A Four Carbon Organonitrate as a Significant Product of Secondary Isoprene Chemistry

Abstract Oxidation of isoprene by nitrate radicals (NO3) or by hydroxyl radicals (OH) under high NOx conditions forms a substantial amount of organonitrates (ONs). ONs impact NOx concentrations and consequently ozone formation while also contributing to secondary organic aerosol. Here we show that the ONs with the chemical formula C4H7NO5 are a significant fraction of isoprene‐derived ONs, based on chamber experiments and ambient measurements from different sites around the globe. From chamber experiments we found that C4H7NO5 isomers contribute 5%–17% of all measured ONs formed during nighttime and constitute more than 40% of the measured ONs after further daytime oxidation. In ambient measurements C4H7NO5 isomers usually dominate both nighttime and daytime, implying a long residence time compared to C5 ONs which are removed more rapidly. We propose potential nighttime sources and secondary formation pathways, and test them using a box model with an updated isoprene oxidation scheme.


Introduction 44
Details on the isoprene oxidation experiments in the atmospheric simulation chamber 45 SAPHIR are provided in section S1. Sensitivities for the quantification of the organonitrates 46 are discussed in section S2. Section S3 provides details on the ambient measurements.

47
Section S4 discusses the updated chemical mechanism of the C4H7NO5. The last section, 48 S5, gives details on the complimentary experiments at the Go:PAM flow reactor.

49
Text S1. Experiments in the atmospheric simulation chamber SAPHIR

50
The experiments were conducted in the atmospheric simulation chamber SAPHIR (Rohrer 51 et al., 2005;Fuchs et al., 2017) in Jülich, Germany, in August 2018 to improve our knowledge 52 on the gas and particle phase products of isoprene oxidation by NO3 radicals (Dewald et 53 al., 2020;Brownwood et al., 2021;Wu et al., 2021;Vereecken et al., 2021). Here we selected 54 four experiments to scrutinize the formation of ONs (Table S1). The major loss of peroxy 55 radicals was the reaction with HO2. However, different chemical conditions enhanced 56 different chemical regimes (Brownwood et al., 2021). In experiment 1 HO2 formation was 57 enhanced by propene ozonolysis and CO addition to favor the RO2 + HO2 regime, whereas 58 experiment 2 favored the RO2 + RO2 regime. In the other two experiments we simulated 59 nighttime to daytime transition exposing the nighttime products to OH oxidation and  Table S1). In a last step, after total isoprene consumption, only ozone and NO2 67 were added to enhance further oxidation of the products (exp. 1 and exp. 2). In exp. 3 and 68 4 the chamber roof was open to test the daytime effect on the products after the third 69 injection. The potential aerosol contributions of isoprene products were scrutinized by 70 addition of ammonium sulfate as aerosol seeds in exp. 3. However, the focus on the 71 present study is on the gas-phase processes where a high-resolution time-of-flight 72 chemical ionization mass spectrometer (HR-ToF-CIMS, Aerodyne Research Inc.) (hereafter 73 I-CIMS) using iodide as the primary reagent ion (Lee et al., 2014) was used to measure the 74 gas-phase oxidation products. A filter inlet for gases and aerosols (FIGAERO) (Lopez-75 Hilfiker et al., 2014) was also coupled to the I-CIMS during the experiment with aerosol 76 seeds to measure the particle phase oxidation products. For clarity, the particle data was 77 removed from the time trends shown in Figure 1c. The I-CIMS was placed in an air-78 conditioned container under the chamber. A four-meter long PFA (Swagelok, 6mm 79 diameter) line and four-meter long copper tubing (12mm diameter) were used as gas and 80 particle inlets respectively. Both were insulated to avoid condensation in the lines. The 81 measured signal of the 64 identified ONs was corrected for background, normalized to 82 iodide signal and converted to ppt using a bulk sensitivity of 4.8 ncps ppt -1 for all measured 83 ONs (see Section S2). No loss corrections have been applied to the data set. A CIMS using 84 bromide as the reagent ion and coupled with a customized inlet (Albrecht et al., 2019) 85 attached directly at the bottom of the chamber was also deployed (Wu et al., 2021).

89
(hydroxy nitrates, IHN) were estimated based on isoprene consumption (Table S2). The 90 measured ONs signal was converted to ppt using a bulk ON calibration factor of 4.8 ncps 91 ppt -1 with a standard deviation of 0.7 ncps ppt -1 ( Figure S1).

102
Recently,  showed that secondary chemistry processes in the ion

132
The ion counts, i.e. normalized counts per second (ncps), derived for each species using 133 the I-CIMS can be converted to concentration units using appropriate instrumental 134 sensitivities, which can be derived from standards. Unfortunately, for the detected ONs

135
there is a lack of standards making direct quantification difficult. However, two methods 136 to derive limits on sensitivities were applied, i.e. using bulk sensitivity or relative sensitivity.

137
These two methods constrain and strengthen the various conclusions of the overall and 138 relative importance of the various organonitrates discussed in this work.

139
To derive bulk-sensitivities for the ONs a Thermal Dissociation Cavity Ring-down

202
The C5H9NO4 relative sensitivity is lower than the others but closer to values that has been 203 reported before using synthesized standards of different C5H9NO4 (IHN) isomers (Lee et 204 al., 2014). This difference indicates that we may underestimate the C5H9NO4 concentration.

205
It should be noted that the relative sensitivities were estimated to understand the potential 206 variability of I-CIMS sensitivity to the different ONs. The bulk ON calibration factor of 4.8 207 ncps ppt -1 with a standard deviation of 0.7 ncps ppt -1 was utilized for all conversions to 208 ppt using GU-CIMS.

218
The measured ONs signal has been converted using the bulk sensitivity factor for the ONs

224
The measurements in Gothenburg, Sweden, took place in the city's port in October, 2014.

225
The measurement in Changping and Hong Kong were part of the project "Photochemical  Table S5 and Table S6, respectively.

273
The oxidation products with the chemical formula C4H7NO5 consist of different isomers.

274
The chemical structure of most relevant isomers and the naming convention using the 275 Master Chemical Mechanism (MCM) are depicted in Table S7.

276
The FZJ-NO3-isoprene mechanism was published recently in the study by Vereecken et al. 277 (2021). There, the focus was on the initial reaction of isoprene with the NO3 radical and the 278 resulting peroxy-and alkoxy-radicals. This also leads to additional sources of HC4ACHO  Figure S5 were constructed. The rate 282 of the reaction with NO3 was calculated as shown in (Kerdouci et al., 2014), with a 65% 283 preference for addition on the secondary carbon as used for the detailed description of 284 the addition reaction in the MCM. The rate coefficients and branching ratios for the 285 bimolecular peroxy radical reaction pathways (reaction with NO, NO3, HO2 and other RO2) 286 are calculated according to (Jenkin et al., 2019), the unimolecular reactions as given in 287 (Vereecken and Nozière, 2020). The rates for the alkoxy radicals are calculated as given in 288 (Vereecken and Peeters, 2009;Novelli et al., 2021), with only the competitive reactions 289 implemented. It should be noted that this is the first expansion of the FZJ-NO3-isoprene 290 mechanism to such late-stage chemistry. The mechanism expansion itself with all relevant   The comparison of the modeled and the measured C4H7NO5 time profiles are in good 344 agreement for experiments 2 and 3 ( Figure S7 and S8) (see also Figure 3 and Table S1).

345
However, the estimated concentrations differ between the model and the measurements.

346
The discrepancy varies depending on the experiment by a factor of 10 to 22. The lowest 347 difference was observed for experiment 2 (a factor of 10). One may note that in experiment