The manuscript entitled “Validation of a new cavity ring-down spectrometer for measuring tropospheric gaseous hydrogen chloride” by Furlani et al. discusses the implementation of a commercial spectroscopic HCl instrument for ambient air measurements. The study indicates that the Picarro HCl CRDS instrument can be implemented with low detection limits, high precision, high time resolution, and similar/better accuracy compared to a cadre of other trace HCl measurement techniques. Discussion of sampling and analytical issues is thorough and thoughtful. Statistical treatment and reporting is adequate. Studies of inlet effects is detailed and provides important information for potential users in the future.

This paper is high quality and should be published in Atmospheric Measurement Techniques pending consideration of a small number of comments provided below.

Line 229 and Figure 1: Please provide a reference (within the associated text) to the approach of using the Allan-Werle deviation.

Lines 246-248: “spectroscopic techniques offer a distinct advantage as they are absolute measurements and accuracy determinations rely on propagating uncertainty in measured parameters” This statement is then met with the reality of the present study [in subsequent lines of this paragraph] in which concurrent denuder measurements are used to determine accuracy. Why not stand upon the ‘distinct advantage’ that spectroscopic measurements provide? [See also: next comment.]

Lines 248-251: “In the absence of determining accuracy of the CRDS system from its operating parameters…” Is this due to lacking knowledge of these parameters due to the commercial nature of the instrument or are the intercomparison differences seen to be greater in importance, greater in magnitude, or both? Overall -- please clarify the motivation for the accuracy determination approach that was undertaken.

Calibration questions: What is the impact of uncertainty in the permeation tube HCl concentration on the mixing ratio determination by the CRDS instrument? For instance, why aren’t the error bars in the horizontal direction in Figure 2 comparable to those in the vertical direction? How can a measurement have lower uncertainty than the standard from which it was calibrated? Please clarify this issue. Perhaps I am confused about details of the study, but so too may be a future reader.

Figures 2 and onward: Numbering of figures in textual references is often (always?) incorrect.

Line 362: Please clarify whether the effect of sampling line flowrate on τ₁ was investigated experimentally or if this point is a supposition.

Line 363: Change to “An additional set of experiments was…”

In this paper, the authors present a CRDS for HCl and evaluate its performance with an appropriate set of laboratory evaluations. They also report measurements of ambient HCl and compare these with another measurement technique; this comparison is done extremely well.

The findings are within the scope of AMT, substantive conclusions are reached about the instrument performance and field comparison, and the paper is well written. The instrument itself is not particularly novel since an extremely similar instrument was presented in AMT almost a decade ago (Hagen et al., 2014). This instrument does seem better than that one in some ways (mainly in the 30 s precision of 6 pptv, which is quite impressive), so I suppose publishing on this instrument is justified.

Still, there are some pieces that are missing or need revision. I discuss these in ‘Broader Comments.’ I’m calling these minor, but they do approach the threshold between ‘minor’ and ‘major’ revisions. I also have some decidedly minor suggestions, which follow as ‘Specific Comments.’ I recommend accepting for publication after these aspects are addressed.

Broader Comments:

1) The introduction should have a slighty more comprehensive discussion of HCl measurement techniques. There have been many cavity-enhanced HCl instruments developed in the past decade or so that are not mentioned here but should be. Los Gatos Research and Tiger Optics have already developed commercial, cavity-enhanced HCl spectrometers. Hagen et al., 2014 and Wilkerson et al., 2021 have also published on cavity-enhanced HCl instruments. Hagen’s instrument is even a CRDS that measures HCl via its first overtone—the same as the instrument discussed here. So you should also explain how this instrument compares and contrasts with Hagen’s since they are so similar.

2) There is no CRDS instrument schematic in this CRDS instrument paper. I understand that proprietary issues may limit how specific this can be, but please include at least a rough schematic of the instrument layout as either a main figure or supplemental figure.

3) You state that tropospheric HCl mixing ratios vary between 10-1000 pptv. Yet, the data in Fig. 2 extends all the way up to 45 ppbv HCl and doesn’t go below 10 ppbv HCl. You really need to be evaluating these effects closer to relevant, atmospheric levels of HCl. Ideally, you would dip below ppbv levels since those are the most relevant (as you further demonstrate in Fig. 3). Hagen et al., 2014 goes slightly below 1 ppbv HCl when they produced their correlation line (Fig. 8); they proposed an instrument extremely similar to this one and published that discussion in this very same journal (same with Wilkerson et al., 2021). There is, therefore, a clear precedent for a more rigorous evaluation than this. Having said that, I appreciate that laboratory assessments of sub-ppbv levels for HCl can get quite difficult. At bare minimum, you should remove the data above 30 ppbv and add at least two iterations of this experiment below 10 ppbv HCl.

Specific Comments:

Line 24: Recent reports of response times for HCl instruments are usually presented as 90% response times. Is that the case with this range? If not, please change this to a 90% response time range, so your instrument can be more easily compared with other recently reported ones (or at least explicitly call your current range a X% response time, where X is whatever this reported range corresponds to).

Line 29: You should probably drop ‘response time’ from this list. First, a response time of 2-6 minutes is quite worse than recent HCl instruments discussed in scientific literature (Hagen et al., 2014; Wilkerson et al., 2021) and a bit worse than commercial HCl instruments (e.g. LGR’s HCl analyzer is ~1 minute: http://www.lgrinc.com/documents/HCl_Datasheet2019.pdf). Second, you don’t display response times in Table 1 where evidence of a favorable response time compared to other instruments would presumably be presented to the reader.

Line 49: Add just a couple sentences to explain where you would expect to see atmospheric HCl levels of 10 and 1000 pptv. For example, my understanding is that atmospheric HCl is elevated near oceans because sea salt spray provides a large source of chlorine to the atmosphere. (And if I’m wrong, then that’s more evidence to support adding a couple sentences to briefly explain this to readers).

Line 118: The first overtone has a very low intensity compared to the fundamental transition for HCl, and the fundamental transition is by far the most common transition used in HCl spectrometers (there are also ro-vibrational lines in the fundamental transition that are spectroscopically isolated from other atmospherically relevant species like H2O). You should clarify the language ‘relatively high intensity’ so the reader knows what this intensity is relative to.

Line 234: The 30-second precision of 6 pptv is great—surprisingly great. Hagen et al. 2014 discussed a CRDS HCl instrument that also engaged a line in HCl’s first overtone. This is pretty similar to your setup. Yet, their 60 s precision is 20 pptv. You should include a compelling explanation in this paper as to how your precision is so much better than any other HCl instrument, especially one that is so similar to yours.

Line 234: Your response time is 2-6 minutes, but the precisions you report skip from 30 seconds to 1 hour. Can you provide a precision for an averaging time that is within range of or only slightly above your response time (e.g. 5 minutes or 10 minutes)?

Line 240: Change to, “This CRDS has many advantages compared to…” since you are comparing this instrument to another CRDS in Table 1.

Line 253: Again, this response time is not that fast compared to many recent HCl instruments (see earlier comment). Can you better contextualize this assertion?

Line 261: Why does the Iodide Cl-HR-ToF-MS have a higher LOD (30 pptv) than precision (53.3 pptv)?

Line 261: You allude to two recent cavity-enhanced HCl instruments in the main text that were published on in AMT (Hagen et al., 2014; Wilkerson et al., 2021), but only one is present in the table. There are also other commercial, cavity-enhanced instruments beside the one in this paper (LGR and Tiger Optics). If possible, can you add these missing ones to your table? (I say 'If possible' because their descriptions may be missing too much information to warrant inclusion).

Line 261: Hagen et al., 2014 has an accuracy <10%, not >10% (>10% reads as if the accuracy is at best 10% but could be infinitely worse). I believe that the Tandem mist chamber and IC-CD is the same situation (>25% when it should be <25%). Please correct these or just report the number without a ‘<’ sign.

Line 304: I think you mean to refer to Figure 3a.

Line 425: Some of these references successfully used a silicon coating on their instrument, as you state on Line 422. Yet, your focus for the rest of this paragraph is on all the unpleasantness associated with fluorinated materials. What’s your evaluation of silicon coatings, which lack all of these disadvantages? Please include that (even if it’s just a few sentences).