Emissions of fine particles, NOx, and CO from on-road vehicles in Finland

ARTICLE IN PRESS

Atmospheric Environment 39 (2005) 6696–6706 www.elsevier.com/locate/atmosenv

Emissions of fine particles, NOx, and CO from on-road vehicles in Finland Tarja Yli-Tuomia,, Pa¨ivi Aarnioa, Liisa Pirjolab,c, Timo Ma¨kela¨d, Risto Hillamod, Matti Jantunena a

Department of Environmental Health, KTL, National Public Health Institute, P.O. Box 95, FIN-70701 Kuopio, Finland b Helsinki Polytechnic Stadia, P.O. Box 4020, FIN-00099 Helsinki, Finland c Department of Physical Sciences, University of Helsinki, P.O. Box 64, FIN-00014 Helsinki, Finland d Finnish Meteorological Institute, Air Quality Division, P.O. Box 503, FIN-00101 Helsinki, Finland Received 9 December 2004; received in revised form 12 May 2005; accepted 26 July 2005

Abstract Real-time particle number size distributions and NO, NO2, NOx, CO, and CO2 concentrations were measured with a mobile laboratory van in October 2003 on the streets and highways of the Helsinki metropolitan area, Finland. A bimodal particle size distribution was observed with about 85% of the particles being smaller than 29 nm. Real-time fuel-based emission factors for size-resolved particle numbers, CO, and NOx were determined. Wide distributions of emission factors were obtained for all pollutants. In addition, PM2.5 samples were collected and the elemental compositions were analysed. Relative to fixed site urban PM2.5, street air PM2.5 concentrations of Cu, BC, Fe, and Zn were elevated. Weather and road conditions influenced PM concentrations more than the differences between the city and highway traffic environments. r 2005 Elsevier Ltd. All rights reserved. Keywords: Traffic; Particles; Carbon monoxide; Nitrogen oxides; Emission factors

1. Introduction Exposure in traffic contributes considerably to the total human exposure to air pollutants (Duci et al., 2003; Adams et al., 2001; Gee and Raper, 1999), and knowledge of this exposure is important for the assessment of health effects of pollutants. In traffic environments, the concentrations of traffic-related pollutants are higher than in other environments and a considerable amount of time, on average from 4% to 8% of total hours of the day, is spent in traffic in Corresponding author. Tel./fax: +358 17 201184.

E-mail address: tarja.yli-tuomi@ktl.fi (T. Yli-Tuomi).

developed countries (Eurostat, 2004; Jantunen et al., 1998; Jenkins et al., 1992). In-vehicle exposure to PM2.5 has been reported to cause cardiovascular effects in healthy young patrol officers in North Carolina, US (Riediker et al., 2003). Most health endpoints were associated to a PM2.5 source factor that reflects speedchanging traffic conditions with a high loading of copper, aldehydes, and sulphur (Riediker et al., 2004). The most widely used methods for evaluating vehicle tailpipe emissions are dynamometer tests, which involve measurements of emissions from selected vehicles using standardized driving cycles under controlled conditions. A key concern with these tests is that they do not fully represent real-world driving conditions and emissions.

1352-2310/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.atmosenv.2005.07.049

ARTICLE IN PRESS T. Yli-Tuomi et al. / Atmospheric Environment 39 (2005) 6696–6706

In city traffic, the driving cycle, fleet age, and type distributions, as well as dilution conditions are highly variable compared to the fixed test cycles. Also, the nontailpipe emissions such as those from the wear of brake linings, tyres and asphalt, and resuspension of road and soil dust are not covered by this test method. Alternative approaches to determine the emissions from motor vehicles are tunnel studies, remote sensing techniques, and model calculations combined with monitoring results. Recently, mobile laboratories have emerged as useful tools for real-world emission measurement (Kittelson et al., 2004; Pirjola et al., 2004a; Vogt et al., 2003; Bukowiecki et al., 2002). Most of the on-road motor vehicle emission inventories have been travel based (g km
1). A major disadvantage of this method is that the travel-based emission factors vary much more than the fuel-based emission factors (g (kg fuel)
1) as driving modes change. Fuel-based methods have been used in on-road emissions measurements such as those from remote sensors (Pokharel et al., 2002; Singer and Harley, 1996, 2000), tunnel (Kirchstetter et al., 1999, 2002), and near-road (Shi et al., 2002) studies, as well as measurements with mobile laboratories (Kittelson et al., 2004). In this study, the fuel-based approach with CO2 variation as a tracer of exhaust dilution in real conditions is used on mobile laboratory measurements. In Finland, winter tyres are compulsory in motor vehicles during the winter months and approximately 90% of cars and 0% of buses and trucks have winter tires equipped with metal studs. Also, sand and de-icing salt are used to prevent road slipperiness. These factors have a considerable effect on the wear of road surfaces and resuspension of particles from them (Kupiainen and Tervahattu, 2004). In 2003 in the Helsinki metropolitan area street traffic volume, the share of private cars was 84% of the total mileage and the shares of vans, trucks, and buses were 9%, 5%, and 2%, respectively. The share of private cars with three-way catalysts was 70% of the mileage of gasoline driven cars. Diesel vehicles corresponded to 18% of the mileage of private cars and 92% of the vans (VTT, 2005). Gasoline and diesel fuels sold in Finland are technically sulphur-free (o10 ppm). In Finland, 100% of the total automobile gasoline consumption has been unleaded since 1995. The objective of this study is to determine the emission factors for CO, NOx, and particle numbers in different size classes for the car fleet and traffic conditions in the Helsinki metropolitan area. Also, the particle number concentrations and the chemical composition of fine particles in the midst of traffic in relation to urban ambient levels are determined. Size-fractioned data has not been published before although it is urgently needed for dispersion modelling and exposure assessment.

6697

2. Methods A mobile laboratory called ‘‘Sniffer’’ was utilized to measure real-time traffic emissions in actual traffic conditions. The design of the Volkswagen LT35 diesel van laboratory has been described in detail elsewhere (Pirjola et al., 2004a, b). The mobile laboratory provides measurements of particle total number concentration and size distribution, some gaseous species (CO, CO2, NO, NO2, and NOx), meteorological (T, RH, and wind speed and direction at the roof of the van) and geographical parameters (by a GPS system) with high spatial and time resolutions. Two different inlet systems opening towards the driving direction were constructed, one above the windshield at a height of 2.4 m for general traffic monitoring and the other above the bumper at a height of 0.7 m for chasing a single vehicle. In this study, the sample air was drawn through an inlet above the windscreen. Separate sampling lines were used for particles and gases. In addition to the air pollution measurement devices, the van was equipped with a video camera that recorded the driving conditions through the windscreen. The concentrations of air pollution components were measured in traffic for 3 days on streets of the Helsinki city centre (22–24 October 2003) and for 3 days on highways in the Helsinki metropolitan area (27–29 October 2003). On both environments, the total 12.5 h measurement time included three morning peak periods (between 7:13 and 10:38) and two afternoon peak periods (between 14:33 and 17:22) (Table 1). Particle number concentration and size distribution were measured by the Electrical Low Pressure Impactor (ELPI, Dekati Ltd.) with a flow rate of 10 L min
1. ELPI (with the electrical filter stage) enables real-time particle size distribution in the size range of 7 nm–10 mm (aerodynamic diameter) with 12 channels. The particles are charged, size-classified by inertial impaction, and electrically detected (Keskinen et al., 1992). The offset currents were always checked before and after the measurement. In this work, results of sub-micrometre particles are presented. LI-6262 CO2/H2O Analyzer (LI-COR, inc.) was used for real-time CO2 monitoring. The detection range for CO2 was 0–3000 ppm and accuracy 71 at 350 ppm. The monitor was calibrated against a new factory-calibrated GM70 (Vaisala hand-held CO2 meter). Model CO12M (Environnement S.A.) CO gas monitor had a detection range of 0.05–20 ppm. In the NOx analyser (Model APNA 360, Horiba) the measuring cycle is composed of three different parts: NO-, NOx-, and zero level measurements. The NOx detection limit is about 0.5 ppb and the measurement range reaches up to 4 ppm. CO and NOx analysers were calibrated with five different certified standard gas concentrations. Zero and span checks were performed a couple of times per day for both analysers.

ARTICLE IN PRESS T. Yli-Tuomi et al. / Atmospheric Environment 39 (2005) 6696–6706

6698

Table 1 Driving conditions during the measurement periods Code

Date (October 2003)

Time

Road condition

Temperature (1C)

RH (%)

Speed (km h
1) Average

Centre 22m 22e 23m 23e 24m

Maximum

22 22 23 23 24

08:19–10:38 14:35–17:13 07:28–10:29 14:51–17:16 07:43–10:07

Dry Dry Dry Dry Snowing


3
1
2 0
2

70 60 60 58 80

20 16 19 18 19

55 55 54 58 53

Highway 27m 27 27e 27 28m 28 28e 28 29m 29

07:13–10:12 14:33–17:02 07:35–09:46 14:52–17:22 07:20–09:40

Snow and salt on the road Wet, salt on the road Wet, salt on the road Wet Wet


8 1 6 7 5

87 68 88 80 95

43 50 38 44 47

101 83 82 90 86

Particle number concentration and size distribution, as well as the concentrations of the gases were recorded every second. However, because of the differences in the instrumental reaction times, 10 s average was used for the determination of the emission factors. PM2.5 mass was collected on Teflon (Teflo, 47 mm, 2 mm pore size) and quartz filters using PQ100 pumps with 16.7 L min
1 flow rate and EPA-WINS PM2.5 impactors that were installed into the van for this work. In order to ensure an adequate mass for elemental analysis, PM2.5 was collected on one Teflon filter from both street and highway environments. One quartz filter was used per day. The Teflon filters were weighted for mass and analysed by ED–XRF at the Antwerp University for elemental composition. The quartz filters were analysed for black carbon (BC) in the Finnish Meteorological Institute with a thermal-optical transmission method. The emission factors were calculated not for a single vehicle but for the whole fleet and traffic situations along the routes driven during this study. A short averaging time (10 s) enables reasonable separation of different traffic situations, i.e., short-term values for individual vehicles and driving conditions, and reveals the wide range of emission factors in real life. However, simultaneous influence of more than one vehicle leads to underestimation of the higher and overestimation of the lower emission factors. A fuel-based emission factor (EFP) for pollutant (P) is defined as the ratio of the mass (or number count) of pollutant emitted (MP) per mass of fuel consumed (Mfuel). Since the carbon mass fraction of the fuel (CMFfuel) is the carbon mass of the fuel (CMfuel) divided by its mass (Mfuel), emission factor EFP ¼ CMFfuel

MP . CMfuel

(1a)

The carbon mass emitted by the vehicle equals the carbon mass of the fuel consumed. Therefore, the fuel carbon mass (CMfuel) can be replaced by the carbon mass of the exhaust, which is the sum of the mass of its main carbon-containing components, weighted by their respective carbon mass fractions. Since gasoline engine combustion efficiencies in normal driving situations are better than 90% and diesel even better, carbon mass in vehicle exhaust is mostly in the form of CO2. The emission of CO depends significantly on the vehicle type (catalyst or not, diesel or gasoline), but because the portion of CO is low, a change in CO emission has only a small impact (o10%) on the CO2 emission. The roles of HC, particulate OC, and BC are much smaller. Thus, the contribution of CO and other carbon-containing compounds was neglected, yielding EFP ¼ CMFfuel

MP , M CO2 CMFCO2

(1b)

where M CO2 is the mass of CO2 emitted and CMFCO2 is the carbon mass fraction of CO2, that is 27.3% based on the atomic masses. Moosmuller et al. (2003) used CMFgasoline ¼ 86.7% and CMFdiesel ¼ 85.6% based on empirical fuel formulas yielding 3176 g CO2 (kg gasoline)
1 and 3136 g CO2 (kg diesel)
1. These are in close agreement with data of the Technical Research Centre of Finland (VTT, 2004) with 3133 g CO2 (kg gasoline)
1 and 3148 g CO2 (kg diesel)
1 for reformulated fuels. In Finland, about 10% of gasoline is estimated to be oxygenated with slightly different values. An average of VTT values was used: CMFfuel/ CMFCO2 ¼ 3141 g CO2 (kg fuel)
1. Mass (or number count) of emitted pollutant MP was defined as a difference between measured concentration in traffic [P]t (10 s average) and background concentration [P]bg. The background concentration was defined separately for each measured peak period as the average

ARTICLE IN PRESS

1270 240 2050 3 620000 3 460000 188 000 68 000 52 000 27 500 9190 1840 330

— —

0 0 30 10 000 10 000 1000 1000 600 200 70 10 0 530 120 910 570 000 510 000 25 000 18 000 13 800 5500 1660 380 90

— —

40 30 110 30 000 20 000 2000 2000 1700 700 230 50 10 740 210 60 170 000 140 000 9500 7800 5800 2200 640 150 36

—

1980 390 3100 2 540000 2 420000 104 000 53 000 50 200 22 800 7790 1330 270 0 0 10 10 000 5000 40 70 90 20 10 0 0 640 190 1150 470 000 400 000 33 000 24 000 17 500 6600 1870 330 80

—

210 80 410 140 000 110 000 12 000 9300 5700 1900 450 77 25 200 50 340 160 000 140 000 10 000 7000 5400 2200 650 120 30

— — —

30 30 90 40 000 30 000 3000 2000 1400 500 120 30 10

—

1 3869 3868 3868 4302 4302 4302 4302 4302 4302 4302 4302 4302 23 270 90 500 180 000 150 000 15 000 11 000 7300 2600 700 120 30 1 3455 3455 3455 3704 3704 3703 3704 3704 3704 3704 3699 3695 PM2.5 NO NO2 NOx Ntot N0.007–0.0286 N0.0286–0.056 N0.056–0.0943 N0.0943–0.158 N0. 158–0.266 N0.266–0.388 N0.388–0.621 N0.621–0.960

Minimum 95th % 5th % Median Standard

NOx calculated as NO2. Concentrations of gases in mg m
3, particle numbers in cm
3.

Median Standard n Average n

Maximum

Highway City centre

Table 2 Pollutant concentrations during on-road measurements (10 s averages)

Concentrations of particles larger than 7 nm varied from 1  104 to 4  106 cm
3 with an average of 2  105 cm
3 (Table 2). The average fine particle size distributions during the morning and afternoon rush hours are presented in Fig. 1a. The distributions are similar in the city centre and on the highway, but the rapidly varying traffic situations cause relatively high standard deviations (Fig. 1b). A bimodal structure typical for traffic-related fine particles can be observed. On average, 85% of total particles were smaller than 30 nm and 91% smaller than 56 nm. Unfortunately, the nucleation mode geometric mean diameter cannot be determined due to the lack of detection of particles smaller than 7 nm (the filter stage of the ELPI can detect only particles larger than 7 nm with the upper limit of 28.6 nm determined by the lowest impactor stage; geometric mean diameter is then 14.1 nm). The geometric mean diameters for the soot modes are located at 80–100 nm. A high correlation (R2 40:95) was observed between the concentration of BC (n ¼ 6) and the average number concentration of particles in the size range of 94–158 nm. For more detailed analysis, the median size distributions in linear scale were plotted (Fig. 1c). Two important features can be observed: Firstly, higher nucleation mode median concentrations are measured on the highway than in the city with an exception of the low concentration on the afternoon of 28th October for which no clear reason has been found. Secondly, in three days out of four, the median nucleation mode was lower in the morning than in the afternoon. On the highway in the morning of 27th October, the high nucleation peak might be caused by cold temperature along with very busy traffic of passenger cars. The highest soot mode in the city centre on 24th October might originate from the diesel buses and delivery vehicles and the first snowfall that kept the traffic congested through the day. Correlation of the particle number concentration with CO2 was relatively weak, R2 below 0.5, for all particle sizes. The connection between changes in the CO2 and particle number concentrations was the best for particles in the size range from 56 to 266 nm (R2 in city 0.45, on highway 0.37). The correlation was nearly as good for 266–960 nm-sized particles on both environments (R2 in city 0.37, on highway 0.32). This indicates that traffic is

5th %

3.1. Particle numbers

—

95th %

3. Results

160 30 260 220 000 210 000 11 000 5000 4400 1900 630 140 30

(1c)

Minimum

½Pt
½Pbg CMFfuel . CMFCO2 ½CO2 t
½CO2 bg

Average

EFP ¼

—

of the measurements with the lowest 5% of the measured CO2 concentrations

6699

16 240 70 430 220 000 190 000 11 000 9000 6500 2500 760 180 40

Maximum

T. Yli-Tuomi et al. / Atmospheric Environment 39 (2005) 6696–6706

ARTICLE IN PRESS 6700

T. Yli-Tuomi et al. / Atmospheric Environment 39 (2005) 6696–6706

an important emitter of 56–960 nm particles. Nanoparticles (7–56 nm) correlated only weakly with CO2 (R2 in city 0.16, on highway 0.14). The estimated emission factors by size classes for submicrometre particles are presented in Table 3. The total particle number (Ntot) emissions are dominated by the emissions of the smallest (7–28 nm) particles. The distributions of Ntot emission factors are presented in Fig. 2 for streets and highways. The 10 s emission factor varied in the range of (0.002–720)  1015 particles (kg fuel)
1. Although the overall range of the emission factors was wide, 90% of the values were within (0.6–20)  1015 in the city centre and (1–24)  1015 on the highways. 3.2. CO and NOx The average NO2 concentration was 90 mg m
3 during the city centre measurements and 70 mg m
3 during the highway measurements (Table 2). Compared to the concentrations measured at the same time at the Vallila monitoring site, about 14 m from a road, the concentration of primary pollutant NO was 8 times higher on the road and concentration of NO2, a secondary pollutant, 2 times. The emission factors for CO, NO, NO2, and NOx (calculated as NO2) are shown in Table 3. 3.3. PM2.5

Fig. 1. (a) Average size distribution measurements during the morning (m) and afternoon (a) rush hours in the city centre (c) and on highways (h). (b) Calculated mean values of the 10 s size distributions along with standard deviations in the city centre and on highways. (c) Median size distributions in linear scale.

The road conditions, air temperature, relative humidity, and speed differed substantially between the city centre and highway measurements (Table 1). The wet conditions during the highway measurements seemed to reduce concentration of traffic-related PM2.5, much of which originates from road dust, brakes, tyres, and other mechanical sources in addition to the exhaust emissions. The reduction was confirmed by the data from urban air quality monitoring stations—the average of PM2.5 concentration was higher during the dry than the wet conditions. Furthermore, the dry/wet condition ratio of PM2.5 was higher in stations affected by traffic emissions (YTV, 2003). The effect of wet conditions on resuspension can be seen also in the chemical composition (Table 4) with 63–80% lower concentrations of crustal elements Ti, Si, Ca, and K on the highway. The concentration of Fe, another crustal element, decreased only by 33% and thus, Fe is likely to have other non-crustal sources in the traffic. De-icing salt, mainly NaCl, was applied on the highways between Friday 24th afternoon and early Tuesday 28th morning. During highway measurements, the salt concentration varied from 0 to 1.1 g m
2 with an average of 0.2 g m
2 at Pirkkola monitoring station located on the route (Finnish Road Administration,

Table 3 Estimated short-term (10 s) emission factors City centre Average

Standard Median

5th %

95th %

Minimum Maximum n

Average

Standard Median

5th %

95th %

Minimum Maximum

2906 2288 2058 2308 3099 3040 3137 3072 2945 2787 2692 2927 3075

44 11 4 19 7.2E+15 6.2E+15 5.2E+14 3.5E+14 2.3E+14 8.2E+13 2.1E+13 3.2E+12 9.6E+11

74 30 9 51 1.5E+16 1.4E+16 9.6E+14 6.1E+14 4.2E+14 1.6E+14 4.1E+13 6.5E+12 1.7E+12

3 1 0.14 1 5.5E+14 3.8E+14 5.7E+13 4.2E+13 2.0E+13 6.4E+12 1.3E+12 2.0E+11 9.7E+10

130 29 11 52 2.0E+16 1.8E+16 1.4E+15 8.7E+14 5.6E+14 2.0E+14 5.3E+13 8.4E+12 2.5E+12

0.01 0.01 0.00 0.02 2.4E+12 5.2E+11 3.4E+11 5.7E+11 2.7E+11 4.0E+09 1.8E+09 2.8E+09 5.7E+08

40 10 2 16 9.3E+15 8.4E+15 4.3E+14 3.1E+14 2.3E+14 8.3E+13 2.3E+13 5.3E+12 1.4E+12

65 19 5 32 1.7E+16 1.6E+16 6.6E+14 4.5E+14 3.3E+14 1.3E+14 3.8E+13 8.8E+12 2.3E+12

4 1 0.07 1 9.7E+14 8.0E+14 5.7E+13 5.5E+13 3.9E+13 1.4E+13 3.2E+12 6.5E+11 1.7E+11

110 25 5 40 2.4E+16 2.2E+16 1.1E+15 7.1E+14 5.2E+14 2.0E+14 5.9E+13 1.3E+13 3.6E+12

0.01 0.00 0.00 0.01 1.6E+12 1.1E+13 1.4E+12 3.4E+10 7.5E+11 8.0E+10 5.7E+09 1.6E+10 8.7E+07

27 5 2 10 4.0E+15 3.2E+15 3.2E+14 2.4E+14 1.6E+14 5.6E+13 1.4E+13 2.0E+12 6.5E+11

1890 650 210 1210 2.8E+17 2.8E+17 2.2E+16 1.4E+16 1.1E+16 4.5E+15 1.0E+15 2.4E+14 5.6E+13

3335 3528 2854 3541 3820 3762 3897 3956 3956 3932 3912 3880 3863

25 6 1 10 6.2E+15 5.4E+15 3.1E+14 2.4E+14 1.7E+14 6.1E+13 1.6E+13 3.7E+12 9.6E+11

NOx calculated as NO2. Emission factors of gases in g (kg fuel)
1. Particle numbers in (kg fuel)
1.

1470 560 160 960 7.2E+17 6.5E+17 2.5E+16 2.1E+16 1.5E+16 5.0E+15 1.3E+15 3.2E+14 9.5E+13

Highway 27–29 October 2003

6701

City centre 22–24 October 2003

— 1 65 132 24 507 45 5 1 14 4 290 133 3 — 28 8 16

Fig. 2. Distribution of Ntot emission factors for streets and highways.

163 1 174 23 31 752 118 9 1 19 7 393 583 15 2 33 11 23

Table 4 Chemical composition of PM2.5 on streets and highways

Al Br Ca Cl Cu Fe K Mn Ni P Pb S Si Ti V Zn BC Mass

Mass and BC in mg m
3, others in ng m
3.

2004). Because of the de-icing salt, the Cl concentration in the highway PM2.5 sample was high.

4. Discussion

4.1. Particle numbers

Higher nucleation mode concentrations were found in the morning than in the afternoon on three days out of four. High number concentrations in traffic environments in the morning have been reported previously in

ARTICLE IN PRESS

n

T. Yli-Tuomi et al. / Atmospheric Environment 39 (2005) 6696–6706

CO NO NO2 NOx Ntot N0.007
0.0286 N0.0286
0.056 N0.056
0.0943 N0.0943
0.158 N0. 158
0.266 N0.266
0.388 N0.388
0.621 N0.621
0.960

Highway

ARTICLE IN PRESS T. Yli-Tuomi et al. / Atmospheric Environment 39 (2005) 6696–6706

6702

other studies (Molnar et al., 2002; Wehner et al., 2002; Wahlin et al., 2001). Charron and Harrison (2003) also observed a shift of the nuclei mode diameter from about 23 nm at the beginning of morning rush hour to about 31 nm at afternoon rush hour. They suggested that small particles are formed in the morning or during the night and grow afterward by coagulation and/or condensation of gases. Relatively weak correlations of particle number concentration with CO2 were expected because particle emissions per fuel mass vary considerably for different vehicles and traffic situations. For example, in the highway measurements two episodes with 5–10 times higher Ntot/CO2 ratios compared to other data were observed. These high emission factors were assigned to two trucks. The weakest correlations were found for nanoparticles (7–56 nm), which are secondary particles formed via nucleation processes from semivolatile compounds in the exhaust gas. Nucleation is affected by several external physical (temperature and the availability of larger particles as competing condensation surfaces) and chemical factors, which weaken the direct connection to primary traffic emissions. Similar observations for 11–30 nm particles have been reported by Charron and Harrison (2003). Values of the fuel-based particle number concentrations and emission factors found from the literature are presented in Table 5. Comparison of the results is difficult because not only the lower cut point of the particle counters but also the ambient temperature (California in summer, Helsinki and Minnesota in winter) as well as study design (steady speed in tunnel vs. variable speed in traffic, weaker dilution in tunnel) are likely to affect the measured concentrations and emission factors. Considering the difference in the measured particle size, the results of this study are comparable to the aerosol particle number concentrations and emission factors measured on a Minnesota

highway by Kittelson et al. (2004) as well as those based on measurements by Kirchstetter et al. (2002, 1999). In an on-road measurement campaign in France (Gouriou et al., 2004), particle concentration (D430 nm) was about 3 times higher than concentrations of particles larger than 29 nm in Helsinki. According to Kittelson et al. (2004), engine load, exhaust temperature, and exhaust flow increase at high vehicular speeds resulting in higher nanoparticle emission. In Minnesota highway measurements, lower speeds produced lower number concentrations and larger particles. In the Helsinki metropolitan area, the average Ntot emission factor for particles larger than 7 nm (Table 2) was slightly higher (9.3  1015 particles (kg fuel)
1) on the highway than in the street traffic (7.2  1015 particles (kg fuel)
1). However, weather conditions affected the results and further measurements are needed to reveal differences of the traffic environments. 4.2. CO and NOx The emission factors for both CO and NOx were higher (1.2–1.3 fold) in the street traffic compared to those on the highway. The likely reasons for this were the differences in the driving cycles and the higher portion of heavy-duty diesel vehicles, especially buses and delivery trucks with high NOx emissions, in the city centre. The effect of temperature and humidity on the CO and NOx emission is considered low. Values of fuelbased emission factors for CO, NO, and NOx found from the literature are presented in Table 6. As 3-way catalyst-equipped vehicles are replacing the older fleet, there is a clear trend of decreasing CO emission factors from the early 1990 to 2000 measurements. A similar trend has been reported by Singer and Harley (2000) based on emission measurements of different vehicle model years. There is less data on NOx emission factors, but according to VTT (2004) data in Table 7, NOx

Table 5 Average particle number concentrations and emission factors from different studies Study

Place

Sampling

Type

Particle size (nm)

Concentration (cm
3)

Emission (kg
1)

This study

Helsinki, Finland Helsinki, Finland Minnesota, USA California, USA Rouen, France

October 2003

Mobile

47

2  105

8.3  1015

October 2003

Mobile

429

3.3  104

—

November 2000

Mobile

43

4  105

(2.2–11)  1015

July–August 1997 May 2002

Tunnel

410

(1.6–4.0)  105

Mobile

430

9.5  104

6.3  1015 HD, 4.6  1014 LDa —

This study Kittelson et al. (2004) Kirchstetter et al. (1999, 2002) Gouriou et al. (2004)

Particle number concentration highly depends on the smallest detected size. a HD ¼ heavy-duty diesel truck, LD ¼ light-duty vehicle.

ARTICLE IN PRESS T. Yli-Tuomi et al. / Atmospheric Environment 39 (2005) 6696–6706

6703

Table 6 Comparison of average CO, NO, and NOx emission factors from the literature and present work Study

Place

Year

Type of study

CO (g kg
1)

NO (g kg
1)

NOx (g kg
1)

This study Pokharel et al. (2002) Singer and Harley (2000) Bradley et al. (2000) Bradley et al. (2000) Kirchstetter et al. (1999, 2002) Pokharel et al. (2002) Fraser et al. (1998) Singer and Harley (2000)

Helsinki, Finland Denver, USA Los Angeles, USA Denver, USA Denver, USA California, USA Denver, USA Los Angeles, USA Los Angeles, USA

2003 2001 1997 1997 1997 1997 1996 1993 1991

Mobile Remote sensing Remote sensing Near-road On-road Tunnel Remote sensing Tunnel Remote sensing

55 53 100a 100 118

11 6

20

42 HD, 9 LD 12

84 164a 136a

HD ¼ heavy-duty diesel truck, LD ¼ light-duty vehicle. a g L
1 converted to g kg
1 using density 1.26 L kg
1.

Table 7 VTT data on CO and NOx emission factors by vehicle age group (VTT, 2004) CO (g (kg uel)
1)

NOx (g (kg fuel)
1)

Personal cars Gasoline/street Gasoline/highway Diesel/street Diesel/highway

-1989 non cat 294 198 13 9

200025 14 8 4

-1989 non-catalyst 29 54 19 19

20001 3 11 11

Heavy-duty diesel vehicles Bus/street delivery Truck/street

-1991 32 29

19996 5

-1991 49 48

199920 18

emissions of the new vehicles are also substantially smaller. Since most of the earlier studies have been completed during the 1990s, smaller emission factors were expected and observed in this study. 4.3. Impact of high emitters A small number of high emitting vehicles is responsible for a disproportionately large fraction of mobile emissions. Also, congestion could play a significant role in peak emissions (Seakins et al., 2002; Cadle et al., 1997; Zhang et al., 1994). For both CO and HC, it is typical to find that less than 10% of the fleet is responsible for half of the total exhaust emissions and that less than 20% of the fleet is responsible for 80% of the emissions (Singer and Harley, 1996). The total emission along the driving routes can be estimated as the sum of the products of simultaneous fuel consumptions and emission factors. If a constant fuel consumption (g s
1) was assumed for all vehicles in all traffic situations, the highest 10% of the emission factors would be responsible for 36% of the total emission in the city centre and 30% on the highway. Based on the video screening, the highest emission factors were connected to the older heavy-duty diesel vehicles

accelerating from traffic lights or on uphill. Since the fuel consumption at acceleration is essentially higher than on lower engine loads, and the overall consumption of heavy-duty vehicles is higher than that of personal cars, the share of the high polluters is even higher than the estimated 30–36%. 4.4. PM2.5 Unfortunately, simultaneous data on the off-road PM2.5 chemical composition were not available for comparisons. Instead, the chemical composition of PM2.5 samples (n ¼ 22) collected at the Vallila urban air quality monitoring station as a part of the PAMTOX project from 13 January to 20 April 2000 was used (Sillanpaa et al., 2005). The Vallila monitoring site is located 2 km northeast from the centre of Helsinki, and about 14 m from a road, which has an average traffic load of 14 000 vehicles per day on working days. Compared to the Vallila roadside average concentrations, Cu and BC were clearly elevated with 8–11 times higher concentrations on traffic lanes. In addition, the on-road Fe and Zn concentrations were 2–3 times higher than the average roadside concentrations. However, the peak Fe and Zn concentrations at Vallila during the

ARTICLE IN PRESS 6704

T. Yli-Tuomi et al. / Atmospheric Environment 39 (2005) 6696–6706

Table 8 Traffic-related components in the particle phase recently reported in the literature Study

Type

Size

Traffic related components

Comment

Harrison et al. (2003)

Roadside

PM10

Cu, Zn, Pb, Ba, Mo

Vehicle wear

Sternbeck et al. (2002)

Tunnel

TSP

Cu, Zn, Pb, Ba, Cd, Sb, (Fe)

Mainly from wear; Cu, Ba, Sb brakes

Laschober et al. (2004)

Tunnel

TSP

TC, BC, Cu, Zn, Pb

TC, BC combustion; Cu, Zn, Pb brakes

Allen et al. (2001)

Tunnel

PM10, PM1.9

Zn, Ba, Al, Cr, Fe, Hg, La, Mg, Mn, Na, Sb, Sc, V, OC, BC

Cu and Pb not analysed

Roemer and van Wijnen (2001)

Roadside

TSP

BC

Chellam et al. (2005)

Tunnel

PM2.5

Cu, Zn, Ba

spring dust episode at the end of March 2000 were higher than the on-road concentrations in October 2003. There is a wide temporal variation in the concentrations of PM2.5 chemical components at Vallila (Pakkanen et al., 2001). However, 8–11 times higher on-road Cu and BC concentrations indicate that traffic is one source of these components. Also, Fe and Zn seem to have traffic-related non-crustal sources. Indeed, in several other studies Cu, BC, Fe, and Zn have been reported to be related to particulate traffic emissions (Chellam et al., 2005; Laschober et al., 2004; Harrison et al., 2003; Sternbeck et al., 2002; Allen et al., 2001; Roemer and van Wijnen, 2001; Garg et al., 2000). A variety of other elements have also been connected to traffic emissions from tailpipe, brake and tyre wear, or resuspension (Table 8). In future studies, simultaneous measurements from background areas are needed for thorough analysis. Exposure in traffic can contribute significantly to the exposure to metals, which are important considering the health effects, e.g., cardiopulmonary outcomes (Karlsson et al., 2005; Dai et al., 2002; Costa and Dreher, 1997). Since the concentrations of traffic emission components are higher on the road than few metres away from the road, on-road measurements are needed in order to estimate the exposure to and the health effects of particles while in traffic. Information on the organic compounds in traffic is limited, and more research is needed also to find out the seasonal variation of the PM mass and composition in traffic.

5. Conclusions The particle number concentrations, size distributions, and emission factors do not differ much between

Cu, Zn engine oil and brake wear; Zn tyres; Ba brakes and diesel fuel

the street and highway traffic in the Helsinki metropolitan area. In addition to the traffic type and density, also the road and weather conditions affect the concentrations, especially the traffic-related PM2.5 concentration. Traffic planning should be used to minimize the need for speed changes; but most importantly, emission restrictions are needed also for old vehicles. Because the emissions of new vehicles are already well below those of the average car fleet, further reductions in these emissions would be inefficient in controlling the street traffic emissions in comparison to controlling the respective emissions of the older and sometimes poorly maintained smoking vehicles. On-road measurements are needed for assessment of exposure and health effects: a considerable amount of time is spent in the traffic and the high on-road concentrations of traffic-related pollutants can contribute significantly to the exposure to BC and several transition metals.

Acknowledgements This work has been supported by intramural funding from KTL and Stadia.

References Adams, H.S., Nieuwenhuijsen, M.J., Colvile, R.N., McMullen, M.A.S., Khandelwal, P., 2001. Fine particle (PM2.5) personal exposure levels in transport microenvironments, London, UK. Science of the Total Environment 279, 29–44. Allen, J.O., Mayo, P.R., Hughes, L.S., Salmon, L.G., Cass, G.R., 2001. Emissions of size-segregated aerosols from

ARTICLE IN PRESS T. Yli-Tuomi et al. / Atmospheric Environment 39 (2005) 6696–6706 on-road vehicles in the Caldecott Tunnel. Environmental Science & Technology 35 (21), 4189–4197. Bradley, K.S., Brooks, K.B., Hubbard, L.K., Popp, P.J., Stedman, D.H., 2000. Motor vehicle fleet emissions by OP–FTIR. Environmental Science & Technology 34 (5), 897–899. Bukowiecki, N., Dommen, J., Prevot, A.S.H., Richter, R., Weingartner, E., Baltensperger, U., 2002. A mobile pollutant measurement laboratory—measuring gas phase and aerosol ambient concentrations with high spatial and temporal resolution. Atmospheric Environment 36 (36–37), 5569–5579. Cadle, S.H., Mulawa, P.A., Ball, J., Donase, C., Weibel, A., Sagebiel, J.C., Knapp, K.T., Snow, R., 1997. Particulate emission rates from in use high emitting vehicles recruited in Orange County, California. Environmental Science & Technology 31 (12), 3405–3412. Charron, A., Harrison, R.M., 2003. Primary particle formation from vehicle emissions during exhaust dilution in the roadside atmosphere. Atmospheric Environment 37 (29), 4109–4119. Chellam, S., Kulkarni, P., Fraser, M.P., 2005. Emissions of organic compounds and trace metals in fine particulate matter from motor vehicles: a tunnel study in Houston, Texas. Journal of the Air & Waste Management Association 55 (1), 60–72. Costa, D.L., Dreher, K.L., 1997. Bioavailable transition metals in particulate matter mediate cardiopulmonary injury in healthy and compromised animal models. Environmental Health Perspectives 105, 1053–1060. Dai, J., Xie, C., Churg, A., 2002. Iron loading makes a nonfibrogenic model air pollutant particle fibrogenic in rat tracheal explants. American Journal of Respiratory Cell and Molecular Biology 26 (6), 685–693. Duci, A., Chaloulakou, A., Spyrellis, N., 2003. Exposure to carbon monoxide in the Athens urban area during commuting. Science of the Total Environment 309 (1–3), 47–58. Eurostat, 2004. How Europeans Spend their Time—Everyday Life of Women and Men—Data 1998–2002. 2004 ed. Office for the Official Publications of the European Communities, Luxembourg. Finnish Road Administration, 2004. Road Weather Conditions Database. Fraser, M.P., Cass, G.R., Simoneit, B.R.T., 1998. Gas-phase and particle-phase organic compounds emitted from motor vehicle traffic in a Los Angeles roadway tunnel. Environmental Science & Technology 32 (14), 2051–2060. Garg, B.D., Cadle, S.H., Mulawa, P.A., Groblicki, P.J., Laroo, C., Parr, G.A., 2000. Brake wear particulate matter emissions. Environmental Science & Technology 34 (21), 4463–4469. Gee, I.L., Raper, D.W., 1999. Commuter exposure to respirable particles inside buses and by bicycle. Science of the Total Environment 235 (1–3), 403–405. Gouriou, F., Morin, J.-P., Weill, M.-E., 2004. On-road measurements of particle number concentrations and size distributions in urban and tunnel environments. Atmospheric Environment 38 (18), 2831–2840. Harrison, R.M., Tilling, R., Callen Romero, M.S., Harrad, S., Jarvis, K., 2003. A study of trace metals and polycyclic

6705

aromatic hydrocarbons in the roadside environment. Atmospheric Environment 37 (17), 2391–2402. Jantunen, M.J., Ha¨nninen, O.O., Katsouyanni, K., Kno¨ppel, H., Ku¨nzli, N., Lebret, E., Maroni, M., Saarela, K., Sra´m, R.J., Zmirou, D., 1998. Air pollution exposure in European cities: the EXPOLIS-study. Journal of Exposure Analysis and Environmental Epidemiology 8 (4), 495–518. Jenkins, P.L., Phillips, T.J., Mulberg, E.J., Hui, S.P., 1992. Activity patterns of Californians: use of and proximity to indoor pollutant sources. Atmospheric Environment 26A, 2141–2148. Karlsson, H.L., Nilsson, L., Moller, L., 2005. Subway particles are more genotoxic than street particles and induce oxidative stress in cultured human lung cells. Chemical Research in Toxicology 18 (1), 19–23. Keskinen, J., Pietarinen, K., Lehtimaki, M., 1992. Electrical low pressure impactor. Journal of Aerosol Science 23 (4), 353–360. Kirchstetter, T.W., Harley, R.A., Kreisberg, N.M., Stolzenburg, M.R., Hering, S.V., 1999. On-road measurement of fine particle and nitrogen oxide emissions from light- and heavy-duty motor vehicles. Atmospheric Environment 33 (18), 2955–2968. Kirchstetter, T.W., Harley, R.A., Kreisberg, N.M., Stolzenburg, M.R., Hering, S.V., 2002. Corrigendum to ‘‘On-road measurement of fine particle and nitrogen oxide emissions from light- and heavy-duty motor vehicles’’ [Atmospheric Environment 33 (18) (1999) 2955–2968.]. Atmospheric Environment 36 (39–40), 6059. Kittelson, D.B., Watts, W.F., Johnson, J.P., 2004. Nanoparticle emissions on Minnesota highways. Atmospheric Environment 38 (1), 9–19. Kupiainen, K., Tervahattu, H., 2004. The effect of traction sanding on urban suspended particles in Finland. Environmental Monitoring and Assessment 93 (1–3), 287–300. Laschober, C., Limbeck, A., Rendl, J., Puxbaum, H., 2004. Particulate emissions from on-road vehicles in the Kaisermuhlen-tunnel (Vienna, Austria). Atmospheric Environment 38 (14), 2187–2195. Molnar, P., Janhall, S., Hallquist, M., 2002. Roadside measurements of fine and ultrafine particles at a major road north of Gothenburg. Atmospheric Environment 36 (25), 4115–4123. Moosmuller, H., Mazzoleni, C., Barber, P.W., Kuhns, H.D., Keislar, R.E., Watson, J.G., 2003. On-road measurement of automotive particle emissions by ultraviolet lidar and transmissometer: instrument. Environmental Science & Technology 37 (21), 4971–4978. Pakkanen, T.A., Loukkola, K., Korhonen, C.H., Aurela, M., Makela, T., Hillamo, R.E., Aarnio, P., Koskentalo, T., Kousa, A., Maenhaut, W., 2001. Sources and chemical composition of atmospheric fine and coarse particles in the Helsinki area. Atmospheric Environment 35 (32), 5381–5391. Pirjola, L., Parviainen, H., Hussein, T., Valli, A., Hameri, K., Aalto, P., Virtanen, A., Keskinen, J., Pakkanen, T.A., Makela, T., Hillamo, R.E., 2004a. ‘‘Sniffer’’—a novel tool for chasing vehicles and measuring traffic pollutants. Atmospheric Environment 38 (22), 3625–3635. Pirjola, L., Parviainen, H., Lappi, M., Ha¨meri, K., Hussein, T., 2004b. A novel mobile laboratory for ‘‘chasing’’ city traffic. SAE Paper No. 2004-01-1962.

ARTICLE IN PRESS 6706

T. Yli-Tuomi et al. / Atmospheric Environment 39 (2005) 6696–6706

Pokharel, S.S., Bishop, G.A., Stedman, D.H., 2002. An on-road motor vehicle emissions inventory for Denver: an efficient alternative to modeling. Atmospheric Environment 36 (33), 5177–5184. Riediker, M., Williams, R., Devlin, R., Griggs, T., Bromberg, P., 2003. Exposure to particulate matter, volatile organic compounds, and other air pollutants inside patrol cars. Environmental Science & Technology 37 (10), 2084–2093. Riediker, M., Devlin, R.B., Griggs, T.R., Herbst, M.C., Bromberg, P.A., Williams, R.W., Cascio, W.E., 2004. Cardiovascular effects in patrol officers are associated with fine particulate matter from brake wear and engine emissions. Particle and Fibre Toxicology 1 (1), 2. Roemer, W.H., van Wijnen, J.H., 2001. Differences among black smoke, PM10, and PM1.0 levels at urban measurement sites. Environmental Health Perspectives 109 (2), 151–154. Seakins, P.W., Lansley, D.L., Hodgson, A., Huntley, N., Pope, F., 2002. New directions: mobile laboratory reveals new issues in urban air quality. Atmospheric Environment 36 (7), 1247–1248. Shi, J.P., Harrison, R.M., Evans, D.E., Alam, A., Barnes, C., Carter, G., 2002. A method for measuring particle number emissions from vehicles driving on the road. Environmental Technology 23 (1), 1–14. Sillanpaa, M., Hillamo, R., Salonen, R.O., 2005, personal communication. Singer, B.C., Harley, R.A., 1996. A fuel-based motor vehicle emission inventory. Journal of the Air & Waste Management Association 46 (6), 581–593.

Singer, B.C., Harley, R.A., 2000. A fuel-based inventory of motor vehicle exhaust emissions in the Los Angeles area during summer 1997. Atmospheric Environment 34 (11), 1783–1795. Sternbeck, J., Sjodin, A., Andreasson, K., 2002. Metal emissions from road traffic and the influence of resuspension—results from two tunnel studies. Atmospheric Environment 36 (30), 4735–4744. Vogt, R., Scheer, V., Casati, R., Benter, T., 2003. On-road measurement of particle emission in the exhaust plume of a diesel passenger car. Environmental Science & Technology 37 (18), 4070–4076. VTT, LIPASTO, 2004 http://lipasto.vtt.fi/indexe.htm, (accessed 7 July 2004). VTT, LIISA, 2005. http://www.lipasto.vtt.fi/lipasto/liisa/index.htm, (accessed 8 April 2005). Wahlin, P., Palmgren, F., Van Dingenen, R., 2001. Experimental studies of ultrafine particles in streets and the relationship to traffic. Atmospheric Environment 35 (Suppl. 1), 63–69. Wehner, B., Birmili, W., Gnauk, T., Wiedensohler, A., 2002. Particle number size distributions in a street canyon and their transformation into the urban-air background: measurements and a simple model study. Atmospheric Environment 36 (13), 2215–2223. YTV (Helsinki Metropolitan Area Council), 2003. Air Quality Database. Zhang, Y., Bishop, G.A., Stedman, D.H., 1994. Automobile Emissions Are Statistically Gamma-Distributed. Environmental Science & Technology 28 (7), 1370–1374.