Environ. Sci. Technol. 2006, 40, 1911-1915
Monitoring the Soot Emissions of Passing Cars A. KURNIAWAN AND A. SCHMIDT-OTT* Delft University of Technology, Faculty of Applied Sciences, Nanostructured Materials, Julianalaan 136, NL-2628BL Delft, The Netherlands
We report on the first application of a novel fast on-road sensing method for measurement of particulate emissions of individual passing passenger cars. The study was motivated by the shift of interest from gases to particles in connection with strong adverse health effects. The results correspond very much to findings by Beaton et al. (Science, May 19, 1995) for gaseous hydrocarbon and CO emissions: A small percentage of “superpolluters” (here 5%) account for a high percentage (here 43%) of the pollution (here elemental carbon). We estimate that up to 50% of the particulate emissions of vehicles could be avoided on the basis of the present legislation, if on-road monitoring would be applied to enforce maintenance. Our fast sensing method for particles is based on photoelectron emission from the emitted airborne soot particles in combination with a CO2 sensor delivering a reference.
Introduction Numerous epidemiological and toxicological studies have found correlations of health effects with airborne particulate mass (PM), which is primarily soot in the cities (e.g., ref 1). Epidemiological studies published over the past 10 years suggest that the current level of ambient particle concentration is very closely associated with cardio-respiratory problems as well as cancer, and that it can even cause premature death. For example, Dutch epidemiological studies have estimated that PM emissions annually cause the premature death of 1700-3000 people in The Netherlands (population 16 million) (2). Long-term effects of chronic exposure could even result in an estimated premature mortality of 1000015000 per year (2). A joint study of Austrian, French, and Swiss organizations shows that about 50% of the total air pollution in their countries is traffic related (2). This study also concludes that 6% of total mortality is caused by air pollution, which means that 3% is traffic related. Regarding such numbers, particle emissions from cars are currently a hot topic in European media. There is special concern in Europe, because 40% of the fuel consumed is diesel, and diesel vehicles are known to be the major particle source. Adverse health effects of air pollution in cities have already caused concern decades ago. A comprehensive study in California showed more than 10 years ago that few poorly maintained cars dominate pollution with respect to volatile hydrocarbons (HC) and CO. The results of this study were published in Science in 1995 (3). They convincingly show that properly maintained vehicles of any age are relatively small contributors to exhaust pollution and that most of the pollution is caused by a small percentage of badly maintained * Corresponding author phone: + 31 (0) 15 278 3540; fax: +31 (0) 15 278 4945; e-mail:
[email protected]. 10.1021/es051140h CCC: $33.50 Published on Web 02/02/2006
2006 American Chemical Society
cars. The article strongly suggests that a targeted repair program would be highly effective. This would require inuse surveillance involving on-road measurements. However, due to strong epidemiological and clinical evidence, the main attention has moved from gaseous emissions to soot particles during the past decade. While the technology for assessment of gas phase pollutants has existed for a long time, measurement of PM is a much more complex task, especially for time-resolved on-road measurements. Thus a study dedicated to particles in line with the California study has never been presented. A goal of the present study is to find out whether the conclusion that a small percentage of “superpolluting” cars determine most of the particulate pollution is also valid for PM in a European city. At the same time, it introduces a feasible approach of assessing the soot emissions of individual passing cars, which could be an inexpensive key to a drastic improvement of urban air. The study was carried out in the city of Delft, The Netherlands. In this country, periodic exhaust gas tests are prescribed, which also include a crude measurement of soot for diesel vehicles. Compared to that from countries where such tests are not legally required, the contribution of superpolluters in terms of particles are expected to be at the low end. The typical size distribution of particle emissions for diesel vehicles ranges from 30 nm up to 400 nm, whereas particles from gasoline vehicles have sizes from below 10 nm to about 200 nm (4). The larger end of these size distributions is a result of agglomeration of smaller primary particles (10-20 nm). These carbonaceous particles are commonly known as soot. Several studies have linked effects on health to chemical composition, particle size, particle surface, and number concentration. For example, particles with diameters below 100 nm (ultrafine particles) showed stronger toxicity or biological activity than larger particles (19, 20). Other studies report correlations with elemental carbon (21, 22) and polyaromatic hydrocarbons (PAH) (23). The degree of damage to health and the mechanisms are still unclear and under strong discussion. In particular, the small size of the primary insoluble particles coming from automobiles is strongly suspected to have a critical influence. In any case, those facts already secured are convincing enough to make all the possible moves for reduction of PM emissions from automobiles. PM is considered to be the major pollutant in ambient air, and governments are continuously tightening the new car particle emission standards for diesels. While gasoline cars have been excluded from such measures, our study does not distinguish diesel from gasoline cars, because particulate emissions from badly maintained gasoline cars may also contribute.
Methods The airborne elemental carbon (EC) concentration is chosen as a measure of the particulate emission because it is relevant to health and, in contrast to the total mass, it has the advantage of being rather independent of the kind of fuel. Furthermore, it remains unchanged when condensation onto the particles occurs in the cooling and dilution process (5). Typically, EC in diesel exhaust is accompanied by about the same amount of organic matter (see, e.g., ref 25). The measurement is based on the following concept: The soot cloud associated with a passing car is detected with a sufficient time resolution and sensitivity at a stationary point at the side of the street. Thus each vehicle gives rise to a peak in the airborne elemental carbon, which is recorded. The integrated peak depends on the car’s soot emission but also VOL. 40, NO. 6, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1911
TABLE 1. Particulate Emission Standards allowed PM emission
FIGURE 1. Measurement principle. on the extent of dilution by ambient air. This dilution factor varies from case to case, and the CO2 concentration in the exhaust gas was used as a reference. Ambient air is sucked through a duct with the opening facing the street, which splits the flow into the EC detector and the CO2 detector (see Figure 1). Thus the elemental carbon mass concentration m′EC from and the CO2 mass concentration m′CO2 from an identical sample volume are recorded simultaneously. The ratio of the measured quantities is independent of the degree of dilution by ambient air. The emission factor is defined by the ratio of the emitted elemental carbon mass and the carbon mass in the emitted CO2 and given by the measured values through
f)
mEC mEC MCO2 m′EC MCO2 ) ) mC mCO2 MC m′CO2 MC
(1)
Here MCO2 and MC stand for the molar masses of CO2 and C, respectively. We determine f for each vehicle by forming the integral of m′EC(t) and m′CO2(t) over the peak lengths and then forming the ratio. Relation of the Measured Value of f to Existing Standards. The amount of carbon in the fuel converted to soot can be considered negligible with respect to the amount of carbon converted into CO2. Under this condition we get the EC mass emitted per unit volume of burnt fuel as follows:
mEC MC ) Ffuel f Vfuel Mfuel
(2)
where Mfuel stands for the molar mass of the fuel per carbon atom. For the EC emission per kilometer, we obtain
m*EC ) V*fuelFfuel
MC f Mfuel
(3)
* where Vfuel is the fuel consumption per kilometer. A common value for the density of diesel fuel is Ffuel ) 830 g/L. The molar mass of C is MC ) 12 g, and the molar mass of a typical diesel fuel per C atom is Mfuel ) 13.7 g, since it is typically composed of 80 wt % aliphatic (-CH2-) and 20 wt % aromatic * (-C10H8-) constituents. We use Vfuel ) 6 L/100 km as a typical value for diesel passenger vehicles and obtain the estimated EC mass per km as
m*EC ≈ 4.4 × 104 ‚ f [mg/km]
(4)
European regulations (Euro standards) for passenger cars refer to the total diesel particulate mass emitted per kilometer of path driven, m*DPM. We use m*EC/m*DPM ) 0.4 as an approximate value according to several reports (e.g., refs 24 and 25). With this assumption we obtain the estimation
m*DPM ≈ 1.1 × 105 ‚ f [mg/km]
(5)
which allows an approximate conversion of the Euro standards to emission factors according to our definition. These values are summarized in Table 1 together with the 1912
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 6, 2006
standard
year of introduction
m*DPM [g/km]
mDPM/ Vfuel [g/L]
f (estimated) [mg EC/mg C]
APK Euro1 Euro2 (DI)a Euro2 (IDI)a Euro3 Euro4
1992 1997 1997 2001 2006
0.14 0.10 0.08 0.05 0.025
2.3 1.7 1.3 0.8 0.4
3.5 × 10-3 1.3 × 10-3 7.4 × 10-4 9.2 × 10-4 4.6 × 10-4 2.3 × 10-4
a Two Euro2 standards are applicable, one for indirect-direct injection (IDI) diesel engine and one for direct injection (DI) diesel engine.
years when they became effective. Using eqs 4 and 5, the corresponding EC emission values per liter of diesel fuel are approximated and also indicated in the table. The yearly test required in The Netherlands for cars older than 3 years and registered after 1980 (APK test) sets a limit to the emission of diesel cars in terms of the so-called Bosch number, or soot number B. The blackening of a filter paper must not exceed B ) 3 on a scale from 0 to 10 in a standard procedure. B is a measure for elemental carbon emission and has been related to m*EC by a number of reports (7, 12). From these data we conclude that to meet B ) 3, m*EC ) 165 mg EC/m3 represents the upper limit of elemental carbon permitted. We obtain f from this value through eq 1. The value calculated for Table 1 assumes 9.5 vol % of CO2 in diesel exhaust gas (8). Experimental Setup. A suitable quantitative measurement of EC emitted from passing cars in continuous traffic flow requires a sensor with a short response time on the order of 1 s. State-of-the-art sensors or methods suitable for EC detection are the Aethalometer (9), laser-induced incandescence (LII) (29), and photoacoustic spectroscopy (28). The response times obtained with the Aethalometer and photoacoustic measuring instruments in diluted exhaust gases are significantly longer than 1 s. Laboratory setups based on the photoacoustic principle do deliver response times of 1 s or shorter but have only been achieved with expensive high-intensity lasers. LII would also exhibit the required response time, but with today’s laser technology the sophisticated LII systems presently in use do not have the potential of being developed into portable devices suitable for on-road measurements. The Aethalometer was used on passing cars in 1990 (10). A time interval of about 1 min between two successive cars was required, which evidently made it difficult to collect data for a large population, and that study was restricted to 60 cars. In view of the strong reduction of emission limits since 1990 and to be able to measure at the roadside in flowing traffic, we required a sensor that was more sensitive as well as faster than conventional solutions. The measuring device we applied consists of a photoelectric aerosol sensor (PAS) (11, 13, 16) in combination with a commercial Binos 4B.1 (Leybold Heraeus), an NDIR-based analyzer. It detects CO2 concentrations in the range of 0-500 ppmV or 0-1000 ppmV. Our EC sensor (see Figure 2) is based on photoelectron emission from aerosol particles under UV radiation (11, 13). The probed volume flows through a quartz tube 1 cm in diameter and a few centimeters long which is irradiated by a low-pressure UV lamp (185 nm). Former work has shown that under such conditions soot particles from combustion of organic material emit electrons due to the photoelectric effect. This effect is quite selective for soot (13, 27) with respect to other particulate constituents in the atmosphere. The electrons migrate to the wall, while the particles remain with a positive charge. In a steady flow, the electric current associated with this charge is determined in a so-called
FIGURE 2. Principle of the photoelectric detector. aerosol electrometer. More recent work has shown that the signal of such sensors can be calibrated in units of mass of elemental carbon per unit volume for cars (5, 6, 15, 26, 27). PAS instruments have originally been developed and commercialized as PAH sensors, because for a specific combustion source there is also a correlation between the PAH concentration and the PAS response (16). As EC sensors, they have been heavily criticized because, for a number of new cars with exhaust filters, the correlation between particulate mass and the PAS signal turned out to be bad (17, 18). This finding caused confusion and hindered further development in the direction of applying the PAS as an EC detector. We have identified the reason for this problem: The filter strongly reduces the soot emitted. At the same time, volatile constituents such as H2 SO4 are retained, and much of the vapor condenses on the few remaining soot particles. Despite the low concentration of sulfur in today’s fuels, the ratio of sulfur and soot emitted is much higher for the new, filter-equipped cars (5) than for the older models. We conclude that the sulfur compounds deposited on the particle surfaces quench photoemission, and the signal becomes highly sensitive to temperature and vapor conditions in the car’s exhaust filter configuration. The present study is based on the solid evidence that for diesel vehicles without filters there is an excellent association between the signal and the EC concentration and that the instrument underestimates the emissions of vehicles equipped with properly functioning soot filters due to quenching by the condensate. We can neglect this error for the purpose of the present study because practically no passenger cars were equipped with soot filters in The Netherlands when the data were taken. Furthermore, the filter typically reduces the emission to 1/100 (5), so that the total contribution of such cars can certainly be neglected. In view of an expected increase of soot filter equipped vehicles in the future, modification of the present EC sensor to include desorption of photoelectrically inactive surface layers on the particles are in progress. The EC instrument we applied was an LQ1-L detector (Matter Engineering, Switzerland), which has not been produced any more since 1996. The instrument has been calibrated for elemental carbon using a diesel source without after-treatment. It has a response time below 1s. The aerosol flowing through this device is slightly heated below 50 °C to ensure particle surfaces free from condensed water. The equipment was set up at the roadside 60 m from a crossing. The speed of most cars was around 50 km/h, which is also the speed limit. Because a crossing lies ahead, the cars do not accelerate at the measurement point. Particle emissions from diesel engines would strongly increase upon acceleration, and so the detected soot is assumed to underestimate the average soot emission of each car in city traffic. The measurements were made during 4 days after 6 pm, when the traffic was not too busy. The particle emissions of 1255 passenger cars including minivans were detected. There were seldom events of 2 or more vehicles following each other too closely to be resolved individually. Such clusters were treated like single vehicles, which has an averaging effect. An example of the raw data obtained with the EC sensor and the CO2
FIGURE 3. Example of the recorded elemental carbon concentration, mEC ′ , and CO2 concentration, m′CO2, vs time. sensor is shown in Figure 3. Here we see the peaks of 3 cars following each other in time intervals between cars of about 9 s, which is well resolvable. The signal peaks were integrated graphically. There is a rise of the apparent background concentrations due to the mixing of the exhaust plumes. Only the areas of the well distinguishable peaks above the interpolated background concentration line were determined. Since this is done for EC as well as CO2, neglect of the “mixed” area does not introduce any error. The short response time below 1 s of both instruments leads to a sharp separation of the peaks. The measurement time for each car is between 5 and 10 s and determined by the process of turbulent mixing and transport of the exhaust plume past the measuring site. Only cars, including minivans, were measured, because they represent the largest source of particulate matter. Diesel and gasoline cars were not distinguished, but we assume that the highest emitters were the diesels.
Results and Discussion We divided the cars measured into emission classes indicated on the horizontal axis of Figure 4. Note that the scale is logarithmic. The horizontal lines of Figure 4 represent the number of passenger cars in each emission class. The measurements resulted in emission factors from f ) 6.3 × 10-6 to 3.3 × 10-2, which means that they range over more than 3 orders of magnitude. The average value of f was 1200 × 10-6. The bars indicate the corresponding distribution of emission contributions calculated as the sum of all values of f in each emission class. The estimated European standard limits according to Table 1 are indicated as well as the estimated limit value in the Dutch APK test for diesel. It is to be noted that APK uses a free acceleration test. Diesel cars typically show lower soot emissions at constant speed or deceleration, which is the condition under which we measured. Thus the APK indication in the figure must be seen as a rather crude indication, and the actual limit is deemed more to the left. The striking conclusion is that the highest polluting 5% of all cars (dark shaded area in Figure 4) contribute to (43 ( 2)% of the particulate pollution of passenger cars. This group of “superpolluters” would certainly not pass the annual VOL. 40, NO. 6, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1913
FIGURE 4. Horizontal lines stand for the number of cars vs the emission factor class. The numbers indicate the number of cars in the respective emission class. The bars give the emission contribution (sum of the emission factors in each class) vs the emission factor class. The dark area on the right side marks the highest emitting 5%. The standards tabulated in Table 1 are indicated in the figure. The Euro2 standard in the figure corresponds to Euro2-DI. required test (APK). The statistical error ((2%) of our 1255 samples was determined by a bootstrap procedure. Cars registered before 1980 are exempt from the APK particle emission test. We are certain that none of such cars passed our measurement site, because they are distinguished by black license plates. We emphasize that our measurement underestimates diesel emissions with respect to the standards for the following reasons: (a) The standards Euro1-4 refer to driving cycles that include acceleration, where diesel particle emissions are typically much higher than at constant speed. The APK soot test refers to free acceleration. In contrast, we measured under a constant speed or even decelerating condition. (b) The photoelectric detector underestimates the emissions from filter-equipped cars. We conclude that about half of the present particulate pollution by passenger cars would be avoidable if adequate monitoring would enable enforcement of existing laws. The situation is deemed to be similar with trucks, busses, and 2-wheel vehicles. The effect of tightened standards (Euro14) to be met by new vehicles is covered by the dominance of superpolluters. We expect that the situation is comparable in other countries and that The Netherlands are probably even on the low emission side, due to the compulsory annual particle emission test. We propose to use our measurement method for assessment of particulate emissions by cars. It could be developed into an automatic system to identify superpolluters and to enforce proper maintenance of those cases. Since measurement and analysis of the exhaust peak corresponding to a certain car takes between 5 and 10 s, the changed position of the cars after that time could complicate unambiguous identification. To avoid any errors in allocation, an automated system would scan all license plates at the measuring position and label those of the superpolluters after the corresponding peak has been recorded a few seconds later. Such a system 1914
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 6, 2006
would not be more complex than a speed check and would lead to very substantial improvement of air qualityson the basis of today’s legislation.
Acknowledgments We are grateful to Matter Engineering for providing the PAS sensor, to Heinz Burtscher, Saul Lemkowitz, and Konstantin Siegmann for helpful discussions, and to Peter Verheijen for help in determining the statistical error.
Literature Cited (1) Lippmann, M.; Frampton, M.; Schwartz, J.; Dockery, D.; Schlesinger, R.; Koutrakis, P.; Froines, J.; Nel, A.; Finkelstein, J.; Godleski, J.; Kaufman, J.; Koenig, J.; Larson, T.; Luchtel, D.; Liu, L. J. S.; Oberdo¨rster, G.; Peters, A.; Sarnat, J.; Sioutas, C.; Suh, H.; Sullivan, J.; Utell, M.; Wichmann, E.; Zelikoff, J. The US Environmental Protection Agency particulate matter health effects research centers program: A midcourse report of status, progress, and plans. Environ. Health Perspect. 2003, 111 (8), 1074-1092. (2) Buringh, E.; Opperhuizen, A. On health risks of ambient PM in The Netherlands. RIVM publication, October 2002, report no. 650010032. (3) Beaton, S. P.; Bishop, G. A.; Zhang, Y.; Ashbaugh, L. L.; Lawson, D. R.; Stedman, D. H. On-road vehicle emissions: regulations, costs and benefits. Science 1995, 268, 991. (4) Kauffeldt, Th.; Schmidt-Ott, A.; Hild, O.; Ba¨cker, H. On line measurement of the soot particle size distribution during the combustion process inside a diesel engine. J. Aerosol Sci. 1997, 28, Suppl. 1, S45-S46. (5) Burtscher, H. Physical characterization of particulate emissions from diesel engines: a review. J. Aerosol Sci. 2005, 36, 896-932. (6) BRISKA “Baseler Risikostudie Aussenluft”, Der Photoelektrische Aerosolsensor zur Bestimmung polyzyklischer aromatischer Kohlenwasserstoffe (PAH) im Vergleich mit anderen Messmethoden, Lufthygieneamt beider Basel, CH-4410 Liestal, April 1999. (7) Christian, R.; Knopf, F.; Jaschek, A.; Schindler, W. Eine neue Messmethodik der Bosch-Zahl mit erhoehter Empfindlichkeit. Motortech. Z. 1993, 54, 16-22.
(8) UNEP/WHO, (United Nations Environmental Programme/ World Health Organization). Urban air pollution in mega cities of the world, Blackwell, London, 1992. (9) Hansen, A. D. A.; Rosen, H.; Novakov, T. Real time measurement of the absorption coefficient of aerosol particles. Appl. Opt. 1982, 21, 3060. (10) Hansen A. D. A.; Rosen H. Individual measurements of emission factor of aerosol black carbon in automobile plumes. JAWMA 1990, 40, 1654-1657. (11) Schmidt-Ott, A.; Federer, B. Photoemission from small particles suspended in a gas. Surf. Sci. 1980, 106, 538. (12) Alkidas, A. C. Relationship between smoke measurements and particulate measurements. SAE Technical Paper Series 1984, no. 840412. (13) Burtscher, H.; Scherrer, L.; Schmidt-Ott, A.; Siegmann, H. C. Probing aerosols by photoelectric charging. J. Appl. Phys. 1982, 53, 3787. (14) Burtscher, H.; Siegmann, H. C. Monitoring PAH-emissions from combustion processes by photoelectric charging, Combust. Sci. Technol. 1994, 101, 327-332. (15) Przybilla, K. R. Investigation of particles from combustion with special consideration of elemental carbon. Dissertation submitted to the Swiss Federal Institute of Technology Zurich for degree of Doctor of Natural Sciences. 2002. Diss. ETH no. 14624. (16) McDow, S. R.; Giger, W.; Burtscher, H.; Schmidt-Ott, A.; Siegmann, H. C. Polycyclic aromatic hydrocarbon and combustion aerosol photoemission. Atmos. Environ. 1990, 24A, 2911. (17) Wessels, S.; Eiterer, H.; Schneider, M.; Sebesse, W. ADA Report 10/2002, from Abgaszentrum Der Automobilindustrie GbR, Porschestraβe, 71287 Weissach, Germany. (18) OICA, Industry Comments on Proposed Particulate Measurement Techniques, OICA Contribution to PMP Part 1: Synthesis Report. 16 Dec., 2002; http://www.oica.net/htdocs/WWH/ Particulates/ OICA%20contribution%20to%20PMP%20-%20Synthesis.pdf. (19) Li, N.; Kim, S.; Wang, M.; Froines, J.; Sioutas, C.; Nel, A. Use of a stratified oxidative stress model to study the biological effects
(20) (21) (22)
(23)
(24)
(25) (26) (27) (28)
(29)
of ambient concentrated and diesel exhaust particulate matter. Inhalation Toxicol. 2002, 14 (5), 459-486. Oberdo¨rster, G. Toxicology of ultrafine particles: in vivo studies. Philos. Trans. R. Soc. London, Ser. A 2000, 358 (1775), 27192739. Mar, T. F.; Norris, G. A.; Koenig, J. Q.; Larson, T. V. Associations between air pollution and mortality in Phoenix, 1995-1997. Environ. Health Perspect. 2000, 108 (4), 347-353. Metzger, K. B.; Tolbert, P. E.; Klein, M.; Peel, J. L.; Flanders, W. D.; Todd, K.; Mulholland, J. A.; Ryan, P. B.; Frumkin, H. Ambient air pollution and cardiovascular emergency department visits. Epidemiology 2004, 15 (1), 46-56. Dejmek, J.; Solansky, I.; Benes, I.; Lenicek, J.; Sram, R. J. The impact of polycyclic aromatic hydrocarbons and fine particles on pregnancy outcome. Environ. Health Perspect. 2000, 108 (12), 1159-1164. Fruin, S. A.; Winer, A. M.; Rodes, C. E. Black carbon concentrations in California vehicles and estimation of in-vehicle diesel exhaust particulate matter exposures. Atmos. Environ. 2004, 38, 4123-4133. Sharma, M.; Agarwal, A. K.; Bharathi, K. V. L. Characterization of exhaust particulates from diesel engine. Atmos. Environ. 2005, 39, 3023-3028. Dahmann, D.; Mosimann, T.; Matter, U. Validation of on-line sensors for monitoring occupational exposures from diesel engines, J. Aerosol Sci. 2000, 31, Suppl. 1, S21-S22 Burtscher, H.; Siegmann, H. C. Combust. Sci. Technol. 1994, 101, 327. Schindler, W.; Haisch C.; Beck H. A.; Niessner, R.; Jakob, E.; Rothe, D. Photoacoustic Sensor System for Time Resolved Quantitative Measurement of Diesle Soot Emission, SAE 200401-0968, 2004. Schraml, S.; Will, S.; Leipertz, A.; Zens, T.; d’Alonso, N. Performance Characteristics of TIRE-LII Soot Diagnostics in Exhaust Gases of Diesel Engines, SAE 200-01-2002.
Received for review June 16, 2005. Revised manuscript received December 23, 2005. Accepted January 6, 2006. ES051140H
VOL. 40, NO. 6, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1915