Chemical characterisation of benzene oxidation products under high- and low-NO _x conditions using chemical ionisation mass spectrometry

Aromatic hydrocarbons are a class of volatile organic compounds associated with anthropogenic activity and make up a significant fraction of urban volatile organic compound (VOC) emissions that contribute to the formation of secondary organic aerosol (SOA). Benzene is one of the most abundant species emitted from vehicles, biomass burning and industry. An iodide time-of-flight chemical ionisation mass spectrometer (ToF-CIMS) and nitrate ToF-CIMS were deployed at the Jülich Plant Atmosphere Chamber as part of a series of experiments examining benzene oxidation by OH under high- and low-NOinline-formula_x conditions, where a range of organic oxidation products were detected. The nitrate scheme detects many oxidation products with high masses, ranging from intermediate volatile organic compounds (IVOCs) to extremely low volatile organic compounds (ELVOCs), including Cinline-formula₁₂ dimers. In comparison, very few species with Cinline-formula_≥6 and Oinline-formula_≥8 were detected with the iodide scheme, which detected many more IVOCs and semi-volatile organic compounds (SVOCs) but very few ELVOCs and low volatile organic compounds (LVOCs). A total of 132 and 195 CHOpage3474 and CHON oxidation products are detected by the iodide ToF-CIMS in the low- and high-NOinline-formula_x experiments respectively. Ring-breaking products make up the dominant fraction of detected signal and 21 and 26 of the products listed in the Master Chemical Mechanism (MCM) were detected. The time series of highly oxidised (Oinline-formula_≥6) and ring-retaining oxidation products (Cinline-formula₆ and double-bond equivalent inline-formula= 4) equilibrate quickly, characterised by a square form profile, compared to MCM and ring-breaking products which increase throughout oxidation, exhibiting sawtooth profiles. Under low-NOinline-formula_x conditions, all CHO formulae attributed to radical termination reactions of first-generation benzene products, and first-generation auto-oxidation products are observed. Several N-containing species that are either first-generation benzene products or first-generation auto-oxidation products are also observed under high-NOinline-formula_x conditions. Hierarchical cluster analysis finds four clusters, of which two describe photo-oxidation. Cluster 2 shows a negative dependency on the inline-formula $M12inlinescrollmathmlchem{\mathrm{normal\; NO}}_{\mathrm{normal\; 2}}/{\mathrm{normal\; NO}}_x$ 51pt14ptsvg-formulamathimg16bd505c82589b769a43bc19e5d0ff23 acp-21-3473-2021-ie00001.svg51pt14ptacp-21-3473-2021-ie00001.png ratio, indicating it is sensitive to NO concentration and thus likely to contain NO addition products and alkoxy-derived termination products. This cluster has the highest average carbon oxidation state (inline-formula $M13inlinescrollmathml\stackrel{normal\; ‾}{{\mathrm{normal\; OS}}_{\mathrm{normal\; C}}})$ 28pt16ptsvg-formulamathimg9c420eac3cbefe99ce398da965a11047 acp-21-3473-2021-ie00002.svg28pt16ptacp-21-3473-2021-ie00002.png and the lowest average carbon number. Where nitrogen is present in a cluster member of cluster 2, the oxygen number is even, as expected for alkoxy-derived products. In contrast, cluster 1 shows no dependency on the inline-formula $M14inlinescrollmathmlchem{\mathrm{normal\; NO}}_{\mathrm{normal\; 2}}/{\mathrm{normal\; NO}}_x$ 51pt14ptsvg-formulamathimgf706ed434c2af35e5f9a2839852764c8 acp-21-3473-2021-ie00003.svg51pt14ptacp-21-3473-2021-ie00003.png ratio and so is likely to contain more NOinline-formula₂ addition and peroxy-derived termination products. This cluster contains fewer fragmented species, as the average carbon number is higher and inline-formula $M16inlinescrollmathml\stackrel{normal\; ‾}{{\mathrm{normal\; OS}}_{\mathrm{normal\; C}}}$ 24pt16ptsvg-formulamathimg25988ee4f5117ec202e7346cd5f9ceb1 acp-21-3473-2021-ie00004.svg24pt16ptacp-21-3473-2021-ie00004.png lower than cluster 2, and more species with an odd number of oxygen atoms. This suggests that clustering of time series which have features pertaining to distinct chemical regimes, for example, inline-formula $M17inlinescrollmathmlchem{\mathrm{normal\; NO}}_{\mathrm{normal\; 2}}/{\mathrm{normal\; NO}}_x$ 51pt14ptsvg-formulamathimgf5e255e70dc3320fd8419e25a488b144 acp-21-3473-2021-ie00005.svg51pt14ptacp-21-3473-2021-ie00005.png perturbations, coupled with a priori knowledge, can provide insight into identification of potential functionality.