General Comments:

Zu et al. investigate the influence of dimethylamine (DMA) on HIO₃-HIO₂ cluster formation using quantum chemical methods and atmospheric cluster dynamics simulations. This is an excellent and natural extension of the previous studies on iodine oxoacids by the same group.

A funneling approach is used to identify the cluster configurations lowest in free energy. The final cluster structures are calculated using density functional theory (ωB97X-D/6-311++G (3df,3pd)) and the single point energy is calculated using RI-CC2/aug-cc-pVTZ calculations. The calculated thermochemistry is applied as input to the atmospheric cluster dynamics code (ACDC) to simulate new particle formation rates in various marine regions (Mace Head, Zhejiang and Aboa). The main finding is that the HIO₃-HIO₂ cluster formation rates does not correspond to the NPF observations, but DMA enhance the cluster formation rate by several orders of magnitude, thereby increasing the agreement between the modelling and the observations.

I only have some minor quarrels with the applied methodology. The cluster formation simulations are very sensitive to the quantum chemical data, so some sensitivity runs should be performed to see how robust the conclusions are to the applied level of theory. In addition, the influence of other nucleation precursors (SA, MSA, multiple bases, ect) should be further discussed in the manuscript to emphasize that the HIO₃-HIO₂-DMA mechanism is not the only explanation for the gap between theory and measurements.  However, the authors do not need to carry out the actual calculations, just discuss the potential importance of other species.

Overall, I believe the chosen systems and current study at hand is an interesting addition to the literature. The manuscript is easy to follow and the topic falls within the scope of ACP.

Specific Comments:

Introduction: I am missing some introduction to what we, in general, know about cluster formation from previous quantum chemical studies. Please put the current study into context of the whole field and not just iodine studies. What vapours have previously been studied and found important and what are the main findings of previous work?

Line 38: How high are the HIO₃ and HIO₂ concentrations measured at Mace Head? Please state the concentrations here as well.

Line 73-74: ”Firstly, the ABCluster program (Zhang and Dolg, 2015) was performed to generate up to 120000 initial isomer structures using the artificial bee algorithm.”

Where does the number 120000 come from? Is this the ABCluster population value (SN) times the number of generations? Some more information on the ABCluster parameters would be a useful addition.

Line 75-91: I am missing some comments on the accuracy of the applied configurational sampling methodology and the applied quantum chemical methods.

• Only saving 1000 local minima from the ABCluster search sounds a bit low. How certain are the authors that they have located the global minimum?
• As the UFF forcefield cannot handle bond breaking a more diverse pool of clusters is usually desirable. This is usually done by performing ABCluster runs with ionic monomers as well (see Kubečka et al., https://doi.org/10.1021/acs.jpca.9b03853).
• Only selecting the lowest 100 cluster configurations based on PM7 could lead to the global minimum cluster being missed (see Kurfman et al., https://doi.org/10.1021/acs.jpca.1c00872). Could the authors comment on this aspect?
• How accurate are the RI-CC2/aug-cc-pVTZ calculations? The leading terms in the CC2 equations are MP2-like, at the cost of N⁵. Hence, you could get accurate DLPNO-CCSD(T)/aug-cc-pVTZ calculations at essentially the same computational cost.
• The authors admit that previous agreement with experiments is caused by random cancellation of errors. Our previous work has shown (Schmitz et al., https://doi.org/10.1021/acsomega.0c00436) that RI-CC2/aug-cc-pVTZ is severely overbinding, i.e. yielding too negative binding energies, thus leading to too stable clusters. Where is the remaining error cancellation coming from? All missing effects in the simulations (hydration, ionic effects, anharmonicity, potential inadequate sampling, ect …) would make the clusters more stable and hence make the current results agree less with experiments.
• ACDC simulations are extremely sensitive to the applied QC methods, and we can essentially get whatever we want by tweaking the level of theory. Hence, some further information on how we can trust the results is warranted. How robust are the conclusions to the applied level of theory? I suggest the authors test if the ωB97X-D/6-311++G(3df,3pd) calculations without RI-CC2 are yielding the same conclusions. Hence, this would not require additional calculations, but vastly improve the reliability of the study.

Line 113-115: Please also mention the boundary conditions here in the main text. Setting the boundary clusters as clusters consisting of only six molecules could lead to artefacts in the ACDC simulations, thereby yielding too high cluster formation rates (see Besel et al., https://doi.org/10.1021/acs.jpca.0c03984). For some acid-base systems the “critical cluster” is already found within the initial 2x2 cluster system, however this depend on the given base (https://doi.org/10.1021/acs.jpca.3c00068). Overall, I am not entirely convinced that the boundary conditions are adequate in the current study and might yield too high cluster formation rates. Please elaborate on this aspect.

Line 118-132: I do not see what Figure 1 is contributing with to the present study. It is simple chemistry to identify the donor/acceptor groups in molecules. No need for electrostatic potential maps for doing this. Please remove this part.

Line 137: “… , which proves the prediction of electrostatic potential analysis”

Please remove.

Line 143: “… and HIO₂ can also act as a stabilizing base in the neutral nucleation process of HIO₃-HIO₂” and “Hence, the participation of DMA may potentially lead to a competition between two basic molecules for proton transfer reaction.”

I am not completely comfortable calling iodous acid a base (in the Brøndsted-Lowry acid-base formalism). I agree that the HIO₂ show peculiar proton transfer dynamics, but I would refrain from calling it a base. I guess it is technically amphoteric.

Line 153-154: “… which possesses relatively stronger basicity than HIO₂ in the process of proton transfer.”

What is the gas-phase basicity and pK_a values of DMA and HIO₂ respectively?

Line 180: What is the absolute cluster formation rate for the conditions given in Figure 3? How does this compare to the conventional SA-DMA system?

Line 201-202: “… and the HIO₃-HIO₂-DMA ternary nucleation is critical in explaining the missing sources of new particles especially in the place where the concentrations of HIO₂ and DMA are similar.”

I would be careful stating that the HIO₃-HIO₂-DMA mechanism is the “critical” missing link. It might contribute, but other mechanisms might also be important. Some discussion on the potential other species (SA, MSA, multibases, water, ect) that might contribute to cluster formation in marine environments would be a welcome addition to the manuscript.

Line 204-205: “This is the first time that a combined influence of multiple bases has been discovered in the nucleation process driven by HIO₃, …”

Again, I am not comfortable calling HIO₃-HIO₂-DMA a “two”-base system. Please remove this sentence.

Line 215: I really like Figure 4. It is an excellent way to show at what conditions the different pathways dominate.

Line 231: “The J of HIO₃-HIO₂-DMA and HIO₃-HIO₂ in Mace Head are shown …”

To avoid misinterpreting this as actual measurements at Mace Head, please specify that these are simulations of conditions corresponding to Mace Head.

Section 3.3 – cluster formation rates: I understand the rationale behind Figure 5. However, I believe it would be worth to more clearly state in the text that this is just a mechanism, potentially one out of many, that increases the rates such that they match the observations. The measurements are essentially the sum of all possible nucleation pathways. This means that all possible nucleating precursor vapours contribute to the measured J-value. For instance, how would the results be influenced if your simulations included water, sulfuric acid or base synergy such as having both ammonia and DMA present? All these factors would yield clusters lower in free energy, increasing the cluster formation rates, and hence push the agreement further away from the observations.

Line 293-295: “However, considering the conditions of humidity in oceanic atmosphere and the complexity of marine NPF events, future research should investigate the role of water molecules and other crucial precursors to establish a comprehensive multi-component nucleation mechanism in the marine atmosphere.”

I believe this is a very important point, that should be mentioned and discussed much earlier and not just as an outline.

Technical corrections:

Line 12: derive -> drive

Line 13: broad marine regions -> various? marine regions

Line 51: Quelever -> Quéléver

Line 86: Kuerten should be Kürten.

Line 176: Kurten -> Kurtén. Please check the spelling of all Finnish authors in the references as many umlauts are missing.

The paper investigates the nucleation mechanism involving iodine oxoacids and dimethylamine (DMA). The authors suggest that DMA can enhance iodine oxoacid nucleation in ambient conditions and compare their results with ambient measurements.

The mechanism explored by the authors is new; therefore, it may be worthy of publication in ACP. However, the authors fail to support their statement that their mechanism is important in broad marine regions. None of the three field observations supports their statements. In fact, two out of three observations strongly oppose the idea that HIO3-HIO2-DMA is an important mechanism, while the last one shows great ambiguity.

A major revision is clearly needed. The authors should revise their manuscript so that ambient observations are correctly interpreted, and the implications of their results should be constrained properly before this manuscript can be accepted.

Major comments:

While the subject is of great interest considering recent ambient and laboratory studies of iodine nucleation, I am not convinced by the authors that DMA is essential for pristine marine aerosol nucleation processes. The authors mention two field measurements from Ireland and Antarctica sites. However, my brief glance over these studies suggests that the evidence from these sites is clearly against their hypothesis, i.e., DMA is not important in the nucleation events observed there.

In the Mace Head case (Sipila 2016), nucleation from iodine species was proposed by the authors to be the dominant mechanism there, and other nucleating species, e.g., SA, NH3, DMA, played a minor role there. In fact, the APi-TOF data did not even show the presence of NH3 and DMA. Therefore, the authors’ conclusion that iodine nucleation could not explain the observation is more likely a problem with the authors’ calculations. Additionally, the authors’ way of saying HIO3-HIO2 is not sufficient to explain the nucleation in Mace Head is rather flawed. The authors use an example case from Figure 1 of Sipila 2016. They adopted the corresponding nucleation rate as the bottom line of their range in Figure 5A. They further introduced an independent study (O’Dowd 2002) which reported an extraordinary 1e6 cm-3s-1 nucleation rate as the upper limit in Figure 5A. However, it is rather clear from Sipila 2016, Figure 3, that the example case represented the upper limit of the field observation. Therefore, the nucleation rate range is much lower than what the authors presented in their Figure 5A. In conclusion, their conclusion that HIO3-HIO2-DMA is important in the Mace Head station is not reasonable.

In the Aboa station, the authors (Jokinen 2018) also clearly suggested that the nucleation mechanism there was dominated by the SA-NH3 mechanism instead of iodine-dominated mechanisms. The authors measured little to no DMA-containing clusters in the APi-TOF, and therefore, HIO3-HIO2-DMA is not an important mechanism there.

The authors propose that the reason why DMA was not measured in Mace Head and Aboa was that the Nitrate-CIMS in these campaigns was not sensitive enough to measure DMA. However, they ignored the fact that the APi-TOF instrument was used in these campaigns and is extremely sensitive to DMA. The low (or no) DMA signals measured in these campaigns simply suggest that DMA was not important there, and the title and content of this paper exaggerate the implication of this manuscript.

The proposed mechanism, at maximum, could be important for polluted coastal environments where clean marine air masses meet DMA-containing polluted air masses. The authors also used an observation from a Chinese coastal site to imply the HIO3-HIO2-DMA nucleation mechanism is important. However, there appear to be no iodine measurements in the mentioned study (Yu 2019). Therefore, their statement that HIO3-HIO2 nucleation mechanism is not sufficient while HIO3-HIO2-DMA is, is not reasonable since we do not know the exact nucleation mechanism there in the first place. I strongly suggest the authors restrict the scope of their study and try to avoid overstatements.

Additionally, the scientific quality of this study is not high compared to the authors' previous studies and other similar studies. Most such studies will calculate the clusters with up to 8 monomers, while this study only calculated clusters with up to 5 monomers. How large are these clusters? Have they reached the critical size? I can understand the difficulty in getting clusters with 8 monomers for the three-component system, but 5 monomers are clearly not sufficient, and the authors are urged to discuss whether this can be improved.

Minor comments:

Line 11: Journal name should not be presented. The reference style should be consistent throughout the manuscript.

Line 12-13: Is there clear evidence supporting the author's claim that the mentioned mechanism cannot explain NPF in broad marine regions? Or is it just that it may be insufficient to explain all cases? It sounds like quite a strong statement the authors are trying to make.

Line 32: “are originated” → “originate”

Line 33: are thought

Line 34: coastal areas

Line 40: reference, how did Yu 2019 measure HIO2? Could the authors confirm this?

Line 43-45: The statement seems different than the definitive statement in the abstract. Could the authors clarify: is iodine nucleation not explaining broad marine NPF or is it sometimes insufficient? Do we know where iodine is important and when they are not? Are there field evidence?

Lines 45-46: I have a big problem with the statement that DMA is a common nucleation precursor in oceanic atmospheres. See major comments.

Line 136: references to tables S1-S4 appear to be missing. Should they be added somewhere? Or otherwise, reorder the tables to make S5 the S1.

Lines 150-159: Interesting observations here related to the proton transfer. Since iodine is a halogen, could the authors also comment on the role of halogen bonding in cluster formation? Is the halogen bonding important? Previous studies that the authors have cited (e.g., Zhang 2022) seem to point to a strong involvement of halogen bonds. Therefore, I encourage the authors to have a detailed session discussing the halogen bonds and a further session comparing the halogen bonds and hydrogen bonds, which are more important?

Additionally, the authors suggest that DMA competes for a proton from HIO2, giving the impression that they are separate things while in reality, it appears to be synergistic effects (Figure 4), instead of competition (i.e., adding DMA would never reduce the nucleation rates).

Lines 161-176: What is the nucleation rate simulated by this study for Figure 3? Does it agree with the field observation they refer to? Analyzing the branching ratio has to be based on a reasonable agreement between the field observation and their simulation.

Lines 173-176: The discussion here about DMA detection makes no sense. Even if the nitrate-CIMS did not measure DMA1, the APi-TOF deployed in the mentioned paper should have captured DMA. Have Sipila et al. (2016) measured DMA with their APi-TOF? Note, APi-TOF would capture DMA if there were over 5e5 molec. cm-3 levels of DMA.

Lines 255-257: Are there iodine measurements at the mentioned site? Where did the author get the concentrations? If there were no measurements of iodine species, how did the author derive the conclusion that the nucleation rates are too high to be explained?

Figure 5: When the authors calculate the HIO3-HIO2 rates, which they suggest citing from another study, did the author use the same size of clusters as the HIO3-HIO2-DMA system (5 monomers in this system)? This will also influence the nucleation rates of these systems.

Figure 5A: The field observation range is very misleading. How is it possible that the nucleation rate in the ambient reaches 1e6? Did the O’Dowd 2002 paper measure iodine species? If not, what is the acid and nucleation rate range in the mentioned Sipila 2016 paper? The authors should use a dataset that has both acid and nucleation rate measurements instead of assembling different datasets to prove their points. The authors mentioned one case from the Sipila 2016 with a 1e4 nucleation rate and 1e8 HIO3; it appears it agrees with their high-end HIO3-HIO2 simulation and low-end of HIO3-HIO2-DMA simulation. How can they conclude from this?

Figure 5: What are the shades in the figure? It is very confusing since I do not find an explanation for this. Please clarify. Additionally, the authors should compare their HIO3-HIO2 nucleation rate with the experimental work (He 2021?) to see whether their rates are reasonable before deriving any meaningful conclusions.

Lines 264-272: The original paper from Jokinen et al. 2018 (10.1126/sciadv.aat9744) clearly suggests that SA-NH3 was the nucleation mechanism there and did not measure DMA in the APi-TOF (Fig 2 of that study). This study's calculation has no relevance to the conditions there. Also, please cite the mentioned paper for the discussion here.

References:

(1)       Simon, M.; Heinritzi, M.; Herzog, S.; Leiminger, M.; Bianchi, F.; Praplan, A.; Dommen, J.; Curtius, J.; Kürten, A. Detection of Dimethylamine in the Low Pptv Range Using Nitrate Chemical Ionization Atmospheric Pressure Interface Time-of-Flight (CI-APi-TOF) Mass Spectrometry. Atmospheric Meas. Tech. 2016, 9 (5), 2135–2145. https://doi.org/10.5194/amt-9-2135-2016.