- Email: [email protected]
Vertical variations of particle number concentration and size distribution in a street canyon in Shanghai, China
Science of the Total Environment 378 (2007) 306 – 316 www.elsevier.com/locate/scitotenv
Vertical variations of particle number concentration and size distribution in a street canyon in Shanghai, China X.L. Li a,b,⁎, J.S. Wang a , X.D. Tu a , W. Liu a,b , Z. Huang a a b
Center for Combustion and Environmental Technology, Shanghai Jiao Tong University, Shanghai, 200030, China School of Environmental and Science and Engineering, Shanghai Jiao Tong University, Shanghai, 200030, China Received 24 June 2006; received in revised form 29 January 2007; accepted 28 February 2007 Available online 5 April 2007
Abstract Measurements of particle number size distribution in the range of 10–487 nm were made at four heights on one side of an asymmetric street canyon on Beijing East Road in Shanghai, China. The result showed that the number size distributions were bimodal or trimodal and lognormal in form. Within a certain height from 1.5 to 20 m, the particle size distributions significantly changed with increasing height. The particle number concentrations in the nucleation mode and in the Aitken mode significantly dropped, and the peaking diameter in the Aitken mode shifted to larger sizes. The variations of the particle number size distributions in the accumulation mode were less significant than those in the nucleation and Aitken modes. The particle number size distributions slightly changed with increasing height ranging from 20 to 38 m. The particle number concentrations in the street canyon showed a stronger association with the pre-existing particle concentrations and the intensity of the solar radiation when the traffic flow was stable. The particle number concentrations were observed higher in Test I than in Test II, probably because the small pre-existing particle concentrations and the intense solar radiation promoted the formation of new particles. The pollutant concentrations in the street canyon showed a stronger association with wind speed and direction. For example, the concentrations of total particle surface area, total particle volume, PM2.5 and CO were lower in Test I (high wind speed and step-up canyon) than in Test II (low wind speed and wind blowing parallel to the canyon). The equations for the normalized concentration curves of the total particle number, CO and PM2.5 in Test I and Test II were derived. A power functions was found to be a good estimator for predicting the concentrations of total particle number, CO and PM2.5 at different heights. The decay rates of PM2.5 and CO concentrations were lower in Test I than in Test II. However, the decay rate of the total particle number concentration in Test I was similar to that in Test II. No matter how the wind direction changed, for example, in the step-up case or wind blowing parallel to the canyon, the decay rates of the total particle number concentration were larger than those of PM2.5 and CO concentrations. For example, CO concentrations decreased by 0.33 and 0.69 at the heights ranging from 1.5 to 38 m in Test I and Test II, while the total particle number concentrations decreased by 0.72 and 0.85 within the same height ranges in Test I and Test II. It is concluded that the coagulation process, besides the dilution process, affected the total particle number concentration. © 2007 Elsevier B.V. All rights reserved. Keywords: Street canyon; Ultrafine particle; Number concentration; Size distribution; Concentration decay rate
⁎ Corresponding author. Center for Combustion and Environmental Technology, Shanghai Jiao Tong University, Shanghai, 200030, China. Tel.: +86 21 6407 4085; fax: +86 21 6407 8095. E-mail addresses: [email protected], [email protected] (X.L. Li). 0048-9697/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2007.02.040
1. Introduction Recent epidemiology studies have suggested that there is a correlation between exposure to ambient ultrafine
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
particles with higher number concentration and adverse health effects (Peters and Wichmann, 2001). Ultrafine particles in urban environments primarily result from the vehicle emissions (Schauer et al., 1996; Shi et al., 1999). Vehicles emit appreciable amounts of organic semivolatile compounds (Kittelson et al., 2000) and small amounts of sulphur compounds (Schneider et al., 2005) that may be involved in the formation of new particles in the atmosphere. On-road chase experiments by Kittelson et al. (2000) showed a strong dependence of particle size distributions on dilution conditions, with larger concentrations of nanoparticles formed during dilution at lower ambient temperatures. Another on-road chase experiment by Schneider et al. (2005) showed that the exhaust particle mode around 100 nm was markedly smaller than in the high-sulphur case (360 ppm, S) when low sulphur fuel (30 ppm, S) was used at 100 km h− 1 vehicle speed. Recently, several groups have measured the traffic-related particle number concentrations in the roadside environment e.g. (Shi et al., 1999; Woo et al., 2001; Charron and Harrison, 2003; Mönkkönen et al., 2004). They found that the pre-existing particle concentration and the solar radiation significantly affected the formation of new particles. Charron and Harrison (2003) found that “cleaner atmospheres” created by stronger winds and rain evens promoted the occurrence of high numbers of ultrafine particles. These particles are not primarily emitted by vehicles but formed via nucleation process during the cooling and dilution of vehicle exhausts in the atmosphere. Mönkkönen et al. (2004) measured the aerosol number concentration and the PM10 mass concentration of urban background aerosols in different seasons in New Delhi, Indian. They found, the number concentration increased with the mass concentration up to 0.3 mg m− 3. However, above this point, the number concentration decreased even if the mass concentration increased. This was due to the high coagulation scavenging ratio with high concentrations of pre-existing particles. Measurements of the aerosol size distribution in the range from 3 nm to 2 μm diameter were carried out over a 24 month period in Atlanta, GA by Woo et al. (2001). They observed some events where pronounced peaks in the size range of 3–10 nm occurred. These events typically occurred around noon, when the solar radiation was high. A temporal association between nanoparticles in the size range of 3–7 nm and the solar radiation was also observed by Shi et al. (1999) in the urban background air when no other local sources were influential. A relatively narrow street with buildings lined up continuously along both sides of a street is a common configuration in an urban district. This typical configuration is the so-called street canyon (Nicholson, 1975).
307
The street canyon might be classified as symmetric, if the buildings flanking the street have the same height approximately, or asymmetric. Assimakopoulos et al. (2003) studied the wind field and pollution dispersion patterns in both of symmetric and asymmetric canyons with perpendicular background winds with a computational fluid dynamics code. A main vortex covered approximately the top 2/3 of the street canyon depth, while a secondary weak counter rotating vortex was inferred close to the floor in a symmetric street canyon where the street width to building height aspect ratio was 1:2. When the upwind building was reduced, the main vortex was shifted upwards, its centre being almost higher than the upwind building roof with Hu/Hd = 1/2. Furthermore, the flow within the canyon was isolated from the flow above the buildings, since the main vortex acts as a cap. The flow inside the canyon was nearly stagnant and the pollutants diffuse in an upward manner. Boddy et al. (2005a) measured CO and airflow on both sides of two street canyons. The study showed evidence that near-parallel wind flow played a significant role in the dispersion of CO in both street canyons. High CO concentrations on each side of the street axis coincided with nearparallel background flows. This indicated the presence of helical vortices along the street canyon. Helical flows were shown to cause significantly higher concentrations on the leeward street side than parallel flows. Recently, experimental investigations of the spatial transformation of the ultrafine particle number concentration and size distribution in the street canyon and their transformation into the urban-air background were conducted (e.g. Väkevä et al., 1999; Wehner et al., 2002; Longley et al., 2003; Longley et al., 2004). Väkevä et al. (1999) measured the gas and aerosol pollutants in a symmetric street canyon in Lahti, Finland. Monitoring was undertaken at two heights: at street level (1.5 m) and at rooftop (22 m). The result showed that dilution factor was on average 5 between street and rooftop levels. Using a theoretical expression based on the measured data, they inferred that the nucleation probability was significant at rooftop (22 m) because of the low pre-existing particles at rooftop level. Wehner et al. (2002) measured the particle number size distributions in a street canyon in Leipzig, Germany. They found the maximum in the number size distribution was at a diameter of about 15 nm in the street canyon and occurred during rush hour, while in the urban background, this maximum was about 20–25 nm. The increase in the maximum indicated that during transport from the street canyon to the urban background, the particle diameters had enough time to grow by aerosol dynamical processes. Longley et al. (2003) measured the particle number concentration and size distribution in an
308
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
Fig. 1. Map of the location of the experimental canyon in the Shanghai area.
asymmetric street canyon of Princess Street, Manchester. They found, at a fixed measurement site, the average particle number concentrations increased about 3 times when the wind direction turned from “Up-canyon” (the wind blowing parallel to the canyon) to “SW-perpendicular” (perpendicular or near perpendicular flow of synoptic winds) due to the combination of poor dispersion in low winds and the shorter distance from the source to the receptor over which dilution might occur. Longley et al. (2004) further studied particle number size distributions at different heights in the street canyon. The result showed that the total number concentration at 17 m was generally 50% those at 4 m during the day. In “SW-perpendicular” flow, the number spectrum at 17 m was slightly broadened towards larger particles compared with that at 4 m. In this work, a field study was undertaken in a street canyon. The particle number concentrations and size distributions at four different heights with wind directions perpendicular or blowing parallel to the street were measured. The concentrations of PM2.5 and CO in the street canyon were also measured. The aim was to study the effects of variations of wind speed, wind direction and height on the particle number concentration and size distribution in the street canyon.
north latitude and 121°29′ east longitude. The city is adjacent to East China Sea and the climate varies with four seasons. The prevailing wind direction shows south wind in summer and north wind in winter. Mean annual wind is 3.1 m s− 1. A field monitoring was conducted on Beijing East Road, an arterial street with busy traffic and east to west direction. The map of the location of the canyon in Shanghai is shown in Fig. 1. Fig. 2 shows a map of the immediate neighborhood of the measurement site. There is a path leading to the west end of the segment, in which there is very little traffic. Two near-parallel narrow streets
2. Sampling and instruments 2.1. Site Shanghai, the largest city in the east of China, has a population of 13.3 million. The city is situated at 31°14′
Fig. 2. Map of the immediate neighborhood of the experimental site.
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
309
Fig. 3. Plan of experimental canyon site.
lie to the south and north of Beijing East Road. Both narrow streets had a negligible amount of traffic. Henan Middle Road is at the east end of the segment with little traffic compared with that in Beijing East Road. The chosen segment of the street is 125 m long and asymmetric as shown in Fig. 3. This segment is 21 m wide with five lanes of traffic and the sidewalks on both sides are 3 m and 6 m, respectively. The buildings along the south side have a height of 18 m, and a variety of buildings on the north side of heights are 39–50 m. 2.2. Instruments A Scanning Mobility Particle Sizer (SMPS, TSI Model 3034) was used to measure the particles in the range from 10 to 487 nm by separating particles based on their electrical mobility. Particles of a selected size are detected optically, using a detection technology through which the small particle visibility is enhanced by “growing” the particles in condensing butyl alcohol vapour. The device used for particle separation is referred to as Differential Mobility Size Analyzer or DMA. The particle counter is referred to as a Condensation Particle Counter or CPC due to the particle growth mechanism. The entire system is automated. The total particle concentration is 102 to 107 particles/cm3, and a scanning time is 3 min. A DustTrak, TSI Model 8520 Aerosol Monitor, was used to measure the mass concentration of PM2.5. A tapered element oscillating microbalance (TEOM) of 1400a ambient particle monitor (Rupprecht and Patashnick Co., Inc., Albany, NY) was used to recalibrate the reading of the DustTrak before the field measurement on Beijing East Road. CO Concentrations were measured with a portable infrared CO monitor (Model GXH-3011, Beijing Huayun Instru-
ments Inc., precision of 0. 10 ppm). The CO monitor was calibrated with standard gases in lab. Wind speed and direction were measured by using a three-cup aerovane (Model FYF-1, Shanghai Fengyun Meteorology Instruments Inc.). 2.3. Field measurement methods The particle number concentration and size distribution, CO and PM2.5 concentrations were measured on 30–31May (named Test I) and 23–25 November 2005 (named Test II). The sample started at 9:00 a.m. and ended at 17:00 p.m. The distribution of monitoring sites is shown in Fig. 4. The experiments were only undertaken
Fig. 4. Inferred air flow field in Test I and the distribution of monitoring sites in the vertical section and within the street canyon.
310
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
Table 1 Average meteorological and traffic conditions −1
Traffic (vehicles h ) Number of trucks (%) Wind speed (m s− 1) Temperature (°C) Relative humidity (%)
30 May
31 May
23 November
24 November
25 November
1780 ± 240 8±2 3.6 ± 0.9 30.1 ± 1.4 38.5 ± 6
1660 ± 150 8±2 2.9 ± 1.0 32.8 ± 1.8 41.1 ± 4
1727 ± 140 8±1 1.3 ± 0.2 21.1 ± 1.2 38.2 ± 3
1794 ± 214 8±2 1.4 ± 0.3 18.1 ± 3.1 30.7 ± 5
1659 ± 180 7±1 1.2 ± 0.3 19.2 ± 3.0 36.2 ± 5
on the north side of the street canyon. The procedures of the measurement are given here: All the instruments, including the SMPS, the DustTrak and the CO monitor, were placed on a trolley, and the trolley was pushed to each monitoring site by an elevator. Storage batteries were used to provide continuous source of power for all the instruments. Measurements were taken at the heights of 1.5, 8, 20 and 38 m above the pavement. Sampling tubes of each instrument were extended out horizontally from the window, and the distance from the sampling inlets to the wall of the north side was 1.5 m. At each location,
three size distribution samples were taken with the SMPS. The particle number concentrations and size distributions, CO and PM2.5 concentrations were monitored simultaneously at each location. It took about 15 min to complete one sampling at each location and about 60 min to complete one set at all the four locations. Eight to ten sets of the measurements were performed everyday. The urban background concentrations were measured on the roof of a building 400 m to the south of Beijing East Road, an area with less traffic. The building is 55 m high. The background measurement was conducted before and after the canyon measurement everyday, and each measurement lasted 1 h. The three-cup aerovane was erected on the roof of the building on the north side. The
Fig. 5. Wind speed and direction at sampling site in: (a) Test I and (b) Test II.
Fig. 6. Average particle number size distribution vertical variations in: (a) Test I and (b) Test II.
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
height of this building is greater than the average height in Shanghai city center and is not overlooked. The measurements also included meteorological parameters such as temperature, atmospheric pressure and relative humidity on the roof. 2.4. Meteorological conditions and traffic volume The average values of the meteorological and traffic parameters and their standard deviations are presented in Table 1.The traffic flow was estimated by counting the numbers of light duty vehicles, medium duty vehicles and heavy duty vehicles passing by in random 5 min periods during the measurement periods. The average traffic volumes per hour during the measurement periods were generally 1700–1800 vehicles h− 1 and the number of the light duty vehicles, medium duty vehicles and heavy duty vehicles accounted for about 70%, 22% and 8% of the total vehicles, respectively. The meteorological conditions in Test I and Test II were significantly different. The mean wind speed was about 3 m s − 1 and the prevailing wind blew perpendicularly to the street from south in Test I (Fig. 5a), which means that the wind approached over the shorter
311
south wall of the street canyon. This configuration was the typical step-up street canyon. The prevailing wind blew parallel to the canyon from the west and the mean wind speed was about 1.6 m s− 1 in Test II (Fig. 5b). The relative humidity was about 30–40% during all of the measurement periods. In Test I, the environmental temperature was about 30 °C, in Test II about 20 °C. 3. Results 3.1. Changes in particle number size distribution with increasing height Generally, the size range definitions for atmospheric particles are: nucleation mode (diameter b 0.02 μm), Aitken mode (0.02 μm b diameter b 0.1 μm), accumulation mode (0.1 μm b diameter b 1 μm) and coarse particles (diameter N 1 μm). The nucleation mode particles are thought to be formed through nucleation process directly from gas phase species such as sulphuric acid. The Aitken mode particles arise from the growth or coagulation of smaller particles, and are also produced in high numbers by primary combustion
Fig. 7. Comparison of the average particle number size distributions in Test I and Test II at: (a) 1.5 m, (b) 8 m, (c) 20 m, and (d) 38 m.
312
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
sources such as vehicle exhausts. The accumulation mode particles are mainly due to the coagulation of particles with diameters smaller than 0.1 μm and condensation of vapours onto pre-existing particles, and the source of coarse particles may be re-suspension by traffic and traffic-induced turbulence in the atmosphere. Fig. 6a and b depicts the average particle number size distributions at 1.5, 8, 20 and 38 m heights in the street canyon in Test I and Test II. The data were averaged for all applicable sampling data for each height in Test I and Test II. Three distinct modes peaking at 10.4 nm (belonging to nucleation mode), 28.4 nm (belonging to the Aitken mode) and 89.8 nm (belonging to accumulation mode) were observed at 1.5 m above the pavement in Test I (Fig. 6a). The particle size distributions in the nucleation mode changed significantly with increasing height. Within the height ranging from 1.5 to 8 m above the pavement, the peak number concentrations of the nucleation mode and Aitken mode dropped sharply with increasing height. Within the range from 8 to 20 m, the peak number concentration of the nucleation mode went on dropping significantly, while that of the Aitken mode dropped insignificantly. However, the peak diameter of the nucleation mode, fairly stabled in the position of 10.4 nm within the height range from 1.5 to 8 m was not observed at height N 20 m above the pavement. Within the height range from 1.5 to 20 m, the peak diameter of the Aitken mode shifted to larger particle diameters with increasing height, which shifted from 28.4 nm (at 1.5 m) to 35.4 nm (at 20 m). The particle size distribution in the nucleation and the Aitken modes changed insignificantly within the range from 20 to 38 m. The study showed that the size distribution of the particles in the accumulation mode changed insignificantly within the range from 1.5 to 38 m. Three distinct modes, the Aitken1, Aitken2 and accumulation modes were also observed at 1.5 m above the pavement in Test II (Fig. 6b). The Aitken1 mode was peaking at 22.9 nm, the Aitken2 mode at 35.2 nm and the accumulation mode as shown in Fig. 6b. The Aitken2 mode disappeared at 8 m. The value of mode in the Aitken1 mode shifted to larger particle diameters with increasing height, which shifted from 22.9 nm (at 1.5 m) to 28.4 nm (at 20 m). The particle size distribution in the Aitken1 mode changed insignificantly within the height ranging from 20 to 38 m. The peak number concentration of the accumulation mode dropped slightly within the height ranging from 1.5 to 38 m and its size distribution changed insignificantly. Fig. 7a–d gave the comparison of the average particle number size distributions in Test I and Test II
Fig. 8. Fitted and measured average particle size distributions at the height of 1.5 m in: (a) Test I and (b) Test II.
at identical heights. In general, the total particle number concentrations were higher in Test I than in Test II, especially in the nucleation mode. However, this situation reversed above about 150 nm. Fig. 8a and b fits the lognormal distributions of the average particle number concentrations at 1.5 m in Test I and Test II. As shown in the figures, both of them are trimodal and lognormal in form. 3.2. Vertical characteristic of total particle number, total particle surface area, total particle volume, PM2.5 and CO concentrations In order to further study the vertical characteristic of the ultrafine particles, Fig. 9a–e illustrates the average concentrations of CO, PM2.5, total particle number, total particle surface area and total particle volume with their standard deviations as a function of heights in Test I and Test II. The average concentrations of CO, PM2.5, total particle surface area and total particle volume were about 1.5–3 times larger in Test II than in Test I at identical heights. However, compared with that of other pollutants, the total particle number concentration,
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
313
Fig. 9. Averaged concentrations of (a) CO, (b) PM2.5, (c) total particle number, (d) total particle surface and (e) total particle volume with their standard deviations as a function of height in Test I and Test II.
showing an unusual behaviour, was about 1.5 times larger in Test I than in Test II. Based on the measurements, the equations for the concentration decay curves with increasing height in the street canyon were derived. To eliminate the urban background effect on the street canyon, the pollutant concentrations were normalized by the formula. Cnorm ¼
Ci Cb C1:5 Cb
Where Ci refers to the pollution concentrations at different heights, i = 1.5, 8, 20 and 38 m; Cb is urban background concentrations. Fig. 10a–c show the normalized decay curves of concentrations of the total particle number, CO and PM2.5 in Test I and Test II. The concentration decay rates of the CO and PM2.5 were significantly lower in Test I than in Test II. However, the concentration decay rate of the total particle number in Test I was similar to that in Test II. The concentration decay rate of the total particle number was larger
314
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
Fig. 10. Curves of the normalized concentrations of (a) CO, (b) PM2.5, and (c) total particle number as a function of the height in Test I and Test II.
than those of CO and PM2.5 both in Test I and Test II (Fig. 10a–c). For example, the CO concentration decreased by 0.33 and 0.69 within the height ranging from 1.5 m to 38 m in Test I and Test II. However, the total particle number concentration decreased by 0.72 and 0.85 within the identical range of height in Test I and Test II. 4. Discussions In Test I and Test II, at a certain height in the range from 1.5 to 20 m, with increasing height, the peak
number concentrations in the nucleation mode and Aitken mode significantly dropped and the peak diameter in the Aitken mode shifted to larger diameters, but the peak number concentration and the peak diameter in the accumulation mode varied slightly, which suggests that the coagulation affects the smallest ultrafine particles more significantly than large particles. Although the data of the traffic flow of the two periods were approximately equal to each other, pollutant concentrations and vertical gradients in Test I greatly differed from those in Test II. The difference between the nucleation mode in Test I and that in Test II may be explained by the nucleation process of the volatile matter emitted from the vehicle. The sulphur reacts with excess oxygen to form SO2 gas, and a few fractions of SO2 further turn out SO3. The SO3 reacts with H2O to form H2SO4 vapours. H2SO4–H2O nucleation forms new particles (0.1 nm b diameter b 1 nm) even if the vapours pressure is very low (Baumgard and Johnson, 1996). However, with the onset of the solar radiation, photochemical reactions resulting in the production of condensable species (e.g., SO2 + OH → H2SO4) are started (Weingartner et al., 1998). If the pre-existing aerosol surface area concentration is low, the homogeneous nucleation is favoured and new particles are formed. It was fine in Test I but cloudy in Test II, so the solar radiation was more intense in Test I than in Test II. The intense solar radiation may be a reason why a large number of fine particles appeared in Test I. This is identical to the work of Woo et al. (2001) and Shi et al. (1999). Vehicles emit amounts of semi-volatile species, such as heavy hydrocarbons (Kittelson, 1998), small amounts of sulphur compounds (Allen et al., 2001) and ammonia. These matters are potentially involved in new particle formation process. The exhaust gases are cooled and diluted in the atmospheric environment. The saturation ratio of low volatility compounds is influenced by environmental meteorological parameters, such as temperatures and humidity. When the saturation ratio reaches the critical value, the process of nucleation leading to formation of new particles occurs. However, the high concentration of pre-existing particles may promote the condensation of the semi-volatile compounds and suppress the growth of fresh nuclei (Kerminen et al., 2001). Small pre-existing aerosol concentrations favour the production of new particles and their growth to a detectable size in the atmosphere (Kulmala et al., 2000). It was also found that the total particle surface area and PM2.5 concentrations were about 1.5–2 times higher in Test II than in Test I, which means that the pre-existing aerosol surface area concentration was larger in Test II than in Test I. The
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
condensation of emitted condensable gases onto existing particles is likely to be the most favourable process in Test II and the process of nucleation was suppressed. Therefore, the lower pre-existing particle density may be another reason why a large number of fine particles appeared in Test I. This is identical with the work of Charron and Harrison (2003) and Mönkkönen et al. (2004). The difference between the concentrations of total particle surface area, total particle volume, CO and PM2.5 in Test I and those in Test II may be caused by the influence of the wind speed and direction. In Test I, the upwind building was shorter than the downwind building and the canyon street configuration was similar to the step-up canyon as described by Assimakopoulos et al. (2003). In Test II, the wind speed was lower, the average wind speed was about 1.5 m s− 1, and the wind blew parallel to the canyon. Huang et al. (2000) performed numerical tests of more practical cases, such as tests with the inclined inflow wind or unequal building heights. They concluded that the pollutants released inside street canyons would be diluted more effectively in step-up streets. In addition, increasing inflow turbulence intensities lowered the street-level pollutant concentration. More limited vertical mixing of pollutants in parallel conditions was observed by Boddy et al. (2005b). They found that the difference in CO with height on one side of the street was significant with the winds from the direction for which the background wind was upstream of a long section of the street canyon. The favourable wind velocity and direction may lead to the decrease in the concentrations of total particle surface area, total particle volume, CO and PM2.5 concentrations in Test I, compared with those in Test II. No matter how the wind direction changed, as shown in the step-up case or the case of wind blowing parallel to the canyon, the normalized total particle number concentration decay rates were larger than those of CO and PM2.5. Concentrations of PM2.5 were far higher than those generally reported in some street canyon aerosol studies e.g. (Ketzel et al., 2003), and thus the coagulation time scale relatively decreased, while the rate of the coagulation scavenging increased in this case. 5. Concluding remark A field experiment has been conducted to measure the particle number size distributions at four heights in one side of an asymmetric street canyon on Beijing East Road in Shanghai, China. It is exhibited that particle number size distributions were bimodal or trimodal and lognormal in form. At a certain height ranging from 1.5
315
to 20 m, particle number concentrations and size distributions significantly varied with increasing height. It is showed that the peak number concentrations in the nucleation and Aitken modes significantly dropped, and the peak diameter in the Aitken mode shifted to larger diameters. The particle size distribution in the accumulation mode varied slightly. With further increase in height, no significant variation was observed in the particle size distribution. The particle number density in the street canyon seems to depend on the pre-existing particles and the intensity of the solar radiation with the stable traffic flow. The high number densities in Test I may be the result of new particle formation from low volatility vapours because of the low pre-existing particles and intense the solar radiation. On the contrary, high pre-existing particle concentrations and the weak solar radiation in Test II may promote the condensation of the semivolatile vapours and disfavour the formation of the new particles. Due to the effect of the wind speed and direction, the concentrations of PM2.5, CO, total particle surface area and total particle volume were lower in Test I (in step-up street with high wind speed) than in Test II (with low wind speed and wind blowing parallel to the canyon). The equations for the normalized concentration curves of total particle number, CO and PM2.5 in Test I and Test II were derived. A power function was found to be a good estimator for the decrease of the pollutant concentration decay rates of total particle number, CO and PM2.5. The decay rates of PM2.5 and CO concentrations were lower in Test I than in Test II. However, the concentration decay rate of the total particle number in Test I was similar to that in Test II. No matter how the wind direction changed, as shown in the step-up case or the case of wind blowing parallel to the canyon, the decay rates of the total particle number concentrations were larger than those of PM2.5 and CO concentrations. For example, CO concentration decreased by 0.33 and 0.69 from 1.5 to 38 m in Test I and Test II. However, the total particle number concentration decreased by 0.72 and 0.85 within the same height ranges in Test I and Test II, which clearly shows that the coagulation process, besides the dilution process, affects the total particle number concentration. References Allen JO, Mayo PR, Hughes LS, Salmon LG, Cass GR. Emissions of size-segregated aerosols from on road vehicles in the Caldecott tunnel. Environ Sci Technol 2001;35:4189–97. Assimakopoulos VD, ApSimon HM, Moussiopoulos N. A numerical study of atmospheric pollutant dispersion in different two-
316
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
dimensional street canyon configurations. Atmos Environ 2003;37: 4037–49. Baumgard KJ, Johnson JH. The effect of fuel and engine design on diesel exhaust particle size distributions. SAE; 1996 [Paper number 960131]. Boddy JWD, Smalley RJ, Dixona NS, Tate JE, Tomlin AS. The spatial variability in concentrations of a traffic-related pollutant in two street canyons in York, UK—Part I: the influence of background winds. Atmos Environ 2005a;39:3147–61. Boddy JWD, Smalley RJ, Goodman PS, Tate JE, Bell MC, Tomlin AS. The spatial variability in concentrations of a traffic-related pollutant in two street canyons in York, UK—Part II: the influence of traffic characteristics. Atmos Environ 2005b;39:3163–76. Charron A, Harrison RM. Primary particle formation from vehicle emissions during exhaust dilution in the roadside atmosphere. Atmos Environ 2003;37:4109–19. Huang H, Akutsu Y, Arai M, Tamura M. A two-dimensional air quality model in an urban street canyon: evaluation and sensitivity analysis. Atmos Environ 2000;34:689–98. Kerminen V-M, Pirjola L, Kulmala M. How significantly does coagulational scavenging limit atmospheric particle production? J Geophys Res 2001;106:24119–25. Ketzel M, Wåhlin P, Berkowicz R, Palmgren F. Particle and trace gas emission factors under urban driving conditions in Copenhagen based on street and roof-level observations. Atmos Environ 2003;37:2735–49. Kittelson DB. Engines and nanoparticles: a review. J Aerosol Sci 1998;29:575–88. Kittelson DB, Johnson J, Watts W, Wei Q, Drayton M, Paulsen D, et al. Diesel aerosol sampling in the atmosphere. SAE; 2000 [Paper number 2000-01-2212]. Kulmala M, Pirjola L, Makela JM. Stable sulphate clusters as a source of new atmospheric particles. Nature 2000;404:66–9. Longley ID, Gallagher MW, Dorsey JR, Flynn M, Allan JD, Alfarra MR, et al. A case study of aerosol (4.6 nm b Dp b 10 μm) number and mass size distribution measurements in a busy street canyon in Manchester, UK. Atmos Environ 2003;37:1563–71.
Longley ID, Gallagher MW, Dorsey JR, Flynn M, Bower KN, Allan JD. Street canyon aerosol pollutant transport measurements. Sci Total Environ 2004;334–335:327–36. Mönkkönen P, Uma R, Srinivasan D, Koponen IK, Lehtinen KEJ, Hämeri K, et al. Relationship and variations of aerosol number and PM10 mass concentrations in a highly polluted urban environment— New Delhi, India. Atmos Environ 2004;38:425–33. Nicholson SE. A pollution model for street-level air. Atmos Environ 1975;9:19–31. Peters A, Wichmann HE. Epidemiological evidence on the health effects of ultrafine particles. Epidemiology 2001;12:544. Schauer JJ, Rogge WF, Hildemann LM, Mazurek MA, Cass GR, Simoneit BRT. Source apportionment of airborne particulate matter using organic compounds as tracers. Atmos Environ 1996;30: 3837–55. Schneider J, Hock N, Weimer S, Borrmann S. Nucleation particles in diesel exhaust: composition inferred from in-situ mass spectrometric analysis. Environ Sci Technol 2005;39:6153–61. Shi JP, Khan AA, Harrison RM. Measurements of ultrafine particle concentration and size distribution in the urban atmosphere. Sci Total Environ 1999;235:51–64. Väkevä M, Hämeri K, Kulmala M, Lahdes R, Ruuskanen J, Laitinen T. Street level versus rooftop concentrations of submicron aerosol particles and gaseous pollutants in an urban street canyon. Atmos Environ 1999;33(9):1385–97. Wehner B, Birmili W, Gnauk T, Wiedensonhler A. Particle number size distributions in a street canyon and their transformation into the urban-air background: measurements and simple model study. Atmos Environ 2002;36:2215–23. Weingartner E, Nyeki S, Streit N, Baltensperger U. Seasonal variation of size distributions and fine particle formation of continental aerosol at high alpine site. J Aerosol Sci 1998;29:S509–10. Woo KS, Chen DR, Puid YH, Mcmurry PH. Measurement of Atlanta aerosol size distributions: observations of ultrafine particle events. Aerosol Sci Technol 2001;34(1):75–87.
Vertical variations of particle number concentration and size distribution in a street canyon in Shanghai, China X.L. Li a,b,⁎, J.S. Wang a , X.D. Tu a , W. Liu a,b , Z. Huang a a b
Center for Combustion and Environmental Technology, Shanghai Jiao Tong University, Shanghai, 200030, China School of Environmental and Science and Engineering, Shanghai Jiao Tong University, Shanghai, 200030, China Received 24 June 2006; received in revised form 29 January 2007; accepted 28 February 2007 Available online 5 April 2007
Abstract Measurements of particle number size distribution in the range of 10–487 nm were made at four heights on one side of an asymmetric street canyon on Beijing East Road in Shanghai, China. The result showed that the number size distributions were bimodal or trimodal and lognormal in form. Within a certain height from 1.5 to 20 m, the particle size distributions significantly changed with increasing height. The particle number concentrations in the nucleation mode and in the Aitken mode significantly dropped, and the peaking diameter in the Aitken mode shifted to larger sizes. The variations of the particle number size distributions in the accumulation mode were less significant than those in the nucleation and Aitken modes. The particle number size distributions slightly changed with increasing height ranging from 20 to 38 m. The particle number concentrations in the street canyon showed a stronger association with the pre-existing particle concentrations and the intensity of the solar radiation when the traffic flow was stable. The particle number concentrations were observed higher in Test I than in Test II, probably because the small pre-existing particle concentrations and the intense solar radiation promoted the formation of new particles. The pollutant concentrations in the street canyon showed a stronger association with wind speed and direction. For example, the concentrations of total particle surface area, total particle volume, PM2.5 and CO were lower in Test I (high wind speed and step-up canyon) than in Test II (low wind speed and wind blowing parallel to the canyon). The equations for the normalized concentration curves of the total particle number, CO and PM2.5 in Test I and Test II were derived. A power functions was found to be a good estimator for predicting the concentrations of total particle number, CO and PM2.5 at different heights. The decay rates of PM2.5 and CO concentrations were lower in Test I than in Test II. However, the decay rate of the total particle number concentration in Test I was similar to that in Test II. No matter how the wind direction changed, for example, in the step-up case or wind blowing parallel to the canyon, the decay rates of the total particle number concentration were larger than those of PM2.5 and CO concentrations. For example, CO concentrations decreased by 0.33 and 0.69 at the heights ranging from 1.5 to 38 m in Test I and Test II, while the total particle number concentrations decreased by 0.72 and 0.85 within the same height ranges in Test I and Test II. It is concluded that the coagulation process, besides the dilution process, affected the total particle number concentration. © 2007 Elsevier B.V. All rights reserved. Keywords: Street canyon; Ultrafine particle; Number concentration; Size distribution; Concentration decay rate
⁎ Corresponding author. Center for Combustion and Environmental Technology, Shanghai Jiao Tong University, Shanghai, 200030, China. Tel.: +86 21 6407 4085; fax: +86 21 6407 8095. E-mail addresses: [email protected], [email protected] (X.L. Li). 0048-9697/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2007.02.040
1. Introduction Recent epidemiology studies have suggested that there is a correlation between exposure to ambient ultrafine
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
particles with higher number concentration and adverse health effects (Peters and Wichmann, 2001). Ultrafine particles in urban environments primarily result from the vehicle emissions (Schauer et al., 1996; Shi et al., 1999). Vehicles emit appreciable amounts of organic semivolatile compounds (Kittelson et al., 2000) and small amounts of sulphur compounds (Schneider et al., 2005) that may be involved in the formation of new particles in the atmosphere. On-road chase experiments by Kittelson et al. (2000) showed a strong dependence of particle size distributions on dilution conditions, with larger concentrations of nanoparticles formed during dilution at lower ambient temperatures. Another on-road chase experiment by Schneider et al. (2005) showed that the exhaust particle mode around 100 nm was markedly smaller than in the high-sulphur case (360 ppm, S) when low sulphur fuel (30 ppm, S) was used at 100 km h− 1 vehicle speed. Recently, several groups have measured the traffic-related particle number concentrations in the roadside environment e.g. (Shi et al., 1999; Woo et al., 2001; Charron and Harrison, 2003; Mönkkönen et al., 2004). They found that the pre-existing particle concentration and the solar radiation significantly affected the formation of new particles. Charron and Harrison (2003) found that “cleaner atmospheres” created by stronger winds and rain evens promoted the occurrence of high numbers of ultrafine particles. These particles are not primarily emitted by vehicles but formed via nucleation process during the cooling and dilution of vehicle exhausts in the atmosphere. Mönkkönen et al. (2004) measured the aerosol number concentration and the PM10 mass concentration of urban background aerosols in different seasons in New Delhi, Indian. They found, the number concentration increased with the mass concentration up to 0.3 mg m− 3. However, above this point, the number concentration decreased even if the mass concentration increased. This was due to the high coagulation scavenging ratio with high concentrations of pre-existing particles. Measurements of the aerosol size distribution in the range from 3 nm to 2 μm diameter were carried out over a 24 month period in Atlanta, GA by Woo et al. (2001). They observed some events where pronounced peaks in the size range of 3–10 nm occurred. These events typically occurred around noon, when the solar radiation was high. A temporal association between nanoparticles in the size range of 3–7 nm and the solar radiation was also observed by Shi et al. (1999) in the urban background air when no other local sources were influential. A relatively narrow street with buildings lined up continuously along both sides of a street is a common configuration in an urban district. This typical configuration is the so-called street canyon (Nicholson, 1975).
307
The street canyon might be classified as symmetric, if the buildings flanking the street have the same height approximately, or asymmetric. Assimakopoulos et al. (2003) studied the wind field and pollution dispersion patterns in both of symmetric and asymmetric canyons with perpendicular background winds with a computational fluid dynamics code. A main vortex covered approximately the top 2/3 of the street canyon depth, while a secondary weak counter rotating vortex was inferred close to the floor in a symmetric street canyon where the street width to building height aspect ratio was 1:2. When the upwind building was reduced, the main vortex was shifted upwards, its centre being almost higher than the upwind building roof with Hu/Hd = 1/2. Furthermore, the flow within the canyon was isolated from the flow above the buildings, since the main vortex acts as a cap. The flow inside the canyon was nearly stagnant and the pollutants diffuse in an upward manner. Boddy et al. (2005a) measured CO and airflow on both sides of two street canyons. The study showed evidence that near-parallel wind flow played a significant role in the dispersion of CO in both street canyons. High CO concentrations on each side of the street axis coincided with nearparallel background flows. This indicated the presence of helical vortices along the street canyon. Helical flows were shown to cause significantly higher concentrations on the leeward street side than parallel flows. Recently, experimental investigations of the spatial transformation of the ultrafine particle number concentration and size distribution in the street canyon and their transformation into the urban-air background were conducted (e.g. Väkevä et al., 1999; Wehner et al., 2002; Longley et al., 2003; Longley et al., 2004). Väkevä et al. (1999) measured the gas and aerosol pollutants in a symmetric street canyon in Lahti, Finland. Monitoring was undertaken at two heights: at street level (1.5 m) and at rooftop (22 m). The result showed that dilution factor was on average 5 between street and rooftop levels. Using a theoretical expression based on the measured data, they inferred that the nucleation probability was significant at rooftop (22 m) because of the low pre-existing particles at rooftop level. Wehner et al. (2002) measured the particle number size distributions in a street canyon in Leipzig, Germany. They found the maximum in the number size distribution was at a diameter of about 15 nm in the street canyon and occurred during rush hour, while in the urban background, this maximum was about 20–25 nm. The increase in the maximum indicated that during transport from the street canyon to the urban background, the particle diameters had enough time to grow by aerosol dynamical processes. Longley et al. (2003) measured the particle number concentration and size distribution in an
308
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
Fig. 1. Map of the location of the experimental canyon in the Shanghai area.
asymmetric street canyon of Princess Street, Manchester. They found, at a fixed measurement site, the average particle number concentrations increased about 3 times when the wind direction turned from “Up-canyon” (the wind blowing parallel to the canyon) to “SW-perpendicular” (perpendicular or near perpendicular flow of synoptic winds) due to the combination of poor dispersion in low winds and the shorter distance from the source to the receptor over which dilution might occur. Longley et al. (2004) further studied particle number size distributions at different heights in the street canyon. The result showed that the total number concentration at 17 m was generally 50% those at 4 m during the day. In “SW-perpendicular” flow, the number spectrum at 17 m was slightly broadened towards larger particles compared with that at 4 m. In this work, a field study was undertaken in a street canyon. The particle number concentrations and size distributions at four different heights with wind directions perpendicular or blowing parallel to the street were measured. The concentrations of PM2.5 and CO in the street canyon were also measured. The aim was to study the effects of variations of wind speed, wind direction and height on the particle number concentration and size distribution in the street canyon.
north latitude and 121°29′ east longitude. The city is adjacent to East China Sea and the climate varies with four seasons. The prevailing wind direction shows south wind in summer and north wind in winter. Mean annual wind is 3.1 m s− 1. A field monitoring was conducted on Beijing East Road, an arterial street with busy traffic and east to west direction. The map of the location of the canyon in Shanghai is shown in Fig. 1. Fig. 2 shows a map of the immediate neighborhood of the measurement site. There is a path leading to the west end of the segment, in which there is very little traffic. Two near-parallel narrow streets
2. Sampling and instruments 2.1. Site Shanghai, the largest city in the east of China, has a population of 13.3 million. The city is situated at 31°14′
Fig. 2. Map of the immediate neighborhood of the experimental site.
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
309
Fig. 3. Plan of experimental canyon site.
lie to the south and north of Beijing East Road. Both narrow streets had a negligible amount of traffic. Henan Middle Road is at the east end of the segment with little traffic compared with that in Beijing East Road. The chosen segment of the street is 125 m long and asymmetric as shown in Fig. 3. This segment is 21 m wide with five lanes of traffic and the sidewalks on both sides are 3 m and 6 m, respectively. The buildings along the south side have a height of 18 m, and a variety of buildings on the north side of heights are 39–50 m. 2.2. Instruments A Scanning Mobility Particle Sizer (SMPS, TSI Model 3034) was used to measure the particles in the range from 10 to 487 nm by separating particles based on their electrical mobility. Particles of a selected size are detected optically, using a detection technology through which the small particle visibility is enhanced by “growing” the particles in condensing butyl alcohol vapour. The device used for particle separation is referred to as Differential Mobility Size Analyzer or DMA. The particle counter is referred to as a Condensation Particle Counter or CPC due to the particle growth mechanism. The entire system is automated. The total particle concentration is 102 to 107 particles/cm3, and a scanning time is 3 min. A DustTrak, TSI Model 8520 Aerosol Monitor, was used to measure the mass concentration of PM2.5. A tapered element oscillating microbalance (TEOM) of 1400a ambient particle monitor (Rupprecht and Patashnick Co., Inc., Albany, NY) was used to recalibrate the reading of the DustTrak before the field measurement on Beijing East Road. CO Concentrations were measured with a portable infrared CO monitor (Model GXH-3011, Beijing Huayun Instru-
ments Inc., precision of 0. 10 ppm). The CO monitor was calibrated with standard gases in lab. Wind speed and direction were measured by using a three-cup aerovane (Model FYF-1, Shanghai Fengyun Meteorology Instruments Inc.). 2.3. Field measurement methods The particle number concentration and size distribution, CO and PM2.5 concentrations were measured on 30–31May (named Test I) and 23–25 November 2005 (named Test II). The sample started at 9:00 a.m. and ended at 17:00 p.m. The distribution of monitoring sites is shown in Fig. 4. The experiments were only undertaken
Fig. 4. Inferred air flow field in Test I and the distribution of monitoring sites in the vertical section and within the street canyon.
310
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
Table 1 Average meteorological and traffic conditions −1
Traffic (vehicles h ) Number of trucks (%) Wind speed (m s− 1) Temperature (°C) Relative humidity (%)
30 May
31 May
23 November
24 November
25 November
1780 ± 240 8±2 3.6 ± 0.9 30.1 ± 1.4 38.5 ± 6
1660 ± 150 8±2 2.9 ± 1.0 32.8 ± 1.8 41.1 ± 4
1727 ± 140 8±1 1.3 ± 0.2 21.1 ± 1.2 38.2 ± 3
1794 ± 214 8±2 1.4 ± 0.3 18.1 ± 3.1 30.7 ± 5
1659 ± 180 7±1 1.2 ± 0.3 19.2 ± 3.0 36.2 ± 5
on the north side of the street canyon. The procedures of the measurement are given here: All the instruments, including the SMPS, the DustTrak and the CO monitor, were placed on a trolley, and the trolley was pushed to each monitoring site by an elevator. Storage batteries were used to provide continuous source of power for all the instruments. Measurements were taken at the heights of 1.5, 8, 20 and 38 m above the pavement. Sampling tubes of each instrument were extended out horizontally from the window, and the distance from the sampling inlets to the wall of the north side was 1.5 m. At each location,
three size distribution samples were taken with the SMPS. The particle number concentrations and size distributions, CO and PM2.5 concentrations were monitored simultaneously at each location. It took about 15 min to complete one sampling at each location and about 60 min to complete one set at all the four locations. Eight to ten sets of the measurements were performed everyday. The urban background concentrations were measured on the roof of a building 400 m to the south of Beijing East Road, an area with less traffic. The building is 55 m high. The background measurement was conducted before and after the canyon measurement everyday, and each measurement lasted 1 h. The three-cup aerovane was erected on the roof of the building on the north side. The
Fig. 5. Wind speed and direction at sampling site in: (a) Test I and (b) Test II.
Fig. 6. Average particle number size distribution vertical variations in: (a) Test I and (b) Test II.
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
height of this building is greater than the average height in Shanghai city center and is not overlooked. The measurements also included meteorological parameters such as temperature, atmospheric pressure and relative humidity on the roof. 2.4. Meteorological conditions and traffic volume The average values of the meteorological and traffic parameters and their standard deviations are presented in Table 1.The traffic flow was estimated by counting the numbers of light duty vehicles, medium duty vehicles and heavy duty vehicles passing by in random 5 min periods during the measurement periods. The average traffic volumes per hour during the measurement periods were generally 1700–1800 vehicles h− 1 and the number of the light duty vehicles, medium duty vehicles and heavy duty vehicles accounted for about 70%, 22% and 8% of the total vehicles, respectively. The meteorological conditions in Test I and Test II were significantly different. The mean wind speed was about 3 m s − 1 and the prevailing wind blew perpendicularly to the street from south in Test I (Fig. 5a), which means that the wind approached over the shorter
311
south wall of the street canyon. This configuration was the typical step-up street canyon. The prevailing wind blew parallel to the canyon from the west and the mean wind speed was about 1.6 m s− 1 in Test II (Fig. 5b). The relative humidity was about 30–40% during all of the measurement periods. In Test I, the environmental temperature was about 30 °C, in Test II about 20 °C. 3. Results 3.1. Changes in particle number size distribution with increasing height Generally, the size range definitions for atmospheric particles are: nucleation mode (diameter b 0.02 μm), Aitken mode (0.02 μm b diameter b 0.1 μm), accumulation mode (0.1 μm b diameter b 1 μm) and coarse particles (diameter N 1 μm). The nucleation mode particles are thought to be formed through nucleation process directly from gas phase species such as sulphuric acid. The Aitken mode particles arise from the growth or coagulation of smaller particles, and are also produced in high numbers by primary combustion
Fig. 7. Comparison of the average particle number size distributions in Test I and Test II at: (a) 1.5 m, (b) 8 m, (c) 20 m, and (d) 38 m.
312
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
sources such as vehicle exhausts. The accumulation mode particles are mainly due to the coagulation of particles with diameters smaller than 0.1 μm and condensation of vapours onto pre-existing particles, and the source of coarse particles may be re-suspension by traffic and traffic-induced turbulence in the atmosphere. Fig. 6a and b depicts the average particle number size distributions at 1.5, 8, 20 and 38 m heights in the street canyon in Test I and Test II. The data were averaged for all applicable sampling data for each height in Test I and Test II. Three distinct modes peaking at 10.4 nm (belonging to nucleation mode), 28.4 nm (belonging to the Aitken mode) and 89.8 nm (belonging to accumulation mode) were observed at 1.5 m above the pavement in Test I (Fig. 6a). The particle size distributions in the nucleation mode changed significantly with increasing height. Within the height ranging from 1.5 to 8 m above the pavement, the peak number concentrations of the nucleation mode and Aitken mode dropped sharply with increasing height. Within the range from 8 to 20 m, the peak number concentration of the nucleation mode went on dropping significantly, while that of the Aitken mode dropped insignificantly. However, the peak diameter of the nucleation mode, fairly stabled in the position of 10.4 nm within the height range from 1.5 to 8 m was not observed at height N 20 m above the pavement. Within the height range from 1.5 to 20 m, the peak diameter of the Aitken mode shifted to larger particle diameters with increasing height, which shifted from 28.4 nm (at 1.5 m) to 35.4 nm (at 20 m). The particle size distribution in the nucleation and the Aitken modes changed insignificantly within the range from 20 to 38 m. The study showed that the size distribution of the particles in the accumulation mode changed insignificantly within the range from 1.5 to 38 m. Three distinct modes, the Aitken1, Aitken2 and accumulation modes were also observed at 1.5 m above the pavement in Test II (Fig. 6b). The Aitken1 mode was peaking at 22.9 nm, the Aitken2 mode at 35.2 nm and the accumulation mode as shown in Fig. 6b. The Aitken2 mode disappeared at 8 m. The value of mode in the Aitken1 mode shifted to larger particle diameters with increasing height, which shifted from 22.9 nm (at 1.5 m) to 28.4 nm (at 20 m). The particle size distribution in the Aitken1 mode changed insignificantly within the height ranging from 20 to 38 m. The peak number concentration of the accumulation mode dropped slightly within the height ranging from 1.5 to 38 m and its size distribution changed insignificantly. Fig. 7a–d gave the comparison of the average particle number size distributions in Test I and Test II
Fig. 8. Fitted and measured average particle size distributions at the height of 1.5 m in: (a) Test I and (b) Test II.
at identical heights. In general, the total particle number concentrations were higher in Test I than in Test II, especially in the nucleation mode. However, this situation reversed above about 150 nm. Fig. 8a and b fits the lognormal distributions of the average particle number concentrations at 1.5 m in Test I and Test II. As shown in the figures, both of them are trimodal and lognormal in form. 3.2. Vertical characteristic of total particle number, total particle surface area, total particle volume, PM2.5 and CO concentrations In order to further study the vertical characteristic of the ultrafine particles, Fig. 9a–e illustrates the average concentrations of CO, PM2.5, total particle number, total particle surface area and total particle volume with their standard deviations as a function of heights in Test I and Test II. The average concentrations of CO, PM2.5, total particle surface area and total particle volume were about 1.5–3 times larger in Test II than in Test I at identical heights. However, compared with that of other pollutants, the total particle number concentration,
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
313
Fig. 9. Averaged concentrations of (a) CO, (b) PM2.5, (c) total particle number, (d) total particle surface and (e) total particle volume with their standard deviations as a function of height in Test I and Test II.
showing an unusual behaviour, was about 1.5 times larger in Test I than in Test II. Based on the measurements, the equations for the concentration decay curves with increasing height in the street canyon were derived. To eliminate the urban background effect on the street canyon, the pollutant concentrations were normalized by the formula. Cnorm ¼
Ci Cb C1:5 Cb
Where Ci refers to the pollution concentrations at different heights, i = 1.5, 8, 20 and 38 m; Cb is urban background concentrations. Fig. 10a–c show the normalized decay curves of concentrations of the total particle number, CO and PM2.5 in Test I and Test II. The concentration decay rates of the CO and PM2.5 were significantly lower in Test I than in Test II. However, the concentration decay rate of the total particle number in Test I was similar to that in Test II. The concentration decay rate of the total particle number was larger
314
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
Fig. 10. Curves of the normalized concentrations of (a) CO, (b) PM2.5, and (c) total particle number as a function of the height in Test I and Test II.
than those of CO and PM2.5 both in Test I and Test II (Fig. 10a–c). For example, the CO concentration decreased by 0.33 and 0.69 within the height ranging from 1.5 m to 38 m in Test I and Test II. However, the total particle number concentration decreased by 0.72 and 0.85 within the identical range of height in Test I and Test II. 4. Discussions In Test I and Test II, at a certain height in the range from 1.5 to 20 m, with increasing height, the peak
number concentrations in the nucleation mode and Aitken mode significantly dropped and the peak diameter in the Aitken mode shifted to larger diameters, but the peak number concentration and the peak diameter in the accumulation mode varied slightly, which suggests that the coagulation affects the smallest ultrafine particles more significantly than large particles. Although the data of the traffic flow of the two periods were approximately equal to each other, pollutant concentrations and vertical gradients in Test I greatly differed from those in Test II. The difference between the nucleation mode in Test I and that in Test II may be explained by the nucleation process of the volatile matter emitted from the vehicle. The sulphur reacts with excess oxygen to form SO2 gas, and a few fractions of SO2 further turn out SO3. The SO3 reacts with H2O to form H2SO4 vapours. H2SO4–H2O nucleation forms new particles (0.1 nm b diameter b 1 nm) even if the vapours pressure is very low (Baumgard and Johnson, 1996). However, with the onset of the solar radiation, photochemical reactions resulting in the production of condensable species (e.g., SO2 + OH → H2SO4) are started (Weingartner et al., 1998). If the pre-existing aerosol surface area concentration is low, the homogeneous nucleation is favoured and new particles are formed. It was fine in Test I but cloudy in Test II, so the solar radiation was more intense in Test I than in Test II. The intense solar radiation may be a reason why a large number of fine particles appeared in Test I. This is identical to the work of Woo et al. (2001) and Shi et al. (1999). Vehicles emit amounts of semi-volatile species, such as heavy hydrocarbons (Kittelson, 1998), small amounts of sulphur compounds (Allen et al., 2001) and ammonia. These matters are potentially involved in new particle formation process. The exhaust gases are cooled and diluted in the atmospheric environment. The saturation ratio of low volatility compounds is influenced by environmental meteorological parameters, such as temperatures and humidity. When the saturation ratio reaches the critical value, the process of nucleation leading to formation of new particles occurs. However, the high concentration of pre-existing particles may promote the condensation of the semi-volatile compounds and suppress the growth of fresh nuclei (Kerminen et al., 2001). Small pre-existing aerosol concentrations favour the production of new particles and their growth to a detectable size in the atmosphere (Kulmala et al., 2000). It was also found that the total particle surface area and PM2.5 concentrations were about 1.5–2 times higher in Test II than in Test I, which means that the pre-existing aerosol surface area concentration was larger in Test II than in Test I. The
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
condensation of emitted condensable gases onto existing particles is likely to be the most favourable process in Test II and the process of nucleation was suppressed. Therefore, the lower pre-existing particle density may be another reason why a large number of fine particles appeared in Test I. This is identical with the work of Charron and Harrison (2003) and Mönkkönen et al. (2004). The difference between the concentrations of total particle surface area, total particle volume, CO and PM2.5 in Test I and those in Test II may be caused by the influence of the wind speed and direction. In Test I, the upwind building was shorter than the downwind building and the canyon street configuration was similar to the step-up canyon as described by Assimakopoulos et al. (2003). In Test II, the wind speed was lower, the average wind speed was about 1.5 m s− 1, and the wind blew parallel to the canyon. Huang et al. (2000) performed numerical tests of more practical cases, such as tests with the inclined inflow wind or unequal building heights. They concluded that the pollutants released inside street canyons would be diluted more effectively in step-up streets. In addition, increasing inflow turbulence intensities lowered the street-level pollutant concentration. More limited vertical mixing of pollutants in parallel conditions was observed by Boddy et al. (2005b). They found that the difference in CO with height on one side of the street was significant with the winds from the direction for which the background wind was upstream of a long section of the street canyon. The favourable wind velocity and direction may lead to the decrease in the concentrations of total particle surface area, total particle volume, CO and PM2.5 concentrations in Test I, compared with those in Test II. No matter how the wind direction changed, as shown in the step-up case or the case of wind blowing parallel to the canyon, the normalized total particle number concentration decay rates were larger than those of CO and PM2.5. Concentrations of PM2.5 were far higher than those generally reported in some street canyon aerosol studies e.g. (Ketzel et al., 2003), and thus the coagulation time scale relatively decreased, while the rate of the coagulation scavenging increased in this case. 5. Concluding remark A field experiment has been conducted to measure the particle number size distributions at four heights in one side of an asymmetric street canyon on Beijing East Road in Shanghai, China. It is exhibited that particle number size distributions were bimodal or trimodal and lognormal in form. At a certain height ranging from 1.5
315
to 20 m, particle number concentrations and size distributions significantly varied with increasing height. It is showed that the peak number concentrations in the nucleation and Aitken modes significantly dropped, and the peak diameter in the Aitken mode shifted to larger diameters. The particle size distribution in the accumulation mode varied slightly. With further increase in height, no significant variation was observed in the particle size distribution. The particle number density in the street canyon seems to depend on the pre-existing particles and the intensity of the solar radiation with the stable traffic flow. The high number densities in Test I may be the result of new particle formation from low volatility vapours because of the low pre-existing particles and intense the solar radiation. On the contrary, high pre-existing particle concentrations and the weak solar radiation in Test II may promote the condensation of the semivolatile vapours and disfavour the formation of the new particles. Due to the effect of the wind speed and direction, the concentrations of PM2.5, CO, total particle surface area and total particle volume were lower in Test I (in step-up street with high wind speed) than in Test II (with low wind speed and wind blowing parallel to the canyon). The equations for the normalized concentration curves of total particle number, CO and PM2.5 in Test I and Test II were derived. A power function was found to be a good estimator for the decrease of the pollutant concentration decay rates of total particle number, CO and PM2.5. The decay rates of PM2.5 and CO concentrations were lower in Test I than in Test II. However, the concentration decay rate of the total particle number in Test I was similar to that in Test II. No matter how the wind direction changed, as shown in the step-up case or the case of wind blowing parallel to the canyon, the decay rates of the total particle number concentrations were larger than those of PM2.5 and CO concentrations. For example, CO concentration decreased by 0.33 and 0.69 from 1.5 to 38 m in Test I and Test II. However, the total particle number concentration decreased by 0.72 and 0.85 within the same height ranges in Test I and Test II, which clearly shows that the coagulation process, besides the dilution process, affects the total particle number concentration. References Allen JO, Mayo PR, Hughes LS, Salmon LG, Cass GR. Emissions of size-segregated aerosols from on road vehicles in the Caldecott tunnel. Environ Sci Technol 2001;35:4189–97. Assimakopoulos VD, ApSimon HM, Moussiopoulos N. A numerical study of atmospheric pollutant dispersion in different two-
316
X.L. Li et al. / Science of the Total Environment 378 (2007) 306–316
dimensional street canyon configurations. Atmos Environ 2003;37: 4037–49. Baumgard KJ, Johnson JH. The effect of fuel and engine design on diesel exhaust particle size distributions. SAE; 1996 [Paper number 960131]. Boddy JWD, Smalley RJ, Dixona NS, Tate JE, Tomlin AS. The spatial variability in concentrations of a traffic-related pollutant in two street canyons in York, UK—Part I: the influence of background winds. Atmos Environ 2005a;39:3147–61. Boddy JWD, Smalley RJ, Goodman PS, Tate JE, Bell MC, Tomlin AS. The spatial variability in concentrations of a traffic-related pollutant in two street canyons in York, UK—Part II: the influence of traffic characteristics. Atmos Environ 2005b;39:3163–76. Charron A, Harrison RM. Primary particle formation from vehicle emissions during exhaust dilution in the roadside atmosphere. Atmos Environ 2003;37:4109–19. Huang H, Akutsu Y, Arai M, Tamura M. A two-dimensional air quality model in an urban street canyon: evaluation and sensitivity analysis. Atmos Environ 2000;34:689–98. Kerminen V-M, Pirjola L, Kulmala M. How significantly does coagulational scavenging limit atmospheric particle production? J Geophys Res 2001;106:24119–25. Ketzel M, Wåhlin P, Berkowicz R, Palmgren F. Particle and trace gas emission factors under urban driving conditions in Copenhagen based on street and roof-level observations. Atmos Environ 2003;37:2735–49. Kittelson DB. Engines and nanoparticles: a review. J Aerosol Sci 1998;29:575–88. Kittelson DB, Johnson J, Watts W, Wei Q, Drayton M, Paulsen D, et al. Diesel aerosol sampling in the atmosphere. SAE; 2000 [Paper number 2000-01-2212]. Kulmala M, Pirjola L, Makela JM. Stable sulphate clusters as a source of new atmospheric particles. Nature 2000;404:66–9. Longley ID, Gallagher MW, Dorsey JR, Flynn M, Allan JD, Alfarra MR, et al. A case study of aerosol (4.6 nm b Dp b 10 μm) number and mass size distribution measurements in a busy street canyon in Manchester, UK. Atmos Environ 2003;37:1563–71.
Longley ID, Gallagher MW, Dorsey JR, Flynn M, Bower KN, Allan JD. Street canyon aerosol pollutant transport measurements. Sci Total Environ 2004;334–335:327–36. Mönkkönen P, Uma R, Srinivasan D, Koponen IK, Lehtinen KEJ, Hämeri K, et al. Relationship and variations of aerosol number and PM10 mass concentrations in a highly polluted urban environment— New Delhi, India. Atmos Environ 2004;38:425–33. Nicholson SE. A pollution model for street-level air. Atmos Environ 1975;9:19–31. Peters A, Wichmann HE. Epidemiological evidence on the health effects of ultrafine particles. Epidemiology 2001;12:544. Schauer JJ, Rogge WF, Hildemann LM, Mazurek MA, Cass GR, Simoneit BRT. Source apportionment of airborne particulate matter using organic compounds as tracers. Atmos Environ 1996;30: 3837–55. Schneider J, Hock N, Weimer S, Borrmann S. Nucleation particles in diesel exhaust: composition inferred from in-situ mass spectrometric analysis. Environ Sci Technol 2005;39:6153–61. Shi JP, Khan AA, Harrison RM. Measurements of ultrafine particle concentration and size distribution in the urban atmosphere. Sci Total Environ 1999;235:51–64. Väkevä M, Hämeri K, Kulmala M, Lahdes R, Ruuskanen J, Laitinen T. Street level versus rooftop concentrations of submicron aerosol particles and gaseous pollutants in an urban street canyon. Atmos Environ 1999;33(9):1385–97. Wehner B, Birmili W, Gnauk T, Wiedensonhler A. Particle number size distributions in a street canyon and their transformation into the urban-air background: measurements and simple model study. Atmos Environ 2002;36:2215–23. Weingartner E, Nyeki S, Streit N, Baltensperger U. Seasonal variation of size distributions and fine particle formation of continental aerosol at high alpine site. J Aerosol Sci 1998;29:S509–10. Woo KS, Chen DR, Puid YH, Mcmurry PH. Measurement of Atlanta aerosol size distributions: observations of ultrafine particle events. Aerosol Sci Technol 2001;34(1):75–87.