the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 25 Sep 2026
Precision and reproducibility of cosmogenic 10Be and 26Al measurements assessed from a multi-laboratory compilation of duplicate and reference material analyses
Abstract. To quantify the reproducibility and precision of cosmogenic 10Be and 26Al analyses, we analysed a suite of liquid and quartz reference materials, together with duplicate preparations of unknown research samples. Comparison among these materials allows us to constrain uncertainties associated with AMS measurement, chemical extraction, and quartz purification, and to assess whether additional variability is introduced at successive stages of the analytical process. Samples were prepared in three chemistry laboratories at the Australian Nuclear Science and Technology Organisation (ANSTO), the University of Wollongong (UOW), and Imperial College London, and subsequently analysed by accelerator mass spectrometry (AMS) using the 6 MV SIRIUS accelerator at ANSTO. All three laboratories showed intra-laboratory variability below the routine 2% analytical uncertainty, while measured values were consistent with the nominal values of the reference materials. These results indicate that both methodological reproducibility and inter-laboratory variability are also within the analytical uncertainty, and provide no evidence for additional uncertainty associated with processing quartz relative to liquid reference materials. The implications for cosmogenic-nuclide research are that: (1) the reproducibility and precision of cosmogenic 10Be and 26Al analyses are typically 2–3% when sufficient counting statistics are achieved; (2) at lower isotope concentrations, counting statistics and/or background corrections may limit the achievable measurement uncertainty, although the resulting variability remains consistent with the reported analytical precision; and (3) precision of approximately 1% is achievable for high-concentration liquid reference materials, as demonstrated by Be-KN-5-4 and Al-KN-4-3, but has not yet been demonstrated for quartz samples.
- Preprint
(1433 KB) - Metadata XML
-
Supplement
(78 KB) - BibTeX
- EndNote
Status: open (until 06 Nov 2026)
- RC1: 'Comment on egusphere-2026-5512', Ashley Corbett, 22 Oct 2026 reply
-
RC2: 'Comment on egusphere-2026-5512', Anonymous Referee #2, 23 Oct 2026
reply
This study describes important work to assess the reproducibility of 10Be and 26Al analyses from samples prepared across different three different sample prep labs from liquid reference materials, quartz reference materials, and unknown samples. The work performed aimed to break down the sample preparation process and determine which, if any, processes involved could impact reproducibility.
The primary stated outcomes of these efforts were the findings that 1) high ratio reference materials can have precision approaching 1%, 2) reproducibility of ~2-3% is achievable with sufficient counting statistics in both liquid and quartz samples, and that the different processes involved with liquid reference materials and quartz dissolution and chemical separation do not lead to any resolvable additional uncertainties, and that 3) the main limit to precision for lower ratio samples lies solely in counting statistics, not in any sample preparation procedures.
Overall, I found manuscript to be structured well and very easy to read, while presenting useful insights. I only have minor comments, which I’ve broken down by relevant lines or figure, listed below.
Line 101- what might lead to the variability in transmission between individual samples, assuming that they went through the same sample preparation procedures?
Line 118 – The window used for the ionization chamber is mentioned for 26Al, but not for 10Be. Is the absorber cell used for 10B suppression acting as the window? If so, what is the material used? Though this is potentially requesting too much AMS technical detail in a non-strictly AMS methods paper.
Line 141 – Why was the amount of HF used for the UVM-A reference material chosen? Is this based on the average amount of quartz digested to make an approximate analogue to the solid sample preparation for the liquid carrier?
Line 160 – what is the cause of the different preparation requirements between the Nishiizumi and UVM-A standards?
Lines 220-235 – Units, presumed to be atoms/g of quartz, are missing on the reported 10Be and 26Al concentrations. For clarity, either list units for each value, or for concision, state clearly beforehand the standard unit used for the reported concentrations.
Figure 8,A – The caption states that 10Be samples are divided into groups of xi > 10^5, 10^5> xi > 10^4, and xi < 10^4 atoms/g, though the rightmost bar in the 10Be data set is labeled as 10^3< xi <10^4. Which is correct? E.g., does the rightmost bar truly represent no less than 10^3 atoms?
Figure 8,B – Should the right most bar in the 26Al dataset say xi < 10^4 rather than >?
Citation: https://doi.org/10.5194/egusphere-2026-5512-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 407 | 148 | 63 | 618 | 70 | 66 | 74 |
- HTML: 407
- PDF: 148
- XML: 63
- Total: 618
- Supplement: 70
- BibTeX: 66
- EndNote: 74
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review of: Precision and reproducibility of cosmogenic 10Be and 26Al measurements assessed from a multi-laboratory compilation of duplicate and reference material analyses
In this manuscript, the authors explore repeated 10Be and 26Al analyses of (1) liquid reference materials, (2) quartz reference materials, and (3) unknown quartz materials. The goal of the study is to assess the uncertainties associated with AMS measurements of cosmogenic nuclides, as well as to determine which parts of the process (quartz purification, Be/Al analysis, or AMS analysis) could introduce variability.
Overall, these are important questions to explore, and they yield meaningful conclusions for many laboratories around the world. The authors document 2-3% reproducibility of cosmogenic nuclide measurements when sufficient counting statistics are achieved, which has been suggested in plenty of other studies, but is well supported and conclusively documented here.
I found the approach particularly interesting and unique, since most past studies have focused on either liquid or quartz reference materials, but not both. This multi-pronged approach allows the authors to assess whether uncertainty is compounded throughout progressive stages of sample processing and analysis. They conclude that it is not.
The paper is well-written and easy to read. The structure is clear and follows a logical progression of ideas. Individual paragraphs have strong topic sentences. As a whole, this was a pleasure to read.
Below, I’ve detailed some bigger-picture comments, as well as some line-by-line suggestions. Overall, I commend the authors on an interesting study that uses a very interesting, unique, and thorough experimental design.
Lee Corbett
***
Big-picture comments:
Study design. The paper would benefit from a very clear, concise description of the study design immediately following the introduction. At present, it dives right into AMS methods, which might lose some readers. A flow-chart or other schematic could really help with visualizing all the different components of the study.
Relevance of methodological details. The paper contains very detailed methods, which is fine for this sort of piece. However, it would be useful to the reader to lay out why these methodological details are important. For example, section 2.1 would be impenetrable to anyone other than the AMS community, but it never mentions the most important big-picture element: all samples were analyzed on the same AMS, thereby omitting AMS-to-AMS variability from the study.
Methods order. I found it confusing that the method details appear to go in reverse order; AMS measurement is discussed before sample preparation. It might be easier to read if the flow of the writing followed the flow in the lab.
Beam current variability. I’m surprised to see such a large range of beam currents for samples that are nominally all the same (and, presumably, all clean). It would be useful to discuss why, even though your data show that it doesn’t ultimately matter. I think people interested in methods, and especially people who run prep labs, will be curious.
Bulk dissolutions. The authors performed bulk dissolutions of the quartz reference materials, making them more analogous to the liquid reference materials. This is a clever approach, but right now it’s buried in the results details. Implementing a study design section, like I’ve suggested above, would help this interesting subtlety to pop out more.
26Al/10Be ratios. I assume the authors thought about this, but I wonder if it would be useful to extend the analysis to 26Al/10Be ratios, where the data exist? Of course they’re related to the two isotopes individually, but people in the community who do multi-isotope work might find it helpful to see.
***
Minor Comments:
Lines 28-30: I suggest rewording to focus on what the paper DOES address, rather than what it DOESN’T.
Lines 40-41: It would be helpful to include just a few words, perhaps in a parenthetical, defining what “not good” means. At present it’s vague.
Lines 153-155: This sentence is about the study design; it feels misplaced in the results section. As described in my bigger-picture comments, it would be helpful to expand this significantly into a full-fledged study design section (and then not have it in the results).
Lines 160-161: Can you describe why the two liquids are prepared differently? I assume it’s because the Nishiizumi series doesn’t contain small amounts of Fe, Ti, etc.? This seems like an important detail for readers.
Lines 229-230: It would be useful to remind the reader (just in a parenthetical maybe?) of why these two should be related.
Lines 318-319: This is especially interesting and important. It might be worth nothing this conclusion more explicitly in the abstract, or at least the conclusions.