General comments:

This manuscript describes a combined experimental and theoretical study on the observation of (hydro)peroxides using PTR-based mass spectrometers. The authors argue that due to fragmentation of the peroxides upon addition of the ionizing proton, the (quantifiable) detectability of peroxides is severely hampered, and does not appear to be a viable approach. They clearly document their reasons for this assertion, based on observations of peroxides (or the lack thereof) and an extensive set of theoretical calculations showing facile decomposition pathways for a myriad of protonated peroxides. The paper is well written, reads easily with a clear message founded in high-quality results. I support publication of the manuscript, with only some minor comments below. Note, however, that I am not able to properly assess the experimental part of the paper and focus mostly on the theoretical part.

Specific comments:

p. 12, line 285: "This may change if an oxygenated substituent is present, since very low or no reaction barriers were observed from calculations involving proton transfer to and from these groups."

Do the authors mean to say that migration to adjacent oxygen groups is slow, but transfer to more distant groups is faster? If so, at what distance would the transfer become sufficiently fast to occur under the conditions in this paper ?

The use of molecular dynamics to determine energy transfer parameters is not new, and it might be useful to reference earlier efforts. Examples include

Barker, J. R.: Energy transfer in master equation simulations: A new approach, Int. J. Chem. Kinet., 41, 748–763, https://doi.org/10.1002/kin.20447, 2009.

Barker, J. R. and Weston, R. E., Jr.: Collisional Energy Transfer Probability Densities P(E, J; E′, J′) for Monatomics Colliding with Large Molecules, J. Phys. Chem. A, 114, 10619–10633, https://doi.org/10.1021/jp106443d,

Jasper, A. W., Pelzer, K. M., Miller, J. A., Kamarchik, E., Harding, L. B., and Klippenstein, S. J.: Predictive a priori pressure-dependent kinetics, Science, 346, 1212–1215, https://doi.org/10.1126/science.1260856, 2014.2010.

SI: "In particular, it is apparent that the models used by MESMER significantly underestimate the collision rate between nitrogen gas and a large gas-phase molecule, and therefore when using the default parameters the rate of thermalization is in turn underestimated." and

"The selected parameter values, in special σ, are much higher than those offered by MESMER as default (ε/kB = 50 K and σ = 5 Å)"

I don't understand the emphasis on the default collision model parameters in MESMER. It can only have a single default value, but only by chance would these values be appropriate for any study. Default values should not be used as a black box but rather require explicit deliberation, given that it is well known that collision parameters are dependent on the molecule considered, and extensive work exists determining appropriate values for many compounds experimentally and theoretically. The manual of Multiwell, for example, even has a handy table of collision parameters for a variety of molecular sizes. The values obtained by the authors in the current work are by no means surprising for molecules of the size studied here. I propose the authors remove this awkward consideration that default values in MESMER might somehow have been appropriate, and rather look in collisional energy transfer literature for studies that examine the dependence of the parameters on molecular size to state that the values obtained from the MD calculations are not that unexpected (the references in the Multiwell manual would be good start). There are also methods for estimating such molecule-specific parameters that are not based on quantum chemical or MD calculations.

SI: "The overall temperature equilibration must therefore arise from the relatively small asymmetry in this histogram, which we see from Figure S4 becomes more symmetric as the simulation proceeds and the temperatures of the gas and analyte become closer."

I would avoid stating that "the models used by MESMER significantly underestimate the collision rate", as the problem appears to be not so much the models but the inappropriate use of black box defaults. The histograms derived from the MD calculations confirms the collisional models in MESMER, i.e. that the amount of energy transfered is roughly exponential for both up- and down-transitions, and that it is the small asymmetric between these curves that causes thermalization. This has been already described extensively in the literature. Some references to this literature would be appropriate.

The authors do not discuss an E-dependence of delta-Edown. Is this because they do not have sufficient data to derive the dependence, or did they find only a negligible dependence ?

SI, p. 12, start of lower paragraph: "Reaction dynamics simulations for MeOOH2+ revealed that this species decomposes only at thermal rates".

Add : "across the pressure range considered here"

Typographic comments:

Add page numbers to SI

p. 6, line 163: 10-5 Hartree : superscript on exponent -5

SI, p. 2, 3th line from bottom: "th eory" ; remove spaces

SI: throughout the text "kcal.mol-1". SI notation of units does not use periods (neither for the abbreviation nor for the multiplication)

SI : "Molecular structures carrying the double-dagger symbol ‡ are transition states." This is evident from the schemes and need not be indicated explicitly.

This manuscript shows the reason why recently developed PTR-based mass spectrometers are not easily able to detect ROOR and ROOH species via extensive experiments and theoretical calculations. The authors used a-pinene and cyclohexene ozonolysis experiments as well as model peroxide, hydroperoxide molecules for the calculation. Among various factors, they have investigated Pressure of FIMR, DC voltage, model compound, and substitutes were investigated. In the experiment, they showed dimers are hardly detected and confirmed that it was not because of low transmission on that mass per charge range. With increasing pressure, less fragmentation was observed and calculated. That was because of more frequent collisions with gas molecules which release excess internal energy. In addition, higher DC voltage in FIMR had a negligible effect on fragmentation. Varying model molecules showed several different results. For example, MeOOMeH+ showed less fragmentation because of their inductive effect and hyperconjugation which stabilize the transition state. Oxygen-based substitutes have a lower level of fragmentation because their initial protonation isomer has a lower barrier for isomerization thus more stable. For hydroperoxides, because of the higher degree of polarizability and smaller proton affinity, they were not easy to be detected by the instrument. This systematic study of model molecules in both experiment and theoretical calculation is well conducted and the manuscript is well written, and the interpretation of the results is reasonable. It is a very informative paper and falls well within the scope of AMT. Therefore, I recommend this paper to publish with some very minor comments and some questions.

1. Do you have any idea why a-pinene dimers were not detected at all but some cyclohexene dimers are detected?

2. I think in the introduction, not much about hydroperoxides is mentioned.

3. Would it be different if the experiment is conducted in higher relative humidity like 50 %?

4. Page 7.

How is the reaction ratio between RO2+RO2 / RO2+HO2?

Do you also have a similar experiment at higher RO2 + HO2? Like in the presence of CO?

5. Page 17, line 383

Can you explain more about why the opposite trend was observed for hydroperoxide?

6. Supplement S2.4 Page 7. Result and discussion, end of it (with the same page with Figure S4 and Table S1)

“The overall temperature equilibration must therefore arise from the relatively small asymmetry in this histogram, which we see from Figure S4 becomes more symmetric as the simulation proceeds and the temperature of the gas and analyte become closer.”

--> It is hard to see how asymmetry (I agree it is small) become more symmetric as simulation proceed. No systematic change with increasing time. Can you give us more explanation about this?

7. Supplement S2.4 Page 8(Fitting MESMER parameters)

I still did not get why cyclohexene has a longer time for the molecule temperature to be equilibrium as surrounding gas than EtOOHEt+ besides higher collision frequency and larger energy transfer. Can you explain more about this?

8. This paper can be added as a reference. This provides basic information about the fragmentation patterns from PTR-MS.

Pagonis D, Sekimoto K, de Gouw J. A Library of Proton-Transfer Reactions of H3O+ Ions Used for Trace Gas Detection. J Am Soc Mass Spectrom. 2019 Jul;30(7):1330-1335. DOI: 10.1007/s13361-019-02209-3. Epub 2019 Apr 29. PMID: 31037568.

For correction

1. Supplement page 2, line 6
   
   theory remove empty spaces
2. Add page numbers in supplement
3. Two S2.4 (Result and discussion, Fitting MESMER parameters)
4. Remove dot from kcal.mole^-1
5. Supplement Page 11 8^th line from the bottom
   
   “stabilised” --> stabilized (unify UK/US English)