The authors investigated the SOA formation from the ozonolysis of b-caryophyllene (BCP) at different temperatures in an atmospheric simulation chamber and showed that the SOA particle yields increase with decreasing temperatures. The authors attributed this to the decrease of the vapour pressure of the oxidation products at reduced temperatures and supported this by showing that a larger fraction of products with lower number of oxygen atoms were present at lower temperatures compared to higher, where higher O:C was observed. They also provide, for the first time, reaction rate coefficients for the reaction of BCP with ozone at different temperatures. Furthermore, in each experiment, after the initial amounts of BCP were consumed, additional BCP, ozone and nitrogen dioxide were added to the chamber to form SOA in the presence of nitrate radicals. Advanced mass spectrometry techniques were used to investigate the chemical composition of the gas and particle phase in each experimental setting and showed that the chemical composition had significant temperature dependence.

Overall, I believe that the manuscript falls within the scope of ACP as it provides new information for the SOA formation, composition and kinetics from the oxidation of BCP and importantly, their temperature dependence. I recommend this work for publication after a few major and some minor considerations have been addressed.

General comments:

1. The study suggests that the experiments were conducted under representative of the real atmosphere conditions, which might be true for the selected temperature and relative humidity conditions. However, the study uses unrealistically high oxidant and precursor concentrations that could have altered the fate of the radicals. I am therefore wondering how the resulted chemical regime of the experiments could have affected the results presented and their implications to the real atmosphere.
2. All the experiments aside from the different temperatures were also conducted under substantially different relative humidity (RH) conditions (13-97%). It was recently shown that the BCP-SOA chemical composition could be considerably affected by the RH levels (e.g., Kundu et al., 2017), therefore how the difference in the RH between the experiments conducted in this study could have affected the reported results? I understand that the authors try to capture the variation of the temperature and RH found in the different layers of the atmosphere. However, I believe that the potential implications of the different RH levels should be included in the interpretation of the results.
3. The study compares the SOA yields obtained in this study with values obtained previously in the literature and attributes the observed differences to the potentially different oxidative conditions. Whilst this, at least in part might be true, it may be worthwhile considering the potentially different losses and partitioning of the semi-volatile vapours in the different chambers - even so considering the different materials of the chambers (i.e., Teflon vs Aluminum). Additional discussion should probably be included when interpreting and discussing the SOA yield results.
4. In my opinion, a lot more information is required for the operation and data analysis of the FIGARO-CIMS dataset. The authors report a considerably large fraction of products with particularly high molecular weight (>400 Th). It would be helpful for the readers to know the range the mass calibration was conducted, the associated peak assignment errors and thereby, the confidence in the results presented. Furthermore, additional information about the particle-phase background subtraction and sampling strategy would be beneficial.
5. I believe that the absorptive partitioning should be considered when discussing and interpreting the results and particularly, the different levels of absorptive mass in each experimental setting. For example, higher SOA yields and lower O:C have been observed at the lower temperature experiments opposed to those conducted at higher temperatures (that exhibited lower yield + higher O:C). Could this behaviour be attributed to the higher levels of absorptive mass present at the higher yield experiments, enabling the partitioning of the less oxygenated (and consequently more volatile species) to the condensed phase, thus decreasing the average O:C? To better understand this, I think that it would be beneficial, at least in the supporting information, to show the time-series of the SOA mass in each experimental setting. This effect could be even more pronounced when the authors compare the ozonolysis experiments with those in the presence of NO3. The ozonolysis experiments entailed the formation of SOA from the nucleation of the oxidation products, whereas those formed in the presence of NO3 involve the condensation of the species on the top of those pre-existing particles. Intuitively, the partitioning behaviour of the species in each of these cases could be significantly altered due to the significantly different absorptive mass present and thereby could have affected the reported results and their comparison.
6. Further to the above, all the experiments were conducted in the absence of seed particles. Could the potentially different particle number (and thereby surface area) in each experimental setting have affected the partitioning of the species between the particle phase and the chamber walls? What would be the implications on the results? Again, time-series of the particle number/surface area in each experiment might be beneficial.

Specific comments

L57, L87 and L217: As from 2020, IUPAC has released updated values for the reaction rate coefficients for major organic compounds, including BCP that can be found on Cox et al., (2020) and Mellouki et al., (2021). I recommend those values to be used throughout the manuscript, including the calculations performed.

L126: “saturated with its vapour” how this was confirmed?

L140: Why the authors decided to get a slower decay only in the 273K experiments and not the rest? How this could have affected the results?

L171: I think that it would be beneficial for the readers if the authors provide a more detailed description of their error propagation estimates.

L183: What was the relative ionisation efficiency of the AMS and how it was obtained?

L191: where the remaining 0.003m3/min are going? I presume to the exhaust but it would be nice if this was clarified in the revised manuscript.

L204-207: In line with the general comment #4, it is unclear how the background subtraction was conducted. I think that additional information for the data processing would be beneficial.

L231: It is unclear to me how the COSIMA accounts for the losses of semi-volatile vapours. Furthermore, to what extent the interactions of the aluminium walls of the AIDA with the losses of particles and the partitioning of the semi-volatile vapours were considered? Does this have any implications for the reported results?

L246-250: BCP secondary additions are not visible in Fig.2, while the third is not described in the methods section nor related data are shown in table 1. I think that such information should be detailed in the experimental conditions section.

L259: I generally do not support the calculation of the total mass concentrations from the FIGAREO-CIMS assuming maximum sensitivity for all the compounds detected, given that a lot of work has been devoted to illustrate the issue of the differential sensitivity in the instrument along with potential ways to constrain such issues (e.g., Lopez-Hilfiker et al., 2016). Nonetheless, I do see some value in comparing the estimated trends with those derived from the HR-AMS. I would recommend however, definite quantitative statements such as those presented on that sentence to be soften.

L279: Do you have any evidence that the SOA particle yields from the OH oxidation should be higher than the ozonolysis? For other precursors, such as the a-pinene, the SOA yields from the O3 oxidations have been found to be higher than the OH-initiated oxidation.

L400: How the contribution of the NO3 oxidation was estimated at different temperatures, given that there are no available BCP reaction rates with temperature? What would be the error in those estimates?  Perhaps additional information about the box modelling should be included in the manuscript.

L411: Why the formation of C15H25O7N indicates that BCP is reacting directly with NO3 and the formation and not that certain pre-existing O3-initiated oxidation products are reacting with NO3? Perhaps a more descriptive approach will benefit the readers.

L421: How the mass of the organic nitrates was calculated from the HR-AMS measurements? It is known that it is challenging to retrieve N-containing species from that instrument (e.g., Farmer et al., 2010), so it would be helpful if more information was provided.

L451: Perhaps adding a reference about the potential thermal instability of N-containing compounds at 313K would be beneficial here. If so, how about their decomposition in the FIGAERO-CIMS? 

Technical corrections

L101: “big condensation potential”, I think that the authors mean that at lower temperatures the vapour pressure of the components is lower and thereby altering the partitioning behaviour of the species. Probably this sentence needs some rephrasing and appropriate references.

L103: this sentence is a bit difficult to understand, please re-write.

L115: “bottom”, I assume of the chamber. Please re-write.

L168: please remove the parenthesis from the contribution of the major ions

L381-386: these are not annotated in any of the figures and are a bit difficult to follow, please re-write.

L390: From figure 2 second addition of BCP appears to be between 150-160 min instead of 180-190 min.

L430: This sentence is not very clear could you please re-write? Also, could this shift be attributed to the different total absorptive mass present enabling the partitioning of the less oxygenated and consequently more volatile species to the particle phase (see also general comment #5)?

L435-436 and L439-440: these sentences are a bit difficult to understand, please re-write.

L457-458: I am not sure if the “BCP ozonolysis in the presence of NO3 radicals” is the right wording here. If I understand correctly, the BCP oxidation in the last stage of the experiment occurs in the presence of both oxidants.

Fig. S3 I struggle to differentiate the modelled from the measured ozone. Additionally, the reaction rates should have different units?

Please also re-check figure numbering.

References

Cox et al., (2020), Evaluated kinetic and photochemical data for atmospheric chemistry: Volume VII – Criegee intermediates. Atmos. Chem. Phys., 20, 13497–13519, https://doi.org/10.5194/acp-20-13497-2020

Farmer et al., (2010), Response of an aerosol mass spectrometer to organonitrates and organosulfates and implications for atmospheric chemistry, P. Natl. Acad. Sci. USA, 107, 6670–6675, https://doi.org/10.1073/pnas.0912340107

Kundu et al.., (2017), Molecular formula composition of β-caryophyllene ozonolysis SOA formed in humid and dry conditions, Atmospheric Environment, 154, 2017, 70-81, https://doi.org/10.1016/j.atmosenv.2016.12.031.

Lopez-Hilfiker et al., (2016), Constraining the sensitivity of iodide adduct chemical ionization mass spectrometry to multifunctional organic molecules using the collision limit and thermodynamic stability of iodide ion adducts, Atmos. Meas. Tech., 9, 1505–1512, https://doi.org/10.5194/amt-9-1505-2016

Mellouki et al., (2021), Evaluated kinetic and photochemical data for atmospheric chemistry: volume VIII – gas-phase reactions of organic species with four, or more, carbon atoms ( ≥  C4). Atmos. Chem. Phys., 21, 4797–4808, https://doi.org/10.5194/acp-21-4797-2021

In the study by Gao et al., the authors explore the oxidation of b-caryophyllene (BCP) by dark ozonolysis performed in the AIDA atmospheric simulation chamber with and without the presence of nitrogen oxides at different temperatures. From state-of-the art analytical techniques, the authors report on the formation and composition of resulting gas- and particle-phase secondary organic aerosol (SOA). Presented results show that temperature effects both SOA yields as well as chemical composition which the authors attribute to temperature-dependent difference in vapour pressure of the BCP oxidation products. Following initial ozonolysis of BPC, the influence of nitrogen oxides is examined by the addition of nitrogen dioxide and additional BPC to the ozone-filled chamber. In the presence of nitrogen oxides, the authors show formation of organonitrates contribution to the SOA with higher contributions of more oxygenated species observed at higher temperatures.

The manuscript provides new and important findings on the temperature-dependent formation of SOA from BCP. The applied analytical techniques are comprehensive including both gas and particle phase characterization of the formed SOA. The manuscript is well written, the results are clearly presented and discussed, and the topic falls within the scope of ACP. I thus recommend this work for publication once the following comments have been addressed.

General comments:

1. Discussions on the chemical compositions is mostly based on data from experiment 1a (213K) and 5a (313K). However, in experiment 1a knowledge of the experimental conditions are incomplete from the lack of BCP measurements. In general, the reviewer finds that all experiments vary in their execution and experimental conditions other than temperature. In particularly, large variation in BCP concentrations, RH and BCP/Ozone ratios are noted. How do the authors justify comparison between experiments and in particular to exp. 1a?
2. Could the author comments on the expected phase-state of the formed SOA particles under the studied conditions? With the large temperature and RH span, the authors should consider this for two at least two reasons; 1) the partitioning of oragnics (e.g. semi-volatile org and org-Ns species) to the preexisting SOA particles (e.g. Bastelberger, 2017) and 2) particle-bound (surface or bulk) reactions (e.g. Shiraiwa, 2011). With respect to the latter, particle phase-state (solid or liquid) could affect both surface oxidation processes by ozone and OH-radicals but might also affect the prosed formation of dimers through esterification.
3. The authors calculate OH-radical yield under the studied conditions and find significantly higher yields at elevated temperatures (5% at 243K vs 15% at 313K). However, due to the fast reaction of ozone with the endocyclic double bond of BCP 91-92% of BCP are calculated to react with ozone under the studied conditions, hence rendering contributions of OH-oxidation of BCP minor. However, considering that the authors attribute formation of higher oxidized compounds to further gas-phase oxidations of first-generation products, it would be worthwhile to discuss the influence of OH-radicals in this regard. In Witkowski (2019), several oxidation products from the OH + β-caryophyllonic acid has been identified including many found in the gas and particle phase of the current study, including C₄H₆O₄ and C₁₄H₂₂O₄ ,with the former identified in gas-phase SOA in 298K and 313K experiments only (Fig. 5).
4. On the composition of both gas and particle phase SOA data are presented as normalized signals to total gas or particle C_xH_yO_z(%), respectively, at the given temperature. Whilst this provide a clear picture of the changes to the chemical composition, it fails to report on the differences in concentration of individual species between experiments. For example, in Fig. 5, showing the average CIMS gas phase mass spectra, even at low 213K and 243K gas-phase species seems to be abundant. It would be beneficial to include (in the SI) the absolute signals of the identified compounds, thus to be able to note the differences in abundance of these species between the experiments. As shown in Fig. 5, one might conclude that the gas-phase contain less organic compounds thus undermining the statement of gas-phase oxidation of first generation oxidation products as possible source to more oxidized monomers in particles formed at elevated temperatures. Also, as no BCP could be detected in the gas-phase at 213K due to wall-losses, how do the authors explain the detection of the oxidized species in Fig. 5?

Specific comments:

1. The general experimental protocol is unclear and seems to vary between experiments. From Fig.2 and table 1, two ozonolysis experiments without NO2 (2a and 4a) include more than one addition of BCP. What is the reason for this?
2. Fig 2 indicates that the initial oxidation of the added BCP (65 µg m-3) was performed using a lower ozone concentration (25 ppb) than stated in table 1 (325 ppb). Please clarify.
3. It is not apparent to the reviewer, whether results from yield calculations and chemical analysis relates to SOA formed after the initial (time 0, low ozone, Fig. 2) ozonlysis or following the second addition of BCP at much higher O3 concentration (300 ppb).
4. Despite the typical high time resolution and sensitivity of the PTR-ToF-MS no BCP measurements are reported during the two additions of BCP in Fig. 2. Why is this? If the rate by which BCP is added is the same during the two additions, it would have been useful to see how the loss of BCP changes under the studied conditions.
5. Line 126-128: the authors refer to the low vapour pressure and strong wall losses as possible explanation for the lacking BCP measurements at low temperatures. If all BCP is lost to the chamber walls, from where do the SOA particle mass form? Reactions on the chamber surfaces? if so, how can the authors account for this in their experiments?
6. Line 147-148: Do the authors expect any issues from operating all instruments at 296K? This is significantly higher than experimental conditions of 213K and lower than 313K, thus may produce bias in the gas and particle phase measurements from evaporation and condensation of semi-volatile species during sampling.   
7. Line 191-195: How many particle samples were collected for each experiment and how often were these collected? If multiple samples were collected, showing the evolution of the spectra or specific species (i.e. dimers and trimers) over time could be beneficial.
8. Line240-241: it would be useful if the authors provided similar figures as Fig. 2 (in SI) for all experiments conducted. Also, the authors should state the maximum particle number concentration and particle size (Table 1 og SI).
9. Line 251: Did the authors observe new particle formation following the last injection of BCP in all experiments?
10. Line 278-279: What was the reasoning behind not applying OH-scavengers and seed particles to the experiments?
11. Line 354-355: How do the concentration/relative signal of dimers and monomer change over time? If dimers are formed from esterification of monomers, this could be evident from continuous increase of dimeric compounds after BCP depletion. In SI only time resolved data of monomeric species are shown.
12. Line 356-358: Why are the dimeric molecules not observed at temperatures above 273K, despite the presence of the monomeric precursors?

Technical corrections

Line 354: Remove punctuation mark after “…vapor pressure”

References

Witkowski B, Al-sharafi M, Gierczak T (2019) Kinetics and products of the aqueous-phase oxidation of β-caryophyllonic acid by hydroxyl radicals, Atmospheric Environment, 213, 231-238,

Bastelberger S, Krieger UK, Luo B, & Thomas P (2017) Diffusivity measurements of volatile organics in levitated viscous aerosol particles. Atmospheric Chemistry and Physics 17(13):8453-8471

Shiraiwa M, Ammann M, Koop T, & Pöschl U (2011) Gas uptake and chemical aging of semisolid organic aerosol particles. Proceedings of the National Academy of Sciences 108(27):11003-11008.