Xin and co-workers present a new set of kinetic measurements of the oxidation of a variety of aliphatic hydrocarbons by the OH radical. This work was performed using a temperature controlled reaction chamber in which the kinetics were determined using the relative rate method and the relative decays of hydrocarbon reactants were tracked using a GC-FID instrument. As the authors point out, the oxidation of hydrocarbons is a subject that is relevant to atmospheric chemistry and the publication of new kinetic measurements is of fundamental importance if we are to assess the effects of these reactions upon atmospheric chemistry. Therefore (in principle) I am in favor of the publication of papers such as this in Atmospheric Chemistry and Physics. However, I do have some serious comments to make, and I would insist that the authors address each one of them to the fullest extent, only then would I support publication.

Major comments:

1. The authors assessment of the kinetics literature is currently incomplete. This is made clear by the fact that they take credit for making the first temperature-dependent measurements in cases where measurements are clearly available, as well as the incomplete literature data presented in Table 2. I would suggest that the authors make use of the database paper of McGillen et al. (2020), and download the accompanying database. This will achieve two things: 1. It will give the authors a more comprehensive knowledge of the kinetics literature for OH + hydrocarbon reactions. 2. It will provide these authors with critically evaluated rate coefficients for many of the species that are contained within their paper. Regarding the latter, I would strongly encourage the authors to use these recommendations as their reference rate constants where applicable and reanalyze their data accordingly, and if not, I would expect the authors to justify why they do not accept these recommendations.
2. The presentation of the data/ quality of the data is unsatisfactory. When I inspect the contents of Table 1, taking the reference rate constants as provided on page 7 of the manuscript, I am able to reproduce kOH for the first 3 entries (i.e. propane in N2 with n-hexane, cyclohexane and n-octane as references). Following this, the next 3 entries (propane in air) are inconsistent. I noticed that throughout this table there are many problems of this type. In my judgement, this is not acceptable for a paper whose principle subject is kinetic data and it undermines your experimental work. What is the purpose of this data, if your readers cannot trust it? For this reason, I insist that the authors return to their spreadsheets, remake this table correctly and triple check its contents.
3. The authors insistence on Arrhenius parameters for this selection of reactions in Section 3.3. is unjustified. With all the high-temperature data available for this collection of compounds, it is plainly obvious that the Arrhenius equation is insufficient to describe the temperature dependencies of any of these reactions. The only reason for using such an equation would be for datasets spanning a small temperature range (or where data precision is insufficient). I strongly encourage the authors to consider fitting their data within the context of the available measurements, because I don’t think these new Arrhenius parameters add any value to your paper.
4. Problems with general consistency. This is an example (but there are several others), Figure 6(a) shows temperature-dependent literature measurements for OH + methylcyclopentane. In the abstract, it states that “… Arrhenius expressions (in units of cm3·molecule-1·s-1) for the reactions of various cyclo- and branched alkanes with OH were determined for the first time: methylcyclopentane…”. I can interpret this in one of two ways: 1. The authors don’t appear to be aware that there is temperature-dependent data, even though they have presented it in their figures. 2. The authors are discounting the work of Sprengnether et al. because it is not presented in Arrhenius form. If the former, then the authors should organize their manuscript more carefully. If the latter, then I find this to be a strange idea (after all, if you don’t like their equation, just re-fit their data with an Arrhenius equation!)…
5. It is not clear to me why the authors chose to consider the bath-gas as an important aspect of these measurements. The literature has many examples of measurements that were conducted in a wide variety of bath gases (helium, argon, nitrogen, air etc.). As far as I am aware, no dependence on bath gas has been noted. In fact, a small amount of oxygen would be necessary in relative rate experiments such as yours, otherwise alkyl radicals formed from the hydrogen abstraction reaction would themselves abstract hydrogen from other alkanes in the system, re-forming the original alkane, consuming some of the other hydrocarbons and confusing your results. In practice, it is difficult to remove oxygen to such an extent, and I would therefore assume that your experiments are not affected by this anoxic chemistry. Either way, I suggest that bath gas is an irrelevance in this paper and can be ignored.
6. There are some advantages to studying so many compounds simultaneously, the main one being that it can save you some time. However, there are also some possible problems. The main one would be the formation of products which could interfere with some of your analyte peaks. I see no discussion of any sort regarding products of the reaction. Do you see any product peaks in the GC-FID? If not, why not?

Minor comments:

General: the symbol for rate constant in the kinetics literature is an italicized lower-case k. It is not to be confused with an upper-case K (which is reserved for units of kelvin), or an italicized upper-case K (which is normally reserved for equilibrium constants).

Abstract: revise according to suggestions above.

Introduction:

Line 44. why are you making comparisons with NO3? It is well known that the abstraction reactions are unimportant for the alkanes. Chlorine on the other hand may become important in some environments.

Line 44. “Dehydrogenation” of alkanes leads to alkenes. You mean to say “hydrogen abstraction”.

Line 46. I assume by “rate constants”, the authors mean “room temperature rate constants”. You should specify this.

Line 47. Assuming that the authors have by now become more familiar with the kinetic database, you will of course know that the range of reactivity of alkanes goes from 6.36E-15 (methane) to 2.16E-11 (n-hexadecane) at the time of writing. The range provided is therefore misleading.

Line 47. “mol” is absolutely not an abbreviation of “molecule”. “mol” is an abbreviation of “mole”, which would be highly misleading.

Line 48. Rate constants are not faster or slower than other rate constants. They are larger or smaller.

Line 66. In fact, precise measurements are highly desirable in the relative rate method. Low precision in your GC-FID measurements would lead to scatter in your relative rate plots. Therefore, this statement is misleading.

Methods:

General comment: several tests have been made with respect to dark losses, photolytic losses etc. However, as far as I can tell, no tests were performed regarding storage in the 1 litre sample bags. In your experiments, your samples are stored for some time before they are analysed by the GC-FID are they not? During this time, your samples are subjected to conditions of higher surface area to volume ratios, and it is here, where I would expect to see the most wall loss.

Line 115. Excess with respect to what?

Line 145. “self-developed”.

Line 155. It is not clear to me how you have improved or expanded the work of Shaw et al. Furthermore, it is also not clear to me how the results of this work are significantly different from any other relative rate study in which several reference compounds are considered.

Line 173. There is a certain irony to this statement, because I would strongly recommend that you should use the available expert-evaluated rate constants wherever possible.

Line 193. Was the H2O2 purified?

Lines 211-212. I don’t know what the authors mean by general error is 2 sigma… Do you mean that the uncertainties for SAR estimates are generally within a factor of two? This is possibly true in a global sense, but it is considerably lower for the alkane dataset (the subject of this paper).

Line 238. I don’t know what an “error strip” is. I think the authors mean “error bars”.

Line 238. What is fitting dispersion?

Results:

Figure 3. It would be useful to include literature recommendations for each of these rate constants where available, allowing us to see how well the experiments are performing.

Table 1. As noted above, the errors in this table are unacceptable at present. In addition to this, the formatting of this table is confusing and should be rethought to help the readers understand it better.

Structure-activity relationships section:

You only compare with one SAR (the Atkinson SAR). This is a missed opportunity, there are several others to choose from (some examples include: Neeb, 2000; Jenkin et al., 2018; McGillen et al., 2024). In the case of Neeb, this is of particular relevance because there is some critical discussion on the use of a ring-strain correction factor on page 6 (300) of that study, the authors should consider this in their discussion of SAR performance for cyclic compounds. Incidentally, the recent SAR of McGillen et al. (2024) is not currently configured for cyclic compounds, however, I have assessed its performance on the selection of compounds of this paper. This is very easily done by running the Python code of that paper, from which I find that it performs marginally better than Kwok and Atkinson (1995). This, at least for this limited selection of compounds, supports Neeb’s statements about ring strain corrections. However, it is most likely the case that more data on cyclic alkanes of differing ring size would be very useful in assessing this further.

References:

McGillen et al. (2020) DOI: 10.5194/essd-12-1203-2020

Sprengnether et al. (2009) DOI: 10.1021/jp810412m

Neeb (2000) DOI: 10.1023/A:1006278410328

Jenkin et al. (2018) DOI: 10.5194/acp-18-9297-2018

McGillen et al. (2024) DOI: 10.1039/D3EA00147D

Kwok and Atkinson (1995) DOI: 10.1016/1352-2310(95)00069-B

see attached file

General comments:

The article entitled “Rate coefficients for the reactions of OH radical with C3-C11 alkanes determined by the relative rate technique” by Xin et al., reports a series of rate coefficients measurements for the reactions of 25 alkanes with OH radicals using a relative rate technique. The research employed the GC-FID technique and a simulation chamber to evaluate the rate coefficients values at 25 C and Arrhenius expressions in the 273-323 K temperature range. The authors also explored the structure activity relationship approach to evaluate the OH radical rate coefficients of the alkane series. The main objective of the author’s research was to provide multiple kinetic data studying simultaneously a large number of compounds, but even achieved, however, the lack of higher accuracy, quality and novelty for released rate coefficients do not justify the publication of the article in the present form. An extensive revaluation of the data could help to approach this study from a different perspective and enhance the quality of the results.

Major comments:

1. The introduction provides a large number of literature data; however, a very chaotic presentation induces to readers the feeling of jumping from one study to another, all of them mostly with very general information regarding the kinetic of alkanes. The studies mentioned in the introduction are presented without an effective detail of the rate coefficient information. I would write the introduction assessing the importance of alkanes for air quality with their impact to atmosphere, then I would add information about their concentrations in troposphere and the impact on potential ozone formation, potential SOA formation and their sources and sinks. One important point of the study is to highlight the importance of accurate kinetic rate coefficients for database, global model atmospheric processes and degradation mechanisms. As one of the reviewers mentioned already, the McGillen et al., (2018) database is very important to be used as a start in the literature data assessment and comparison.
2. The reason of selected those 25 alkanes is not presented in the study and their selection looks arbitrary. An organized evaluation based on their separation on straight-chain, branching and cyclic alkane structure would be more interesting and helpful to get valuable information. First class of alkanes including linear compounds could provide information regarding reactivity of each CH₂ group added to their structure and how the rate coefficient value would change over increasing alkane chain. Secondly, from the branching alkanes the authors could have more information related to CH groups added to the alkane structure. The third class in the evaluation would be cyclic alkanes where the authors could extract information regarding the reactivity increase with the cycle size from cyclobutane to cyclodecane. The authors should include in their evaluation the alkanes rate coefficients studied in present study and those existing in the literature, to release more complete discussion on their behaviour and reactivity (figure 3 and 4). As an example, there are a lot of data which could be included, mentioning here only a few (k_OH for cyclooctane, bicyclo octane, methyl octane, etc.). 
3. Since the authors highlight their first study on 3-methylheptane why they do not cover other not studied yet branched alkanes (2,2,3-trimethylpentane, etc.)
4. The data reported in the Table 2 should be reevaluated and presented in more concise and understandable form. There are multiple examples of inconsistency of the data with many average data not well calculated (i.e. isopentane in N2).
5. With the extensive interpretation of data including the existing literature rate coefficients, the authors would be able to evaluate the accurate reactivity trends for the class of alkane toward OH radicals and calculate new factors for the SAR method and then to improve the SAR method. A simple comparison with the Kwok and Atkinson SAR method is not worth to do. Evaluation of existing SAR approaches in the literature, with discussion about the influence on the substituent factors, is necessary. (McGillen et al., 2024 (doi.org/10.1039/D3EA00147D), Jenkin et al., 2018 (doi.org/10.5194/acp18-9297-2018)
6. The reaction channel of the OH radical initiated degradation of alkanes is strictly related to hydrogen abstraction in the presence of oxygen. There are clearly correlations on the alkane reactivity with OH radicals and Cl atoms for all saturated class of VOCs. Please evaluate a log-log correlation of k_CL and k_OH as presented by Calvert et al., 2011 (Calvert, J., Mellouki, A., Orlando, J., Pilling, M., and Wallington, T.: Mechanisms of Atmospheric Oxidation of the Oxygenates, Oxford University Press) for alkanes, saturated alcohols and ethers and also by Tovar et al. (2022) (doi.org/10.5194/acp-22-6989-2022) for saturated epoxides.
7. The importance of the bath-gas is over highlighted in this study and a single example for a selected alkane would be enough to prove that is no bath-gas effect on the rate coefficient value. A revaluation of the paper consistency should be performed.
8. Please add more information regarding the conditions needed for relative rate techniques. Also include more advantages and disadvantages of the absolute and relative techniques.
9. The authors should highlight the atmospheric implication and the impact of their research as requested by a scientific journal as “Atmospheric Chemistry and Physics“. Please add information about the alkane lifetime in the atmosphere toward the OH radicals. A more extensive conclusion and atmospheric implication should be performed.

Minor comments:

Both reports of the  previous  reviewers have included a detailed list of minor comments and only additional comments I would add here:

Line 84: Finlayson-Pitts

Line 127: “at253”

Line 130: please avoid given values in the form of 0.00013 or 0.00048. Change the units to pptv/h.

Line 283 and 285: please add units

Line 460: please revise the rate coefficient

Line 104: The figure of the simulation chamber shows a ratio of 200:50 of N2:O2 mixture. The study used synthetic air or this mixture shown in the figure?