Atmos. Chem. Phys., 20, 12721–12740, 2020 https://doi.org/10.5194/acp-20-12721-2020 © Author(s) 2020. This work is distributed under the Creative Commons Attribution 4.0 License. Size-segregated particle number and mass concentrations from different emission sources in urban Beijing Jing Cai1,2, Biwu Chu1,2,3,4, Lei Yao2, Chao Yan1,2, Liine M. Heikkinen1,2, Feixue Zheng1, Chang Li1, Xiaolong Fan1, Shaojun Zhang5, Daoyuan Yang5, Yonghong Wang2, Tom V. Kokkonen1,2, Tommy Chan1,2, Ying Zhou1, Lubna Dada1,2, Yongchun Liu1, Hong He3,4, Pauli Paasonen1,2, Joni T. Kujansuu1,2, Tuukka Petäjä1,2, Claudia Mohr6, Juha Kangasluoma1,2, Federico Bianchi1,2, Yele Sun7, Philip L. Croteau8, Douglas R. Worsnop2,8, Veli-Matti Kerminen1,2, Wei Du1,2, Markku Kulmala1,2, and Kaspar R. Daellenbach1,2 1Aerosol and Haze Laboratory, Beijing Advanced Innovation Center for Soft Matter Science and Engineering, Beijing University of Chemical Technology, Beijing, 100029, China 2Institute for Atmospheric and Earth System Research, Faculty of Science, University of Helsinki, Helsinki, 00014, Finland 3Center for Excellence in Regional Atmospheric Environment, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen, 361021, China 4State Key Joint Laboratory of Environment Simulation and Pollution Control, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing, 100085, China 5School of Environment, Tsinghua University, Beijing, 100084, China 6Department of Environmental Science, Stockholm University, Stockholm, 11418, Sweden 7State Key Laboratory of Atmospheric Boundary Layer Physics and Atmospheric Chemistry, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, 100029, China 8Aerodyne Research, Inc., Billerica, MA 01821, USA Correspondence: Wei Du (wei.du@helsinki.fi), Markku Kulmala (markku.kulmala@helsinki.fi), and Kaspar R. Daellenbach ([email protected]) Received: 14 March 2020 – Discussion started: 12 May 2020 Revised: 1 September 2020 – Accepted: 14 September 2020 – Published: 4 November 2020 Abstract. Although secondary particulate matter is reported 4:2±3:0×103 cm−3), were resolved by both methods. Con- to be the main contributor of PM2:5 during haze in Chi- verted mass concentrations from particle size distributions nese megacities, primary particle emissions also affect par- components were comparable with those from chemical fin- ticle concentrations. In order to improve estimates of the gerprints. Size distribution source apportionment separated contribution of primary sources to the particle number and vehicular emissions into a component with a mode diameter mass concentrations, we performed source apportionment of 20 nm (“traffic-ultrafine”) and a component with a mode analyses using both chemical fingerprints and particle size diameter of 100 nm (“traffic-fine”). Consistent with similar distributions measured at the same site in urban Beijing day- and nighttime diesel vehicle PM2:5 emissions estimated from April to July 2018. Both methods resolved factors re- for the Beijing area, traffic-fine particles, hydrocarbon-like lated to primary emissions, including vehicular emissions OA (HOA, traffic-related factor resulting from source ap- and cooking emissions, which together make up 76 % and portionment using chemical fingerprints) and black carbon 24 % of total particle number and organic aerosol (OA) mass, (BC) showed similar diurnal patterns, with higher concentra- respectively. Similar source types, including particles re- tions during the night and morning than during the afternoon lated to vehicular emissions (1:6 ± 1:1 µgm−3; 2:4 ± 1:8 × when the boundary layer is higher. Traffic-ultrafine particles 103 cm−3 and 5:5 ± 2:8 × 103 cm−3 for two traffic-related showed the highest concentrations during the rush-hour pe- components), cooking emissions (2:6±1:9 µgm−3 and 5:5± riod, suggesting a prominent role of local gasoline vehicle 3:3×103 cm−3) and secondary aerosols (51±41 µgm−3 and emissions. In the absence of new particle formation, our re- Published by Copernicus Publications on behalf of the European Geosciences Union. 12722 J. Cai et al.: Size-segregated particle number and mass concentrations from emission sources sults show that vehicular-related emissions (14 % and 30 % cal mass balance (CMB) (Zhang et al., 2013; Y. Zhang et al., for ultrafine and fine particles, respectively) and cooking- 2017; Tao et al., 2017). Among them, many focused on the activity-related emissions (32 %) dominate the particle num- sources of OA due to its large contribution to fine PM, its ber concentration, while secondary particulate matter (over complex mix of origins and its tracers from different sources. 80 %) governs PM2:5 mass during the non-heating season in The development of aerosol mass spectrometer technologies Beijing. has allowed identification of the primary sources of OA in Beijing, namely traffic emissions, cooking activities, biomass burning and coal combustion (Xu et al., 2019; Hu et al., 2017; Sun et al., 2013b, 2018). Generally, in OA source apportion- 1 Introduction ment using chemical fingerprints from mass spectrometers (aerosol chemical speciation monitor, ACSM, and aerosol Even though it is commonly recognized that secondary mass spectrometer, AMS), particle size distributions are dis- aerosol mass governs haze formation in megacities in China regarded. Results of source contributions focus on mass and (Huang et al., 2014; Zhang et al., 2013; Tao et al., 2017; Sun not on number, which has however shown to be of impor- et al., 2018), the contributions of primary (direct) particle tance from a health perspective as well. Aerosol mass spec- sources cannot be neglected. Previous studies have demon- trometers are blind to particles smaller than ∼ 70 nm (Xu strated that primary emission sources, such as residential et al., 2017). Apportioning smaller particles to their sources heating, traffic and cooking activities, can contribute signif- is therefore crucial for air quality mitigation. icantly to both particle number and mass concentrations in Size-distribution-based source apportionment can pro- the urban atmosphere in China (He et al., 2004a; Xu et al., vide, though less applied and with more uncertainties, size- 2014; Du et al., 2017; Wang et al., 2013; Sun et al., 2018). segregated particle number concentrations of sources and It was recently reported that traffic could be a major source processes. Until now, size distribution source apportionment of nanoclusters (sub-3 nm) in urban environments (Ronkko has successfully been applied to data from US, European and et al., 2017). On average, 13 %–24 % of the total fine or- Chinese cities, such as London (Harrison et al., 2011; Bed- ganic aerosol (OA) mass concentration can be attributed to dows and Harrison, 2019), New York (Ogulei et al., 2007), cooking activities and 11 %–20 % to traffic emissions in Bei- Barcelona (Vu et al., 2015) and Beijing (Wang et al., 2013; jing, China (Hu et al., 2016, 2017). Together with direct Du et al., 2017; Liu et al., 2014). Different sources, such as particle emissions, many identified primary sources co-emit different types of traffic, cooking, road dust, combustion, re- high concentrations of volatile organic compounds (VOCs), suspension and secondary sulfate and nitrate have been iden- which in turn contribute to secondary organic aerosol (SOA) tified in Beijing by this approach (Du et al., 2017; Liu et al., mass formation (T. Liu et al., 2017a, b). Therefore, it is im- 2014; Vu et al., 2015; Wang et al., 2013). The application portant to identify primary particle sources and disentangle of size PMF from previous literature is summarized in Ta- them from the secondary organic and inorganic aerosol (SOA ble S1 in the Supplement. Yet, until now, very few studies and SIA) whose precursors were co-emitted, with the goal to have combined particle number size distribution source ap- better understand their contributions in highly complex urban portionment with chemical speciation source apportionment atmospheres for advising air pollution control policies. and compared their results in a comprehensive manner. In Beijing, a megacity with a population of 20 million, has general, size-distribution-based source apportionment results suffered from severe fine particulate matter (PM) pollution tend to lack validation as well as comparison to other meth- for several decades (He et al., 2001; Tao et al., 2017). Due to ods, which results in larger uncertainties and a necessity to its impact on human health and the climate, fine particulate combine it with chemical speciation source apportionment. matter has gained increased attention (Lelieveld et al., 2015; In this study, we aim to better constrain the chemical and Huang et al., 2014). To study the fine PM sources in Beijing, physical properties of primary organic aerosol in Beijing us- numerous receptor source apportionment studies have been ing particle number size distribution and chemical speciation conducted (Ding et al., 2016; Tao et al., 2017; Hu et al., 2016; source apportionment approaches. We applied both chemi- Zhang et al., 2013). These studies can be grouped into two cal fingerprints (OA PMF) and particle size distribution (size approaches: the widely applied chemical component method PMF) analyses to resolve the particle mass and number con- (Xu et al., 2019; Hu et al., 2017; Sun et al., 2013b, 2018) and tributions from various sources during the same period. Com- the less-applied size distribution method (Wang et al., 2013; bining physical information