Novel Method for On-Road Emission Factor Measurements Using a

consists of a plume capture trailer towed behind a test vehicle. The trailer collects a sample of the naturally diluted plume in a 200 L conductive ba...
0 downloads 0 Views 187KB Size
Download PDF
Environ. Sci. Technol. 2007, 41, 574-579

Novel Method for On-Road Emission Factor Measurements Using a Plume Capture Trailer L . M O R A W S K A , * ,† Z . D . R I S T O V S K I , † G.R. JOHNSON,† E.R. JAYARATNE,† AND K. MENGERSEN‡ International Laboratory for Air Quality and Health, Queensland University of Technology, GPO Box 2434, Brisbane, QLD 4001, Australia, and School of Mathematical Sciences, Queensland University of Technology, GPO Box 2434, Brisbane, QLD 4001, Australia

The method outlined provides for emission factor measurements to be made for unmodified vehicles driving under real world conditions at minimal cost. The method consists of a plume capture trailer towed behind a test vehicle. The trailer collects a sample of the naturally diluted plume in a 200 L conductive bag and this is delivered immediately to a mobile laboratory for subsequent analysis of particulate and gaseous emissions. The method offers low test turnaround times with the potential to complete much larger numbers of emission factor measurements than have been possible using dynamometer testing. Samples can be collected at distances up to 3 m from the exhaust pipe allowing investigation of early dilution processes. Particle size distribution measurements, as well as particle number and mass emission factor measurements, based on naturally diluted plumes are presented. A dilution profile relating the plume dilution ratio to distance from the vehicle tail pipe for a diesel passenger vehicle is also presented. Such profiles are an essential input for new mechanistic roadway air quality models.

Introduction There has been an ongoing debate on methods for measuring vehicle emission factors and the limitations presented by current methods. The two main approaches used are dilution tunnel measurement and on-road sampling, with each of these presenting different advantages but also different limitations. In particular, dilution tunnel measurements of particle emissions do not reproduce real world dilution exactly, and can produce artifact nanoparticle modes through desorption and/or pyrolysis of organic material deposited on the walls of the exhaust transfer tubing, causing gross changes in the number of emitted particles (1). Although non-artifact nucleation modes have been clearly observed both near roadways and in dynamometer studies, their study in dynamometer tests has in some cases been complicated by the presence of artifact nucleation modes (2). Maricq et al. (3) compared conventional dilution tunnel, direct tailpipe sampling, and wind tunnel measurements which they viewed as progressively more representative of * Corresponding author phone: +61 7 3864 2616; e-mail: [email protected]. † International Laboratory for Air Quality and Health. ‡ School of Mathematical Sciences. 574

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 2, 2007

real-world emissions. They found that dilution tunnel measurements could exhibit artifact nucleation modes with concentrations 2-4 orders of magnitude larger than the nucleation modes which can arise in the wind tunnel measurements of the exhaust plume behind a vehicle and less frequently in direct tailpipe measurements where the exhaust is diluted to a fixed ratio of the order of 10 or 100 immediately (within ∼0.03 s) at the tailpipe using ejector pump diluters. Lyyranen et al. (4) observed distinct nucleation modes most frequently with dilution techniques which involved rapid cooling and mixing of the exhaust with the dilution air. They concluded that while such artifacts at present only significantly affect particle number emission rates, new measures introduced to curb particulate emissions may lead to a situation where the artifacts also represent a significant fraction of the collected particulate mass. Direct tailpipe dilution also differs significantly from real world dilution in that dilution is essentially instantaneous to a fixed dilution ratio and occurs using clean air or nitrogen whereas real world dilution uses ambient air and is a continuous process. Mariq et al. (3) concluded that to reduce artifacts one should dilute as close as possible to the tailpipe and minimize contact between the hot exhaust and material surfaces. Clearly the ideal approach in terms of most accurately approaching real world dilution is to sample the exhaust plume after it has been emitted, and this should preferably be done during normal on-road driving conditions. A number of on-road sampling techniques have been developed to obtain particulate matter emissions data from moving vehicles. A mobile emission laboratory (MEL) (5-8) has been used to obtain a variety of emissions data including particulate matter mass and chemical composition as well as submicrometer size distribution from vehicle exhaust plumes obtained by scanning mobility particle sizer (SMPS). The SMPS measurements are conducted on plume aerosol drawn into and confined in a 200 L conductive bag. The Aerodyne Research Inc (ARI) mobile laboratory (9) allows CPC and gaseous emission measurements at distances ranging from 3 to 15 m from the test vehicle tailpipe with a delivery residence time of 7 s. A mobile pursuit laboratory (10) performs SMPS size distribution measurements on sampled exhaust plume with a 7-8 s delivery system residence time. Vogt et al. (11) conducted SMPS measurements with a delivery system residence time of 1.2 s from a mobile laboratory while pursuing a diesel passenger vehicle and compared the results to those obtained with the same vehicle using a chassis dynamometer and dilution tunnel. Pursuit methods become considerably more difficult when smaller dilution ratios are required, as, for example, where higher concentrations are required from low emission vehicles. The method developed here avoids the expense associated with wind tunnel measurements and dedicated mobile laboratories while providing a true representation of real world dilution behind a moving vehicle and allowing measurements to be conducted at dilution ratios in the range from 1 to 1000. While it is not without limitations, it presents an attractive alternative, particularly for large-scale emission factor measurements, since it enables cost-effective, accurate, and safe testing of up to twenty vehicles per day.

Experimental Section The sample capture device consists of a conductive bag, which is filled while driving. The bag, located on a lightweight trailer attached to the test vehicle, is connected via a valve 10.1021/es060179z CCC: $37.00

 2007 American Chemical Society Published on Web 12/08/2006

FIGURE 1. Plume capture trailer. to a sample probe, aligned with the exhaust. Exhaust entering the probe can be directed into the bag or to a bypass duct depending on the position of the valve. Sample capture occurs in a matter of seconds and the bag is delivered immediately to a roadside laboratory where it is removed for immediate testing and replaced in preparation for the next sample. Plume Capture Trailer Description. A schematic side view of the plume capture trailer (PCT) is shown in Figure 1. See Figure A, Supporting Information for a photograph of the completed PCT. The position (d, r, φ) of the 150 mm diameter probe tube can be adjusted to allow alignment of the probe mouth with the vehicle exhaust pipe at distances (d) ranging from 0.5 to 5 meters. The probe is coupled via a section of flexible ducting to a central duct consisting of removable 0.5 m long sections which couple the probe to the diverter valve. The diverter valve directs the sample to one of two outlets, which open into a rigid bag enclosure, which is vented at the rear. The 200 L conductive Velostat (3M) sample collection bag is attached to a removable coupling which connects to the upper valve outlet. The forward part of the enclosure is divided horizontally by a partition, which supports the sample collection bag. The partition shields the collection bag from the bypass outlet flow. The partition does not extend for the full length of the chamber so that air can freely flow between the upper and lower sections at the rear of the chamber. A vent at the rear of the chamber allows the free passage of air to the outside. This design ensures that the alternate sample paths are essentially equivalent, thereby minimizing disturbance of the sample probe and central duct flows when the valve is activated. Trapped volumes of air which can dilute or contaminate the sample were minimized in the valve design, resulting in a volume not exceeding 6.3 L or 3% of the sample volume, based on the valve and coupling dimensions. The diverter valve flap is moved by a pneumatic actuator in response to a radio signal triggered by the capture system software operating on a laptop computer located in the passenger compartment. See Supporting Information for a more detailed description of the control system. The valve is returned to the bypass position after a preset delay or alternatively once the bag is filled sufficiently to interrupt a light beam at a preset height on the side walls of the enclosure. In practice, typical sample capture periods range from 4 to 14 s depending on the distance of the plume intake from the vehicle and the vehicle speed. The plume is assumed to undergo dilution in two distinct zones. After an initial brief period of rapid dispersal (zone A) when the plume dispersion is governed by turbulence associated with the exhaust gas exit velocity (12) and

entrainment in the vehicle wake, the plume dispersion becomes dominated by traffic or wind induced turbulence and plume development. Dilution Ratio Measurement. Dilution ratio (D) was determined using the NOx concentrations as in eq 1.

D)

CNOX,exhaust

(1)

CNOX,bag

where CNOX,exhaust ) NOx concentration measured in the exhaust pipe; and CNOX,bag ) NOx concentration measured in the bag. Emission Factor Measurement. Emission factor measurements using the PCT were conducted for a four wheel drive, passenger vehicle (manufacturer’s stated specifications: engine size 4.164 L, 6 cylinders, maximum power 96 kW, 5 speed, mass 2205 kg) operating on diesel fuel with a sulfur level of 300 ppm without a catalytic converter. The capture tests were carried out at Willowbank Raceway, Willowbank, Queensland, on a 1.2 km long sealed service road located in a rural area. The tests were conducted when no other vehicles were operating in the area. Each sample collection run commenced with driving the vehicle and trailer from a mobile laboratory stationed at one end of the test track to the start point during which the onboard gas analyzer system sampled ambient air. The vehicle then accelerated to the desired test speed. After verifying that the vehicle speed and exhaust gas concentrations were stable the capture intake valve was opened for a set period corresponding to the minimum time required to fill the capture bag at the test speed after which the valve was closed. The vehicle then continued on to the mobile laboratory where the analysis of the contents of the bag commenced immediately. The test vehicle engine and exhaust system were conditioned prior to each round of measurements as follows. Prior to completion of the first exhaust sample collection, the vehicle was driven repeatedly over the test circuit at a driving speed of 60 km‚h-1, for a total distance of 8 km before returning to the mobile testing laboratory. The vehicle then remained idling throughout the sample analysis process between subsequent vehicle test runs. The mobile laboratory was equipped with a scanning mobility particle sizer (TSI 3934 SMPS) (sample flowrate 0.7 Lpm) for particle size distribution measurements, a separate condensation particle counter (TSI 3022A CPC) (0.3 Lpm), a TSI 8520 DustTrack aerosol monitor (1.7 Lpm) for particle mass measurements, a NOx analyzer (Ecotech ML9841A) (0.64 Lpm), and sensors for CO2, humidity, and temperature (TSI VOL. 41, NO. 2, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

575

QTrak 8552/8554) (0.3 Lpm). The data from each of these instruments was continually logged. All instruments were connected to a common sampling manifold, the intake to which was connected alternately to an ambient air intake tube drawing air from upwind of the laboratory or to the plume sample container. Thus a record of the corresponding ambient data was recorded alternately with the captured sample data so that appropriate background subtractions could be performed. The analysis of the exhaust samples and background air was performed approximately 100 m from the exhaust sample collection point. Throughout the measurements the sample concentrations were between 20 and 600 times larger than the measured background concentration. The interval between intake valve closure and the commencement of the first size distribution measurement ranged from 90 to 180 s. Analysis of the sample including size distribution and gas concentration measurements took a total of 300 s. The submicrometer particle number emission factor (EFN) is defined as the number of particles emitted per kilometer. Assuming that number concentration in the exhaust pipe is conserved, the above number emission measurement can be determined from the particle number concentration in the exhaust pipe, the exhaust flow rate and the vehicle speed as per eq 2.

EFN )

QexhNexh vvehicle

(2)

where Qexh ) exhaust flow rate, Nexh ) particle concentration in exhaust at time of capture, and vvehicle ) speed of the vehicle. If the dilution ratio (D) and captured sample particle number concentration at the time of capture (Nbag) are known, then the number concentration of the exhaust can be calculated as Nexh ) DNbag. Assuming that the exhaust flow rate can be approximated by the air intake flow rate to the engine, an approximation can be made as per eq 3.

EFN ≈

QairDNbag vvehicle

(3)

where Qair ) air flow rate to engine. Particle mass emission factors can be calculated by replacing the number concentrations by the PM1.0 mass concentrations measured by the DustTrak, usually in units of µg m-3. All emission factor evaluations in these experiments were performed using the above approach. The DustTrak used for the mass emission factor measurements was calibrated (13) against a gravimetric equivalent method (tapered element oscillating microbalance). The engine air flow rate was calculated based on the duct dimensions and air velocity measurements performed using hot wire anemometry (TSI model 8330 air velocity meter). The method of approximating the exhaust flow rate to the air intake flow rate has been validated in a previous study carried out on a chassis dynamometer using a Ford Falcon passenger sedan operating on unleaded petrol. The intake flow rate was measured with a TSI model 8330 air velocity meter, as in the present study. The exhaust flow rate was determined by measuring the pressure drop across a restriction orifice within a long tube attached to the tailpipe of the vehicle, appropriate corrections being made for temperature differences. The estimated exhaust flow rates at five speeds between idle and 100 km‚h-1 agreed with the measured intake 576

9

FIGURE 2. Average dilution ratio versus vehicle speed vcar at various capture distances d, based on three tests for each d,vcar combination. Error bars represent the standard errors of the means of the three measurements.

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 2, 2007

FIGURE 3. Dilution ratio versus capture distance for various vehicle speeds. Also presented are dilution ratios obtained by Vogt et al. (11) for a diesel passenger vehicle closely followed by a mobile laboratory. air flow rate within an error margin of less than 5%. Engine flowrates appear in the Supporting Information.

Results and Discussion Dispersion of the Plume. Figure 2 shows the average dilution ratio versus vehicle speed for a range of capture distances. It can be concluded from this figure that, while constant for the largest capture distance of 3 m, the dilution ratio decreases with increasing velocity for smaller capture distances (d e 2 m). Figure 3 shows the dilution ratio for all speeds versus the capture distance. Also shown are the plume dilution ratios obtained by Vogt et al. (11) for a diesel passenger car followed by a mobile testing laboratory. The combined data give an approximate model for near field plume dispersion for such a vehicle represented by the fitted curve D ) 26.8d1.3, where d is the capture distance in meters and D is the dilution ratio. Such dilution profile measurements are required as input for models such as that being developed by Zhang and Wexler (14). The dilution profile is presented in Table 1 along with two others referred to by Zhang and Wexler (14). These profiles determine the rate of cooling of the exhaust plumes and therefore probably affect the occurrence of nucleation (14). It is interesting to observe that a similar slope was observed in our measurements and in those from the literature (14). Particle Concentration in the Bag at Time of Capture. Particle Losses in the PCT. The size distribution of a confined aerosol can change primarily due to coagulation and wall deposition (15). Particle losses due to coagulation can be controlled via the capture distance and therefore the particle concentration. Wall losses are expected to be dominated by diffusion of the particles to the container walls, and this effect is expected to be most prominent for the smallest particles as these have the largest diffusion coefficient. In order to

TABLE 1. Reported Dilution Profiles dilution profile D(x)

vehicle type

measurement technique

26.8 x1.28

diesel passenger vehicles

17.6 x1.3

330 hp heavy duty truck

plume capture trailer (this study) and data from a pursuing mobile laboratory (11); no other traffic present. wind tunnel

7.01 x0.955

heavy duty truck

on road chase experiments

determine how quickly nucleation mode particles are lost from the captured aerosol, the particle loss rate in the nucleation mode size range was examined by injecting 20 nm mono-disperse size classified aerosols into the bag and measuring the decay in particle concentration with time. See Supporting Information for details of these measurements. The results show that approximately 10% of 20 nm particles would typically be lost in the 300 s interval between capture and completion of sample analysis. Therefore, while nucleation mode particles will be lost during storage, the mode should be readily detectable and these particles will be well represented in particle concentration measurements. This approach to particle loss characterization could be extended to construct a size-dependent wall loss model for the capture system which might then be applied to correct the measured particle size distribution to derive the size distribution at the time of capture, however such an approach is questionable. The degree of inflation of the capture system can vary from sample to sample or during analysis, and such variations will alter the wall loss characteristics of the system. Furthermore, the plume temperature and the ambient temperature may influence thermophoretic particle loss behavior and the contribution of coagulation to particle loss will depend on the concentration and particle size distribution of a given sample. To avoid the necessity of generalizing from laboratory based measurements conducted under controlled conditions to measurements conducted under real world conditions, a more directly empirical approach was adopted for the onroad measurements. This approach consisted of measuring, on a sample by sample basis, the time-dependent decay of the overall particle number or mass concentration and extrapolating this back to the time of capture. The approach is well suited to practical, rapid onroad measurements of particle number emission factors and particle mass emission factors as the instruments’ measurement interval can be very short (1 s) in comparison to size distribution measurements which require 1-2 min per scan. Thus a statistically useful number of data points can be collected to accurately establish the decay trend. To obtain the initial concentration values for the captured samples, linear regression was used to extrapolate back to the time of capture from the measured concentration values for each captured sample. This procedure involved first plotting the CPC concentration time series data recorded from the bag and fitting a straight line. The equation of this line was then used to extrapolate to the time of capture. A typical example of this extrapolation process is given in the Supporting Information. Using this approach, the uncertainty in the line equation value for the time of capture based on a 95% confidence interval was less than 1% in all cases. Assuming that this linear approximation remains valid for the extrapolated regime, the main source of error in the captured particle concentrations at capture time will be due to the instrumental errors associated with the bag concentration measurement. Particle Number Emission Factor Measurement. Figure 4 shows the particle number emission factor versus vehicle

source this study and data from Vogt et al. (11) Doghee Kim, pers. commun. cited in Zhang & Wexler (14) Kittelson et al. (16)

FIGURE 4. Average submicrometer particle number emission factor vs vcar for various capture distances d based on three tests for each d,vcar combination. Error bars represent the standard errors of the means of the three measurements.

FIGURE 5. Average submicrometer particle number emission factor versus capture distance d for each vehicle speed vcar based on three tests for each d,vcar combination. Error bars represent the standard errors of the means of the three measurements. speed for this vehicle for various values of the capture distance d, while Figure 5 shows the emission factor versus the capture distance for each vehicle speed. Analysis of variance using a 95% confidence interval shows that there is a significant dependence of emission factor (p ) 0.015) on the vehicle’s speed but not on capture distance (p ) 0.69). On te basis of sets of three repeated measurements for each vehicle speed and capture distance combination, the standard error of the mean emission factors increased as the capture delay increased, ranging from 8% at 1 m and 40 km/h to 32% at 3 m and 100 km/h. Particle Size Distribution. Figure 6 shows the particle size distribution recorded at different distances from the source for each vehicle speed. Each curve represents the average of three repeats of the plume capture run for the given capture distance and vehicle speed. The results show that the particle size distribution was unimodal in the range from 9 to 400 nm with no evidence of a nucleation mode at any of the examined driving speeds and capture distance combinations. Figure 7 shows the count median diameter of the size distribution mode versus the captured sample overall number concentration for all of the samples. The count median VOL. 41, NO. 2, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

577

FIGURE 8. Particle mass emission factors at four vehicle speeds. dependence of mass emission factor (p ) 0.14) on the vehicle speed. The estimated mass emission factors across the four speeds used ranged from about 37 to 75 mg km-1.

Acknowledgments This project is supported by the Australian Research Council grant DP0211775. Assistance with data collection provided by the following people is gratefully acknowledged: Nick Holmes, Barbara D’Anna, and Arinto Wardoyo. We also thank Jeff Rice and the Willowbank Raceway (www. willowbank-raceway.com.au) for their valued assistance, and for the use of their facilities.

Supporting Information Available

FIGURE 6. Size distribution, corrected for dilution, for aerosol captured at distances of 1, 2, and 3 m, for vehicle speeds of (a) 40 km/h, (b) 60 km/h, (c) 80 km/h, and (d) 100 km/h.

Design criteria and constraints; photograph of the PCT; description and schematic of capture system and on-board exhaust analyzer control system; examination of particle deposition in the PCT; engine air flowrate dependence on driving speed, gear and engine speed; example of extrapolation of particle concentration in the PCT bag to time of capture. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited

FIGURE 7. Count median diameter. diameter shows a distinct increase for concentrations exceeding 106 cm-3. Such concentrations occurred only where the minimum capture distance of 1 m was used with the highest vehicle speed (100 km‚h-1). This effect appears to be due to coagulation which is expected to be more prominent at high concentrations. Particle Mass Emission Factor Measurements. Figure 8 shows the particle mass emission factors estimated for four vehicle speeds: 40, 60, 80, and 100 km h-1. Each value plotted represents the mean of three runs at that particular speed. In these experiments, the capture distance was fixed at 2.0 m behind the tailpipe. The error bars show the standard deviations about the mean. Analysis of variance using a 95% confidence interval showed that there was no significant 578

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 2, 2007

(1) Maricq, M.; Chase, R. E.; Podsiadlik, D. H.; Vogt, R. SAE Paper No. 1999-01.-1461; Society of Automotive Engineers: Warrendale, PA, 1999. (2) Vogt, R.; Kirchner, U.; Scheer, V.; Hinz, K. P.; Trimborn, A.; Spengler, B. Identification of diesel exhaust particles at an Autobahn, urban and rural location using single-particle mass spectrometry. J. Aerosol Sci. 2003, 34, 319-337. (3) Maricq, M. M.; Chase, R. E.; Xu, N. A comparison of tailpipe, dilution tunnel, and wind tunnel data in measuring motor vehicle PM. J. Air Waste Manage. 2001, 51, 1529-1537. (4) Lyyranen, J.; Jokiniemi, J.; Kauppinen, E. I.; Backman, U.; Vesala, H. Comparison of different dilution methods for measuring diesel particle emissions. Aerosol Sci. Technol. 2004, 38, 12-23. (5) Bukowiecki, N.; Kittelson, D. B.; Watts, W. F.; Burtscher, H.; Weingartner, E.; Baltensperger, U. Real-time characterization of ultrafine and accumulation mode particles in ambient combustion aerosols. J. Aerosol Sci. 2002, 33, 1139-1154. (6) Kittelson, D., Watts, W., Johnson, J. Diesel Aerosol Sampling Methodology; CRC E-43 Final Report; University of Minnesota Department of Mechanical Engineering: Minneapolis, MN, 2002. (7) Cocker, D. R.; Shah, S. D.; Johnson, K.; Miller, J. W.; Norbeck, J. M. Development and application of a mobile laboratory for measuring emissions from diesel engines. 1. Regulated gaseous emissions. Environ. Sci. Technol. 2004, 38, 2182-2189. (8) Kittelson, D. B.; Watts, W.; Johnson, J. In 6th ETH Conference on Nanoparticle-Measurement; Burtscher, A., Ed.: Zu ¨ rich, 2002. (9) Canagaratna, M. R.; Jayne, J. T.; Ghertner, D. A.; Herndon, S.; Shi, Q.; Jimenez, J. L.; Silva, P. J.; Williams, P.; Lanni, T.; Drewnick,

(10)

(11) (12)

(13)

F.; Demerjian, K. L.; Kolb, C. E.; Worsnop, D. R. Chase studies of particulate emissions from in-use New York City vehicles. Aerosol Sci. Technol. 2004, 38, 555-573. Pirjola, L.; Parviainen, H.; Hussein, T.; Valli, A.; Hameri, K.; Aaalto, P.; Virtanen, A.; Keskinen, J.; Pakkanen, T. A.; Makela, T.; Hillamo, R. E. “Sniffer”-a novel tool for chasing vehicles and measuring traffic pollutants. Atmos. Environ. 2004, 38, 3625-3635. Vogt, R.; Scheer, V.; Casati, R.; Benter, T. On-road measurement of particle emission in the exhaust plume of a diesel passenger car. Environ. Sci. Technol. 2003, 37, 4070-4076. Chan, T. L.; Dong, G.; Cheung, C. S.; Leung, C. W.; Wong, C. P.; Hung, W. T. Monte Carlo simulation of nitrogen oxides dispersion from a vehicular exhaust plume and its sensitivity studies. Atmos. Environ. 2001, 35, 6117-6127. Jamriska, M.; Morawska, L.; Thomas, S.; He, C. Diesel Bus Emissions Measured in a Tunnel Study. Environ. Sci. Technol. 2004, 38, 6701-6709.

(14) Zhang, K. M.; Wexler, A. S. Evolution of particle number distribution near roadways-Part I: analysis of aerosol dynamics and its implications for engine emission measurement. Atmos. Environ. 2004, 38, 6643-6653. (15) Baron, P. A.; Willeke, K. Aerosol Measurement, Principles Techniques and Applications, 2nd ed.; John Wiley & Sons, Inc: Brisbane, 2001. (16) Kittelson, D. B.; Kadue, P. A.; Scherrer, H. C.; Loverien, R. E. Characterization of Diesel Particles in the Atmosphere; Coordinating Research Council AP-2 Project Group Final Report; 1988.

Received for review January 25, 2006. Revised manuscript received September 6, 2006. Accepted October 20, 2006. ES060179Z

VOL. 41, NO. 2, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

579