Effect of NOX, O3 and NH3 on sulfur isotope composition during heterogeneous oxidation of SO2: a laboratory investigation

Sulfate aerosol is a major fraction of haze, playing an important role in aerosol formation and aging processes. In order to understand the mechanism of sulfate formations, the characteristics of sulfur isotope composition were determined during different heterogeneous oxidation reactions of sulfur dioxide. Although NH3 was more beneficial to the formation of sulfate, compared with NOX and O3, 34S tended to enrich the lighter sulfur isotopes in the presence of NH3. Furthermore, in 15 consideration of the potential competitive effects of NOX, O3, and NH3 in the heterogeneous oxidation processes, the contributions of each gas were evaluated via Rayleigh distillation model. Notably, NOX oxidation contributed 67.5±10 % of the whole sulfate production, which is higher than O3 (13.3±10 %), and NH3 oxidation (19.2±10 %) on the basic of the average fractionation factor. The observed δ34S values of sulfate aerosols were negatively correlated with sulfur oxidation ratios, owing to the sulfur isotopic fractionations during the sulfate formation processes. Given the isotope mass balance, the 20 overall δ34Ssulfate approached the δ34Semission as oxidation of SO2 progressed, suggesting that NOX played a major rather than a sole role in the different heterogeneous oxidation processes of SO2.

emission from fossil burning is the main source of anthropogenic sulfate aerosol (Haywood et al., 2000). SO2 can be converted to sulfate via the gas-phase oxidation and heterogeneous reactions Kong et al., 2014). Sulfate aerosol not only affects the global climate through direct and indirect radiative forcing (Anderson et al., 2003;Rosenfeld et al., 2014), but also threatens the human respiratory system (Liu et al., 2019). Studying the source and formation mechanism 35 of sulfate aerosol in the atmosphere is of great significance for clarifying the climate change characteristics (Chen et al., 2015). However, the formation mechanism of sulfate is still unclear. Guo et al. (2017) reported that catalytic oxidation of metal ions contributed the most to the formation of sulfate. It was confirmed that SO2, NOX, and NH3 synergistically accelerated the formation of sulfate under high humidity (Wang et al., 2016). Li et al. (2018) demonstrated that the oxidation of S(IV) in haze in China was driven by the HONO/NO2generated due to the consumption of NO2 on the surface of 40 aerosols. Besides, aqueous NO2 serves as the dominant oxidant of SO2 at highly elevated NOX levels (Xue et al., 2019).
Moreover, the synergistic effect between NO2 and SO2 on the surface of mineral promotes the conversion of SO2 to sulfate. Therefore, the oxidation processes of SO2 cannot be summarized by a simple oxidation mechanism.
The heterogeneous oxidation is the core process of the formation of secondary aerosol, having a guiding role in revealing the formation mechanism of the compound pollution (Lu et al., 2018). The measured sulfur isotopic fractionation 45 showed that about -9 ‰ is for homogeneous oxidation of SO2 and up to +16.5 ‰ is for heterogeneous oxidation of SO2 (Chen et al., 2017). In addition, the pathways of SO2 oxidation in aqueous-phase systems include reactions with O3, H2O2, NO2, and by O2 via catalyst (Hung et al., 2015). The formation of sulfate was demonstrated that it was mainly affected by a synergistic effect between NOX and SO2 (Gao et al., 2020). Sulfate formations measured during autumn were mainly related to excessive O2 with Fe 3+ as catalyst (Guo et al., 2014). Therefore gaseous oxides are the key oxidation pathways in the 50 formation of sulfate during the heterogeneous oxidation process (Chen et al., 2017).
The formation process of the secondary sulfate and sources of sulfur in the atmosphere can be investigated by sulfur isotope ratios (Han et al., 2017), as the characteristics of different sulfur sources can be represented by sulfur isotopic signatures ( 32 S and 34 S) (Winterholler et al., 2008). Besides, sulfur isotopes also exhibit distinctive isotope fractionations for different oxidation processes of SO2 (Han et al., 2016a), which can be applied in the study of formation processes of sulfate.

55
The enrichment of heavy sulfur isotopes in sulfate may be caused by heterogeneous oxidation of SO2, whereas light sulfur isotopes in sulfate may be attributed to homogeneous oxidation (Chen et al., 2017). Nevertheless, isotope fractionations can only be used to roughly distinguish between the heterogeneous and homogeneous oxidation. The specific oxidation path of SO2 cannot be discriminated due to the similarity of the sulfur isotope fractionation under different conditions during the oxidation processes of SO2. 60 Some investigations about sulfur isotope composition and fractionation have been performed for understanding the formation pathways of sulfate aerosols (Harris et al., 2013b;Yang et al., 2018;Chen et al., 2017). However, most of these observational and modeling studies investigate the sulfur isotope composition in real atmosphere without physical boundaries (Han et al., 2016b;Li et al., 2020). Yet to date, a few experiments were performed in limited physical boundaries to explore the mechanism of sulfur isotope fractionation on the microscale (Harris et al., 2012a;Harris et al., 2012b). To our the sulfur isotopic fractionations were investigated with the Rayleigh distillation model to understand the relative contribution of each SO2 oxidation pathway. It may not only provide a theoretical basis for the causes of subsequent sulfates, 70 but also be crucial for improving the air quality and studying the regional climate change.

Material and methods
Sampling site was located on the roof of the library in Nanjing University of Information Science & Technology (32.1°N, 118.5°E). PM2.5 samples were collected by using a high volume sampler (TH-1000H, Tianhong Co., Wuhan) with a 75 flow rate of 1.05 m 3 min -1 from 9 am to 9 pm per day from 26 th Feb. 2016 to 6 th Apr. 2016.
Hematite (α-Fe2O3) in the experiments was prepared according to the previous studies (Legodi et al., 2007;Fu et al., 2006). A plate with evenly dispersed α-Fe2O3 powder was loaded into the experimental apparatus. On the basis of SO2-Ar, different proportions of NOX (O3 or NH3) were added to the reactor, combined with/without O2 and/or light. The NOX was composed of NO2 and NO with volume ratio of 2:1. The flow rates of Ar and SO2 were 95 and 2 mL min -1 respectively. The 80 flow rates of NOX (O3 or NH3) varied from 2 to 16 mL min -1 depending on the ratio of SO2 to NOX (O3 or NH3). The wavelength of ultraviolet (UV) light is 303 nm, and the light intensity is 25 μW cm -2 . All of the experiments were conducted for 2 h at a temperature of 298 K with a relative humidity of nearly 40 %. The values of temperature and humidity in the experiments were set similarly to those in the air during this period.
The obtained samples were soaked in 50 mL of Milli-Q water and sonicated for 30 minutes to extract sulfate. The 85 samples were centrifuged to separate the sulfate supernatant. The dissolved sulfate in the supernatant was precipitated as BaSO4 by adding 1 mol L -1 BaCl2 solution. The BaSO4 precipitates were separated with 0.22 μm acetate membrane and rinsed with 150 mL Milli-Q water to remove Cl -. The BaSO4 powder was calcined at 1123 K in a muffle furnace for 2 h to obtain the final pure BaSO4 sample.

90
The δ 34 S value was determined by using isotope mass spectrometer (IRMS, Delta V Plus, Finningan) and Elemental analyzer (EA, Flash 2000, Thermo). BaSO4 was used as an analytical sample of sulfur isotope composition. The result was with respect to international standard V-CDT, and the accuracy was better than ±0.2 %.

Results and discussion
First all, for confirming the sulfur isotope composition and fractionation, PM2.5 samples from 26th Feb. to 6th Apr. were 95 collected and the δ 34 S values of them were measured. δ 34 S value may change when SO2 is oxidized into the sulfate in the atmosphere via different oxidation pathways. δ 34 S aerosol and SO2 Oxidation Ratio (SOR = SO4 2-/(SO4 2-+SO2)) calculated throughout the sample period were shown in Fig. 1. It can be found that the δ 34 Saerosol values showed a~3.9 ‰ variation, and displayed a negative correlation with SOR. The variation of δ 34 Saerosol values was attributed to the isotope fractionation during the oxidation processes of SO2 (Li et al., 2020). To shed light on the mechanism of SO2 oxidation in sulfur isotope 100 fractionation and sulfate formation, the SO2 oxidation processes on the surface of α-Fe2O3 in the presence of NOX, O3, and NH3 were carried out in laboratory.
NOX is the most important oxidant during the heterogeneous oxidation of SO2 taking the impact of ion strength into account. As shown in Fig. 2, the yield of SO4 2ranged from 0.0097 to 0.7795 g and the values of δ 34 S were 2.9-4.8 ‰. It was noteworthy that there was an obvious discrepancy between the yield of SO4 2and the sulfur isotope values. Few sulfates were 105 formed via the heterogeneous reaction between NOX and SO2 on the surface of mineral in the dark, whereas the formation of sulfate was enhanced under light, suggesting that UV light could promote the oxidation of SO2. Besides, in the presence of O2, NOX and mineral oxides could act as catalysts to increase the conversion rate of SO2 on the surface of mineral oxides (Gao et al., 2020). In addition, the increase in the amount of NOX was another key factor that led to the acceleration of sulfate formation (Cheng et al., 2016). When NOX and SO2 coexisted, the content of sulfite on the surface of all oxides was 110 reduced significantly . Simultaneously, HONO was formed by NO2 and subsequent hydrolysis in thin films of water coating boundary layer surfaces according to reaction (R1) (Kebede et al., 2016). The release of HONO may help to sustain the efficient sulfate production and droplet acidity. Moreover, the co-adsorption of the oxidant N2O4 formed from nitrate under the action of S(IV) further led to the formation of sulfate (Cheng et al., 2016).
The different proportion of SO2 and NOX made a great growth of the sulfur isotope values, which suggested that the O3 also exerted an influence on the formation of sulfate and sulfur isotope values. The yield of SO4 2ranged from 0.0081 to 0.6712 g with the δ 34 S values of 1.6-2.9 ‰ (Fig. 3). The rapid growth sulfate yields indicated that the oxidation of SO2 was sensitive to the concentration of ozone. Ozone, as a very efficient oxidant, could react with the sulfite to release 130 oxygen, promoting the subsequent oxidation of SO2, which was described as: HSO3 -+ O3 → HSO4 -+ O2, Whether the oxidation processes were on the surface of Fe2O3 mineral dust or not, the coexistence of O3 can convert could generate a mass of hydroxyl radicals (Ran et al., 2014;Cheng et al., 2016). As adsorption sites for water, surface hydroxyls were the principal reactive sites on metal oxides. In turn, the adsorbed water was either dissociated into more hydroxyls at oxygen vacancies or hydrogen-bonded to surface O-H groups, which was in favor of the heterogeneous oxidation of SO2 .
145  (Silvern et al., 2017). Herein, the results showed that the reaction of NH3 with acid may lead to an increase in the formation of sulfate. In addition, surface Lewis basicity might be provided by NH3 for SO2 absorption on the mineral, increasing the amount of condensed water on the secondary aerosols and enhancing the formation of sulfate (Chu et al., 2016). Moreover, the oxygen vacancies in α-Fe2O3 may lead to the formation of sulfate on α-Fe2O3 .
Of note, the effect of NH3 on the sulfur isotope composition was not apparent with the relatively stable overall trend.

165
Under only-light, NH3, which increased the alkalinity by producing OHfrom hydrolysis, dominated in the reaction, leading to an increase of δ 34 S values (Jiang et al., 2017 Among them, R= 34 S/ 32 S. Thus, α > 1 indicates that the reactions are more inclined to enrich heavy isotopes. The δ 34 S value of SO2 sources (δ 34 Semission) strongly depends on the origin of SO2, thus it is difficult to constrain (Li et al., 2018). On the basis of the literature, +2.7±2.0 ‰ is used as the δ 34 Semission value during our sampling period (Li et al., 2020). Therefore, we used a Rayleigh distillation model and the isotopic enrichment factor (ԑ = (α-1)*1000‰) to explain the Where Rp and Ri are the ratios of 34 S/ 32 S for the product sulfate and the initial SO2 gas respectively and f (1-SOR) is the 190 fraction of remaining SO2.
The measured ԑobs values as a result of mixing oxidation pathways of SO2 + NOX + O3 + NH3 were +1.3±1.4 ‰ by simulations (Fig. 5). NOX oxidation enriched 34 S in the product sulfate with an enrichment factor (ԑNOX) of +2.1 ‰, and oxidation by O3 pathway depleted 34 S (ԑO3=+0.2 ‰) in the product sulfate, and oxidation by NH3 pathway enriched 34 S in the product sulfate with a ԑNH3 value of -0.6 ‰. Considering the isotope mass balance, the overall ԑobs value (+1.3 ‰) fell in 195 between NOx, O3, and NH3 values and approached δ 34 Semission as oxidation of SO2 progressed, indicating that NOX, O3, and NH3 all had a certain influence on the heterogeneous oxidation of SO2.