Quantification, diel variation and photochemistry of inorganic chlorine trace gases in continental Germany

Abstract

Understanding the sources, distribution, and lifetime of inorganic chlorine-containing trace gases is crucial to assessing their tropospheric impacts. We report in situ measurements of Cl₂, HOCl, ClNO₂, and ClONO₂ using iodide chemical ionization mass spectrometry during a 2.5-week campaign in June 2024 at a rural continental site in central Germany. Air masses that had passed over “marine-anthropogenic” regions (≈400 km distant) showed significantly higher mixing ratios of chlorine-containing gases than “continental-unpolluted” air masses. From the marine-anthropogenic period, we provide the first quantitative observations of ClONO₂ in the lower troposphere (up to 59.8 pptv during daytime). Persistent nonzero ClONO₂ at night implies a non-photochemical source of ClONO₂ or its precursor ClO, and/or that heterogeneous loss is slower than laboratory uptake coefficients suggest. ClNO₂ levels were consistent with production via N₂O₅ uptake on chloride-containing particles; both ClNO₂ and Cl₂ were enhanced when O₃-/N₂O₅-rich air entrained into the nocturnal boundary layer. Photolysis of ClNO₂, Cl₂, and HOCl yielded mean maximum Cl atom production rates of 1.0 × 10⁶ cm⁻³ s⁻¹ under marine-anthropogenic influenced air and 1.6 × 10⁵ cm⁻³ s⁻¹ under continental-unpolluted conditions. In the early morning, Cl production (due to ClNO₂ photolysis) exceeded primary–OH production from O₃ photolysis, while after noon HOCl photolysis was the dominant Cl source. At low solar zenith angles, HOCl photolysis contributed up to 40% of primary OH. These measurements indicate that Cl atoms can strongly influence hydrocarbon oxidation in similar rural regions, with potential regional and global implications (of up to 15%) for the methane lifetime.

---

Environmental significance

Tropospheric, inorganic chlorine-containing trace gases (Cl-species) provide a source of chlorine atoms that impact e.g. on the rates of hydrocarbon oxidation and thus photochemical O₃ formation. We report measurements of Cl₂, ClNO₂, ClONO₂ and HOCl in central Germany, distant from coastal regions and thus not expected to be influenced by marine-emissions of sea-salt. The measurement of ClONO₂, the first ever made in the lower troposphere, enabled estimation of nighttime ClO and Cl radical mixing ratios. The impact of Cl₂, ClNO₂, ClONO₂ and HOCl on Cl and OH radical production is shown to be significant in rural continental regions, when air masses have passed marine and anthropogenically polluted regions. Further estimations suggest a significant influence on the methane lifetime.

1 Introduction

The role of inorganic chlorine trace-gases in atmospheric science is well established for the stratosphere, where molecules such as ClONO₂, HOCl, HCl and thermally unstable Cl₂O₂ play a central role in ozone depletion, most noticeably in polar regions where the catalytic recycling of Cl atoms through a mixture of heterogeneous and photochemical processes leads to the formation of the “ozone-hole”.^1–3 In the stratosphere, the source of reactive chlorine trace gases is the photolysis of organic chlorine-containing species such as CH₃Cl and CCl₂F₂, which were emitted into the lower atmosphere in large quantities prior to their ban in the Montreal Protocol in 1989.^4–7

Chlorine species are highly reactive when in radical forms such as Cl or ClO, resulting in a short lifetime (seconds) and significant chemical effects on their immediate environment. The lifetime of the inorganic radical reservoirs, including photolabile Cl₂, ClNO₂, ClONO₂, and HOCl depends on the actinic flux,^8,9 while the lifetime of HCl depends on its reaction with OH and deposition. While there are several observations of ClNO_2, HOCl, and Cl₂ in the troposphere e.g.,^10–14 ClONO₂ is considered as a stratospheric trace gas, which is only present in the upper troposphere as a result of intrusion from stratospheric air masses.^15–17 In the troposphere, chlorine atoms are important as they contribute significantly to the oxidative capacity and especially the oxidation of hydrocarbons.^18–21

The presence of chlorine radicals derived from inorganic reservoirs such as ClNO₂ and HOCl enhances the rate of oxidation of hydrocarbons (forming HCl and organic peroxyl radicals), which influences air composition and especially photochemical ozone production.^22–24 Understanding the sources, distribution and lifetime of inorganic chlorine trace-gases is thus crucial for evaluating their impact on chemical processes in the troposphere.

ClNO₂ and Cl₂, HOCl and HCl (in both gas and particle phases) are generally believed to be the most abundant, inorganic chlorine-containing gas phase species in the troposphere.²⁵ Various routes to their formation in the atmosphere are illustrated in Fig. 1 and described below.

Fig. 1 A simplified diagram for formation and (photo)chemical transformations of chlorine-containing trace-gases in the troposphere. The links to NO_Y (e.g., NO₂, NO₃ and N₂O₅) are highlighted, as is the coupling with HO_X cycles (OH and HO₂).

The main source of gaseous inorganic chlorine in the troposphere is the release in the form of HCl, e.g., via acid displacement of chloride from sea-salt (mainly NaCl)^26–28 and gaseous emission from biomass burning, waste disposal, and fossil fuel extraction.^29–33 Additionally, HCl can be formed through H-abstraction from volatile organic compounds (VOCs) by Cl atoms that originate from the mechanisms discussed below.²⁵ In a continental atmosphere, NH₃ and HCl react to form NH₄Cl, which partially partitions into the particle phase. The heterogenous uptake of N₂O₅ on chlorinated particles, liberates chloride as ClNO₂ into the gas phase.³⁴

N₂O₅ + Cl⁻ → ClNO₂ + NO₃⁻ (R1)

N₂O₅ is formed in the reaction between NO₂ and NO₃ (the latter formed in the oxidation of NO₂ by O₃) and, due to its thermal equilibrium with photolabile NO₃ (R2), is only present in significant concentrations at night:³⁵

NO₂ + NO₃ + M ↔ N₂O₅ + M (R2)

The abundance of N₂O₅ (and thus ClNO₂) is closely tied to the availability of NO_X and the lifetime of NO₃ and is therefore maximized in nocturnal air masses that are either impacted by anthropogenic emissions with high levels of NO_X or in which the NO₃ reactivity towards unsaturated hydrocarbons is low. NO₃ and thus N₂O₅ lifetimes in forested regions are often very short^36,37 as the abundance of terpenoids, which react rapidly with NO₃, is high. In this case, activation of reactive chlorine from its particle reservoirs is inefficient and the formation of ClNO₂ is suppressed. High levels of ClNO₂ are typically found at night in polluted marine areas such as Asian megacities and industrial metropolitan areas.^38–41 ClNO₂ has also been observed in air masses which have been transported long distances from polluted and marine source regions to continental sites.^10,13,14

Once formed in the multiphase reaction between N₂O₅ and particle-chloride, most of the ClNO₂, which has a low solubility,⁴² is released into the gas phase where it is photolyzed to Cl and NO₂ (R3).^43–45 Other loss processes, such as reaction with OH are negligible by comparison.⁴⁶

ClNO₂ + hv → Cl + NO₂ (R3)

ClNO₂ can also react with acidified (pH < 2), chloride-containing particles to generate gas-phase Cl₂:

ClNO₂ + Cl⁻ → Cl₂ + NO₂⁻ (R4)

Cl₂ is also formed through the uptake of O₃ onto acidic, chloride-containing particles³⁰ and (as shown below) through other heterogeneous processes.⁴⁷ During daytime, Cl₂ is rapidly photolyzed in the troposphere (lifetime ∼10 min) to form two Cl atoms:

Cl₂ + hv → 2Cl (R5)

The ClO radical is formed when Cl atoms react with O₃ (R6).³⁴ ClO radicals then react with HO₂ and NO₂ to form HOCl and ClONO₂, respectively ((R7) and (R8)).

Cl + O₃ → ClO + O₂ (R6)

ClO + HO₂ → HOCl + O₂ (R7)

ClO + NO₂ + M → ClONO₂ + M (R8)

As HO_X radicals are ubiquitous in the atmosphere, the ClO radical is the limiting trace gas for formation of HOCl. Consequently, larger mixing ratios of the latter have been measured near source regions with high chloride load such as marine sites or industrial facilities.^11,48 The formation of ClONO₂ requires the presence of NO₂ and is thus expected to be formed at largest rates in regions with both anthropogenic influence and particulate chloride as precursor for Cl-atoms and thus ClO.

HOCl and ClONO₂ are lost in both gas- and multiphase processes. Similar to ClNO₂, the uptake of HOCl or ClONO₂ to chloride-containing, acidic particles results in the generation of Cl₂ ((R9) and (R10)),^49,50 while heterogenous hydrolysis of ClONO₂ forms HOCl (R11).⁵¹ The photolysis of HOCl leads to the production of Cl atoms and OH radicals (R12), whereas the photolysis of ClONO₂ leads mainly to Cl atoms (R13a) with a minor contribution to ClO formation (R13b).³⁴

HOCl + Cl⁻ → Cl₂ + OH⁻ (R9)

ClONO₂ + Cl⁻ → Cl₂ + NO₃⁻ (R10)

ClONO₂ + H₂O → HOCl + HNO₃ (R11)

HOCl + hv → Cl + OH (R12)

ClONO₂ + hv → Cl + NO₃ (R13a)

ClONO₂ + hv → ClO + NO₂ (R13b)

Measurements of chlorine-containing trace-gases in the troposphere date back to the mid-1980s, whereby chromatographic methods were used to monitor e.g., HCl and HOCl.^26,52 Recent developments in mass spectrometry have enabled sensitive and selective detection of several chlorinated trace gases with chemical-ionization mass spectrometry (CIMS) using iodide as the reagent ion for a wide variety of inorganic chlorine species.^11,23,53,54 The recent development of high-resolution (e.g., Time-of-Flight, ToF) analysers allows identification of non-fragmented chlorine species and therefore offers high selectivity as well as excellent sensitivity. However, the number of studies using such devices to examine tropospheric chlorine chemistry is still limited and those available mostly report measurements in polluted marine regions. The presence of chlorine-species in rural continental sites, where the source of chlorine is obscure, remains largely unexplored.

In this study we present the quantitative measurement of Cl₂, HOCl, ClNO₂ and ClONO₂ using VUV-Iodide-ToF-CIMS during field measurements in central Germany. The formation and loss of inorganic chlorine-containing species throughout the diel cycle are discussed, including the first quantitative measurements of ClONO₂ in the troposphere and potential limitations by interfering secondary ion reactions. The goal of our analysis is to assess the influence of chlorine species on local Cl and OH production for air masses with marine-anthropogenic and continental-unpolluted origin and to estimate the impact on e.g., the lifetimes of methane and non-methane hydrocarbons.

2 Methods: CIMS operation and calibration

The inorganic chlorine trace-gases ClNO₂, ClONO₂, HOCl and Cl₂ as well as N₂O₅ were measured using an Iodide-ToF-CIMS (Aerodyne Research Inc., USA⁵⁵) during the BISTUM24 campaign (Section 2.3) and in further laboratory studies (Section 2.2 and Supplement). A vacuum ultraviolet (VUV) lamp and 20 ppmv CH₃I in N₂ were used to generate iodide primary ions, which form clusters with the inorganic chlorine-containing compounds in the ion molecular reactor (IMR). The IMR was maintained at 170 mbar, heated to 60 °C and continuously humidified (>7000 ppm H₂O). The chlorine-containing trace gases and N₂O₅ were detected as clusters with iodide at the mass-to-charge ratios (m/z) listed in Table 1. During all measurements, the CIMS resolution was m/z/Δm/z = 5000 at m/z 250. Peak assignment and data processing was conducted with Tofware V4.0.3, ensuring that isotopic ratios are in accordance with their natural abundance. Iodide clusters with Cl and ClO were also observed but were minor in magnitude (<3%) of the signal from ClNO₂. These were excluded from further analysis as they most likely originate from the fragmentation of larger chlorine-containing species³⁹ as indicated by high correlations with ClNO₂ (R² = 0.96). The unambiguous first in-field detection of ClONO₂ is shown in the mass spectra in Fig. S2.

Table 1 Calibrated chlorine species with detection limit (LoD, as 3 × σ for an averaging period of 5 min) as found for noise in the laboratory (LoD lab) and in the field (LoD campaign), calibration method and associated uncertainty. Thermal decomposition and non-destructive cavity ring-down spectroscopy (TD-CRDS and CRDS) were used as reference instruments^a

Species	I^−-cluster/m/z	LoD lab/pptv	LoD campaign/pptv	Calibration method	Uncertainty/%
a Uncertainties calculated via error-propagation considering e.g. calibration methods and the linear regression of the calibration curves (Fig. S3 and S4): ^a<7% (mainly uncertainty in Cl2 cross-sections^56 and humidity correction). ^b ∼8% (measurement uncertainty in Cl2 and humidity correction, assumes 100% conversion of HOCl to Cl2.) ^c 28% (same as N2O5, assumes 100% conversion of N2O5 to ClNO2). ^d12% (uncertainty in ClONO2 IR-cross-sections^57 and humidity correction). ^e28% (uncertainty in NO2 cross-sections and conversion of N2O5 to NO2.^58 Uncertainty in dilution factor is based on a flow controller uncertainty of 0.6%. The uncertainty associated with the humidity dependence function is: HOCl: 3.7%, Cl2: 5.3%, ClNO2: 0.9%, ClONO2: 6.1% and N2O5: 1.5%.
Cl2	196.843	0.05	0.03	Direct (Cl2 gas), CRDS	7^a
HOCl	178.877	3.50	2.21	Indirect (HOCl-to-Cl2 conversion)	8^b
ClNO2	207.867	0.19	0.10	Indirect (N2O5-to-ClNO2 conversion), TD-CRDS	28^c
ClONO2	223.862	0.20	0.36	Direct (synthesis), FT-IR	12^d
N2O5	234.886	0.12	0.03	Direct (synthesis), TD-CRDS	28^e

---

2.1 VUV lamp as reagent ion source in CIMS

Iodide reagent ions in the CIMS were generated using a VUV lamp mounted in front of the IMR. The VUV lamp (10.6 eV photon energy, PKR 106, Heraeus, Germany) was powered with a current of 2 mA at 1500 V DC. An electronic pre-filter was used to reduce noise from the switching power supply (Fig. S1).

Two major changes were implemented in our setup in contrast to the VUV lamp assembly described by Ji et al.⁵⁹ First, an acceleration voltage of 5 V was used for the ions to guide them into the IMR, which doubled the reagent ion yield. Second, the reagent ion flow was directed towards the VUV lamp before leaving the photo-ionization region. This has the advantage of maximizing the irradiance of the reagent gas shortly before the ions enter the IMR and requires less accurate alignment of the VUV lamp in its stainless-steel tube. A detailed overview of the differences between the VUV lamp and ²¹⁰Po ion sources with regard to the characteristics of Iodide-CIMS spectra can be found in Ji et al.⁵⁹

2.2 Calibration of chlorine-containing species

Relative humidity dependent calibrations of Cl₂, ClNO₂, ClONO₂, HOCl and N₂O₅ were conducted post-campaign. In addition, we examined the effect of adding O₃ to flows of ClNO₂ and HOCl to assess potential secondary ion chemistry including the formation of IClONO₂⁻ and IHO₂Cl⁻ (presented in SI). A summary of calibration curves for ClNO₂, ClONO₂, Cl₂ and HOCl can be found in Fig. S3. Higher mixing ratios than found in the atmosphere were used to enable simultaneous detection with less sensitive but absolute methods e.g., CRDS and FT-IR with dynamic dilution for the CIMS.

An expression for the humidity-dependence of the calibration signal (for constant trace-gas mixing ratio) is given by polynomial fits (Fig. S4) that cover the full ambient humidity, i.e., IH₂O⁻/(I⁻ + IH₂O⁻) range so that no extrapolation beyond these humidities was necessary. Uncertainties associated with the humidity-dependent calibrations and the detection limits during the campaign and laboratory calibration are presented in Table 1.

2.2.1 Cl₂. A canister of Cl₂ in N₂ was prepared by diluting a 5%, commercially obtained mixture in N₂ to 6 ppm. A small flow from the canister was dynamically diluted prior to simultaneous detection with the CIMS and a cavity ring-down spectrometer⁶⁰ operating at 405 nm where Cl₂ has a well-known absorption cross-section.

2.2.2 N₂O₅. N₂O₅ was synthesized by flowing a mixture of NO and a large excess of O₃ into a 2 L glass reaction chamber where they reacted for ≈1 min.⁶¹ The exhaust, containing N₂O₅ and O₃, passed through a glass tube at −78 °C (dry ice/acetone bath) to trap the N₂O₅ as pure crystals. For calibration of the CIMS, the head-space above the N₂O₅ crystals (at −60 °C) was flushed with N₂ carrier gas to generate a dilute N₂O₅ gas mixture,⁵⁸ which was simultaneously sampled by the CIMS and a thermal dissociation cavity ring down spectrometer designed for measurement of N₂O₅ and NO₃.

2.2.3 ClNO₂. The CIMS was calibrated for ClNO₂ by converting a known amount of N₂O₅ (from a prior N₂O₅ calibration) to ClNO₂. This was achieved flowing the N₂O₅ flow through a glass reactor internally coated with dried NaCl at room temperature. The ClNO₂ generated this way was assumed to be equivalent in concentration to the N₂O₅ lost as shown in laboratory experiments investigating the heterogeneous uptake of N₂O₅ to dry sea-salt.^54,62

2.2.4 ClONO₂. ClONO₂ was synthesized by allowing Cl₂O to react with N₂O₅ for several hours at −10 °C. Cl₂O was prepared by passing a flow of gaseous Cl₂ through a sample of HgO dispersed in glass-beads at −10 °C.⁵⁷ The ClONO₂ sample was used to generate a diluted sample in N₂ in a glass-bulb. The contents of the bulb were analyzed using a FTIR-spectrometer (Bruker Vector 22, Bruker Corporation, USA) at 0.5 cm⁻¹ resolution. The absorption features of ClONO₂ were consistent with those reported by Davidson et al.,⁵⁷ whose integrated band-strengths were used to quantify the ClONO₂ mixing ratio in the glass-bulb to be 200 ppmv. The CIMS was calibrated by sampling ClONO₂ from the storage bulb with known dilutions in N₂.

2.2.5 HOCl. HOCl was generated by bubbling 30 sccm N₂ through a NaOCl solution (11–14% active chlorine) and diluting this flow with 10 sLpm N₂ (Fig. S5) The gaseous mixture was then directed through a glass U-tube at −40 °C to condense moisture prior to sampling by the CIMS. To quantify the HOCl content, the HOCl flow was alternatively directed through a U-tube containing a frozen, diluted HCl solution at −40 °C, which converts HOCl to Cl₂. The amount of HOCl was determined by measuring the depletion of HOCl and the generation of Cl₂ (for which the CIMS was already calibrated) according to Lawler et al.¹¹ It was crucial to remove moisture before directing the HOCl gas to the iced HCl surface, to prevent Cl₂ depletion due to uptake on moist surfaces. The uncertainty associated with the HOCl calibration does not consider the potential non-stoichiometric conversion to Cl₂ in the second U-tube, which, given the high vapor pressure and non-polarity of Cl₂ is expected to be negligible.

2.3 BISTUM24 field campaign: measurement site and instrumentation

The BISTUM24 field campaign took place from June 6 to 22, 2024 in Spielberg, central Germany (50°19′N, 9°15′E, Fig. S6). The rural measurement site, which is located 391 m above mean sea level on top of a hill in a treeless clearing, was selected because air masses encountered at the site originate mainly from rural regions and are rarely influenced by urban pollution despite the proximity to the Frankfurt metropolitan area (∼45 km). With a distance of ∼400 km to the next coastline, the North and Baltic Seas, the site may be described as continental with no direct marine influence (Fig. 2a and b).

Fig. 2 Upper panels: 48 h-backward trajectories indicate that air masses in the period June 6 to 14, 2024 passed over marine and industrialised source regions (a). Between June 16 to 21, 2024 the air passed over rural, continental regions (b). Lower panel: time series of irradiance (c), chlorine trace gas precursor compounds particulate chloride (pCl⁻) and N₂O₅, (d) and gaseous chlorine-containing species (e and f) measured during BISTUM24. The period between the 1st and 2nd dotted lines is designated “marine-anthropogenic”, the period between the 3rd and 4th dotted lines is “continental-unpolluted”.

The instruments used in this study sampled from inlets at a height of 6 m (Fig. S7) and were housed in the “Mobile Laboratory”.⁶³ These included the Iodide-ToF-CIMS and an Aerosol Mass Spectrometer (AMS, Aerodyne Research Inc., USA), as well as an O₃ monitor (TB205, Monitor 205 Dual Beam Ozone Monitor, 2B Technologies, Inc., USA) and a pyranometer (CMP3 Pyranometer Sensor, Campbell Scientific Inc., UK). The surface area and volume of aerosol particles were calculated from a merged size distribution from different particle sizers (Fast Mobility Particle Sizer Model 3091, TSI, Inc., USA and Optical Particle Counter Model 1.109, Grimm Aerosoltechnik GmbH, Germany).⁶⁴ Local meteorological information was provided by a meteorological station (WXT520, Vaisala Oyj, Finland).

The CIMS was connected to the inlet via 6 m of PTFE tubing for which the residence time was ∼1 s (corresponding to a total flow of 32 L min⁻¹). The CIMS sub-sampled 2 L min⁻¹ from this flow. The inlet line also included a PTFE-particle filter, which was changed after every 20 000 liters of sampled air. While heterogeneous conversion of chloride to chlorine trace gases on filters or inlet lines can be an issue in coastal regions with high sea-salt loadings,^11,65 this is not expected to occur at the present location where chloride is not associated with sea-salt. The non-observation of ClNO₂ formation despite the presence of high concentration of N₂O₅ and the use of “aged” inlet lines supports this assumption. Zero measurements were performed every 8 h for 20 min during the campaign using humidified N₂. The CIMS data were zero-corrected, and the signal at each mass analyzed was normalized to the sum of the most abundant reagent ions, I⁻ and IH₂O⁻. Calibration factors were applied, determined as described in Section. 2.2.

3 Chlorine-containing compounds and air mass origin

Fig. 2 provides a time series of the chlorine-containing trace gases ClNO₂, ClONO₂, HOCl, Cl₂ and N₂O₅, measured by the CIMS, as well as particulate chloride (pCl⁻) as measured by the AMS. Although the AMS does not efficiently detect pCl⁻ from sea salt, it can detect pCl⁻ from, e.g., ammonium chloride. For this reason, pCl⁻ does not necessarily represent the total amount of particulate chloride present in the particle phase. Information on airmass origin was gained using HySplit^66,67 with which 48-h backward trajectories ending 30 m above ground level at the measurement site were calculated.

In the first phase (June 6 to 14), back-trajectories showed that the air passed over the English Channel and industrialized regions in Belgium and the Netherlands, as well as the German Ruhr metropolitan area before reaching the measurement site. The transport times were 12 to 36 h for UK and Benelux and 4 to 6 h for the Ruhr region, respectively. During transport, trajectories rarely exceeded altitudes of 500 m above mean sea level, indicating that the air was transported within the boundary layer. We refer to the air masses in this period as having a “marine-anthropogenic” influence as evidenced by the enhanced levels of pCl⁻ and N₂O₅ (Fig. 2d).

In contrast, backward trajectories for the period June 16 to 21 indicate that the air sampled during this period travelled at altitudes > 1000 m above mean sea level for several days before descending close to the measurement site. The vast majority of these trajectories passed over the relatively sparsely populated, non-industrialized center of France and rural south-west Germany (Fig. 2b and S8b). We designate these air masses as “continental-unpolluted”. In these air masses, N₂O₅ was generally below the detection limit and concentrations of pCl⁻ were much lower than those of the other air masses.

As expected, the mixing ratios of Cl₂ and ClNO₂ were largest during the “marine-anthropogenic” period, with ClNO₂ exceeding 70 pptv every night and Cl₂ exceeding 0.5 pptv every day (Fig. 2e). HOCl and ClONO₂ were less abundant than ClNO₂, with mixing ratios of up to 60 pptv (Fig. 2f). During this period, mean mixing ratios of 56.2 pptv ClNO₂, 0.64 pptv Cl₂, 41.3 pptv HOCl and 4.8 pptv ClONO₂ were observed, whereas during the “continental-unpolluted” period, 2.25 pptv ClNO₂, 0.17 pptv Cl₂ and 11.1 pptv HOCl were measured and ClONO₂ levels were below the detection limit.

Worldwide, the mixing ratios of inorganic chlorine species vary considerably depending on the air masses sampled and the geographical location. Our ClNO₂ mixing ratios are consistent with those observed at continental sites with similar (indirect) marine-influenced air masses, which show peak ClNO₂ concentrations of 300–700 pptv.^13,14 These mixing ratios are higher than those measured in maritime air,^22,48 which results from a longer residence time over landmasses with a greater anthropogenic influence, which is crucial for the formation of ClNO₂ (see Section. 1). Consequently, for highly polluted sites like in Hong Kong or Los Angeles up to 3.5 ppbv of ClNO₂ have been reported.^68,69

During both periods of the BISTUM24 campaign, the Cl₂ levels were generally low compared to those reported in other studies, especially those influenced by highly polluted urban regions.^48,70 Cl₂ mixing ratios in the marine boundary layer vary from ∼2 to ∼500 pptv^{11,48,69,71,72} and are enhanced compared to rural continental sites ranging from ∼2 to ∼33 pptv,^22,73 depend heavily on the agricultural land use in region. There are no previous reports of ClONO₂ measurements in lower troposphere with which to compare our data.

Results from our HOCl measurements are broadly consistent with the 2–120 pptv HOCl range reported for continental and marine air masses.^11,48 The large variability is related to air mass history, even when only marine regions are considered.

Fig. 2 reveals substantial differences in the diel profiles of the trace gases measured. N₂O₅ and ClNO₂ were observed predominantly at nighttime whereas HOCl and Cl₂ were observed predominantly during the day. The absence of N₂O₅ during the day is expected ((R2), Section 1) and, as N₂O₅ is the precursor of ClNO₂, the nighttime dominance of the latter is also consistent with present knowledge of its production and loss terms (Section 1). pCl⁻ has a less clear day–night variation which is related to the presence of multiphase equilibria, whereby chloride is present as a dissolved ion in the particle phase or degassed from aerosols as HCl. For the mixing ratios of ClONO₂, the two strongest features were observed after midnight on the nights of 13th and 14th June with strong co-variance with ClNO₂. This was unexpected, as ClONO₂ is believed to be formed exclusively in the reaction of ClO with NO₂, with the ClO-radical being present at significant levels only during the daytime when photochemically generated Cl atoms react with O₃.

While many studies have reported ClNO₂ and Cl₂ in coastal regions^39,48,54,74 or in highly polluted regions where non-marine chloride sources exist,^12,32,33,75 only few^10,13,22 have identified reactive chlorine trace-gases in air masses that may be described both as non-marine (i.e., several hundred km from coastal regions) and also not impacted by local pollution. Of these, only Phillips et al.¹³ investigated chlorine chemistry in mainland Europe. They argued that, as coarse mode sea-salt particles are not efficiently transported hundreds of km, the source of particle chloride was HCl that was released from coarse-mode sea-salt particles close to coastal regions (by, e.g., acid displacement) and subsequently transported as gas phase HCl or particulate ammonium chloride. As Phillips et al.¹³ did not measure other chlorine-containing trace-gases apart from ClNO₂, they did not consider the potential role of HOCl, Cl₂ or ClONO₂ in, e.g., the production rate of Cl atoms or OH radicals.

4 The diel cycle of Cl₂, HOCl, ClNO₂ and ClONO₂

Fig. 3 displays the mean, fractional contribution of each individual chlorine-containing trace gas to the total inorganic chloride mixing ratios during BISTUM24. For the “marine-anthropogenic” influenced phase (a), ClNO₂ is dominant during the night (max. ∼70%), with HOCl being dominant during the day (max. ∼90%) and ClONO₂ and Cl₂ representing only a few percent of overall chlorine-containing species throughout the diel cycle. During the “continental-unpolluted” phase (b), HOCl is the dominant chlorine-containing trace gas both day and night, accounting for > 60% during the night and >80% during the day. Both ClONO₂ and Cl₂ are minor contributors with ∼1% each, while the nocturnal ClNO₂ contribution is relatively low due to N₂O₅-poor conditions (∼30%). The diel profile of ClNO₂ is understood in terms of its generation through multiphase chemistry involving N₂O₅ during the night, followed by its daytime photolysis to release Cl atoms that (via reactions of HO₂ and ClO) lead to daytime HOCl. The significant nighttime contribution of HOCl is likely related to incomplete loss from the previous day rather than nocturnal formation, for which a non-radical mechanism is unknown. The heterogenous loss rate coefficient (k_het) for HOCl with respect to uptake on aerosol particles was calculated with eqn (1):

k_het = 0.25γAc_HOCl (1)

Fig. 3 Diel cycle of the fractional contribution of inorganic chlorine species to the summed mixing ratio. The data are separated into air masses designated “marine-anthropogenic” (a) and “continental-unpolluted” (b). ClONO₂ in (b) is at the detection limit.

with the mean molecular velocity c_HOCl = 34 700 cm s⁻¹, derived from the molecular weight of the gas species for 298 K, the mean aerosol particle surface during the campaign A = 130 µm² cm⁻³, which is representative also for nighttime (Fig. S10), and the uptake coefficient γ < 2 × 10⁻⁴ that is available for particles with pH = 1.³⁴

The calculated lifetime of HOCl due to uptake on acidic particles (k_het⁻¹ ≈ 5 days) is much longer than the lifetime with respect to photolysis (∼1 h, Table 2). Hence, the observed nighttime HOCl was most likely generated photochemically the previous day.

Table 2 Photolysis rate J and lifetime of photolabile species at noon from.⁷⁶

Species	J s^−1	Lifetime min
a Thermolabile N2O5 forms NO3 during daytime in an equilibrium. Fast photolysis of NO3 significantly decreases the N2O5-lifetime.
O3(^1D)	3.52 × 10^−5	473
ClNO2	4.97 × 10^−4	33.5
ClONO2	4.97 × 10^−5	335
Cl2	2.43 × 10^−3	6.9
HOCl	2.86 × 10^−4	58.3
NO3	2.37 × 10^−2	0.7
N2O5	4.78 × 10^−5	349^a

---

4.1 ClNO₂

In the following, we investigate the formation and loss of ClNO₂ through the diel cycle during the “marine-anthropogenic” period and highlight the influence of air mass history and vertical mixing on the ClNO₂ abundance during two individual night–day transitions.

In Fig. 4 we present the median, diel cycles of the mixing ratios of ClNO₂, ClONO₂ and N₂O₅ over the entire “marine-anthropogenic” time period. The diel cycle is divided into three phases which represent (I) nighttime, when N₂O₅ is present, (II) the period just after sunrise when N₂O₅ is rapidly depleted and the mixing ratios of ClNO₂ and ClONO₂ are either constant or increasing, and phase III when irradiance is stronger and ClNO₂ mixing ratios decrease monotonically to zero.

Fig. 4 Diel cycle of ClNO₂ (black) and ClONO₂ (pink), N₂O₅ (blue) and irradiance (yellow) during the “marine-anthropogenic” period of the BISTUM24 campaign in local time. Lines represent the median; the interquartile ranges are given as shaded area.

The formation of ClNO₂ depends on the presence of N₂O₅ and pCl⁻ (R1) and thus we expect net-production only when N₂O₅ and pCl-are present. The data in phase I and II from June 10 and 12, 2024 show that this is not always reflected in the ClNO₂ mixing ratios which remain constant or sometimes even increase during the early morning when N₂O₅ is absent, although the photolytic loss of ClNO₂ increases with intensifying irradiance. Note that, subsequent to a post-dawn peak, the ClNO₂ mixing ratios decrease with a decay constant of 5 × 10⁻⁴ s⁻¹ in accord with J_ClNO2 (Table 2).

At the beginning of the night from June 12, from 21 : 00 to 01 : 30 (Fig. 5a), an approximately linear increase in ClNO₂ of (19 ± 1) pptv h⁻¹ was observed. The ClNO₂ production rate can be calculated from k_het, the N₂O₅ mixing ratio and the fraction α of N₂O₅ that forms ClNO₂:

P(ClNO₂) = [N₂O₅]αk_het (2)

where k_het, was calculated from the uptake coefficient for N₂O₅ to an aqueous surface γ = 0.02,³⁴ the aerosol surface area in this period A = 104 µm² cm⁻³ and the mean molecular velocity of N₂O₅ c_N2O5 = 27 800 cm s⁻¹. Considering a mean [N₂O₅] = 18 pptv, and α = 0.8 given for a 0.22 M chloride in particles,⁶² a ClNO₂ production rate of 7.7 pptv h⁻¹ is obtained. Given the factor 3 uncertainty in γ^34,77 and the AMS uncertainty in pCl⁻ and thus,⁷⁸ the calculated P(ClNO₂) is consistent with the observed, net ClNO₂ formation rates.

Fig. 5 Mixing ratios of ClNO₂ (black), ClONO₂ (dark blue) and pCl⁻ (as mass (pink) and molar (red) concentration) through the diel cycle on June 10 (a) and 12 (b), 2025.

After 01 : 30 the ClNO₂-to-N₂O₅ ratio switches from ∼2 to ∼6, and the high ClNO₂ levels in the next 4 h cannot be explained by local production from N₂O₅. This trend continues in phase II in Fig. 5a, where an increase in pCl⁻ (up to 1 M Cl⁻) was accompanied by an increase in ClNO₂ levels although N₂O₅ levels were close to zero. A similar picture emerges from the data in Fig. 5b, in which post-dawn peaks in ClNO₂ (∼55 pptv between 05 : 00 and 06 : 00 and ∼69 pptv between 08 : 20 and 09 : 20) are observed despite the production term (in the absence of N₂O₅) being zero. In both cases, the in situ production of ClNO₂ from N₂O₅ reaction with pCl⁻ can be ruled out and the increase in ClNO₂ after dawn is attributed to the sampling of air masses in which N₂O₅ had been previously converted to ClNO₂. This simply reflects the fact that ClNO₂ is substantially longer-lived during the day than N₂O₅ as J_ClNO2 is much lower than J_NO3 (which controls the N₂O₅ lifetime due to its equilibrium with N₂O₅).

The phase II, post-dawn peaks in ClNO₂ were found to be enhanced for those days when backward trajectories showed subsiding air masses 1–3 h before the air reached the measurement site (see interquartile range in Fig. 4). The air from higher altitudes entrains O₃-rich residual layer air into the nocturnal boundary layer,⁷⁹ which results in enhanced formation of N₂O₅ and thus of ClNO₂. Additionally, at dawn when the nocturnal boundary layer dissipates through radiative heating and convective mixing of air from the residual layer, O₃-rich air may be transported to the surface. The daily development and breakdown of the nocturnal boundary layer can also cause variability in ClNO₂ levels due to vertical mixing.

4.2 ClONO₂

While ClNO₂, has been measured in the lowermost troposphere on many occasions,^10,12–14 the detection of ClONO₂ during BISTUM24 was unexpected. Previous studies have demonstrated the potential of using CIMS methods with SF₆⁻ and I⁻ to measure ClONO₂ as FClNO₃⁻ and IClNO₃⁻,^15,75 and airborne SF₆⁻-CIMS measurements have quantified ClONO₂ in the upper troposphere where its presence (at levels up to 10 pptv) was explained by the entrainment of stratospheric air.¹⁵

Our study presents the first quantitative ambient ground-level ClONO₂ measurements using Iodide-CIMS. As already mentioned, the accepted mechanism for formation of atmospheric ClONO₂ is (R8), which requires the presence of ClO and NO₂. ClO is a radical that is formed as a result of photochemical processes and expected to be present predominantly during daytime; we therefore expect reaction (R8) to be responsible for the net increase in ClONO₂ observed during the day (Phase III in Fig. 5). However, following this line of reasoning ClONO₂ is not expected to be formed in significant rates during the night, which we nonetheless observe (Phase I in Fig. 5). In the following, we first calculate the ClONO₂ lifetime and then estimate the production rate that would be required to maintain the observed levels. The photolysis frequency of ClONO₂ is 4.97 × 10⁻⁵ s⁻¹, resulting in a lifetime of ∼5–6 h with respect to this loss process ((R13), Table 2). The heterogeneous loss frequency of ClONO₂ was calculated from eqn (1) with the uptake coefficient on aqueous particles γ = 0.027 and c_ClONO2 = 25 590 cm s⁻¹.³⁴ Using a campaign mean particle surface area density A of 130 µm² cm⁻³ results in a loss rate constant of k_het = 2.25 × 10⁻⁴ s⁻¹, or a lifetime of ∼75 min. Given this short lifetime with respect to heterogeneous uptake, ClONO₂ cannot survive from the previous day and, in the absence of sources, its mixing ratios should tend to zero in the first hours during the night. In order to explain the observed mean ClONO₂ mixing ratio of 4.3 pptv at 05 : 00 during the night, we require a nocturnal production rate of ∼2.4 × 10⁴ molecules cm⁻³ s⁻¹, which can be equated to P(ClONO₂) = k_NO2+ClO × [ClO] × [NO₂]. Assuming an NO₂ mixing ratio of 1–2 ppbv during the “marine-anthropogenic” period as measured at a similar site with similar prevailing wind direction⁸⁰ and that the sole source of ClONO₂ is the reaction between ClO and NO₂ (with a rate coefficient, k_NO2+ClO, at 298 K and 1 bar of 2.4 × 10⁻¹² cm³ molecule⁻¹ s⁻¹, IUPAC) we calculate that the ClO concentration necessary to maintain this level of ClONO₂ is [ClO] = (2–4) × 10⁵ molecules cm⁻³, where the spread in values results from the range of NO₂ mixing ratios assumed for the calculation.

A well-known source of gas phase ClO is the reaction between Cl atoms and O₃ (R6). Thus, the presence of nighttime ClONO₂ would require a reaction that generates either Cl atoms or ClO directly, both in the absence of sunlight. The conversion of Cl to ClO is directly influenced by O₃ levels and competing Cl reactions with, e.g., hydrocarbons and NO_X trace gases. The loss rate constant of Cl through its reaction with O₃ is given as k_Cl+O3 × [O₃] = 11.8 s⁻¹ (calculated from the mean average O₃ concentration (40 ppbv) and the rate coefficient k_Cl+O3 = 1.2 × 10⁻¹¹ cm³ molecule⁻¹ s⁻¹)³⁴.

The first-order loss rate constant of Cl atoms reacting with the most abundant alkanes can be calculated from k_Cl+CH4 × [CH₄] + k_Cl+C2H6 × [C₂H₆] + k_Cl+C3H8 × [C₃H₈]. As concentrations of these hydrocarbons were not measured during the campaign, we use reference mixing ratios of 1.9 ppm CH₄, 0.8 ppbv C₂H₆, and 0.3 ppbv C₃H₈ that are typical for rural continental conditions and the corresponding rate coefficients with Cl.^34,81 The calculated total loss rate constant for reaction with these hydrocarbons is ∼6.8 s⁻¹, with the largest single contributions from methane (4.6 s⁻¹, i.e., ∼25% of Cl). We acknowledge that loss of Cl atoms due to terpenoids can be significant, however, their concentration is highly variable (even in rural continental regions) and their contribution is highly uncertain (considering 1 ppbv terpenoids with k_Cl+isoprene = 4.6 × 10⁻¹⁰ molecules cm⁻³ s⁻¹ would result in a loss rate to terpenoids of ∼11.2 s⁻¹,⁸² i.e., the same as the loss rate to O₃ adding a uncertainty factor of 2 to the calculation). Cl atoms react with NO₂ with a rate coefficient of 6.7 × 10⁻¹² cm³ molecule⁻¹ s⁻¹ campaign conditions.⁸³ Assuming (as above) that NO₂ mixing ratios were between 1 and 2 ppbv at night results in a Cl atom loss rate constant of ∼0.3 s⁻¹.

Taking the formation rate of ClO from Cl (11.8 s⁻¹) and the loss rates of Cl through other pathways (6.8 s⁻¹ and 0.3 s⁻¹), we can conclude that approximately 60% of the Cl atoms generated react to form ClO. We can estimate the production rate of ClO by assuming that its concentration is in steady state with respect to its gas phase formation and loss processes at night. In the absence of NO, the dominant loss process for ClO at night is its reaction with NO₂, with a loss rate of 2.4 × 10⁴ molecules cm⁻³ s⁻¹. Using the 60% ClO yield from Cl, this in turn can be converted into a Cl atom production rate P(Cl) = k_NO2+ClO × [ClO] × [NO₂]/0.6 = (2.0–3.9) × 10⁴ atoms cm⁻³ s⁻¹.

We now examine potential reactions that, at nighttime, could lead to Cl atom production rates of this magnitude. The dominant radical at nighttime is often considered to be NO₃, which initiates the oxidation of many biogenic hydrocarbons. However, reactions of NO₃ radicals with sufficiently abundant (non-radical) chlorine species to generate ClO or Cl atoms are either unknown or expected to be too slow to contribute significantly. For example, the rate coefficient for the reaction between NO₃ and HCl (R14) is < 5 × 10⁻¹⁷ cm³ molecule⁻¹ s⁻¹,³⁴ which would result in a maximal Cl atom production rate of 1.2 × 10² atoms cm⁻³ s⁻¹, if we assume typical mixing ratios of 100 pptv HCl^84,85 and 10 to 40 pptv NO₃,⁸⁰ which is a factor ∼350 lower than that required.

HCl + NO₃ → Cl + HNO₃ (R14)

As the reaction between NO₃ and HCl cannot explain the nighttime production of Cl atoms, we consider the potential role of OH radicals, which react much more rapidly with HCl ((R15), k_OH+HCl = 7.8 × 10⁻¹³ cm³ molecule⁻¹ s⁻¹)³⁴.

HCl + OH →Cl + H₂O (R15)

OH radicals are present at night owing to their formation in the ozonolysis of unsaturated hydrocarbons,^86,87 and their abundance has been the subject of much research, with concentrations found to be varying between 9.1 × 10⁴ and 7 × 10⁶ molecules cm⁻³ at night time.^88–91 Due to the absence of OH measurements during BISTUM24, we assume a nighttime concentration of 1 × 10⁵ molecules cm⁻³ that was determined in a similar agricultural rural environment.⁸⁸ For the HCl mixing ratio, we assume a value of 100 pptv HCl (as measured over the continent⁸⁵) and calculate a Cl production rate of 1.9 × 10² atoms cm⁻³ s⁻¹, which is a factor 100–200 smaller than the value of (2.0–3.9) × 10⁴ atoms cm⁻³ s⁻¹ derived above.

The nocturnal ozonolysis of unsaturated hydrocarbons also leads to the formation of Crigee-intermediates.^87,92 Studies of the reaction of HCl with Crigee-intermediates appears however not to form Cl or ClO but R₁R₂OOCl compounds, where R₁ and R₂ are organic groups.⁹³

In addition, we also consider the direct formation of ClO through the reaction of OH with HOCl (R16):

HOCl + OH → ClO + H₂O (R16)

for which the rate coefficient at 298 K is k_OH+HOCl = 5.0 × 10⁻¹³ cm³ molecule⁻¹ s⁻¹.³⁴ Assuming again an OH concentration of 1.0 × 10⁵ molecules cm⁻³ during night⁸⁸ and nocturnal campaign mean mixing ratio of 50 pptv for HOCl we derive a ClO production rate of 1.2 × 10³ ClO cm⁻³ s⁻¹ which is a factor ten larger than the rate of formation of Cl from OH + HCl but falls well short of the value required to explain the observations of nighttime ClONO₂.

We thus conclude that the presence of ClONO₂ at nighttime, despite its rapid heterogeneous loss to particles, cannot be explained by the OH-induced conversion of HCl and HOCl to Cl and ClO, even within the major uncertainties related to the estimations of the HCl and OH levels. We note however, that nighttime Cl levels of 2.0 × 10⁴ atoms cm⁻³ have been reported in the urban UK,⁹¹ although the authors were not able to name a Cl source. Such levels of Cl result in a ClO production rate of 2.4 × 10⁵ molecules cm⁻³ s⁻¹ which is a factor 10 larger than at required to explain our observations.

At this point, we examine the possibility that ClONO₂ is actually significantly longer lived at night than so far calculated based on an uptake coefficient of γ = 0.027,³⁴ which is the result of laboratory experiments performed on inorganic aqueous solutions containing sulphate, halides etc. We now draw analogy to the heterogeneous uptake to aqueous particles of another di-acid anhydride, N₂O₅ for which γ from laboratory studies is also very similar (∼0.02).³⁴ In contrast to ClONO₂, the tropospheric, multiphase chemistry of N₂O₅ has been the subject of many studies and the rate of uptake to “real” atmospheric particles that are internal mixtures of organics, sulphates, nitrates etc. is a strong function of composition, where the presence of organic has been observed to reduce γ significantly.⁹⁴ As the mechanism of uptake of N₂O₅ and ClONO₂ to aqueous particles is expected to be similar (both hydrolyze/ionize on/at the surface) it is very possible that γ and the rate of uptake of ClONO₂ is greatly reduced compared to the values calculated above, and that the source term required to generate the observed mixing ratios of ClONO₂ is actually lower. In the absence of laboratory studies investigating the dependence of γ(ClONO₂) on e.g., particle organic content, this remains speculative. In addition, while no appropriate mechanism is known to us, we cannot rule out alternative (potentially multiphase) routes to ClONO₂ formation at nighttime may be more important than the gas phase reaction between ClO and NO₂.

However, to illustrate the potential importance of nocturnal Cl production we calculate its effect on the methane lifetime. Based on the measurement of ClONO₂, we derived a nocturnal Cl production rate of (2.0–3.9) × 10⁴ atoms cm⁻³ s⁻¹ (see above). We can combine this with the fact that ∼25% of the Cl atoms formed will react with CH₄ (see above) to calculate a CH₄ lifetime of ∼45 years with respect to nocturnal loss via reaction with Cl atoms. Given that the lifetime (i.e., the inverse reaction rate) of the important climate gas methane is of the order of 10 years (predominantly through reaction with OH) this translates to a ∼15% fractional contribution to CH₄ loss. We note that this calculation is based on the assumption that ClONO₂ is formed solely via gas phase reaction of ClO and NO₂. A heterogeneous process forming ClONO₂ that does not require the formation of Cl atoms would have no effect on hydrocarbon oxidation rates. A more rigorous assessment of the global impact of regional chlorine activation, for example regarding the methane lifetime, requires further related observations, inclusion of realistic chlorine activation scenarios and chlorine photochemistry in global models.

4.3 Formation and diel cycle of Cl₂ and HOCl

Median diel cycles of HOCl and Cl₂ are depicted in Fig. 6 and show maxima of 60–70 pptv HOCl and 0.8–1.0 pptv Cl₂ in the afternoon (Fig. 6), implying that both are largely of photochemical origin. Strong correlations of Cl₂ (R² = 0.88) and HOCl (R² = 0.85) with O₃ (Fig. S11) indicate a photochemical source.

Fig. 6 Diel cycle of HOCl (grey) and Cl₂ (blue) mixing ratios and irradiance (yellow). Lines represent the median; the interquartile ranges are given as shaded area. The diel cycles only consider data from the “marine-anthropogenic” period of the BISTUM24 campaign.

Cl₂ is known to be formed in the reaction between gaseous chlorine species such as ClNO₂, ClONO₂ and HOCl with pCl⁻ in acidified aerosol, which may help to explain the timing of the peak Cl₂ mixing ratio in the afternoon (phase III) after maximum irradiance ((R4), (R9), (R10), and Oum et al.³⁰). In addition to the daytime peak, a slight increase in Cl₂ levels was monitored between 04 : 00 to 07 : 00, i.e., between phase I and II. This observation may also be linked to the entrainment of O₃/N₂O₅-rich air from the residual layer into the nocturnal boundary layer (Section. 4.1) in agreement with Xia et al.,⁹⁵ so that particle chloride can be converted into chlorine-containing gases such as Cl₂ and ClNO₂. Cl₂ is also formed in other multiphase processes involving chloride and the intermediary of NO₃⁻, O(³P), Fe(III) and TiO₂.^47,96–99

For HOCl, the recognized and only atmospherically relevant route to its generation in the gas phase is through the reaction between two radical species, ClO and HO₂ (R7) which takes place predominantly during the daytime when active photochemistry leads to high radical concentrations. The particle phase hydrolysis of ClONO₂ (R11) also leads to the formation of HOCl, which can either degas from the particles or react with particle chloride to form Cl₂. As Cl₂ is weakly soluble, this will likely transfer to the gas phase where it is rapidly photolyzed (Table 2) to reform Cl atoms that, in turn, reform HCl and the HOCl precursor ClO. As HOCl is also photolabile, its gas phase mixing ratios are determined by chemical cycles that are strongly dependent on actinic flux and the availability of, e.g., O₃.

5 Daytime Cl-atom production rates

While the OH radical is the primary driver of oxidation of the vast majority of atmospheric trace gases, the oxidation rates of certain trace gases can be augmented by the presence of Cl atoms. In this section, we examine Cl-atom production rates P(Cl) from the photolysis of ClNO₂, ClONO₂, HOCl and Cl₂ ((R3), (R5), (R12) and (R13)) over the diel cycle:

P(Cl) = J_ClNO2[ClNO₂] + J_ClONO2[ClONO₂] + 2J_Cl2[Cl₂] + J_HOCl[HOCl] (3)

The photolysis frequencies J of ClNO₂, ClONO₂, HOCl and Cl₂ were derived for each trace gas through the diel cycle (data provided by the TUV calculator for cloudless conditions for the BISTUM24 location⁷⁶), corrected for cloud-reduced irradiance, measured using a local on-site pyranometer. We recognize that clouds do not impact irradiance in the same way for all wavelengths; however, considering cloudiness provides a more accurate estimate of real photolysis rates.

The results of the calculations are summarized in Fig. 7 which shows that Cl production follows the actinic flux, reaching a maximum of (1.0 ± 0.47) × 10⁶ atoms cm⁻³ s⁻¹ in the period in which we sampled “marine-anthropogenic” air. The maximum for “continental-unpolluted” air is a factor ∼6 lower (1.6 ± 0.7) × 10⁵ atoms cm⁻³ s⁻¹. The daytime maximum Cl production rates are broadly consistent with summertime observations from Los Angeles¹⁰⁰ (6.2 × 10⁵ atoms cm⁻³ s⁻¹), Houston²² (4.7 × 10⁵ atoms cm⁻³ s⁻¹), Beijing⁷⁵ (1.6 × 10⁵ atoms cm⁻³ s⁻¹) and wintertime Manchester-UK³⁹ (3.0 × 10⁵ atoms cm⁻³ s⁻¹). Note however, that the values for Houston considered only Cl₂ and ClNO₂ as Cl sources,²² while additionally, for Los Angeles HCl + OH contributes significantly.¹⁰⁰ In the Beijing-study a minor role was attributed to HOCl and ClONO₂ (not quantified and considered in Le Breton et al.⁷⁵) and ClNO₂ was considered to be the only Cl sources in marine urban Manchester.³⁹ The fact that Cl production rates were observed in polluted and marine influenced regions, as well as in rural continental areas (this work) indicates the role of Cl chemistry could be more widespread than currently assumed.

Fig. 7 Median production rates of Cl (P(Cl), blue) and OH (P(OH), red) as well as the ratio of the production rates (P(OH)/P(Cl), black) for the air masses with marine-anthropogenic (a) and with continental-unpolluted influence (b). The shaded areas represent interquartile ranges, while the dotted line indicates equal production of Cl and OH.

The fractional contribution of each chlorine-containing trace gas to Cl atom production rates is highlighted in Fig. 8. HOCl is the major contributor to Cl production from photolysis, accounting in the morning for 10% for “marine-anthropogenic” air masses and 49% for air masses from “continental-unpolluted” source regions. After ∼12 : 00, the HOCl contribution is in both cases between 70% and 80% (see Fig. 8a and b).

Fig. 8 Fractional contribution of various inorganic chlorine species to the diurnal total chlorine production rate for air masses with marine-anthropogenic (a) and with continental-unpolluted influence (b).

Contrary to HOCl, the contribution of ClNO₂ to Cl production decreases from 85% at sunrise (07 : 00) to 3% after 13 : 00 if the air masses were transported from the “marine-anthropogenic” source region. The contribution of ClNO₂ in continental unpolluted air, meanwhile, is significantly lower, ranging from 6% to 33%. This observation emphasizes the importance of ClNO₂ as a nighttime reservoir species, particularly for air masses affected by polluted areas.

For air masses with marine-anthropogenic signature, Cl₂ contributes approximately 5% to Cl production rate in the morning hours, increasing to 20% from 13 : 00 on. Despite its much lower mixing ratios, after 11 : 00 the contribution of Cl₂ is larger than that of ClNO₂, which is related to the absorption cross section of Cl₂ which overlaps much more with the actinic flux (Fig. S12). The contribution of Cl₂ to Cl production rates from continental unpolluted air remains around 20% throughout the day.

Low daytime concentrations of maximum 3.3 pptv, coupled with low absorption cross-sections at the wavelengths of the actinic flux, mean that ClONO₂ plays a minor role in Cl production. Fig. 8a indicates that ClONO₂ contributes only 0.5% to Cl production, independent of the time of day, which is consistent with conclusions of Le Breton et al.⁷⁵

The impact of a summed, maximum daytime Cl production rate of ∼1 × 10⁶ atoms cm⁻³ s⁻¹ (at ∼14 : 00 local time) can also be illustrated by estimating its impact on CH₄ oxidation. As discussed in Sect. 4.2, ∼25% of the Cl atoms generated under these conditions will react with CH₄. The CH₄ loss due to Cl during one hour at noon is ∼40 pptv. This can be compared to the CH₄ loss in the same period owing to reaction with OH, which is ∼400 pptv, assuming OH concentrations of 1 × 10⁷ molecules cm⁻³ using 1.9 ppmv CH₄. While these estimations highlight the potential of Cl compared to OH regarding CH₄ oxidation, rigorous assessment of the impact of chlorine chemistry on a long-lived climate gas such as CH₄ requires implementing appropriate chemistry into global chemistry-transport models, which is beyond the scope of this manuscript.

6 Photolysis of HOCl as source of OH

As shown in reaction (R12), the photolysis of HOCl generates both a Cl atom and an OH radical. In the following, we compare the rates of formation of OH through the diel cycle airing from HOCl photolysis with the rate of OH formation from the photolysis of O₃ in the presence of water-vapor (R19), traditionally considered the major primary source of atmospheric OH.

O₃ + hv → O(¹D) + O₂ (R19a)

O(¹D) + H₂O → 2OH (R19b)

The primary photochemical production rate of OH from O₃, P_O3(OH) was calculated from the O₃ photolysis frequency J_O(¹D) modified using the rate coefficients for the O(¹D) reaction with H₂O or quenching by N₂ and O₂ in eqn (4). Rate coefficients k were taken from the IUPAC evaluation.³⁴

The diurnal cycle of the Cl and OH production rates during the “marine-anthropogenic” and “continental-unpolluted” periods (Fig. 7) show that for both periods the OH production from O₃ reaches a maximum of ∼7 × 10⁶ molecules cm⁻³ s⁻¹ around 13 : 00. For comparison, the maximum Cl (and thus OH from photolysis of HOCl) production rate is a factor of seven lower for the “marine-anthropogenic” case (Fig. 7a) and a factor 40 smaller for the “continental-unpolluted” case (Fig. 7b). However, in the early morning, in marine-anthropogenically influenced air the Cl production rate is more rapid than that of OH, which results from the red-shifted absorption spectra of Cl₂, HOCl and ClNO₂ compared to O₃ and thus better overlap with the early morning actinic flux (Fig. 7a). This is entirely consistent with the findings of Phillips et al.¹³ at a measurement site 40 km away from our site when marine-influenced air was encountered. In this early morning period, the generation of Cl atoms results in the oxidation of hydrocarbons and thus increases the rate of formation of organic peroxyl radicals that play an important role in daytime, photochemical ozone formation.

Tropospheric OH formation occurs via primary formation from O₃ photolysis as well as by recycling through reaction of HO₂ with NO and (generally less importantly) the photolysis of peroxides and HONO.^101,102 HOCl is generally not considered as a potential contributor to OH production due to a lack of HOCl data. However, Fig. 9a shows that while OH formation from photolysis of O₃ is dominant, OH formation from HOCl (compare eqn (3)) is still significant at some times of the day when the actinic flux at the shorter wavelengths needed to photolyze O₃ is relatively weak (see Fig. 9b). For marine-anthropogenically influenced air masses, the contribution of HOCl photolysis to OH formation approaches 40% at sunrise and sunset, compared to 8% at noon. For the “continental-unpolluted” air masses, the contribution of HOCl to OH ranges from 2% at noon to 17% at sunrise and sunset. This analysis shows that HOCl should be considered as a source of primary OH formation, especially during the morning and the evening hours.

Fig. 9 The median fraction of primary OH originating from HOCl for both marine-anthropogenic and continental-unpolluted air masses (a). The actinic flux (yellow) and photolysis rates of HOCl (J_HOCl, green) and O₃ (J_{O3 (¹D)}, grey) (b).

7 Conclusions

Several inorganic chlorine species (Cl₂, HOCl, ClNO₂ and ClONO₂) were measured with an Iodide-CIMS at a continental measurement site in rural central Germany, ∼400 km from the nearest coastline. The largest mixing ratios were detected when 48 h-backward trajectories indicated that the air mass had passed over marine and polluted regions. The dominant chlorine-containing trace gases were ClNO₂ during the night and early-morning and HOCl throughout the day. A consistent feature of the ClNO₂ diel profile was an early morning peak, which was associated with vertical mixing. In unpolluted air, the levels of HOCl and Cl₂ were strongly enhanced compared to ClNO₂, while ClONO₂ was not observed.

In air masses with marine-anthropogenic influence we observed not only ClNO₂ but also ClONO₂ along with enhanced particulate chlorine during the nighttime. ClONO₂ was thus observed for the first time in tropospheric air masses, albeit at low mixing ratios of 2–3 pptv. Laboratory experiments helped rule out that this observation was not an artefact resulting, e.g., from reaction of ClNO₂ with IO_X⁻ in the IMR of the instrument. Calculations of the total chlorine production rate from the photolysis of the observed species reveal values of 1.0 × 10⁶ atoms cm⁻³ s⁻¹ for air masses with marine polluted source regions, and 1.6 × 10⁵ atoms cm⁻³ s⁻¹ for continental clean ones. Depending on the air mass source region, ClNO₂ contributed 30–90% to Cl atom production up to three hours after sunrise and 5–10% throughout the rest of the day. Additionally, throughout the day, HOCl contributed 10–85% to Cl production, while Cl₂ was responsible for 5–20%.

Cl formation contributed significantly to radical generation (and thus oxidation rates) in the early morning implying significant peroxyl radical and ClO formation rates. The formation of OH radicals from photolysis of HOCl was compared to that by photolysis of O₃ and shown to make a significant contribution for air masses that had passed over maritime and polluted regions before reaching the site with a maximum contribution of 40% of the total primary OH formation at sunrise and sunset.

Consequentially, we show that N₂O₅-induced activation of chlorine from its particle reservoir into photolabile, gas phase forms has repercussions for the lifetimes of many hydrocarbons, highlighting methane in remote continental areas distant to source regions where it affects local chemical processes.

Author contributions

L. M. performed CIMS measurements, trajectory calculation, analyzed the data and drafted the manuscript. F. F. and F. D. performed ground-based measurements with the MoLa and data post-processing. J. N. C. performed CRDS, TD-CRDS and FT-IR measurements, synthesized the ClONO₂ and helped with calibration. All authors discussed the data and the presented results. All co-authors commented on the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: discussions regarding secondary ion chemistry and supporting figures. See DOI: https://doi.org/10.1039/d5ea00142k.

Data for this article, including data from all figures are available at Zenodo at https://doi.org/10.5281/zenodo.18670872.

Acknowledgements

This work was supported by internal funding from the Max Planck Society. L. M. is funded by the Deutsche Forschungs-gemeinschaft (DFG, German Research Foundation)–TRR 301 “TPChange” (Project-ID 428312742). The authors thank Thomas Böttger and Philipp Schuhmann for support during the BISTUM24 campaign, and Laura Wüst for preparation of N₂O₅ crystals (all Max-Planck-Institute for Chemistry). Additionally, the authors thank Yijing Chen (Tsinghua University), Lewis Marden and Jake Willow (both from University of York), Cheng Wu (University of Gothenburg) and Cora Young (York University) for discussions about secondary ion chemistry. Open Access funding provided by the Max Planck Society.

References

1. M. J. Molina and F. S. Rowland, Nature, 1974, 249, 810–812.
2. F. S. Rowland, Environ. Conserv., 1988, 15, 101–115.
3. S. Solomon, Rev. Geophys., 1999, 37, 275–316.
4. J. W. Elkins, T. M. Thompson, T. H. Swanson, J. H. Butler, B. D. Hall, S. O. Cummings, D. A. Fishers and A. G. Raffo, Nature, 1993, 364, 780–783.
5. A. Engel, U. Schmidt and D. McKenna, Geophys. Res. Lett., 1998, 25, 3319–3322.
6. R. G. Prinn, R. F. Weiss, P. J. Fraser, P. G. Simmonds, D. M. Cunnold, F. N. Alyea, S. O'Doherty, P. Salameh, B. R. Miller, J. Huang, R. H. J. Wang, D. E. Hartley, C. Harth, L. P. Steele, G. Sturrock, P. M. Midgley and A. McCulloch, J. Geophys. Res. Atmos., 2000, 105, 17751–17792.
7. A. Engel, M. Strunk, M. Müller, H.-P. Haase, C. Poss, I. Levin and U. Schmidt, J. Geophys. Res. Atmos., 2002, 25(17), 3319–3322.
8. M. K. W. Ko and N. D. Sze, J. Geophys. Res. Atmos., 1984, 89, 11619–11632.
9. T. Von Clarmann and S. Johansson, Atmos. Chem. Phys., 2018, 18, 15363–15386.
10. J. A. Thornton, J. P. Kercher, T. P. Riedel, N. L. Wagner, J. Cozic, J. S. Holloway, W. P. Dubé, G. M. Wolfe, P. K. Quinn, A. M. Middlebrook, B. Alexander and S. S. Brown, Nature, 2010, 464, 271–274.
11. M. J. Lawler, R. Sander, L. J. Carpenter, J. D. Lee, R. Von Glasow, R. Sommariva and E. S. Saltzman, Atmos. Chem. Phys., 2011, 11, 7617–7628.
12. L. H. Mielke, A. Furgeson and H. D. Osthoff, Environ. Sci. Technol., 2011, 45, 8889–8896.
13. G. J. Phillips, M. J. Tang, J. Thieser, B. Brickwedde, G. Schuster, B. Bohn, J. Lelieveld and J. N. Crowley, Geophys. Res. Lett., 2012, 39, L10811.
14. T. J. Bannan, A. M. Booth, A. Bacak, J. B. A. Muller, K. E. Leather, M. Le Breton, B. Jones, D. Young, H. Coe, J. Allan, S. Visser, J. G. Slowik, M. Furger, A. S. H. Prévôt, J. Lee, R. E. Dunmore, J. R. Hopkins, J. F. Hamilton, A. C. Lewis, L. K. Whalley, T. Sharp, D. Stone, D. E. Heard, Z. L. Fleming, R. Leigh, D. E. Shallcross and C. J. Percival, J. Geophys. Res. Atmos., 2015, 120, 5638–5657.
15. T. P. Marcy, R. S. Gao, M. J. Northway, P. J. Popp, H. Stark and D. W. Fahey, Int. J. Mass Spectrom., 2005, 243, 63–70.
16. M. Von Hobe, J. U. Grooß, R. Müller, S. Hrechanyy, U. Winkler and F. Stroh, Atmos. Chem. Phys., 2005, 5, 693–702.
17. R. Atkinson, D. L. Baulch, R. A. Cox, J. N. Crowley, R. F. Hampson, R. G. Hynes, M. E. Jenkin, M. J. Rossi and J. Troe, Atmos. Chem. Phys., 2007, 7, 981–1191.
18. R. Sander and P. J. Crutzen, J. Geophys. Res. Atmos., 1996, 101, 9121–9138.
19. R. von Glasow and P. J. Crutzen, in Treatise on Geochemistry, ed. H. D. Holland and K. K. Turekian, Pergamon, Oxford, 2003, pp. 1–67,  DOI:10.1016/B0-08-043751-6/04141-4.
20. C. Arsene, A. Bougiatioti, M. Kanakidou, B. Bonsang and N. Mihalopoulos, Atmos. Chem. Phys., 2007, 7, 4661–4673.
21. R. Von Glasow, R. Sander, A. Bott and P. J. Crutzen, J. Geophys. Res. Atmos., 2002, 107(D17), ACH 9.
22. C. Faxon, J. Bean and L. Ruiz, Atmosphere, 2015, 6, 1487–1506.
23. N. Bhattacharyya, M. Modi, L. G. Jahn and L. Hildebrandt Ruiz, Environ. Sci.: Atmos., 2023, 3, 1174–1185.
24. Q. Li, D. Meidan, P. Hess, J. A. Añel, C. A. Cuevas, S. Doney, R. P. Fernandez, M. van Herpen, L. Höglund-Isaksson, M. S. Johnson, D. E. Kinnison, J.-F. Lamarque, T. Röckmann, N. M. Mahowald, A. Saiz-Lopez, Q. Li, D. Meidan, P. Hess, J. A. Añel, C. A. Cuevas, S. Doney, R. P. Fernandez, M. van Herpen, L. Höglund-Isaksson, M. S. Johnson, D. E. Kinnison, J.-F. Lamarque, T. Röckmann, N. M. Mahowald and A. Saiz-Lopez, Nat. Commun., 2023, 14(1), 4045.
25. X. Wang, D. J. Jacob, S. D. Eastham, M. P. Sulprizio, L. Zhu, Q. Chen, B. Alexander, T. Sherwen, M. J. Evans, B. H. Lee, J. D. Haskins, F. D. Lopez-Hilfiker, J. A. Thornton, G. L. Huey and H. Liao, Atmos. Chem. Phys., 2019, 19, 3981–4003.
26. W. C. Keene, A. A. P. Pszenny, D. J. Jacob, R. A. Duce, J. N. Galloway, J. J. Schultz-Tokos, H. Sievering and J. F. Boatman, Global Biogeochem. Cycles, 1990, 4, 407–430.
27. W. C. Keene, R. Sander, A. A. P. Pszenny, R. Vogt, P. J. Crutzen and J. N. Galloway, J. Aerosol Sci., 1998, 29, 339–356.
28. C. Volpe, M. Wahlen, A. A. P. Pszenny and A. J. Spivack, Geophys. Res. Lett., 1998, 25, 3831–3834.
29. T. A. Cahill, K. Wilkinson and R. Schnell, J. Geophys. Res. Atmos., 1992, 97(D13), 14513–14520.
30. K. W. Oum, M. J. Lakin, D. O. DeHaan, T. Brauers and B. J. Finlayson-Pitts, Science, 1998, 279.
31. J. M. Lobert, W. C. Keene, J. A. Logan and R. Yevich, J. Geophys. Res. Atmos., 1999, 104, 8373–8389.
32. T. J. Bannan, M. A. H. Khan, M. Le Breton, M. Priestley, S. D. Worrall, A. Bacak, N. A. Marsden, D. Lowe, J. Pitt, G. Allen, D. Topping, H. Coe, G. McFiggans, D. E. Shallcross and C. J. Percival, Geophys. Res. Lett., 2019, 46, 8508–8516.
33. C. G. Masoud, M. Modi, N. Bhattacharyya, L. G. Jahn, K. N. McPherson, P. Abue, K. Patel, D. T. Allen and L. Hildebrandt Ruiz, Environ. Sci. Technol., 2023, 57, 15454–15464.
34. M. Ammann, R. A. Cox, J. N. Crowley, H. Herrmann, M. E. Jenkin, V. F. McNeill, A. Mellouki, M. J. Rossi, J. Troe and T. J. Wallington, Evaluated Kinetic and Photochemical Data Foratmospheric Chemistry, 2025, https://iupac.aeris-data.fr/.
35. S. S. Brown and J. Stutz, Chem. Soc. Rev., 2012, 41, 6405.
36. J. Liebmann, E. Karu, N. Sobanski, J. Schuladen, M. Ehn, S. Schallhart, L. Quéléver, H. Hellen, H. Hakola, T. Hoffmann, J. Williams, H. Fischer, J. Lelieveld and J. N. Crowley, Atmos. Chem. Phys., 2018, 18, 3799–3815.
37. P. Dewald, T. Seubert, S. T. Andersen, G. N. T. E. Türk, J. Schuladen, M. R. McGillen, C. Denjean, J.-C. Etienne, O. Garrouste, M. Jamar, S. Harb, M. Cirtog, V. Michoud, M. Cazaunau, A. Bergé, C. Cantrell, S. Dusanter, B. Picquet-Varrault, A. Kukui, C. Xue, A. Mellouki, J. Lelieveld and J. N. Crowley, Atmos. Chem. Phys., 2024, 24, 8983–8997.
38. X. Wang, H. Wang, L. Xue, T. Wang, L. Wang, R. Gu, W. Wang, Y. J. Tham, Z. Wang, L. Yang, J. Chen and W. Wang, Atmos. Environ., 2017, 156, 125–134.
39. M. Priestley, M. Le Breton, T. J. Bannan, S. D. Worrall, A. Bacak, A. R. D. Smedley, E. Reyes-Villegas, A. Mehra, J. Allan, A. R. Webb, D. E. Shallcross, H. Coe and C. J. Percival, Atmos. Chem. Phys., 2018, 18, 13481–13493.
40. J. D. Haskins, B. H. Lee, F. D. Lopez-Hilifiker, Q. Peng, L. Jaeglé, J. M. Reeves, J. C. Schroder, P. Campuzano-Jost, D. Fibiger, E. E. McDuffie, J. L. Jiménez, S. S. Brown and J. A. Thornton, J. Geophys. Res. Atmos., 2019, 124, 8851–8869.
41. G. Chen, X. Fan, H. Wang, Y. J. Tham, Z. Lin, X. Ji, L. Xu, B. Hu and J. Chen, Atmos. Chem. Phys., 2025, 25, 7815–7828.
42. F. Schweitzer, P. Mirabel and C. George, J. Phys. Chem. A, 1998, 102, 3942–3952.
43. B. J. Finlayson-Pitts, M. J. Ezell and J. N. Pitts, Nature, 1989, 337, 241–244.
44. M.-T. Leu, R. S. Timonen, L. F. Keyser and Y. L. Yung, J. Phys. Chem., 1995, 99, 13203–13212.
45. G. J. Phillips, J. Thieser, M. Tang, N. Sobanski, G. Schuster, J. Fachinger, F. Drewnick, S. Borrmann, H. Bingemer, J. Lelieveld and J. N. Crowley, Atmos. Chem. Phys., 2016, 16, 13231–13249.
46. J. A. Ganske, M. J. Ezell, H. N. Berko and B. J. Finlayson-Pitts, Chem. Phys. Lett., 1991, 179, 204–210.
47. X. Peng, T. Wang, W. Wang, A. R. Ravishankara, C. George, M. Xia, M. Cai, Q. Li, C. M. Salvador, C. Lau, X. Lyu, C. N. Poon, A. Mellouki, Y. Mu, M. Hallquist, A. Saiz-Lopez, H. Guo, H. Herrmann, C. Yu, J. Dai, Y. Wang, X. Wang, A. Yu, K. Leung, S. Lee and J. Chen, Nat. Commun., 2022, 13, 939.
48. G. Chen, X. Fan, S. Yu, Y. J. Tham, Z. Lin, X. Ji, L. Xu and J. Chen, Environ. Sci. Technol., 2025, 59(10), 5164–5171.
49. M. Eigen and K. Kustin, J. Am. Chem. Soc., 1962, 84, 1355–1361.
50. G. Deiber, C. George, S. Le Calvé, F. Schweitzer and P. Mirabel, Atmos. Chem. Phys., 2004, 4, 1291–1299.
51. M. A. Tolbert, M. J. Rossi, R. Malhotra and D. M. Golden, Science, 1987, 238, 1258–1260.
52. R. J. Cicerone, Rev. Geophys., 1981, 19, 123–139.
53. T. E. Caldwell, K. L. Foster, T. Benter, S. Langer, J. C. Hemminger and B. J. Finlayson-Pitts, J. Phys. Chem. A, 1999, 103, 8231–8238.
54. J. P. Kercher, T. P. Riedel and J. A. Thornton, Atmos. Meas. Tech., 2009, 2, 193–204.
55. B. H. Lee, F. D. Lopez-Hilfiker, C. Mohr, T. Kurtén, D. R. Worsnop and J. A. Thornton, Environ. Sci. Technol., 2014, 48, 6309–6317.
56. D. Maric, J. P. Burrows, R. Meller and G. K. Moortgat, J. Photochem. Photobiol., A, 1993, 70, 205–214.
57. J. A. Davidson, C. A. Cantrell, R. E. Shetter, A. H. McDaniel and J. G. Calvert, J. Geophys. Res. Atmos., 1987, 92, 10921–10925.
58. N. Sobanski, J. Schuladen, G. Schuster, J. Lelieveld and J. N. Crowley, Atmos. Meas. Tech., 2016, 9, 5103–5118.
59. Y. Ji, L. G. Huey, D. J. Tanner, Y. R. Lee, P. R. Veres, J. A. Neuman, Y. Wang and X. Wang, Atmos. Meas. Tech., 2020, 13, 3683–3696.
60. J. Thieser, G. Schuster, J. Schuladen, G. J. Phillips, A. Reiffs, U. Parchatka, D. Pöhler, J. Lelieveld and J. N. Crowley, Atmos. Meas. Tech., 2016, 9, 553–576.
61. J. A. Davidson, A. A. Viggiano, C. J. Howard, I. Dotan, F. C. Fehsenfeld, D. L. Albritton and E. E. Ferguson, J. Chem. Phys., 1978, 68, 2085–2087.
62. J. M. Roberts, H. D. Osthoff, S. S. Brown, A. R. Ravishankara, D. Coffman, P. Quinn and T. Bates, Geophys. Res. Lett., 2009, 36, L20808.
63. F. Drewnick, T. Böttger, S. L. von der Weiden-Reinmüller, S. R. Zorn, T. Klimach, J. Schneider and S. Borrmann, Atmos. Meas. Tech., 2012, 5, 1443–1457.
64. J. Pikmann, F. Drewnick, F. Fachinger and S. Borrmann, Atmos. Chem. Phys., 2024, 24, 12295–12321.
65. G. C. Roberts, M. V. Ramana, C. Corrigan, D. Kim and V. Ramanathan, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 7370–7375.
66. A. F. Stein, R. R. Draxler, G. D. Rolph, B. J. B. Stunder, M. D. Cohen and F. Ngan, Bull. Am. Meteorol. Soc., 2015, 96, 2059–2077.
67. G. Rolph, A. Stein and B. Stunder, Environ. Model. Software, 2017, 95, 210–228.
68. Y. J. Tham, C. Yan, L. Xue, Q. Zha, X. Wang and T. Wang, Chin. Sci. Bull., 2014, 59, 356–359.
69. T. P. Riedel, G. M. Wolfe, K. T. Danas, J. B. Gilman, W. C. Kuster, D. M. Bon, A. Vlasenko, S. M. Li, E. J. Williams, B. M. Lerner, P. R. Veres, J. M. Roberts, J. S. Holloway, B. Lefer, S. S. Brown and J. A. Thornton, Atmos. Chem. Phys., 2014, 14, 3789–3800.
70. B. D. Finley and E. S. Saltzman, Geophys. Res. Lett., 2006, 33, L11809.
71. A. A. P. Pszenny, W. C. Keene, D. J. Jacob, S. Fan, J. R. Maben, M. P. Zetwo, M. Springer-Young and J. N. Galloway, Geophys. Res. Lett., 1993, 20, 699–702.
72. C. W. Spicer, E. G. Chapman, B. J. Finlayson-Pitts, R. A. Plastridge, J. M. Hubbe, J. D. Fast and C. M. Berkowitz, Nature, 1998, 394, 353–356.
73. X. Chen, Y. Jiang, Z. Zong, Y. Wang, W. Sun, Y. Wang, M. Xia, L. Guan, P. Liu, C. Zhang, J. Chen, Y. Mu and T. Wang, Environ. Sci. Technol., 2025, 59, 12775–12785.
74. H. D. Osthoff, J. M. Roberts, A. R. Ravishankara, E. J. Williams, B. M. Lerner, R. Sommariva, T. S. Bates, D. Coffman, P. K. Quinn, J. E. Dibb, H. Stark, J. B. Burkholder, R. K. Talukdar, J. Meagher, F. C. Fehsenfeld and S. S. Brown, Nat. Geosci., 2008, 1, 324–328.
75. M. Le Breton, Å. M. Hallquist, R. K. Pathak, D. Simpson, Y. Wang, J. Johansson, J. Zheng, Y. Yang, D. Shang, H. Wang, Q. Liu, C. Chan, T. Wang, T. J. Bannan, M. Priestley, C. J. Percival, D. E. Shallcross, K. Lu, S. Guo, M. Hu and M. Hallquist, Atmos. Chem. Phys., 2018, 18, 13013–13030.
76. J. Zhang and D. Kinnison, TUV Calculator, NCAR (National Center for Atmospheric Research), Boulder, Colorado, 2025, https://www.acom.ucar.edu/Models/TUV/Interactive_TUV/, accessed 15.06.2025.
77. T. H. Bertram and J. A. Thornton, Atmos. Chem. Phys., 2009, 9, 8351–8363.
78. S. R. Zorn, F. Drewnick, M. Schott, T. Hoffmann and S. Borrmann, Atmos. Chem. Phys., 2008, 8, 4711–4728.
79. L. Moormann, F. Fachinger, F. Drewnick and H. Tost, EGUsphere, 2025, pp. 1–23,  DOI:10.5194/egusphere-2025-3862.
80. J. N. Crowley, G. Schuster, N. Pouvesle, U. Parchatka, H. Fischer, B. Bonn, H. Bingemer and J. Lelieveld, Atmos. Chem. Phys., 2010, 10, 2795–2812.
81. Y. Ge, S. Solberg, M. R. Heal, S. Reimann, W. Van Caspel, B. Hellack, T. Salameh and D. Simpson, Atmos. Chem. Phys., 2024, 24, 7699–7729.
82. M. Ragains and B. Finlayson-Pitts, J. Phys. Chem. A, 1997, 101, 1509–1517.
83. J. Burkholder, S. Sander, J. Abbatt, J. Barker, C. Cappa, J. Crounse, T. Dibble, R. Huie, C. Kolb and M. Kurylo, Chemical Kinetics and Photochemical Data for Use in Atmospheric Studies; Evaluation Number 19, Jet Propulsion Laboratory, National Aeronautics and Space, Pasadena, CA, 2020.
84. J. D. Haskins, L. Jaeglé, V. Shah, B. H. Lee, F. D. Lopez-Hilfiker, P. Campuzano-Jost, J. C. Schroder, D. A. Day, H. Guo, A. P. Sullivan, R. Weber, J. Dibb, T. Campos, J. L. Jimenez, S. S. Brown and J. A. Thornton, J. Geophys. Res. Atmos., 2018, 123, 12897–12916.
85. A. A. Angelucci, T. C. Furlani, X. Wang, D. J. Jacob, T. C. Vandenboer and C. J. Young, ACS Earth Space Chem., 2021, 5, 2507–2516.
86. R. Atkinson, S. M. Aschmann, J. Arey and B. Shorees, J. Geophys. Res. Atmos., 1992, 97, 6065–6073.
87. R. A. Cox, M. Ammann, J. N. Crowley, H. Herrmann, M. E. Jenkin, V. F. McNeill, A. Mellouki, J. Troe and T. J. Wallington, Atmos. Chem. Phys., 2020, 20, 13497–13519.
88. I. Bey, B. Aumont and G. Toupance, Geophys. Res. Lett., 1997, 24, 1067–1070.
89. I. Faloona, D. Tan, W. Brune, J. Hurst, D. Barket, T. L. Couch, P. Shepson, E. Apel, D. Riemer, T. Thornberry, M. A. Carroll, S. Sillman, G. J. Keeler, J. Sagady, D. Hooper and K. Paterson, J. Geophys. Res. Atmos., 2001, 106, 24315–24333.
90. X. Ren, H. Harder, M. Martinez, R. L. Lesher, A. Oliger, J. B. Simpas, W. H. Brune, J. J. Schwab, K. L. Demerjian, Y. He, X. Zhou and H. Gao, Atmos. Environ., 2003, 37, 3639–3651.
91. M. A. H. Khan, M. J. Ashfold, G. Nickless, D. Martin, L. A. Watson, P. D. Hamer, R. P. Wayne, C. E. Canosa-Mas and D. E. Shallcross, Atmos. Sci. Lett., 2008, 9, 140–146.
92. R. L. Caravan, M. F. Vansco and M. I. Lester, Commun. Chem., 2021, 4, 44.
93. H. Wang and Y.-P. Lee, J. Phys. Chem. A, 2024, 128, 8690–8698.
94. T. H. Bertram, J. A. Thornton, T. P. Riedel, A. M. Middlebrook, R. Bahreini, T. S. Bates, P. K. Quinn and D. J. Coffman, Geophys. Res. Lett., 2009, 36, L19803.
95. M. Xia, X. Peng, W. Wang, C. Yu, P. Sun, Y. Li, Y. Liu, Z. Xu, Z. Wang, Z. Xu, W. Nie, A. Ding and T. Wang, Atmos. Chem. Phys., 2020, 20, 6147–6158.
96. Y. Li, W. Nie, Y. Liu, D. Huang, Z. Xu, X. Peng, C. George, C. Yan, Y. J. Tham, C. Yu, M. Xia, X. Fu, X. Wang, L. Xue, Z. Wang, Z. Xu, X. Chi, T. Wang and A. Ding, Environ. Sci. Technol. Lett., 2020, 7, 70–75.
97. E. Z. Dalton, E. H. Hoffmann, T. Schaefer, A. Tilgner, H. Herrmann and J. D. Raff, J. Am. Chem. Soc., 2023, 145, 15652–15657.
98. G. Chen, L. Xu, S. Yu, L. Xue, Z. Lin, C. Yang, X. Ji, X. Fan, Y. J. Tham, H. Wang, Y. Hong, M. Li, J. H. Seinfeld and J. Chen, Environ. Sci. Technol., 2024, 58, 22714–22721.
99. J. Dai, T. Wang, H. Shen, M. Xia, W. Sun and G. P. Brasseur, Environ. Sci. Technol., 2025, 59, 2169–2180.
100. T. P. Riedel, T. H. Bertram, T. A. Crisp, E. J. Williams, B. M. Lerner, A. Vlasenko, S.-M. Li, J. Gilman, J. de Gouw, D. M. Bon, N. L. Wagner, S. S. Brown and J. A. Thornton, Environ. Sci. Technol., 2012, 46, 10463–10470.
101. J. Lelieveld, F. J. Dentener, W. Peters and M. C. Krol, Atmos. Chem. Phys., 2004, 4, 2337–2344.
102. G. Chua, V. Naik and L. W. Horowitz, Atmos. Chem. Phys., 2023, 23, 4955–4975.

---