High atmospheric oxidation capacity drives wintertime nitrate pollution in the eastern

Abstract. Nitrate aerosol plays an increasingly important role in wintertime haze
pollution in China. Despite intensive research on wintertime nitrate
chemistry in recent years, quantitative constraints on the formation
mechanisms of nitrate aerosol in the Yangtze River Delta (YRD), one of the
most developed and densely populated regions in eastern China, remain
inadequate. In this study, we identify the major nitrate formation pathways
and their key controlling factors during the winter haze pollution period in
the eastern YRD using 2-year (2018–2019) field observations and detailed
observation-constrained model simulations. We find that the high atmospheric
oxidation capacity, coupled with high aerosol liquid water content (ALWC),
made both the heterogeneous hydrolysis of dinitrogen pentoxide
(N2O5) and the gas-phase OH oxidation of nitrogen dioxide
(NO2) important pathways for wintertime nitrate formation in this
region, with contribution percentages of 69 % and 29 % in urban areas
and 63 % and 35 % in suburban areas during the haze pollution episodes,
respectively. We further find that the gas-to-particle partitioning of
nitric acid (HNO3) was very efficient so that the rate-determining step
in the overall formation process of nitrate aerosol was the oxidation of
NOx to HNO3 through both heterogeneous and gas-phase processes.
The atmospheric oxidation capacity (i.e., the availability of O3 and OH
radicals) was the key factor controlling the production rate of HNO3
from both processes. During the COVID-19 lockdown (January–February 2020),
the enhanced atmospheric oxidation capacity greatly promoted the oxidation
of NOx to nitrate and hence weakened the response of nitrate aerosol to
the emission reductions in urban areas. Our study sheds light on the
detailed formation mechanisms of wintertime nitrate aerosol in the eastern
YRD and highlights the demand for the synergetic regulation of atmospheric
oxidation capacity and NOx emissions to mitigate wintertime nitrate and
haze pollution in eastern China.



S1. Hygroscopicity correction of aerosol volume and surface area concentrations
The hygroscopicity parameter kappa (κ) of ambient particles was estimated based on the measured chemical composition and an empirical parameterization proposed by Liu et al. (2014): κ = 0.01 + 0.63fNH + 4 + 0.51fNO -3 + 0.81fSO 2-4 + 0.18fWSOC (S1) where fx represents the mass fraction of component x in the particles. During the observation period, the mass fraction of OC in PM2.5 was 8-13% when PM2.5 mass concentration was above 35 μg m -3 , and the water soluble fraction (WSOC) could be smaller. Therefore, we did not consider the contribution of WSOC to κ in our study.
According to the definition of κ (Farmer et al., 2015), we can get the diameter of the wet particle: Where Dp,dry and Dp,wet are the dry and wet diameters of particle, respectively; σs is surface tension of the solution/air interface; ρw and Mw are the density and molecular weight of water; R is the ideal gas constant and T is the temperature (in K).

S2. The parameterization of the major heterogeneous production pathways of HONO
In this study, we parameterized the major heterogeneous HONO production pathways to estimate the HONO budget during the pollution episodes (see Table 1 in the main text). For the photolysis frequency of particulate nitrate (jNO -3 ), previous studies suggested that it had a similar diurnal variation with the photolysis frequency of HNO3 (Romer et al., 2018;Xue et al., 2020). Considering the fact that the photolysis rate of particulate nitrate is faster than that of HNO3, an enhancement factor (EF= jNO -3 /jHNO3) was employed to parameterized the photolysis process of particulate nitrate. We also added the heterogeneous reaction between SO2 and NO2 on aqueous aerosols (R.S1), which is also a source of HONO in the atmosphere (Wang et al., 2016;Wang et al., 2020). In the model, the rate of this reaction was calculated using eq. S3: SO2 (g) + 2NO2 (g) + 2H2O (aq) → SO 2-4 (aq) + 2 H + (aq) + 2 HONO (g) (R.S1) where kaq is the aqueous reaction rate of SO2 and NO2, which is 1.4 × 10 5 M -1 s -1 for pH < 5 and 2 × 10 6 M -1 s -1 for pH > 6, with a linear interpolation between the two pH values (Lee and Schwartz, 1983;Wang et al., 2020); HSO2 and HNO2 are the Henry's Law coefficient of NO2 and SO2 in water, with a value of 1.23 M atm -1 and 1.2×10 -2 M atm -1 at 298K, respectively; Kα1 and Kα2 are the firstand second-order dissociation constant of SO2·H2O, with a value of 1.3×10 -2 and 6.6×10 -8 at 298K, respectively. The H values at various temperatures can be derived by eq. S4: Where ΔHA is the enthalpy change of dissollution at constant temperature and pressure. At 298 K, the value of ΔHA is -6.25 kcal mol -1 for SO2 and -5.0 kcal mol -1 for NO2 (Seinfeld and Pandis, 2016).
T is the temperature (in K).
In addition, the dissociation constant of SO2·H2O at different temperatures can be derived by eq. S5: Where ΔH is the enthalpy change of dissociation at constant temperature and pressure. At 298 K, the value of ΔH is -4.16 and -2.23 kcal mol -1 for disocciation of SO2·H2O and HSO -3 , respectively (Seinfeld and Pandis, 2016).

S3. Analysis of the time series of pollutants at the Qingpu site in the winter of 2019
The time series of PM2.5, nitrate, and other related parameters at the Qingpu site in 2019 are shown in Figure S2. The variation trends of the pollutants at the Qingpu site were similar to those at the Pudong site, but the concentrations were much higher. Nitrate was also the dominant component in PM2.5 during the pollution episodes, and the relatively higher nitrate concentration at the Qingpu site might be due to the higher NOx emissions (8-263 ppb). The O3 concentration ranged between 1-65 ppb with an average of 22 ppb. The Ox concentration ranged from 22 to 85 ppb and was often higher than 40 ppb during the observation period. The high atmospheric oxidation capacity led to the high NOR at the Qingpu site, which was up to 0.54. Similarly, the ALWC was also high due to the high RH in the eastern YRD, and sometimes could also exceed 200 μg m -3 , which would make an important contribution to the nitrate formation.

S4. Case studies of the model simulation during the pollution episodes at the Qingpu site
Different from the Pudong site, the increase of nitrate concentration at the Qingpu site in case 1 occurred during the daytime, from 19.2 μg m -3 at 6:00 to 39.1 μg m -3 at 14:00 on 30 December, 2019, with an average growth rate of 2.5 μg m -3 h -1 ( Figure S6a). The OH radical concentrations was high during the nitrate-increasing period, and the maximum values even reached 2.9 × 10 6 molecules cm -3 , while the N2O5 concentration was close to 0 ppb. This high OH concentration made the gas-phase OH + NO2 process a dominant nitrate formation pathway in this case. After excluding data under RH > 95% conditions, the simulated average production rate of HNO3 from the gas-phase OH + NO2 process during the daytime reached 6.9 μg m -3 h -1 .
In episode 2 (see Figure S6b), the nitrate concentration was maintained at a high level (30-40 μg m -3 ) from the noon of 11 January to the midnight of 14 January, 2020. It then had a rapid increase from 36.1 μg m -3 at 01:00 to 74.9 μg m -3 at 10:00 on 14 January, 2020, with an average growth rate of 4.3 μg m -3 h -1 . Similar to the Pudong site, the heterogeneous hydrolysis of N2O5 made the major contribution to the HNO3 formation during this episode, with the average production rate of 4.0 μg m -3 h -1 , twice that by the gas-phase process.

S5. Sensitivity analyses for key parameters of heterogeneous HONO formation and dilution process in the model
As significant uncertainties remain in the key parameters of the heterogeneous HONO formation pathways used in the model (see Table 1 in the main text), which could affect the prediction of the OH concentration and thereby HNO3 production via gas-phase OH + NO2 reaction, we conducted sensitivity analyses for such parameters to evaluate their influences on HNO3 production during two typical pollution episodes at the Pudong site (see Figure S7). In the base case simulation where a best guess of kinetic parameters was used (see Table 1), the formation of nitrate had comparable contributions from the gas-phase and heterogeneous processes (45% vs. 53%) during the episode 1, while it was dominated by the heterogeneous process (79%) during episode 2. The sensitivity analyses show that although the dark uptake coefficient of NO2 on ground surfaces (γNO2-dk-gs) had the largest influence on HONO concentration during nighttime (-40%/+196%, Figures S8a, d), the photo-enhanced uptake coefficient of NO2 on ground surfaces (γNO2-hv-gs) had the greatest influence on the overall HONO formation as well as HNO3 production via the gas-phase process ( Figures S7b, c, e, f). Specifically, varying the γNO2-hv-gs value by a factor of 5, the gas-phase HNO3 production rate had a change within -13%/+38% and -22%/+63% compared to the base scenario for the episodes 1 and 2, respectively. Correspondingly, the contribution of gas-phase processes to the total HNO3 formation varied within -3%/+8% and -4%/+8%, respectively. It should be noted that variations in these kinetic parameters did not significantly affect heterogeneous HNO3 production. These results suggest that the parameterizations of the heterogeneous HONO formation pathways in the model could provide robust constraints on the relative contributions of both gasphase and heterogeneous processes to nitrate formation during haze pollution events.
Considering the uncertainty in the dilution rate constant (kdil), we also performed a sensitivity analysis for kdil by varying its value from 0.028 h -1 to 0.2 h -1 (corresponding to a dilution lifetime of 5 hours to 36 hours) to evaluate its influence on HNO3 production in a typical pollution episode at the Pudong site (see Figure S8). As the dilution lifetime varied from 5 hours to 36 hours, the average concentrations of N2O5 and OH radicals changed within -23%/+0.8% and -21.6%/+10.8%, respectively ( Figure S8a, d), compared to the base case (dilution lifetime: 24 hours) during the episode. Accordingly, the HNO3 production rates from the heterogeneous hydrolysis of N2O5 and gas-phase OH + NO2 reactions changed within -17%/+1.2% and -33%/+12% ( Figure S8b, e) and the relative contributions of the two pathways changed within -2.5%/+5.5% and -5%/+2.3% ( Figure S8c, f), respectively. The relatively small changes in the rates and relative contributions of the two HNO3 production pathways upon variations in kdil from 0.028 h -1 to 0.2 h -1 suggest that the simplified parameterization of the dilution process using a constant kdil would not result in significant uncertainty in the model results.

S6. Influence of monoterpenes on HNO3 production
The consumption of NO3 radicals by monoterpenes during nighttime can influence the budget of NO3 radicals and N2O5 and thereby the formation of HNO3. We have conducted a sensitivity test for monoterpenes to evaluate their influence on the HNO3 formation. It should be noted that we only have the observation data of monoterpenes obtained using a proton transfer reaction time-of-flight mass spectrometry (PTR-ToF-MS, Vocus, Tofwerk) at an urban site in Shanghai in early November, 2019. We selected the data on 9 November as the ambient temperature (which strongly affects monoterpene emissions) on this day was relatively low (average: 13.3 °C), close to the temperature in winter. The wind speed was also low (average: 0.76 m s -1 ) on this day, which limits the transport and dilution of monoterpene emissions. The monoterpene concentration on this day ranges from 0.009 ppb to 0.070 ppb, with an average of 0.038 ppb. The sensitivity analysis shows that when the monoterpene chemistry was considered, the N2O5 concentration and HNO3 production rate from N2O5 hydrolysis (pHNO3(N2O5)) both had a decrease, especially during the nighttime with high N2O5 concentration ( Figure S9a, b). However, such decrease was relatively small; the average N2O5 concentration and pHNO3(N2O5) decreased by 23% and 12% during the nighttime, respectively. In addition, the contribution of heterogeneous N2O5 hydrolysis to HNO3 formation only decreased by 2.7% ( Figure S9c). Notably, the average temperature in the selected winter haze episode was 8.1 °C, which was lower than the temperature on 9 November, so the concentration of monoterpenes may be smaller, as is their impact on the HNO3 formation.

Heterogeneous 1σ
Gas-phase process 1σ Figure S7 Sensitivity of (a, d) HONO concentration and production rates of (b, e) HONO and (c, f) HNO3 to the variations in the values of key parameters of the heterogeneous HONO formation pathways in the model. Episode 1 (a-c) was from 26 to 31 December, 2019. Episode 2 (d-f) was from 11 to 15 January, 2020. The base case was simulated using the best guess of the parameters as listed in Table 1 in the main text. Figure S8 Sensitivity of N2O5 and OH radical concentrations, production rates of HNO3 from different pathways, as well as their contributions to the HNO3 production to the variations in the value of dilution lifetime from 5 hours to 36 hours in the model. The chosen pollution episode was from 26 to 31 December, 2019. In the base case, a typical dilution lifetime of 24 hours was assumed. Figure S9 Sensitivity of N2O5 concentration, production rates of HNO3 from N2O5 hydrolysis (pHNO3(N2O5)), as well as its contribution to the HNO3 formation (pHNO3(N2O5)/ pHNO3(total)), to the inclusion of monoterpenes in the model simulation. The chosen episode was from 26 to 31 December, 2019. The base case did not consider the effect of monoterpenes.
Figure S10 Production rate of HNO3 from the heterogeneous hydrolysis of N2O5 (the grey line with markers) as a function of γN2O5 during the six haze pollution episodes at the Pudong site in the winter of 2019 (not including the data with RH > 95%). The red line is an "S" curve fitted to the HNO3 production rate and the shaded area is the standard deviation. The blue circle indicates the median of γN2O5 (0.022) during the six pollution episodes, which is located in the region where the heterogeneous production of HNO3 is insensitive to the variation in the value of γN2O5. This suggests that the uptake of N2O5 by aerosols was very efficient so that it was not the rate-determining step in the heterogeneous HNO3 formation during the haze pollution periods.