Soni et al. report an updated mechanism for the community atmospheric chemistry box model CAABA/MECCA, adding some 35 gas-phase reactions, 15 aqueous-phase reactions, and 4 reactions governing the partitioning of gases between gas and aerosol phase. The focus is on chemistry of chlorine compounds (Cl₂, ClNO₂, etc.). Reactive intermediates such as the nitronium ion (NO₂⁺) are treated explicitly. The revised model's was tested using two recent urban data sets collected in New Delhi (India) and Leicester (UK).

The paper is written well. I have a few concerns that the authors will hopefully be able to address in revision.

General comments.

(1) More explanation is needed to justify the chlorine nitrite chemistry in the mechanism, considering this molecule has not been unambiguously observed in ambient air. The authors are correct that chlorine nitrite may form from Cl+NO₂ (e.g., Golden, J. Phys. Chem. A 2007, 111(29), 6772–6780, https://doi.org/10.1021/jp069000x and Niki et al., Chem. Phys. Lett. 1978, 59(1), 78-79, https://doi.org/10.1016/0009-2614(78)85618-8) - these papers should be cited. However, its chemistry is incomplete. ClONO is metastable and converts to ClNO₂ (Janowski et al., Berichte der Bunsengesellschaft für physikalische Chemie 1977, 81(12), 1262-1270, https://doi.org/10.1002/bbpc.19770811212; Niki et al. Chem. Phys. Lett. 1978, 59(1), 78-79, https://doi.org/10.1016/0009-2614(78)85618-8) - a reaction that should be added to the mechanism. Note that Niki et al. report a ClONO lifetime of ~150 s.

Be it as it may, the reaction between Cl and NO₂ is generally thought to be negligible compared to reaction of N₂O₅ on chloride containing aerosol, except for unusual environments such as Delhi in winter. This paper thus seems to be tailored towards specific data sets, which should be mentioned in the introduction, and the relevant measurement papers (e.g., Haslett et al., Atmos. Chem. Phys. Disc., https://egusphere.copernicus.org/preprints/2023/egusphere-2023-497/) should be cited.

(2) More discussion as to the applicability of this model is needed.

For example, a limitation of this study is that all species are assumed to be well-mixed. In reality, there will be vertical gradients for most species evaluated here, in particular radical reservoir species such as HONO and ClNO₂ (e.g., Young et al., Environm. Sci. Technol. 2012, 46(20), 10965-10973, https://doi.org/10.1021/es302206a). This limitation should be discussed.

Transport phenomena should also be acknowledged (since the model does not include them) and assumptions should be clearly stated. For example, it is well established that ClNO₂ formation occurs in the nocturnal residual layer (which contains less NO than the surface layer), and ClNO₂ then mixes downward in the morning when the convective mixed layer forms (e.g., Bannan et al., J. Geophys. Res. 2015, 120(11), 5638-5657, https://doi.org/10.1002/2014jd022629; Tham et al. Atmos. Chem. Phys. 2016, 16(23), 14959-14977, https://doi.org/10.5194/acp-16-14959-2016).

Specific comments

Title - specify season of study in title of paper (winter)

line 27 - The authors differentiate between nitryl chloride (ClNO₂) and chlorine nitrite (ClONO), citing a the MCM modeling study by Riedel et al (2014). More explanation is needed here since Riedel et al. do not mention chlorine nitrite in their paper.

Chlorine nitrite may form from Cl+NO₂ (e.g., Golden, J. Phys. Chem. A 2007, 111, 29, 6772–6780, https://doi.org/10.1021/jp069000x and Niki et al., Chemical Physics Letters 1978, 59(1), 78-79, https://doi.org/10.1016/0009-2614(78)85618-8); however, the reaction between Cl and NO₂ is generally thought to be negligible compared to reaction of N₂O₅ on chloride containing aerosol. It would thus be informative to add the relative contributions of ClONO and ClNO₂ to the bottom trace of Figure 2.

Furthermore, ClONO is metastable and converts to ClNO₂ (Janowski et al., Berichte der Bunsengesellschaft für physikalische Chemie 1977 81(12), 1262-1270, https://doi.org/10.1002/bbpc.19770811212; Niki et al (1978)) - a reaction that should be added to Table 1.

line 3. Please define the abbreviation CAABA/MECCA.

line 80. Please state here that the full mechanism is shown in the SI.

Page 4 - Table 1. Please state the units for the reaction rate constants.

For the photolysis reactions, please state the maximum (noon) j values.

Page 6, Table 2 - reaction A13 - please subscript the 3 in CH₃COO.

Line 100 - The Sander et al. (2014) reference is inappropriate. Cite Ghosh et al., J. Phys. Chem. A 2012, 116, 5796-5805, https://doi.org/10.1021/jp207389y, please.

pg 7 - Figure 1 - The nitronium ion (NO₂⁺) is a potent nitrating agent, and there are many more organic molecules in the aerosol-phase than shown here. Please discuss the limitations of the abridged mechanism.

pg 10 - Figure 2. It would be informative to add the relative contributions of ClONO and ClNO₂ to the bottom trace of Figure 2.

The NO₃ and N₂O₅ peaks at noon local time are highly unusual. Consider adding a brief explanatory note to the caption.

pg 10 - Figure 3. The left-hand side graphs suggest that there is no nitryl chloride formation from heterogeneous uptake of N₂O₅ in Delhi. This seems unlikely considering non-zero mixing ratios of N₂O₅ are shown in Figure 2 (see also the next comment).

pg 10 line 201. Morning ClNO₂ peaks are generally due to vertical transport of ClNO₂ produced in the residual layer to the surface. Please cite relevant literature here and discuss.

General Comments

Soni et al present new model results for air quality simulations that include the impacts of chlorine chemistry. The manuscript reports on updates to the chemical mechanism of CAABA/MECCA and discusses the impact of chlorine chemistry in two disparate regions (Leicester and Delhi). I have several general and specific comments that should be addressed prior to publication.

• The model reports a surprisingly large conversion of ClNO2 to Cl2. This is because of the large, condensed phase rate constant for ClNO2 + Cl- that was implemented in the model (Roberts et al 2008). More recent analyses have shown that this rate is likely too large. For example, the analysis of Haskins et al., JGR 2019, using field observations of ClNO2 and Cl2 suggested that this rate must be significantly smaller (of order 1E4 s-1). I think the authors should look at a sensitivity test to this rate to highlight that the selection of the ClNO2+Cl- rate constant has a really significant impact on the Cl production rate.
• As I understand the model treatment of heterogeneous and multiphase reactions here is quite different than what is in most models. Specifically, it appears that N2O5 is equilibrated between the gas and condensed phase using an equilibrium constant then permitted to react. It would be very helpful if the authors compared (and contrasted) this approach to the more common approach of using a reactive uptake coefficient for N2O5 chemistry that is sensitive to the chemical composition and phase of the aerosol particles. I expect that the two approaches would yield quite different results both with respect to magnitude and temporal trends. Since N2O5 chemistry is central to this study, this should be discussed.
• Since the model has a comprehensive gas phase chemical mechanism, and the authors draw conclusions about the role of OH vs Cl radicals in VOC oxidation, it would be a nice opportunity to comment on the production of oVOC that stem directly (and only) from Cl+VOC reactions as they could be used in the future for testing the role of Cl chemistry. For example, what is the mixing ratio of chloroacetone?

Specific Comments

Line 55: This reaction is listed as H1, but H2 (and so on) are all Henry’s Law Equilibriums. Should this be R1?

Line 56: I would remove “recent studies” as it was shown 1997 by Behnke et al that the heterogeneous reaction of N2O5 could form ClNO2 on aqueous chloride containing films.

Table 2: It would be helpful to add the Henry’s Law constants (and references) to the table for the molecules studied. The Henry’s Law constant for most of these gases have never been measured, thus the values reported in the literature are based on model fits to measured reactive uptake coefficients.

Section 3: How is aerosol surface area and aerosol liquid water treated in the model? I appreciate that this may be described in one of the cited references (Rosanka?) But given its central importance to the science discussed here, I think it would be helpful if there was an explicit discussion.

Line 203: I think it would be helpful to cast the ClNO2 loss rates in units of per second as it is easier for readers to compare them to other locations. The ClNO2 + Cl- loss rate is enormous, and it would be helpful to see how that compares to other locations. Specifically, the ClNO2 uptake coefficient is quite small (< 1E-5) even on acidic aerosol. The surface area here must be enormous to drive a loss rate that is 10x that of photolysis (3hr lifetime). I think it would be very helpful to expand the discussion here to think more directly about this comparison.

The manuscript "Comprehensive multiphase chlorine chemistry in the box

model CAABA/MECCA: Implications to atmospheric oxidative capacity" by

Soni et al. describes an expansion of the MECCA chemical mechanism to

include chlorine chemistry. Using published data from India and the

UK, the authors show how the inclusion of this additional chemistry

leads to improved modelling results and present an interesting

analysis of the oxidation chemistry in these two very different

locations.

The manuscript is generally well written, although the English could

benefit by some tweaking, and clearly laid out. I have only a few

comments, and after the authors have addressed them, I recommend

publication.

My main suggestion is to change figures 2 and 6. I think it would make

the whole paper much clearer if they both show the base model, the

base model with added chemistry, and the measurements. To keep the

figures in a manageable size I would suggest having all radical

species in one figure and all non-radicals species in the other

figure. Likewise, I suggest introducing earlier in the paper the three

mechanisms that are now discussed only from section 4.3 onwards. In

this way, it will be easier for the reader to understand how the model

results have changed with the addition of the new Cl chemistry.

line 36: I wouldn't say that the limitation in our understanding of Cl

chemistry is "mostly" due to the limitations of the models. These

processes are also understudied in laboratory/chamber experiments, not

to mention that the database of ambient observations is rather

limited.

lines 125-127. I suggest moving to line 121 the explanation of why the

winter season was chosen for the model simulations, and also add a

note explaining why the Leicester and Delhi datasets were used for

this study.

figure 2: the isoprene mixing ratio in Leicester looks constant. I

assume it is an estimate of some sort, and in an average sense that

may be fine, but the profile is likely unrealistic. The authors should

consider how this affect their results and the related discussion.

line 211: "indicating", rather than "representing"?

line 219: "Cl- concentrations"?

line 226: why are the rate constants for OH + X reactions not taken

from MECCA, like those for Cl + X reactions?

figure 3, and related discussion: the model suggests that the gas

phase reaction Cl + NO2 can be a significant source of ClNO2. As far

as I am aware, most studies indicate the aqueous-phase reaction as the

major (if not only) source of ClNO2, so this may be a potentially

interesting/important finding. Can the authors expand the discussion

on this point? For instance, how well is this reaction known? Have

previous studies considered it?

figure 4, and related discussion: I find it a bit odd that Cl is so

important for the AOC in Leicester when the model predicts significant

concentrations of Cl only around 8am. Likewise the levels of Cl in

Delhi during the night are expected to be very small. Perhaps the

authors should comment on this point.

lines 251-257: it is not clear to me how the base model differ from

the base model without chlorine chemistry. Up until this point I was

under the impression that chlorine chemistry was not present in the

"original" MECCA. Can you please clarify here, and in the Introduction

if necessary, what are the differences in the various mechanisms?