The manuscript presents the build and initial testing of an instrument to measure H₂ and Ne concentrations in ice core samples. The instrument was designed to deploy to the field and measure ice core samples immediately after drilling to minimise the time that the air bubbles in the cores can equilibrate with modern atmosphere with respect to the small diameter molecules H₂ and Ne, which readily diffuse through the ice matrix. The system is novel in that it is the first example (to my knowledge) of a field-ready work-around for this type of gas loss. Briefly, the system extracts the air from ice core air bubbles using a typical melt technique, then injects the air to a GC using an in-house built piston system that represents a clever work-around to the need for high-pressure samples from the small volume of air typically extracted from ice cores. The authors describe the design of the sample extraction device and piston injection, the calibration of measurements using custom in-house prepared standards, and the deployment of the instrument to the field site (Summit, Greenland) including ice core measurements.

This work represents an advance in the larger field of ice core gas analytics. The paper is well-written and quite straightforwardly presented. I think more information could be provided about long-term performance of the system (even if only on standard runs), as well as the calibration (I cannot tell from the text, for example, what is the range of compositions of the various prepared working reference gases used to calibrate the instrument). Besides needing to address minor and technical comments, I think this work should be published in AMT. I would be happy to review the paper again if necessary, though in my opinion this is probably not necessary assuming the below points are sufficiently addressed.

General comments

It would be helpful to see a photo of the whole instrument setup including the piston/ linear actuator device.

One of the stated goals of the instrument was to determine the rate of equilibration of ice core samples with modern air, but data toward this end are not shown. Are there preliminary results that can be included in this manuscript? If not, I recommend rephrasing this (line 42) so as not to build up the expectation for this paper.

It would be helpful to gauge system performance if more data were shown that characterise the long-term stability of the system. Figure 6 is useful for demonstrating the linearity and gives some indication of the range of total variation, but it would be useful to see, e.g., the linearity-corrected standard data over a long period of time. I’m a little unclear how the sensitivity correction was determined. Perhaps showing the drift in a figure would help the reader gauge the timescale and magnitude of the sensitivity drift.

Regarding calibration, the main text reads as if the three high-pressure cylinders had the same (or nearly the same) composition, though I think this is a mistake. Did you not measure standards of different compositions (especially standards with compositions like the ice core air) against one another to verify your calibration scheme? This seems necessary to me, especially given the lower concentrations of H₂ and Ne measured in the samples.

Specific comments

I may have just missed it, but please make sure the ice sample size is stated somewhere.

Line 38 This is a bit pedantic, but I’m not sure in situ is exactly appropriate here given the cores are first drilled and extracted from the ice sheet prior to measurement.

You should state what is the heating element of the oven in section 2.6. I didn’t realise it was an oven until line 196.

It would be helpful for the reader to see the temporal variability in the blank to judge for themselves how significant it is. Perhaps just show the spline fit and blank data (lines 233-237).

You might consider putting the calibration equations on Figure 6 instead of listing them in the main text in lines 243-260.

Line 261 – Does “calibration uncertainty” refer to the uncertainty in the standard mixing ratios? Please specify if so, and if not please also address this source of uncertainty.

Line 264 – I would change this to say, “The factors limiting system precision are different for H₂ versus Ne.”

In Figure 7, please state what the red circles mean in the caption.

Lines 267-268 There is only one sentence in section 2.4 that states the reasons for the detector limiting Ne precision (lines 168-169), so you might as well just say again here what those reasons are.

Line 272-273 Change to something like, “…, suggesting the atmospheric H₂ levels were constant within the measurement uncertainty over this period.”

Technical corrections

On the Figure 1 caption, also state what “MV” stands for.

Figure 1 and Figure 2 could be combined to save space if desired.

Line 99-100 Change to read, “This outgassing is associated with movement of the O-ring.”

Line 101 “…polymer due to compressive or shear…”

Line 103 “the O-ring did not outgas…”

Line 145 “This prevented(?) major constituents from air…”

Line 193 “…we constructed a tubular aluminum frame oven with insulated walls…”

Line 204 “…and evacuation to the vapor pressure of ice”

Figure 6 caption – “non-linear” rather than “non=linear”

Figure 7 caption – references to Mitchell et al. and Rhodes et al. should have proper formatting. The age range 150-1750 is different than what is stated in the text on line 271 (1600-1730 CE).

General comments:

The paper describes the setup of a sampling module, coupled to a gas chromatograph and a pulsed discharge detector, for the special purpose of extraction and analysis of Hydrogen and Neon from ice core samples. The aim is to develop an instrument that can be deployed directly at remote polar sites to minimize the time between drilling the ice core and analysis. The motivation is to avoid exchange of the targeted gases in the sample with today’s atmosphere. A particular focus is placed on describing the building of an extraction module with a custom-made pump, required for handling the small sample volumes from ice cores. Tests for characterising instrument performance and a calibration scheme are described. Measurements from the first successful field deployment are presented.

The underlying scientific question, i.e. to extend the atmospheric Hydrogen and Neon records to pre-industrial times to better understand the temporal development of these gases, is relevant and of future interest with the progression of a Hydrogen economy. This paper describes a specialized adaptation of a known basic instrumental setup, supplemented by newly developed parts such as the piston pump. The instrument is used for the generation of (to my knowledge) novel scientific data. Therefore, the concept, methods, and results are novel, and I recommend this article to be published in AMT.

The paper is written well and concise, with a clear structure. The instrument and method development are described in such a way that a reader can nicely follow and could reproduce the setup. Nevertheless, in my opinion, some parts need to be described in more detail. I will point these out in detail in the “Specific comments” section, however, to summarise briefly: My questions mainly relate to the materials used within the setup, as Hydrogen is quite material sensitive. Also, I think a flow chart of the GC-PDD would be helpful to follow the descriptions in the text. More background information needs to be added on the preparation of standards and on the implementation of the calibration. This is necessary to fully trace the entire process of deriving the sample concentrations. At last, I was wondering whether the instrumental drift was associated with the running time of the instrument, or could be characterized in any other way?

I consider my comments to be minor and think that the paper should be published once they have been sufficiently addressed.

Specific comments:

• Line 38: I think with “in situ” the authors intended to emphasize that the measurements are happening at site, so I would slightly rephrase to “determining H₂ and Ne levels in situ”
• Line 52: It is known that Hydrogen is very material sensitive, also to aluminium, which can be exposed if a Synflex line is slightly damaged at a union for example. Further down in the text, the authors also describe how the material of the piston seal matters regarding influences on the Hydrogen concentration in the sample. In various wetted parts of the setup non-stainless-steel materials are used. Did the authors make material choices based on tests they conducted, or based on references? I think it is worth emphasizing at a suitable place in the text (possibly in the section describing the blanks) why PFA, PTFE, PEEK, etc. are readily usable; or give some references where that was shown.
• Figure 1: Please make sure that the reader knows that “ultra-torr” means an “ultra-torr adapter”, like it is phrased in the text.
• Line 62: I have some difficulty understanding the sentence: “The lower portion of the chamber is a flask constructed from one side of a 75 mm borosilicate glass O-ring joint (Ace Glass, 7646-18) with the tube end closed.” What does “from one side” and “the tube end closed” mean? Could the authors please rephrase this sentence. Or, if this is a special technical term, then it would probably help if a real picture or some labels were added to the schematic.
• Line 63: If there is a specific reason for the choice of a ground flat O-ring joint, please briefly mention it.
• Line 65: “limit compression” in general, or on a specific part? If meant for a specific part, please mention, e.g. “limit compression on the O-ring”
• Line 66: To avoid confusion between the plastic rings and the O-ring, please write “The outer plastic ring….”.
• Line 96: This may be lack of knowledge on my side, but was a standardized test developed to check on the state of the seal or on which criterium is it replaced every 1-2 months?
• Line 103: Please quickly check that when speaking of permeation rates in polymers, you really mean polymers in general, and not the polymer types that were tested in this study (as PTFE is also a polymer, but this was used in the setup).
• Line 142: Maybe change the title to something including the detector as well.
• Line 142: I think a flow schematic of the GC-PDD setup (including getters, flow controllers, valves, etc.) would be really helpful (if there is still allowance for another display item).
• Line 143: A 10 port two-position valve?
• Line 145: High purity helium of grade 5 or 6?
• Line 145: How many getters of each model?
• Line 146: Were the flow controllers already described before? Where are they located?
• Line 152/ Line 154: Do I understand it right, that water vapour and CO₂ do not even reach the main column, but O₂, N₂ and Ar do? If so, then please specify, e.g. “and other high boiling compounds from reaching the main column and the detector”.
• Line 152: Was it tested how much water vapour is still left in the sample after Nafion drying? Would this have an influence on the targeted gases in the sample?
• Line 156: This is not a two-position valve, correct? Is this also heated?
• Line 157: Please specify “for roughly 3.5 minutes (minute xxx-xyz through the run) during elution…”
• Line 159: Is it ever mentioned somewhere at what temperature the GC oven and the PDD are operated?
• Line 160: Is the temperature controller for monitoring and setting the PDD temperature? Is it not done via the GC?
• Line 166: “orders of magnitude”, including the range of atmospheric H₂ abundance?
• Line 174 & Line 183: Hydrogen can drift depending on the cylinder material and cylinder batch. Was there some sort of longer-term stability testing of the used cylinders or comparison to a sample filled into another type of cylinder to exclude drift in the cylinders? Or could you exclude simultaneous drift in the three cylinders with the comparisons you made and this question I already covered in Line 183?
• Line 176: Do I understand it right, that you prepared your own scale? Or is it planned to reference (at least Hydrogen) to an external scale like the WMO scale? I am not sure if an established scale exists for Neon. Is this preparation of calibration cylinders a refined procedure in your lab (e.g. at the level of a metrological institute) and is there a literature reference that describes the details? What was taken into account to derive the 2% uncertainty? I think this section either needs referencing or should be described in a bit more detail.
• Line 179: What were the resulting concentrations in the working standards? How many working standards did you produce and use? Same question for Line 238 with the part of the sentence saying “several concentrations”. Please specify at least in one of these text passages.
• Line 180: Did you see a difference whether N₂ or zero air is used as the balance gas, or does it not matter?
• Line 183: Were working standards analyzed once or multiple times daily? Please specify in the text (or in Line 238). How was the necessary frequency of working standard analysis assessed?
• Line 183: I assume that the weekly preparation of working standards has to do with limitations at this remote site? If there was a method-related reason, please specify in the text.
• Line 187: connected to the “vacuum line” (I presume)
• Figure 4: Please also mention the calibrated volume briefly in the text – or is this meant by the stainless-steel vacuum line?
• Line 198: What was the eventual temperature range that the aluminium frame box could create?
• Line 212: I understand that the GC and detector were always restarted anew before an analysis batch? The stabilization of a GC-PDD can take some time after restart. Did you observe this issue? In case yes, I assume the GC and the detector were already running some test runs before the actual sample analysis, to have the baseline stabilised?
• Line 218: The level of detector output/ baseline varies over the runs. Is this the drift you are addressing later in the text and that needed correcting? What is the reason for this, i.e. is there a connection with the operating time of the instrument, or parameters like temperature or pressure, or is it a random pattern?
• Line 240-259: I am not familiar with this procedure for calibrating measurements. I presume the purpose of all these steps is to, firstly, establish a calibration curve and secondly, to apply a drift correction. Please specify this at the start of the description of the steps, e.g. in Line 239. Also, if possible, please reference to respective literature for this data treatment method.
• Line 245: If injection loop temperature should actually also be accounted for, why was it not included, even if it did not improve or even worsened the regression statistics? If including it did not have a large effect on the statistics, then it could still have been included, or the sentence should at least be rephrased saying “did not change the regression statistics” or similar.
• Line 252: I have already commented on this above, but could this daily variation of residuals be an artefact of restarting the instrument every day? Or is the instrument running continuously, or is it in idle over night?

Technical corrections:

• Line 8: the abbreviation needs to be “He-PDD”, not “He-PPD”
• Line 101: a “to” is missing, “due to compressive”
• Line 103: a “the” is missing, “because the O-ring”
• Line 192: a “to” is missing, “due to”
• Line 204: an “of” is missing, “of the ice”
• Line 246: I think “Ne” has to be “neon”, as this is used throughout the text and to make the naming consistent; please check for consistent naming throughout the manuscript.
• Line 248: I cannot find section 2.1.3, please correct the number in the cross-reference.
• Figure 7: Maybe I missed it, but what are the red circles in the plot? Please add description to caption.
• Figure 7: I think the depth range “75-150 m” in the caption does not match the x-axis of the plot.
• Figure 7: The gas age range “150-1750” in the caption does not match the age range in Line 271.