- Email: [email protected]
Temporal distribution and other characteristics of new particle formation events in an urban environment
Environmental Pollution 233 (2018) 552e560
Contents lists available at ScienceDirect
Environmental Pollution journal homepage: www.elsevier.com/locate/envpol
Temporal distribution and other characteristics of new particle formation events in an urban environment* Buddhi Pushpawela, Rohan Jayaratne, Lidia Morawska* International Laboratory for Air Quality and Health, Queensland University of Technology, GPO Box 2434, Brisbane, QLD 4001, Australia
a r t i c l e i n f o
a b s t r a c t
Article history: Received 12 May 2017 Received in revised form 25 October 2017 Accepted 25 October 2017 Available online 5 November 2017
Studying the characteristics of new particle formation (NPF) is important as it is generally recognized as a major contributor to particle pollution in urban environments. We investigated NPF events that occurred during a 1-year period in the urban environment of Brisbane, Australia, using a neutral cluster and air ion spectrometer (NAIS) which is able to monitor both neutral and charged particles and clusters down to a size of 0.8 nm. NPF events occurred on 41% of days, with the occurrence rate of 7% greater in the summer than in the winter. We derived the first diurnal event distribution of NPF events anywhere in the world and showed that the most probable starting time of an NPF event was near 08:30 a.m., being about an hour earlier in the winter than in the summer. During NPF days, 10% of particles were charged. The mean neutral and charged particle concentrations on NPF days were, respectively, 49% and 14% higher than those on non-event days. The mean formation rate of 2e3 nm particles during an NPF event was 20.8 cm3 s1. The formation rate of negatively charged particles was about 10% higher than that of positively charged particles. The mean particle growth rate in the size range up to 20 nm was 6.2 nm h1. These results are compared and contrasted with corresponding values that have been derived with the scanning mobility particle sizer (SMPS) at the same location and with values that have been reported with the NAIS at other locations around the world. This is the first comprehensive study of the characteristics of NPF events over a significantly long period in Australia. © 2017 Elsevier Ltd. All rights reserved.
Keywords: New particle formation Nucleation Atmospheric aerosols Atmospheric ions
1. Introduction Secondary particles are formed by the homogeneous nucleation of gaseous precursors such as sulphuric acid, ammonia and volatile organics, in the atmosphere (Kulmala et al., 2013). This process is also known as new particle formation (NPF) and has been observed in many different environments including urban, industrial, agricultural and coastal sites, as well as in boreal forests and the polar regions (Kulmala et al., 2004). The large majority of these observations were made in the northern hemisphere, mostly in Europe. In the atmosphere, gas molecules cluster together and remain stable until they reach a size of about 1.6 nm. These are known as clusters (Hirsikko et al., 2011). Beyond this size, if the gaseous saturation remains high enough, they go on to form particles. NPF typically occurs around mid-day with high intensity of solar radiation, high concentration of gaseous precursors and low
*
This paper has been recommended for acceptance by Dr. Chen Da. * Corresponding author. E-mail address: [email protected] (L. Morawska).
https://doi.org/10.1016/j.envpol.2017.10.102 0269-7491/© 2017 Elsevier Ltd. All rights reserved.
concentration of pre-existing aerosols (Birmili and Wiedensohler, 2000). The role of ions in this process has also been proposed (Iida et al., 2006; Yu and Turco, 2000). Ions attach to neutral gas molecules in the air to form charged molecular clusters smaller than 1.6 nm in size. These “cluster ions” attach to aerosol particles in the air, producing charged particles. Charged particles in the size range 1.6e7.5 nm are known as intermediate ions and are used as an indicator of particle formation in the atmosphere as their concentration shows a sharp increase during NPF events. There have been several studies of NPF in Brisbane. For example, Guo et al. (2008) reported NPF events on 7 out of 20 days (35%), while Cheung et al. (2011) identified NPF events on 65 out of 252 days (26%). In contrast, Salimi et al. (2017) reported 219 NPF events out of 285 days of measurements at 25 sites across Brisbane. This occurrence rate of 77% is significantly higher than any of the values found previously in Brisbane and at any other location in the world. These three studies were carried out using a scanning mobility particle sizer (SMPS) with a lower size limit of about 10 nm. In the present study, we collected data of charged and uncharged particle concentrations in the urban environment of
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
Brisbane using a Neutral cluster and Air Ion Spectrometer (NAIS) over a one year time period. The NAIS is designed to measure particles down to a size of 0.8 nm and does not have the limitations of the SMPS in identifying NPF events. We have also previously used a NAIS in Brisbane to investigate various characteristics of NPF events. Crilley et al. (2014) monitored particle number concentration (PNC) and the chemical composition of aerosols using a NAIS and a time-of-flight aerosol mass spectrometer (TOF-AMS) and showed that NPF events were observed on 20 out of 36 days (55%). Airborne PNC, charged particles, sulphate and ammonium concentrations were highly correlated on NPF days. Jayaratne et al. (2015b) showed that cluster ion concentrations were suppressed by high PNCs and NPF events were more likely to form when the air was cleaner. Jayaratne et al. (2016b) investigated the charging state of aerosols during NPF events and showed that NPF occurred on 45% of days observed. The coefficient of unipolarity (positive/ negative ion concentration) was 1.37 for cluster ions and 1.17 for charged particles. The positive cluster ion concentration was 40% higher than negative while the positive charged particle concentration was 20% higher than negative. The cluster ion concentration was higher during the night than in the daytime, while the charged particle concentration was higher during the daytime than in the night. There was a positive correlation between PNC and charged particle concentration and a negative correlation between PNC and cluster ion concentration. The percentage of particles carrying a charge was 10e15% during the night and 5e10% during the day. In this study, we derived the characteristics of NPF such as the particle formation rates and growth rates using PNC down to the smallest particle sizes for the first time in Brisbane. In addition, we estimated the short and long term temporal variations of the occurrences of NPF and derived a diurnal distribution plot of NPF events for the very first time anywhere in the world. 2. Methods 2.1. The study area The measurements were conducted at the Garden Point campus of the Queensland University of Technology in Brisbane, Australia, which is a typical urban environment, over one calendar year. Monitoring was carried out through the window of a building at a height of approximately 10 m above the ground, about 100 m away from a busy highway. The dominant anthropogenic sources at this site are motor vehicle exhaust emissions. Other anthropogenic sources affecting the site are emissions from the Brisbane port and two oil refineries located about 20 km to the north-east of the city. 2.2. Instrumentation The NAIS is manufactured by Airel Ltd, Estonia (Manninen et al., 2016). It is designed to detect neutral and charged cluster and particles in the electrical mobility range from 3.2 to 0.0013 cm2 V1 s1. This corresponds to a particle size range from 0.8 to 42 nm. It has a high-resolution time down to 1 s. The instrument cycles between the ion mode, where it measures naturally charged particles, and the particle mode in which it measures both charged and uncharged particles. Two parallel mobility analysers enable both positive and negative particles to be monitored simultaneously. Manninen et al. (2011) and Manninen et al. (2016) have recently pointed out that the NAIS has some difficulty in differentiating between charged particles smaller than 2.0 nm and corona ions generated by the charger in the instrument. Therefore, as recommended by these workers, we have limited the smallest particle that can be detected by the NAIS to 2.0 nm. Since the NAIS does not
553
detect any particles larger than 42 nm, all PNC values quoted in this paper refer to the size range 2e42 nm. A scanning mobility particle sizer (SMPS), including a TSI 3071 differential mobility analyser and a TSI 3782 condensation particle counter, was used to measure the particle size distribution in the range 9e415 nm. The data obtained with the SMPS were used to calculate the condensation sinks. 2.3. Data analysis 2.3.1. Classification of NPF events NPF events were identified using the rate of change of total particle concentration, dN/dt, where N is the number of particles in the size range 2.0e10.0 nm and using the classification scheme developed by Zhang et al. (2004). Events with N > 10,000 cm3 for at least 1 h and dN/dt > 10,000 cm3 h1 were classified as NPF events. These events generally exhibit a “banana” shape in the contour plot of PNC. Days where the above criteria are not fulfilled were classified as “non-event” days. 2.3.2. Starting time of NPF events In the absence of NPF, the PNC in the intermediate size range 2.0e7.5 nm is very small. When an NPF event begins, this PNC shows a sharp increase. This increase may be used to identify the starting time of an NPF event (Leino et al., 2016). In this study, we used the time of first occurrence of dN/dt > 10,000 cm3 h1, where N is the number of particles in the size range 2.0e10.0 nm, as the starting time of an event. 2.3.3. Calculation of condensation sink (CS) and coagulation sink (CoagS) The condensation sink determines the condensation rate of the vapour onto the aerosol particles. This parameter is controlled by the diffusion properties of condensing vapour and the surface area of aerosol particles. The CS of particles is defined as
X CS ¼ 2 p D bm dp;i dp;i Ni
(1)
i
where D is the diffusion coefficient of the condensing vapour and bm is the transition correction factor for mass flux (Dal Maso et al., 2002; Lehtinen et al., 2003; Dal Maso et al., 2005; Salma et al., 2011; Kulmala et al., 2012). In this expression, dp,i and Ni are the diameter and the number concentration of particles in the size bin i, respectively. The unit of CS is per second. In this study, we calculated the CS from the number concentration of particles reported by the SMPS in 107 size bins in the range 9e415 nm and by the NAIS in 7 size bins in the range 2e9 nm. Several studies have shown that sulphuric acid was the main condensing vapour in the urban atmosphere (Stanier et al., 2004; Zhang et al., 2004). Studies in Brisbane have confirmed this result (Crilley et al., 2014; Salimi et al., 2014). We calculated the diffusion coefficient of sulphuric acid using the expression given by Jeong (2009):
D ¼ 5:0032 106 þ 1:04 108 T þ 1:64 1011 T 2 1:566 1014 T 3 (2) where D has the units of square metres per second and temperature T is in Kelvin. Here, we used T ¼ 300 K which is a good estimate of the average daytime temperature in Brisbane. The value of D
554
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
calculated using Equation (2) at T ¼ 300 K was 0.0918 cm2 s-1. The transition correction factor is calculated using the FuchsSutugin expression (Fuchs and Sutugin, 1971)
bm
Kn þ 1 ¼ 1 þ 34a þ 0:337 Kn þ 34a Kn2
GR ¼ (3)
Where
Kn ¼
2l dp
0 a 1:
and
Here Kn, the Knudsen number describes the nature of the suspending vapour relative to particle, l is the mean free path of suspending vapour molecule and dp is the diameter of the particle (Seinfeld and Pandis, 2006). The mass accommodation coefficient (sticking coefficient) a describes the probability of sticking a vapour molecule to a surface of a particle during vapour-particle interaction (Seinfeld and Pandis, 2006). In this study, we assumed a ¼ 1. We used the derived values of CS to calculate the coagulation sink using the expression given by Lehtinen et al. (2007):
CoagSdp ¼ CS:
dp 0:71
(4)
2.3.4. Calculation of particle formation rate (FR) The formation rate is the number of particles formed per cubic centimetre per second at a particular size, usually the size of the smallest particle measured by the instrument, e.g 2e3 nm for the NAIS. The nucleation size is typically referred to thermodynamically stable clusters in the size range 1.0e2.0 nm (Kulmala et al., 2007). Kulmala et al. (2012) has defined the FR of neutral particles as
dNdp dt
þ CoagSdp :Ndp þ
GR
:Ndp
Ddp
(5)
and the FR of charged particles as
J ± dp ¼
dN± dp dt
ddp dp2 dp1 ¼ dt t2 t1
(7)
where dp2 and dp1 are the diameters of particles at times t1 and t2 (Kulmala et al., 2012). We calculated GR using the maximum concentration method described in Kulmala et al. (2012). This method examines the time of maximum particle concentrations during NPF events for different particle sizes. First, we exported the number concentrations of particles obtained by the NAIS in 14 bins in the size range 2.0e42.0 nm. Next, we identified the time of maximum concentration during an NPF event for each particle size range. GR was estimated from the slope of the best-fitted line on the graph of mid-point diameter of particle in each size bin versus the time of maximum concentration. The unit of the GR is nanometre per hour. 3. Results 3.1. Identification of NPF events
m
where the exponent m varies from 1.75 to 1.5 with a mean of 1.7. The value 0.71 is the diameter of a hydrated sulphuric acid molecule. Here we used a value of m ¼ 1.7 (Dal Maso et al., 2008) and l ¼ 50 nm. The latter value is based on the mean free path of an air molecule in the atmosphere (68 nm) provided by Jennings (1988). The unit of CoagS is per second.
Jdp ¼
2.3.5. Calculation of particle growth rate (GR) The GR of particles is defined as
þ CoagSdp :N ± dp þ
GR N ± dp þ a:N ± dp :N H < dp Ddp
c:Ndp :N± dp (6) where Ndp and N±dp are the number concentration of neutral and charged particles in the size range dp and (dp þ Ddp) respectively. The CoagSdp represents the loss of the particles due to coagulation and GR is the growth rate of particles. a is the ion-ion recombination coefficient (1.6 106 cm3 s1) and c is the ion-aerosol attachment coefficient (0.01 106 cm3 s1). Here the superscript þ and e refer to positive and negative charged particles. The subscript < dp refers the size smaller than dp. In this study, to calculate the FR of particles in the size range 2e3 nm, we used Equations (5) and (6) with the number concentrations of particles obtained by the NAIS in 14 bins in the size range 2.0e42.0 nm. The unit of FR is per cubic centimetre per second.
Fig. 1 shows a PNC spectragram obtained by the NAIS on a day with a strong NPF event that began around 7.30 a.m. and lasted for 4e5 h. Growth of particles is clearly indicated by the typical “banana shape” of the concentration contours. The markers indicate the particle size. 3.2. Annual distribution of NPF events During the one-year period of monitoring, NAIS data were available for full 24-h periods on a total of 253 days. On the other days, the instrument was unavailable, as it was required for other projects. A few days were lost due to instrument cleaning or malfunction such as during power failures. The distribution of the observed days and the NPF days over the year is schematically shown in Fig. 2. Out of the 253 days, NPF events were observed on 104 days, which comprised an occurrence rate of 41%. 3.3. Starting time of NPF events The starting times of all 104 NPF events were determined by using the method described in section 2.3.2. This is the time when the particles in the intermediate size range show a sharp increase. In the event shown in Fig. 1, it is clear that the intermediate size PNC begins to increase at approximately 7.30 a.m. We applied this method to all NPF events and found that NPF events occurred mainly during the morning (Fig. 3a). The plot in Fig. 3b is a histogram of the number of events observed in each 30 min period between 6.00 and 12.00 h. There is a high likelihood of occurrence between 8.00 and 10.00 a.m. Of the 104 events, 55 began during this 2-h period. 3.4. Condensation sinks and coagulation sinks The condensation sinks during the NPF events, calculated from Equation (1), assuming the growth to be due to sulphuric acid, varied between 2.6 x 103 s1 and 8.4 103 s1 with a mean and a standard deviation of (5.3 ± 1.7) x 103 s1. The corresponding coagulation sink, calculated from the condensation sink using a particle diameter of 2 nm in Equation (4) varied between 4.5 104 s1 and 14.5 104 s1 with a mean and a standard deviation of (9.1 ± 2.9) x 104 s1.
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
555
Fig. 1. Spectragram of an NPF event showing the typical “banana shape”. The colour contours represent the PNC and the markers represent the times at which the PNC reached its maximum value at each particle size. The unit of PNC is per cubic centimetre. Data below 2.0 nm should be treated with caution due to instrumentation limitations as described in the text. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 2. Summary of all observational days and days with NPF events.
3.5. Formation rate
3.7. Concentrations of particles
The formation rates of charged and neutral particles in the size range 2e3 nm during all NPF events were calculated using the equations in section 2.3.4 and a summary of the results are presented in Table 1.
Fig. 4 shows an overlay of the PNC obtained with the NAIS between 4.00 and 18.00 h on the NPF day that was shown in Fig. 1. The white curve shows the total PNC in the size range 2.0e42 nm. The black curve shows the corresponding values in the size range 2.0e10 nm. During the NPF event, the total PNC increased from about 4.5 104 cm3 at 7.30 h to over 1.9 105 cm3 at 9.30 h. The particle formation rate was 7.2 105 cm3 h1. Thereafter, the particles continued to grow in size for several hours when the PNC decreased to about 3 104 cm3 at 18.00 h. A summary of the maximum number concentrations of charged and neutral particles during NPF events, and the mean concentrations during NPF days and non-event days are shown in Table 3. The concentration values on NPF and non-event days are mean values over each 24 h day. The individual mean values on each of the NPF days are represented by the bars in the graphs in Fig. 5. The mean PNC on NPF days was 49% greater than on non-event days. However, the corresponding increase in charged particle concentration was only about 14%. The percentage of particles that carried a
3.6. Particle growth rate The markers in the NPF event shown in Fig. 1 represent the midpoint diameter of particles in each size bin at each time calculated according to the method described in section 2.3.5. In this event, the particles showed a high growth rate of about 7 nm h1 in the size range 2.0e20 nm, falling to about 4 nm h1 between 20 and 42 nm. Table 2 shows a summary of growth rates of charged and neutral particles of all NPF events. The growth rate was generally highest soon after formation, and then decreased as the particles grew.
556
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
Fig. 3. (a) Summary of starting times of NPF events (b) Histogram of the number of events starting within each 30 min period.
Table 1 Formation rates of neutral particles, positively and negatively charged particles in the size range 2e3 nm during NPF events. Formation Rate (cm3 s1) Particles
Minimum
Maximum
Mean
Standard Deviation
Neutral Positively Charged Negatively Charged
10.0 0.07 0.08
33.8 0.56 0.56
20.8 0.22 0.24
6.7 0.14 0.14
Table 2 Growth rates of charged particles and neutral particles during NPF events. Size Range (nm)
Minimum
Maximum
Mean
Standard Deviation
16.4 8.0
9.7 6.0
5.0 1.4
Charged Particles (nm h1) 2.0e20.0 20.0e42.0
4.0 4.1
Neutral Particles (nm h1) 2.0e20.0 20.0e42.0
4.1 4.1
21.6 8.3
12.1 6.2
6.5 1.5
charge on NPF days was about 10%. In agreement with Jayaratne et al. (2015b), this value decreased significantly during the actual event times.
3.8. Percentage of particles charged Fig. 6 shows the mean percentage of particles charged as a function of particle size during all NPF events over the entire measurement period. The percentages were calculated for each of seven size bins in the detectable particle size range 2.0e42 nm. The results were compared with the curve corresponding to the steady state charge distribution of particles in the bipolar ion atmosphere from Wiedensohler's approximation of Fuchs theory (Wiedensohler, 1988). In charge equilibrium, the percentage of particles charged increase strongly with increasing particle diameter, from about 4% at 5 nm to about 42% at 50 nm. At any given particle size, a value greater than the equilibrium value indicates that the particles are overcharged, while a value smaller than the equilibrium value indicates that the particles are undercharged. Fig. 6 demonstrates that, during NPF events, particles smaller than 5 nm were overcharged while particles larger than 5 nm were undercharged. Similar observations have been reported previously (Jayaratne et al., 2016b) and (Vana et al., 2006).
4. Discussion The NPF occurrence rate of 41% is considerably higher than the
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
557
Fig. 4. Spectragram of the NPF event in Fig. 1, showing the time variation of the PNC in the two different size ranges 2.0e42 nm (white curve) and 2.0e10 nm (black curve).
Table 3 The number concentration of charged and neutral particles during NPF events, NPF days and non-event days. Number Concentration (cm3)
NPF Event (Maximum) NPF Days (Mean) Non-Event Days (Mean)
Charged Particles (x 103)
Neutral Particles (x 104)
% Charged
8.7 ± 2.8 3.2 ± 0.8 2.8 ± 0.6
15.3 ± 7.5 3.0 ± 1.7 1.9 ± 0.4
5 10 12
corresponding values of 35% and 26% found with the SMPS by Guo et al. (2008) and Cheung et al. (2011) respectively, and significantly lower than the value of 77% reported by Salimi et al. (2017), all in Brisbane. These earlier results were derived with an SMPS. The wide range of values may be attributed to the greater efficiency of the NAIS over the SMPS in identifying NPF events as well as to the different criteria used to identify an event. In particular, the particle size range used to determine the value of N used in the term dN/dt is critically important in the criterion. Recent studies using the NAIS by Crilley et al. (2014) and Jayaratne et al. (2016a) have provided much closer occurrence rates of 56% and 45%, respectively. The
Crilley et al. (2014) study was conducted during the summer months, when the occurrence rate was generally higher than at other times of the year, while the Jayaratne et al. (2016a) study covered an entire year. Here too, it should be noted that the criteria used to define an NPF event were slightly different in the two papers. Jayaratne et al. (2016a), as in the present study, stipulated an additional requirement for the minimum time during which N was greater than a certain threshold value. NPF events are not expected to be common in urban environments due to high loading of pre-existing aerosols particles that provide large condensation and coagulation sinks. The observed occurrence rate of NPF event in other urban cities such as Birmingham (Alam et al., 2003), Paris (Dos Santos et al., 2015), Budapest (Salma et al., 2011), and Helsinki (Hussein et al., 2008) varied between 5 and 27%. However, in some other cities the observed occurrence rate of NPF events was higher. For example, in St Louis (Qian et al., 2007), Pittsburgh (Stanier et al., 2004), and Beijing (Yu et al., 2015) with 36%, 50% and 47% respectively. It should be noted that the population densities in these three cities exceed 1000 km2 which is much higher than the equivalent figure of 145 km2 in Brisbane. The higher the population, the greater will be the number of motor vehicles and other sources of precursor
Fig. 5. Mean number concentration of particles on all NPF days. The charged particles are shown in red and the neutral particles are in blue. The mean values of the charged and total number concentrations on NPF and non-event days over the entire period of observation are indicated on the right margin. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
558
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
Fig. 6. Mean number percentage of particles carrying a charge during NPF events as a function of their size. The line shows the percentage of particles charged at equilibrium according to the Wiedensohler model.
gases, which may explain the higher rates of NPF occurrence. The results presented in Fig. 2 need to be considered with some care as they are based on just one calendar year, where the number of observational days varied from month to month. January and March are particularly uncertain as they consisted of less than 10 observational days each. Despite this limitation, we can derive several important characteristics. We note that the minimum percentage of NPF events occurred in December, a month during which there were several days with smoke from back-burning activities in the northern suburbs of Brisbane. NPF events were suppressed during the course of this month due to the prevalent high condensation sink (Jayaratne et al., 2015b). In spite of this suppression, the probability of occurrence of NPF was highest during the summer (43.8% in NovembereFebruary) and lowest during the winter (32.1% in MayeAugust). Some caution is needed in comparing these values with other locations because the difference in mean temperature in Brisbane between summer and winter is not as great as in more temperate climates. Manninen et al. (2010) analysed NPF events in 12 European cities and found that the highest occurrence rate was in late spring and the lowest in winter. Salma et al. (2011) found the highest in the spring (44%) and the lowest in the winter (7%) in Budapest, Hungary. Qian et al. (2007) found a marked difference between the summer (36%) and winter (8%) in St Louis, USA. In Fig. 3, the most probable starting times of NPF (8e9 am) coincides very well with the morning rush hour traffic. Although the PNC is relatively high at this time, the concentration of condensable vapours produced by motor vehicles is also at a maximum. Thus, despite the higher CS, NPF events are still observed to occur. In fact, we have previously shown that while NPF events are not observed in Brisbane when the PM10 in the atmosphere is higher than 20 mg m3 (Jayaratne et al., 2015a) in more polluted Beijing they are observed at PM10 concentrations as high as 43 mg m3 (Jayaratne et al., 2017). Studies have shown that the main precursor for NPF in urban environments is sulphuric acid from industry and motor vehicle emissions (Stanier et al., 2004; Zhang et al., 2004; Crilley et al., 2014). We also found that the NPF starting times during the winter was about 2 h earlier than during the summer. This is surprising as the sun rises up to 2 h earlier in the summer and the air temperature and solar irradiance are significantly higher than during winter at any given time in the morning. A possible explanation is that the supersaturation of the precursor gases such as
sulphuric acid is considerably higher in the early morning winter hours when the temperature is lower (Zhang et al., 2004). The value of the CS in Brisbane, (5.3 ± 1.7) x 103 s1, is considerably lower than the values reported at some other urban cities such as Paris (14.3 103 s1) (Dos Santos et al., 2015) and Budapest (16 103 s1) (Salma et al., 2011). This is to be expected as the air quality in Brisbane is relatively better than in most other cities of similar or larger size. The coagulation sink for 2 nm particles determined here was (9.1 ± 2.9) x 104 s1. This parameter has not been determined for Brisbane before. We may compare it with the value of 9.9 104 s1 for 3 nm particles found in Beijing where the condensation sink is relatively much higher than in Brisbane (Wu et al., 2011). The mean formation rate of neutral particles during NPF events, calculated using the PNC in the range 2e3 nm, was 20.8 cm3 s1. This is comparable to the value of 17 cm3 s1 in St Louis, USA, determined by Qian et al. (2007) and within the range 0.7e32.4 cm3 s1 reported at 12 European cities by Manninen et al. (2010). Our results also showed that the formation rate of negatively charged particles was about 10% higher than that of positively charged particles. This is also in agreement with Manninen et al. (2010) in European cities. The growth rates of neutral particles found were 12.1 and 6.2 nm h1 in the size ranges 2.0e20 and 20e42 nm, respectively. There was a clear decrease in growth rates with increasing particle size. This explains why data obtained with an SMPS with a higher detection size threshold, usually provide lower growth rates than that derived by the NAIS. For example, the average value of growth rate obtained by Cheung et al. (2011) in Brisbane using an SMPS was 4.6 nm h1, which is significantly lower than the values obtained with the NAIS. The growth rates determined in our study may be compared with those from two other urban locations: 3e20 nm h1 in Atlanta, USA (Stolzenburg et al., 2005) and 2e13 nm h1 in Budapest, Hungary. In the present study, there was a significant difference between the growth rates of neutral and charged particles in the size range 2.0e20 nm but not so in the higher size range 20e42 nm (Table 2). In most NPF events, the charged particles formed a few tens of minutes before the neutral particles and suggested a possible role of ion-induced nucleation as previously proposed by Jayaratne et al. (2016a). The growth rates of the two types of particles attained
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
similar values after about 1.5 h. 5. Summary and conclusions Particles have not been monitored at their very smallest sizes in Brisbane until now. This study, conducted with a NAIS over a period of 1 year, has yielded more precise characteristics of NPF than has been possible before. It is also the first comprehensive study of the characteristics of NPF events in Australia. A summary of the findings of the study are listed below. NPF events were observed on 41% of the observational days with a higher occurrence rate in the summer (November to February) than in the winter (May to August). The mean PNC on NPF days was 49% higher than those of non NPF days. Events occurred mainly during the mornings, with the highest probability of occurrence between 8.00 and 10.00 a.m. The mean condensation and coagulation sinks during the events were estimated to be 5.3 103 s1 and 9.1 104 s1, respectively. There was a significant difference between the growth rates of neutral and charged particles in the size range 2.0e20 nm but not so in the higher size range 20e42 nm. The mean formation rates of 2e3 nm particles during NPF event was 20.8 cm3 s1. The formation rate of negatively charged particles was about 10% higher than the positively charged particles, in the size range 2e3 nm. The mean charged particle concentration on NPF days was 14% higher than those of non NPF days. On NPF days, 10% of particles were charged. Particles smaller than 5 nm were found to be overcharged in comparison with the steady state during NPF events. Particles larger than 5 nm were found to be undercharged in comparison with the steady state during NPF events. These results, based on data obtained at the very sizes where particles formed, have enhanced our knowledge of NPF events in urban environments. Some results, such as the diurnal variation of the occurrence of NPF events and the growth rates of charged particles are reported for the first time. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Acknowledgements We thank Dr. Xuan Ling for his assistance with the data acquisition. References Alam, A., Shi, J.P., Harrison, R.M., 2003. Observations of new particle formation in urban air. J. Geophys. Res. Atmos. 108. Birmili, W., Wiedensohler, A., 2000. New particle formation in the continental boundary layer: meteorological and gas phase parameter influence. Geophys. Res. Lett. 27, 3325e3328. Cheung, H., Morawska, L., Ristovski, Z., 2011. Observation of new particle formation in subtropical urban environment. Atmos. Chem. Phys. 11, 3823. Crilley, L.R., Jayaratne, E.R., Ayoko, G.A., Miljevic, B., Ristovski, Z., Morawska, L., 2014. Observations on the formation, growth and chemical composition of aerosols in an urban environment. Environ. Sci. Technol. 48, 6588e6596. €, J., Aalto, P., O'Dowd, C., 2002. Dal Maso, M., Kulmala, M., Lehtinen, K.E., M€ akela Condensation and coagulation sinks and formation of nucleation mode particles in coastal and boreal forest boundary layers. J. Geophys. Res. Atmos. 107. Dal Maso, M., Kulmala, M., Riipinen, I., Wagner, R., Hussein, T., Aalto, P.P., Lehtinen, K.E., 2005. Formation and growth of fresh atmospheric aerosols: eight years of aerosol size distribution data from SMEAR II, Hyytiala, Finland. Boreal Environ. Res. 10, 323. €rinen, A., Komppula, M., Tunved, P., KERMINEN, V., Dal Maso, M., Hyva Lihavainen, H., Viisanen, Y., HANSSON, H.C., Kulmala, M., 2008. Annual and interannual variation in boreal forest aerosol particle number and volume concentration and their connection to particle formation. Tellus B 60, 495e508. Dos Santos, V., Herrmann, E., Manninen, H., Hussein, T., Hakala, J., Nieminen, T.,
559
Aalto, P., Merkel, M., Wiedensohler, A., Kulmala, M., 2015. Variability of air ion concentrations in urban Paris. Atmos. Chem. Phys. 15, 13717e13737. Fuchs, N.A., Sutugin, A.G., 1971. High-dispersed Aerosols, Topic in Current Aerosol Research. Pergamon, NY. Guo, H., Ding, A., Morawska, L., He, C., Ayoko, G., Li, Y.S., Hung, W.T., 2008. Size distribution and new particle formation in subtropical eastern Australia. Environ. Chem. 5, 382e390. , S., Lehtipalo, K., Manninen, H.E., Ehn, M., Hirsikko, A., Nieminen, T., Gagne ~rrak, U., Kerminen, V.M., Laakso, L., McMurry, P.H., Mirme, A., Mirme, S., Ho Pet€ aj€ a, T., Tammet, H., Vakkari, V., Vana, M., Kulmala, M., 2011. Atmospheric ions and nucleation: a review of observations. Atmos. Chem. Phys. 11, 767e798. https://doi.org/10.5194/acp-11-767-2011. Hussein, T., Martikainen, J., Junninen, H., Sogacheva, L., Wagner, R., Dal Maso, M., Riipinen, I., Aalto, P.P., Kulmala, M., 2008. Observation of regional new particle formation in the urban atmosphere. Tellus B 60, 509e521. Iida, K., Stolzenburg, M., McMurry, P., Dunn, M.J., Smith, J.N., Eisele, F., Keady, P., 2006. Contribution of ion-induced nucleation to new particle formation: methodology and its application to atmospheric observations in Boulder, Colorado. J. Geophys. Res. Atmos. 111. Jayaratne, E.R., Clifford, S., Morawska, L., 2015a. Atmospheric visibility and PM10 as indicators of new particle formation in an urban environment. Environ. Sci. Technol. 49, 12751e12757. https://doi.org/10.1021/acs.est.5b01851. Jayaratne, E.R., Pushpawela, B., Morawska, L., 2016a. Temporal evolution of charged and neutral nanoparticle concentrations during atmospheric new particle formation events and its implications for ion-induced nucleation. Front. Environ. Sci. Eng. 10, 13. Jayaratne, R., Ling, X., Morawska, L., 2016b. Charging state of aerosols during particle formation events in an urban environment and its implications for ion-induced nucleation. Aerosol Air Qual. Res. 16, 348e360. Jayaratne, R., Pushpawela, B., He, C., Li, H., Gao, J., Chai, F., Morawska, L., 2017. Observations of particles at their formation sizes in Beijing, China. Atmos. Chem. Phys. 17, 8825e8835. https://doi.org/10.5194/acp-17-8825-2017. Jayaratne, R.E., Ling, X., Morawska, L., 2015b. Suppression of cluster ions during rapidly increasing particle number concentration events in the environment. Aerosol Air Qual. Res. 15, 28e37. Jennings, S.G., 1988. The mean free path in air. J. Aerosol Sci. 19, 159e166. https:// doi.org/10.1016/0021-8502(88)90219-4. Jeong, K., 2009. Condensation of Water Vapor and Sulfuric Acid in Boiler Flue Gas. ProQuest. €j€ Kulmala, M., Vehkam€ aki, H., Peta a, T., Dal Maso, M., Lauri, A., Kerminen, V.-M., Birmili, W., McMurry, P.H., 2004. Formation and growth rates of ultrafine atmospheric particles: a review of observations. J. Aerosol Sci. 35, 143e176. Kulmala, M., Riipinen, I., Sipil€ a, M., Manninen, H.E., Pet€ aj€ a, T., Junninen, H., Dal Maso, M., Mordas, G., Mirme, A., Vana, M., 2007. Toward direct measurement of atmospheric nucleation. Science 318, 89e92. €ja €, T., Nieminen, T., Sipil€ Kulmala, M., Peta a, M., Manninen, H.E., Lehtipalo, K., Dal Maso, M., Aalto, P.P., Junninen, H., Paasonen, P., 2012. Measurement of the nucleation of atmospheric aerosol particles. Nat. Protoc. 7, 1651e1667. Kulmala, M., Kontkanen, J., Junninen, H., Lehtipalo, K., Manninen, H.E., Nieminen, T., €, M., Schobesberger, S., Rantala, P., 2013. Direct observations of Pet€ aj€ a, T., Sipila atmospheric aerosol nucleation. Science 339, 943e946. Lehtinen, K.E., Korhonen, H., Maso, M., Kulmala, M., 2003. On the concept of condensation sink diameter. Boreal Environ. Res. 8, 405e412. Lehtinen, K.E., Dal Maso, M., Kulmala, M., Kerminen, V.-M., 2007. Estimating nucleation rates from apparent particle formation rates and vice versa: revised formulation of the KermineneKulmala equation. J. Aerosol Sci. 38, 988e994. €ja €, T., Kerminen, V.-M., Kulmala, M., Leino, K., Nieminen, T., Manninen, H.E., Peta 2016. Intermediate ions as a strong indicator of new particle formation bursts in a boreal forest. Boreal Environ. Res. 21 (3-4), 274e286. , S., H€ Manninen, H., Nieminen, T., Asmi, E., Gagne akkinen, S., Lehtipalo, K., Aalto, P., Vana, M., Mirme, A., Mirme, S., 2010. EUCAARI ion spectrometer measurements at 12 European siteseanalysis of new particle formation events. Atmos. Chem. Phys. 10, 7907e7927. Manninen, H., Franchin, A., Schobesberger, S., Hirsikko, A., Hakala, J., Skromulis, A., Kangasluoma, J., Ehn, M., Junninen, H., Mirme, A., 2011. Characterisation of corona-generated ions used in a neutral cluster and air ion spectrometer (NAIS). Atmos. Meas. Tech. 4, 2767. Manninen, H.E., Mirme, S., Mirme, A., Pet€ aj€ a, T., Kulmala, M., 2016. How to reliably detect molecular clusters and nucleation mode particles with Neutral cluster and Air Ion Spectrometer (NAIS). Atmos. Meas. Tech. Discuss. 9 (8), 3577e3605. https://doi.org/10.5194/amt-9-3577-2016. Qian, S., Sakurai, H., McMurry, P.H., 2007. Characteristics of regional nucleation events in urban East St. Louis. Atmos. Environ. 41, 4119e4127. Salimi, F., Ristovski, Z., Mazaheri, M., Laiman, R., Crilley, L.R., He, C., Clifford, S., Morawska, L., 2014. Assessment and application of clustering techniques to atmospheric particle number size distribution for the purpose of source apportionment. Atmos. Chem. Phys. 14, 11883e11892. Salimi, F., Rahman, M., Clifford, S., Ristovski, Z., Morawska, L., 2017. Nocturnal new particle formation events in urban environments. Atmos. Chem. Phys. 17, 521e530. s, T., Weidinger, T., Aalto, P., Hussein, T., Dal Maso, M., Kulmala, M., Salma, I., Borso 2011. Production, growth and properties of ultrafine atmospheric aerosol particles in an urban environment. Atmos. Chem. Phys. 11, 1339. Seinfeld, J.H., Pandis, S.N., 2006. Atmospheric Chemistry and Physics. Wiley, Hoboken, NJ.
560
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
Stanier, C.O., Khlystov, A.Y., Pandis, S.N., 2004. Nucleation events during the Pittsburgh Air Quality Study: description and relation to key meteorological, gas phase, and aerosol parameters special issue of aerosol science and technology on findings from the fine particulate matter supersites program. Aerosol Sci. Technol. 38, 253e264. Stolzenburg, M.R., McMurry, P.H., Sakurai, H., Smith, J.N., Mauldin, R.L., Eisele, F.L., Clement, C.F., 2005. Growth rates of freshly nucleated atmospheric particles in Atlanta. J. Geophys. Res. Atmos. 110. Vana, M., Tamm, E., Horrak, U., Mirme, A., Tammet, H., Laakso, L., Aalto, P., Kulmala, M., 2006. Charging state of atmospheric nanoparticles during the nucleation burst events. Atmos. Res. 82, 536e546. Wiedensohler, A., 1988. An approximation of the bipolar charge distribution for particles in the submicron size range. J. Aerosol Sci. 19, 387e389.
Wu, Z., Hu, M., Yue, D., Wehner, B., Wiedensohler, A., 2011. Evolution of particle number size distribution in an urban atmosphere during episodes of heavy pollution and new particle formation. Sci. China Earth Sci. 54, 1772. Yu, F., Turco, R.P., 2000. Ultrafine aerosol formation via ion-mediated nucleation. Geophys. Res. Lett. 27, 883e886. Yu, H., Zhou, L., Dai, L., Shen, W., Zheng, J., Ma, Y., Chen, M., 2015. Nucleation and growth of sub-3 nm particles in the polluted urban atmosphere of a megacity in China. Atmos. Chem. Phys. Discuss. 15. Zhang, Q., Stanier, C., Canagaratna, M., Jayne, J., Worsnop, D., Pandis, S., Jiminez, J., 2004. Insights into the chemistry of new particle formation and growth events in Pittsburgh based on aerosol mass spectrometry. Environ. Sci. Technol. 38, 4797e4809.
Contents lists available at ScienceDirect
Environmental Pollution journal homepage: www.elsevier.com/locate/envpol
Temporal distribution and other characteristics of new particle formation events in an urban environment* Buddhi Pushpawela, Rohan Jayaratne, Lidia Morawska* International Laboratory for Air Quality and Health, Queensland University of Technology, GPO Box 2434, Brisbane, QLD 4001, Australia
a r t i c l e i n f o
a b s t r a c t
Article history: Received 12 May 2017 Received in revised form 25 October 2017 Accepted 25 October 2017 Available online 5 November 2017
Studying the characteristics of new particle formation (NPF) is important as it is generally recognized as a major contributor to particle pollution in urban environments. We investigated NPF events that occurred during a 1-year period in the urban environment of Brisbane, Australia, using a neutral cluster and air ion spectrometer (NAIS) which is able to monitor both neutral and charged particles and clusters down to a size of 0.8 nm. NPF events occurred on 41% of days, with the occurrence rate of 7% greater in the summer than in the winter. We derived the first diurnal event distribution of NPF events anywhere in the world and showed that the most probable starting time of an NPF event was near 08:30 a.m., being about an hour earlier in the winter than in the summer. During NPF days, 10% of particles were charged. The mean neutral and charged particle concentrations on NPF days were, respectively, 49% and 14% higher than those on non-event days. The mean formation rate of 2e3 nm particles during an NPF event was 20.8 cm3 s1. The formation rate of negatively charged particles was about 10% higher than that of positively charged particles. The mean particle growth rate in the size range up to 20 nm was 6.2 nm h1. These results are compared and contrasted with corresponding values that have been derived with the scanning mobility particle sizer (SMPS) at the same location and with values that have been reported with the NAIS at other locations around the world. This is the first comprehensive study of the characteristics of NPF events over a significantly long period in Australia. © 2017 Elsevier Ltd. All rights reserved.
Keywords: New particle formation Nucleation Atmospheric aerosols Atmospheric ions
1. Introduction Secondary particles are formed by the homogeneous nucleation of gaseous precursors such as sulphuric acid, ammonia and volatile organics, in the atmosphere (Kulmala et al., 2013). This process is also known as new particle formation (NPF) and has been observed in many different environments including urban, industrial, agricultural and coastal sites, as well as in boreal forests and the polar regions (Kulmala et al., 2004). The large majority of these observations were made in the northern hemisphere, mostly in Europe. In the atmosphere, gas molecules cluster together and remain stable until they reach a size of about 1.6 nm. These are known as clusters (Hirsikko et al., 2011). Beyond this size, if the gaseous saturation remains high enough, they go on to form particles. NPF typically occurs around mid-day with high intensity of solar radiation, high concentration of gaseous precursors and low
*
This paper has been recommended for acceptance by Dr. Chen Da. * Corresponding author. E-mail address: [email protected] (L. Morawska).
https://doi.org/10.1016/j.envpol.2017.10.102 0269-7491/© 2017 Elsevier Ltd. All rights reserved.
concentration of pre-existing aerosols (Birmili and Wiedensohler, 2000). The role of ions in this process has also been proposed (Iida et al., 2006; Yu and Turco, 2000). Ions attach to neutral gas molecules in the air to form charged molecular clusters smaller than 1.6 nm in size. These “cluster ions” attach to aerosol particles in the air, producing charged particles. Charged particles in the size range 1.6e7.5 nm are known as intermediate ions and are used as an indicator of particle formation in the atmosphere as their concentration shows a sharp increase during NPF events. There have been several studies of NPF in Brisbane. For example, Guo et al. (2008) reported NPF events on 7 out of 20 days (35%), while Cheung et al. (2011) identified NPF events on 65 out of 252 days (26%). In contrast, Salimi et al. (2017) reported 219 NPF events out of 285 days of measurements at 25 sites across Brisbane. This occurrence rate of 77% is significantly higher than any of the values found previously in Brisbane and at any other location in the world. These three studies were carried out using a scanning mobility particle sizer (SMPS) with a lower size limit of about 10 nm. In the present study, we collected data of charged and uncharged particle concentrations in the urban environment of
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
Brisbane using a Neutral cluster and Air Ion Spectrometer (NAIS) over a one year time period. The NAIS is designed to measure particles down to a size of 0.8 nm and does not have the limitations of the SMPS in identifying NPF events. We have also previously used a NAIS in Brisbane to investigate various characteristics of NPF events. Crilley et al. (2014) monitored particle number concentration (PNC) and the chemical composition of aerosols using a NAIS and a time-of-flight aerosol mass spectrometer (TOF-AMS) and showed that NPF events were observed on 20 out of 36 days (55%). Airborne PNC, charged particles, sulphate and ammonium concentrations were highly correlated on NPF days. Jayaratne et al. (2015b) showed that cluster ion concentrations were suppressed by high PNCs and NPF events were more likely to form when the air was cleaner. Jayaratne et al. (2016b) investigated the charging state of aerosols during NPF events and showed that NPF occurred on 45% of days observed. The coefficient of unipolarity (positive/ negative ion concentration) was 1.37 for cluster ions and 1.17 for charged particles. The positive cluster ion concentration was 40% higher than negative while the positive charged particle concentration was 20% higher than negative. The cluster ion concentration was higher during the night than in the daytime, while the charged particle concentration was higher during the daytime than in the night. There was a positive correlation between PNC and charged particle concentration and a negative correlation between PNC and cluster ion concentration. The percentage of particles carrying a charge was 10e15% during the night and 5e10% during the day. In this study, we derived the characteristics of NPF such as the particle formation rates and growth rates using PNC down to the smallest particle sizes for the first time in Brisbane. In addition, we estimated the short and long term temporal variations of the occurrences of NPF and derived a diurnal distribution plot of NPF events for the very first time anywhere in the world. 2. Methods 2.1. The study area The measurements were conducted at the Garden Point campus of the Queensland University of Technology in Brisbane, Australia, which is a typical urban environment, over one calendar year. Monitoring was carried out through the window of a building at a height of approximately 10 m above the ground, about 100 m away from a busy highway. The dominant anthropogenic sources at this site are motor vehicle exhaust emissions. Other anthropogenic sources affecting the site are emissions from the Brisbane port and two oil refineries located about 20 km to the north-east of the city. 2.2. Instrumentation The NAIS is manufactured by Airel Ltd, Estonia (Manninen et al., 2016). It is designed to detect neutral and charged cluster and particles in the electrical mobility range from 3.2 to 0.0013 cm2 V1 s1. This corresponds to a particle size range from 0.8 to 42 nm. It has a high-resolution time down to 1 s. The instrument cycles between the ion mode, where it measures naturally charged particles, and the particle mode in which it measures both charged and uncharged particles. Two parallel mobility analysers enable both positive and negative particles to be monitored simultaneously. Manninen et al. (2011) and Manninen et al. (2016) have recently pointed out that the NAIS has some difficulty in differentiating between charged particles smaller than 2.0 nm and corona ions generated by the charger in the instrument. Therefore, as recommended by these workers, we have limited the smallest particle that can be detected by the NAIS to 2.0 nm. Since the NAIS does not
553
detect any particles larger than 42 nm, all PNC values quoted in this paper refer to the size range 2e42 nm. A scanning mobility particle sizer (SMPS), including a TSI 3071 differential mobility analyser and a TSI 3782 condensation particle counter, was used to measure the particle size distribution in the range 9e415 nm. The data obtained with the SMPS were used to calculate the condensation sinks. 2.3. Data analysis 2.3.1. Classification of NPF events NPF events were identified using the rate of change of total particle concentration, dN/dt, where N is the number of particles in the size range 2.0e10.0 nm and using the classification scheme developed by Zhang et al. (2004). Events with N > 10,000 cm3 for at least 1 h and dN/dt > 10,000 cm3 h1 were classified as NPF events. These events generally exhibit a “banana” shape in the contour plot of PNC. Days where the above criteria are not fulfilled were classified as “non-event” days. 2.3.2. Starting time of NPF events In the absence of NPF, the PNC in the intermediate size range 2.0e7.5 nm is very small. When an NPF event begins, this PNC shows a sharp increase. This increase may be used to identify the starting time of an NPF event (Leino et al., 2016). In this study, we used the time of first occurrence of dN/dt > 10,000 cm3 h1, where N is the number of particles in the size range 2.0e10.0 nm, as the starting time of an event. 2.3.3. Calculation of condensation sink (CS) and coagulation sink (CoagS) The condensation sink determines the condensation rate of the vapour onto the aerosol particles. This parameter is controlled by the diffusion properties of condensing vapour and the surface area of aerosol particles. The CS of particles is defined as
X CS ¼ 2 p D bm dp;i dp;i Ni
(1)
i
where D is the diffusion coefficient of the condensing vapour and bm is the transition correction factor for mass flux (Dal Maso et al., 2002; Lehtinen et al., 2003; Dal Maso et al., 2005; Salma et al., 2011; Kulmala et al., 2012). In this expression, dp,i and Ni are the diameter and the number concentration of particles in the size bin i, respectively. The unit of CS is per second. In this study, we calculated the CS from the number concentration of particles reported by the SMPS in 107 size bins in the range 9e415 nm and by the NAIS in 7 size bins in the range 2e9 nm. Several studies have shown that sulphuric acid was the main condensing vapour in the urban atmosphere (Stanier et al., 2004; Zhang et al., 2004). Studies in Brisbane have confirmed this result (Crilley et al., 2014; Salimi et al., 2014). We calculated the diffusion coefficient of sulphuric acid using the expression given by Jeong (2009):
D ¼ 5:0032 106 þ 1:04 108 T þ 1:64 1011 T 2 1:566 1014 T 3 (2) where D has the units of square metres per second and temperature T is in Kelvin. Here, we used T ¼ 300 K which is a good estimate of the average daytime temperature in Brisbane. The value of D
554
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
calculated using Equation (2) at T ¼ 300 K was 0.0918 cm2 s-1. The transition correction factor is calculated using the FuchsSutugin expression (Fuchs and Sutugin, 1971)
bm
Kn þ 1 ¼ 1 þ 34a þ 0:337 Kn þ 34a Kn2
GR ¼ (3)
Where
Kn ¼
2l dp
0 a 1:
and
Here Kn, the Knudsen number describes the nature of the suspending vapour relative to particle, l is the mean free path of suspending vapour molecule and dp is the diameter of the particle (Seinfeld and Pandis, 2006). The mass accommodation coefficient (sticking coefficient) a describes the probability of sticking a vapour molecule to a surface of a particle during vapour-particle interaction (Seinfeld and Pandis, 2006). In this study, we assumed a ¼ 1. We used the derived values of CS to calculate the coagulation sink using the expression given by Lehtinen et al. (2007):
CoagSdp ¼ CS:
dp 0:71
(4)
2.3.4. Calculation of particle formation rate (FR) The formation rate is the number of particles formed per cubic centimetre per second at a particular size, usually the size of the smallest particle measured by the instrument, e.g 2e3 nm for the NAIS. The nucleation size is typically referred to thermodynamically stable clusters in the size range 1.0e2.0 nm (Kulmala et al., 2007). Kulmala et al. (2012) has defined the FR of neutral particles as
dNdp dt
þ CoagSdp :Ndp þ
GR
:Ndp
Ddp
(5)
and the FR of charged particles as
J ± dp ¼
dN± dp dt
ddp dp2 dp1 ¼ dt t2 t1
(7)
where dp2 and dp1 are the diameters of particles at times t1 and t2 (Kulmala et al., 2012). We calculated GR using the maximum concentration method described in Kulmala et al. (2012). This method examines the time of maximum particle concentrations during NPF events for different particle sizes. First, we exported the number concentrations of particles obtained by the NAIS in 14 bins in the size range 2.0e42.0 nm. Next, we identified the time of maximum concentration during an NPF event for each particle size range. GR was estimated from the slope of the best-fitted line on the graph of mid-point diameter of particle in each size bin versus the time of maximum concentration. The unit of the GR is nanometre per hour. 3. Results 3.1. Identification of NPF events
m
where the exponent m varies from 1.75 to 1.5 with a mean of 1.7. The value 0.71 is the diameter of a hydrated sulphuric acid molecule. Here we used a value of m ¼ 1.7 (Dal Maso et al., 2008) and l ¼ 50 nm. The latter value is based on the mean free path of an air molecule in the atmosphere (68 nm) provided by Jennings (1988). The unit of CoagS is per second.
Jdp ¼
2.3.5. Calculation of particle growth rate (GR) The GR of particles is defined as
þ CoagSdp :N ± dp þ
GR N ± dp þ a:N ± dp :N H < dp Ddp
c:Ndp :N± dp (6) where Ndp and N±dp are the number concentration of neutral and charged particles in the size range dp and (dp þ Ddp) respectively. The CoagSdp represents the loss of the particles due to coagulation and GR is the growth rate of particles. a is the ion-ion recombination coefficient (1.6 106 cm3 s1) and c is the ion-aerosol attachment coefficient (0.01 106 cm3 s1). Here the superscript þ and e refer to positive and negative charged particles. The subscript < dp refers the size smaller than dp. In this study, to calculate the FR of particles in the size range 2e3 nm, we used Equations (5) and (6) with the number concentrations of particles obtained by the NAIS in 14 bins in the size range 2.0e42.0 nm. The unit of FR is per cubic centimetre per second.
Fig. 1 shows a PNC spectragram obtained by the NAIS on a day with a strong NPF event that began around 7.30 a.m. and lasted for 4e5 h. Growth of particles is clearly indicated by the typical “banana shape” of the concentration contours. The markers indicate the particle size. 3.2. Annual distribution of NPF events During the one-year period of monitoring, NAIS data were available for full 24-h periods on a total of 253 days. On the other days, the instrument was unavailable, as it was required for other projects. A few days were lost due to instrument cleaning or malfunction such as during power failures. The distribution of the observed days and the NPF days over the year is schematically shown in Fig. 2. Out of the 253 days, NPF events were observed on 104 days, which comprised an occurrence rate of 41%. 3.3. Starting time of NPF events The starting times of all 104 NPF events were determined by using the method described in section 2.3.2. This is the time when the particles in the intermediate size range show a sharp increase. In the event shown in Fig. 1, it is clear that the intermediate size PNC begins to increase at approximately 7.30 a.m. We applied this method to all NPF events and found that NPF events occurred mainly during the morning (Fig. 3a). The plot in Fig. 3b is a histogram of the number of events observed in each 30 min period between 6.00 and 12.00 h. There is a high likelihood of occurrence between 8.00 and 10.00 a.m. Of the 104 events, 55 began during this 2-h period. 3.4. Condensation sinks and coagulation sinks The condensation sinks during the NPF events, calculated from Equation (1), assuming the growth to be due to sulphuric acid, varied between 2.6 x 103 s1 and 8.4 103 s1 with a mean and a standard deviation of (5.3 ± 1.7) x 103 s1. The corresponding coagulation sink, calculated from the condensation sink using a particle diameter of 2 nm in Equation (4) varied between 4.5 104 s1 and 14.5 104 s1 with a mean and a standard deviation of (9.1 ± 2.9) x 104 s1.
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
555
Fig. 1. Spectragram of an NPF event showing the typical “banana shape”. The colour contours represent the PNC and the markers represent the times at which the PNC reached its maximum value at each particle size. The unit of PNC is per cubic centimetre. Data below 2.0 nm should be treated with caution due to instrumentation limitations as described in the text. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 2. Summary of all observational days and days with NPF events.
3.5. Formation rate
3.7. Concentrations of particles
The formation rates of charged and neutral particles in the size range 2e3 nm during all NPF events were calculated using the equations in section 2.3.4 and a summary of the results are presented in Table 1.
Fig. 4 shows an overlay of the PNC obtained with the NAIS between 4.00 and 18.00 h on the NPF day that was shown in Fig. 1. The white curve shows the total PNC in the size range 2.0e42 nm. The black curve shows the corresponding values in the size range 2.0e10 nm. During the NPF event, the total PNC increased from about 4.5 104 cm3 at 7.30 h to over 1.9 105 cm3 at 9.30 h. The particle formation rate was 7.2 105 cm3 h1. Thereafter, the particles continued to grow in size for several hours when the PNC decreased to about 3 104 cm3 at 18.00 h. A summary of the maximum number concentrations of charged and neutral particles during NPF events, and the mean concentrations during NPF days and non-event days are shown in Table 3. The concentration values on NPF and non-event days are mean values over each 24 h day. The individual mean values on each of the NPF days are represented by the bars in the graphs in Fig. 5. The mean PNC on NPF days was 49% greater than on non-event days. However, the corresponding increase in charged particle concentration was only about 14%. The percentage of particles that carried a
3.6. Particle growth rate The markers in the NPF event shown in Fig. 1 represent the midpoint diameter of particles in each size bin at each time calculated according to the method described in section 2.3.5. In this event, the particles showed a high growth rate of about 7 nm h1 in the size range 2.0e20 nm, falling to about 4 nm h1 between 20 and 42 nm. Table 2 shows a summary of growth rates of charged and neutral particles of all NPF events. The growth rate was generally highest soon after formation, and then decreased as the particles grew.
556
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
Fig. 3. (a) Summary of starting times of NPF events (b) Histogram of the number of events starting within each 30 min period.
Table 1 Formation rates of neutral particles, positively and negatively charged particles in the size range 2e3 nm during NPF events. Formation Rate (cm3 s1) Particles
Minimum
Maximum
Mean
Standard Deviation
Neutral Positively Charged Negatively Charged
10.0 0.07 0.08
33.8 0.56 0.56
20.8 0.22 0.24
6.7 0.14 0.14
Table 2 Growth rates of charged particles and neutral particles during NPF events. Size Range (nm)
Minimum
Maximum
Mean
Standard Deviation
16.4 8.0
9.7 6.0
5.0 1.4
Charged Particles (nm h1) 2.0e20.0 20.0e42.0
4.0 4.1
Neutral Particles (nm h1) 2.0e20.0 20.0e42.0
4.1 4.1
21.6 8.3
12.1 6.2
6.5 1.5
charge on NPF days was about 10%. In agreement with Jayaratne et al. (2015b), this value decreased significantly during the actual event times.
3.8. Percentage of particles charged Fig. 6 shows the mean percentage of particles charged as a function of particle size during all NPF events over the entire measurement period. The percentages were calculated for each of seven size bins in the detectable particle size range 2.0e42 nm. The results were compared with the curve corresponding to the steady state charge distribution of particles in the bipolar ion atmosphere from Wiedensohler's approximation of Fuchs theory (Wiedensohler, 1988). In charge equilibrium, the percentage of particles charged increase strongly with increasing particle diameter, from about 4% at 5 nm to about 42% at 50 nm. At any given particle size, a value greater than the equilibrium value indicates that the particles are overcharged, while a value smaller than the equilibrium value indicates that the particles are undercharged. Fig. 6 demonstrates that, during NPF events, particles smaller than 5 nm were overcharged while particles larger than 5 nm were undercharged. Similar observations have been reported previously (Jayaratne et al., 2016b) and (Vana et al., 2006).
4. Discussion The NPF occurrence rate of 41% is considerably higher than the
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
557
Fig. 4. Spectragram of the NPF event in Fig. 1, showing the time variation of the PNC in the two different size ranges 2.0e42 nm (white curve) and 2.0e10 nm (black curve).
Table 3 The number concentration of charged and neutral particles during NPF events, NPF days and non-event days. Number Concentration (cm3)
NPF Event (Maximum) NPF Days (Mean) Non-Event Days (Mean)
Charged Particles (x 103)
Neutral Particles (x 104)
% Charged
8.7 ± 2.8 3.2 ± 0.8 2.8 ± 0.6
15.3 ± 7.5 3.0 ± 1.7 1.9 ± 0.4
5 10 12
corresponding values of 35% and 26% found with the SMPS by Guo et al. (2008) and Cheung et al. (2011) respectively, and significantly lower than the value of 77% reported by Salimi et al. (2017), all in Brisbane. These earlier results were derived with an SMPS. The wide range of values may be attributed to the greater efficiency of the NAIS over the SMPS in identifying NPF events as well as to the different criteria used to identify an event. In particular, the particle size range used to determine the value of N used in the term dN/dt is critically important in the criterion. Recent studies using the NAIS by Crilley et al. (2014) and Jayaratne et al. (2016a) have provided much closer occurrence rates of 56% and 45%, respectively. The
Crilley et al. (2014) study was conducted during the summer months, when the occurrence rate was generally higher than at other times of the year, while the Jayaratne et al. (2016a) study covered an entire year. Here too, it should be noted that the criteria used to define an NPF event were slightly different in the two papers. Jayaratne et al. (2016a), as in the present study, stipulated an additional requirement for the minimum time during which N was greater than a certain threshold value. NPF events are not expected to be common in urban environments due to high loading of pre-existing aerosols particles that provide large condensation and coagulation sinks. The observed occurrence rate of NPF event in other urban cities such as Birmingham (Alam et al., 2003), Paris (Dos Santos et al., 2015), Budapest (Salma et al., 2011), and Helsinki (Hussein et al., 2008) varied between 5 and 27%. However, in some other cities the observed occurrence rate of NPF events was higher. For example, in St Louis (Qian et al., 2007), Pittsburgh (Stanier et al., 2004), and Beijing (Yu et al., 2015) with 36%, 50% and 47% respectively. It should be noted that the population densities in these three cities exceed 1000 km2 which is much higher than the equivalent figure of 145 km2 in Brisbane. The higher the population, the greater will be the number of motor vehicles and other sources of precursor
Fig. 5. Mean number concentration of particles on all NPF days. The charged particles are shown in red and the neutral particles are in blue. The mean values of the charged and total number concentrations on NPF and non-event days over the entire period of observation are indicated on the right margin. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
558
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
Fig. 6. Mean number percentage of particles carrying a charge during NPF events as a function of their size. The line shows the percentage of particles charged at equilibrium according to the Wiedensohler model.
gases, which may explain the higher rates of NPF occurrence. The results presented in Fig. 2 need to be considered with some care as they are based on just one calendar year, where the number of observational days varied from month to month. January and March are particularly uncertain as they consisted of less than 10 observational days each. Despite this limitation, we can derive several important characteristics. We note that the minimum percentage of NPF events occurred in December, a month during which there were several days with smoke from back-burning activities in the northern suburbs of Brisbane. NPF events were suppressed during the course of this month due to the prevalent high condensation sink (Jayaratne et al., 2015b). In spite of this suppression, the probability of occurrence of NPF was highest during the summer (43.8% in NovembereFebruary) and lowest during the winter (32.1% in MayeAugust). Some caution is needed in comparing these values with other locations because the difference in mean temperature in Brisbane between summer and winter is not as great as in more temperate climates. Manninen et al. (2010) analysed NPF events in 12 European cities and found that the highest occurrence rate was in late spring and the lowest in winter. Salma et al. (2011) found the highest in the spring (44%) and the lowest in the winter (7%) in Budapest, Hungary. Qian et al. (2007) found a marked difference between the summer (36%) and winter (8%) in St Louis, USA. In Fig. 3, the most probable starting times of NPF (8e9 am) coincides very well with the morning rush hour traffic. Although the PNC is relatively high at this time, the concentration of condensable vapours produced by motor vehicles is also at a maximum. Thus, despite the higher CS, NPF events are still observed to occur. In fact, we have previously shown that while NPF events are not observed in Brisbane when the PM10 in the atmosphere is higher than 20 mg m3 (Jayaratne et al., 2015a) in more polluted Beijing they are observed at PM10 concentrations as high as 43 mg m3 (Jayaratne et al., 2017). Studies have shown that the main precursor for NPF in urban environments is sulphuric acid from industry and motor vehicle emissions (Stanier et al., 2004; Zhang et al., 2004; Crilley et al., 2014). We also found that the NPF starting times during the winter was about 2 h earlier than during the summer. This is surprising as the sun rises up to 2 h earlier in the summer and the air temperature and solar irradiance are significantly higher than during winter at any given time in the morning. A possible explanation is that the supersaturation of the precursor gases such as
sulphuric acid is considerably higher in the early morning winter hours when the temperature is lower (Zhang et al., 2004). The value of the CS in Brisbane, (5.3 ± 1.7) x 103 s1, is considerably lower than the values reported at some other urban cities such as Paris (14.3 103 s1) (Dos Santos et al., 2015) and Budapest (16 103 s1) (Salma et al., 2011). This is to be expected as the air quality in Brisbane is relatively better than in most other cities of similar or larger size. The coagulation sink for 2 nm particles determined here was (9.1 ± 2.9) x 104 s1. This parameter has not been determined for Brisbane before. We may compare it with the value of 9.9 104 s1 for 3 nm particles found in Beijing where the condensation sink is relatively much higher than in Brisbane (Wu et al., 2011). The mean formation rate of neutral particles during NPF events, calculated using the PNC in the range 2e3 nm, was 20.8 cm3 s1. This is comparable to the value of 17 cm3 s1 in St Louis, USA, determined by Qian et al. (2007) and within the range 0.7e32.4 cm3 s1 reported at 12 European cities by Manninen et al. (2010). Our results also showed that the formation rate of negatively charged particles was about 10% higher than that of positively charged particles. This is also in agreement with Manninen et al. (2010) in European cities. The growth rates of neutral particles found were 12.1 and 6.2 nm h1 in the size ranges 2.0e20 and 20e42 nm, respectively. There was a clear decrease in growth rates with increasing particle size. This explains why data obtained with an SMPS with a higher detection size threshold, usually provide lower growth rates than that derived by the NAIS. For example, the average value of growth rate obtained by Cheung et al. (2011) in Brisbane using an SMPS was 4.6 nm h1, which is significantly lower than the values obtained with the NAIS. The growth rates determined in our study may be compared with those from two other urban locations: 3e20 nm h1 in Atlanta, USA (Stolzenburg et al., 2005) and 2e13 nm h1 in Budapest, Hungary. In the present study, there was a significant difference between the growth rates of neutral and charged particles in the size range 2.0e20 nm but not so in the higher size range 20e42 nm (Table 2). In most NPF events, the charged particles formed a few tens of minutes before the neutral particles and suggested a possible role of ion-induced nucleation as previously proposed by Jayaratne et al. (2016a). The growth rates of the two types of particles attained
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
similar values after about 1.5 h. 5. Summary and conclusions Particles have not been monitored at their very smallest sizes in Brisbane until now. This study, conducted with a NAIS over a period of 1 year, has yielded more precise characteristics of NPF than has been possible before. It is also the first comprehensive study of the characteristics of NPF events in Australia. A summary of the findings of the study are listed below. NPF events were observed on 41% of the observational days with a higher occurrence rate in the summer (November to February) than in the winter (May to August). The mean PNC on NPF days was 49% higher than those of non NPF days. Events occurred mainly during the mornings, with the highest probability of occurrence between 8.00 and 10.00 a.m. The mean condensation and coagulation sinks during the events were estimated to be 5.3 103 s1 and 9.1 104 s1, respectively. There was a significant difference between the growth rates of neutral and charged particles in the size range 2.0e20 nm but not so in the higher size range 20e42 nm. The mean formation rates of 2e3 nm particles during NPF event was 20.8 cm3 s1. The formation rate of negatively charged particles was about 10% higher than the positively charged particles, in the size range 2e3 nm. The mean charged particle concentration on NPF days was 14% higher than those of non NPF days. On NPF days, 10% of particles were charged. Particles smaller than 5 nm were found to be overcharged in comparison with the steady state during NPF events. Particles larger than 5 nm were found to be undercharged in comparison with the steady state during NPF events. These results, based on data obtained at the very sizes where particles formed, have enhanced our knowledge of NPF events in urban environments. Some results, such as the diurnal variation of the occurrence of NPF events and the growth rates of charged particles are reported for the first time. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Acknowledgements We thank Dr. Xuan Ling for his assistance with the data acquisition. References Alam, A., Shi, J.P., Harrison, R.M., 2003. Observations of new particle formation in urban air. J. Geophys. Res. Atmos. 108. Birmili, W., Wiedensohler, A., 2000. New particle formation in the continental boundary layer: meteorological and gas phase parameter influence. Geophys. Res. Lett. 27, 3325e3328. Cheung, H., Morawska, L., Ristovski, Z., 2011. Observation of new particle formation in subtropical urban environment. Atmos. Chem. Phys. 11, 3823. Crilley, L.R., Jayaratne, E.R., Ayoko, G.A., Miljevic, B., Ristovski, Z., Morawska, L., 2014. Observations on the formation, growth and chemical composition of aerosols in an urban environment. Environ. Sci. Technol. 48, 6588e6596. €, J., Aalto, P., O'Dowd, C., 2002. Dal Maso, M., Kulmala, M., Lehtinen, K.E., M€ akela Condensation and coagulation sinks and formation of nucleation mode particles in coastal and boreal forest boundary layers. J. Geophys. Res. Atmos. 107. Dal Maso, M., Kulmala, M., Riipinen, I., Wagner, R., Hussein, T., Aalto, P.P., Lehtinen, K.E., 2005. Formation and growth of fresh atmospheric aerosols: eight years of aerosol size distribution data from SMEAR II, Hyytiala, Finland. Boreal Environ. Res. 10, 323. €rinen, A., Komppula, M., Tunved, P., KERMINEN, V., Dal Maso, M., Hyva Lihavainen, H., Viisanen, Y., HANSSON, H.C., Kulmala, M., 2008. Annual and interannual variation in boreal forest aerosol particle number and volume concentration and their connection to particle formation. Tellus B 60, 495e508. Dos Santos, V., Herrmann, E., Manninen, H., Hussein, T., Hakala, J., Nieminen, T.,
559
Aalto, P., Merkel, M., Wiedensohler, A., Kulmala, M., 2015. Variability of air ion concentrations in urban Paris. Atmos. Chem. Phys. 15, 13717e13737. Fuchs, N.A., Sutugin, A.G., 1971. High-dispersed Aerosols, Topic in Current Aerosol Research. Pergamon, NY. Guo, H., Ding, A., Morawska, L., He, C., Ayoko, G., Li, Y.S., Hung, W.T., 2008. Size distribution and new particle formation in subtropical eastern Australia. Environ. Chem. 5, 382e390. , S., Lehtipalo, K., Manninen, H.E., Ehn, M., Hirsikko, A., Nieminen, T., Gagne ~rrak, U., Kerminen, V.M., Laakso, L., McMurry, P.H., Mirme, A., Mirme, S., Ho Pet€ aj€ a, T., Tammet, H., Vakkari, V., Vana, M., Kulmala, M., 2011. Atmospheric ions and nucleation: a review of observations. Atmos. Chem. Phys. 11, 767e798. https://doi.org/10.5194/acp-11-767-2011. Hussein, T., Martikainen, J., Junninen, H., Sogacheva, L., Wagner, R., Dal Maso, M., Riipinen, I., Aalto, P.P., Kulmala, M., 2008. Observation of regional new particle formation in the urban atmosphere. Tellus B 60, 509e521. Iida, K., Stolzenburg, M., McMurry, P., Dunn, M.J., Smith, J.N., Eisele, F., Keady, P., 2006. Contribution of ion-induced nucleation to new particle formation: methodology and its application to atmospheric observations in Boulder, Colorado. J. Geophys. Res. Atmos. 111. Jayaratne, E.R., Clifford, S., Morawska, L., 2015a. Atmospheric visibility and PM10 as indicators of new particle formation in an urban environment. Environ. Sci. Technol. 49, 12751e12757. https://doi.org/10.1021/acs.est.5b01851. Jayaratne, E.R., Pushpawela, B., Morawska, L., 2016a. Temporal evolution of charged and neutral nanoparticle concentrations during atmospheric new particle formation events and its implications for ion-induced nucleation. Front. Environ. Sci. Eng. 10, 13. Jayaratne, R., Ling, X., Morawska, L., 2016b. Charging state of aerosols during particle formation events in an urban environment and its implications for ion-induced nucleation. Aerosol Air Qual. Res. 16, 348e360. Jayaratne, R., Pushpawela, B., He, C., Li, H., Gao, J., Chai, F., Morawska, L., 2017. Observations of particles at their formation sizes in Beijing, China. Atmos. Chem. Phys. 17, 8825e8835. https://doi.org/10.5194/acp-17-8825-2017. Jayaratne, R.E., Ling, X., Morawska, L., 2015b. Suppression of cluster ions during rapidly increasing particle number concentration events in the environment. Aerosol Air Qual. Res. 15, 28e37. Jennings, S.G., 1988. The mean free path in air. J. Aerosol Sci. 19, 159e166. https:// doi.org/10.1016/0021-8502(88)90219-4. Jeong, K., 2009. Condensation of Water Vapor and Sulfuric Acid in Boiler Flue Gas. ProQuest. €j€ Kulmala, M., Vehkam€ aki, H., Peta a, T., Dal Maso, M., Lauri, A., Kerminen, V.-M., Birmili, W., McMurry, P.H., 2004. Formation and growth rates of ultrafine atmospheric particles: a review of observations. J. Aerosol Sci. 35, 143e176. Kulmala, M., Riipinen, I., Sipil€ a, M., Manninen, H.E., Pet€ aj€ a, T., Junninen, H., Dal Maso, M., Mordas, G., Mirme, A., Vana, M., 2007. Toward direct measurement of atmospheric nucleation. Science 318, 89e92. €ja €, T., Nieminen, T., Sipil€ Kulmala, M., Peta a, M., Manninen, H.E., Lehtipalo, K., Dal Maso, M., Aalto, P.P., Junninen, H., Paasonen, P., 2012. Measurement of the nucleation of atmospheric aerosol particles. Nat. Protoc. 7, 1651e1667. Kulmala, M., Kontkanen, J., Junninen, H., Lehtipalo, K., Manninen, H.E., Nieminen, T., €, M., Schobesberger, S., Rantala, P., 2013. Direct observations of Pet€ aj€ a, T., Sipila atmospheric aerosol nucleation. Science 339, 943e946. Lehtinen, K.E., Korhonen, H., Maso, M., Kulmala, M., 2003. On the concept of condensation sink diameter. Boreal Environ. Res. 8, 405e412. Lehtinen, K.E., Dal Maso, M., Kulmala, M., Kerminen, V.-M., 2007. Estimating nucleation rates from apparent particle formation rates and vice versa: revised formulation of the KermineneKulmala equation. J. Aerosol Sci. 38, 988e994. €ja €, T., Kerminen, V.-M., Kulmala, M., Leino, K., Nieminen, T., Manninen, H.E., Peta 2016. Intermediate ions as a strong indicator of new particle formation bursts in a boreal forest. Boreal Environ. Res. 21 (3-4), 274e286. , S., H€ Manninen, H., Nieminen, T., Asmi, E., Gagne akkinen, S., Lehtipalo, K., Aalto, P., Vana, M., Mirme, A., Mirme, S., 2010. EUCAARI ion spectrometer measurements at 12 European siteseanalysis of new particle formation events. Atmos. Chem. Phys. 10, 7907e7927. Manninen, H., Franchin, A., Schobesberger, S., Hirsikko, A., Hakala, J., Skromulis, A., Kangasluoma, J., Ehn, M., Junninen, H., Mirme, A., 2011. Characterisation of corona-generated ions used in a neutral cluster and air ion spectrometer (NAIS). Atmos. Meas. Tech. 4, 2767. Manninen, H.E., Mirme, S., Mirme, A., Pet€ aj€ a, T., Kulmala, M., 2016. How to reliably detect molecular clusters and nucleation mode particles with Neutral cluster and Air Ion Spectrometer (NAIS). Atmos. Meas. Tech. Discuss. 9 (8), 3577e3605. https://doi.org/10.5194/amt-9-3577-2016. Qian, S., Sakurai, H., McMurry, P.H., 2007. Characteristics of regional nucleation events in urban East St. Louis. Atmos. Environ. 41, 4119e4127. Salimi, F., Ristovski, Z., Mazaheri, M., Laiman, R., Crilley, L.R., He, C., Clifford, S., Morawska, L., 2014. Assessment and application of clustering techniques to atmospheric particle number size distribution for the purpose of source apportionment. Atmos. Chem. Phys. 14, 11883e11892. Salimi, F., Rahman, M., Clifford, S., Ristovski, Z., Morawska, L., 2017. Nocturnal new particle formation events in urban environments. Atmos. Chem. Phys. 17, 521e530. s, T., Weidinger, T., Aalto, P., Hussein, T., Dal Maso, M., Kulmala, M., Salma, I., Borso 2011. Production, growth and properties of ultrafine atmospheric aerosol particles in an urban environment. Atmos. Chem. Phys. 11, 1339. Seinfeld, J.H., Pandis, S.N., 2006. Atmospheric Chemistry and Physics. Wiley, Hoboken, NJ.
560
B. Pushpawela et al. / Environmental Pollution 233 (2018) 552e560
Stanier, C.O., Khlystov, A.Y., Pandis, S.N., 2004. Nucleation events during the Pittsburgh Air Quality Study: description and relation to key meteorological, gas phase, and aerosol parameters special issue of aerosol science and technology on findings from the fine particulate matter supersites program. Aerosol Sci. Technol. 38, 253e264. Stolzenburg, M.R., McMurry, P.H., Sakurai, H., Smith, J.N., Mauldin, R.L., Eisele, F.L., Clement, C.F., 2005. Growth rates of freshly nucleated atmospheric particles in Atlanta. J. Geophys. Res. Atmos. 110. Vana, M., Tamm, E., Horrak, U., Mirme, A., Tammet, H., Laakso, L., Aalto, P., Kulmala, M., 2006. Charging state of atmospheric nanoparticles during the nucleation burst events. Atmos. Res. 82, 536e546. Wiedensohler, A., 1988. An approximation of the bipolar charge distribution for particles in the submicron size range. J. Aerosol Sci. 19, 387e389.
Wu, Z., Hu, M., Yue, D., Wehner, B., Wiedensohler, A., 2011. Evolution of particle number size distribution in an urban atmosphere during episodes of heavy pollution and new particle formation. Sci. China Earth Sci. 54, 1772. Yu, F., Turco, R.P., 2000. Ultrafine aerosol formation via ion-mediated nucleation. Geophys. Res. Lett. 27, 883e886. Yu, H., Zhou, L., Dai, L., Shen, W., Zheng, J., Ma, Y., Chen, M., 2015. Nucleation and growth of sub-3 nm particles in the polluted urban atmosphere of a megacity in China. Atmos. Chem. Phys. Discuss. 15. Zhang, Q., Stanier, C., Canagaratna, M., Jayne, J., Worsnop, D., Pandis, S., Jiminez, J., 2004. Insights into the chemistry of new particle formation and growth events in Pittsburgh based on aerosol mass spectrometry. Environ. Sci. Technol. 38, 4797e4809.