Environ. Sci. Technol. 2004, 38, 5704-5711
New Method for Time-Resolved Diesel Engine Exhaust Particle Mass Measurement U . L E H M A N N , * , † V . N I E M E L A¨ , ‡ A N D M. MOHR† Laboratory for Internal Combustion Engines, EMPA (Swiss Federal Laboratories for Materials Testing and Research), CH-8600 Du ¨ bendorf, Switzerland, and Dekati Ltd., Osuusmyllynkatu 13, FIN-33700 Tampere, Finland
The Dekati mass monitor (DMM; Dekati Ltd., Finland), a relatively new real-time mass measurement instrument, was investigated in this project. In contrast to the existing gravimetric filter method also used as a standard for regulation purposes, this instrument provides second-by-second data on mass concentration in the engine exhaust gas. The principle of the DMM is based on particle charging, inertial and electrical size classification, and electrical detection of aerosol particles. This study focuses on the instrument’s practical performance. Details on calibration and the theory of operation will be published elsewhere. The exhaust emissions of two heavy-duty engines complying with the Euro III emission standard were measured on a dynamic engine test bench. We looked at the particle number and mass emissions of the engines in different transient test cycles and steady-state conditions. The ability to follow transient test cycles and the response times of the DMM were investigated. The aerosol mass concentration measured by the DMM was compared with the mass concentration obtained by the standard gravimetric filter method with Teflon-coated glass fiber filters. The total mass concentration (integral over the whole cycle) measured by the DMM is about 20% higher than that measured by the standard gravimetric filter method. The total mass concentration from the DMM was also compared with the volume concentration calculated from the electrical lowpressure impactor (ELPI) measurements. Correlations were made with other particle measuring systems. The DMM correlates very well with the particulate mass (R2 ) 0.95) and exhibits good linearity and repeatability. The response time to a well-defined change in exhaust concentration was observed to be fast and stable. The DMM was able to follow transient test cycles and provides good results on a second-by-second basis. The instrument used in this study was still under development, and there is therefore no complete scientific background reference for the DMM. This study therefore focuses more on the measurements than on the scientific background. The measurements have shown that the DMM is an adequate instrument for measuring the mass concentration of engine exhaust, with results comparable to those from * Corresponding author phone: +41-1-823 42 19; fax: +41-1-823 40 41; e-mail:
[email protected]. † EMPA (Swiss Federal Laboratories for Materials Testing and Research). ‡ Dekati Ltd. 5704
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004
the standard gravimetric filter method. In addition, the DMM provides real-time second-by-second data of the mass concentration during transient test cycles.
Introduction Numerous studies have been conducted to characterize airborne particles in the atmosphere and to identify their sources (1). A growing number of studies are documenting the risk of emissions from diesel engines to human health and the environment (e.g., ref 2). Diesel exhaust particles appear to have a significant effect on human health (3). Epidemiological studies have shown similar associations between adverse health effects and particulate pollution, wherever it is measured by mass or number. There is limited knowledge as to whether particle mass, surface area, or number is more responsible for effects on respiratory symptoms (4). Diesel particle emissions primarily produced by automotive traffic and especially by diesel engines are one source under investigation. A study on heavy-duty engines (5) shows that the correlation between particulate mass (PM) and total particle number is influenced by the engine and seems to be unaffected by the different transient test cycles. The relationship between particle mass and number of ambient air particles in the submicrometer size range was investigated by Morawska et al. (6). They found that due to the uncertainty in the density of submicrometer particles, it is not possible to estimate the submicrometer mass concentration with better than 60% uncertainty. Up to now, regulations (7) for automotive particle emissions have been based solely on PM with a gravimetric filter, resulting only in a total, cumulative mass emission over the entire test cycle. Current air quality standards are expressed in terms of particle mass concentration as PM10 or PM2.5, which are mass concentrations of particles with diameter less than 10 or 2.5 µm, respectively (8). Other time-resolved mass measurement methods with particle collection on a filter followed by analysis by the thermal/optical reflectance carbon method (TOR) are reported by Moosmu ¨ ller et al. (9), whose investigations focused on elemental carbon (EC) and organic carbon (OC) measurements. Such filter methods cannot be used to study emissions during transient operating conditions. Particulate light absorption with a photoacoustic instrument and determination of light extinction with a smoke meter were investigated. They conclude that for a better understanding of the actual composition of automotive emissions, instruments for measuring particulate mass emissions with a time resolution on the order of 1 s are needed. On the other hand, the gravimetric filter measurements are approaching their detection limits for modern low-emission diesel engines. New measuring techniques as well as new dilution systems have been investigated (9). A better time resolution of particle mass emission is needed to identify high concentrations in transient engine behavior. The new Dekati mass monitor (DMM; Dekati Ltd., Finland) is a real-time particle measurement instrument that provides second-by-second information on particle mass concentration. In this study, a prototype of the DMM was used to measure the aerosol mass concentrations of two heavy-duty diesel engines in five different transient test cycles. The integrated aerosol mass concentration is compared with the results of the regulated gravimetric filter method. The aerosol mass concentrations are also compared with particle number 10.1021/es035206p CCC: $27.50
2004 American Chemical Society Published on Web 09/29/2004
current carried on these particles is measured. For the case of unimodal and symmetrical particle size distribution, the ratio between this current and the total current measured from the impactor defines the median mobility diameter (13, 14). Equation 2 defines the relationship between current ratio and mobility diameter. This equation is based on numerical simulation of known distributions and known mobility analyzer collection efficiency and is verified by actual distributions. The entire calibration method and data will be published elsewhere. The average effective density of particles is calculated from the difference between aerodynamic and mobility median diameters:
dp ) 59
FIGURE 1. Principle of the Dekati mass monitor (DMM).
TABLE 1. Impactor Calibration Values for Dekati Mass Monitor (DMM) stage
1
2
3
4
5
6
D50% (µm) 0.0296 0.0512 0.0860 0.114 0.237 0.533
7 (precut) 1.290
concentrations, measured simultaneously by a condensation particle counter (CPC) with and without thermodesorber (573 K) and by an electrical low-pressure impactor (ELPI). We also compare these measurements with different measurement techniques.
Experimental Methods Dekati Mass Monitor. The principle of the DMM is based on the ELPI (10, 11), also manufactured by Dekati Ltd. (Finland). This instrument is based on particle charging, inertial and electrical size classification, and electrical detection of aerosol particles. The sample passes through a unipolar positive-polarity triode-type diffusion charger where the particles are electrically charged by ions produced by corona discharge. An electrical field is then used to collect the smallest particle sizes (particles smaller than 30 nm) on an electrode measuring the current carried on these particles (Figure 1). Charger calibration defines the size-dependent average charge per particle and charger losses. The calibration method is similar to that for the ELPI and is explained in ref 12, resulting in the charging efficiency given in
PneQ )
{
15.885dp2.294 dp e 0.068 µm 0.847dp1.204 dp > 0.068 µm
(1)
where P is the charger penetration, n is the average charge per particle, e is the electronic charge unit (1.602 × 10-19 C), Q is the flow rate (10.3 lpm), and dp is the particle mobility diameter. After the aerosol passes the diffusion charger, a six-stage cascade low-pressure impactor classifies the particles according to their aerodynamic diameter (Figure 1), and the current carried on these particles is measured. The impactor calibration method is also described by Marjama¨ki et al. (12) and the cut point diameters (D50%) for the instrument are given in Table 1. ELPI calibration is based on the concentration measurement (12). Therefore, a new method for particle density measurement is applied to DMM, allowing more reliable conversion from measured electrical current to particle mass concentration. This method is based on a comparison of aerodynamic and electrical mobility diameters, described in ref 13. A simplified mobility diameter measurement is based on a zeroth-degree particle mobility analyzer, which collects the smallest particles by use of a static electrical field. The
(
0.938
Imob - 0.124 Itot
-1
)
(1/2.13)
(2)
where dp is the median particle mobility diameter (in micrometers), Imob is the mobility channel current, and Itot is the total current measured from the mobility electrode and impactor (Imob + Iimpactor). The measurement method of the DMM was developed at Tampere University of Technology, aerosol physics laboratory and will be published elsewhere. The mobility diameter measurement algorithm is based on the fact that the size distribution of the diesel aerosol soot fraction can be approximated as log-normal and unimodal (15). If volatile material is not removed from the sample, nucleation phenomena can cause a nonsymmetrical or bimodal size distribution (16). The size distribution modality is examined continuously in the DMM, and if a nonsymmetrical distribution is detected, the mass concentration is calculated from a constant, pre-defined density value (unit density value). Thermal treatment of the sample (heated dilution, thermodesorber, or catalytic stripper) is therefore recommended to prevent nucleation when a DMM is used. The DMM calculates the mass concentration of particles smaller than 1.3 µm on a real-time second-by-second basis. The nominal volumetric flow rate is 10 L/min and the concentration range is from 3 up to about 5000 µg/m3. The instrument used in this study is still in the development phase and the calibration values are not identical with commercially available units. Initial measurements on different diesel engines by this new instrument were presented at the IAC 2002 conference (14). Gravimetric Determination of Particulate Matter. PM of the diesel engine exhaust gas was obtained by collection on Teflon-coated glass-fiber filters (Pall, T60A20) and gravimetric determination in accordance with the regulation (7). An aerosol flow of 70 L/min was separated from the fullflow CVS tunnel and passed through a secondary dilution unit with fixed dilution ratio. The diluted sample was drawn through a filter, which was kept at a maximum temperature of 325 K. The total dilution ratio for the PM measurement was in the range between 6 and 7 depending on the exhaust flow rate. Engines and Sampling Method. Two heavy-duty diesel engines complying with Euro III emission technology level were employed in this study; both were turbocharged intercooled engines used in commercial road transport in Europe. The parameters of these engines are shown in Table 2. All tests were performed with standard diesel fuel [density (at 288 K) 381.5 kg m-3, sulfur content 346 mg kg-1]. Measurements with engine 2 were carried out on a CRT system in order to reach a level corresponding to about 60% of the future Euro IV emission level regarding PM emissions in the European transient test cycle (ETC). For our measurements, the engines were run on a transient heavy-duty test bench (see Figure 2). This facility consists of VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5705
FIGURE 2. Measurement setup of transient heavy-duty test bench.
TABLE 2. Euro III Heavy-Duty Diesel Engine brand no. of cylinders displacement technology rated engine power max torque year of registration mileage driven exhaust system
engine 1
engine 2
MB V6 12 L turbocharger/ intercooler in-line pump with ECU 260 kW at 1800 rpm
Volvo in-line, 6 7L turbocharger/ intercooler
1730 Nm 2001 100 000 km muffler
190 kW at 2200 rpm 1030 Nm nn nn CRT system/ bypass
an asynchronous motor, a full-flow dilution tunnel for the regulatory measurement of particulate matter, and standard exhaust emission analyzers. The exhaust of the engine was piped by a 9-m long heated tube (diameter 125 mm, temperature not less than 453 K) to the 4.6-m long full-flow dilution tunnel (diameter 450 mm). The exhaust gas from engine 1 was measured by taking the sample for the DMM and the particle number measurements from the dilution tunnel, near the location of the sample probe for the gravimetric measurement (Figure 2). The sample was immediately diluted in an additional ejectortype dilution unit (17). The transfer line (15 cm), the diluter, and the dilution air were heated to about 343 K. The entire setup is made of stainless steel. Particle mass was measured by a prototype of the new DMM instrument. An electrical low-pressure impactor (ELPI, Dekati Ltd., Finland) (10) was configured in parallel to measure the particle number size distribution. The ELPI used for these measurements was not equipped with a filter stage and covered a size range from 0.032 to 10 µm (50% cut point diameters) for a flow rate of 9.8 L/min. The particle size measured by the ELPI is an aerodynamic equivalent diameter, the diameter of a sphere with uniform density and with the same settling velocity as the particle considered. To compare the ELPI with the DMM, the ELPI results were converted to a volume-based mass concentration by assuming a spherical shape and unit density. The total number of emitted particles was measured by two condensation particle counters (CPC model 3022A; TSI Inc.) (18). One CPC measured the total nonvolatile particle concentration downstream of a thermodesorber (19), heated 5706
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004
to 573 K. In addition, the total concentration of the emitted particles was measured by another CPC. For the steady-state tests, the same CPC was also used in combination with a differential mobility analyzer (DMA; TSI Inc.) as a scanning mobility particle sizer (SMPS; TSI Inc.). The particle size measured by the SMPS is an electrical mobility-equivalent diameter, the diameter of a sphere with the same mobility as the particle in question. Further details of the method are described elsewhere (20, 21). Additional measurements were performed with engine 2. For this configuration, the sample for the DMM and the CPC was taken directly from the raw exhaust gas line (before entering into the CVS dilution tunnel), whereas the probe for the gravimetric filter measurement was taken out of the CVS tunnel in accordance with regulations. The exhaust gas sample for the DMM and CPC was diluted by a double-stage ejector with a heated first stage (473 K).
Test Cycles Heavy-duty engine test cycles are defined according to relative engine torque and speed. The test cycles (with the exception of the FTP cold test cycle) were started after a defined warmup procedure. This procedure consists of a 30-min warmup of the engine, followed by a 20-min constant run at B100 (1537 Nm, 248 kW, highest temperature in the dilution tunnel). The test was started immediately after this procedure. Different transient cycles were run within this program: the European transient cycle (ETC), the U.S. Federal Test Procedure (FTP), and the transient test cycle developed by the Technical University of Graz (TUG). The FTP was started once with and once without warmup of the engine prior to the test (referred as FTP cold and FTP hot). In addition, the ETC cycle was split into the subcycles urban, rural, and highway. All of the test cycles used are described elsewhere in ref 5. The representative worldwide harmonized heavy-duty transient test cycle (WHTC) relates the engine speed to the same characteristic engine speed values that were used in the drive train model. The substitution model was tested against the drive train model and found to be equivalent (22). The parameters of the investigated transient cycles and subcycles are listed in Table 3. In addition to the transient test cycles, two steady-state modes were tested at one moderate engine speed (1530 rpm) but with two different engine loads: B100 (1537 Nm, 248 kW; full load) and B25 (384 Nm, 62 kW; 25% load). These two modes were tested one after the other, beginning with 100%
TABLE 3. Cycle Parameters cycle time (s) rel. engine speed (%) (mean values) rel. engine torqued (%) (mean values)
ETCa
ETC urbana
ETC rurala
ETC highwaya
TUGb
FTPc
WHTC
1800 50.9 36.7
600 40.3 35.8
600 54.2 47.7
600 58.1 26.6
2187 55.0 31.2
1199 41.5 24.1
1800 41.6 24.9
a Scaling in accordance with Directive 2001/27/EC. b Scaling based on Directive 2001/27/EC. c Scaling in accordance with the Code of Federal Regulations, Office of Federal Register National Archives and Records Administration, parts 86-99 (1994). d Negative torque values are set to zero for calculation of the mean relative torque value.
FIGURE 3. Coefficient of variance (COV) of different instruments for three test cycles (left panel) and COV for two instruments on different test cycles (right panel) and at lower emission level (see text). The number of test repetitions is given in parentheses. load, for about 20 min each. After 5 min of stabilization, the particle measurements by DMM and CPC were started.
Results and Discussion The main focus of these measurements was to obtain information on the repeatability, transient behavior, and response time of the DMM during transient cycles. We looked at correlation between the mass concentration measured by the standard gravimetric filter method and the mass concentration determined by the DMM. Simultaneously, the total number of emitted particles was measured by the CPC and the ELPI and compared to the DMM measurements. All measurements were converted to standard conditions (101.3 kPa and 273.15 K). Repeatability of Measurements. The coefficient of variance (COV) of the instruments on different transient test cycles (ETC, FTP hot, and FTP cold) is shown in Figure 3 (left). The measurements were carried out on different days and the number of repetitions is indicated in parentheses. The variation in engine cycle work is less than 0.2%. The standard gravimetric filter method and the DMM measured the mass concentration of the emitted particles. A condensation particle counter with thermodesorber at an operating temperature of 573 K (CPC with TD) and a CPC without thermodesorber (CPC without TD) measured the total particle number concentration. The particle number concentration measured by the ELPI was converted into a volume concentration by assuming a spherical shape and unit density (1 g cm-3). The COVs of the transient cycles are low for most instruments. The relatively high COV for the ELPI volume (Figure 3, left) is due to the volume calculation from the number concentration. The most repeatable result was achieved by measuring with the CPC with the thermodesorber (CPC with TD) removing volatile materials. To obtain more information on the repeatability of the DMM, the ETC cycle and four steady-state conditions were performed at much lower emission levels (about 7 times lower; emission level is about 2 mg Nm-3 over total ETC cycle). These measurements were performed with engine 2. The ETC test cycle was run 7 times on three different days. The standard deviation of the DMM (Figure 3, right) changes little for the different steady-state conditions (SM). The highest emission level is at 25% load (B25; about 4.6 mg
Nm-3) and the lowest at 75% load (C75; about 1.7 mg Nm-3). For the transient cycles, the standard deviation of the DMM is comparable to the standard deviation of the gravimetric analysis. Ability To Follow the Transient Cycle/Time-Resolved Real-Time Measurement. An interesting feature of the new DMM mass measurement system is the ability to follow transient cycles and to provide real-time second-by-second mass concentration data. An example of time-resolved mass concentration is shown in Figure 4 for the ETC cycle. For comparison, the real-time particle number concentration measured by the CPC with and without thermodesorber is shown as well. The particle mass concentration and the particle number concentration are related to fast acceleration/engine power changes. The DMM as well as the CPC showed a good time response and was able to follow the transient cycle. At the beginning of the cycle, the DMM and the CPC without thermodesorber (CPC without TD) are in good agreement (Figure 4). Each change in engine torque and the resulting change in particle emission is measured by the DMM as well as by the CPC without TD. Behavior during the sequence between 410 and 430 s and between 870 and 1000 s (dashed vertical boxes in Figure 4) is different. During these periods, both engine torque and engine speed are stable. The DMM measures a stable concentration, whereas the CPC without TD measures an increasing number of particles. This leads to the conclusion that nucleation phenomena might play a more important role in the measurement by CPC without TD. The use of a thermodesorber (CPC with TD) reduces such effects but also results in a smoothed measurement due to the additional volume of the thermodesorber. The response time of the DMM was examined by measuring a well-defined change in engine power switching between 90% and 10% load at a constant engine speed. The calculated time response data cannot be taken as general characteristic values of the instruments, as they are influenced by the behavior of the engine. An example of a second-bysecond concentration measurement and the definition of the characteristic times are shown in Figure 5 (right). This cycle was repeated 15 times and an average of the characteristic times is given on the left in Figure 5. VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5707
FIGURE 4. Time-resolved on-line particulate emissions data over the entire ETC test cycle. Left scale: DMM mass/CPC total number.
FIGURE 5. Characteristic times of the examined DMM compared to those of the CPC.
FIGURE 6. Correlation between mass measured by DMM and by gravimetry for different units. The response of the DMM to a defined concentration change was observed to be fast and stable within a few seconds. Taking into account the time constant of the whole system, the characteristic time t90-10 (time between the detection of 10% and 90% of the measured end concentration) is found to be less than 8 s. The characteristic times are about 30% of the corresponding times of the CPC. The DMM is able to follow a transient test cycle and to detect individual peaks in particle concentration due to brief load peaks during the transient cycles. Correlation between Particle Mass Measured by DMM and Gravimetry. The correlation between the absolute values of the total mass measured by gravimetric measurement and DMM is shown in Figure 6, calculated per unit sample volume (left) and per unit engine cycle work (right). If we look only at the transient test cycles, a good correlation (Figure 6) can be seen by comparing the total particle mass measured by DMM and gravimetry in mass 5708
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004
per unit volume (correlation coefficient R2 > 0.95) and in mass per unit engine cycle work (R2 > 0.91). Absolute values were in good agreement in most measurements, but a linear trend line has a slope of 1.2 and 1.4, and a negative bias is evident in both comparisons. The reason for these discrepancies is the differing treatment of volatile material in the gravimetric and DMM measurements. If volatile material, such as water or light hydrocarbons, deposits on the surface of existing soot particles, it increases particle size and mass, or it can nucleate and form new particles. The dilution ratio and temperature change in the standard dilution tunnel are favorable for such phenomena (23, 24). According to the regulations, the gravimetric filter is conditioned in a temperature-controlled environment before weighing (7), evaporating a part of the volatile material from the filter medium. Online instruments measure all particles as they are on entering the instrument, thus indicating a
FIGURE 7. Ratio between particle mass concentrations measured by DMM and by gravimetry.
FIGURE 8. Correlation between mass measured by DMM and by gravimetry for different units.
FIGURE 9. Comparison of different instruments relative to gravimetric PM measurement (normalized). higher mass concentration. The ejector diluter at CVS temperature decreases this difference, as some of this material is evaporated, but different load conditions produce different quantities of volatile material, resulting in different discrepancies. This can be seen in Figure 7, where the ratio between DMM and PM is shown for all measured cycles. The factor between these two instruments is between 0.89 and 1.15 for the transient cycles and concentrations per unit volume.
Engine speed and load affect the correlation between DMM and gravimetry, as seen in ETC urban, rural, and highway phases and single-mode measurements (B25 and B100), as well as temperature (FTP hot and cold starts). Engine speed changes the dilution ratio and load affects the temperature change, resulting in a change in partial vapor pressure and dew point of the volatile material. Another set of measurements was carried out during the UN/GRPE PMP program (25), where the correlation between VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5709
FIGURE 10. Correlation between mass measured by DMM and by LII and PASS. DMM and gravimetric mass was examined for engine 2 at two different exhaust emission levels. For engine emission level 1, PM emissions over the total ETC cycle were about 2 ng Ncm-3, while for engine emission level 2, PM emissions were found to be about 0.3 ng Ncm-3. A total of 12 ETC test cycles were measured. The correlation is shown in Figure 8 (left). The correlation is similar to that for engine 1 (Figure 6), and a negative bias in the correlation is more evident. At very low emission levels, the gravimetric measurement yields significantly higher mass values than the DMM. This effect is due to gas-phase hydrocarbon adsorption on the filter medium, as discussed in refs 26-28, for example. In terms of aerosol particle measurement, the result from the DMM is more accurate, but in comparison with the regulatory measurement, there is a significant difference, especially at very low PM emission levels. The ratio between the examined instruments and the gravimetric filter method is shown in Figure 9 for all transient cycles. The ratios are normalized with respect to concentration over the total ETC cycle. For the CPC, the ratio varies much more without a thermodesorber in comparison with the measurement with a thermodesorber. In addition, the CPC with TD measured fewer particles than the CPC without TD. As a result of the normalization, the ratio of CPC with TD to PM is sometimes larger than that of CPC without TD to PM. Comparison between CPC results with and without a thermodesorber indicates that there is a different quantity of volatile material depending on driving cycles (Figure 9). On the highway part of the ETC cycle, the quantity of volatile materials is higher in comparison with the urban part. Increased load results in a larger quantity of condensed and nucleated materials. The measurement with a thermodesorber is much less influenced by nucleation phenomena. Of all of the on-line instruments evaluated in this study, the DMM shows the best correlation with regulated PM emissions. Large variations in volume-weighted ELPI concentration indicate that the density information used for DMM mass calculation improves the ELPI measurement principle. Large variations in number-based measurements is due to the fact that particle mass is a function of diameter and thus changes with particle size. Within this project, the DMM was compared with other mass-based particle measuring instruments. The correlation between DMM and the laser-induced incandescence soot analyzer (LII; Esytec, Germany) (29) is shown on the left side in Figure 10. Good linearity and correlation were found between the DMM and the EC mass-based instrument LII (R2 ) 0.99). The different sampling method, treatment of volatile material, and measuring principle influenced the agreement of absolute values. Figure 10 (right) shows the comparison between DMM and the photoacoustic absorption measurement technique (PASS; Technical University of Munich, Germany) measuring EC mass concentration (30). A good correlation can be seen (R2 ) 0.98), but the different 5710
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004
sampling location and the different measured quantity make a comparison of the absolute values more difficult. The DMM also exhibited good correlation with other mass-based instruments, although the correlation of absolute values was influenced by different sampling locations.
Acknowledgments This research was supported by the Swiss Agency for the Environment, Forests and Landscape (SAEFL, Bern) and by the Swiss Federal Roads Authority (ASTRA, Bern). We thank Roland Graf (EMPA) and Rolf Ziegler (EMPA) for their assistance in preparing and running the engines. We also thank Pirita Mikkanen (Dekati) and Mikko Moisio (Dekati) for their fruitful discussions.
Literature Cited (1) Gehrig, R.; Hueglin, C.; Hofer, P.; Kobler, J.; Stahel, W. A.; Baltensperger, U.; Monn, C. Int. J. Vehicle Des. 2001, 27, 55-64. (2) Kwon, S.; Lee, K.; Saito, K.; Shinozaki, O.; Seto, T. Environ. Sci. Technol. 2003, 37 (9), 1794-1802. (3) Kagawa, J. Toxicology 2002, 181, 349-353. (4) Osunsanya, T.; Prescott, G.; Seaton, A. Occup. Environ. Med. 2001, 58, 154-159. (5) Lehmann, U.; Mohr, M.; Schweizer, T.; Rutter, J. Atmos. Environ. 2003, 37, 5247-5259. (6) Morawska, L.; Johnson, G.; Ristovski, Z. D.; Agranovski, V. Atmos. Environ. 1999, 33, 1983-1990. (7) Directive 2001. (8) U.S. Environmental Protection Agency, 1997. (9) Moosmueller, H.; Arnott, W. P.; Rogers, C. F.; Bowen, J. L.; Gillies, J. A.; Pierson, W. R. Environ. Sci. Technol. 2001, 35, 1935-1942. (10) Keskinen, J.; Pietarinen, K.; Lehtimaki, M. J. Aerosol Sci. 1992, 23, 353-360. (11) van Gulijk, C.; Schouten, J. M.; Marijnissen, J. C. M.; Makkee, M.; Moulijn, J. A. J. Aerosol Sci. 2001, 32, 1117-1130. (12) Marjamaki, M.; Keskinen, J.; Chen, D.-R.; Pui, D. Y. H. J. Aerosol Sci. 2000, 31, 249-261. (13) Ristimaki, J.; Virtanen, A.; Marjamaki, M.; Rostedt, A.; Keskinen, J. J. Aerosol Sci. 2002, 33, 1541-1557. (14) Moisio, M.; Niemela¨, V. IAC 2002. (15) Harris, S. J.; Maricq, M. M. J. Aerosol Sci. 2001, 32, 749-764. (16) Mathis, U.; Ristima¨ki, J.; Mohr, M.; Keskinen, J.; Ntziachristos, L.; Samaras, Z.; P., M. Aerosol Sci. Technol. (submitted for publication). (17) Koch, W.; Lo¨dding, H.; Mo¨lter, W.; Munzinger, F. StaubReinhaltung der Luft 1988, 48, 341-344. (18) Agarwal, J. K.; Sem, G. J. J. Aerosol Sci. 1980, 11, 343-357. (19) Burtscher, H.; Baltensperger, U.; Bukowiecki, N.; Cohn, P.; Huglin, C.; Mohr, M.; Matter, U.; Nyeki, S.; Schmatloch, V.; Streit, N.; Weingartner, E. J. Aerosol Sci. 2001, 32, 427-442. (20) Knutson, E. O.; Whitby, K. T. J. Aerosol Sci. 1975, 6, 443-451. (21) Wang, S. C.; Flagan, R. C. J. Aerosol Sci. 1989, 20, 1485-1488. (22) GRPE (Working Party on Pollution and Energy) World Forum for Harmonisation of Vehicle Regulations (WP.29) 2001, TRANS/ WP.29/GRPE/2001/2. (23) MacDonald, J. S.; Plee, S. L.; D’Arcy, J. B.; Schreck, R. M. SAE Tech. Pap. Ser. 1980, No. 800185. (24) Ntziachristos, L.; Samaras, Z.; Mohr, M.; Mathis, U.; Keskinen, J.; Ristma¨ki, J.; Mikkanen, P.; Vogt, R. SAE Tech. Pap. Ser. 2004. (25) GRPE (Working Party on Pollution and Energy). 2003.
(26) McDow, S. R.; Huntzicker, J. J. Atmos. Environ. 1990, 24A, 25632571. (27) Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T. Environ. Sci. Technol. 1999, 33, 1578-1587. (28) Chase, R. E.; Duszkiewicz, G. J.; Richert, F. O.; Lewis, D.; Maricq, M.; Xu, N. SAE Tech. Pap. Ser. 2004, No. 2004-01-0967. (29) Schraml, S.; Will, S.; Zens, T.; D’Alfonso, N. SAE Trans. 2000, 109, 1935-1942.
(30) Beck, H.; Niessner, R.; Haisch, C. Anal. Bioanal. Chem. 2003, 375, 1136-1143.
Received for review October 29, 2003. Revised manuscript received August 6, 2004. Accepted August 11, 2004. ES035206P
VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5711
