Characterization of isoprene-derived secondary organic aerosols at a rural site in North China Plain with implications for anthropogenic pollution effects

Isoprene is the most abundant non-methane volatile organic compound (VOC) and the largest contributor to secondary organic aerosol (SOA) burden on a global scale. In order to examine the influence of high concentrations of anthropogenic pollutants on isoprene-derived SOA (SOAi) formation, summertime PM2.5 filter samples were collected with a three-hour sampling interval at a rural site in the North China Plain (NCP), and determined for SOAi tracers and other chemical species. RO2+NO pathway derived 2-methylglyceric acid presented a relatively higher contribution to the SOAi due to the high-NOx (~20 ppb) conditions in the NCP that suppressed the reactive uptake of RO2+HO2 reaction derived isoprene epoxydiols. Compared to particle acidity and water content, sulfate plays a dominant role in the heterogeneous formation process of SOAi. Diurnal variation and correlation of 2-methyltetrols with ozone suggested an important effect of isoprene ozonolysis on SOAi formation. SOAi increased linearly with levoglucosan during June 10–18, which can be attributed to an increasing emission of isoprene caused by the field burning of wheat straw and a favorable aqueous SOA formation during the aging process of the biomass burning plume. Our results suggested that isoprene oxidation is highly influenced by intensive anthropogenic activities in the NCP.

which could also be an important contributor to isoprene-derived SOA (SOA i ) in the atmosphere 18,19 . Moreover, many studies have found that relative humidity (RH) 20,21 , aerosol acidity [22][23][24] , anthropogenic pollutants such as gaseous NO x and SO 2 25-27 and particulate sulfate 13,14 have pronounced effects on SOA i formation. In comparison with other countries, the atmospheric environment of China has been suffering from high levels of SO 2 , NO x , O 3 , VOCs, NH 3 and particulate matter (PM) especially in North China Plain (NCP). Due to the lack of efficient emission controls during the fast industrialization and urbanization processes, severe haze episodes frequently occurred in the region in the past decades 28,29 . In addition, field open burning is still a common activity for disposal of crop residue in the rural area of NCP, especially in summertime wheat harvest period, which releases huge amount of pollutants into the atmosphere and leads to significant impacts on the air quality and aerosol properties 30-32 . Many previous studies in the NCP focused on anthropogenic source distributions of gaseous and particle pollutants, and their environmental, ecological and human health effects [33][34][35][36] . However, only few studies examined how biogenic SOA is affected by such high loadings of anthropogenic pollutants in the region [37][38][39] . In the current study, PM 2.5 filter samples were collected with a three-hour sampling interval at a rural site in the central part of NCP during June 10 th to 25 th , 2013, and determined for SOA i tracers to explore the influence of anthropogenic pollution on SOA i formation. We firstly investigated the temporal variation and chemical composition of SOA i in NCP , and then examined the impacts of sulfate, ozone and biomass burning on the SOA i formation.
Under high-NOx conditions isoprene can be oxidized into methacrolein (MACR) through RO 2 +NO reaction, which undergoes further oxidation with OH and NOx radicals to form methacryloyl peroxynitrate (MPAN) 25,42 . MPAN may react with OH radical to produce MAE and HMML. Those two compounds undergo further reactions under acid conditions, forming 2-methylglyceric acid and some related dimers, organic nitrates and organosulfates 17,27 . Field studies conducted at low-NOx sites (0.1-2 ppb) in the US 27,43,44 and Germany 45 showed that 2-methylglyceric acid concentrations are almost one order of magnitude lower than 2-methyltetrols and C 5 -alkene triols. In contrast, NOx levels (around 20 ppb) 33 in NCP are much higher. Thus, the role of RO 2 +NO pathway in SOA i formation is more important in the region. As shown in Table 1 46 . Both results highlight the important role of the severe anthropogenic pollution in isoprene-derived SOA formation process in the NCP atmosphere. All the SOA i tracers showed a similar temporal variation pattern to organic carbon (OC) (Fig. 1e); concentration of total SOA i tracers linearly correlated with that of OC with a coefficient of r 2 = 0.61, suggesting that isoprene oxidation is an important source of organic aerosols in the rural area of NCP.
Influence of sulfate, aerosol acidity and water content. Acid-catalyzed reactive uptake and subsequent particle-phase reactions of IEPOX (for 3-MeTHF-3,4-diols, C 5 -alkene triols and 2-MTs) and MAE or HMML (for 2-MGA) are the most important chemical pathways of SOA formation from isoprene 12,16 . Several previous studies found that aerosol particles of varying acidity and liquid water contents can affect the SOA i yield 22,23,47,48 . In order to examine the aerosol aqueous chemistry during the SOA i formation, ISORROPIA-II, a thermodynamic model, was performed to estimate the aerosol acidity (i.e., In-situ pH, pH IS , or H + concentration in the aqueous phase, H + aq ) and liquid water content (LWC). As mentioned in Methods section, pH values could approximately be underestimated by one unit due to the lack of NH 3 data 49,50 . In the current study, pH values of PM 2.5 ranged from 0.2 to 4.1 with an average of 2.1 ± 0.6 at the Gucheng site ( Fig. 1b), indicating that the acidity of aerosol in the rural region of NCP is much weaker than that in southeast US (0.5-2.0) 49 , mainly due to the abundant alkaline NH 3 in the region 29 . In contrast, aerosol water content (42 ± 50 μg m −3 ) at the Gucheng site is about 10 times higher than that (5.1-8.4 μg m −3 ) in southeastern US 48 , which can be ascribed to the high loading of inorganic ions in NCP.
Based on a multivariate linear regression analysis of SOA i tracers (r 2 = 0.41, p < 0.001) with sulfate, particle acidity (represented as H + aq ), and water content, we found that only sulfate has a statistically significant (p < 0.001) positive linear relationship with SOA i tracers (Table S1). Moreover, Fig. 2 showed that all the detected SOA i tracers present moderate or good linear correlation (r 2 = 0.3~0.6) with sulfate concentration, again demonstrating the enhancing effect of sulfate on SOA i formation, which is consistent with recent studies in the southeast US 43,44,48,51 . Such an effect can be primarily explained by two reasons: 1) acidic sulfate provides a surface that is favorable for acid-catalyzed reactive uptake and ring-opening reaction of the key intermediates in the gas phase 14,43 , 2) salting-in effect of sulfate increases the solubility of polar organic compounds like IEPOX, MAE and HMML 48 . However, as shown in Fig. 2, we found that sulfate and SOA i tracers present stronger correlations and higher slopes in lower acidity conditions (pH > 2 in the current study, the red dots in Fig. 2), which indicates that reactive acidic uptake of SOA i precursors is more responsible for sulfate growth in the atmosphere when aerosols acidity is weak. This phenomenon may be related to the effects of sulfate on IEPOX reaction probability (γ IEPOX ), aerosol acidity 52,53 , surface area 54 , and the exact explanation needs further research.
In agreement with the multivariate regression analysis, no significant correlations of the detected SOA i tracers with particle acidity and liquid water content (LWC) of aerosol were found in the current study (Fig. 1b), consisting with the results observed in southeastern US 43,48 . This is because that the influences of aerosol acidity and water content on isoprene oxidation are very complicated processes in ambient environment, which can also involve in viscosity or morphology changes and liquid-liquid phase separations of aerosols, dilution effects on ion strength and acidity by particle-phase water, and competitions from gas phase oxidation 48,55 . In addition, particle acidity and liquid water content may vary regionally, which can also influence their relationships to SOA i concentration in the sampling site when SOA are formed during a long-distance transport process. Figure 3 displays diurnal variations of the SOA i tracers at Gucheng station during June 11-19, 2013, which was a period without precipitation. Both 3-MeTHF-3,4-diols and C 5 -alkene triols presented higher concentrations in the nighttime than in the daytime ( Fig. 3a and b) with a statistic significance (p < 0.05). The planetary boundary layer (PBL) height decrease (Fig. 3d) due to the nighttime lower temperature (Fig. 1) would be the primary reason. On the other hand, as showed in Fig. 3c, higher concentration of sulfate can enhance the yield of isoprene-derived SOA by promoting gas-aerosol phase conversion of its precursors and the subsequent aqueous phase chemistry. Thus, the diurnal variation of both 3-MeTHF-3,4-diols and C 5 -alkene triols showed a maximum at 21:00-24:00 p.m. and a minimum in the afternoon (12:00-15:00 p.m.).

Influence of ozone.
However, 2-methyltetrols showed a diurnal variation pattern different from that of 3-MeTHF-3,4-diols and C 5 -alkene triols. Concentrations of 2-methyltetrols displayed an obvious increase in the afternoon, peaking at 15:00-18:00 p.m. In addition to •OH-initiated oxidation, isoprene ozonolysis is also a potential pathway of SOA formation 56 . Previous studies 19,57 proposed that initial formation of isoprene primary ozonides leads to stabilized sCIs, which can further react in the gas phase to form higher molecular weight products that subsequently partition to the aerosol phase and form SOA. Based on their chamber study, Riva et al. 18 found that in the presence of wet acidic aerosols isoprene ozonolysis yields 2-methyltetrols and organosulfates but not produces C 5 -alkene triols, 3-MeTHF-3,4-diols and 2-methylglyceric acid. They tentatively proposed that hydroperoxides formed in the gas phase from isoprene ozonolysis potentially partition to wet acidic sulfate aerosols and hydrolyze to 2-methyltetrols and related organosulfates. Recently, Rattanavaraha et al. 44 verified that isoprene ozonolysis plays a role in 2-methyltetrols formation process. During the current study period, ozone showed a highest concentration in the afternoon (12:00-15:00 p.m.) (Fig. 3d), and weakly correlated with 2-methyltetrols (r 2 = 0.21, Fig. 4a) in the daytime, suggesting an important role of isoprene ozonolysis in the SOA formation process in NCP. Additionally, the non-correlation of ozone with 3-MeTHF-3,4-diols and C 5 -alkene triols and the different diurnal variation patterns of those SOA i tracers suggest that isoprene ozonolysis would have limited influence on these compounds.
In consistent with the field study by Rattanavaraha et al. 44 , 2-methylglyceric acid, represented as MAE/ HMML-derived SOA, also presented a moderate correlation (r 2 = 0.34, Fig. 4b) with O 3 concentration during the daytime, and was highest during 15:00-18:00 p.m., because the formation of MACR (a precursor to MAE and HMML) was enhanced by oxidation of isoprene by O 3 44 .    (Fig. 5a), we found that intensive emissions from the wheat straw burning occurred during the first half of the sampling period (i.e., June 10-18, 2013) in the NCP region. As shown in Fig. 5b, 77% of the air masses arriving at the Gucheng station during June 10-18 were transported via long distances from the southeast part of NCP. Some emission inventory studies revealed that biomass burning could release large amounts of isoprene to the atmosphere 58,59 . Zhu et al. 60 examined the amounts of VOCs in the air at another rural site (Yucheng, Shandong Province, China) in the NCP region, and found that isoprene concentration during the wheat straw burning period is around double of that in normal periods. Thus, in order to discuss the influence of the regional intense burning of biomass fuels on isoprene-derived SOA in the rural areas of NCP, the relationships between SOA i tracers and biomass burning tracer (i.e., levoglucosan) were analyzed ( Fig. 4c and d). During June 10-18, the detected SOA i tracers positively correlated with levoglucosan with a high coefficient (r 2 = 0.67 in the daytime and 0.68 in the nighttime) (Fig. 4c), suggesting that emissions from the biomass burning enhanced the isoprene-derived SOA formation in the NCP region.
As shown in Fig. 5c and d, during the late half of sampling period (i.e., June 19-25), the number of fire spots decreased sharply, suggesting that field burning biomass during the period was insignificant. Thus the correlation between the SOA i tracers and levoglucosan during the period was not as significant as that during June 10-18 (Fig. 4d). However, the SOA i tracers and levoglucosan still had a moderate relationship (r 2 = 0.52) in the nighttime during June 19-25. As shown in Fig. 5d, most air masses (62%) of Gucheng during the late half period were transported from the local area. Because at high isoprene emission from natural plants is insignificant and the local residential biomass burning is the major source, thus the correlation of the SOA i tracers with levoglucosan was only observed at night. Interestingly, the nighttime SOA i tracer concentrations during June 10-18 (150 ± 115 ng m −3 ) were obviously higher than those in the period of June 19-25 (85 ± 53 ng m −3 ) (Table S2), although the nighttime concentrations of levoglucosan during the two phases (250 ± 182 and 218 ± 208 ng m −3 ) were comparable. Recently, Gilardoni et al. 61 directly observed a strong formation of aqueous secondary organic aerosol from biomass burning emissions. At the Gucheng site nighttime concentration (26 ± 16 μg m −3 ) of sulfate during June 10-18 was more than 2 times higher than that (12 ± 11 μg m −3 ) in the remaining period (Table S2), which can be attributed to a continuous aqueous oxidation of SO 2 from biomass burning and other anthropogenic activities during the long-range transport at the first period. As discussed in Section 3.2, sulfate plays an important role in promoting aqueous phase oxidation of isoprene-derived products. Thus, the higher SOA i concentration during the nighttime of June 10-18 was attributed to an enhancing aqueous-phase chemistry of isoprene-derived organics during the biomass burning smoke transport process.  Figure S1), which is located in a rural area of Gucheng, Hebei Province. Detailed information of the station and some observation results of meteorological condition and air quality in the area were described elsewhere 33 . Briefly, the Gucheng site is in the central zone of Beijing-Tianjin-Hebei Region, about 110 km southwest to Beijing (See Figure S1). Taihang Mountains, with a range extending over 400 km from north to south, is about 30 km away from the west of the site. The sampling station is surrounded by the farms and villages, where the main crops are corn and wheat in summer. Weather conditions during the observation period are characteristic of warm and humid (24 ± 4 °C and 67 ± 19%). In all the seasons, the prevailing surface wind directions were northeasterly and southwesterly. A previous study showed that the concentrations of NO x , SO 2 , and O 3 in June (2007) were around 20 ppb, 10 ppb, and 50 ppb, respectively, indicating severe air pollution in summer in the region 33 . PM 2.5 samples were collected on the rooftop (about 10 m above the ground) of a three-story building at the site of the Gucheng station. The sampling was conducted from June 10 th to 25 th , 2013 by using a high volume (1.13 m 3 min −1 ) sampler (Anderson) with a three-hour sampling interval. A total of 118 field samples were collected onto pre-baked (450 °C, 6-8 hr) quartz fiber filters (20.3 cm * 25.4 cm). Three field blank samples were also collected by mounting blank filters onto the sampler for about 15 min without pumping any air. After sampling, the sample filter was individually sealed in pre-baked aluminum foil bags and stored in a freezer (−20 °C) prior to analysis.

Chemical analysis.
A punch of the filter (including both field and blank samples, N = 121) with an area of 12.5-25 cm 2 was extracted three times with a mixture of dichloromethane and methanol (2:1, v/v) each for 10 min under ultrasonication. The extracts were concentrated using a rotary evaporator under vacuum conditions and then blow down to dryness using pure nitrogen. After reaction with 50 μL N,O-bis-(trimethylsilyl) trifluoroacetamide (BSTFA) with 1% of trimethylsilyl chloride and 10 μL of pyridine at 70 °C for 3 hr, the derivatives were determined using gas chromatography/electron ionization mass spectrometry (GC/EI-MS) 55 .
OC and EC in the PM 2.5 samples were analyzed using DRI Model 2001 Carbon analyzer following the Interagency Monitoring of Protected Visual Environments (IMPROVE) thermal/optical reflectance (TOR) protocol. Water soluble inorganic ions including SO 4 2− , NO 3 − , NH 4 + , Cl − , Ca 2+ , K + , and Mg 2+ in samples were analyzed using ion chromatography. Detailed methods for OC, EC, and ions determination are provided in the supplementary material.
Estimations of aerosol pH and water content. ISORROPIA-II 66,67 , a thermodynamic model, was used to estimate aerosol pH and water content, because a direct measurement of aerosol acidity (i.e., pH) is not possible at present. Following by the assumptions of Weber et al. 49 , the PM 2.5 aerosol at the Gucheng rural site was considered as internally mixed and in a single aqueous phase that contained the inorganic species, and no compositional dependence on particle size. In applying ISORROPIA-II, "forward" mode calculation, in which inputs to the model include temperature (T), relative humidity (RH), and the total (gas+aerosol) concentrations of aerosol precursors in the air parcel, would be more suitable to the aerosol pH and water content estimation, although the concentration of NH 3 , HNO 3 , and HCl in the gas phase are unavailable in the current study 50 . Therefore, the pH results given by the current work would be systematically underestimated by approximately one unit 49,50 .