Technical Note: Stable Chloride Measurements at CAMS

Anderson, Tyler S.; Hidy, Alan

doi:10.5194/egusphere-2026-4091

Preprints

https://doi.org/10.5194/egusphere-2026-4091

Preprints

27 Jul 2026

| 27 Jul 2026

Technical Note: Stable Chloride Measurements at CAMS

Tyler S. Anderson and Alan Hidy

Abstract. We present a new method for preparing silicate samples for stable chloride analysis using the low-energy injection components of an accelerator mass spectrometry (AMS) system. This method uses Nb over AgBr as a bulking agent during sample preparation which drastically reduces background Cl signals (ion source memory) by a factor of ~100 and allows for faster, cheaper, and more precise analyses on low-Cl samples (<10 μg Cl). We demonstrate this approach with measurements of stable chlorine isotope ratios (³⁵Cl/³⁷Cl) performed on silicates, prepared in both AgBr and Nb matrices, that characterize chlorine concentrations of samples containing ~1–500 ppm Cl. While this method was developed primarily to interrogate silicates, it performs equally well on other commonly analyzed sample matrices including carbonates and water. This Nb-based method is now the standard procedure at the Center for Accelerator Mass Spectrometry (CAMS) for preparing all stable Cl targets, including those accompanying ³⁶Cl analyses.

Received: 11 Jul 2026 – Discussion started: 27 Jul 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Tyler S. Anderson and Alan Hidy

Status: final response (author comments only)

RC1:
'Comment on egusphere-2026-4091', Anonymous Referee #1, 09 Sep 2026

The manuscript “Technical Note: Stable Chloride Measurements at CAMS” describes new sample preparation method and mass spectrometric approach to measure stable Cl concentration. The measurement of stable Cl concentration via isotope dilution is somewhat finicky task and advances in the field are warmly welcomed. Introduction of Nb matrix instead of commonly used AgBr seem to have merit, alas it comes with a cost. I’m supportive of the manuscript but have few concerns and recommendations for improvement.
First, and this is mostly a matter for the editor to decide, the benefits of the manuscript lie in the improvement of the low concentration Cl measurements. This in turn has very limited impact on exposure age dating with 36Cl given improving the measurement precision and/or accuracy of low Cl concentrations have very little impact on 36Cl derived ages. That is, if the Cl concentration is low there is very small contribution of 36Cl being produced via neutron capture on 35Cl, and uncertainties in the number of target nuclides have no bearing on final 36Cl ages. As such relevance of the manuscript to Geochronology is peripheral at best. Having said this I’m supportive of the manuscript but think that it would be better suited to more mass spectrometry focused journal, where more detailed discussions of the background corrections and uncertainty analysis would be at home also.
My second overarching criticism is in the style of the manuscript, and again this might be a matter for the editor. That is, currently I feel the manuscript reads more like laboratory notebook rather than self-contained journal article. For example, sections called ‘Experiment 1’ and ‘Experiment 2’ might be unfortunate hindsight, but they also imply that ‘Experiment 3’ and ‘Experiment 4’ will be reported next month. Whilst titled as ‘Technical note:’ I still feel the authors could have been more complete in their presentation, for example they could have used high-energy measurement of 35/37 ratios that are free of molecular interferences, and/or mass scans to identify molecular fragments, and thus show sufficient removal of fluorides that justifies use of low energy measurement in the future. I understand that high-energy measurement might not be desired as end goal as it might be less stable than low-energy measurement, but it can still be used for diagnostic purposes.
My last general comment refers to laboratory jargon. The descriptions of the measurement sequence/process are currently difficult to follow. I understand this is common feature as each laboratory develops its own language and terms that might be hard for others to follow. For example, my understanding of the measurement process is that each sample is measured once, and this measurement is artificially divided into 15 internal subdivisions that are then used to calculate the internal spread and standard deviation for the that measurement. The final error is then elevated if multiple measurements on standards have higher standard deviation to this value. Either way consider simplifying the terms and being clearer on the process.
I believe figures would benefit from few improvements:
Figure 1 and the associated discussions. Figure shows increasing 37Cl- current as a function of the cathode voltage from 50ug sample without specifying if this is AgBr or Nb target, presumably latter but it is pertinent to mention this at someplace. Operational limits are not demonstrated by this data, I don’t doubt them, just wondering the value of this figure. It is well known that the current output typically increases as a function of cathode voltage in Cs sputter sources, see e.g. early work by Middleton et al. Is the response similar/different to AgBr targets, what if one increases Cs delivery, or ioniser power? Presumably what is important here is the signal to noise or more specifically signal to source memory, which is discussed later. Does AgBr and Nb need to be run at the same ion source settings? Can the figure be expanded to contain more general conclusion, e.g. 5 times increase in Cl- output with Nb vs AgBr?
Figure 2: I have real trouble understanding this figure. Please elaborate and/or device alternative method of demonstrating the advantage of the reduced memory. Currently both x and y axis are ambiguous: integrated current – over what time/samples, acquisition sequence – what does this mean in terms of time and samples measured? What are the spikes in the graphs? Why does the 37Cl- with Nb integrated charge never go below 1 nC? Note: my understanding of this figure improved after I finished the entire manuscript but presume other readers would have similar challenges, so I left this comment as is. I would recommend just to make a simple plot of source memory 35 or 37 current as a function of time after running Cl targets with both AgBr and Nb and then repeat with reverse order to make sure build up from Nb is not reason for elevated memory from AgBr.
Figure 3, 6, 7: all the mass scans. I’d recommend using log scale so that the achieved interference suppression becomes visible.
Figure 5: not sure this is needed. It is basic and can easily be referenced.
Figure 9: introduce small shift between multiple samples on x-axis to make errors visible.
Specific comments:
-Paragraph lines 26-31: gives the impression that 10-100 mg of AgCl is still needed for measurement. Mention that samples with Cl masses around 1mg are routinely measured for 36Cl.
-Paragraph between 32-46 currently oversells. Measuring stable Cl from separate vs combined with 36Cl targets has a long history and each have their benefits; there doesn’t have to be ultimate answer here either. Also, ANU AMS laboratory has been measuring stable Cl form separate targets for a long time, I suggest a reference would be appropriate here. Similarly note such as “significantly reduces the amount of expensive isotopically enriched Cl carrier” is little misleading in the context. That is, I agree with the statement, but feel that currently the reader is misled by implying from the above paragraph that close to 100mg of AgCl is needed, or that pure 37Cl carrier is needed, when ~1mg of 35Cl enriched carrier suffices. Then the sentence “This method provides additional flexibility because the carrier can be enriched in either 35Cl or 37Cl, …” is simply not true. I routinely measure combined stable Cl and 36Cl samples that are enriched with 35Cl without significant 36Cl background from the 35Cl enriched carrier, there is no difference here.
-Line 87: “Collectively, these early findings demonstrated the potential benefits from using Nb as a matrix for stable Cl isotope analyses of silicates.” So far there have been no evidence or mention of why the method would be beneficial for silicates, all the presented evidence is simply about Cl currents and source memory and applies to any Cl measurement, not just silicates.
-Line 140: “centrifuging us referred” typo, should be is.
-Around lines 200-215 there is mention of ‘pass’ few times. Consider if this is needed. The magnet bounce cycle is given, and one could simply just state typical measurement time of X min instead of introducing the term ‘pass’.
-Line 233: “Measured standard ratios are typically ~0.3% higher than the natural ratio, resulting in Fnorm of ~0.997.” Please state what the normalisation factor is for pure enriched carrier samples to validate that the instrumental fractionation is constant between endmembers.
-Line 235: “Method one calculates the intra-target standard deviation of the 15 individual analyses that make up one measurement ‘pass’” This sounds like the sample is measured once and arbitrarily divided into 15 units that are then used to calculate standard deviation. The measurement could be equally divided into any number of units. This approach is fine for estimating the stability of a single measurement, but the 15 units shouldn’t be considered as independent analyses.
In summary, despite my lengthy and perhaps somewhat critical comments I’m highly supportive of the manuscript and would like to see the work published as it offers a new approach to stable Cl measurements with increased sensitivity, that would have a benefit for the community. This work has potential to enable new studies into stable Cl concentration and/or fractionation. Many of my comments, I believe, are rather easy to address, e.g. improving the figure clarity, and will improve readability of the manuscript. My recommendation would be to consider more mass spectrometry specific journal but leave this up to the editor and authors to decide.
Best regards.

Citation: https://doi.org/10.5194/egusphere-2026-4091-RC1
- AC1: 'Reply on RC1', Tyler S Anderson, 17 Oct 2026
  
  We thank the reviewer for their detailed and constructive comments which will lead to significant improvements in clarity and scientific quality. We will revise the manuscript to include your comments, which are reproduced below in italics followed by our responses to each point.
  First, and this is mostly a matter for the editor to decide, the benefits of the manuscript lie in the improvement of the low concentration Cl measurements. This in turn has very limited impact on exposure age dating with 36Cl given improving the measurement precision and/or accuracy of low Cl concentrations have very little impact on 36Cl derived ages. That is, if the Cl concentration is low there is very small contribution of 36Cl being produced via neutron capture on 35Cl, and uncertainties in the number of target nuclides have no bearing on final 36Cl ages. As such relevance of the manuscript to Geochronology is peripheral at best. Having said this I’m supportive of the manuscript but think that it would be better suited to more mass spectrometry focused journal, where more detailed discussions of the background corrections and uncertainty analysis would be at home also.
  We appreciate this perspective on our manuscript’s relevance to Geochronology. We also agree that the improved precision in low-Cl concentration will often have little impact on conventional exposure ages where neutron capture on 35Cl contributes only a small fraction of total 36Cl production. However, the practical benefits of the method extend beyond improving exposure age precision.
  An important distinction is between low Cl concentration and low total Cl mass. The ability to measure smaller quantities of native Cl allows concentration measurements on smaller subaliquots, preserving material for subsequent 36Cl analysis when sample abundance is limited. Stable Cl concentration also remains necessary for characterizing the neutron capture contribution, even when that contribution proves negligible. Measuring it beforehand also supports sample screening and selection of an appropriate mass for 36Cl extraction. These benefits build directly on the separate aliquot workflow descripted by Lesnek et al. (2024)--published in Geochronology.
  This method also has relevance to emerging burial-dating applications using 36Cl (e.g., Balco et al. 2019—published in Geochronology). At depths sufficient to suppress spallogenic production, 36Cl production via capture of locally generated radiogenic neutrons continues while the inherited cosmogenic inventory decays away. Consequently, this contribution can become increasingly important during burial, even in samples with low Cl concentration. Accurate low-Cl measurements therefore help constrain a component of the burial-age calculation whose relative importance may be substantially greater than in surface exposure dating. We will highlight this benefit in the revised version of the manuscript.
  Our choice of Geochronology therefore reflects the method’s direct application to an established 36Cl dating workflow and its relevance to researchers already using that workflow. We appreciate that these connections could be stated more clearly and will strengthen their treatment in the manuscript.
  My second overarching criticism is in the style of the manuscript, and again this might be a matter for the editor. That is, currently I feel the manuscript reads more like laboratory notebook rather than self-contained journal article. For example, sections called ‘Experiment 1’ and ‘Experiment 2’ might be unfortunate hindsight, but they also imply that ‘Experiment 3’ and ‘Experiment 4’ will be reported next month. Whilst titled as ‘Technical note:’ I still feel the authors could have been more complete in their presentation, for example they could have used high-energy measurement of 35/37 ratios that are free of molecular interferences, and/or mass scans to identify molecular fragments, and thus show sufficient removal of fluorides that justifies use of low energy measurement in the future. I understand that high-energy measurement might not be desired as end goal as it might be less stable than low-energy measurement, but it can still be used for diagnostic purposes.
  We thank the reviewer for these suggestions concerning presentation and analytical validation. The numbered experiment labels were intended as shorthand, but we agree that descriptive subsections headings would better communicate the structure and purpose of the comparisons more effectively. We will replace these labels throughout the manuscript and related captions and clarify how the comparisons collectively evaluate the preparation method.
  We also agree that high energy 35Cl/37Cl measurements could provide a useful independent check for molecular interference. We instead used diagnostics available in the low energy configuration we employ for routine stable Cl analysis. Specifically, the mass scans in Figs. 6 and 7 show substantial reduction of 19-F^-following the revised prep procedure. Additionally, the HF carrier blank in Fig. 6 yields a 35Cl/37Cl ratio consistent with the known ratio of the carrier. These observations collectively provide evidence for effective fluoride removal and thus suitability for low energy stable Cl analysis. The goal of these tests was to evaluate the procedure vs. identify individual F-bearing molecular species. Though, we recognize that suppression of 19-F^- alone does not specifically quantify residual 16-O19F- at mass = 35. We will explain the diagnostic evidence and its limitations more explicitly and revise categorical statements about complete F removal/interference elimination to reflect these observations in the revision. Although high energy measurements would provide complementary validation, we consider the existing mass scans and carrier-ratio checks appropriate for evaluating the revised sample preparation procedure within the scope of this technical note.
  My last general comment refers to laboratory jargon. The descriptions of the measurement sequence/process are currently difficult to follow. I understand this is common feature as each laboratory develops its own language and terms that might be hard for others to follow. For example, my understanding of the measurement process is that each sample is measured once, and this measurement is artificially divided into 15 internal subdivisions that are then used to calculate the internal spread and standard deviation for the that measurement. The final error is then elevated if multiple measurements on standards have higher standard deviation to this value. Either way consider simplifying the terms and being clearer on the process.
  This comment was very helpful for us to see that the language we have adopted is not immediately clear to outside readers. We will take care to generalize and improve the language describing specific or nuanced laboratory procedures, as well as the techniques used to analyze and report analytical data.
  Figure 1 and the associated discussions. Figure shows increasing 37Cl- current as a function of the cathode voltage from 50ug sample without specifying if this is AgBr or Nb target, presumably latter but it is pertinent to mention this at someplace. Operational limits are not demonstrated by this data, I don’t doubt them, just wondering the value of this figure. It is well known that the current output typically increases as a function of cathode voltage in Cs sputter sources, see e.g. early work by Middleton et al. Is the response similar/different to AgBr targets, what if one increases Cs delivery, or ioniser power? Presumably what is important here is the signal to noise or more specifically signal to source memory, which is discussed later. Does AgBr and Nb need to be run at the same ion source settings? Can the figure be expanded to contain more general conclusion, e.g. 5 times increase in Cl- output with Nb vs AgBr?
  Thank you for highlighting the inadvertently omitted sample and ion source information. The manuscript will be updated to clarify that the 37Cl- current measurements were performed in Nb. The current response should indeed be similar to figure 1 whether the sample matrix is AgBr or Nb, except that at higher cathode voltages and Cs temperatures (to increase Cs delivery) the AgBr matrix becomes volatile to the point of melting out of the cathode, resulting in inconsistent and volatile currents that lead to a short in the ion source. Ionizer power does not have an impact on Cl- output within our nominal range of 130-133 W, which we maintain to preserve the lifetime of the ionizer. As a result, ionizer power is not a free variable for our source operation. While signal to noise is indeed important, the intended point of figure 1 is more in line with your suggested more general conclusion of higher Cl- output and will be adjusted to make this clearer.
  Figure 2: I have real trouble understanding this figure. Please elaborate and/or device alternative method of demonstrating the advantage of the reduced memory. Currently both x and y axis are ambiguous: integrated current – over what time/samples, acquisition sequence – what does this mean in terms of time and samples measured? What are the spikes in the graphs? Why does the 37Cl- with Nb integrated charge never go below 1 nC? Note: my understanding of this figure improved after I finished the entire manuscript but presume other readers would have similar challenges, so I left this comment as is. I would recommend just to make a simple plot of source memory 35 or 37 current as a function of time after running Cl targets with both AgBr and Nb and then repeat with reverse order to make sure build up from Nb is not reason for elevated memory from AgBr.
  We will elaborate in the manuscript to clarify any misunderstandings from figure 2. The x-axis integrated current is the collected charge for one of the 15 individual charge collections, or acquisitions, that make up a single 3-minute measurement of one blank sample. The acquisition sequence is merely an index from the beginning of all measurements, where the first sample begun measurement at acquisition 1, was measured through acquisition 15, sample was measured from 16-30, etc.
  We appreciate the comment raising the possibility of measurement order effects on memory. The wheels were measured consecutively (Nb first and AgBr the following day). The background measured at the end of the Nb wheel was substantially lower than that observed during the subsequent AgBr measurements. This order was important as it allowed the Nb background to be characterized without carryover from the much higher memory from the AgBr targets. Reversing the sequence would indeed provide info about memory carryover from AgBr into the Nb measurements, but that memory carryover would complicate characterization of the background associated with Nb operation. A more direct check of whether preceding Nb measurements could explain the elevated AgBr memory comes from independent operating sequences (of which we have ample process knowledge). Historic AgBr measurements performed following a source clean, without preceding Nb analyses, consistently show backgrounds similar to those reported here. Conversely, subsequent routine operation exclusively with Nb has shown no comparable background increase over multiple consecutive wheels. We therefore propose adding representative supporting measurements from these independent operating sequences to the revised manuscript. These would address the reviewer’s concern about Nb-derived memory buildup directly without introducing AgBr carryover into the Nb comparison via a reversed sequence.
  Figure 3, 6, 7: all the mass scans. I’d recommend using log scale so that the achieved interference suppression becomes visible.
  All mass scan plots will be updated with a log-scale y-axis.
  Figure 5: not sure this is needed. It is basic and can easily be referenced.
  We agree that this figure is basic; it was included to allow this manuscript to be a standalone description of the Nb stable Cl technique and to remove the need to find the figure in another paper for those unaware of the experimental setup. We are open to removing it if the editor agrees it is unneeded.
  Figure 9: introduce small shift between multiple samples on x-axis to make errors visible.
  Small x-axis offsets will be applied to all samples to improve legibility of error bars.
  -Paragraph lines 26-31: gives the impression that 10-100 mg of AgCl is still needed for measurement. Mention that samples with Cl masses around 1mg are routinely measured for 36Cl.
  This will be clarified with an addition to the corresponding text.
  -Paragraph between 32-46 currently oversells. Measuring stable Cl from separate vs combined with 36Cl targets has a long history and each have their benefits; there doesn’t have to be ultimate answer here either. Also, ANU AMS laboratory has been measuring stable Cl form separate targets for a long time, I suggest a reference would be appropriate here. Similarly note such as “significantly reduces the amount of expensive isotopically enriched Cl carrier” is little misleading in the context. That is, I agree with the statement, but feel that currently the reader is misled by implying from the above paragraph that close to 100mg of AgCl is needed, or that pure 37Cl carrier is needed, when ~1mg of 35Cl enriched carrier suffices. Then the sentence “This method provides additional flexibility because the carrier can be enriched in either 35Cl or 37Cl, …” is simply not true. I routinely measure combined stable Cl and 36Cl samples that are enriched with 35Cl without significant 36Cl background from the 35Cl enriched carrier, there is no difference here.
  We thank the reviewer for these comments and agree that separate and combined stable Cl and 36Cl measurements are established approaches with different advantages. Our intention here was to describe the benefits of the workflow adopted at CAMS rather than imply that separate target measurements are new or universally preferred. We will clarify this distinction in the revision.
  We believe there has been a misunderstanding of our intended carrier saving comparison and recognize that the transition from historical discussion could be clearer. The large (100 mg) AgCl masses mentioned in the preceding paragraph refer to early AMS measurements and not contemporary ones. Also, we refer to 37Cl enriched carrier, not isotopically pure 37Cl. The savings described apply to reducing the mass of Cl added by enriched carrier from ~1 mg per sample in our previous workflow to ~50 µg for a separate stable Cl aliquot. Natural ratio carrier can then be used, as required, for the separate 36Cl target. For a given enriched carrier stock, this represents an ~20 fold reduction in consumption and related per sample carrier cost. This benefit would apply whether the carrier is enriched in 35Cl or 37Cl and is not an estimate of the reduction in total analytical cost.
  We also agree that the ability to use either enriched carrier (35Cl or 37Cl) is not unique to separate-target measurements and will clarify that wording. The flexibility we intended to emphasize is in regards to the carrier’s 36Cl background vs. which stable isotope is enriched. We did not intend to suggest that suitable low-background carriers can’t be obtained. Rather, existing carrier stocks with unsuitable 36Cl backgrounds (which are prevalent in many labs that prepare 36Cl samples) can still be used for separate stable Cl determination because it is not introduced into the 36Cl sample. The availability of low-background 35Cl carrier does not remove this benefit for labs possessing carriers with elevated backgrounds.
  We thank the reviewer for drawing attention to ANU’s separate target practice. We were aware that they analyzed the 35Cl/37Cl separately on the same target (e.g., Davie et al. 1987) and will identify and cite a publication that also explicitly documents the separate target workflow in the revision.
  -Line 87: “Collectively, these early findings demonstrated the potential benefits from using Nb as a matrix for stable Cl isotope analyses of silicates.” So far there have been no evidence or mention of why the method would be beneficial for silicates, all the presented evidence is simply about Cl currents and source memory and applies to any Cl measurement, not just silicates.
  This will be clarified in the text to make clearer that the motivation for silicates was to be able to apply the demonstrated general benefits to a broader range of sample types, including the more challenging silicates, rather than the benefits being for silicates specifically over any other rock type.
  -Line 140: “centrifuging us referred” typo, should be is.
  Corrected.
  -Around lines 200-215 there is mention of ‘pass’ few times. Consider if this is needed. The magnet bounce cycle is given, and one could simply just state typical measurement time of X min instead of introducing the term ‘pass’.
  Agreed, ‘pass’ is redundant and will be removed from the manuscript. This will be clarified and changed to refer instead to a complete measurement and reference the measurement time.
  -Line 233: “Measured standard ratios are typically ~0.3% higher than the natural ratio, resulting in Fnorm of ~0.997.” Please state what the normalisation factor is for pure enriched carrier samples to validate that the instrumental fractionation is constant between endmembers.
  We will add clarification that the value of Fnorm is consistent across both natural and enriched standard materials.
  -Line 235: “Method one calculates the intra-target standard deviation of the 15 individual analyses that make up one measurement ‘pass’” This sounds like the sample is measured once and arbitrarily divided into 15 units that are then used to calculate standard deviation. The measurement could be equally divided into any number of units. This approach is fine for estimating the stability of a single measurement, but the 15 units shouldn’t be considered as independent analyses.
  We will clarify the text to remove mention of ‘individual’ analyses to clarify that a complete measurement is composed of 15 sequential measurements and not 15 independent analyses.
  In summary, despite my lengthy and perhaps somewhat critical comments I’m highly supportive of the manuscript and would like to see the work published as it offers a new approach to stable Cl measurements with increased sensitivity, that would have a benefit for the community. This work has potential to enable new studies into stable Cl concentration and/or fractionation. Many of my comments, I believe, are rather easy to address, e.g. improving the figure clarity, and will improve readability of the manuscript. My recommendation would be to consider more mass spectrometry specific journal but leave this up to the editor and authors to decide.
  We thank you again for your very detailed comments – we believe the manuscript will be greatly improved by them – and thank you for your support of the manuscript.
  
  Lesnek, A. J., Licciardi, J. M., Hidy, A. J., & Anderson, T. S. (2024). Optimizing the in situ cosmogenic 36 Cl extraction and measurement workflow for geologic applications. Geochronology, 6(3), 475-489.
  Balco, G., Blisniuk, K., & Hidy, A. (2019). Chlorine-36/beryllium-10 burial dating of alluvial fan sediments associated with the Mission Creek strand of the San Andreas Fault system, California, USA. Geochronology, 1(1), 1-16.
  Davie, R. F., Bird, J. R., Calf, G. E., Topham, S., Fifield, L. K., Ophel, T. R., ... & Chivas, A. R. (1987, November). The chlorine-36 measurement program at the ANU. In Fifth Australian conference on nuclear techniques of analysis: proceedings (No. INIS-mf--11495, pp. 6-9).
  
  Citation: https://doi.org/10.5194/egusphere-2026-4091-AC1
RC2:
'Comment on egusphere-2026-4091', Anonymous Referee #2, 10 Sep 2026

general comments
The authors of the "Technical Note: Stable Chloride Measurements at CAMS" present a systematic study to perform stable isotope measurements with an AMS ion source. Their progress to reduce the ion source memory for a highly volatile element such as Cl by very simple means is remarkable and highly useful to the community. Ion source memory is an important parameter in addition to general efficiency and output as well as potential suppression of the isobaric interference, which typically are the main criteria to optimize the composition of an AMS target. The study expanded the use of the Nb technique from silicates to diverse original sample matrices, which is highly relevant for labs that are engaged in analysis of other samples, e.g. ice or water.

specific comments
chapter 1.1. Explicitly state the materials used for the cathode and the wheel.

Line 83-85 The term "Nb wheel" may give the wrong perception of a wheel facibrated of Nb. Rephrase to clarify that you ran wheels containing targets with different matrices.

Fig. 2: What is the time scale on the x-axis? The term acquisition sequence is not defined anywhere else. How long is one acquisition sequence? Are they continuous or followed by target changes?

Line 153-155: Specify Nb powder properties (mesh size, purity) and clarify the sequence of addition for AgNO3 relative to the Nb powder. Is the AgNO3 added quasi simultaneously to the solution as the Nb powder?

chapter 2.2.: Please give more information on the stability of ratios over time when sputtering a sample, e.g. in a plot. Are there any effects from cratering? The measurements seem to take only few minutes. How much of the material is used up in this time? How many "passes" are performed per measurement? Please clarify the temporal measurement process, how the "pass" relates to the acquisition sequence.

Please also include some graphics to document the temporal evolution of the output for typical samples and blank cathodes and the magnitude of the blank correction. This could be combined with the data in Fig. 2.

line 233: State the precision of F_norm~0.997. Is the deviation from 1 significant and what is the reason? Could it be related to different focusing during the fast bouncing or a different sputtering yield?

line 335: Define ion source recovery time. How does this compare to values published for other ion sources (e.g. in Finkel et al. NIMB 294, 121–125 and Pavetich et al. NIMB 329, 22–29? )

line 337: What is the criterion for a "high-Cl sample"?

line 339/340: What is the typical amplitude and period of the current fluctuations?

line 373: Please comment in more detail, maybe somewhere else in the paper, on the difference in Sulfur output for Nb targets compared to AgBr targets.

technical corrections
Line 50: It should read "switching from AgBr to Nb"?

Citation: https://doi.org/10.5194/egusphere-2026-4091-RC2
- AC2: 'Reply on RC2', Tyler S Anderson, 17 Oct 2026
  
  We thank the reviewer for their helpful comments. We will revise the manuscript to include your comments, which are reproduced below in italics followed by our responses to each point.
  
  specific comments
  chapter 1.1. Explicitly state the materials used for the cathode and the wheel.
  We will clarify the text to include that both cathode and wheel are fabricated out of stainless steel.
  
  Line 83-85 The term "Nb wheel" may give the wrong perception of a wheel facibrated of Nb. Rephrase to clarify that you ran wheels containing targets with different matrices.
  We will rephrase this section to clear up the ambiguity about sample matrices and wheel material from our initial wording.
  
  Fig. 2: What is the time scale on the x-axis? The term acquisition sequence is not defined anywhere else. How long is one acquisition sequence? Are they continuous or followed by target changes?
  We will clarify the discussion toward the end of lines 190-209 to include mention of ‘acquisitions’ when describing how measurements are performed. One ‘acquisition’ takes 12.5 seconds, so the time scale on the x-axis is roughly 12.5 seconds per acquisition, not including pre-sputtering (~100s, the first time a sample is measured) or the time it takes for the sample wheel to move between different sample positions. A complete measurement is composed of 15 acquisitions and takes ~ 3 minutes. A single sample will be measured for 3 minutes, comprised of 15 acquisitions, before changing to another sample. Using Nb, every standard and unknown measurement is bracketed by a complete measurement of a Nb blank.
  
  Line 153-155: Specify Nb powder properties (mesh size, purity) and clarify the sequence of addition for AgNO3 relative to the Nb powder. Is the AgNO3 added quasi simultaneously to the solution as the Nb powder?
  We will add details regarding the Nb powder and will rephrase the sequence of reagent addition to remove ambiguity and make clear that niobium is added, followed by AgNO3, typically quasi simultaneously.
  
  chapter 2.2.: Please give more information on the stability of ratios over time when sputtering a sample, e.g. in a plot. Are there any effects from cratering? The measurements seem to take only few minutes. How much of the material is used up in this time? How many "passes" are performed per measurement? Please clarify the temporal measurement process, how the "pass" relates to the acquisition sequence.
  
  We will provide additional information on sample stability in the revised text. While we have not ever run a stable Cl sample to exhaustion, we have analyzed a Nb blank for over 400 acquisitions and were nowhere near sputtering to exhaustion. We have not observed any effects from cratering. You are correct that for a single complete measurement a sample is sputtered for ~4.5 minutes (~100 pre-sputtering ‘warm up’ plus ~3 minute charge collection and analysis), during which time very little material is consumed; typically a sputtering pit has just begun to be visible. One ‘complete measurement’ (which is the same as a ‘pass’ and will be clarified in the revised text) is composed of 15 12.5s acquisitions, and is typically all that is required to achieve permille precision; returning to a sample to measure it more than once is typically only done for the lowest Cl-content samples where collected Cl- charges are just above background from a Nb blank.
  
  Please also include some graphics to document the temporal evolution of the output for typical samples and blank cathodes and the magnitude of the blank correction. This could be combined with the data in Fig. 2.
  We will investigate the best way to present this additional information in the revised manuscript. We will make clearer in the text the magnitude of the blank correction. The collected charge from a Nb blank is typically on the order of 10 nC, whereas a 50 µg standard will have on the order of 10^4 – 10^5 nC of collected charge. The blank correction only becomes significant for small samples ~ 10 µg Cl or less.
  
  line 233: State the precision of F_norm~0.997. Is the deviation from 1 significant and what is the reason? Could it be related to different focusing during the fast bouncing or a different sputtering yield?
  The precision of F_norm is typically between 0.3 – 1 permille. While a normalization value of 1 is usually within uncertainty, the F_norm value is consistently 0.995 – 0.998. Calculating the standard error for F_norm across multiple wheels improves this precision to ~< 0.001. While we are unsure of the exact reason for the slight observed fractionation, we believe it to most likely be more related to something like how the two ions are focused differently rather than different sputtering yields.
  
  line 335: Define ion source recovery time. How does this compare to values published for other ion sources (e.g. in Finkel et al. NIMB 294, 121–125 and Pavetich et al. NIMB 329, 22–29? )
  In the first provided example paper, Finkel et al. describe it taking several minutes for Cl-currents to reach a plateau which was accurately indicative of the sample’s 35Cl/37Cl ratio. In our studies, except for the highest Cl samples (described in more detail below, but e.g., having 37Cl- current of 10-30 uA compared to a 50 µg Cl standard material with ~1-3 uA of current), we found a typical time to reach a plateau in blank current to be less than the time it takes for one complete blank measurement, i.e., < 3 minutes. For the aforementioned highest Cl samples we have measured, the plateau is observed during a second complete measurement of the Nb blank, or < 6 minutes.
  
  line 337: What is the criterion for a "high-Cl sample"?
  We try to choose the lowest maximum current setting for our current integrator that will avoid saturation of the electronics for all knowns and unknowns, as this minimizes the charge collected during background Nb blank measurements, increasing the signal to noise ratio. Accordingly, we generally define a ‘high-Cl’ sample to be one containing > ~500 µg, which is a result of the maximum Cl- current output of the sample exceeding the selected maximum current range on the current integrator, which was chosen based upon the Cl mass of the standards (generally 50 or 20 µg Cl).
  
  line 339/340: What is the typical amplitude and period of the current fluctuations?
  The period is short, on the timescale of ~100 ms, within one or two bounce cycles; effectively instantaneous on the scale of the electronics. The amplitude can be as much as a doubling of the current.
  
  line 373: Please comment in more detail, maybe somewhere else in the paper, on the difference in Sulfur output for Nb targets compared to AgBr targets.
  The sulfur output into the detector is not too high to measure Cl while using AgBr as a matrix, and it is far too high to allow into the detector while using Nb. There is no resolvable current at the faraday cup immediately before the detector for either Nb or AgBr. Given this, it is difficult to quantify post-acceleration 36S content in Nb relative to AgBr.
  
  technical corrections
  Line 50: It should read "switching from AgBr to Nb"?
  Yes, thank you, the typo has been corrected.
  Thank you again for your insightful and helpful comments, we believe the revised manuscript will be greatly improved with your suggested changes.
  
  Citation: https://doi.org/10.5194/egusphere-2026-4091-AC2

Tyler S. Anderson and Alan Hidy

Viewed

Total article views: 301 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
175	104	22	301	22	23

HTML: 175
PDF: 104
XML: 22
Total: 301
BibTeX: 22
EndNote: 23

Views and downloads (calculated since 27 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	37	11	2	50
Aug 2026	32	22	4	58
Sep 2026	67	42	11	120
Oct 2026	39	29	5	73

Cumulative views and downloads (calculated since 27 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	37	11	2	50
Aug 2026	32	22	4	58
Sep 2026	67	42	11	120
Oct 2026	39	29	5	73

Viewed (geographical distribution)

Total article views: 281 (including HTML, PDF, and XML) Thereof 281 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 18 Oct 2026

Short summary

We have developed a new method to measure stable Cl isotope ratios in rocks, improving on the standard method by preparing our samples in niobium instead of a silver bromide. This technique improves signal to noise by a factor of ~100, reduces sample measurement time and cost, and makes isotopic analyses of as little as 1 microgram of Cl extracted from a rock possible. This improves measurement of ³⁶Cl and makes small or chlorine-poor samples accessible for stable isotope analysis.


Total:	0
HTML:	0
PDF:	0
XML:	0