D. V.; Colome, S. D.; Fruin, S. A. Relative importance of school bus-related microenvironments to children s pollutant exposure. J. Air Waste Manage. Assoc. 2005, 55, (4) Sabin, L. D.; Kozawa, K.; Behrentz, E.; Winer, A. M.; Fitz, D. R.; Pankratz, D. V.; Colome, S. D.; Fruin, S. A. Analysis of real-time variables affecting children s exposure to diesel-related pollutants during school bus commutes in Los Angeles. Atmos. Environ. 2005, 39, (5) Marshall, J. D.; Behrentz, E. Vehicle self-pollution intake fraction: Children s exposure to school bus emissions. Environ. Sci. Technol. 2005, 39, (6) Hill, B. L.; Zimmerman, N. J.; Gooch, J. A Multi-City Investigation of the Effectiveness of Retrofit Emissions Controls in Reducing Exposures to Particulate Matter in School Buses; Clean Air Task Force: Boston, MA, January 2005; reports/catf-purdue_multi_city_bus_study.pdf. (7) Gilliam, A. R.; Reeves, U.-R. A Safer Ride to School: How to Clean Up School Buses and Protect our Children s Health; Southern Alliance for Clean Energy: Atlanta, GA, January (8) Behrentz, E.; Fitz, D. R.; Pankratz, D. V.; Sabin, L. D.; Colome, S. D.; Fruin, S. A.; Winer, A. M. Measuring self-pollution in school buses using a tracer technique. Atmos. Environ. 2004, 38, (9) Ireson, R. G.; Easter, M. D.; Lakin, M. L.; Ondov, J. M.; Clark, N. N.; Wright, D. B. Estimation of diesel particulate matter concentrations in a school bus using a fuel-based tracersa sensitive and specific method for quantifying vehicle contributions. Transport. Res. Record 2004, 1880, (10) Wargo, J.; Brown, D. Children s Exposure to Diesel Exhaust on School Buses; Environmental and Human Health, Inc.: New Haven, CT, February (11) Solomon, G. M.; Campbell, T. R.; Feuer, G. R.; Masters, J.; Samkian, A.; Paul, K. A. No Breathing in the Aisles; Natural Resources Defense Council, Coalition for Clean Air: New York, January (12) What You Should Know About Diesel Exhaust and School Bus Idling; EPA420-F , U.S. Environmental Protection Agency, Office of Transportation and Air Quality: Washington, DC, June (13) Kinsey, J. S.; Mitchell, W. A.; Squier, W. C.; Wong, A.; Williams, C. D.; Logan, R.; Kariher, P. H. Development of a new mobile laboratory for characterization of the fine particulate emissions from heavy-duty diesel trucks. J. Automobile Eng. 2006, 220, D3, (14) Kinsey, J. S.; Mitchell, W. A.; Squier, W. C.; Linna, K.; King, F. G.; Logan, R.; Dong, Y.; Thompson, G. J.; Clark, N. N. Evaluation of methods for the determination of diesel-generated fine particulate matter: Physical characterization results. J. Aerosol Sci. 2006, 37, (15) Dong, Y.; Hays, M. D.; Smith, N. D.; Kinsey, J. S. Inverting cascade impactor data for size-resolved characterization of fine particulate source emissions. J. Aerosol Sci. 2004, 35, (16) Toback, A. T.; Hearne, J. S.; Kuritz, B.; Marchese, A. J.; Hesketh, R. P. The Effect of Ambient Temperature and Humidity on Measured Idling Emissions from Diesel School Buses; Paper , Society of Automotive Engineers: Warrendale, PA, ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 14, 2007

8 (17) Turrio-Baldassarri, L.; Battistelli, C. L.; Conti, L.; Crebelli, R.; DeBerardis, B.; Iamiceli, A. L.; Gambino, M.; Iannaccone, S. Evaluation of emission toxicity of urban bus engines: Compressed natural gas and comparison with liquid fuels. Sci. Total Environ. 2006, 355, (18) Kwon, S.-B.; Lee, K. W.; Saito, K.; Shinozaki, O.; Seto, T. Sizedependent volatility of diesel nanoparticles: Chassis dynamometer experiments. Environ. Sci. Technol. 2003, 37, (19) Kittleson, D. B.; Watts, W. F.; Johnson, J. P. On-road and laboratory evaluation of combustion aerosolsspart 1: Summary of diesel engine results. J. Aerosol Sci. 2006, 37, (20) Arnott, W. P.; Zielinska, B.; Rogers, C. F.; Sagebiel, J.; Park, K.; Chow, J.; Moosmuller, H.; Watson, J. G. Evaluation of 1047-nm photoacoustic instruments and photoelectric aerosol sensors in source sampling of black carbon aerosol and particle-bound PAHs from gasoline and diesel powered vehicles. Environ. Sci. Technol. 2005, 39, (21) Lim, M. C. H.; Ayoko, G. A.; Morawska, L.; Ristoviski, Z. D.; Jayaratne, E. R. Effect of fuel composition and engine operating conditions on polycyclic aromatic hydrocarbon emissions from a fleet of heavy-duty diesel buses. Atmos. Environ. 2005, 39, (22) Gajendran, P; Clark, N. N. Effect of truck operating weight on heavy-duty diesel emissions. Environ. Sci. Technol. 2003, 37, Received for review October 18, Revised manuscript received March 29, Accepted April 30, ES VOL. 41, NO. 14, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY