2.1 Sample collection

As part of the bilateral Sweden–China framework research program on “Photochemical smog in China”, an intensive field campaign was conducted in Beijing, aiming to improve the understanding of secondary chemistry during photochemical smog events in China (Hallquist et al., 2016). The campaign was conducted at Changping (40.14^∘ N, 116.11^∘ E), a regional site northeast of the Beijing urban area, from 15 May to 5 June 2016. During this period, the site was influenced by anthropogenic pollutants from Beijing urban areas and under high-NO_x conditions, as suggested by field measurement evidence reported in previous publications related to this campaign (Tang et al., 2018; H. C. Wang et al., 2018). During 17 May–5 June, the daily average concentrations of benzene, toluene, and NO_x were 66–922, 47–1344, and 4.0–32.3 ppb, respectively.

Day and night ambient PM_2.5 (particles with aerodynamic diameter less than 2.5 µm) samples were collected on prebaked quartz fiber filters (Whatman Inc.) and Teflon filters (Whatman Inc.) using a high-volume sampler (TH-1000C, Tianhong, China) and a four-channel sampler (TH-16A, Tianhong, China). The sampling flow rates were 1.05 m³ min⁻¹ and 16.7 L min⁻¹, respectively. The daytime samples were collected from 08:30 to 17:30 LT (UTC+8) and the nighttime ones from 18:00 to 08:00 LT (UTC+8) the next morning. Field blank samples were collected by placing filters in the samplers with the pump off for 30 min.

2.2 Quantification of NACs

An aliquot of 25 cm² was removed from each quartz fiber filter sample and extracted in an ultrasonic bath three times using 3, 2, and 1 mL methanol containing 30 µL saturated EDTA solution in methanol–acetic acid consecutively, each time for 30 min. The extracts were then filtered through a 0.25 µm polytetrafluoroethylene (PTFE) syringe filter (Pall Life Sciences), combined, and evaporated to dryness under a gentle stream of high-purity nitrogen. The dried samples were re-dissolved in 50 µL methanol ∕ water (1:1) containing 100 ppb 4-nitrophenol-2,3,5,6-d₄ as the internal standard. The solution was centrifuged and the supernatant was used for analysis, using the Agilent 1260 LC system (Palo Alto, CA) coupled to the QTRAP 4500 (AB Sciex, Toronto, Ontario, Canada) mass spectrometer. The LC-MS system was equipped with an electrospray ionization (ESI) source operated in negative mode. More details of the extraction and optimized MS parameters have been given in our previous study (Chow et al., 2016).

Chromatographic separation was performed on an Acquity UPLC HSS T3 column (2.1 mm × 100 mm, 1.8 µm particle size; Waters, USA) with a guard column (HSS T3, 1.8 µm). The column temperature was kept at 45 ^∘C and the injection volume was 5.0 µL. The mobile eluents were (a) water containing 0.1 % acetic acid (v∕v) and (b) methanol (v∕v) containing 0.1 % acetic acid at a flow rate of 0.19 mL min⁻¹. The gradient elution was set as follows: started with 1 % (b) for 2.7 min; increased to 54 % (b) within 12.5 min and held for 1.0 min; then increased to 90 % (b) within 7.5 min and held for 0.2 min; and finally decreased to 1 % (b) within 1.8 min and held for 17.3 min until the column was equilibrated. Chromatograms of NAC standards and an ambient sample are shown in Fig. S1 in the Supplement.

The quantified NAC species are listed in Table 1. The NACs were identified and quantified using the [M-H]⁻ ions in the extracted ion chromatogram (EIC), using authentic standards or surrogates with the same molecular formula (Table 1). The standards included 4-nitrocatechol (4NC), 4-nitrophenol (4NP), 2-methyl-4-nitrophenol (2M4NP), 3-methyl-4-nitrophenol (3M4NP), and 2,6-dimethyl-4-nitrophenol (2,6DM4NP) from Sigma-Aldrich (St. Louis, MO, USA); and 4-methyl-5-nitrocatechol (4M5NC) from Santa Cruz Biotech (Dallas, TX, USA). The recoveries of the target NACs were 91 %–106 %. 4M5NC was employed as a surrogate standard to quantify 3M5NC and 3M6NC. However, a recent study suggested that no 3M6NC could be detected in ambient aerosols and the MNC isomer could be an incorrect assignment of 3M4NC as 3M6NC (Frka et al., 2016). We cannot exclude the possibility of the MNC isomer as 3M4NC due to a lack of authentic standards. Employing 4M5NC as a surrogate standard, the concentrations of 3M6NC could be obviously underestimated due to its poor ionization under ESI conditions compared with that of 4M5NC (Frka et al., 2016). The concentration of dimethyl-nitrophenol (DMNP) was the sum of three isomers. 2,6DM4NP was identified based on its retention time matching that of the authentic standard (Fig. S1), while we cannot exclude the possibility of the other two DMNP isomers as ethylnitrophenols or methoxylated isomers.

Table 1The quantified nitro-aromatic compounds in this study.

Download Print Version

2.3 Other online and offline measurements

Other online and offline instruments were also employed to obtain a related database, which has been introduced in detail in our previous paper (Y. Wang et al., 2018). In brief, a high-resolution time-of-flight aerosol mass spectrometer (AMS) was used to measure the chemical composition of PM₁ (Zheng et al., 2017). The aerosol surface area was calculated based on the measurements of particle number and size distribution by a scanning mobility particle sizer (SMPS, TSI 3936) and an aerosol particle sizer (APS, TSI 3321) (Yue et al., 2009; H. C. Wang et al., 2018). VOCs were measured by a proton-transfer-reaction mass spectrometer (PTR-MS). Gaseous NH₃ was measured using a NH₃ analyzer (G2103, Picarro, California, USA) (Huo et al., 2015). Meteorological parameters, including relative humidity (RH), temperature, wind direction, and wind speed (WS), were continuously monitored by a weather station (Met one Instrument Inc.) during the whole campaign.

Organic carbon (OC) and element carbon (EC) were measured on the quartz fiber filter samples using a thermal/optical carbon analyzer (Sunset Laboratory). The organic matter (OM) concentration was calculated by multiplying OC by 1.6 (Turpin and Lim, 2001). The Teflon filter samples were used to quantify the water-soluble inorganic ions by an ion chromatograph (IC, DIONEX, ICS2500/ICS2000) following procedures described in Guo et al. (2010). Aerosol acidity and liquid water content (ALWC) were then calculated using the ISORROPIA-II thermodynamic model. ISORROPIA-II was operated in forward mode, assuming the particles are “metastable” (Hennigan et al., 2015; Weber et al., 2016; Guo et al., 2015). The input parameters included ambient RH, temperature, particle-phase inorganic species (${\mathrm{SO}}_{\mathrm{4}}^{\mathrm{2}-}$, ${\mathrm{NO}}_{\mathrm{3}}^{-}$, Cl⁻, ${\mathrm{NH}}_{\mathrm{4}}^{+}$, K⁺, Na⁺, Ca²⁺, Mg²⁺), and gaseous NH₃. More details and validation of the thermodynamic calculations have been described in our previous paper (Y. Wang et al., 2018).

2.4 Estimation of the gas-phase NACs

The concentrations of gas-phase NACs were not measured in this study. They were calculated based on the measured particle-phase NAC concentrations and equilibrium absorption partitioning theory (Pankow, 1994a, b; Pankow et al., 2001) (Eqs. 1 and 2):

where F_p is the fraction of NACs in the particle phase. C_OA is the concentration of organic aerosols (OA), calculated as OC multiplied by 1.6. c_g and c_p are the concentrations of NACs in the gas phase and particle phase, respectively. C^∗ is the effective saturation mass concentration (µg m⁻³) and is calculated using Eq. (2):

where M is the molecular weight of NACs (g mol⁻¹). ζ is the activity coefficient of the species (assumed = 1). R is the gas constant (8.314 J (mol K)⁻¹), T is the temperature (K), and P_v (Pa) is the saturation pressure. P_v at the average temperature during the campaign (296 K) is calculated using the multiphase system online property prediction tool developed by the University of Manchester (UManSysProp, http://umansysprop.seaes.manchester.ac.uk, last access: 31 May 2019). The vapor pressures were estimated using the Nannoolal approach (Nannoolal et al., 2008), and the boiling points were estimated using the Joback and Reid approach (Joback and Reid, 1987).

The estimated P_v, F_p, and gas-phase concentrations of NACs are listed in Table S1 in the Supplement. 4NP and methyl-nitrophenols (2M4NP and 3M4NP) were predicted to be mainly in the gas phase (F_p < 10 %), while DMNP, 4NC, and MNC (3M6NC, 3M5NC, and 4M5NC) were mainly in the particle phase (F_p > 60 %). The gas-phase DMNP and MNC (F_p > 95 %) will not be further discussed in this study. While the equilibrium model gives reasonable estimation of F_p and gas-phase concentrations for nitrocatechols, it overestimates the vapor pressure of NPs by several orders of magnitude (Bannan et al., 2017). The estimated F_p(0.83 %) was obviously lower than the measured values for 4NP. For example, Cecinato et al. (2005) measured F_p of 4NP and 3M4NP to be 82 % and 78 %, respectively, in downtown Rome; Le Breton et al. (2018) reported F_p of nitrophenol at ∼17 % using the Chemical-Ionization Mass Spectrometer (CIMS) coupled with the Filter Inlet for Gases and AEROsols (FIGAERO) during this campaign. We note that CIMS could not distinguish the isomers (e.g., 2NP) of 4NP; however, the measured F_p values showed us the range of particulate fractions of 4NP during the campaign. The equilibrium absorption partitioning model could underestimate the F_p of 4NP by ∼20 times. Thus, the gas-phase 4NP concentration was roughly calculated using the measured F_p (17 %) by FIGAERO-CIMS (Le Breton et al., 2018).

Gas-phase NACs could also dissolve into the aqueous-phase particles. The concentrations dissolved into the aqueous phase (C_aq) were estimated by Henry's law (Sander, 2015). Henry constants were obtained from Sander (2015) and ALWC was estimated using ISORROPIA-II (see Sect. 2.4). The estimated C_aq of 4NP and 3M4NP were $\mathrm{4.4}\times {\mathrm{10}}^{-\mathrm{4}}$ and $\mathrm{2.4}\times {\mathrm{10}}^{-\mathrm{5}}$ ng m⁻³, contributing < 0.02 % to their concentrations in the particle phase. The contribution of dissolution into aqueous-phase particles for NC and MNC is expected to be lower, due to the much lower gas-phase concentrations than that of 4NP. For this reason, we will not further consider the dissolution of NACs into the aqueous phase of particles.

3.1 Concentration and composition of NACs

The average concentration of quantified NACs was 6.63 ng m⁻³, ranging from 1.27 to 17.70 ng m⁻³ in summer in Beijing. Figure 2 compares the total NAC concentrations across this and prior studies, and the individual NAC concentrations are compared in Fig. S2 and Table S2. The total NAC concentration in this work was higher than those measured in other studies conducted in summer in mountain, rural, or urban environments (Teich et al., 2017; Kitanovski et al., 2012; Kahnt et al., 2013; Zhang et al., 2013; Chow et al., 2016; L. Wang et al., 2018), and comparable to those reported in the studies in summertime Wangdu, China (Teich et al., 2017; L. Wang et al., 2018) (Fig. 2). Most NAC species (NC, MNP, and MNC), except for DMNP and NP, also showed elevated concentrations in Changping compared with those reported in other summertime studies (Fig. S2). Influenced by the outflow from urban Beijing air masses, the site was under typical high-NO_x conditions (H. C. Wang et al., 2018), implying abundant potential secondary formation of NACs during the observation period. A recent study suggested that nocturnal biogenic VOC (BVOC) oxidation would transfer from low- to high-NO_x regimes and that nearly all the BVOCs would be oxidized by NO₃ radicals, at a NO_x/BVOCs ratio higher than 1.4 (Edwards et al., 2017). If we approximate the BVOC concentrations to be the sum of isoprene, MVK+MACR (methyl vinyl ketone and methacrolein), and monoterpenes, the NO_x∕BVOC ratios were higher than 8 (nighttime ratios higher than 20) (Fig. S3). If we further consider the major anthropogenic VOCs (toluene, benzene), NO_x/VOCs ratios were higher than 5 (nighttime ratios higher than 10) (Fig. S3). The high-NO_x conditions during the campaign were expected to facilitate the oxidation of aromatic hydrocarbons and the subsequent secondary formation of NACs. Other emissions from biomass burning and coal combustion were also observed to be contributors of organic aerosols during the campaign (Tang et al., 2018), and they could also be the precursor sources of NACs. Biomass burning episodes occurred during the Wangdu campaign, indicating NAC emissions from biomass burning (Teich et al., 2017; Tham et al., 2016), which explains the high NAC levels in summer in Wangdu. The NAC concentrations during summer (including this study) are generally lower than those during spring, autumn, or winter, which could be due to stronger contributions from combustion sources (e.g., biomass burning and coal combustion) during spring, autumn, or winter than those during summer (Chow et al., 2016; L. Wang et al., 2018; Kitanovski et al., 2012; Kahnt et al., 2013).

Figure 2Summary of NAC concentrations across this and prior studies (see Table S2 for the data and references therein). The NAC concentrations in summer correspond to the left axis and other seasons correspond to the right axis.

Download

The NAC compositions are shown in the inserted pie chart in Fig. S2. 4-Nitrophenol and 4-nitrocatechol were the most abundant ones among all the quantified NAC species, accounting for 32.4 % and 28.5 % of the total quantified NACs, followed by methyl-nitrocatechols (4M5NC, 3M5NC, and 3M6NC, 16.2 %), methyl-nitrophenol (2M4NP and 3M4NP, 15.6 %) and dimethyl-nitrophenol (8.3 %) (Table 1). The contribution of NP and NC could be larger when considering both gas and particle phases. The average concentration of 4NC in both gas and particle phases was estimated to be 2.2ng m⁻³ using the equilibrium absorption partitioning model. The total concentration of 4NP (13 ng m⁻³) in both the gas and particle phases was approximated using the measured F_p (17 %) by FIGAERO-CIMS (Le Breton et al., 2018). Nitrophenols and nitrocatechols were generally reported among the most abundant NAC species in previous studies (Table S2 and the references therein). Nitrophenols could be formed via the oxidation of anthropogenic VOCs (e.g., benzene) in the presence of NO₂, and nitrocatechols were found to dominate the composition of NAC products formed in the benzene–NO_x system in laboratory studies (Xie et al., 2017; Harrison et al., 2005; Yuan et al., 2016; Sato et al., 2007; Ji et al., 2017; Olariu et al., 2002). Thus, it is not a surprise to observe the high concentrations of nitrophenol and nitrocatechol in the typical high-NO_x and anthropogenic VOC dominated environments in summer in Beijing.

The contribution of NP among the total NACs at Changping was higher than that in summer in Hong Kong, while that of MNC was lower (Table S2 and the inserted pie charts in Fig. S2). This NAC composition difference between Changping and Hong Kong may be a result of different formation pathways for NPs and MNC and different environmental conditions at the two sites. The gas-phase oxidation of aromatic hydrocarbons (e.g., phenol, benzene) in the presence of NO₂ is a major source of NPs (Harrison et al., 2005; Yuan et al., 2016; Sato et al., 2007; Ji et al., 2017; Olariu et al., 2002), while aqueous-phase oxidation represents the important formation pathway for atmospheric MNC (Frka et al., 2016; Vidovic et al., 2018). The ambient RH in Hong Kong (> 70 %) was significantly higher than that in summer in Beijing (5 %–81 %, 37 % on average), and thus the relative contribution of aqueous-phase pathways could be more dominant in Hong Kong, promoting the aqueous-phase formation of MNC. The influence of ambient RH on NAC formation will be further discussed in Sect. 3.4. In comparison, more abundant gas-phase formation of nitrophenol was expected in summer in Beijing, under higher anthropogenic VOCs and high-NO_x and low-RH conditions. In addition, the lower temperature in summer in Changping was more favorable for the partitioning of nitrophenols from gas phase into particle phase.

Figure 3Time series of (a) wind speed (WS) and relative humidity (RH), (b) benzene and toluene, mass concentrations of (c) K⁺, (d) organics and nitrate, and (e) NACs. The pollution episodes, with elevated organic aerosols, are marked by gray shading.

Download

3.2 Temporal variations and sources of NACs

Temporal variations of the total quantified NAC concentrations are shown in Fig. 3, along with particulate organics, nitrate, potassium ion, toluene, benzene, acetonitrile, wind speed, and RH. During the field campaign, four pollution episodes (episodes I, II, III, and IV), marked by gray shading in Fig. 3, were identified through observation of elevated organic aerosols. Elevated NAC concentrations were observed during pollution episodes, coinciding with the increasing of toluene, benzene, acetonitrile, and potassium. The correlations between NACs and other chemical components are shown in Table S3. The potassium ion was employed to indicate particulate emissions from biomass burning. As the biomass burning-derived immediate VOC precursors to NACs were not detected in this study, acetonitrile was used to track the variations of VOCs released by biomass burning. It was noticed that NACs showed stronger correlations with toluene ($r={\mathrm{0.70}}^{**}$), benzene ($r={\mathrm{0.64}}^{**}$), or acetonitrile ($r={\mathrm{0.61}}^{**}$) than with potassium ($r={\mathrm{0.49}}^{**}$). The “${}^{**}$” following the numerical value denotes significant correlation at the 0.01 level. This appeared to suggest that the NO_x oxidation of anthropogenic VOCs and precursor VOCs from biomass burning was a more important source of NACs than primary biomass burning emission in summer in Beijing. A lower correlation between particulate NACs and EC (Table S3, $r={\mathrm{0.39}}^{**}$) was also in agreement with the suggestion of the lesser importance of primary emissions to NACs during the campaign. We note that only particulate NAC concentrations were used to do the correlation analysis. Two atmospheric processes, namely photolysis and gas-to-particle partitioning, could influence the abundance of particle-phase NACs, especially for NP and MNPs, since the majority of them were expected to be in the gas phase (Table S1). As such, correlations of particle-phase NP and MNPs with other species may less reliably reflect the underlying associations with the correlation species. As for the relative importance of anthropogenic VOCs and biomass burning-derived VOCs, we do not have direct field measurement data for the differentiation. However, previous studies suggested that the sources of anthropogenic VOCs in summer in Beijing were dominated by vehicle emissions (> 50 %), with minor contributions from solvent evaporation and biomass burning (Wang et al., 2014; Liu et al., 2008). Modelling studies incorporating emission inventories of the relevant VOC precursors could address this issue and are suggested in future investigation of NAC sources. We note that biomass burning could often be of an anthropogenic origin. Within this work, the term “anthropogenic VOCs” does not include VOCs from human-caused biomass burning activities.

To further investigate the formation of NACs, we examined the time series and day–night variations of individual NAC species (Figs. 4, S4, and S5). Daytime enhancements of 4NP, 2M4NP, and DMNP, and nighttime enhancements of 3M4NC and 4M5NC, were observed, and other NAC species did not show discernible day–night variations (Figs. 4, S4, and S5), indicating different formation pathways among NAC species. Good inter-species correlations were observed among nitrophenol and its derivatives (2M4NP, 3M4NP, DMNP, r=0.56–0.88), as well as among nitrocatechol and its derivatives (3M6NC, 3M5NC, 4M5NC, r=0.49–0.84). This showed that the formation and loss pathways as well as the influence factors were likely similar within NP and NC groups. In comparison, the correlations of NACs across the two groups, i.e., between nitrophenol derivatives (MNP, DMNP) and nitrocatechol derivatives (MNC, r=0.05–0.45), were lower (Table S3), suggesting different formation pathways and influence factors. NC and its derivatives showed stronger correlations with toluene, benzene, acetonitrile, and K⁺, compared with NP and its derivatives (Table S3). This was more likely associated with the fact that particle-phase NPs only account for a minor part of the atmospheric NP abundance due to the high vapor pressure of NPs (Table S1). The abundance of particulate NP could largely depend on gas-to-particle partitioning, which is strongly affected by temperature, as well as their gas-phase loss pathways (e.g., photolysis) (Bejan et al., 2007; Yuan et al., 2016; Sangwan and Zhu, 2018). NC and MNC were mainly present in the particle phase (Table S1). The oxidation degradation rates and photolysis of NC and MNC were therefore much lower. A recent laboratory study found that OH uptake by MNC particles was suppressed by a factor of 4 at RH 15 %–30 % in comparison with the dry condition, as a result of competitive co-adsorption of water molecules that occupied reactive sites (Slade and Knopf, 2014). During the campaign, the ambient RH was 37 %. Such an RH condition rendered the OH uptake by particles suppressed, and therefore heterogeneous oxidation of MNC was likely not important.

Figure 4Time series of (a) NO_x, (b) J(O¹D), (c) 4-methyl-5-nitrocatechol (4M5NC), (d) 4-methyl-5-nitrocatechol (3M5NC), (e) 4-nitrophenol (4NP), (f) 2-methyl-4-nitrophenol (2M4NP), and (g) dimethyl-nitrophenol (DMNP). The gray background denotes the nighttime and the white background denotes the daytime.

Download

Nighttime enhancements of 4M5NC and 3M5NC were observed during the whole observation period (Fig. 4). A strong correlation between 4M5NC and 3M5NC and their similar temporal variations likely indicated shared similarity in their formation pathways. Previous studies suggested that aqueous-phase oxidation (including photooxidation and nighttime oxidation) is an important formation pathway for atmospheric MNC, especially in polluted high-NO_x environments and relatively acidic particles (pH around 3) (Vidovic et al., 2018; Frka et al., 2016). 4M5NC and 3M5NC showed relatively stronger correlations with RH compared with other NAC species (Table S3), implying the importance of water in their formation processes and the aqueous-phase pathway. During the campaign, the acidic particles (a pH in the range of 2.0–3.7) and the high-NO_x conditions (H. C. Wang et al., 2018; Y. Wang et al., 2018) provided suitable environments for the aqueous-phase oxidation formation of MNC. The nighttime enhancements of 4M5NC and 3M5NC were more obvious during episode I than during episodes II–IV (nighttime/daytime concentrations at 1.9–3.1 vs. 0.9–1.5) (Fig. 4), which suggested that nighttime aqueous-phase formation pathways played more important roles during the first episode. The daytime correlations between 4M5NC or 3M5NC and RH or NO₂ were stronger than the nighttime ones (Table S4). The aqueous-phase NO_x oxidation could be more dependent on ambient RH and NO₂ levels during the daytime, due to the lower RH and NO₂ concentrations than those at night (Figs. 3, S3). MNCs also showed good correlations with acetonitrile and potassium, as MNCs could also be formed via the oxidation of biomass burning-derived VOC precursors (e.g., cresol) (Iinuma et al., 2010; Finewax et al., 2018; Olariu et al., 2002). 3M6NC (or 3M4NC isomer) showed different temporal variations from 4M5NC or 3M5NC (Figs. 4, S4) and their correlations were lower than that between 4M5NC and 3M5NC (Tables S3, S4), possibly suggesting different formation pathway for 3M6NC (or 3M4NC isomer) from those of 4M5NC or 3M5NC. Quantum calculations have predicted the formation of 3M5NC via aqueous-phase electrophilic substitution and nitration by ${\mathrm{NO}}_{\mathrm{2}}^{+}$, while the formation of 3M6NC was negligible due to higher activation barriers for nitration of 3-methylcatechol to form 3M6NC (Frka et al., 2016). A dominant presence of 3M5NC in ambient aerosols was also expected according to theoretical predictions (Frka et al., 2016). The 3M5NC concentration was higher than that of 3M6NC in summer in Beijing, consistent with the suggestion from computation study by Frka et al. (2016).

Different from the nighttime enhancements of 4M5NC and 3M5NC, 4NP, 2M4NP and DMNP showed daytime enhancements during the whole campaign (Figs. 4, S5). Previously, Yuan et al. (2016) also suggested the daytime gas-phase oxidation of aromatics could represent the major source of NPs, while the contribution from nighttime reaction of phenol with NO₃ radicals was relatively lower (Yuan et al., 2016). The daytime enhancements of NP and its derivatives (2M4NP, DMNP) were more prominent during episodes II–IV than episode I (daytime/nighttime concentrations at 3.1–4.5 vs. 1.8–2.0) (Fig. 4), which indicated that gas-phase photochemical oxidation plays a more important role during the later period of the campaign. We did not find good correlation between 4NP and NO₂ when considering the whole campaign period (Table S3), while good correlations were observed when treating the daytime and nighttime conditions separately (Table S4). The strong correlations between 4NP and benzene, toluene or NO₂ during daytime and nighttime indicated its formation via oxidation of benzene and toluene in the presence of NO₂ (Table S4). The formation mechanisms of nitrophenol were different during daytime (OH-initiated photooxidation of aromatics in the presence of NO₂) and nighttime (NO₃-initiated oxidation of aromatics) (Harrison et al., 2005; Yuan et al., 2016; Sato et al., 2007; Ji et al., 2017; Olariu et al., 2002), thus the role and influence of NO₂ on NAC formation were different. For DMNP, 2M4NP and 3M4NP, they also showed good correlations with benzene, toluene and NO₂ during daytime, but the correlations were absent at night. Instead, their correlations with RH were higher at night, implying the possible formation via aqueous-phase pathways.

3.3 The NO₂ control of NAC formation

The analysis in Sect. 3.2 suggests that NO_x oxidation of anthropogenic VOC precursors represented the dominant sources of NACs in summer in Beijing. To further investigate the impacts of NO₂ on NAC secondary formation, we plot the concentrations of NACs, nitrate (${\mathrm{NO}}_{\mathrm{3}}^{-}$), and the ${\mathrm{NO}}_{\mathrm{3}}^{-}/\mathrm{NAC}$ ratios as a function of NO₂ levels (Fig. 5). The variation of $\left({\mathrm{NO}}_{\mathrm{3}}^{-}\right)/\mathrm{NACs}$ ratios was employed to illustrate the relative abundance of inorganic nitrate and oxidized organic nitrogen. The variations during daytime and nighttime were separately considered due to the different atmospheric conditions and oxidation mechanisms.

Figure 5Concentration of NACs, nitrate and ${\mathrm{NO}}_{\mathrm{3}}^{-}/\mathrm{NAC}$ ratios as a function of NO₂ concentration bins during daytime and nighttime. The markers represent the mean values and whiskers represent the 25th and 75th percentiles.

Download

Generally, higher concentrations of NACs and nitrate were observed with elevated NO₂ concentration levels, in a nonlinear fashion (Fig. 5). During the daytime, NACs increased with NO₂, and ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations and $\left({\mathrm{NO}}_{\mathrm{3}}^{-}\right)/\mathrm{NACs}$ ratios were lower under low-NO_x conditions (NO₂ < 20 ppb). As NO₂ increased to higher than 20 ppb, NAC concentrations did not increase with NO₂ anymore, signaling the transition from NO_x-sensitive to NO_x-saturated regimes for NAC secondary formation. At the same time, the ${\mathrm{NO}}_{\mathrm{3}}^{-}$ concentrations and $\left({\mathrm{NO}}_{\mathrm{3}}^{-}\right)/\mathrm{NACs}$ ratios showed increasing trends compared with those under low-NO_x conditions (NO₂ < 20 ppb) (Fig. 5a, b, c). It was likely that the daytime NO₂ was in excess for the oxidation of ambient VOCs and the NAC formation at NO₂ > 20 ppb. Then the excess NO₂ would be oxidized to form inorganic nitrate, producing a shift of products from organic- to inorganic-dominated conditions. Similarly, during nighttime a transition was observed at NO₂ ∼25 ppb in which oxidation products were shifted from organic to inorganic dominance (Fig. 5d, e, f). At NO₂ > 25 ppb, the nighttime NAC formation became independent of NO₂ concentrations and inorganic nitrate dominated the NO_x oxidation products. The simplified mechanisms and schematic diagram of the competing formation of inorganic nitrates and NACs are shown in Fig. S6. The nighttime NO₂ transition value (∼25 ppb) was higher than the daytime one (∼20 ppb). The higher concentrations of anthropogenic VOC precursors (Fig. S3) and different oxidation mechanisms (Fig. 1) were the potential reasons for elevated NO₂ transition values at night.

The compositional variation of inorganic nitrate and NACs described in this work serves as an example in illustrating that the transition from low- to high-NO_x regimes and the corresponding oxidation products shifting from organic- to inorganic-dominated conditions exist in polluted urban atmospheres that are characterized by high NO_x and anthropogenic VOCs. However, the mechanisms as well as transition thresholds were less understood compared with the well-known BVOC–NO_x atmospheres. More comprehensive investigation in urban atmospheres is needed to develop more quantitative understanding of the NO_x regime transition. As only a limited number of VOC species were measured in this study, the NO_x regime transition value was expressed by NO₂ concentrations rather than NO₂∕VOC or NO_x∕VOC ratios. We also note that the NO_x regime transition values in other atmospheres could be quite different. The NO_x regime transition values deserve further investigation through comprehensive lab simulation and field observations to seek a more robust parameter that can be applied to various atmospheric environments.

The analysis in the previous section indicates that the formation pathways of different NAC species vary from each other; thus the role and influence of NO₂ on their formation are different. The NAC compositions under similar NO₂ concentration levels were averaged, with a bin size of 10 ppb NO₂. The variation of NAC compositions as a function of NO₂ levels is shown in Fig. 6 to investigate the influence of NO₂ on NAC compositions. The contributions of NCs (standard deviation < 12 % within each NO₂ bin) increased and those of NPs (standard deviation < 12 % within each NO₂ bin) decreased at elevated NO₂ concentrations. The NAC composition remained relatively constant at NO₂ > 20 ppb, which was approximately the transition value from low- to high-NO_x regimes. The role of elevated NO₂ in promoting formation of NCs was more obvious than that for NPs. The oxidation of aromatics (e.g., benzene, toluene and VOCs emitted from biomass burning) in the presence of NO₂ represents the major formation pathway of NCs. The formation of NCs would increase with increasing of ambient NO₂. As particle-phase NP and MNP were strongly dependent on the gas-to-particle partitioning and gas-phase loss (e.g., photolysis), their increasing trends as a function of NO₂ were not as obvious as those of NC and MNC.

Figure 6Variation of NAC composition as a function of NO₂ concentration bins.

Download

Figure 7Concentration of NACs as a function of toluene and benzene concentration bins during daytime and nighttime. The markers represent the mean values and whiskers represent the 25th and 75th percentiles.

Download

3.4 Other influence factors on NAC formation

Nitration of aromatic hydrocarbons (e.g., benzene and toluene) represents the major source of NACs in summer in Beijing. NACs generally increased with the increasing of anthropogenic toluene and benzene (Fig. 7). During daytime, when toluene was higher than 0.6 ppb and benzene higher than 0.4 ppb, the NAC concentrations did not increase further with VOC concentrations (Fig. 7a, b). It was likely that toluene or benzene was in excess and that the NAC formation became independent of these precursors. Similarly, the nighttime formation of NACs would become insensitive to these precursors when toluene was higher than 1 ppb and benzene higher than 0.6 ppb (Fig. 7c, d). The transition value of toluene or benzene was higher at night than during the daytime. This could be due to the significantly higher NO₂ levels (significant at the p=0.01 level) (Fig. S3), with higher capacity to oxidize VOC precursors and different oxidation mechanisms at night.

Figure 8Correlation analysis (a) between (3M5NC+4M5NC) and RH, (b) between NACs and aerosol surface area, and (c) between NACs and J(O¹D).

Download

Figure 9(3M5NC+4M5NC) ∕ 4NP concentration ratio as a function of ambient RH during (a) daytime and (b) nighttime.

Download

Though the total NACs did not show good correlations with ambient RH, good correlations between 3M4NC, 4M5NC, and RH were observed (Table S3, Fig. 8). Nitrophenols and methyl-nitrophenols, dominated by gas-phase formation pathways, were less affected by ambient RH. Aqueous-phase oxidation represented the major formation pathway of 3M4NC and 4M5NC during the campaign, based on the analysis in Sect. 3.2 and previous studies (Vidovic et al., 2018; Frka et al., 2016). Elevated ambient RH would favor the water uptake of aerosols and decrease the aerosol viscosity, which favors the uptake of organic precursors or other gas molecules into the particles, mass diffusion of reactants, and chemical reactions within the particles (Vaden et al., 2011; Booth et al., 2014; Renbaum-Wolff et al., 2013; Shrestha et al., 2015; Zhang et al., 2015), and thereby enhance the formation of 3M4NC and 4M5NC in aqueous phase.

The (3M4NC+4M5NC) ∕ 4NP mass concentration ratios were employed to indicate the relative contribution of aqueous-phase and gas-phase pathways to NAC formation. The variations of (3M4NC+4M5NC) ∕ 4NP ratios as a function of ambient RH during daytime and nighttime are shown in Fig. 9. During daytime, this ratio increased with RH when RH < 30 %, indicating elevated contribution of aqueous-phase pathways to NAC formation with higher RH conditions. The ratio remained stable at RH > 30 % during both daytime and nighttime, suggesting the relative contribution of aqueous-phase and gas-phase pathways would not increase further with increasing RH beyond RH > 30 % (Fig. 9a and b). The ratio during the nighttime was obviously higher than during the daytime, indicating that the aqueous-phase oxidation played more important roles for NAC formation at night. The results implied the importance of aqueous-phase oxidation for the secondary formation of oxidized organic nitrogen at elevated ambient RH. Due to the limited sample number obtained by filter-based analysis in this study, the influence of RH or aerosol liquid water content on NAC formation needs further confirmative investigation using controlled laboratory studies.

The NAC concentrations also showed good correlations with aerosol surface area (Fig. 8b). Higher aerosol surface area would facilitate the partitioning of gas-phase NAC products or precursors into particle phase and the aqueous phase or heterogeneous oxidation processes (Kroflic et al., 2015; Bauer et al., 2004; Fenter et al., 1996; Vidovic et al., 2018). Photolysis is an important loss pathway of NACs and could be the dominant sink for nitrophenols in the gas phase (Bejan et al., 2007; Yuan et al., 2016). The highest value of J(O¹D) of each day was used to roughly represent the photolysis intensity. The daytime NAC concentrations showed negative correlations with J(O¹D) (Fig. 8c, Table S4), suggesting photolysis as an important sink for NACs during the daytime.