Elucidate the formation mechanism of particulate nitrate based on direct radical observations in the Yangtze River Delta summer 2019

Particulate nitrate (NO${}_{\mathrm{3}}^{-})$ is one of the dominant components of fine particles in China, especially during pollution episodes, and has a significant impact on human health, air quality, and climate. Here a comprehensive field campaign that focuses on the atmospheric oxidation capacity and aerosol formation and their effects in the Yangtze River Delta (YRD) was conducted from May to June 2019 at a regional site in Changzhou, Jiangsu Province in China. The concentrations of NO${}_{\mathrm{3}}^{-}$, OH radicals, N₂O₅, NO₂, O₃, and relevant parameters were measured simultaneously. We showed a high NO${}_{\mathrm{3}}^{-}$ mass concentration with 10.6 ± 8.9 µg m⁻³ on average, which accounted for 38.3 % of total water-soluble particulate components and 32.0 % of total PM_2.5, followed by the proportion of sulfate, ammonium, and chloride by 26.0 %, 18.0 %, and 2.0 %, respectively. This result confirmed that the heavy nitrate pollution in eastern China happened not only in winter, but also in summer. This study's high nitrate oxidation ratio (NOR) emphasized the solid atmospheric oxidation and fast nitrate formation capacity in the YRD. It was found that OH + NO₂ during daytime dominated nitrate formation on clean days, while N₂O₅ hydrolysis vastly enhanced and became comparable with that of OH + NO₂ during polluted days (67.2 % and 30.2 %, respectively). The updated observed-constraint Empirical Kinetic Modeling Approach (EKMA) was used to assess the kinetic controlling factors of both local O₃ and NO${}_{\mathrm{3}}^{-}$ productions, which indicated that the O₃-targeted scheme (VOCs : NO_x = 2:1) is adequate for mitigating the O₃ and nitrate pollution coordinately during summertime in this region. Our results promote the understanding of nitrate pollution mechanisms and mitigation based on field observation and model simulation and call for more attention to nitrate pollution in the summertime.

1 Introduction

Chemical compositions of fine particles have been measured in China during the past 20 years, and secondary inorganic aerosol is regarded as one of the dominant species in aerosol (Cao et al., 2012; Hagler et al., 2006; Zhao et al., 2013; Andreae et al., 2008). Since the Air Pollution Prevention and Control Action Plan, there has been a significant decrease in SO₂, NO₂, and PM_2.5 concentrations in China, while the inorganic nitrate ratio in PM_2.5 increased and became the considerable component in PM_2.5 (Shang et al., 2021; Zhang et al., 2022). Therefore, a comprehensive understanding of the particulate nitrate formation mechanism is essential and critical for mitigating haze pollution in China.

Massive research has been done in China to investigate nitrate formation mechanisms, and a basic framework has been established (Sun et al., 2006; Chang et al., 2018; Wu et al., 2019). In the daytime, NO₂ + OH radical oxidation (Reaction 1) is the major particulate nitrate formation pathway. The product (HNO₃) reacts with alkaline substances in aerosol, generating particulate nitrate. This pathway is mainly controlled by precursor concentrations as well as the gas-particle partition of gaseous nitric acid, and particulate nitrate depends on temperature, relative humidity (RH), NH₃ concentration, and aerosol acidity (Wang et al., 2009; Song and Carmichael, 2001; Meng et al., 2020; Zhang et al., 2021). At night, N₂O₅ uptake is a vital nitrate formation pathway (Reaction 4) (Chen et al., 2020; Wang et al., 2022). N₂O₅ is formed through NO₂ + NO₃ (Reaction 3), and there exists a quick thermal equilibrium balance ($K_{\text{eq}}=\mathrm{5.5}\times {\mathrm{10}}^{-\mathrm{17}}$ cm⁻³ molec.⁻¹ s⁻¹, 298 K). However, two problems remain ambiguous in quantifying the contribution of N₂O₅ uptake to nitrate formation. The first is that the N₂O₅ heterogeneous uptake coefficient (γ) on ambient aerosol is highly varied, with a range from 10⁻⁴ to 10⁻¹ based on previous lab and field measurements (Bertram and Thornton, 2009; Brown et al., 2009; Z. Wang et al., 2017; Wang and Lu, 2016). The other one is the ClNO₂ production yield which influences nitrate contribution due to the extensive variation range (Phillips et al., 2016; Staudt et al., 2019; Tham et al., 2018). Both parameters are complex to well-predicted by current schemes. No₂ heterogeneous uptake has been found to be non-negligible for nitrate formation, which can be a vital pathway during heavy haze events according to recent studies (Qiu et al., 2019; Chan et al., 2021). The uptake coefficient and nitrate yield remain uncertain, as with the N₂O₅ heterogeneous reaction. In addition, N₂O₅ homogeneous hydrolysis and NO₃ radical oxidation have a minor contribution to particulate nitrate under ambient conditions (Brown et al., 2009; Seinfeld and Pandis, 2016).

$$\begin{array}{}\text{(R1)}& {\displaystyle }{\mathrm{NO}}_{\mathrm{2}}+\mathrm{OH}\to {\mathrm{HNO}}_{\mathrm{3}}\text{(R2)}& {\displaystyle }{\mathrm{NO}}_{\mathrm{2}}+{\mathrm{O}}_{\mathrm{3}}\to {\mathrm{NO}}_{\mathrm{3}}+{\mathrm{O}}_{\mathrm{2}}\text{(R3.1)}& {\displaystyle }{\mathrm{NO}}_{\mathrm{2}}+{\mathrm{NO}}_{\mathrm{3}}+M\to {\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}+M\text{(R3.2)}& {\displaystyle }{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}+M\to {\mathrm{NO}}_{\mathrm{2}}+{\mathrm{NO}}_{\mathrm{3}}+M\text{(R4)}& {\displaystyle }{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}+\left({\mathrm{H}}_{\mathrm{2}}\mathrm{O}\text{or}{\mathrm{Cl}}^{-}\right)\to (\mathrm{2}-\mathit{\phi }){\mathrm{NO}}_{\mathrm{3}}^{-}+\mathit{\phi }{\mathrm{ClNO}}_{\mathrm{2}}\end{array}$$

As a critical area of China's economy and industry, the Yangtze River Delta (YRD) has suffered severe air pollution during past decades, and fine particle pollution in the YRD has raised widespread concern (Guo et al., 2014; Zhang et al., 2015, 2017; Ming et al., 2017; Xue et al., 2019). However, most research focuses on wintertime PM_2.5 pollution and lacks measurements of critical intermediate species and radicals to assess the importance of each nitrate formation pathway. In this study, with the direct measurements of hydroxyl radicals and the reactive nitrogen compounds and chemical box model analyses, we explore the characteristics of nitrate and precursors in the YRD in the summer of 2019, the importance of particulate nitrate formation pathways is quantified, and the controlling factors are explored. A further suggestion for summer pollution prevention and control in the local area is proposed.

2 Site description and methods

2.1 The campaign site

This campaign took place at a suburban sanatorium from 30 May to 18 June 2019 in Changzhou, China. Changzhou (119.95^∘ E, 31.79^∘ N) is located in Jiangsu Province and about 150 km northwest of Shanghai. The sanatorium, located 420 m east of Lake Ge (one of the largest lakes in Jiangsu Province, 164 km²), is surrounded by farmland and fishponds. With the closest arterial traffic 1 km away, several industry zones are 4 km to the east. The prevailing wind was from the southern and southeastern sectors (about 30 % of the time) compared to 20 % from the western sector, of which only 15 % came from the east. The wind speed was usually lower than 5 m s⁻¹, with faster speed from the west. This site was influenced by anthropogenic and biological sources with occasional biomass burning.

Figure 1 The location of the campaign site (red star), Changzhou, is 150 km on the northwestern side of Shanghai.

2.2 The instrumentation

Multiple gaseous and particulate parameters were measured simultaneously during the campaign to comprehensively interpret the nocturnal atmospheric capacity and aerosol formation. The related instruments are listed in Table 1. N₂O₅ and particle number and size distribution (PNSD) were measured on the fourth floor of the sanatorium, which is the top of the building. Other instruments were placed in containers on the ground 170 m northeast of the building and sampling inlets at circa 5 m above the ground through the containers' roof.

N₂O₅ was measured by a cavity-enhanced absorption spectrometer (CEAS) based on the Lambert–Beer law which was developed by Wang et al. (2017b). Briefly, air samples were drawn through the window and reached out of the wall 30 cm to prevent influence from surface deposition. The aerosol membrane filter was deployed before the perfluoroalkoxy alkane (PFA) sampling tube and changed every 2 h at night to avoid a decrease in N₂O₅ transmission efficiency due to the increased loss of N₂O₅ from the accumulated aerosols on the filter. N₂O₅ was decomposed into NO₃ and NO₂ by preheating tube heat at 130^∘ and detected within a PFA-coated resonator cavity heated at 110^∘ to prevent the formation of N₂O₅ by reversible reaction subsequently. At the end of each sampling cycle (5 min), a 30 s injection of high-concentration NO (10 ppm, 20 mL min⁻¹) mixed with sample air was set to eliminate NO₃–N₂O₅ from the system. The NO titration spectra were adopted as the dynamic background spectrum by assuming no H₂O concentration variation in a single sampling cycle. The loss of N₂O₅ in the sampling system and filter was also considered during data correction. The limit of detection (LOD) was estimated to be 2.7 pptv (1σ) with an uncertainty of 19 %.

OH radical measurement was conducted by fluorescence assay by gas expansion laser-induced fluorescence techniques (FAGE-LIF). Ambient air was expanded through a 0.4 mm nozzle to low pressure in a detection chamber, where the 308 nm laser pulse irradiated OH radicals at a repetition rate of 8.5 kHz (Chen et al., 2018). NO_x and O₃ were monitored by commercial monitors (Thermo-Fisher 42i and 49i). Volatile organic compounds (VOCs) were measured by an automated gas chromatograph equipped with a mass spectrometer and flame ionization detector (GC-MS) with a time resolution of 60 min. The photolysis frequencies were determined from the spectral actinic photon flux density measured by a spectroradiometer (Bohn et al., 2008).

PM_2.5 concentration was obtained by a Tapered Element Oscillating Microbalance (TEOM 1405, Thermo Scientific Inc). Aerosol surface concentration (S_a) was converted from the particle number and size distribution, which was measured by a Scanning Mobility Particle Sizer (SMPS, TSI 3936) and an Aerosol Particle Sizer (APS, TSI 3321) and modified to the wet particle-state S_a with a hygroscopic growth factor (Liu et al., 2013). The uncertainty of the wet S_a was ∼30 %. Meanwhile, water-soluble particulate components and their gaseous precursors were analyzed through the Monitor for AeRosols and GAses in ambient air (MARGA, Chen et al., 2017). Meteorological data were also available, including the temperature, RH, pressure, wind speed, and wind direction.

Table 1 The observed gas and particle parameters during the campaign.

^a Laser-induced fluorescence. ^b Chemiluminescence. ^c Photolytic converter. ^d Tapered Element Oscillating Microbalance. ^e Spectroradiometer. ^f The Monitor for AeRosols and GAses in ambient air.

2.3 The Empirical Kinetic Modeling Approach

A box model coupled with the Regional Atmospheric Chemical Mechanism version 2 (RACM2, Goliff, Stockwell and Lawson, 2013) is used to conduct the mitigation strategy studies. The model is operated at 1 h time resolution with measurement results of temperature, relative humidity, pressure, CO, NO₂, H₂O, photolysis frequencies, and aggregated VOC input to constrain the model. It should be noted that HONO concentration is calculated by NO₂ × 0.02, as suggested by Elshorbany et al. (2012), and has been used in the box model before (Lou et al., 2022). Long-lived species such as H₂ and CH₄ are assumed to be constants (550 and 1900 ppbv, respectively). Moreover, a 13 h constant loss rate of unconstrained intermediate and secondary products, the result of synthetic evaluating secondary simulation of secondary species, is set for representing the multiple effects of deposition, transformation, and transportation.

The approaches to the chemical production of O₃ (P(O₃)) and inorganic nitrate (P(NO${}_{\mathrm{3}}^{-}\left)\right)$ are described in previous articles (Tan et al., 2021, 2018) and expressed as Eqs. (1) and (4).

$$\begin{array}{}\text{(1)}& {\displaystyle }\mathrm{P}\left({\mathrm{O}}_{\mathrm{3}}\right)=\mathrm{F}\left({\mathrm{O}}_{\mathrm{3}}\right)-\mathrm{D}\left({\mathrm{O}}_{\mathrm{3}}\right)\text{(2)}& {\displaystyle }\mathrm{F}\left({\mathrm{O}}_{\mathrm{3}}\right)=k_{{\mathrm{HO}}_{\mathrm{2}}+\mathrm{NO}}\left[\mathrm{NO}\right]\left[{\mathrm{HO}}_{\mathrm{2}}\right]+k_{({\mathrm{RO}}_{\mathrm{2}}+\mathrm{NO})\mathrm{eff}}\left[\mathrm{NO}\right]\left[{\mathrm{RO}}_{\mathrm{2}}\right]\text{(3)}& {\displaystyle }\begin{array}{rl}\mathrm{D}\left({\mathrm{O}}_{\mathrm{3}}\right)=& k_{\mathrm{OH}+{\mathrm{NO}}_{\mathrm{2}}}\left[\mathrm{OH}\right]\left[{\mathrm{NO}}_{\mathrm{2}}\right]+(k_{\mathrm{OH}+{\mathrm{O}}_{\mathrm{3}}}\left[\mathrm{OH}\right]\\ +& k_{{\mathrm{HO}}_{\mathrm{2}}+{\mathrm{O}}_{\mathrm{3}}}\left[{\mathrm{HO}}_{\mathrm{2}}\right]+k_{\mathrm{alkenes}+{\mathrm{O}}_{\mathrm{3}}}\left[\mathrm{alkenes}\right])\left[{\mathrm{O}}_{\mathrm{3}}\right]\end{array}\text{(4)}& {\displaystyle }\mathrm{P}\left({\mathrm{NO}}_{\mathrm{3}}^{-}\right)=\mathrm{P}\left({\mathrm{HNO}}_{\mathrm{3}}\right)+\mathrm{P}(p{\mathrm{NO}}_{\mathrm{3}}^{-})\text{(5)}& {\displaystyle }\mathrm{P}\left({\mathrm{HNO}}_{\mathrm{3}}\right)=k_{\mathrm{OH}+{\mathrm{NO}}_{\mathrm{2}}}\left[\mathrm{OH}\right]\left[{\mathrm{NO}}_{\mathrm{2}}\right]\text{(6)}& {\displaystyle }\mathrm{P}\left(p{\mathrm{NO}}_{\mathrm{3}}^{-}\right)=\mathrm{0.25}(\mathrm{2}-\mathit{\phi })C{\mathit{\gamma }}_{S_{\text{a}}}\left[{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}\right]\end{array}$$

Briefly, P(O₃) is net ozone production, which is calculated by peroxyl radial + NO oxidation (Eq. 2) minus the chemical loss of O₃ and NO₂ (Eq. 3). P(NO${}_{\mathrm{3}}^{-})$ is constituted by reaction OH + NO₂ (Eq. 5) and N₂O₅ heterogeneous uptake (Eq. 6). Here, rate constants of reactions are obtained from the NASA JPL publication or RACM2 (Goliff et al., 2013). γ is the N₂O₅ uptake coefficient calculated from parameterization (γ_P; more details in Sect. 3.3). φ represents ClNO₂ production yield through N₂O₅ hydrolysis, and the mean value reported by Xia et al. (2020) is used in this work.

The Empirical Kinetic Modeling Approach (EKMA) was invented to study the effects of precursor (VOC and NO_x) reactivity on the region's ozone pollution by Kanaya et al. (2008) and helps recognize the region's susceptibility to precursors by weight and became a prevalent tool to study the process of ozone formation (Tan et al., 2018; D. Yu et al., 2020; Kanaya et al., 2008). The prevention and control problem of pollutant generation can be transformed through the EKMA curve to reduce its precursors' emissions. Furthermore, the precursor reduction scheme needed for total pollutant control is given qualitatively. P(NO${}_{\mathrm{3}}^{-})$ can also be analyzed through EKMA for the nonlinear secondary formation relationship with precursor reactivity. Here, an isopleth diagram of the net ozone production rate as functions of the reactivities of NO_x and VOCs can be derived from EKMA. In detail, 0.01 to 1.2 emission reduction strategy assumptions are exponentially interpolated into 20 kinds of emission situations of NO_x and VOCs, respectively, which counts 400 scenarios.

2.4 The calculation of aerosol liquid water content

Aerosol liquid water content (ALWC) is calculated through ISORROPIA II (Fountoukis and Nenes, 2007). Forward mode is applied in this study. Furthermore, water-soluble particulate components in PM_2.5 and gaseous species (NH₃ + HNO₃ + HCl) obtained from MARGA, along with RH and T, are input as initial input. In addition, a metastable aerosol state is chosen due to high RH during this campaign.

3 Result and discussion

3.1 Overview of measurements

The time used in this study is China Standard Time (UTC + 8), and the local sunrise and sunset times during the campaign were around 17:00 and 19:00, respectively. The whole campaign period is divided into four PM_2.5 clean periods and four PM_2.5 polluted periods (9 out of 14 d; the latter polluted period days refer to PM_2.5 pollution except for a specified description) according to the Chinese National Air Quality Standard (CNAAQS) Grade I of daily PM_2.5 concentrations (<35.0 µg m⁻³). Figure 2 shows the meteorological parameters and gas-phase and particulate species time series during the observation. During the campaign, the temperature was high; the maximum reached 34.5^∘, with an average of 25.1 ± 3.7^∘. RH changed drastically from 21 % to 88 %, with a mean value of 58.9 ± 14.0 %. The mean NO₂ concentration was 14.8 ± 9.5 ppbv. Meanwhile, the O₃ average was 54.6 ± 28.8 ppbv, exceeding CNAAQS Grade II for a maximum daily average of 8 h ozone (160 µg m⁻³) on 14 out of 19 d and exceeding 200 µg m⁻³ on 6 d.

Figure 2 Time series of NO₂, O₃, N₂O₅, OH radicals, PM_2.5, and water-soluble particulate components, temperature, and RH. The vertical dotted line represents the zero clock. The black horizontal solid line in the O₃ and PM_2.5 panels represents Chinese national air quality standards for O₃ and PM_2.5, respectively. The top panel color blocks represent the PM_2.5 clean day (light green) and PM_2.5 polluted day (salmon).

Daytime OH radicals ranged from 2 × 10⁶ to 8 × 10⁶ molec. cm⁻³ with a daily peak over 3 × 10⁶ molec. cm⁻³. Maximum OH radicals reached 8.18 × 10⁶ molec. cm⁻³ in this campaign. Compared with other summertime OH radicals observed in China, the OH radical concentration at this site is relatively low but still on the same order of magnitude (Lu et al., 2012, 2013; Ma et al., 2022; Tan et al., 2017; Woodward-Massey et al., 2020; Yang et al., 2021). N₂O₅ mean concentration was 21.9 ± 39.8 pptv with a nocturnal average of 61.0 ± 63.1 pptv and a daily maximum of over 200 pptv on eight nights. The maximum concentration of N₂O₅ (477.2 pptv, 5 min resolution) appeared at 20:47 on 8 June. The average NO₃ radical production rate P(NO₃) is 2.1 ± 1.4 ppbv h⁻¹, with nocturnal average P(NO₃) 2.8 ± 1.6 ppbv h⁻¹ and daytime P(NO₃) 2.2 ± 1.4 ppbv h⁻¹. P(NO₃) is about twice the documented value in Taizhou and the North China Plain (Wang et al., 2017a, 2018b, 2020a) but close to another result in the YRD before (Chen et al., 2019). The average PM_2.5 was 34.6 ± 17.8 µg m⁻³, with a maximum reach of 163.0 µg m⁻³. The water-soluble particulate components of PM_2.5 are displayed as well. The average NO${}_{\mathrm{3}}^{-}$ concentration was 10.6 µg m⁻³, which accounts for the 38.3 % mass concentration of water-soluble particulate components and 32.0 % total PM_2.5, while the proportions of sulfate, ammonium, and chloride are 26.0 %, 18 %, and 2.0 %, respectively. To sum up, during the campaign period, the pollution of PM_2.5 would be generally exacerbated on high O₃ and NO₂ days. Precipitation occurred during four clean processes and receded pollutant concentration; otherwise, the pollution condition remained severe.

The mean diurnal variations (MDCs) of temperature, RH, NO₂, O₃, P(NO₃), N₂O₅, OH radicals, and PM_2.5 in different air quality are shown in Fig. 3. The temperature, RH, and OH radical MDCs show indistinct differences between clean days (CDs) and polluted days (PDs). The MDC of NO₂ has two concentration peaks that appear at 06:00 and 21:00 on CDs, while on PDs, its peak appears at 20:00 and maintains a high level during the whole night. The O₃ diurnal pattern reflects a typical urban-influenced character with a maximum O₃ peak that lasts 3 h from 14:00 to 17:00, while the PD O₃ peak concentration is 1.2 higher than the CD O₃ peak concentration. P(NO₃) grows after the O₃ peak, and the maximum P(NO₃) shows at 19:00 with an average value of 1.7 ppbv h⁻¹ on CDs. By contrast, the mean PD P(NO3) is 2.6 ppbv h⁻¹, and the maximum value reaches 4.7 ppbv h⁻¹. In contrast, the CD N₂O₅ has a higher average and maximum concentration than PDs, which suggests a faster removal process during PDs. PM_2.5 has a similar trend to P(NO₃) and has a higher concentration during nighttime.

Figure 3 The mean diurnal variations of temperature, RH, NO₂, O₃, PM_2.5, OH radicals (orange), N₂O_5, and P(NO₃) of clean days and polluted days.

3.2 The evolution of nitrate pollution

Figure 4a shows the relationship between nitrate and sulfate with water-soluble particulate components. Nitrate positively correlates with total water-soluble particulate components, while the sulfate ratio has an inverse correlation. With PM_2.5 concentration increasing, nitrate proportion increases rapidly and keeps high weight at a heavy PM_2.5 period, while the sulfate ratio shows the opposite phenomenon. Once the mass concentration of the total water-soluble particulate component is over 30 µg m⁻³, the mass fraction of nitrate in total water-soluble particulate components is up to 50 % on average. This result illustrates that particulate nitrate is one of the vital sources of explosive growth particulate matter.

Figure 4 (a) Particulate ion mass concentration ratio of nitrate and sulfate to water-soluble ion. (b) NOR against RH, colored with temperature. (c) SOR against RH, colored with temperature.

To further assess the conversion capacity of nitrate and sulfate at this site, the sulfur oxidation ratio (SOR) and the nitrogen oxidation ratio (NOR) are used to indicate the secondary transformation ratio of SO₂ and NO₂, respectively (Sun et al., 2006). SOR and NOR are estimated using the formulae below:

$$\begin{array}{}\text{(7)}& {\displaystyle }\text{SOR}={\displaystyle \frac{n{\mathrm{SO}}_{\mathrm{4}}^{\mathrm{2}-}}{n{\mathrm{SO}}_{\mathrm{4}}^{\mathrm{2}-}+n{\mathrm{SO}}_{\mathrm{2}}}},\text{(8)}& {\displaystyle }\text{NOR}={\displaystyle \frac{n{\mathrm{NO}}_{\mathrm{3}}^{-}}{n{\mathrm{NO}}_{\mathrm{3}}^{-}+n{\mathrm{NO}}_{\mathrm{2}}}}.\end{array}$$

n refers to the molar concentration, and the higher SOR and NOR represent more oxidation of gaseous species into a secondary aerosol. As depicted in Fig. 4b–c, NOR rapidly increases at RH < 45 %, remains constant at 45 % < RH < 75 %, and ends with a sharp increase at RH > 75 %. During the study period, not only is the average concentration of NO₂ higher among PDs, but there is also a significant difference between PD and CD NOR. The average values of NOR are 0.32 on PDs and 0.25 on CDs, respectively, which manifests in the more secondary transformation and pollution potential on PDs. In contrast, the SOR stays constant at a high value (∼0.5) during the whole RH scale, which shows a different pattern from previous research (Li et al., 2017; Zheng et al., 2015). One possible explanation is that SO₂ concentration stays low during the whole campaign (4.4 ± 2.4 ppbv on average), and SO₂ oxidation depends on the limit of SO₂ instead of oxidation capability. Meanwhile, the mean SOR in both situations is over 0.5 (0.52 on CDs and 0.56 on PDs), further supporting the SO₂-limited hypothesis. In addition, Table 2 summarizes NOR and SOR values in the YRD. NOR and SOR in this study are similar to values reported in other YRD research (Shu et al., 2019; Y. Zhang et al., 2020; Qin et al., 2021; Zhao et al., 2022), except for values in 2013, but higher than the North China study which emphasizes the solid atmospheric oxidation capacity in the YRD region.

Table 2 Statistical result of NOR and SOR in the YRD.

3.3 The derivation of the N₂O₅ uptake coefficient

Statistical analysis of the observation above highlights the rapid formation of particulate nitrate. To assess the contribution of N₂O₅ hydrolysis to particulate nitrate formation, two methods are applied to calculate the N₂O₅ uptake coefficient. The first method is a stationary-state approximation (Brown et al., 2003). By assuming that the rates of production and loss of N₂O₅ are approximately in balance, the total loss rate of N₂O₅ ($k_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}}$) can be calculated through Eq. (9). $k_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}}$ is mainly dominated by N₂O₅ heterogeneous uptake, since homogeneous hydrolysis of N₂O₅ contributes little (Brown and Stutz, 2012). The N₂O₅ uptake coefficient through the steady state (denoted as γ_S) is derived as Eq. (10). Here C is the mean molecular speed of N₂O₅, and S_a is the aerosol surface concentration.

$$\begin{array}{}\text{(9)}& {\displaystyle }{\mathit{\tau }}_{\mathrm{ss}}\left({\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}\right)={\displaystyle \frac{\left[{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}\right]}{k_{\mathrm{R}\mathrm{3.1}}\left[{\mathrm{NO}}_{\mathrm{2}}\right]\left[{\mathrm{O}}_{\mathrm{3}}\right]}}={\left(k_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}}+{\displaystyle \frac{k_{{\mathrm{NO}}_{\mathrm{3}}}}{K_{\mathrm{eq}}\left[{\mathrm{NO}}_{\mathrm{2}}\right]}}\right)}^{-\mathrm{1}}\text{(10)}& {\displaystyle }{k}_{{\mathrm{N}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{5}}}=\mathrm{0.25}C\mathit{\gamma }\mathit{\_}\text{S}{S}_{\text{a}}\end{array}$$

Due to the fast variety of NO₃ loss rates from VOCs, the steady-state method has been unattainable under conditions affected by emission interferences. During the whole campaign, we only retrieve three valid fitting results. As shown in Fig. 5, the fitted γ_S ranged from 0.057 to 0.123, which is comparable with Taizhou (0.041, Wang et al., 2020a) and much higher than other results in China (C. Yu et al., 2020; Wang et al., 2018a, 2020b, 2017a). The calculated $k_{{\mathrm{NO}}_{\mathrm{3}}}$ ranged from 0.002 to 0.16 s⁻¹, representing drastic VOC change during this campaign.

Figure 5 Derived N₂O₅ uptake coefficients from N₂O₅ steady lifetime (γ_S) with NO₂ and S_a; plots (a–c) represent the linear fitting results on the nights of 30 May, 10 and 11 June, respectively.

The other approach is the parameterization by C. Yu et al. (2020), which is depicted as follows.

$$\begin{array}{}\text{(11)}& \begin{array}{rl}& \mathit{\gamma }\mathit{\_}\text{P}={\displaystyle \frac{\mathrm{4}}{c}}{\displaystyle \frac{V_{\mathrm{a}}}{{\mathrm{S}}_{\mathrm{a}}}}{K}_{\mathrm{H}}\times \mathrm{3.0}\times {\mathrm{10}}^{\mathrm{4}}\times \left[{\mathrm{H}}_{\mathrm{2}}\mathrm{O}\right]\\ & \left(\mathrm{1}-{\displaystyle \frac{\mathrm{1}}{\left(\mathrm{0.033}\times \frac{\left[{\mathrm{H}}_{\mathrm{2}}\mathrm{O}\right]}{\left[{\mathrm{NO}}_{\mathrm{3}}^{-}\right]}\right)+\mathrm{1}+\left(\mathrm{3.4}\times \frac{\left[{\mathrm{Cl}}^{-}\right]}{\left[{\mathrm{NO}}_{\mathrm{3}}^{-}\right]}\right)}}\right)\end{array}\end{array}$$

$V_{\text{a}}/S_{\text{a}}$ is the measured aerosol volume–surface area ratio by the SMPS, K_H is Henry's law coefficient, which is set to 51 as recommended, [NO${}_{\mathrm{3}}^{-}$] and [Cl⁻] are aerosol inorganic concentrations measured by MARGA, and [H₂O] is aerosol water content calculated by ISORROPIA II. The valid parameterization-calculated N₂O₅ uptake coefficient (denoted as γ_P) from 30 May to 8 June 2019 shows in Fig. 6 a good consistency between the trends of γ_P and aerosol water content. Nighttime γ_P varies from 0.001 to 0.024 with an average of 0.069 ± 0.0050 under polluted conditions and 0.0036 ± 0.0026 under clean conditions. The N₂O₅ uptake coefficient shows a good correlation between RH and aerosol water content. For the N₂O₅ uptake coefficient, although particulate nitrate mass concentration increased during the pollution event, an antagonistic effect on the N₂O₅ uptake coefficient was not obvious for the nitrate molarity decrease.

Furthermore, we compare the difference between γ_S and γ_P. Taking the night of 30 May as an example, γ_S is 0.089, while γ_P ranges from 0.024 to 0.057 with an average value of 0.013 ± 0.0051. The difference between steady state and parameterization is significant; one possible explanation is uncertainty for stationary-state approximation caused by local NO or VOC emission (Brown et al., 2009; Chen et al., 2022). Another reason is that the parameterization by C. Yu et al. (2020) ignores the impact of organic matter on the fine particles. The difference in aerosol composition between this work and C. Yu et al. (2020) may also bring uncertainty. Overall, γ_P will be chosen for the N₂O₅ heterogeneous uptake coefficient in later analysis and discussion.

Figure 6 Results of N₂O₅ uptake coefficients through parameterization (γ_P). Panel (a) shows time series of γ_P and ISORROPIA II results of aerosol water content (AWC). Panel (b) is the box plot of the polluted days and clean days, the hollow square represents the mean value, and the solid line across the box shows the median score for the data set, while the top and bottom whiskers represent the 90 % and 10 % values of γ_P, respectively.

3.4 Quantifying the contribution of nitrate formation pathways

After the N₂O₅ uptake coefficient is counted, nitrate production potential P(NO${}_{\mathrm{3}}^{-})$ can be calculated. Here the N₂O₅ uptake coefficient is set to 0.036 on clean days and 0.069 on polluted days, respectively, which are the average values derived from parameterization. The production ratio of NO${}_{\mathrm{3}}^{-}$ (by considering the ClNO₂ yield of 0.54) is set to 1.46 in the former study (Xia et al., 2020). Gas-particle distribution is considered by the result of the particulate nitrate and gas-phase nitrate by MARGA (input ${\mathrm{HNO}}_{\mathrm{3}}/{\mathrm{NO}}_{\mathrm{3}}^{-}$ ratio to the model as the OH + NO₂ nitrate production rate). The NO₂ heterogeneous uptake coefficient is set to 5.8 × 10⁻⁶ depending on the report by Yu et al. (2021), which is the result of 70 % RH on urban grime.

The mean diurnal variations of the nitrate production potential of clean and polluted days are depicted in Fig. 7. The OH + NO₂ pathway shows no significate difference between clean and polluted days and dominates CD nitrate formation potential, since the levels of OH and NO₂ are less affected by the fine particle level. However, the rapid increase in the N₂O₅ heterogeneous uptake pathway on polluted days is fatal, and its peak formation rate at night over the OH + NO₂ pathway can be used to explain nighttime nitrate explosive growth.

As shown in Fig. 7c, OH + NO₂ dominates nitrate production on clean days, while the N₂O₅ uptake pathway only contributes 13.6 µg m⁻³. On polluted days, the ability of N₂O₅ uptake grows quickly, reaching 50.1 µg m⁻³, while the OH pathway does not change much. There is no distinct difference in the daytime pathway (OH +  NO₂) between clean days and polluted days, while the nighttime pathway ratio rises from 38.1 % on clean days to 67.2 % on polluted days. NO₂ heterogeneous uptake increases from 0.93 µg m⁻³ on clean days to 2.0 µg m⁻³ on polluted days, but the contribution proportion does not change obviously. Both the higher N₂O₅ uptake coefficient and the higher S_a on polluted days increase the contribution of N₂O₅ hydrolysis to particulate nitrate under pollution conditions.

Figure 7 The mean diurnal variations of the nitrate production potential of clean days (a) and polluted days (b) and the P(NO${}_{\mathrm{3}}^{-})$ distribution of clean days and polluted days (c).

3.5 Mitigation strategies of particulate nitrate and ozone productions

We selected two pollution episodes (Episode I, 30 May 2019 00:00–2 June 2019 00:00, and IV, 14 June 2019 17:30–17 June 2019 12:00) to explore the mitigation way of ozone and nitrate pollution. Figure 8 shows the EKMA of P(O₃) and P(NO${}_{\mathrm{3}}^{-})$ of these two periods, O₃ located in the VOC-controlling area in the two pollution episodes, which is consistent with a previous YRD urban ozone sensitivity study (Jiang et al., 2018; K. Zhang et al., 2020; Xu et al., 2021). The best precursor reduction for O₃ is VOC : NO_x = 2:1, while nitrate is located in the transition area, which means either of the precursors' reductions will mitigate nitrate pollution. For the regional and complex air pollution characteristics in this region, a fine particle-targeting reduction scheme will aggravate O₃ pollution. In contrast, the O₃-targeting scheme can mitigate O₃ and fine particles simultaneously.

Figure 8 Isogram of P(O₃) and P(NO${}_{\mathrm{3}}^{-}$) of polluted episodes IV (30 May 2019 00:00–6 June 2019 00:00) and I (14 June 2019 17:30–17 June 2019 12:00) with different NO_x and VOC reduction degrees. The grey dashed line represents the ridge line.

4 Conclusions

A comprehensive campaign was conducted to interpret the atmospheric oxidation capacity and aerosol formation from 30 May to 18 June 2019 in Changzhou, China. The high O₃ and PM_2.5 concentrations confirm complex air pollution characteristics in Changzhou, and nitrate accounts for 38.3 % mass concentration of total water-soluble particulate components and 32.0 % of total PM_2.5. In addition, the average values of NOR are 0.32 on PDs and 0.25 on CDs. The positive correlation between NOR and RH and the inverse correlation refer to the contribution of N₂O₅ heterogeneous uptake to nitrate formation.

Based on field observations of OH and related parameters, we show that OH oxidation of the NO₂ pathway steadily contributes to nitrate formation no matter the clean or polluted period and dominates CD nitrate production (about 22 µg m⁻³). N₂O₅ heterogeneous uptake contribution proliferated on polluted days, from 13.6 µg m⁻³ (38.1 %) on clean days to 50.1 µg m⁻³ (67.2 %) on polluted days. NO₂ heterogeneous uptake contributes little to nitrate formation (2.6 %).

The precursor reduction simulation suggests the reduction ratio of VOC : NO_x equaling 2:1 can simultaneously and effectively mitigate O₃ and fine particle pollution during the summertime complex pollution period in Changzhou. To more precisely and delicately establish a cooperative control scheme for regional O₃ and nitrate, the regional and long-time field campaigns are needed in the future to analyze the seasonal and interannual variation of O₃ and nitrate and relevant parameters.