Synergetic effects of NH 3 and NO x on the production and optical absorp-tion of secondary organic aerosol formation from toluene photooxidation

Abstract. NH3 is the most important alkaline gas in the atmosphere and one of the
key species affecting the behaviors of atmospheric aerosols. However, the
impact of NH3 on secondary organic aerosol (SOA) formation remains
poorly understood, especially the dynamic evolution of chemical compositions
in the SOA formation process. In this study, a series of chamber experiments
were performed to probe the individual and common effects of NH3 and
NOx on toluene SOA formation through OH photooxidation. The chemical
compositions of toluene SOA were characterized using the Aerodyne
high-resolution time-of-flight aerosol mass spectrometer (AMS). The SOA
yield increased from 28.1 % in the absence of NH3 to 34.7 % in the
presence of NH3 but decreased to 19.5 % in the presence of NOx.
However, the highest SOA yield of 42.7 % and the lowest carbon oxidation
state (OSC) occurred in the presence of both NH3 and NOx,
indicating that the higher-volatility products that formed in the presence
of NOx could partition into the particle phase when NH3 was added.
This resulted in a synergetic effect on SOA formation when NH3 and
NOx co-existed. The heterogeneous reaction was the main pathway by
which NH3 participated in SOA formation in the photooxidation
process. The synergetic effect of NH3 and NOx was also observed in
SOA optical absorption. A peak at 280 nm, which is characteristic of
organonitrogen imidazole compounds, was observed in the presence of NH3,
and its intensity increased when NOx was added into the chamber. This
work improves our understanding of how the synergistic interactions between
NH3 and NOx influence SOA formation and offers new insights into
mitigating haze pollution.



S1 OH Concentration Calculation Process
The OH concentration was calculated based on the decay ratio of toluene concentrations and the known rate constant with respect to OH. The change of toluene concentration over time can be expressed as: Where, KOH is the reaction rates constant of OH radicals with toluene (KOH=5.7× 10 -12 cm 3 molecule -1 s -1 ). Assuming that the concentration of hydroxide did not change during the experiment, then we can get:  Fig.1(b) was not a straight line. This is because the OH is consumed as the reaction goes on. The evolution of OH concentration at experiment conditions was shown in Fig.S2. The different experiment conditions in this study did not affect the OH concentration obviously. The highest OH concentration of 1.0 × 10 8 molecule cm -3 was observed at the beginning of the reaction. The average OH concentration over the entire reaction period is 5.9 × 10 7 molecule cm -3 .

S2 OSC calculation
In most previous studies, OSC was estimated from the O/C and H/C data (Liu et al., 2015;Chen et al., 2019;Chhabra et al., 2011;Docherty et al., 2018;Kroll et al., 2011 Average OSC in Exp. 3 was calculated as: OSC = 2 O/C -H/C + σNH3 × 3 N/C -σNOx × 5 N/C. Here, σNH3 is the contribution rate of NH3 to total organic nitrogen in SOA, and σNOx is the contribution rate of NOx to total organic nitrogen in SOA."

S3 PMF results
Positive matrix factorization (PMF) is a receptor model and multivariate factor analysis tool (Paatero and Tapper, 1994;Paatero, 1997). Recently, the PMF model was used to provide better separation of different organic components through highresolution (HR) mass spectra data (Liu et al., 2014). This model was expressed as below: xij = Σpgipfpj + eij where i and j refer to values of j species in i samples, respectively, p is the number of factors, and used a least-squares fitting process, minimizing a quality of fit parameter.
In our study, CU AMS PMF Execute Tool v 3.04A, which was developed by Ulbrich et al. (Ulbrich et al., 2009), was used for the PMF analysis. High-resolution ion fragments at m/z from 12-160 were used. We generated the organic data matrices and the corresponding error matrices from PIKA v 1.15D. Ions were classified and downweighted according to the signal-to-noise ratios (SNR). 0.2<SNR<2 was classified as the weak ions and down-weighted by a factor of 2, SNR<0.2 was bad ions and removed from the analysis. Since O + , HO + , H2O + and CO + are related proportionally only to CO2 + in the fragmentation table, the error values for each of these m/z were multiplied to avoid excessive weighting of CO2 + . The data were analyzed using the PMF2 algorithm (Paatero et al., 2002) with f peak varying between -1 and 1.
A summary of the PMF results is presented in Fig. S1 The direct comparisons of the mass spectra and time series of 3-factor solution are shown in Fig. S4. The 3-factor solution splits the High-nitrogen OA (Hi-NOA) into two components for which we cannot offer a physically meaningful interpretation. While the results of 2-factor solution are also used in the familiar chamber study (Chen et al., 2021;Chen et al., 2019). We therefore choose the 2-factor solution. Tables   Table S1. The content of NO3and NH4 + in the particle-phase for each experiment  Fig. S1 The evolution of toluene concentration for each experiment.   Fig. S4 The 2-factor solution for the toluene OH-oxidation in the presence of NH3.