Molecular Formation in Low-Metallicity Hot Cores
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsRevision of English language is required
Comments on the Quality of English LanguageReferee report on the paper "Molecular Formation in Low-Metallicity
Hot Cores" by Y. Sobhy, H. Nomura, T. Yamamoto, and O. Shalabeia
This paper presents a detailed study of the chemistry of low-metallicity
hot cores, and a comparison with abundances measured in several hot
cores in the Large Magellanic Cloud (LMC), a satellite galaxy of the Milky
Way with an average metallicity about a factor of 2 lower than that
of our Galaxy.
The main originality of this study is a quantitative evaluation of
the effect of cosmic-ray ionization on the chemical abundances.
This is especially important, because the LMC is characterized by
a lower density of high-energy CRs (those producing gamma-ray
emission) than the local ISM in the Milky Way. It is therefore
interesting to check whether this deficiency also extends to lower
energy CRs (those affecting the chemistry via ionization). According
to the author's conclusion, no reduction of the CR ionization rate
with respect to Galactic values is required to match the LMC abundance
data. Rather, the best-fit value of the CR ionization rate is a
factor of about 10 larger than the standard value adopted in most
chemical networks.
These are interesting results, however I have some major concerns
about this paper. The first is the lack of information about the
adopted radiation field and its effects on the chemical evolution.
What is the assumed UV flux in the model? What is the impact of
photochemistry on the abundances?
Another concern is the rapid time evolution of methanol and other
O-bearing COMs that disappear in a short time, contrary to observations
in hot cores. There is no explanation of this behaviour, except a
comment on the effect of the stationary conditions assumed in the
model. Is this a general result? As far as I know this is in
contrast with previous investigations, as e.g. in Barger & Garrod
(2020, ApJ, 888, 38). The authors should clarify the reasons for
the disappearance of COMs.
Minor points:
1) On p. 3, line 100, the statement that "the globally averaged CR
ionization rate in the LMC could be as low as 1e-18 s-1" is attributed
to ref. [22]. However, as far as I can tell, in that reference the
CR ionization rate is never mentioned.
2) Numbers and letters in the figures are very small, unreadable
in the printed copy, and only readable on the screen after considerable
zooming-in. I suggest to increase the font size in all figures.
3) in the bibliograpy, the names of all journals are missing (perhaps
a machine-dependent printing issue)
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe manuscript, “Molecular Formation in Low-Metallicity Hot Cores” aims to investigate the role of the density, temperature and cosmic ray ionization rate (CRIR) on the chemistry of hot cores in low(er) metallicity environments such as the LMC. The main result presented in their manuscript is that the CRIR is crucial for models of the chemistry in the low environments. Chemical models of lower metallicity environments will be of high interest to the community as JWST and ALMA continue to explore the molecular chemistry of the interstellar medium in galaxies across cosmological time. However, I have several key points which must be addressed to make the manuscript sufficient for publication, especially since they are concerns regarding the underlying model itself.
1. First, the most crucial part is the underlying chemical model. In fact, with regard to their actual chemical model, they only state that they “used the gas-phase chemical network of the dark cloud model…”. But, the authors are not modelling a dark cloud, they are modelling a hot core. There are multiple public, robust gas-grain astrochemical codes which have successfully been used to model hot core environments, such as Nautilus and UCLCHEM. Using a free-fall one zone models that is common in, e.g. UCLCHEM models, one would be able to simulate the full gas-grain chemistry followed by the heat up. Performing the one-zone calculations including the freezing and following heating and evaporation would ensure an accurate and complete model. The problems with matching observations for the complex organics may likely be the lack of any gas-grain chemical model and adhoc initial conditions. The simulations were evolved for 1E8 years, which is far longer than the age of any star forming region. Chemical models which do this are typically aiming for a steady state model (as is the case of PDR calculations). Reducing this integration time for a reasonable age of cores would enable more models to be run.
2. Second, the initial conditions seem problematic. They appear to use the initial conditions of the 1D hot core model from Nomura and Millar 2004 and scale it. The authors utilise a simple linear scaling of the atomic elements between the MW and the LMC and multiply the initial assumed atomic and molecular abundances each species by its composition. I have not personally seen this type of setup before, and it would need to be demonstrated that decreasing metallicity scales the molecular abundances in this way. Then, the authors state they add the abundances of certain molecules, such as CH3OCH3 in ways comparable to observations (in fact, they later state “If we do not input HNCO and CH3CN initially, the model calculations do not reproduce the observational abundance as shown in Fig. 5.”). This is not a correct way of addressing the initial conditions: the goal of the model is to predict the observed abundances, and therefore one cannot use the observed abundances of a predicted molecule as it’s initial condition. Unless the chemical model were initialised in chemical equilibrium (and thus already solved), the abundance would drift away from what is wanted to be predicted. It appears they use this initial abundance model since they only use a gas-phase network but the initial conditions follow chemistry after ice evaporation. I would be hesitant in doing such, as this is changing the underlying chemical network and thus the evolutionary model.
The models they perform are 0D. However, the authors in principle could perform a 1D calculation with the same physical setup as the Nomura and Millar 2004 study but for a lower metallicity and modern gas-grain chemical models. This may be a future planned work. For using 0D models, one can perform a broader and more finely grained parameter search since 0D models run very quickly.
3. Third, the main conclusion is somewhat self evident since they are doing warm (but not hot) gas-phase chemistry, but do not include any sources of FUV. At their warmer temperatures, ion-neutral reactions are likely still playing a significant role. Therefore, one would already natively expect that orders of magnitude change in the CRIR is more noticeable than factors of few in temperature.
Minor notes:
1. The authors refer many times to the CRIR. However, I could not find anywhere in the paper which CRIR they were referring to, especially for the values they adopt in the variations. Is this the primary proton CRIR? The hydrogen or molecular hydrogen total ionization rate? There should be clarity on this, as the chemical network they use will likely be using one of these definitions.
2. In their Figures 1, 2 and 3, there are mixes of models for different molecules when the plots are supposedly showing the variation in one parameter. For instance, in Figure 1, it is supposed to be highlighting the variation with the CRIR. However, the panels for SO,OCS and H2CS are for a density of 2E5 and temperature of 50 K; NH3, NO, CH3OH, CH3OCH3, HCOOCH3 are for a density of 2E7 and temperature of 200 K, and CN is for a density of 2E5 and a temperature of 200 K. These types of inconsistencies are also in Figure 2 and 3. It is not clear why in plots meant to highlight the variation in one parameter, they are showing models across the parameter space.
3. The authors state that one reason for the higher ionization rate than the Fermi would be a SNR in the neighbourhood of the cores. It may benefit the results of their models for the authors to check if there are observed indicates of a SNR near the core.
4. The authors state that H2S is destroyed by reaction with H, helping to initiate the Sulfur chemistry. Using the KIDA interface, for the reaction H2S + H -> H2 + HS, the reaction rate at 50 K is O(1E-19) and at 200 K O(1E-13). The authors may want to double check for their colder models if this is still the main route for Sulfur chemistry.
Comments on the Quality of English LanguageEnglish language itself in the text is fine in general.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsThe adopted visual extinction of 640 mag. corresponds to a column density of about 10^24 cm^-2. Since low-energy CRs are attenuated by this column density by one/two orders of magnitude (see e.g. Padovani et al. 2009, A&A, 501, 619 or Fig. B1 in Padovani et al. 2024, A&A, 682, A131), the authors should perhaps note that the CR ionization rates of their models imply a much higher galactic CR ionization rate, possibly by one or two orders of magnitude, in more diffuse and less extincted environments.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsI thank the authors for their replies to my concerns on their manuscript. While many of the comments I feel have been adequately and thoughtfully addressed, I still have a concern with the consistently on the initial conditions which can hopefully be readily remedied.
In their replies to my initial comments, the authors state “The goal of this paper is not to reproduce the observed abundances, but to understand the physical and chemical conditions which reproduce the observations.” In order to accomplish the latter goal must do the former (how can you understand the physical and chemical conditions which reproduce the observations if your underlying model does not reproduce observed chemistry - or if it doesn’t, a big question is then why). It seems then counter-intuitive to me for either purpose to choose, for some of the molecules, HNCO, CH3CN, CH3OCH3 and HCOOCH3, to set the initial condition to be the observed value. As such, some initial values come from an assumed initial condition model, while the above molecule initial conditions are coming from the latter evolution time the chemical modelling is seeking to match. The chi-squared fitting for finding the best chemical model is comparing the output of the chemical model evolution to the observations, but some of the species thus start from that observed value instead. Chemical networks are inherently non-linear, so the initial conditions used can matter greatly for the evolution. Indeed, they state their chemical model - for species not in the list above - that best fits the observed values is at 1E5 yrs, but the above molecule abundances are initiated at t = 0 to the observed values. Are these species not in the chemical model that was used for the initial condition? If not, I would recommend using initial models which have all of the species that you are interested in or need for the chemical evolution. It may be that the code or network being used is just not the correct model for the complex chemistry in hot cores.
Minor:
- The authors state at line 148 “Fig. 1 shows how the abundances of different molecules are affected by varying the CRIRs with keeping the values of temperature and density constant”. All their models keep the temperature and density constant with time, so one would assume this means that all of the subplots are the same density and temperature, but they are not. In their replies, they stated that they were changed “in order to reproduce the observed molecular abundances best”. Please change the text in the manuscript or the figures to make them constituent.
- The authors state they use the “gas-phase chemical network of the dark cloud model”. So, are the authors utilizing the dark cloud code (downloadable on the site), or are they using the base gas-phase network from rate12 in another code?
- As a general comment, I believe the paper could be more impactful if more discussion was given on the differences in the chemistry itself of the LMC hot cores. The authors seem to focus in the abstract and elsewhere on the best fit value from their coarse grid. However, a thoughtful discussion of the differences in the chemistry itself would, I believe, provide a much higher impact result for the astrochemical community. Do these molecules go through similar formation pathways? Do new pathways open up at these metallicities? The answers to these types of questions in detail would fall in line with the stated goals, and would be of great importance with interpreting newer observations.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 3
Reviewer 1 Report
Comments and Suggestions for AuthorsThe authors have replied to all my questions and added several clarifications. In my opinion the paper can be published in its present form.
Reviewer 2 Report
Comments and Suggestions for AuthorsI thank the authors for their response. While I still disagree a bit with the choice of initial conditions for the complex molecules, as they stated in their latest response, it does not significantly impact their final conclusions. Since the paper is not focused on the complex molecules, I think it is fine then.
