Photochemical method for removing methane interference for improved gas analysis

Abstract

The development of laser spectroscopy has made it possible to measure minute changes in the concentrations of trace gases and their isotopic analogs. These single or even multiply substituted species occur at ratios from percent to below parts per million and contain important information concerning trace gas sources and transformations. Due to their low abundance, minimizing spectral interference from other gases in a mixture is essential. Options including traps and membranes are available to remove many specific impurities. Methods for removing ${\mathrm{CH}}_{\mathrm{4}}$, however, are extremely limited as methane has low reactivity and adsorbs poorly to most materials. Here we demonstrate a novel method for ${\mathrm{CH}}_{\mathrm{4}}$ removal via chlorine-initiated oxidation. Our motivation in developing the technique was to overcome methane interference in measurements of ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ isotopic analogs when using a cavity ring-down spectrometer. We describe the design and validation of a proof-of-concept device and a kinetic model to predict the dependence of the methane removal efficiency on the methane concentration [${\mathrm{CH}}_{\mathrm{4}}$], chlorine photolysis rate $J_{{\mathrm{Cl}}_{\mathrm{2}}}$, chlorine concentration [${\mathrm{Cl}}_{\mathrm{2}}$] and residence time $t_{\mathrm{R}}$. The model was validated by comparison to experimental data and then used to predict the possible formation of troublesome side products and by-products including ${\mathrm{CCl}}_{\mathrm{4}}$ and $\mathrm{HCl}$. The removal of methane could be maintained with a peak removal efficiency $>\mathrm{98}$ % for ambient levels of methane at a flow rate of 7.5 mL min${}^{-\mathrm{1}}$ with [${\mathrm{Cl}}_{\mathrm{2}}$] at 50 ppm. These tests show that our method is a viable option for continuous methane scrubbing. Additional measures may be needed to avoid complications due to the introduction of ${\mathrm{Cl}}_{\mathrm{2}}$ and formation of $\mathrm{HCl}$. Note that the method will also oxidize most other common volatile organic compounds. The system was tested in combination with a cavity ring-down methane spectrometer, and the developed method was shown to be successful at removing methane interference.

1 Introduction

Infrared absorption is a fast, convenient and non-destructive approach for measuring gas composition that is used in a wide range of applications. High-resolution instruments based on specific rovibrational transitions are becoming available to characterize the abundance of rare isotopocules within gases. Laser spectroscopy has entered territory that has been the exclusive domain of mass spectrometry. While recent advances in the field can give the impression that new laser-based instruments can be used in a “plug and play” manner, there are still limitations to the accuracy and reproducibility of the measurements.

In a recent study investigating the performance of currently available laser spectroscopic ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ isotope analyzers , a number of interferences from other trace gases were identified, arising from spectral overlap of ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ and the rovibrational spectra of the other gases. The consequence was an offset in the measured isotopocule abundance value arising exclusively from ambient levels of methane for a Picarro G5131-i cavity ring-down-based instrument that determines ${\mathit{\delta }}^{\mathrm{15}}\mathrm{N}$, ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\alpha }}$, ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\beta }}$ and ${\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$ for ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$. These instruments are often used to measure isotopic signatures of ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ emitted from soils, , which can help to differentiate distinct microbial and abiotic production pathways.

${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ formation in soils is commonly accompanied by production and/or uptake of other trace gases such as ${\mathrm{CH}}_{\mathrm{4}}$, ${\mathrm{CO}}_{\mathrm{2}}$ and water vapor . These variations complicate measurements. An example of the relevant variation of ${\mathrm{CO}}_{\mathrm{2}}$ and ${\mathrm{CH}}_{\mathrm{4}}$ can be found in the work of in which the background level of ${\mathrm{CH}}_{\mathrm{4}}$ and ${\mathrm{CO}}_{\mathrm{2}}$ at 1.8 and 380 ppm can change suddenly to levels above 3.6 and 560 ppm. For the instrument described in , these variations will result in an observed offset in the measured ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\alpha }}$ of 4.0 ‰ and ${\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$ of 1.1 ‰ . The change in ${\mathrm{CH}}_{\mathrm{4}}$ results in an apparent increase of 4.6 ‰ and 2.2 ‰ for ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\alpha }}$ and ${\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$, respectively, while the change in ${\mathrm{CO}}_{\mathrm{2}}$ results in a decrease of 0.6 ‰ and 1.1 ‰ for ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\alpha }}$ and ${\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$, respectively. As the effect of variation in these two trace gases leads to opposing offsets in the measured isotopologues, it greatly decreases both the accuracy and precision of the G5131-i. It is therefore essential for accurate measurements to account for these interferences.

One solution is multi-line analysis or careful measurement of the interfering gas(es) with a second instrument. These options are not desirable for all applications as they either require a redesign of the instrument or investment in additional equipment, and these corrections can introduce additional uncertainty. A more direct and practical method would be to remove the interfering species from the sample. For discrete sampling the best method would be to separate the ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ from the sample matrix and release it into a well-defined matrix for interference-free measurements.

For online measurements, well-established methods including chemical traps and membranes are readily available for the removal of ${\mathrm{CO}}_{\mathrm{2}}$, $\mathrm{CO}$ and humidity. However, to the best of our knowledge, no method for continuous removal of methane is available with the exception of catalyzed combustion , which requires high temperatures and the addition of oxygen, thereby altering the gas matrix. It was desired to develop a method for removing ${\mathrm{CH}}_{\mathrm{4}}$ and potentially other volatile organic compounds (VOCs) in a manner that would only introduce minimal changes to the matrix composition.

Inspiration for the method investigated in this work was taken from the oxidation pathways taking place in the atmosphere . The majority of methane is oxidized through an initial reaction with OH radicals that results in the formation of ${\mathrm{H}}_{\mathrm{2}}\mathrm{O}$ and ${\mathrm{CH}}_{\mathrm{3}}$ radicals. However, the chlorine radical is a potentially important agent for initiating chain reactions: generally, the reaction rates of $\mathrm{Cl}$ with VOCs exceed the analogous ones with OH by at least 1 order of magnitude. The rate constant for methane reaction with $\mathrm{Cl}$ radicals is $\mathrm{1.07}\times {\mathrm{10}}^{-\mathrm{13}}$ cm${}^{\mathrm{3}}$ s${}^{-\mathrm{1}}$ and with hydroxyl radicals is $\mathrm{6.20}\times {\mathrm{10}}^{-\mathrm{15}}$ cm${}^{\mathrm{3}}$ s${}^{-\mathrm{1}}$ . The reason for the limited role of chlorine in the global atmosphere is that its concentration on average is 3 or 4 orders of magnitude lower than OH, although it can have an impact in the stratosphere and in marine and polar environments. The mechanism for Cl-initiated methane oxidation technology proposed in this study is outlined in Reactions ()–(). $$\begin{array}{}\text{R1}& {\displaystyle }{\mathrm{Cl}}_{\mathrm{2}}+hv\to \mathrm{2}\mathrm{Cl}\text{R2}& {\displaystyle }\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{4}}\to {\mathrm{CH}}_{\mathrm{3}}+\mathrm{HCl}\text{R3}& {\displaystyle }{\mathrm{CH}}_{\mathrm{3}}+{\mathrm{O}}_{\mathrm{2}}+M\to {\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}+M\text{R4}& {\displaystyle }{\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}+\mathrm{Cl}\to {\mathrm{CH}}_{\mathrm{3}}\mathrm{O}+\mathrm{ClO}\text{R5}& {\displaystyle }{\mathrm{CH}}_{\mathrm{3}}\mathrm{O}+{\mathrm{O}}_{\mathrm{2}}\to \mathrm{HCHO}+{\mathrm{HO}}_{\mathrm{2}}\text{R6}& {\displaystyle }\mathrm{HCHO}+\mathrm{Cl}+{\mathrm{O}}_{\mathrm{2}}\to \mathrm{CO}+\mathrm{HCl}+{\mathrm{HO}}_{\mathrm{2}}\end{array}$$

We demonstrate a novel method for ${\mathrm{CH}}_{\mathrm{4}}$ removal through chlorine-initiated oxidation. Using four experimental setups, we show that methane removal is highly dependent on the flow, chlorine mixing ratio and light source. We developed a simple kinetic model to predict the removal efficiency as a function of the four key parameters in the system: [${\mathrm{CH}}_{\mathrm{4}}$], $J_{{\mathrm{Cl}}_{\mathrm{2}}}$, [${\mathrm{Cl}}_{\mathrm{2}}$] and residence time $t_{\mathrm{R}}$. The model includes essential reactions and additional estimated radical wall reactions. Two approaches for estimating the photodissociation rate of ${\mathrm{Cl}}_{\mathrm{2}}$ are presented. The goal is to determine the effect of these variables and achieve the desired methane removal efficiencies by optimizing the parameters. The goal is to achieve removals above 99 % for methane at low to ambient concentrations. With the method developed and refined, a final set of experiments is conducted using a Picarro CRDS model G5131-i capable of measuring ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ mixing ratio and its isotopic abundance. The measured values of ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\alpha }}$ and ${\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$, subject to methane interference, are compared to data corrected for methane levels, as these corrected isotopologue levels remained stable across the experiment.

2 Method

2.1 Experimental approach

2.1.1 Methane experiments

Four different variations of the setup seen in Fig.  are used during our experiments and are summarized in Table  together with the experiments they were used for.

Table 1

Table summarizing experiments and setups. See Fig.  for an overview. FC: flow-controlled. CWL: chlorine waste line. PC: pressure-controlled.

Setup	Description	Experiments
1	High-pressure xenon lamp with FC CWL	A
2	Single-tube hexagonal photochemical device with FC CWL	B
3	Singlet-tube hexagonal photochemical device with PC CWL	C, D, E
4	Multiple-tube hexagonal photochemical device with PC CWL	F, G, H, I

Table 2

Table summarizing experimental conditions.

Flask name	${\mathrm{CH}}_{\mathrm{4}}$ / ppm	${\mathrm{Cl}}_{\mathrm{2}}$ / ppm	${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ / $\mathrm{ppb}$	Matrix composition	Flow range / (mL min${}^{-\mathrm{1}}$)
I	0	$\mathrm{100}\pm \mathrm{2.5}$	0	$>\mathrm{99}$ % ${\mathrm{N}}_{\mathrm{2}}$	6–23
II	$\mathrm{2.003}\pm \mathrm{5}\times {\mathrm{10}}^{-\mathrm{4}}$	0	323	Atmospheric air	1–29
II	$\mathrm{78}\pm \mathrm{2}$	0	0	20.95 % ${\mathrm{O}}_{\mathrm{2}}+>\mathrm{79}$ % ${\mathrm{N}}_{\mathrm{2}}$	0.3–1.2
IV	0	0	509	0.95 % $\mathrm{Ar}+\mathrm{20.95}\mathit{\%}{\mathrm{O}}_{\mathrm{2}}+>\mathrm{78}$ %	28–50

Figure 1

General setup. ACT: activated carbon trap. MFM: mass flowmeter. MFC: MKS mass flow controller GE50A. MFC: manual flow controller. Table : gas flask. Four variations of the general setup are performed. The setup variations and the experiments performed with the setups are shown in Table . Setup 1 uses a xenon lamp as the photochemical device. Setups 2–4 use the same photochemical device, which consists of 420 LEDs. The chamber tube used in setups 1–3 is one quartz tube (20 cm length $\times$ 12.7 mm o.d.), while setup 4 uses seven smaller quartz tubes: five with the size 8.33 mm (o.d.), 6.33 mm (i.d.) and 20 cm (L), as well as two with the size 8.33 mm (o.d.), 6.33 mm (i.d.) and 25 cm (L). The setups also differ in the chlorine waste line. Setups 1–2 use a flow-controlled chlorine waste line, while setup 3.4 uses a pressure-controlled chlorine waste line.

[Figure omitted. See PDF]

The system (Fig. ) has a manifold combining flows from two channels: the sample channel and the chlorine gas channel. [${\mathrm{Cl}}_{\mathrm{2}}$] is supplied from an external tank labeled flask I (see Table  for gas flask). Atmospheric air in flask II is combined with an enriched source of [${\mathrm{CH}}_{\mathrm{4}}$] in flask III to generate various levels of [${\mathrm{CH}}_{\mathrm{4}}$] for the sample channel. A chlorine sensor is placed outside the main flow line to reduce the volume of the setup and allow for increased time resolution. The flow containing methane and chlorine gas is split at a T-piece, where the main flow proceeds through the photochemical device with excess gas going past a ${\mathrm{Cl}}_{\mathrm{2}}$ sensor (chlorine gas detector 0–20 ppm ${\mathrm{Cl}}_{\mathrm{2}}$). ${\mathrm{Cl}}_{\mathrm{2}}$ concentrations above 20 ppm are estimated from the flow rate ratios.

The photochemical device

Setup 1 uses a high-pressure xenon lamp (ILC Technology R100-IB) as the photochemical device. Setups 2–4 use a photochemical chamber consisting of 420 LEDs with peak emission at 365 nm with the circuit board mounted together in a hexagonal cylinder (illustrated in Fig. ). The 420 LEDs are connected in parallel. At the maximum voltage of 3.8 V each consumes 13.2 mA, resulting in a total power of 21 W.

A single quartz tube with 20 cm length and 12.7 mm outer diameter is used as the chamber tube for setups 1–3. In setup 4, the $t_{\mathrm{R}}$ in the chamber is increased by a factor of 2.7 by substituting a single quartz tube with seven smaller quartz tubes in hexagonal shape for optimal packing comprising five tubes with the following dimensions: o.d. 8.33 $\mathrm{mm}$, i.d. 6.33 $\mathrm{mm}$, length 20 $\mathrm{cm}$. An additional two tubes are used with the following dimensions: o.d. 8.00 $\mathrm{mm}$, i.d. 6.00 $\mathrm{mm}$, length 25 $\mathrm{cm}$. The tubes were connected in series via Tygon tubes (Tygon R3603) of length 5 $\mathrm{cm}$. The insides of these tubes were coated with Krytox (DuPont GPL 205 Krytox Performance Grease) to prevent reaction with ${\mathrm{Cl}}_{\mathrm{2}}$.

Post-photolysis scrubbing

After the photochemical device the sample passes through a 35 cm Nafion membrane (TT-030 from Perma Pure LLC). The dried sample then passes through an ascarite trap consisting of a central layer of $\mathrm{NaOH}$ between two layers of $\mathrm{Mg}({\mathrm{ClO}}_{\mathrm{4}}{)}_{\mathrm{2}}$ separated by glass wool. These types of traps are normally used for the removal of ${\mathrm{CO}}_{\mathrm{2}}$ and ${\mathrm{H}}_{\mathrm{2}}\mathrm{O}$ , but they were found to likewise remove $\mathrm{HCl}$ and ${\mathrm{Cl}}_{\mathrm{2}}$. This removal was confirmed by separate experiments, as it was essential that none of the corrosive gases made it to the delicate Picarro instrument. The gas stream then flows into a cavity ring-down spectrometer (CRDS), the Picarro model G1301. A nominal flow of 15 mL min${}^{-\mathrm{1}}$ was maintained with the exception of experiments involving variation in $t_{\mathrm{R}}$ when this flow was changed accordingly. At the outlet of the Picarro G1301 an activated carbon (bead-shaped activated carbon, KUREHA Corporation) trap labeled ACT is attached, which is mainly used for scrubbing chlorinated organic species, such as ${\mathrm{CCl}}_{\mathrm{4}}$, out of health concern .

2.1.2

N${}_{\mathrm{2}}$O experiments

A final set of experiments is conducted using a Picarro CRDS model G5131-i capable of measuring ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ mixing ratio and isotopic abundance. These experiments were performed to validate the effect of the removal of ${\mathrm{CH}}_{\mathrm{4}}$ on the measurement of ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$. These experiments were done in two sets using the setup in Fig. b with and without the sofnocat trap. The difference between the two setups was hence the inclusion of a sofnocat trap in Fig. b. The sofnocat trap is used to oxidize the CO product and was prepared with 1.25 $\mathrm{g}$ of sofnocat contained in a $\mathrm{1}/{\mathrm{4}}^{\prime \prime }$ stainless-steel tube of length 8 $\mathrm{cm}$ kept in place by glass wool.

2.2 Theoretical approach

2.2.1 Kintecus version 6.8

A model is made with the program Kintecus version 6.8 to investigate the reaction mechanisms in the photochemical device. The model contained the relevant reactions with rates for chlorine radical production and removal, methane oxidation, and formation of chlorinated species. The model was kept as simple as possible while still including the relevant reactions. The reactions used in the model are found in Tables –. A simplified reaction scheme is shown in Fig. . A continuous flow was simulated by setting the initial and external concentrations of gases flowing through the chamber to the same value. This is done for the gases ${\mathrm{Cl}}_{\mathrm{2}}$, ${\mathrm{CH}}_{\mathrm{4}}$, ${\mathrm{N}}_{\mathrm{2}}$ and ${\mathrm{O}}_{\mathrm{2}}$. A copy of the model parameters is available in Appendix .

Figure 2

Reaction scheme for the oxidation of methane to ${\mathrm{CO}}_{\mathrm{2}}$. ${\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}$ self-reactions lead to the formation of ${\mathrm{CH}}_{\mathrm{3}}\mathrm{O}$.

[Figure omitted. See PDF]

Radical wall reactions

A set of radical-terminating reactions is incorporated in the model to account for reactions on the walls of the quartz tube. R7$$\mathrm{Cl}\to {\displaystyle \frac{\mathrm{1}}{\mathrm{2}}}{\mathrm{Cl}}_{\mathrm{2}}$$ R8$$\mathrm{ClO}\to {\displaystyle \frac{\mathrm{1}}{\mathrm{2}}}{\mathrm{Cl}}_{\mathrm{2}}+{\displaystyle \frac{\mathrm{1}}{\mathrm{2}}}{\mathrm{O}}_{\mathrm{2}}$$ R9$$\mathrm{OH}\to {\displaystyle \frac{\mathrm{1}}{\mathrm{2}}}{\mathrm{H}}_{\mathrm{2}}\mathrm{O}+{\displaystyle \frac{\mathrm{1}}{\mathrm{4}}}{\mathrm{O}}_{\mathrm{2}}$$ R10$${\mathrm{HO}}_{\mathrm{2}}\to {\displaystyle \frac{\mathrm{1}}{\mathrm{2}}}{\mathrm{H}}_{\mathrm{2}}\mathrm{O}+{\displaystyle \frac{\mathrm{3}}{\mathrm{4}}}{\mathrm{O}}_{\mathrm{2}}$$

The wall reactions are assumed to be diffusion-limited. The diffusion length is calculated as the average distance from the wall. The diffusion length and rate were calculated using Eqs. () and (), respectively. The estimate of the diffusion rate is described in detail in Sect. . The diffusion constants, diffusion lengths and estimated wall reaction rates are shown in Table . 1$$l=r\cdot \left(\mathrm{1}-{\displaystyle \frac{\mathrm{1}}{\sqrt{\mathrm{2}}}}\right)$$ Here, $l$ in the diffusion length and $r$ is the inner radius of the tube. 2$$k={\displaystyle \frac{\mathrm{4}\cdot D}{l^{\mathrm{2}}}}$$ Here, $D$ is the diffusion constant (see Table ).

Model results

The outputs from the model are the photodissociation rate, $J_{{\mathrm{Cl}}_{\mathrm{2}}}$, the abundance of [Cl] and the production of ${\mathrm{CCl}}_{\mathrm{4}}$ as an indicator of the production of unwanted side products.

$J_{{\mathrm{Cl}}_{\mathrm{2}}}$ estimation

The chlorine photolysis rate, $J_{{\mathrm{Cl}}_{\mathrm{2}}}$, is estimated in two ways, which is described in more detail in Sect. . The first approach is to fit $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ to reproduce the observed removal efficiencies from the experimental results. These fits were performed for experiments investigating the effect of power.

A second approach is to estimate $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ by relating it to the electric power going through the circuit, $P_{\mathrm{IN}}$. Based on our observation, a second-order polynomial provided the best fit to describe the effective light output, $P_{\mathrm{eff}}$, as a function of $P_{\mathrm{IN}}$: 3$$P_{\mathrm{eff}}\left(P_{\mathrm{IN}}\right)=(a\cdot P_{\mathrm{IN}}+b)\cdot P_{\mathrm{IN}},$$ where the constants $a$ (${\mathrm{W}}^{-\mathrm{1}}$) and $b$ (unitless) are experiment-dependent constants that scale the effective light output $P_{\mathrm{eff}}$ in watts (W). From the effective power output, the photolysis rate $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ is calculated by Eq. (4). 4$$J_{{\mathrm{Cl}}_{\mathrm{2}}}\left(\mathrm{W}\right)=P_{\mathrm{eff}}\left(P_{\mathrm{IN}}\right)\cdot J_{\mathrm{scale}}$$ $J_{\mathrm{scale}}$ (${\mathrm{J}}^{-\mathrm{1}}$) is the scaling factor and was calculated from the cross section of ${\mathrm{Cl}}_{\mathrm{2}}$, the wavelength distribution of the generated light and the expected photon density. The density of photons depends on the volume and cross section of the tube within the photochemical device. $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ is fit to the data collected for some of the experimental steps for exp. D and I. Exp. D reflects the single-tube system (setups 1–3), while experiment I reflects the optimized multiple-tube system (setup 4). From the fitted $a$, $b$ and calculated $J_{\mathrm{scale}}$ the photolysis rate could be calculated for the other experiments.

3 Results and discussion

3.1 Experimental results

The findings are based on 12 experiments, named A–L, containing multiple steps of turning on the photolysis under different conditions. These steps will be referred to by their experimental letter and their number; e.g., experiment C step 5 would be exp. C5. An overview of the settings and resulting removal efficiencies for experiments C–I can be seen in Table  (see Appendix Tables –). Table  gives an overview of the experiments. As an example of our data, we present the results from experiment H (Fig. ), during which we achieved our highest level of removal. The experiment was carried out with constant [${\mathrm{CH}}_{\mathrm{4}}$]${}_{\mathrm{initial}}$ and [${\mathrm{Cl}}_{\mathrm{2}}$] at $\mathrm{2.000}\pm \mathrm{0.003}$ and 50.5 ppm. The different levels of removal seen reflect stepwise changes to the settings for $t_{\mathrm{R}}$ and $P_{\mathrm{IN}}$. As seen in Fig.  for exp. H1–H4, removal efficiency is improved as the $P_{\mathrm{IN}}$ is increased. Starting with H5 a fan was installed to limit temperature increases. $P_{\mathrm{IN}}$ was kept at the same level, while the residence time in the chamber was decreased. The three steps (H1–H3) were carried out with constant $P_{\mathrm{IN}}$ at 14.8 W with $t_{\mathrm{R}}$ ranging from 164–350 $\mathrm{s}$. $t_{\mathrm{R}}$ was kept at 350 $\mathrm{s}$ for experiments H3–H6. Furthermore, $P_{\mathrm{IN}}$ was varied within the range of 14.8–22.8 W. Two issues affected the results. First, the system was not initially stable. We believe this is due to a build-up of moisture on the glass walls coming to equilibrium after the first step, as can be seen from the slope in step H1. Second, there is a small continuous pressure drop from the ${\mathrm{Cl}}_{\mathrm{2}}$ regulator, which leads to a decrease in ${\mathrm{Cl}}_{\mathrm{2}}$ and an increase in ${\mathrm{CH}}_{\mathrm{4}}$. The reason for this was insufficient drying of the regulator prior to use, which left a layer of moisture to react with chlorine, thus initiating corrosion in the regulator. This is also the reason we needed a chlorine waste line, as a high flow through the regulator was needed to reduce the effect of this loss to the regulator. We have accounted for the effect of the pressure drop, but it contributes to the uncertainty of our reported ${\mathrm{Cl}}_{\mathrm{2}}$. We must stress the importance of proper drying prior to the use of ${\mathrm{Cl}}_{\mathrm{2}}$ gas for those intending to emulate our setup.

Table 3

Removal efficiencies (%) for experiments C–I.

Step RE%	C	D	E	F	G	H	I
1	$\mathrm{45}\pm \mathrm{5}$	$\mathrm{17.6}\pm \mathrm{0.6}$	$\mathrm{22}\pm \mathrm{12}$	$\mathrm{28.0}\pm \mathrm{0.3}$	$\mathrm{47.4}\pm \mathrm{1.2}$	$\mathrm{68}\pm \mathrm{3}$	$\mathrm{46.1}\pm \mathrm{1.8}$
2	$\mathrm{18.9}\pm \mathrm{1.3}$	$\mathrm{24.82}\pm \mathrm{0.5}$	$\mathrm{19.8}\pm \mathrm{1.9}$	$\mathrm{37.3}\pm \mathrm{0.3}$	$\mathrm{54.7}\pm \mathrm{0.5}$	$\mathrm{88.1}\pm \mathrm{1.3}$	$\mathrm{56.6}\pm \mathrm{0.2}$
3	$\mathrm{27.8}\pm \mathrm{1.0}$	$\mathrm{15.4}\pm \mathrm{0.3}$	$\mathrm{16.7}\pm \mathrm{1.2}$	$\mathrm{46.83}\pm \mathrm{0.1}$	$\mathrm{60.8}\pm \mathrm{0.5}$	$\mathrm{92}\pm \mathrm{5}$	$\mathrm{64.29}\pm \mathrm{0.1}$
4	$\mathrm{61}\pm \mathrm{9}$	$\mathrm{6.2}\pm \mathrm{0.3}$	$\mathrm{23.1}\pm \mathrm{1.9}$	$\mathrm{53.77}\pm \mathrm{0.1}$	$\mathrm{66.2}\pm \mathrm{0.6}$	$\mathrm{94}\pm \mathrm{5}$	$\mathrm{70.31}\pm \mathrm{0.1}$
5		$\mathrm{38.2}\pm \mathrm{1.8}$	$\mathrm{35}\pm \mathrm{3}$	$\mathrm{55.2}\pm \mathrm{0.1}$	$\mathrm{69.6}\pm \mathrm{0.3}$	$\mathrm{96}\pm \mathrm{4}$	$\mathrm{75.09}\pm \mathrm{0.1}$
6		$\mathrm{33.6}\pm \mathrm{0.9}$	$\mathrm{29.3}\pm \mathrm{1.6}$	$\mathrm{58.0}\pm \mathrm{0.3}$	$\mathrm{72.0}\pm \mathrm{0.4}$	$\mathrm{98.99}\pm \mathrm{0.1}$	$\mathrm{82.28}\pm \mathrm{0.07}$
7		$\mathrm{37.0}\pm \mathrm{1.3}$	$\mathrm{39}\pm \mathrm{6}$	$\mathrm{59.2}\pm \mathrm{0.3}$	$\mathrm{74.1}\pm \mathrm{0.6}$	$\mathrm{96.7}\pm \mathrm{0.3}$	$\mathrm{83.25}\pm \mathrm{0.04}$
8		$\mathrm{33.1}\pm \mathrm{1.5}$	$\mathrm{40}\pm \mathrm{5}$		$\mathrm{77.2}\pm \mathrm{0.7}$	$\mathrm{87.33}\pm \mathrm{0.1}$
9		$\mathrm{35}\pm \mathrm{15}$	$\mathrm{35}\pm \mathrm{2}$		$\mathrm{60.3}\pm \mathrm{0.4}$	$\mathrm{77.30}\pm \mathrm{0.1}$
10		$\mathrm{37}\pm \mathrm{14}$	$\mathrm{45}\pm \mathrm{4}$		$\mathrm{60.2}\pm \mathrm{0.2}$
11			$\mathrm{47}\pm \mathrm{2}$		$\mathrm{64.0}\pm \mathrm{0.2}$
12			$\mathrm{41.0}\pm \mathrm{1.9}$		$\mathrm{66.1}\pm \mathrm{0.3}$
13			$\mathrm{54}\pm \mathrm{2}$
14			$\mathrm{53}\pm \mathrm{3}$
15			$\mathrm{45}\pm \mathrm{3}$
16			$\mathrm{59}\pm \mathrm{2}$

Figure 3

Exp. H. The $\left[{\mathrm{CH}}_{\mathrm{4}}\right]$ is seen as a function of time. The highlighting indicates the illumination times. In addition, the experimental step is indicated at the top and $P_{\mathrm{IN}}$ ($\mathrm{W}$) is indicated at the bottom.

[Figure omitted. See PDF]

Figure 4

(a) RE% of methane plotted against $t_{\mathrm{R}}$ ($\mathrm{s}$). The result originates from the two experiments C (green) and H (violet). The experiments have different settings in $P_{\mathrm{IN}}$, [${\mathrm{CH}}_{\mathrm{4}}$] and [${\mathrm{Cl}}_{\mathrm{2}}$]. (b) RE% of methane plotted against $P_{\mathrm{IN}}$ ($\mathrm{W}$). The results are from the two experiments F (square) and I (triangle), which have different [${\mathrm{CH}}_{\mathrm{4}}$] settings. (c) The panel presents the methane RE% as a function of the chlorine mixing ratio for exp. E. Step 1, at 30 ppm [${\mathrm{Cl}}_{\mathrm{2}}$], is an example of start-up deviation; therefore, it is removed. The points represent the three different $P_{\mathrm{IN}}$ of the photochemical device. (d) The removal efficiency RE% during exp. G of methane is displayed as a function of the initial methane concentration with the remaining fixed parameters such as [${\mathrm{Cl}}_{\mathrm{2}}$] mixing ratio, $t_{\mathrm{R}}$ and $P_{\mathrm{IN}}$ input. The three red points in the figure represent steps suffering from start-up deviation.

[Figure omitted. See PDF]

Effect of residence time ($t_{\mathrm{R}}$, s)

Increasing the residence time results in increased removal of methane, as shown in Fig. a. The $t_{\mathrm{R}}$ was investigated in the single- and multiple-tube systems. The same flow rate yields a longer $t_{\mathrm{R}}$ for the multiple-tube setup due to the 2.7-fold volume increase. The expected trend of asymptotically approaching 100 % can be seen for exp. H, where the high $P_{\mathrm{IN}}$ approaches more quickly. The effective light output and $t_{\mathrm{R}}$ are lower for experiments B, C and D compared to H. The resulting removal of methane is accordingly lower. Increasing the $t_{\mathrm{R}}$ is an easy way of enhancing the removal but at the expense of a slower response time of the system.

Effect of power input ($P_{\mathrm{IN}}$, w)

The results from experiments with power variations are shown in Fig. b. As presented for exp. F the system reaches a maximum removal efficiency such that increasing the power does not yield significantly higher removal efficiencies. The [${\mathrm{Cl}}_{\mathrm{2}}$] and $t_{\mathrm{R}}$ for experiments F and I are found to be $\mathrm{50}\pm \mathrm{4}$ ppm and 162 mL min${}^{-\mathrm{1}}$, respectively. Comparing exp. F to I it is evident that a higher removal efficiency has been reached thanks to the addition of a fan to distribute the heat and prolong the lifetime of the LEDs.

Effect of [${\mathrm{Cl}}_{\mathrm{2}}$]

Exp. E determined the effect of changing [${\mathrm{Cl}}_{\mathrm{2}}$]; see Fig. c. [${\mathrm{Cl}}_{\mathrm{2}}$] is set between 20 and 70 ppm. Higher [${\mathrm{Cl}}_{\mathrm{2}}$] levels result in an increased methane removal rate. The resulting removal efficiency is still below 60 % and the RE% appears to be linear with [${\mathrm{Cl}}_{\mathrm{2}}$]. Given the result from exp. E the level of [${\mathrm{Cl}}_{\mathrm{2}}$] was set to 50 ppm for the remaining experiments.

Effect of initial [CH${}_{\mathrm{4}}$]

Exp. G, plotted in Fig. d, spans [${\mathrm{CH}}_{\mathrm{4}}$] in the range 1.4–3.8 ppm. Steps G1–G3 are highlighted to indicate the initial instability. The experiment showed high removal of methane at ambient concentrations.

The performance of the experimental setup has been investigated in the aforementioned experiments. The removal efficiencies can be increased by increasing $P_{\mathrm{IN}}$ or [${\mathrm{Cl}}_{\mathrm{2}}$], resulting in an increase in [$\mathrm{Cl}$]. The negative correlation for [${\mathrm{CH}}_{\mathrm{4}}$] is understandable as RE% is a relative value. As expected, the absolute amount of removed methane scales with the [${\mathrm{CH}}_{\mathrm{4}}$].

3.1.1

N${}_{\mathrm{2}}$O experimental results

In Fig. a and b the effects on the isotopic signal of ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\alpha }}$ and ${\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$ from the removal of methane can be seen. The delta values are self-referenced to the gas without the addition of ${\mathrm{CH}}_{\mathrm{4}}$. The results are from experiment L, wherein a sofnocat trap was installed to remove the $\mathrm{CO}$ formed by ${\mathrm{CH}}_{\mathrm{4}}$ oxidation. By applying the trace gas and matrix interference corrections described in in combination with the measurements of ${\mathrm{CH}}_{\mathrm{4}}$, it was found that the isotopologue levels remained stable through the oxidation (grey line). The offset from this corrected value is plotted in red, showing values higher by several per mill. These levels stabilized during the oxidation in accordance with the drop in methane, thus demonstrating the efficiency of the method. The stability of the corrected isotopic values across the experiment shows that the oxidation does not introduce other components that would interfere with the signal, which are not removed by the traps. Variations of roughly 5 % were observed in [${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$] but are accounted for by variations in the flow of [${\mathrm{Cl}}_{\mathrm{2}}$], thus changing the dilution, rather than formation of ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ due to the photochemistry. In Table  the results from the three experiments J, K and L can be seen. In the ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ experiments it was not possible to apply the same conditions that lead to the highest levels of removal presented in the earlier experiments. The reason for this was that the addition of the G5131-i increased the minimum flow through the photochemical device, thus decreasing the maximum residence time. Additionally, not having a high-concentration ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ source capped the dilution, as the ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ needed to remain in the linear range of the G5131-i. The limit on the dilution therefore also limited the concentration of ${\mathrm{Cl}}_{\mathrm{2}}$ available. With a higher-concentration ${\mathrm{Cl}}_{\mathrm{2}}$ source available and a properly prepared regulator, the setup would have been able to deliver sufficient ${\mathrm{CH}}_{\mathrm{4}}$ removal for more than 24 h, at which point the ascarite trap would need replenishment.

Figure 5

(a) Measurements of ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\alpha }}$ during exp. L (‰). Red highlights a 100 $\mathrm{s}$ average measured value corrected for ${\mathrm{O}}_{\mathrm{2}}$, $\mathrm{CO}$ and ${\mathrm{CO}}_{\mathrm{2}}$ effects, while the grey line indicates a 100 s average value corrected for all interference including ${\mathrm{CH}}_{\mathrm{4}}$. The black line shows the ${\mathrm{CH}}_{\mathrm{4}}$ level (ppm). (b) Measurements of ${\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$ from exp. L (‰). Red highlights a 100 s average measured value corrected for ${\mathrm{O}}_{\mathrm{2}}$, $\mathrm{CO}$ and ${\mathrm{CO}}_{\mathrm{2}}$ effects, while the grey line indicates a 100 s average value corrected for all interference including ${\mathrm{CH}}_{\mathrm{4}}$. The black line shows the ${\mathrm{CH}}_{\mathrm{4}}$ level (ppm).

[Figure omitted. See PDF]

Table 4

Experimental data for the ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ experiments using the G5131-i for ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ analysis. Columns: experimental steps, initial [${\mathrm{CH}}_{\mathrm{4}}$] (ppm), residence time in seconds, removal efficiency in percent (%), [${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$] (ppb), ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\alpha }}$, ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\beta }}$ and ${\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$ (‰) refer to the three isotopologue measurements of ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$. Each of the three isotope values have been corrected for the effects of oxygen, $\mathrm{CO}$ and ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ variation according to the method described in . The values have not been bound to an absolute scale by the use of calibration gas, so the daily isotope levels unaffected by methane are shown in the day.

Experiment	${\mathrm{CH}}_{\mathrm{4}}$${}_{\mathrm{initial}}$	$t_{\mathrm{R}}$	RE	${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$	${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\alpha }}$	${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\beta }}$	${\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$
(no.)	($\mathrm{ppm}$)	($\mathrm{s}$)	($\mathrm{\%}$)	($\mathrm{ppb}$)	($\mathrm{‰}$)	($\mathrm{‰}$)	($\mathrm{‰}$)
Exp. J
1	$\mathrm{2.4048}\pm \mathrm{6}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{64}\pm \mathrm{5}$	$\mathrm{28.3}\pm \mathrm{0.5}$	$\mathrm{340.2}\pm \mathrm{0.03}$	$\mathrm{3.2}\pm \mathrm{1.0}$	$-\mathrm{1.7}\pm \mathrm{0.8}$	$\mathrm{2.6}\pm \mathrm{0.4}$
2	$\mathrm{2.4048}\pm \mathrm{6}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{64}\pm \mathrm{6}$	$\mathrm{29.5}\pm \mathrm{0.2}$	$\mathrm{338.3}\pm \mathrm{0.04}$	$\mathrm{4.5}\pm \mathrm{0.9}$	$\mathrm{0.8}\pm \mathrm{0.9}$	$\mathrm{3.7}\pm \mathrm{0.5}$
3	$\mathrm{2.4048}\pm \mathrm{6}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{86}\pm \mathrm{7}$	$\mathrm{34.2}\pm \mathrm{0.2}$	$\mathrm{339.5}\pm \mathrm{0.03}$	$\mathrm{3.6}\pm \mathrm{0.6}$	$\mathrm{0.0}\pm \mathrm{0.8}$	$\mathrm{2.6}\pm \mathrm{0.4}$
4	$\mathrm{2.4048}\pm \mathrm{6}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{128}\pm \mathrm{10}$	$\mathrm{52.2}\pm \mathrm{0.1}$	$\mathrm{338.2}\pm \mathrm{0.04}$	$\mathrm{1.7}\pm \mathrm{0.6}$	$-\mathrm{1.0}\pm \mathrm{0.8}$	$-\mathrm{0.8}\pm \mathrm{0.4}$
5	$\mathrm{2.4048}\pm \mathrm{6}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{513}\pm \mathrm{40}$	$\mathrm{84.8}\pm \mathrm{0.1}$	$\mathrm{354.9}\pm \mathrm{0.02}$	$\mathrm{1.0}\pm \mathrm{0.6}$	$\mathrm{0.4}\pm \mathrm{0.7}$	$-\mathrm{0.8}\pm \mathrm{0.5}$
Exp. K
1	$\mathrm{2.419}\pm \mathrm{1.0}\times {\mathrm{10}}^{-\mathrm{2}}$	$\mathrm{117}\pm \mathrm{9}$	$\mathrm{37.4}\pm \mathrm{2.7}$	$\mathrm{342.5}\pm \mathrm{0.05}$	$\mathrm{4.5}\pm \mathrm{0.5}$	$-\mathrm{1.9}\pm \mathrm{0.7}$	$\mathrm{2.1}\pm \mathrm{0.4}$
2	$\mathrm{2.430}\pm \mathrm{2}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{117}\pm \mathrm{9}$	$\mathrm{44.2}\pm \mathrm{0.3}$	$\mathrm{337.2}\pm \mathrm{0.03}$	$\mathrm{2.3}\pm \mathrm{0.7}$	$\mathrm{0.1}\pm \mathrm{0.7}$	$\mathrm{2.1}\pm \mathrm{0.5}$
Exp. L
1	$\mathrm{2.268}\pm \mathrm{1}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{117}\pm \mathrm{9}$	$\mathrm{43.5}\pm \mathrm{2.0}$	$\mathrm{316.4}\pm \mathrm{0.05}$	$\mathrm{3.7}\pm \mathrm{0.5}$	$-\mathrm{1.7}\pm \mathrm{0.6}$	$\mathrm{1.6}\pm \mathrm{0.4}$
2	$\mathrm{2.406}\pm \mathrm{4.0}\times {\mathrm{10}}^{-\mathrm{2}}$	$\mathrm{89}\pm \mathrm{7}$	$\mathrm{38.0}\pm \mathrm{1.3}$	$\mathrm{329.8}\pm \mathrm{0.09}$	$\mathrm{3.4}\pm \mathrm{1.2}$	$\mathrm{0.5}\pm \mathrm{0.8}$	$\mathrm{2.2}\pm \mathrm{0.8}$
3	$\mathrm{2.406}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{2}}$	$\mathrm{135}\pm \mathrm{10}$	$\mathrm{54.3}\pm \mathrm{6.8}$	$\mathrm{337.8}\pm \mathrm{0.19}$	$\mathrm{2.6}\pm \mathrm{0.5}$	$\mathrm{0.0}\pm \mathrm{0.5}$	$\mathrm{1.5}\pm \mathrm{0.4}$
4	$\mathrm{2.4018}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{86}\pm \mathrm{7}$	$\mathrm{37.3}\pm \mathrm{0.8}$	$\mathrm{337.7}\pm \mathrm{0.14}$	$\mathrm{3.8}\pm \mathrm{1.4}$	$\mathrm{1.6}\pm \mathrm{0.9}$	$\mathrm{1.7}\pm \mathrm{0.7}$
5	$\mathrm{2.4018}\pm \mathrm{7}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{141}\pm \mathrm{11}$	$\mathrm{56.8}\pm \mathrm{0.5}$	$\mathrm{338.2}\pm \mathrm{0.13}$	$\mathrm{1.7}\pm \mathrm{0.5}$	$-\mathrm{0.3}\pm \mathrm{0.6}$	$\mathrm{1.1}\pm \mathrm{0.4}$

3.2 Model results

Parameters $a$ and $b$ in Eq. () were determined from the experimental data. For the single-tube system the values were fitted to steps D2 and D6–D9. Here two linear regimes were found and were fitted by two sets of $a$ and $b$ constants. In this way we could describe the effect of the thermal management system used in later experiments.

Figure 6

RE% for the steps of exp. D, F and H as found experimentally (white stripes) and by the model (grey). (a) Steps D2 and D6–D9 were utilized to generate $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ for the single-tube system. (b) Exp. E and (c) exp. H.

[Figure omitted. See PDF]

The $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ for the single-tube systems is obtained from Eqs. () and () (Fig. c and a). These equations are used to calculate $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ for exp. B, C and D. The comparison between the modeled and experimental efficiency is shown in Fig. .

$J_{{\mathrm{Cl}}_{\mathrm{2}}}$ was determined using the same method. Exp. I is used to obtain model $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ (Fig. b–d and Eq. ).

In Fig. c a comparison of experimental and model results is shown for exp. H, D and E. The model yields good agreement with the experimental results. However, the model slightly underestimates RE% for most of the steps, which is also observed for the other experiments. The initial instability can also be seen for steps D1 and D2 depicted in Fig. a. Problems due to overheating at high $P_{\mathrm{IN}}$ are eliminated with the improved photochemical device, resulting in a power effectiveness at 15 W of 0.6 % for the single tube to 9 % for the multiple-tube system.

Overall, the simple model does a reasonable job of describing the experimental results, although it underestimates the removal efficiency. One issue is that the model does not do a good job of describing the effect of variations of initial methane concentrations in exp. G, as shown in Fig. e.

Additional model runs are used to estimate $J_{\mathrm{Cl}\mathrm{2}}$ of experiments E and F, which are conducted with a modified device; see Eqs () and ()–(), respectively. It is clear that adjusting $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ results in a model that more accurately fits the experimental results.

3.2.1 Parameters simulated and compared with experimental results

Exp. I was chosen as the basis for the final simulation: three parameters are fixed and the fourth varies. The methane removal efficiency, chlorine radical abundance and the resulting abundance of [${\mathrm{CCl}}_{\mathrm{4}}$] are determined. The standard values and the ranges investigated can be seen in Table .

Table 5

Parameter ranges.

Parameter	Standard value	Range
${\mathrm{Cl}}_{\mathrm{2}}$	50 ppm	20–100 ppm
${\mathrm{CH}}_{\mathrm{4}}$	2.04 ppm	0.5–50 ppm
Residence time	165 s	40–400 s
$P_{\mathrm{IN}}$	14.5 W	9–31 W
${\mathrm{O}}_{\mathrm{2}}$	10 %
${\mathrm{N}}_{\mathrm{2}}$	90 %

Figure 7

The removal efficiency of methane (black), [${\mathrm{CCl}}_{\mathrm{4}}$] (red) and [Cl] (grey) is shown in panels (a)–(d). The four parameters are varied while the remaining parameters are kept at the standard parameter presented in Table . (a) The [${\mathrm{Cl}}_{\mathrm{2}}$] is varied. (b) The initial [${\mathrm{CH}}_{\mathrm{4}}$] is varied. (c) The $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ is varied. (d) The $t_{\mathrm{R}}$ is varied.

[Figure omitted. See PDF]

The resulting removal efficiencies as a function of each of the four parameters power input $P_{\mathrm{IN}}$ ($\mathrm{W}$), residence time $t_{\mathrm{R}}$ ($\mathrm{s}$), [${\mathrm{Cl}}_{\mathrm{2}}$] ($\mathrm{ppm}$) and [${\mathrm{CH}}_{\mathrm{4}}$] ($\mathrm{ppm}$) are shown in Fig. . The model results are compared with the experimental results for the parameters $P_{\mathrm{IN}}$ ($\mathrm{W}$), $t_{\mathrm{R}}$ ($\mathrm{s}$) and chlorine mixing ratio ($\mathrm{ppm}$), as shown in Fig. b, a and c, respectively. A good match in the observed response can be seen. The model is too insensitive to methane concentration and fails to recreate the slope observed from the experimental results. The comparison between the model (Fig. d) and the experimental results (Fig. d) shows that the model RE% scale is approximately 1$/$10 that of the experimental results. This may simply be due to the temperature dependence of the methane reaction rate. Simulations with an increased $k_{\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{4}}}$ resulted in better agreement.

The corresponding ${\mathrm{Cl}}_{\mathrm{2}}$ photodissociation rates for the $P_{\mathrm{IN}}$ in Fig. a range from $\mathrm{4.04}\times {\mathrm{10}}^{-\mathrm{3}}$ to $\mathrm{2.37}\times {\mathrm{10}}^{-\mathrm{2}}$ photons s${}^{-\mathrm{1}}$, which is a good match with previous $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ values found for a similar system .

In addition to the RE%, [$\mathrm{Cl}$] and [${\mathrm{CCl}}_{\mathrm{4}}$] are also shown in the aforementioned figures. Chlorinated side products such as ${\mathrm{CH}}_{\mathrm{3}}\mathrm{Cl}$ and ${\mathrm{CCl}}_{\mathrm{4}}$ were investigated as another potential concern due to climate . Figure a shows that an increase in ${\mathrm{Cl}}_{\mathrm{2}}$ concentrations increases the [${\mathrm{CCl}}_{\mathrm{4}}$] production. The amounts of carbon tetrachloride formed are under parts per trillion for initial methane concentrations of tens of parts per million, i.e., yield of the order of less than ${\mathrm{10}}^{-\mathrm{7}}$.

3.2.2 Side reactions and products

The formation of $\mathrm{HCl}$ is unavoidable. As expected, the higher photolysis rate leads to more efficient methane oxidation, and [HCl] rises accordingly. Therefore, scrubber technologies may be necessary, though the use of water bubblers would impose big issues for reliable measurements of isotopologues. The ${\mathrm{NO}}_x$ concentration in our experiments is insignificant, and hence these reactions have not been included in the model.

4 Conclusions

In this study we have described the design, improvement and performance of a process for continuously removing methane from an airstream. The system is based on the photolysis of chlorine gas using UV LEDs to generate chlorine radicals. The performance of the setup was investigated on the basis of four variables: [${\mathrm{CH}}_{\mathrm{4}}$], [${\mathrm{Cl}}_{\mathrm{2}}$], photolysis rate and $t_{\mathrm{R}}$.

A model was built and used to describe the chemistry in more detail, as well as to optimize the performance of the process. In addition, the model found that ${\mathrm{CCl}}_{\mathrm{4}}$ was produced at negligible levels. The highest removal levels achieved experimentally at ambient methane levels were above 98 %, which was maintained under stable conditions. A level above 99.5 % would be achievable by increasing the chlorine concentration or extending the photolysis time. The system was tested using ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ isotope measurements, a case in which methane is known to interfere with measurements of ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\alpha }}$ and ${\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$. With the inclusion of a sofnocat trap to control $\mathrm{CO}$, the setup was able to remove all interference from ${\mathrm{H}}_{\mathrm{2}}\mathrm{O}$, ${\mathrm{CO}}_{\mathrm{2}}$ and $\mathrm{CO}$, and it removed 84.5 % of ${\mathrm{CH}}_{\mathrm{4}}$. While this is not sufficient to remove the effect from ${\mathrm{CH}}_{\mathrm{4}}$, we are confident that with an optimized setup and settings the method can be used to reliably remove $>\mathrm{95}$ % of ${\mathrm{CH}}_{\mathrm{4}}$, thereby enabling continuous accurate measurements of [${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$] and its isotopically substituted analogs using the Picarro G5131-i.

We believe that researchers will be able to use this approach to continuously remove methane from a sample, thereby eliminating interference and improving accuracy.

Appendix A Proof-of-concept experiments – preliminary experiments

Proof-of-concept experiments were conducted to investigate the feasibility of the proposed mechanism.

The ambient air standard was enriched in ${\mathrm{Cl}}_{\mathrm{2}}$ by in situ production of ${\mathrm{Cl}}_{\mathrm{2}}$, ranging from $\mathrm{1}$ to $<\mathrm{20}$ ppm, through electrolysis of a saltwater mixture. Following that, the sample was photolyzed in a photochemical device generating $\mathrm{Cl}$ radicals. The resulting drop in methane was monitored with a cavity ring-down spectrometer (Picarro G1301).

The photochemical device comprised 28 LEDs (385 $\mathrm{nm}$) (UV LED LAMP-VAOL-5EUV8T4) spaced evenly in a polyvinyl chloride plastic housing. The last set of experiments used a high-pressure xenon lamp (ILC Technology R100-IB) equipped with an optical filter at 335 $\mathrm{nm}$. The resulting peak removal efficiencies for the preliminary experiments are presented in Table .

Table A1

Removal efficiencies for the preliminary experiments.

Experimental setup	Highest stable	Initial [${\mathrm{CH}}_{\mathrm{4}}$]
(date)	RE%	($\mathrm{ppm}$)
A (17/4)	68 %	2
A (23/4)	67.75 %	1.98
A (24/4)	76.48 %	1.98
B1 (26/4)	78.52 %	2
B2 (30/4)	80.16 %	2
C2 (26/5)	98.20 %	2

The system yielded an average methane depletion of 86.63 % with a peak depletion at 98.2 %. Various parameters were changed throughout the experiments, and it was determined that the methane depletion is highly dependent on the flow, chlorine production and light source. A better control of these parameters will yield higher and steadier removal of methane.

Figure A1

Experimental setup B2 with the inclusion of an activated carbon trap. Gas flask: ambient air sample, MFM: mass flowmeter, EC: electrolytic device, PC: photochemical device.

[Figure omitted. See PDF]

The experimental setup B2 is presented in Fig. .

A1 The electrolytic device

The experimental setups presented in Table  use an electrolytic device to produce chlorine gas. The electrolytic device is housed in a polycarbonate box. A Nafion membrane (Chemours, Nafion N234) is installed, dividing the volume into two half-cells. Two electrodes are installed, and the two cells consist of two different solutions of $\mathrm{NaCl}$ in Milli-Q water. The average concentration of $\mathrm{NaCl}$ is 1.3 M at the anodic site and 0.13 M at the cathodic site. The electrodes are carbon electrodes with a diameter of 2 mm and a length of 10 cm. On the anodic side ${\mathrm{Cl}}_{\mathrm{2}}$ is produced .

Anode reaction:

AR1$$\mathrm{2}{\mathrm{Cl}}^{-}\to {\mathrm{Cl}}_{\mathrm{2}}\left(g\right)+\mathrm{2}{e}^{-}$$

Cathode reaction: AR2$$\mathrm{2}{\mathrm{H}}^{+}+\mathrm{2}{e}^{-}\to {\mathrm{H}}_{\mathrm{2}}\left(g\right)$$

Overall reaction: AR3$$\mathrm{2}\mathrm{NaCl}+\mathrm{2}{\mathrm{H}}_{\mathrm{2}}\mathrm{O}\to {\mathrm{Cl}}_{\mathrm{2}}+{\mathrm{H}}_{\mathrm{2}}+\mathrm{2}\mathrm{NaOH}$$

The presence of the membrane is essential due to its selectivity to cations. The membrane allows ${\mathrm{Na}}^{+}$ ions move from the anode to the cathode and form $\mathrm{NaOH}$. If the membrane was not present the $\mathrm{NaOH}$ would encounter ${\mathrm{Cl}}_{\mathrm{2}}$ and form hypochlorite. AR4$$\mathrm{2}{\mathrm{H}}_{\mathrm{2}}\mathrm{O}\to {\mathrm{O}}_{\mathrm{2}}\left(g\right)+\mathrm{4}{\mathrm{H}}^{+}\left(\mathrm{aq}\right)+\mathrm{4}{e}^{-}$$

A2 The electrolysis chamber

In the experimental setups A to B2 (Table ), an electrolysis chamber is used to generate ${\mathrm{Cl}}_{\mathrm{2}}$; see Fig. . The chamber is made from PVC plastic; 28 LED (385 nm; UV LED LAMP-VAOL-5EUV8T4) diodes were installed in the chamber, directed at a quartz tube (o.d. 4 mm, length 20 cm) placed through the chamber. The LEDs are connected in parallel with a forward voltage and forward current. The max current is 20 mA for each LED, and the max voltage is 3.6 $\mathrm{V}$. The same voltage runs through the LED and the current is multiplied by the number of lamps, resulting in 0.480 $\mathrm{\AA }$.

The chlorine gas is introduced into the gas stream by using a funnel above the anode. The water level is adjusted to yield optimal conditions for ${\mathrm{Cl}}_{\mathrm{2}}$ to get into the gas stream and avoid chlorine being deposited on the water surface or water getting sucked into the gas stream.

A3 Additional equipment

The Picarro G1301 has a cavity pressure of 18.7 $\mathrm{kPa}$, nominal ambient temperature (DAS temperature) of 30.2 ${}^{\circ }$C and cavity temperature of 45 ${}^{\circ }$C.

We used a cylinder of compressed air with stable mole fractions of ${\mathrm{CH}}_{\mathrm{4}}$ (1.98 ppm), ${\mathrm{CO}}_{\mathrm{2}}$ (376.1 ppm) and ${\mathrm{H}}_{\mathrm{2}}\mathrm{O}$ (1.175 % vol).

The ${\mathrm{Cl}}_{\mathrm{2}}$ sensor used in all experiments is the PG610-CL2 model, which is a chlorine ${\mathrm{Cl}}_{\mathrm{2}}$ gas detector with a gas sound light vibration alarm. The sensor measures chlorine concentrations from 1–20 ppm. The sensor is placed in a 600 mL glass flask.

The general procedure is as follows.

• Prepare solutions.

• Let the system stabilize.

• Turn on the electrochemical device.

• Let the ${\mathrm{Cl}}_{\mathrm{2}}$ concentration stabilize.

• Turn on lamps.

• Let the system stabilize to ensure a stable RE%.

• Take a 10 min measurement with Tenax tube sampling (experiments B1 and B2).

• Turn off the light.

• Let the system stabilize to the initial methane concentration.

A4 Variations in the experimental setups

Experimental setup A is the initial setup. Experimental setup B1 employed Tenax tube sampling for thermal desorption–gas chromatography mass spectrometry (TD-GCMS) measurements of chlorinated species.

Experimental setup B2 follows the same procedure as B1, but with the addition of an activated carbon trap.

Experimental setup C1 uses a high-pressure xenon lamp (ILC Technology R100-IB). The xenon lamp lights up the second photolysis chamber (PC-2), which is equipped with an 8 $\mathrm{mm}$ diameter and 20 $\mathrm{cm}$ length quartz tube. The inner surface of the cylinder is covered with aluminum foil to reflect the light coming in. The xenon lamp emits light in wavelengths from vacuum UV (200 nm) to infrared ; therefore, a 335 nm optical filter is installed.

At the Picarro G1301 outlet the two traps are used for trapping the gases hydrochloric acid, chlorine gas and carbon dioxide.

Experimental setup C2 is similar to C1; however, the ${\mathrm{Cl}}_{\mathrm{2}}$ concentration is diluted to obtain values above the fixed value of 20 ppm. At the electrochemical device outlet a union tee divides the flow into two channels, one to the PC-2 and the other to the sensor chamber. The flow at the outlet of the sensor chamber is measured by a flowmeter (Agilent ADM) to ensure a flow of approximately 40–50 mL min${}^{-\mathrm{1}}$.

Appendix B

Experimental setups (CH${}_{\mathrm{4}}$ and N${}_{\mathrm{2}}$O)

The photochamber for the high-pressure xenon lamp (HPXL) setup uses a quartz tube with dimensions 20 $\mathrm{cm}$ in length and $\mathrm{1}/\mathrm{2}$ in. (12.7 mm) in outer diameter placed in a cylinder coated with aluminum.

Figure B1

(a) General setup for setups 2–4 (Table D4). (b) General setup for ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ experiments. Gas flasks are presented in Table . ACT: activated carbon trap. MFC: mass flow controller.

[Figure omitted. See PDF]

Table B1

Table summarizing the gas flask used in the experiments.

Flask name	${\mathrm{CH}}_{\mathrm{4}}$	${\mathrm{Cl}}_{\mathrm{2}}$	${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$	Matrix composition	Flow range
	(ppm)	(ppm)	(ppb)		(mL min${}^{-\mathrm{1}}$)
I	0	$\mathrm{100}\pm \mathrm{2.5}$	0	$>\mathrm{99}$ % ${\mathrm{N}}_{\mathrm{2}}$	6–23
II	$\mathrm{2.003}\pm \mathrm{5}\times {\mathrm{10}}^{-\mathrm{4}}$	0	0	Atmospheric air	1–29
III	$\mathrm{78}\pm \mathrm{2}$	0	0	20.95 % ${\mathrm{O}}_{\mathrm{2}}$ $+>\mathrm{79}$ % ${\mathrm{N}}_{\mathrm{2}}$	0.3–1.2
IV	0	0	500	Atmospheric air	28–50

The photochemical device (PD; Fig. ) for later experiments (Fig. ) consists of 420 LEDs at 365 nm peak wavelength. The LEDs run in a parallel circuit with a forward voltage and forward current (from positive to negative). The max current is 13.2 mA for each LED, and the max voltage is 3.8 V. The same voltage runs through the LEDs, resulting in a total current across the system of 5.5 A.

Figure B2

Hexagonal photochemical device consisting of connected circuit boards of 420 LEDs at 365 nm.

[Figure omitted. See PDF]

The difference between the two similar setups 2 and 3 is that the forward pressure valve is exchanged with a mass flow controller to allow for a smaller and more stable level of vent flow. The quartz tube of the previous experiments is substituted with seven smaller quartz tubes for setup 4 to yield a longer $t_{\mathrm{R}}$.

B1 Experimental procedure

• Tune the desired flow from flask C for methane and mix it with a flow from flask B equal to the desired flow plus the intended flow from flask A.

• Let the system stabilize.

• Add the desired flow of chlorine from flask A by adjusting the pressure at the flask.

• Reduce the flow from flask B by an equal amount to get the desired mixing ratio.

• Let the system stabilize and confirm that the resulting total flow fits the expected flow. Make sure the chlorine value can be read on the chlorine sensor.

• When a stable methane level has been run for sufficient time, turn on the photochemical device.

• Let the system stabilize to ensure a stable methane RE%.

• Turn off the light.

• Let the system stabilize to the initial methane concentration before the light is turned on.

B2

N${}_{\mathrm{2}}$O experiments

Experiments were conducted with the Picarro model G5131-i, which is used to measure ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ mixing ratio and isotopic abundance. The purpose of the experiments was to confirm that the illumination did not affect ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$. The general experimental setup is shown in Fig. b. The sofnocat trap was prepared with 1.25 g of sofnocat contained in a 6.4 mm diameter tube of length 8 cm and kept in place by glass wool. The trap was installed to prevent effects on the ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ isotope signal from $\mathrm{CO}$, as presented in . The presence of $\mathrm{CO}$ 1 ppm gives rise to an erroneous offset in the observed isotopologue values of 1.2, 2.4 and 0.4 ‰ for ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\alpha }}$, ${\mathit{\delta }}^{\mathrm{15}}\mathrm{N}$${}^{\mathit{\beta }}$ and ${\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$, respectively. The installation of this trap after the ${\mathrm{CO}}_{\mathrm{2}}$ trap allowed us to measure the amount of $\mathrm{CO}$ present. The technical air from flask C was exchanged with a technical air mix with 509 $\mathrm{ppb}$ [${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$], allowing for dilution to the ambient level. The flow ratio between the three different gases was regulated to maintain a mixing ratio of 330 $\mathrm{ppb}$ ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$, 2.4 ppm ${\mathrm{CH}}_{\mathrm{4}}$ and 33 ppm ${\mathrm{Cl}}_{\mathrm{2}}$. Power supply to the lamp was constant at 4.8 V and 5.0 A, and $t_{\mathrm{R}}$ in the chamber was varied between 86, 117 and 145 s.

Appendix C Theoretical models

The model is made with the program Kintecus version 6.8 . The model was developed by describing the relevant reactions with rates for chlorine radical production and/or removal as well as formation of chlorinated species. The model was kept as simple as possible while still including key reactions. The reactions and their rates used in the model are found in Tables –. A simplified reaction scheme is shown in Fig. . The experiments are modeled by choosing both the initial and external concentrations of the species used and the $t_{\mathrm{R}}$ within the chamber. A continuous flow was modeled by setting the initial and external concentrations of gases flowing through the chamber to the same value. This is done for the gases ${\mathrm{Cl}}_{\mathrm{2}}$, ${\mathrm{CH}}_{\mathrm{4}}$, ${\mathrm{N}}_{\mathrm{2}}$ and ${\mathrm{O}}_{\mathrm{2}}$.

The physical parameters are fixed as well: temperature at 298 K, starting integration time to ${\mathrm{10}}^{-\mathrm{6}}$ s (starting step for the integrated model), maximum integration time to 1 s, simulation length equal to $t_{\mathrm{R}}$ plus 5 s and the accuracy of digits to ${\mathrm{10}}^{-\mathrm{4}}$. Furthermore, the energy unit kilocalorie (kcal) was selected, and the unit of concentration was selected to be molecules per cubic centimeter ($\mathrm{molec}.{\mathrm{cm}}^{-\mathrm{3}}$).

C1 Radical wall reactions

As described in the main article a set of radical-terminating reactions was incorporated into the model. The wall reaction rates were estimated based on the diffusion rate of the radicals and the diffusion length. The diffusion length is calculated as the average distance from the wall. Because two different sizes of tubes were used throughout the experiments, the wall reactions reflect that. The diffusion length and the diffusion rate are given in Eqs. () and (), respectively:

C1$$l=\mathrm{2}\cdot (D\cdot t{)}^{\mathrm{0.5}},$$ where $D$ is the diffusion constant and $t$ is time. C2$$k={\displaystyle \frac{l}{t}}={\displaystyle \frac{\mathrm{4}\cdot D}{l^{\mathrm{2}}}}$$ The distance, $l$, is defined as the average distance from the wall, which can alternatively be written as $l=r-d_{\mathrm{c}}$, where $r$ is the radius of the tube, and $d_{\mathrm{c}}$ is distance from a random particle in the cylinder to the center of the circle of the cylinder. Finding the average distance to the wall of an infinite number of randomly located particles in the cylinder can be accomplished by solving Eq. (). The result of Eq. () is used to calculate the resulting diffusion rate with the inclusion of the average distance from the walls of the tube, which is defined in Eq. (). C3$${\displaystyle \frac{\mathrm{1}}{\mathrm{2}}}A=\underset{\mathrm{0}}{\overset{d_{\mathrm{c}}}{\int }}\mathrm{2}r\cdot \mathit{\pi }\mathrm{d}r,$$ Here, $r$ is the radius and $A=r^{\mathrm{2}}\cdot \mathit{\pi }$ is the area. $$\begin{array}{}\text{C4}& {\displaystyle }{d}_{\mathrm{c}}={\left({\displaystyle \frac{A}{\mathrm{2}\mathit{\pi }}}\right)}^{\frac{\mathrm{1}}{\mathrm{2}}}={\displaystyle \frac{r}{\sqrt{\mathrm{2}}}}\text{C5}& {\displaystyle }k={\displaystyle \frac{\mathrm{4}\cdot D}{r-\frac{r}{\sqrt{\mathrm{2}}}}}\end{array}$$ The diffusion constants, diffusion lengths and estimated wall reaction rates are shown in Table .

Table C1

Radical wall reaction parameters.

Setup	Reaction	Diffusion	Reference	Diffusion	Wall reaction
		constant		length (cm)	rate (s${}^{-\mathrm{1}}$)
		(cm${}^{\mathrm{2}}$ s${}^{-\mathrm{1}}$)
Single tube	$\mathrm{Cl}\to \frac{\mathrm{1}}{\mathrm{2}}{\mathrm{Cl}}_{\mathrm{2}}$	0.260		0.146	$\mathrm{1.2}\times {\mathrm{10}}^{\mathrm{2}}$
Multiple tube				0.091	$\mathrm{4.8}\times {\mathrm{10}}^{\mathrm{1}}$
Single tube	$\mathrm{ClO}\to \frac{\mathrm{1}}{\mathrm{2}}{\mathrm{Cl}}_{\mathrm{2}}+\frac{\mathrm{1}}{\mathrm{2}}{\mathrm{O}}_{\mathrm{2}}$	0.184	${}^{\mathrm{b}}$	0.146	$\mathrm{8.8}\times {\mathrm{10}}^{\mathrm{1}}$
Multiple tube				0.091	$\mathrm{3.4}\times {\mathrm{10}}^{\mathrm{1}}$
Single tube	$\mathrm{OH}\to \frac{\mathrm{1}}{\mathrm{2}}{\mathrm{H}}_{\mathrm{2}}\mathrm{O}+\frac{\mathrm{1}}{\mathrm{4}}{\mathrm{O}}_{\mathrm{2}}$	0.217		0.146	$\mathrm{1.0}\times {\mathrm{10}}^{\mathrm{2}}$
Multiple tube				0.091	$\mathrm{4.0}\times {\mathrm{10}}^{\mathrm{1}}$
Single tube	${\mathrm{HO}}_{\mathrm{2}}\to \frac{\mathrm{1}}{\mathrm{2}}{\mathrm{H}}_{\mathrm{2}}\mathrm{O}+\frac{\mathrm{3}}{\mathrm{4}}{\mathrm{O}}_{\mathrm{2}}$	0.139	${}^{\mathrm{a}}$	0.146	$\mathrm{6.7}\times {\mathrm{10}}^{\mathrm{1}}$
Multiple tube				0.091	$\mathrm{2.6}\times {\mathrm{10}}^{\mathrm{1}}$

${}^{\mathrm{a}}$ The diffusion coefficient is estimated from $D_{\mathrm{HO}-\mathrm{Air}}$ and $D_{{\mathrm{HO}}_{\mathrm{2}}-\mathrm{He}}$. ${}^{\mathrm{b}}$ The diffusion model was used to estimate the diffusion constant.

C2

$J_{{\mathrm{Cl}}_{\mathrm{2}}}$ estimation

C2.1 First approach

The first approach is to fit $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ in the model to regenerate the observed removal efficiencies from experimental results. These fits were only produced for experiments investigating the effect of $P_{\mathrm{in}}$. The resulting $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ was related to $P_{\mathrm{in}}$ via the effective power-to-light conversion based on the absorption cross section of ${\mathrm{Cl}}_{\mathrm{2}}$ and the wavelength distribution of the LEDs. $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ was determined in this manner, once for the single-tube systems and once for the multiple-tube systems. The photolysis rate $J$ (photons${}^{-\mathrm{1}}$) can be determined by Eq. ():

C6$$J_{{\mathrm{Cl}}_{\mathrm{2}}}=\int \mathit{\sigma }(\mathit{\lambda },T)\cdot \mathit{\varphi }(\mathit{\lambda },T)\cdot I(\mathit{\lambda },W)\mathrm{d}\mathit{\lambda },$$ where $\mathit{\sigma }(\mathit{\lambda },T)$ is the wavelength-dependent cross section of ${\mathrm{Cl}}_{\mathrm{2}}$ (cm${}^{\mathrm{2}}$ molec.${}^{-\mathrm{1}}$), $\mathit{\varphi }(\mathit{\lambda },T)$ is the quantum yield and $I(\mathit{\lambda },W)$ is the spectral actinic flux density (photons cm${}^{-\mathrm{2}}$ s${}^{-\mathrm{1}}$ nm${}^{-\mathrm{1}}$). The cross section of chlorine dissociation in the range 250–550 nm is defined by Eq. () . C7$$\begin{array}{rl}\mathit{\sigma }(\mathit{\lambda },T)& ={\mathrm{10}}^{-\mathrm{20}}\left(\tanh \right({\displaystyle \frac{\mathrm{402.7}}{T}}){)}^{\mathrm{0.5}}\\ & \cdot (\mathrm{27.3}\cdot e^{-\mathrm{99.0}\cdot \left(\tanh \right(\frac{\mathrm{402.7}}{T}\left)\right)\cdot \left(\ln \right(\frac{\mathrm{329.5}}{\mathit{\lambda }}){)}^{\mathrm{2}}}\\ & +\mathrm{0.932}\cdot e^{-\mathrm{91.5}\cdot \tanh \left(\frac{\mathrm{402.7}}{T}\right)\cdot \left(\ln \right(\frac{\mathrm{406.5}}{\mathit{\lambda }}){)}^{\mathrm{2}}})\end{array}$$ Here, $T$ is the temperature, and $\mathit{\lambda }$ is the wavelength (nm). C8$$I(\mathit{\lambda },W)={\displaystyle \frac{P(\mathit{\lambda },W)\cdot D\left(\mathit{\lambda }\right)\cdot l}{V}}$$ The actinic flux (Eq. C8) is a function dependent on the power output $P(\mathit{\lambda },W)$ from Eq. (), the distribution $D\left(\mathit{\lambda }\right)$ from Eq. () and the tube volume ($V$): C9$$P(\mathit{\lambda },W)=\text{Eff}\left(\mathrm{W}\right)\cdot {\displaystyle \frac{\mathit{\lambda }}{hc}},$$ where $h$ is Planck's constant, and $c$ is the speed of light.

It was observed that the photolysis rate did not scale linearly with the applied power, which we speculate may be due to variation of the efficiency of the lamp with the applied current and operating temperature. This effect was sufficiently accounted for by a linear fit and is defined as Eff ($\mathrm{W}$): C10$$\text{Eff}\left(\mathrm{W}\right)=a\cdot W+b,$$ where $W$ is the power supplied to the diodes, and values for the constants $a$ and $b$ are fitted in the model to match the experiment. The function () accounts for additional variations such as effects due to temperature, the cross-sectional area of the quartz tube, the conductance of the photochamber and the quality of the distribution fit. This is reflected in the constants $a$ and $b$ varying in response to changes in these parameters. As this is used as a simple empirical stand-in function we do not intend to speculate further on how these changes change the constants.

The photon output (Eq. ) from the LED was assumed to follow a normal distribution. For this distribution shown in Eq. (), we assumed a center value of 365 nm and full width at half-maximum of 10 nm. The distribution (Eq. ) is per nanometer (${\mathrm{nm}}^{-\mathrm{1}}$). C11$$D\left(\mathit{\lambda }\right)={\displaystyle \frac{\mathrm{1}}{\mathrm{10}\mathrm{nm}\cdot (\mathrm{2}\mathit{\pi }{)}^{\mathrm{0.5}}\cdot e^{-\mathrm{0.5}\cdot {\left(\frac{\mathit{\lambda }-\mathrm{365}\mathrm{nm}}{\mathrm{10}\mathrm{nm}}\right)}^{\mathrm{2}}}}}$$ The photolysis rate could then be calculated by Eq. () across 250–500 nm at 298 K.

C2.2 Second approach

A second approach for estimating $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ and relating it to $P_{\mathrm{IN}}$ was used. This method estimated $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ by using simplified kinetics and relating it to power via the same method as the model-derived $J_{{\mathrm{Cl}}_{\mathrm{2}}}$. Exp. F reflects the single-tube system, while exp. I reflects the optimized multiple-tube setup. Four main reactions, ()–(), are considered in the simple kinetic model. $$\begin{array}{cc}\text{CR1}& {\displaystyle }& {\displaystyle }{\mathrm{Cl}}_{\mathrm{2}}+hv\stackrel{J_{{\mathrm{Cl}}_{\mathrm{2}}}}{⟶}\mathrm{2}\mathrm{Cl}\text{CR2}& {\displaystyle }& {\displaystyle }\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{4}}\stackrel{k_{\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{4}}}}{⟶}{\mathrm{CH}}_{\mathrm{3}}+\mathrm{HCl}{\displaystyle }& {\displaystyle }[k_{\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{4}}}=\mathrm{1.07}q\times {\mathrm{10}}^{-\mathrm{13}}\cdot \mathrm{molec}{.}^{-\mathrm{1}}{\mathrm{cm}}^{\mathrm{3}}{\mathrm{s}}^{-\mathrm{1}}]\\ \text{CR3}& {\displaystyle }& {\displaystyle }\mathrm{Cl}+\mathrm{Cl}+M\stackrel{k_{\mathrm{self}}}{⟶}{\mathrm{Cl}}_{\mathrm{2}}+M{\displaystyle }& {\displaystyle }[k_{\mathrm{self}}=\mathrm{1.24}q\times {\mathrm{10}}^{-\mathrm{32}}\cdot \mathrm{molec}{.}^{-\mathrm{2}}{\mathrm{cm}}^{\mathrm{6}}{\mathrm{s}}^{-\mathrm{1}}\cdot [M\left]\right]\\ \text{CR4}& {\displaystyle }& {\displaystyle }\mathrm{Cl}\stackrel{k_{\mathrm{wall}}}{⟶}{\displaystyle \frac{\mathrm{1}}{\mathrm{2}}}{\mathrm{Cl}}_{\mathrm{2}}{\displaystyle }& {\displaystyle }[k_{\mathrm{wall}}=\mathrm{124.5}\text{or}\mathrm{48.9}{\mathrm{s}}^{-\mathrm{1}}].\end{array}$$

The $\mathrm{Cl}$ radicals are consumed at a fast rate; therefore, a steady-state approximation for $\mathrm{Cl}$ has been assumed.

C12$$\begin{array}{rl}{\displaystyle \frac{\mathrm{d}\left[\mathrm{Cl}\right]}{\mathrm{d}t}}& =\mathrm{2}\cdot J_{{\mathrm{Cl}}_{\mathrm{2}}}\left[{\mathrm{Cl}}_{\mathrm{2}}\right]-(\mathrm{2}\cdot k_{\mathrm{self}}\cdot [\mathrm{Cl}{]}^{\mathrm{2}}\\ & +k_{\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{4}}}\cdot \left[{\mathrm{CH}}_{\mathrm{4}}\right]\cdot \left[\mathrm{Cl}\right]+k_{\mathrm{wall}}\cdot \left[\mathrm{Cl}\right])=\mathrm{0}\end{array}$$ The photolysis rate for the kinetic calculation is thereby defined in Eq. (). C13$$J_{\mathrm{kin}}={\displaystyle \frac{\mathrm{2}\cdot k_{\mathrm{self}}\cdot [\mathrm{Cl}{]}^{\mathrm{2}}+k_{\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{4}}}\left[{\mathrm{CH}}_{\mathrm{4}}\right]\left[\mathrm{Cl}\right]+k_{\mathrm{wall}}\left[\mathrm{Cl}\right]}{\mathrm{2}\cdot \left[{\mathrm{Cl}}_{\mathrm{2}}\right]}}$$ The photolysis rate is calculated from an estimated [$\mathrm{Cl}$] concentration. This was achieved by assuming that the methane concentration would follow an exponential decay with time (Eq. ). The estimated [$\mathrm{Cl}$] is expressed in Eq. (): C14$$[{\mathrm{CH}}_{\mathrm{4}}{]}_t=[{\mathrm{CH}}_{\mathrm{4}}{]}_{\mathrm{0}}\cdot \exp (-k_{\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{4}}}\cdot \left[\mathrm{Cl}\right]\cdot t),$$ where [${\mathrm{CH}}_{\mathrm{4}}$]${}_t$ is the methane concentration at time $t$, while [${\mathrm{CH}}_{\mathrm{4}}$]${}_{\mathrm{0}}$ is the initial concentration. C15$$\left[\mathrm{Cl}\right]=\ln \left({\displaystyle \frac{\mathrm{1}}{\mathrm{1}-\text{RE}}}\right)/(k_{\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{4}}}\cdot t)$$ The values for $J_{\mathrm{kin}}$ are generated by inserting the experimental values of [${\mathrm{Cl}}_{\mathrm{2}}$], [${\mathrm{CH}}_{\mathrm{4}}$] and the estimated value of [$\mathrm{Cl}$] into Eq. ().

The distribution function $D\left(\mathit{\lambda }\right)$ from Eq. () can be used in combination with the cross section to determine the scale factor $J_{\mathrm{scale}}$. C16$$J_{\mathrm{scale}}(\mathit{\lambda },T)=\underset{\mathrm{250}\mathrm{nm}}{\overset{\mathrm{500}\mathrm{nm}}{\int }}{\displaystyle \frac{\mathit{\lambda }}{hc}}\cdot \mathit{\sigma }(\mathit{\lambda },T)\cdot {\displaystyle \frac{l\cdot D\left(\mathit{\lambda }\right)}{V}}\mathrm{d}\mathit{\lambda }$$ The value of $J_{\mathrm{scale}}$ is calculated from the overlap integral between $\mathit{\sigma }(\mathit{\lambda },T)$ and the emitted photon distribution.

The variable $l$ is the path length across the tube(s) in centimeters (cm), and $V$ is the volume of the tube(s) ($\mathrm{mL}$). $\mathit{\lambda }$ is the wavelength (nm), $h$ is Planck's constant and $c$ is the speed of light. Values for the constants $a$ and $b$ from Eq. () are then fitted to match the photolysis rate in Eq. () with the photolysis rate found from the Kintecus model. C17$$\begin{array}{rl}{P}_{\mathrm{eff}}\left(P_{\mathrm{IN}}\right)& =(a\cdot P_{\mathrm{IN}}+b)\cdot P_{\mathrm{IN}}\\ & =\text{Eff(kin)}\cdot P_{\mathrm{IN}}={\displaystyle \frac{J_{\mathrm{Kin}}}{J_{\mathrm{scale}}\cdot P_{\mathrm{IN}}}}\end{array}$$ Here, $P_{\mathrm{eff}}$ is the effective power, and the constants $a$ and $b$ are setup-dependent constants.

From the effective power output the photolysis rate $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ could be calculated by multiplying $P_{\mathrm{eff}}$ with $J_{\mathrm{scale}}$. C18$$J_{{\mathrm{Cl}}_{\mathrm{2}}}\left(P_{\mathrm{IN}}\right)=P_{\mathrm{eff}}\left(P_{\mathrm{IN}}\right)\cdot J_{\mathrm{scale}}$$

C3

$J_{{\mathrm{Cl}}_{\mathrm{2}}}$ fitted to collected data

The $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ is fit to the data collected for some of the experimental steps for exp. F and I to determine the values for the constants $a$ and $b$.

Exp. F is the single-tube system and exp. I is the optimized multiple-tube system. From the fitted $a$, $b$ and calculated $J_{\mathrm{scale}}$ the photolysis rate could be calculated for the other experiments.

Figure C1

(a) Effectiveness as a function of experimental power input for exp. D. The correlation is used for calculating the effective $P_{\mathrm{IN}}$ for single-tube experiments. (b) Effectiveness as a function of experimental $P_{\mathrm{IN}}$ for exp. I. The correlation is used for calculating the effective power for multiple-tube experiments. (c) Kintecus-obtained $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ as a function of the effective $p_{\mathrm{IN}}$ for exp. D. The effective $p_{\mathrm{IN}}$ is calculated from Fig. a. The combination of the figure with Fig. a is used to calculate the $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ for single-tube experiments by Eqs. () and (). (d) Kintecus-obtained $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ as a function of the effective $p_{\mathrm{IN}}$ in exp. I. The effective $p_{\mathrm{IN}}$ is calculated from Fig. b. The combination of the figure with Fig. b is used to calculate the $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ for multiple-tube experiments by Eq. ().

[Figure omitted. See PDF]

C3.1 Single-tube systems

$J_{{\mathrm{Cl}}_{\mathrm{2}}}$ values are generated on the basis of exp. D2 and D6–D9. The efficiency of $p_{\mathrm{IN}}$ is generated from the $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ model. A correlation between effectiveness ($\mathrm{\%}$) and experimental $p_{\mathrm{IN}}$ ($\mathrm{W}$) is shown in Fig. a as is the correlation with the $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ (Kintecus) values in Fig. c.

The $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ dependence on the $p_{\mathrm{IN}}$ ($\mathrm{W}$) for the single-tube system in exp. B, C and D is given by Eqs. () and (). The equations incorporate a decrease in efficiency of $p_{\mathrm{IN}}$ at higher levels due to overheating of the chamber as seen in Fig. c.

C19$$\begin{array}{rl}J=& \mathrm{2.59}\times {\mathrm{10}}^{-\mathrm{2}}\cdot (\mathrm{2.3}\times {\mathrm{10}}^{-\mathrm{4}}\cdot (P_{\mathrm{IN}}{)}^{\mathrm{2}}\\ & +\mathrm{2.99}\times {\mathrm{10}}^{-\mathrm{3}}\cdot P_{\mathrm{IN}})\\ & \text{if}(P_{\mathrm{IN}}>\mathrm{14.67}\mathrm{W})\end{array}$$ C20$$\begin{array}{rl}J=& \mathrm{2.59}E\times {\mathrm{10}}^{-\mathrm{2}}\cdot (-\mathrm{2.41}\times {\mathrm{10}}^{-\mathrm{4}}\cdot (P_{\mathrm{IN}}{)}^{\mathrm{2}}\\ & +\mathrm{1.15}\times {\mathrm{10}}^{-\mathrm{2}}\cdot P_{\mathrm{IN}})\\ & \text{if}(P_{\mathrm{IN}}<\mathrm{14.67}\mathrm{W})\end{array}$$ The comparison between modeled and experimental efficiency for the single-tube experiments is seen in Fig. a and b.

C3.2 Multiple-tube systems

$J_{{\mathrm{Cl}}_{\mathrm{2}}}$ is generated in the same manner as the experiment for results with multiple tubes. Here exp. I is used to obtain model $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ values (Fig. b and d).

C21$$\begin{array}{rl}J=& \mathrm{5.98}\times {\mathrm{10}}^{-\mathrm{3}}\cdot (\mathrm{2.39}\times {\mathrm{10}}^{-\mathrm{3}}\cdot (P_{\mathrm{IN}}{)}^{\mathrm{2}}\\ & +\mathrm{5.35}E\times {\mathrm{10}}^{-\mathrm{2}}\cdot P_{\mathrm{IN}})\end{array}$$ The overheating at high $p_{\mathrm{IN}}$ is eliminated with the improved photochemical device. This is also apparent when comparing the effectiveness, which is approximately 9 % for the multiple-tube configuration (Fig. b) and approximately 0.6 % for the single-tube system (Fig. a) at the same $p_{\mathrm{IN}}$ of 15 W. Figure e and f show the comparison for exp. G and H, respectively.

C3.3 Exp. E and F

Some experiments cannot be related to the relations presented for the single- and multiple-tube systems. This is due to the optimization done on the photochemical device. A second approach with additional kinetic calculations is therefore used to estimate the $J_{{\mathrm{Cl}}_{\mathrm{2}}}$ of these two experiments. The effectiveness of exp. E is shown in Eq. ().

C22$$P_{\mathrm{eff}}\left(P_{\mathrm{IN}}\right)=P_{\mathrm{IN}}(-\mathrm{4.35}\times {\mathrm{10}}^{-\mathrm{3}}\cdot P_{\mathrm{IN}}+\mathrm{3.26}\times {\mathrm{10}}^{-\mathrm{2}})$$ In the same manner, the effectiveness of exp. F in shown in Eqs. () and (). C23$$\begin{array}{rl}& P_{\mathrm{eff}}\left(P_{\mathrm{IN}}\right)=P_{\mathrm{IN}}\cdot (\mathrm{6.80}\times {\mathrm{10}}^{-\mathrm{4}}\cdot P_{\mathrm{IN}}+\mathrm{4.36}\times {\mathrm{10}}^{-\mathrm{2}})\\ & \text{if}{P}_{\mathrm{IN}}>\mathrm{14.31}\mathrm{W}\end{array}$$ C24$$\begin{array}{rl}{P}_{\mathrm{eff}}\left(P_{\mathrm{IN}}\right)& =P_{\mathrm{IN}}\cdot (-\mathrm{1.57}\times {\mathrm{10}}^{-\mathrm{3}}\cdot P_{\mathrm{IN}}+\mathrm{7.58}\times {\mathrm{10}}^{-\mathrm{2}})\\ & \text{if}{P}_{\mathrm{IN}}<\mathrm{14.31}\mathrm{W}\end{array}$$

Appendix D Settings and experimental results

D1

${\mathrm{CH}}_{\mathrm{4}}$ experimental results

In Tables – the four varying parameters [${\mathrm{CH}}_{\mathrm{4}}$]${}_{\mathrm{initial}}$, [${\mathrm{Cl}}_{\mathrm{2}}$], $t_{\mathrm{R}}$ and $P_{\mathrm{IN}}$ are presented for each experiment alongside the resulting RE%. Table  summarizes the experiments done in the study.

Table D1

Data for exp. A–D. Columns: experimental steps, [${\mathrm{CH}}_{\mathrm{4}}$] (ppm), [${\mathrm{Cl}}_{\mathrm{2}}$] (ppm), residence time $t_{\mathrm{R}}$ (s), power input $p_{\mathrm{IN}}$ (W) and the resulting removal efficiency in %.

Experiment	${\mathrm{CH}}_{\mathrm{4}}$	${\mathrm{Cl}}_{\mathrm{2}}$	Residence time	Power	Removal efficiency
(no.)	($\mathrm{ppm}$)	($\mathrm{ppm}$)	($\mathrm{s}$)	($\mathrm{W}$)	($\mathrm{\%}$)
Exp. A
1	$\mathrm{3.2729}\pm \mathrm{7}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{16.7}\pm \mathrm{1.5}$	$\mathrm{62.4}\pm \mathrm{1.6}$	${}^{*}$	$\mathrm{0.0}\pm \mathrm{0.04}$
2	$\mathrm{2.8327}\pm \mathrm{1.5}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{25}\pm \mathrm{2}$	$\mathrm{62.2}\pm \mathrm{1.6}$	${}^{*}$	$\mathrm{6.4}\pm \mathrm{0.4}$
3	$\mathrm{2.3769}\pm \mathrm{9}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{61.8}\pm \mathrm{1.6}$	${}^{*}$	$\mathrm{27.3}\pm \mathrm{0.4}$
4	$\mathrm{2.9367}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{92}\pm \mathrm{11}$	$\mathrm{62.1}\pm \mathrm{1.6}$	${}^{*}$	$\mathrm{2.7}\pm \mathrm{0.4}$
Exp. B
1	$\mathrm{3.6391}\pm \mathrm{5}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{16.7}\pm \mathrm{1.5}$	$\mathrm{60.7}\pm \mathrm{1.6}$	$\mathrm{17.43}\pm \mathrm{0.03}$	$\mathrm{2.0}\pm \mathrm{0.5}$
2	$\mathrm{3.6598}\pm \mathrm{1}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{16.7}\pm \mathrm{1.5}$	$\mathrm{60.9}\pm \mathrm{1.5}$	$\mathrm{26.13}\pm \mathrm{0.04}$	$\mathrm{6.5}\pm \mathrm{0.2}$
3	$\mathrm{3.7069}\pm \mathrm{1.2}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{16.7}\pm \mathrm{1.5}$	$\mathrm{62.2}\pm \mathrm{1.6}$	$\mathrm{9.91}\pm \mathrm{0.03}$	$\mathrm{6.9}\pm \mathrm{0.2}$
4	$\mathrm{3.7268}\pm \mathrm{6}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{16.7}\pm \mathrm{1.5}$	$\mathrm{62.4}\pm \mathrm{1.6}$	$\mathrm{22.09}\pm \mathrm{0.04}$	$\mathrm{11.0}\pm \mathrm{0.6}$
5	$\mathrm{3.919}\pm \mathrm{1.5}\times {\mathrm{10}}^{-\mathrm{2}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{61.0}\pm \mathrm{1.5}$	$\mathrm{9.92}\pm \mathrm{0.03}$	$\mathrm{23.9}\pm \mathrm{0.8}$
6	$\mathrm{3.945}\pm \mathrm{1.4}\times {\mathrm{10}}^{-\mathrm{2}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{61.4}\pm \mathrm{1.5}$	$\mathrm{16.59}\pm \mathrm{0.03}$	$\mathrm{32.7}\pm \mathrm{0.6}$
Exp. C
1	$\mathrm{3.955}\pm \mathrm{2}\times {\mathrm{10}}^{-\mathrm{2}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{129}\pm \mathrm{4}$	$\mathrm{9.63}\pm \mathrm{0.03}$	$\mathrm{45}\pm \mathrm{5}$
2	$\mathrm{3.957}\pm \mathrm{6}\times {\mathrm{10}}^{-\mathrm{2}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{41.4}\pm \mathrm{1.0}$	$\mathrm{9.63}\pm \mathrm{0.03}$	$\mathrm{18.9}\pm \mathrm{1.3}$
3	$\mathrm{4.0301}\pm \mathrm{5}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{41.4}\pm \mathrm{1.0}$	$\mathrm{17.30}\pm \mathrm{0.03}$	$\mathrm{27.8}\pm \mathrm{1.0}$
4	$\mathrm{3.986}\pm \mathrm{1.0}\times {\mathrm{10}}^{-\mathrm{2}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{128}\pm \mathrm{4.7}$	$\mathrm{17.30}\pm \mathrm{0.03}$	$\mathrm{60}\pm \mathrm{9}$
Exp. D
1	$\mathrm{3.5395}\pm \mathrm{5}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{32}\pm \mathrm{3}$	$\mathrm{62.5}\pm \mathrm{1.5}$	$\mathrm{13.38}\pm \mathrm{0.03}$	$\mathrm{17.6}\pm \mathrm{0.6}$
2	$\mathrm{3.531}\pm \mathrm{1.5}\times {\mathrm{10}}^{-\mathrm{2}}$	$\mathrm{32}\pm \mathrm{3}$	$\mathrm{85}\pm \mathrm{14}$	$\mathrm{13.85}\pm \mathrm{0.03}$	$\mathrm{24.8}\pm \mathrm{0.5}$
3	$\mathrm{3.5570}\pm \mathrm{8}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{32}\pm \mathrm{3}$	$\mathrm{43.4}\pm \mathrm{1.1}$	$\mathrm{13.73}\pm \mathrm{0.03}$	$\mathrm{15.4}\pm \mathrm{0.3}$
4	$\mathrm{3.5405}\pm \mathrm{9}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{32}\pm \mathrm{3}$	$\mathrm{43.0}\pm \mathrm{1.0}$	$\mathrm{13.65}\pm \mathrm{0.03}$	$\mathrm{6.2}\pm \mathrm{0.3}$
5	$\mathrm{3.526}\pm \mathrm{4}\times {\mathrm{10}}^{-\mathrm{2}}$	$\mathrm{32}\pm \mathrm{3}$	$\mathrm{135}\pm \mathrm{58}$	$\mathrm{13.63}\pm \mathrm{0.03}$	$\mathrm{38.2}\pm \mathrm{1.8}$
6	$\mathrm{3.5261}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{32}\pm \mathrm{3}$	$\mathrm{84}\pm \mathrm{2.6}$	$\mathrm{17.65}\pm \mathrm{0.03}$	$\mathrm{33.6}\pm \mathrm{0.9}$
7	$\mathrm{3.564}\pm \mathrm{5}\times {\mathrm{10}}^{-\mathrm{2}}$	$\mathrm{32}\pm \mathrm{3}$	$\mathrm{83}\pm \mathrm{2.2}$	$\mathrm{22.05}\pm \mathrm{0.04}$	$\mathrm{37.0}\pm \mathrm{1.3}$
8	$\mathrm{3.567}\pm \mathrm{5}\times {\mathrm{10}}^{-\mathrm{2}}$	$\mathrm{32}\pm \mathrm{3}$	$\mathrm{83}\pm \mathrm{2.2}$	$\mathrm{28.22}\pm \mathrm{0.04}$	$\mathrm{33.1}\pm \mathrm{1.5}$
9	$\mathrm{3.5729}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{32}\pm \mathrm{3}$	$\mathrm{83}\pm \mathrm{2.2}$	$\mathrm{27.34}\pm \mathrm{0.04}$	$\mathrm{35}\pm \mathrm{15}$
10	$\mathrm{3.5447}\pm \mathrm{7}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{32}\pm \mathrm{3}$	$\mathrm{79}\pm \mathrm{2.1}$	$\mathrm{33.10}\pm \mathrm{0.04}$	$\mathrm{37}\pm \mathrm{14}$

${}^{*}$ The $p_{\mathrm{IN}}$ of the xenon lamp was not varied or determined.

Table D2

Data for exp. E–F. Columns: experimental steps, [${\mathrm{CH}}_{\mathrm{4}}$] ($\mathrm{ppm}$), [${\mathrm{Cl}}_{\mathrm{2}}$] ($\mathrm{ppm}$), residence time $t_{\mathrm{R}}$ ($\mathrm{s}$), power input $p_{\mathrm{IN}}$ ($\mathrm{W}$) and the resulting removal efficiency in %.

Experiment	${\mathrm{CH}}_{\mathrm{4}}$	${\mathrm{Cl}}_{\mathrm{2}}$	Residence time	Power	Removal efficiency
(no.)	($\mathrm{ppm}$)	($\mathrm{ppm}$)	($\mathrm{s}$)	($\mathrm{W}$)	($\mathrm{\%}$)
Exp. E
1	$\mathrm{3.3805}\pm \mathrm{6}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{30}\pm \mathrm{3}$	$\mathrm{62.7}\pm \mathrm{1.6}$	$\mathrm{13.39}\pm \mathrm{0.03}$	$\mathrm{22}\pm \mathrm{12}$
2	$\mathrm{3.3984}\pm \mathrm{2}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{20}\pm \mathrm{2}$	$\mathrm{61.6}\pm \mathrm{1.7}$	$\mathrm{13.36}\pm \mathrm{0.03}$	$\mathrm{19.8}\pm \mathrm{1.9}$
3	$\mathrm{3.3947}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{20}\pm \mathrm{2}$	$\mathrm{61.4}\pm \mathrm{1.5}$	$\mathrm{9.89}\pm \mathrm{0.03}$	$\mathrm{16.7}\pm \mathrm{1.2}$
4	$\mathrm{3.4014}\pm \mathrm{9}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{20}\pm \mathrm{2}$	$\mathrm{61.4}\pm \mathrm{1.5}$	$\mathrm{17.50}\pm \mathrm{0.03}$	$\mathrm{23.1}\pm \mathrm{1.9}$
5	$\mathrm{3.3282}\pm \mathrm{5}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{40}\pm \mathrm{4}$	$\mathrm{61.0}\pm \mathrm{1.5}$	$\mathrm{13.43}\pm \mathrm{0.03}$	$\mathrm{35}\pm \mathrm{3}$
6	$\mathrm{3.3309}\pm \mathrm{5}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{40}\pm \mathrm{4}$	$\mathrm{61.1}\pm \mathrm{1.6}$	$\mathrm{9.92}\pm \mathrm{0.03}$	$\mathrm{29.3}\pm \mathrm{1.6}$
7	$\mathrm{3.3312}\pm \mathrm{4}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{40}\pm \mathrm{4}$	$\mathrm{61.0}\pm \mathrm{1.6}$	$\mathrm{17.47}\pm \mathrm{0.03}$	$\mathrm{39}\pm \mathrm{6}$
8	$\mathrm{3.4096}\pm \mathrm{6}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{60.7}\pm \mathrm{1.5}$	$\mathrm{13.43}\pm \mathrm{0.03}$	$\mathrm{40}\pm \mathrm{5}$
9	$\mathrm{3.4444}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{60.6}\pm \mathrm{1.5}$	$\mathrm{9.90}\pm \mathrm{0.03}$	$\mathrm{35}\pm \mathrm{2}$
10	$\mathrm{3.4377}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{60.6}\pm \mathrm{1.5}$	$\mathrm{17.49}\pm \mathrm{0.03}$	$\mathrm{45}\pm \mathrm{4}$
11	$\mathrm{3.3575}\pm \mathrm{5}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{60}\pm \mathrm{6}$	$\mathrm{60.4}\pm \mathrm{1.5}$	$\mathrm{13.43}\pm \mathrm{0.03}$	$\mathrm{47}\pm \mathrm{2}$
12	$\mathrm{3.3800}\pm \mathrm{7}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{60}\pm \mathrm{6}$	$\mathrm{60.4}\pm \mathrm{1.5}$	$\mathrm{9.90}\pm \mathrm{0.03}$	$\mathrm{41.0}\pm \mathrm{1.9}$
13	$\mathrm{3.3604}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{60}\pm \mathrm{6}$	$\mathrm{60.3}\pm \mathrm{1.6}$	$\mathrm{17.49}\pm \mathrm{0.03}$	$\mathrm{54}\pm \mathrm{2}$
14	$\mathrm{3.4122}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{70}\pm \mathrm{7}$	$\mathrm{60.1}\pm \mathrm{1.6}$	$\mathrm{13.43}\pm \mathrm{0.03}$	$\mathrm{53}\pm \mathrm{3}$
15	$\mathrm{3.4414}\pm \mathrm{1.5}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{70}\pm \mathrm{7}$	$\mathrm{60.0}\pm \mathrm{1.6}$	$\mathrm{9.90}\pm \mathrm{0.03}$	$\mathrm{45}\pm \mathrm{3}$
16	$\mathrm{3.4566}\pm \mathrm{8}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{70}\pm \mathrm{7}$	$\mathrm{59.8}\pm \mathrm{1.6}$	$\mathrm{17.49}\pm \mathrm{0.03}$	$\mathrm{59}\pm \mathrm{2}$
Exp. F
1	$\mathrm{3.5176}\pm \mathrm{6}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{162}\pm \mathrm{3.4}$	$\mathrm{6.75}\pm \mathrm{0.02}$	$\mathrm{28.0}\pm \mathrm{0.3}$
2	$\mathrm{3.5475}\pm \mathrm{1.1}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{162}\pm \mathrm{3.4}$	$\mathrm{9.74}\pm \mathrm{0.02}$	$\mathrm{37.3}\pm \mathrm{0.3}$
3	$\mathrm{3.5668}\pm \mathrm{1.8}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{161}\pm \mathrm{3.4}$	$\mathrm{12.17}\pm \mathrm{0.03}$	$\mathrm{46.83}\pm \mathrm{0.1}$
4	$\mathrm{3.5920}\pm \mathrm{1.0}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{161}\pm \mathrm{3.4}$	$\mathrm{14.63}\pm \mathrm{0.03}$	$\mathrm{53.77}\pm \mathrm{0.1}$
5	$\mathrm{3.6162}\pm \mathrm{1.9}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{161}\pm \mathrm{3.3}$	$\mathrm{17.18}\pm \mathrm{0.03}$	$\mathrm{55.2}\pm \mathrm{0.2}$
6	$\mathrm{3.6425}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{160}\pm \mathrm{3.4}$	$\mathrm{19.73}\pm \mathrm{0.03}$	$\mathrm{58.0}\pm \mathrm{0.3}$
7	$\mathrm{3.6592}\pm \mathrm{1.2}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{160}\pm \mathrm{3.4}$	$\mathrm{22.32}\pm \mathrm{0.03}$	$\mathrm{59.2}\pm \mathrm{0.3}$

Table D3

Data from exp. G–I. Columns: experimental steps, [${\mathrm{CH}}_{\mathrm{4}}$] (ppm), [${\mathrm{Cl}}_{\mathrm{2}}$] (ppm), residence time in seconds, power in watts and the resulting removal efficiency (%).

Experiment	${\mathrm{CH}}_{\mathrm{4}}$	${\mathrm{Cl}}_{\mathrm{2}}$	Residence time	Power	Removal efficiency
(no.)	($\mathrm{ppm}$)	($\mathrm{ppm}$)	($\mathrm{s}$)	($\mathrm{W}$)	($\mathrm{\%}$)
Exp. G
1	$\mathrm{3.5594}\pm \mathrm{1.7}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{167}\pm \mathrm{3.5}$	$\mathrm{14.46}\pm \mathrm{0.03}$	$\mathrm{47.4}\pm \mathrm{1.2}$
2	$\mathrm{3.2339}\pm \mathrm{1.3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{168}\pm \mathrm{3.5}$	$\mathrm{14.49}\pm \mathrm{0.03}$	$\mathrm{54.7}\pm \mathrm{0.5}$
3	$\mathrm{2.9339}\pm \mathrm{9}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{168}\pm \mathrm{3.5}$	$\mathrm{14.56}\pm \mathrm{0.03}$	$\mathrm{60.8}\pm \mathrm{0.5}$
4	$\mathrm{2.684}\pm \mathrm{4}\times {\mathrm{10}}^{-\mathrm{2}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{167}\pm \mathrm{3.6}$	$\mathrm{14.60}\pm \mathrm{0.03}$	$\mathrm{66.2}\pm \mathrm{0.6}$
5	$\mathrm{2.2942}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{164}\pm \mathrm{3.4}$	$\mathrm{14.63}\pm \mathrm{0.03}$	$\mathrm{69.6}\pm \mathrm{0.3}$
6	$\mathrm{1.9817}\pm \mathrm{6}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{164}\pm \mathrm{3.5}$	$\mathrm{14.46}\pm \mathrm{0.03}$	$\mathrm{72.0}\pm \mathrm{0.4}$
7	$\mathrm{1.6982}\pm \mathrm{7}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{166}\pm \mathrm{3.4}$	$\mathrm{14.46}\pm \mathrm{0.03}$	$\mathrm{74.1}\pm \mathrm{0.6}$
8	$\mathrm{1.3899}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{163}\pm \mathrm{3.5}$	$\mathrm{14.46}\pm \mathrm{0.03}$	$\mathrm{77.2}\pm \mathrm{0.7}$
9	$\mathrm{3.8333}\pm \mathrm{7}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{162}\pm \mathrm{3.4}$	$\mathrm{14.70}\pm \mathrm{0.03}$	$\mathrm{60.3}\pm \mathrm{0.4}$
10	$\mathrm{4.1285}\pm \mathrm{1.9}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{161}\pm \mathrm{3.4}$	$\mathrm{14.63}\pm \mathrm{0.03}$	$\mathrm{60.2}\pm \mathrm{0.2}$
11	$\mathrm{3.5053}\pm \mathrm{1.7}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{161}\pm \mathrm{3.4}$	$\mathrm{14.63}\pm \mathrm{0.03}$	$\mathrm{64.0}\pm \mathrm{0.2}$
12	$\mathrm{3.2045}\pm \mathrm{9}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{161}\pm \mathrm{3.4}$	$\mathrm{14.63}\pm \mathrm{0.03}$	$\mathrm{66.1}\pm \mathrm{0.3}$
Exp. H
1	$\mathrm{1.9857}\pm \mathrm{8}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{164}\pm \mathrm{3.4}$	$\mathrm{14.77}\pm \mathrm{0.03}$	$\mathrm{68}\pm \mathrm{3}$
2	$\mathrm{1.9872}\pm \mathrm{1.0}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{261}\pm \mathrm{5.9}$	$\mathrm{14.77}\pm \mathrm{0.03}$	$\mathrm{88.1}\pm \mathrm{1.3}$
3	$\mathrm{1.9955}\pm \mathrm{1.0}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{348}\pm \mathrm{8.3}$	$\mathrm{14.77}\pm \mathrm{0.03}$	$\mathrm{92}\pm \mathrm{5}$
4	$\mathrm{1.9995}\pm \mathrm{8}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{357}\pm \mathrm{8.9}$	$\mathrm{17.36}\pm \mathrm{0.03}$	$\mathrm{94}\pm \mathrm{5}$
5	$\mathrm{2.0099}\pm \mathrm{8}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{342}\pm \mathrm{8.3}$	$\mathrm{19.94}\pm \mathrm{0.03}$	$\mathrm{96}\pm \mathrm{4}$
6${}^{*}$	$\mathrm{2.0021}\pm \mathrm{2}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{342}\pm \mathrm{8.2}$	$\mathrm{22.80}\pm \mathrm{0.03}$	$\mathrm{98.99}\pm \mathrm{0.1}$
7${}^{*}$	$\mathrm{2.0046}\pm \mathrm{3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{265}\pm \mathrm{25}$	$\mathrm{22.80}\pm \mathrm{0.03}$	$\mathrm{96.7}\pm \mathrm{0.3}$
8${}^{*}$	$\mathrm{2.0061}\pm \mathrm{4}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{173}\pm \mathrm{20}$	$\mathrm{22.80}\pm \mathrm{0.03}$	$\mathrm{87.33}\pm \mathrm{0.1}$
9${}^{*}$	$\mathrm{2.0076}\pm \mathrm{6}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{128}\pm \mathrm{10}$	$\mathrm{22.80}\pm \mathrm{0.03}$	$\mathrm{77.30}\pm \mathrm{0.1}$
Exp. I
1	$\mathrm{2.0471}\pm \mathrm{7}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{164}\pm \mathrm{3.4}$	$\mathrm{7.92}\pm \mathrm{0.03}$	$\mathrm{46.1}\pm \mathrm{1.8}$
2	$\mathrm{2.0565}\pm \mathrm{9}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{164}\pm \mathrm{3.5}$	$\mathrm{10.13}\pm \mathrm{0.03}$	$\mathrm{56.6}\pm \mathrm{0.2}$
3	$\mathrm{2.0586}\pm \mathrm{1.0}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{163}\pm \mathrm{3.5}$	$\mathrm{12.54}\pm \mathrm{0.03}$	$\mathrm{64.29}\pm \mathrm{0.1}$
4	$\mathrm{2.0606}\pm \mathrm{1.1}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{163}\pm \mathrm{3.6}$	$\mathrm{15.14}\pm \mathrm{0.03}$	$\mathrm{70.31}\pm \mathrm{0.1}$
5	$\mathrm{2.0627}\pm \mathrm{1.1}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{164}\pm \mathrm{3.6}$	$\mathrm{17.71}\pm \mathrm{0.03}$	$\mathrm{75.09}\pm \mathrm{0.1}$
6	$\mathrm{2.0690}\pm \mathrm{1.4}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{162}\pm \mathrm{3.6}$	$\mathrm{20.63}\pm \mathrm{0.03}$	$\mathrm{80.28}\pm \mathrm{0.07}$
7	$\mathrm{2.0710}\pm \mathrm{1.5}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{161}\pm \mathrm{3.5}$	$\mathrm{23.63}\pm \mathrm{0.03}$	$\mathrm{83.25}\pm \mathrm{0.04}$

${}^{*}$ The [${\mathrm{CH}}_{\mathrm{4}}$] values are calculated based on trend fitting.

Figure D1

Exp. A–I are shown. Each illuminated step has been highlighted. (a) The ${\mathrm{CH}}_{\mathrm{4}}$ is seen as a function of time. The [${\mathrm{Cl}}_{\mathrm{2}}$] is varied. (b) The light intensity and [${\mathrm{Cl}}_{\mathrm{2}}$] are varied. (c) Steps 7–10 highlighted. The light intensity and $t_{\mathrm{R}}$ are varied. (d) Steps 1–10 are highlighted. The light intensity and $t_{\mathrm{R}}$ are varied. (e) Steps 1–16 are highlighted. The light intensity and [${\mathrm{Cl}}_{\mathrm{2}}$] are varied. Following the initial illumination at 13 W the sample was illuminated at three different $p_{\mathrm{IN}}$ for five different chlorine concentrations. The $p_{\mathrm{IN}}$ was of the order 13, 10 and 17 W with chlorine steps 20, 40, 50, 60 and 70 ppm. (f) Steps 1–7 are highlighted. The light intensity is varied. (g) Steps 1–12 are highlighted. The ${\mathrm{CH}}_{\mathrm{4}}$ level is varied, while the light intensity is kept the same. (h) Steps 1–9 are highlighted. The light intensity and $t_{\mathrm{R}}$ are varied. (i) Steps 1–7 are highlighted. The light intensity is varied. Prolonged and stable photolysis was enabled due to cooling. Increasing levels of $p_{\mathrm{IN}}$ for the photochemical chamber define the seven different steps.

[Figure omitted. See PDF]

Figure D2

(a) RE% as a function of $p_{\mathrm{IN}}$ for exp. B1–B4. (b) Experiment steps D2 and D6–D10: RE% as a function of $p_{\mathrm{IN}}$ ($\mathrm{W}$). (c) RE% of methane plotted against $p_{\mathrm{IN}}$ (W). The results from three experiments, D (square), F (circle) and I (arrow), have different settings in $t_{\mathrm{R}}$, [${\mathrm{CH}}_{\mathrm{4}}$] and [${\mathrm{Cl}}_{\mathrm{2}}$]. (d) Experimental steps D1–D5: $t_{\mathrm{R}}$ ($\mathrm{s}$) in the photochemical device as a function of $t_{\mathrm{R}}$ in seconds. (e) The resulting removal efficiencies of exp. H plotted against $t_{\mathrm{R}}$. An additional zoom inset on the four points around 350 s reveals the removal effect plotted against power. (f) RE% of methane plotted against $t_{\mathrm{R}}$ in seconds. The results from three experiments, D (black), C (green) and H (purple), have different settings in $p_{\mathrm{IN}}$, [${\mathrm{CH}}_{\mathrm{4}}$] and [${\mathrm{Cl}}_{\mathrm{2}}$]. (g) RE% as a function of [${\mathrm{Cl}}_{\mathrm{2}}$] (ppm) for the xenon lamp in exp. A. (h) Resulting RE% plotted against [${\mathrm{Cl}}_{\mathrm{2}}$] (ppm) for exp. steps E2–E16. Three different power settings are used: 9.9 W (diamond), 13.4 W (circle) and 17.5 W (square). (i) The RE% is displayed as a function of the initial methane concentration with the remaining fixed parameters such as ${\mathrm{Cl}}_{\mathrm{2}}$ mixing ratio, $t_{\mathrm{R}}$ and $p_{\mathrm{IN}}$. The three points (star) in the figure represent steps suffering from early experimental deviation.

[Figure omitted. See PDF]

Table D4

Table summarizing experiments and setups. FC: flow-controlled, PC: pressure-controlled, CWL: chlorine waste line.

Setup	Description	Experiment
1	High-pressure xenon lamp with FC CWL	A
2	Single-tube hexagonal photochemical device with FC CWL	B
3	Single-tube hexagonal photochemical device with PC CWL	C, D, E
4	Multiple-tube hexagonal photochemical device with PC CWL	F, G, H, I
5	Multiple-tube hexagonal photochemical device with PC CWL (${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$)	J, K
6	Multiple-tube hexagonal photochemical device with PC CWL and sofnocat (${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$)	L

D1.1 Setup 1 (HPXL) experiments

The xenon lamp experiments shown in Fig. a were performed to confirm that the ${\mathrm{Cl}}_{\mathrm{2}}$ added to the gas mix could make it to the photolysis chamber. The RE% of methane was found as a result of varying the [${\mathrm{Cl}}_{\mathrm{2}}$] to 16.7, 25, 50 and 92 ppm as seen in Fig. g. Each concentration step was given 10 min to stabilize before the xenon lamp was turned on for 10 min. The gas provided to the system was a dynamic mix of flows from three different flasks (see Table ). Due to this, it was possible to vary the abundance of chlorine while keeping [${\mathrm{CH}}_{\mathrm{4}}$] constant. The experiment confirmed that the level of ${\mathrm{Cl}}_{\mathrm{2}}$ could be controlled and that higher levels resulted in greater depletion of methane.

D1.2 Setup 2 (single-tube, flow-controlled chlorine waste line) experiments

In exp. B, $P_{\mathrm{IN}}$ was varied in steps 1 to 4, as presented in Fig. b. The aim was to determine the effect of varying light intensity. Figure a shows the RE% as a function of $p_{\mathrm{IN}}$ for experiments 1 to 4. The initial methane concentration is maintained at $\mathrm{3.68}\pm \mathrm{0.02}$ ppm. Steps 1 and 2 are both examples of the start-up deviation. At the time of steps 3 and 4, sufficient flushing had taken place.

Table D5

Exp. B. The three experimental steps clearly show an increasing RE% as the $p_{\mathrm{IN}}$ and the ${\mathrm{Cl}}_{\mathrm{2}}$ mixing ratio are increased.

Step	${\mathrm{CH}}_{\mathrm{4}}$	${\mathrm{Cl}}_{\mathrm{2}}$	Residence time	Power	RE%
(no.)	($\mathrm{ppm}$)	($\mathrm{ppm}$)	($\mathrm{s}$)	($\mathrm{W}$)	($\mathrm{\%}$)
B3	$\mathrm{3.7069}\pm \mathrm{1.1}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{16.7}\pm \mathrm{1.5}$	$\mathrm{62.2}\pm \mathrm{1.5}$	$\mathrm{9.91}\pm \mathrm{0.03}$	$\mathrm{6.87}\pm \mathrm{0.01}$
B5	$\mathrm{3.919}\pm \mathrm{1.4}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{61.0}\pm \mathrm{1.4}$	$\mathrm{9.92}\pm \mathrm{0.03}$	$\mathrm{23.91}\pm \mathrm{0.05}$
B6	$\mathrm{3.945}\pm \mathrm{1.3}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{61.4}\pm \mathrm{1.4}$	$\mathrm{16.59}\pm \mathrm{0.03}$	$\mathrm{32.69}\pm \mathrm{0.04}$

The chlorine concentration was increased from 16.7 to 50 ppm starting with step 5. The four relevant variables and resulting RE% can be seen in Table . [${\mathrm{Cl}}_{\mathrm{2}}$] was increased by a factor of 2.5 between steps 3 and 5. The increase results in a 3.5-fold increase in RE%. Furthermore, the $p_{\mathrm{IN}}$ is increased when going from step 5 to 6, which also leads to an increase in RE%. In a comparison between these three steps, the positive relation for both chlorine concentration and $p_{\mathrm{IN}}$ to the RE% was confirmed.

D1.3 Setup 3 (single-tube, pressure-controlled chlorine waste line) experiments

Three experiments (C, D and E) used this setup. Exp. C presented in Fig. c was carried out with a constant supply of [${\mathrm{Cl}}_{\mathrm{2}}$] at 50 ppm and [${\mathrm{CH}}_{\mathrm{4}}$] at $\mathrm{3.981}\pm \mathrm{0.018}$ ppm. Steps 2 and 3 had the same $t_R$ as steps 1 and 4. In addition, the experiments vary in $p_{\mathrm{IN}}$, as can be seen in Table . Table  shows how the combination of increased $t_{\mathrm{R}}$ and $p_{\mathrm{IN}}$ yields a higher RE%.

Table D6

Exp. C.

Experiment	${\mathrm{CH}}_{\mathrm{4}}$	${\mathrm{Cl}}_{\mathrm{2}}$	Residence time	Power	RE%
(no.)	($\mathrm{ppm}$)	($\mathrm{ppm}$)	($\mathrm{s}$)	($\mathrm{W}$)	($\mathrm{\%}$)
C1	$\mathrm{3.957}\pm \mathrm{5}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{41.4}\pm \mathrm{1.0}$	$\mathrm{9.63}\pm \mathrm{0.03}$	$\mathrm{18.90}\pm \mathrm{0.11}$
C2	$\mathrm{4.0301}\pm \mathrm{5}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{41.4}\pm \mathrm{1.0}$	$\mathrm{17.30}\pm \mathrm{0.03}$	$\mathrm{27.83}\pm \mathrm{0.06}$
C3	$\mathrm{3.955}\pm \mathrm{2}\times {\mathrm{10}}^{-\mathrm{3}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{129}\pm \mathrm{3}$	$\mathrm{9.63}\pm \mathrm{0.03}$	$\mathrm{44.51}\pm \mathrm{0.3}$
C4	$\mathrm{3.986}\pm \mathrm{9}\times {\mathrm{10}}^{-\mathrm{4}}$	$\mathrm{50}\pm \mathrm{5}$	$\mathrm{128}\pm \mathrm{3}$	$\mathrm{17.30}\pm \mathrm{0.03}$	$\mathrm{60.6}\pm \mathrm{0.5}$

Exp. D was carried out with [${\mathrm{Cl}}_{\mathrm{2}}$] kept constant at 32 ppm. The initial methane concentration was maintained at $\mathrm{3.547}\pm \mathrm{0.005}$ ppm. Similarly to exp. C the $t_{\mathrm{R}}$ and $p_{\mathrm{IN}}$ were varied. Steps 1 to 5 are carried out with the same $p_{\mathrm{IN}}$ in the device but with varying residence times; see Fig. f and d. In Fig. f the data for exp. D exhibit clear agreement between $t_{\mathrm{R}}$ and RE%. The longer $t_{\mathrm{R}}$ within the photochamber results in greater removal efficiencies. Steps 2 and 6–10 are carried out with the same $t_{\mathrm{R}}$ but with varying $p_{\mathrm{IN}}$; see Fig. c and b.

The experimental steps of exp. E (Fig. e) were held at the same initial methane concentration of $\mathrm{3.39}\pm \mathrm{0.01}$ ppm and the same $t_{\mathrm{R}}$ of $\mathrm{60.82}\pm \mathrm{0.18}$ s. Throughout the experiments, three levels of $p_{\mathrm{IN}}$ were tested against varied levels of ${\mathrm{Cl}}_{\mathrm{2}}$ mixing ratio spanning in the range 20–70 ppm. Figure h presents, looking at 20 ppm ${\mathrm{Cl}}_{\mathrm{2}}$, the fact that a greater $p_{\mathrm{IN}}$ yields higher RE%.

D1.4 Setup 4 (multiple-tube, pressure-controlled chlorine waste line) experiments

Four experiments (F, G, H and I) were done with this setup. Exp. F (Fig. f) was run at a constant level of [${\mathrm{CH}}_{\mathrm{4}}$]${}_{\mathrm{initial}}$ at $\mathrm{3.593}\pm \mathrm{0.019}$ ppm and [${\mathrm{Cl}}_{\mathrm{2}}$] at 50 $\pm \mathrm{5}$ ppm. At a flow kept at 15.5 mL min${}^{-\mathrm{1}}$ the $t_{\mathrm{R}}$ in the photochamber was maintained at $\mathrm{161.06}\pm \mathrm{3}$ s. Across exp. F stepwise changes were made for $p_{\mathrm{IN}}$ ranging from 6.75–22.92 W. The daily measurement is presented in Fig. f, where the removal for the steps, with the exception of the first step, is characterized by an initial RE%, but this efficiency drops during the first 5 min of illumination. The relationship found between removal and $p_{\mathrm{IN}}$ for exp. F can be seen in Fig. c.

Exp. G was carried out with a stepwise change in [${\mathrm{CH}}_{\mathrm{4}}$]${}_{\mathrm{inititial}}$ in the range 1.39 to 4.13 ppm at constant $t_{\mathrm{R}}$ of 164 s, [${\mathrm{Cl}}_{\mathrm{2}}$] of 50 ppm and $p_{\mathrm{IN}}$ of 14.6 W. The daily result can be seen in Fig. g, where the improvement of silicone removal can be observed from stable levels of RE%. As can be seen in Fig. i decreasing the initial methane concentration yields, as expected, a greater RE%.

Exp. H was carried out with the constant [${\mathrm{CH}}_{\mathrm{4}}$]${}_{\mathrm{inititial}}$ at $\mathrm{2.000}\pm \mathrm{0.003}$ ppm and ${\mathrm{Cl}}_{\mathrm{2}}$ mixing ratio at $\mathrm{50}\pm \mathrm{5}$ ppm, but with mixed settings of $t_{\mathrm{R}}$ and power. Steps H1–H3 were done with constant power at 14.8 W with $t_{\mathrm{R}}$ increasing from 164–350 s. Then, keeping $t_{\mathrm{R}}$ around 350 s, three steps of increasing power were tested ranging from 14.8–22.8 W. Between steps H4 and H5 a fan was installed. The final three steps were kept at 22.8 W and stepped through reduced $t_{\mathrm{R}}$ from 342–130 s.

Exp. I was carried out with [${\mathrm{CH}}_{\mathrm{4}}$]${}_{\mathrm{inititial}}$ maintained around $\mathrm{2.01}\pm \mathrm{0.01}$ ppm, [${\mathrm{Cl}}_{\mathrm{2}}$] at 50 ppm and the $t_{\mathrm{R}}$ held at $\mathrm{163.1}\pm \mathrm{0.4}$ s. The only parameter varied was the $p_{\mathrm{IN}}$ to the photochemical device. The light was turned on at 7.9 W and was left on for the duration of the experiments with a stepwise increase in $P_{\mathrm{IN}}$ after stable removal had been maintained for 5 min. The resulting methane concentration can be seen in Fig. i. [${\mathrm{CH}}_{\mathrm{4}}$] increases throughout the experiments due to the chlorine pressure decline. For the purpose of calculating RE%, the expected [${\mathrm{CH}}_{\mathrm{4}}$] for each of the steps was fitted from the initial [${\mathrm{CH}}_{\mathrm{4}}$] and the end [${\mathrm{CH}}_{\mathrm{4}}$]; ${\mathrm{CH}}_{\mathrm{4}}=\mathrm{0.0002}\cdot t+\mathrm{2.0461}$. The relative median values of initial methane and $t_{\mathrm{R}}$ were chosen in order to best resolve the effects of varying $p_{\mathrm{IN}}$. As the removal effect approaches 100 % asymptotically, the sensitivity to changes will be greater at lower removal values.

The results presented for exp. I in Fig. c can be compared to the results from exp. D and F and represent the improvements implemented to the system. Unlike for those experiments, the trend of exp. I is explained by one trend asymptotically approaching 100 % removal.

D1.5 Comparison

Figure c shows a comparison of three different experiments in which $p_{\mathrm{IN}}$ was varied. When comparing experiments F and I the improvement in performance of the device is clear. However, even if $t_{\mathrm{R}}$ and [${\mathrm{Cl}}_{\mathrm{2}}$] are identical, the initial methane concentration of exp. F is 3.59 ppm compared to exp. I at 2.096 ppm. Exp. D alone shares some $P_{\mathrm{IN}}$ levels and is operated at the same initial methane level as exp. F. The $t_{\mathrm{R}}$ and $\left[{\mathrm{Cl}}_{\mathrm{2}}\right]$ are lower and less removal is accordingly expected. Hence, the main thing to observe is behavior at higher $P_{\mathrm{IN}}$. The efficiency of the photochamber decreases as seen in exp. D and F. The improvements done on the photochamber and installation of a fan to cool the photochemical chamber have prolonged the lifetime of the chamber and improved efficiency.

Figure f shows a comparison of three different experiments in which $t_{\mathrm{R}}$ was varied and in some cases $P_{\mathrm{IN}}$ as well. $t_{\mathrm{R}}$ is improved because the MTH-PD setup made it possible to obtain higher $t_{\mathrm{R}}$ and more efficient use of the photochemical chamber. The experiments with a single tube do not have long residence times. As seen in Fig. f longer $t_{\mathrm{R}}$ greatly improves the RE% and is therefore essential to further improve the setup.

D2

${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ experimental results

From the experiment investigating the compatibility of the removal method and the analysis of ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$, it was found that the oxidization had no effect on the ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ abundance or the isotopic composition. It was, however, discovered that the oxidation path for ${\mathrm{CH}}_{\mathrm{4}}$ terminated at $\mathrm{CO}$, as the isotopic signal changed, matching the interference of $\mathrm{CO}$. To remove this effect a sofnocat trap was implemented, which oxidizes the $\mathrm{CO}$ to ${\mathrm{CO}}_{\mathrm{2}}$. By applying the trace gas and matrix corrections described in , it was found that the isotopic levels remained stable across the oxidation. Variation observed in the ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ was due to the unstable supply of ${\mathrm{Cl}}_{\mathrm{2}}$, resulting in slight shifts in the dilution. The values of ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\alpha }}$ and ${\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$ were both found to approach the unaffected target value during the oxidation as was hoped. Results are shown in Fig. 

Figure D3

Results from the three experiments J, K and L using the G5131-i for ${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$ isotope measurements. The ${\mathrm{CH}}_{\mathrm{4}}$ level is depicted in each row (ppm) along the first $y$ axis. Highlights indicate the several different oxidation settings. Row 1: measurements of ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\alpha }}$ (‰) plotted along the second $y$ axis. Red highlights a 100 s average measured value corrected for ${\mathrm{O}}_{\mathrm{2}}$, $\mathrm{CO}$ and ${\mathrm{CO}}_{\mathrm{2}}$ effects, while grey indicates a 100 $\mathrm{s}$ average value that has been corrected for all interference including ${\mathrm{CH}}_{\mathrm{4}}$. Row 2: measurements of ${\mathit{\delta }}^{\mathrm{15}}{\mathrm{N}}^{\mathit{\beta }}$ (‰) plotted along the second $y$ axis. Red highlights a 100 s average measured value corrected for ${\mathrm{O}}_{\mathrm{2}}$, $\mathrm{CO}$ and ${\mathrm{CO}}_{\mathrm{2}}$ effects, while grey indicates a 100 s average value that has been corrected for all interference including ${\mathrm{CH}}_{\mathrm{4}}$. Row 3: measurements of ${\mathit{\delta }}^{\mathrm{18}}\mathrm{O}$ (‰) plotted along the second $y$ axis. Red highlights a 100 s average measured value corrected for ${\mathrm{O}}_{\mathrm{2}}$, $\mathrm{CO}$ and ${\mathrm{CO}}_{\mathrm{2}}$ effects, while grey indicates a 100 s average value that has been corrected for all trace gas interference including ${\mathrm{CH}}_{\mathrm{4}}$. Row 4: measurements of [${\mathrm{N}}_{\mathrm{2}}\mathrm{O}$] ($\mathrm{ppb}$) shown in blue. Variation observed corresponds to fluctuations in the mixing of the three gases. Exp. J: in this experiment the light was turned on throughout the entire experiment, with the experimental steps corresponding to changes in $t_{\mathrm{R}}$. Exp. K: in this experiment two experimental steps were used with different power settings. Exp. L: in this experiment the sofnocat trap was used in the first three experimental steps, while steps 4 and 5 were completed without. The variation between experimental steps corresponds to changes in $t_{\mathrm{R}}$.

[Figure omitted. See PDF]

Appendix E Kintecus reactions and results

The results from the kinetic model are shown in Fig. .

Figure E1

RE% as found experimentally (grey) and by the model (white stripes). (a) Exp. B and C, (b) exp. D, (c) exp. E, (d) exp. F, (e) exp. G, (f) exp. H.

[Figure omitted. See PDF]

Table E1

JPL: . Hossaini: .

Reaction	Reaction rate coefficient (cm${}^{\mathrm{3}}$ molec.${}^{-\mathrm{1}}$ s${}^{-\mathrm{1}}$)	Reference
${\mathrm{Cl}}_{\mathrm{2}}\to \mathrm{2}\mathrm{Cl}$	X
$\mathrm{Cl}+\mathrm{Cl}+M\to {\mathrm{Cl}}_{\mathrm{2}}+M$	$\mathrm{1.29}\times {\mathrm{10}}^{-\mathrm{32}}{}^{*}$
${\mathrm{O}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{3}}\to {\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{1.79}\times {\mathrm{10}}^{-\mathrm{12}}$
${\mathrm{O}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{O}\to {\mathrm{CH}}_{\mathrm{2}}\mathrm{O}+{\mathrm{HO}}_{\mathrm{2}}$	$\mathrm{1.65}\times {\mathrm{10}}^{-\mathrm{15}}$
${\mathrm{O}}_{\mathrm{2}}+\mathrm{HCO}\to \mathrm{CO}+{\mathrm{HO}}_{\mathrm{2}}$	$\mathrm{5.20}\times {\mathrm{10}}^{-\mathrm{12}}$
${\mathrm{O}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{2}}\mathrm{Cl}\to {\mathrm{CH}}_{\mathrm{2}}{\mathrm{ClO}}_{\mathrm{2}}$	$\mathrm{2.91}\times {\mathrm{10}}^{-\mathrm{12}}$	JPL
$\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{O}\to {\mathrm{CH}}_{\mathrm{2}}\mathrm{O}+\mathrm{HCl}$	$\mathrm{1.91}\times {\mathrm{10}}^{-\mathrm{11}}$
$\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{OH}\to {\mathrm{CH}}_{\mathrm{3}}\mathrm{O}+\mathrm{HCl}$	$\mathrm{5.50}\times {\mathrm{10}}^{-\mathrm{11}}$	JPL
$\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{2}}\mathrm{O}\to \mathrm{HCO}+\mathrm{HCl}$	$\mathrm{7.32}\times {\mathrm{10}}^{-\mathrm{11}}$	JPL
$\mathrm{Cl}+{\mathrm{Cl}}_{\mathrm{2}}\mathrm{O}\to {\mathrm{Cl}}_{\mathrm{2}}+\mathrm{ClO}$	$\mathrm{9.60}\times {\mathrm{10}}^{-\mathrm{11}}$	JPL
$\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{Cl}\to {\mathrm{CH}}_{\mathrm{2}}\mathrm{Cl}+\mathrm{HCl}$	$\mathrm{4.98}\times {\mathrm{10}}^{-\mathrm{13}}$	JPL
$\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{2}}{\mathrm{Cl}}_{\mathrm{2}}\to {\mathrm{CHCl}}_{\mathrm{2}}+\mathrm{HCl}$	$\mathrm{3.57}\times {\mathrm{10}}^{-\mathrm{13}}$	JPL
$\mathrm{Cl}+{\mathrm{CCl}}_{\mathrm{3}}\to {\mathrm{CCl}}_{\mathrm{4}}$	$\mathrm{6.51}\times {\mathrm{10}}^{-\mathrm{11}}$
$\mathrm{Cl}+{\mathrm{CHCl}}_{\mathrm{3}}\to \mathrm{HCl}+{\mathrm{CCl}}_{\mathrm{3}}$	$\mathrm{1.20}\times {\mathrm{10}}^{-\mathrm{13}}$	JPL
$\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}\to {\mathrm{CH}}_{\mathrm{3}}\mathrm{O}+ClO$	$\mathrm{1.60}\times {\mathrm{10}}^{-\mathrm{10}}$	JPL
$\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}\to {\mathrm{CH}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{2}}+\mathrm{HCl}$	$\mathrm{1.60}\times {\mathrm{10}}^{-\mathrm{10}}$	JPL
$\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{4}}\to {\mathrm{CH}}_{\mathrm{3}}+\mathrm{HCl}$	$\mathrm{1.07}\times {\mathrm{10}}^{-\mathrm{13}}$
$\mathrm{Cl}+\mathrm{CHClO}\to \mathrm{HCl}+\mathrm{Cl}+\mathrm{CO}$	$\mathrm{7.79}\times {\mathrm{10}}^{-\mathrm{13}}$
$\mathrm{Cl}+{\mathrm{H}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{2}}\to \mathrm{HCl}+{\mathrm{HO}}_{\mathrm{2}}$	$\mathrm{4.10}\times {\mathrm{10}}^{-\mathrm{13}}$	JPL
$\mathrm{Cl}+{\mathrm{CH}}_{\mathrm{3}}\to {\mathrm{CH}}_{\mathrm{3}}\mathrm{Cl}$	$\mathrm{1.61}\times {\mathrm{10}}^{-\mathrm{12}}$
${\mathrm{Cl}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{2}}\mathrm{Cl}\to {\mathrm{CH}}_{\mathrm{2}}{\mathrm{Cl}}_{\mathrm{2}}+\mathrm{Cl}$	$\mathrm{2.54}\times {\mathrm{10}}^{-\mathrm{13}}$
${\mathrm{Cl}}_{\mathrm{2}}+{\mathrm{CHCl}}_{\mathrm{2}}\to {\mathrm{CHCl}}_{\mathrm{3}}+\mathrm{Cl}$	$\mathrm{2.25}\times {\mathrm{10}}^{-\mathrm{14}}$
${\mathrm{Cl}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{3}}\to {\mathrm{CH}}_{\mathrm{3}}\mathrm{Cl}+\mathrm{Cl}$	$\mathrm{1.55}\times {\mathrm{10}}^{-\mathrm{12}}$
${\mathrm{Cl}}_{\mathrm{2}}+\mathrm{HCO}\to \mathrm{CHClO}+\mathrm{Cl}$	$\mathrm{5.59}\times {\mathrm{10}}^{-\mathrm{12}}$
${\mathrm{Cl}}_{\mathrm{2}}+\mathrm{OH}\to \mathrm{HClO}+\mathrm{Cl}$	$\mathrm{6.42}\times {\mathrm{10}}^{-\mathrm{14}}$

${}^{*}$ Third-order rate expression (units cm${}^{\mathrm{6}}$ molec.${}^{-\mathrm{2}}$ s${}^{-\mathrm{1}}$).

Table E2

JPL: . Hossaini: .

Reaction	Reaction rate coefficient (cm${}^{\mathrm{3}}$ molec.${}^{-\mathrm{1}}$ s${}^{-\mathrm{1}}$)	Reference
$\mathrm{OH}+{\mathrm{CH}}_{\mathrm{4}}\to {\mathrm{CH}}_{\mathrm{3}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}$	$\mathrm{6.30}\times {\mathrm{10}}^{-\mathrm{15}}$
$\mathrm{OH}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{OOH}\to {\mathrm{H}}_{\mathrm{2}}\mathrm{O}+{\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{7.40}\times {\mathrm{10}}^{-\mathrm{12}}$	JPL
$\mathrm{OH}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{OOH}\to {\mathrm{CH}}_{\mathrm{2}}\mathrm{O}+\mathrm{OH}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}$	$\mathrm{7.40}\times {\mathrm{10}}^{-\mathrm{12}}$	JPL
$\mathrm{OH}+{\mathrm{CH}}_{\mathrm{2}}\mathrm{O}\to \mathrm{HCO}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}$	$\mathrm{8.50}\times {\mathrm{10}}^{-\mathrm{12}}$	JPL
$\mathrm{OH}+\mathrm{HCl}\to \mathrm{Cl}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}$	$\mathrm{7.80}\times {\mathrm{10}}^{-\mathrm{13}}$	JPL
$\mathrm{OH}+\mathrm{HClO}\to \mathrm{ClO}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}$	$\mathrm{5.00}\times {\mathrm{10}}^{-\mathrm{13}}$
$\mathrm{OH}+{\mathrm{CH}}_{\mathrm{2}}{\mathrm{Cl}}_{\mathrm{2}}\to {\mathrm{CHCl}}_{\mathrm{2}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}$	$\mathrm{1.00}\times {\mathrm{10}}^{-\mathrm{13}}$	JPL
$\mathrm{OH}+{\mathrm{CHCl}}_{\mathrm{3}}\to {\mathrm{CCl}}_{\mathrm{3}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}$	$\mathrm{1.00}\times {\mathrm{10}}^{-\mathrm{13}}$	JPL
$\mathrm{OH}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{O}\to {\mathrm{CH}}_{\mathrm{2}}\mathrm{O}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}$	$\mathrm{3.01}\times {\mathrm{10}}^{-\mathrm{11}}$	JPL
$\mathrm{OH}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{OH}\to {\mathrm{CH}}_{\mathrm{3}}\mathrm{O}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}$	$\mathrm{1.40}\times {\mathrm{10}}^{-\mathrm{13}}$
$\mathrm{OH}+{\mathrm{CH}}_{\mathrm{3}}\to {\mathrm{CH}}_{\mathrm{3}}\mathrm{OH}$	$\mathrm{9.30}\times {\mathrm{10}}^{-\mathrm{11}}$
$\mathrm{OH}+{\mathrm{CH}}_{\mathrm{2}}\mathrm{ClOOH}\to {\mathrm{CH}}_{\mathrm{2}}{\mathrm{ClO}}_{\mathrm{2}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}$	$\mathrm{3.60}\times {\mathrm{10}}^{-\mathrm{12}}$	Hossaini
$\mathrm{OH}+{\mathrm{CH}}_{\mathrm{2}}\mathrm{ClOH}\to {\mathrm{CH}}_{\mathrm{3}}\mathrm{O}+\mathrm{HClO}$	$\mathrm{4.54}\times {\mathrm{10}}^{-\mathrm{14}}$	Hossaini
$\mathrm{OH}+{\mathrm{H}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{2}}\to {\mathrm{HO}}_{\mathrm{2}}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}$	$\mathrm{1.80}\times {\mathrm{10}}^{-\mathrm{12}}$	JPL
$\mathrm{OH}+\mathrm{CHClO}\to \mathrm{Cl}+\mathrm{CO}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}$	$\mathrm{3.20}\times {\mathrm{10}}^{-\mathrm{13}}$	Hossaini
$\mathrm{OH}+\mathrm{ClO}\to \mathrm{Cl}+{\mathrm{HO}}_{\mathrm{2}}$	$\mathrm{1.80}\times {\mathrm{10}}^{-\mathrm{11}}$	JPL
$\mathrm{OH}+\mathrm{ClO}\to \mathrm{HCl}+{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{1.30}\times {\mathrm{10}}^{-\mathrm{12}}$	JPL

${}^{*}$ Third-order rate expression (units cm${}^{\mathrm{6}}$ molec.${}^{-\mathrm{2}}$ s${}^{-\mathrm{1}}$).

Table E3

JPL: . Hossaini: .

Reaction	Reaction rate coefficient	Reference
	(cm${}^{\mathrm{3}}$ molec.${}^{-\mathrm{1}}$ s${}^{-\mathrm{1}}$)
${\mathrm{HO}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}\to {\mathrm{CH}}_{\mathrm{3}}\mathrm{OOH}+{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{5.12}\times {\mathrm{10}}^{-\mathrm{12}}$	JPL
${\mathrm{HO}}_{\mathrm{2}}+\mathrm{Cl}\to \mathrm{HCl}+{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{3.50}\times {\mathrm{10}}^{-\mathrm{11}}$	JPL
${\mathrm{HO}}_{\mathrm{2}}+\mathrm{Cl}\to \mathrm{ClO}+\mathrm{OH}$	$\mathrm{9.30}\times {\mathrm{10}}^{-\mathrm{12}}$	JPL
${\mathrm{HO}}_{\mathrm{2}}+\mathrm{ClO}\to \mathrm{HClO}+{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{6.90}\times {\mathrm{10}}^{-\mathrm{12}}$	JPL
${\mathrm{HO}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{O}\to {\mathrm{CH}}_{\mathrm{2}}\mathrm{O}+{\mathrm{H}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{5.00}\times {\mathrm{10}}^{-\mathrm{13}}$
${\mathrm{HO}}_{\mathrm{2}}+{\mathrm{HO}}_{\mathrm{2}}\to {\mathrm{H}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{2}}+{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{1.60}\times {\mathrm{10}}^{-\mathrm{12}}$
${\mathrm{HO}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{2}}{\mathrm{ClO}}_{\mathrm{2}}\to {\mathrm{CH}}_{\mathrm{2}}\mathrm{ClOOH}+{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{5.01}\times {\mathrm{10}}^{-\mathrm{12}}$	Hossaini
${\mathrm{HO}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{2}}{\mathrm{ClO}}_{\mathrm{2}}\to \mathrm{CHClO}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}+{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{5.01}\times {\mathrm{10}}^{-\mathrm{12}}$	Hossaini
$\mathrm{ClO}+\mathrm{ClO}\to {\mathrm{O}}_{\mathrm{2}}+{\mathrm{Cl}}_{\mathrm{2}}$	$\mathrm{4.91}\times {\mathrm{10}}^{-\mathrm{15}}$	JPL
$\mathrm{ClO}+\mathrm{ClO}\to \mathrm{2}\mathrm{Cl}+{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{8.00}\times {\mathrm{10}}^{-\mathrm{15}}$	JPL
$\mathrm{ClO}+\mathrm{Cl}+M\to {\mathrm{Cl}}_{\mathrm{2}}\mathrm{O}+M$	$\mathrm{1.56}\times {\mathrm{10}}^{-\mathrm{32}}{}^{*}$
$\mathrm{ClO}+{\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}\to \mathrm{Cl}+{\mathrm{O}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{O}$	$\mathrm{2.40}\times {\mathrm{10}}^{-\mathrm{12}}$	JPL
$\mathrm{ClO}+{\mathrm{CH}}_{\mathrm{3}}\to {\mathrm{CH}}_{\mathrm{3}}\mathrm{OCl}$	$\mathrm{5.69}\times {\mathrm{10}}^{-\mathrm{11}}$
${\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}\to {\mathrm{CH}}_{\mathrm{3}}\mathrm{O}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{O}+{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{3.50}\times {\mathrm{10}}^{-\mathrm{13}}$	JPL
${\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}\to {\mathrm{CH}}_{\mathrm{3}}\mathrm{OH}+{\mathrm{CH}}_{\mathrm{2}}\mathrm{O}+{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{3.50}\times {\mathrm{10}}^{-\mathrm{13}}$	JPL
${\mathrm{CH}}_{\mathrm{3}}+{\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}\to {\mathrm{CH}}_{\mathrm{3}}\mathrm{O}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{O}$	$\mathrm{4.50}\times {\mathrm{10}}^{-\mathrm{11}}$
${\mathrm{CH}}_{\mathrm{3}}\mathrm{O}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{O}\to {\mathrm{CH}}_{\mathrm{2}}\mathrm{O}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{OH}$	$\mathrm{3.85}\times {\mathrm{10}}^{-\mathrm{11}}$
${\mathrm{CH}}_{\mathrm{2}}{\mathrm{ClO}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}\to {\mathrm{CH}}_{\mathrm{2}}\mathrm{ClO}+{\mathrm{CH}}_{\mathrm{2}}\mathrm{O}+{\mathrm{HO}}_{\mathrm{2}}$	$\mathrm{2.50}\times {\mathrm{10}}^{-\mathrm{12}}$	Hossaini
${\mathrm{CH}}_{\mathrm{2}}{\mathrm{ClO}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}\to {\mathrm{CH}}_{\mathrm{2}}\mathrm{ClOH}+{\mathrm{CH}}_{\mathrm{2}}\mathrm{O}+{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{2.50}\times {\mathrm{10}}^{-\mathrm{12}}$	Hossaini
${\mathrm{CH}}_{\mathrm{2}}{\mathrm{ClO}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{3}}{\mathrm{O}}_{\mathrm{2}}\to \mathrm{CHClO}+{\mathrm{CH}}_{\mathrm{3}}\mathrm{OH}+{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{2.50}\times {\mathrm{10}}^{-\mathrm{12}}$	Hossaini
${\mathrm{CH}}_{\mathrm{2}}{\mathrm{ClO}}_{\mathrm{2}}+{\mathrm{CH}}_{\mathrm{2}}{\mathrm{ClO}}_{\mathrm{2}}\to {\mathrm{CH}}_{\mathrm{2}}\mathrm{ClO}+{\mathrm{CH}}_{\mathrm{2}}\mathrm{ClO}+{\mathrm{O}}_{\mathrm{2}}$	$\mathrm{3.50}\times {\mathrm{10}}^{-\mathrm{12}}$	Hossaini
${\mathrm{CH}}_{\mathrm{2}}{\mathrm{Cl}}_{\mathrm{2}}+\mathrm{Cl}\to {\mathrm{CHCl}}_{\mathrm{2}}+\mathrm{HCl}$	$\mathrm{3.57}\times {\mathrm{10}}^{-\mathrm{13}}$
${\mathrm{CHCl}}_{\mathrm{2}}+{\mathrm{Cl}}_{\mathrm{2}}\to {\mathrm{CHCl}}_{\mathrm{3}}+\mathrm{Cl}$	$\mathrm{2.25}\times {\mathrm{10}}^{-\mathrm{14}}$
${\mathrm{CCl}}_{\mathrm{3}}+\mathrm{Cl}\to {\mathrm{CCl}}_{\mathrm{4}}$	$\mathrm{6.51}\times {\mathrm{10}}^{-\mathrm{11}}$
$\mathrm{HCO}+{\mathrm{Cl}}_{\mathrm{2}}\to \mathrm{HC}\left(\mathrm{O}\right)\mathrm{Cl}+\mathrm{Cl}$	$\mathrm{5.59}\times {\mathrm{10}}^{-\mathrm{12}}$
${\mathrm{CH}}_{\mathrm{2}}{\mathrm{O}}_{\mathrm{2}}\to \mathrm{CO}+{\mathrm{H}}_{\mathrm{2}}\mathrm{O}$	$\mathrm{6.00}\times {\mathrm{10}}^{\mathrm{4}}{}^{*}$

${}^{*}$ Third-order rate expression (units cm${}^{\mathrm{6}}$ molec.${}^{-\mathrm{2}}$ s${}^{-\mathrm{1}}$).

Data availability

All data are available from the corresponding author upon request.

Author contributions

MSJ, MP and JBL conceived and planned the experiments. MP and JBL carried out the experiments. MP, JBL and MK planned and carried out the simulations. MP and JBL contributed to the interpretation of the results. MP and JBL wrote the paper in consultation with TB and MSJ. All authors provided critical feedback and helped shape the research, analysis and paper.

Competing interests

The authors declare that they have no conflict of interest.

Disclaimer

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Acknowledgements

We thank the Copenhagen Center for Atmospheric Research (CCAR), the Centre for Ice and Climate, and the University of Copenhagen. The authors thank Silvia Pugliese for her help with the reaction mechanism.

Review statement

This paper was edited by Huilin Chen and reviewed by two anonymous referees.