the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 02 Jun 2026
Coupled K–Ca and Rb–Sr dating by LA-ICP-MS/MS – reaction gas optimisation and geological applications
Abstract. Both the Rb-Sr and K-Ca β-decay isotopic systems can be used to date a range of mica and feldspar group minerals and have the potential to unravel cooling and alteration processes in a wide range of geological settings. The development of LA-ICP-MS/MS has enabled direct in-situ analysis of these β-decay geochronometers via chemical separation with reactive gases within the mass-spectrometer. As well as rapid analysis, the main advantage of in-situ K–Ca dating is that Ca- and Sr-bearing inclusions can be avoided, which are a limiting factor for conventional bulk mineral dating via TIMS. Both Sr and Ca are highly reactive with both SF6 and N2O to form M-F, M-O or M-OH reaction products, while K and Rb are unreactive with either gas, enabling efficient separation of the parent-daughter 40K–40Ca and 87Rb–87Sr isotope pairs. Additionally, mixing a small amount of H2 with SF6 or N2O efficiently eliminates 40Ar based interferences and reduces the background generated by the high ion load in the reaction cell when measuring mass/charge ratio (m/z) 40. This study compares the accuracy, precision and product ion sensitivity between four reaction gas combinations: SF6 only, SF6 plus 2 ml min-1 H2, N2O plus 7 ml min-1 H2, and N2O plus 10 ml min-1 H2, by analysing a range of micas and feldspars with previously constrained dates: MDC & Kola phlogopites, Högsbo and Robins Folly muscovites, G71560 polylithionite, and F-KN and Bohus K-feldspars. Using these gas mixtures we present coupled Rb–Sr and K–Ca dates from a single ablation spot in low Ca-bearing (5–300 ppm) micas and feldspars to within 2 and 5 % age uncertainty, respectively. The direct coupling of Rb–Sr and K–Ca dates within the same ablation volume allows assessment of isotopic disturbances at high spatial resolution. The gas combination of SF6 plus 2 ml min-1 H2 was found to be most effective for coupled K–Ca and Rb–Sr dating in generating the highest sensitivity of reacted species and in reducing the background for reacted 40Ca. Analysis of the FK-N feldspar from Madras, India shows the potential for the two isotopic systems to reveal decoupled dates, with the 515 ± 22 Ma Rb–Sr date representing the crystallisation of the granite and the 437 ± 38 Ma K–Ca date indicating late-stage hydrothermal activity.
- Preprint
(1900 KB) - Metadata XML
-
Supplement
(9048 KB) - BibTeX
- EndNote
Status: open (until 14 Jul 2026)
- RC1: 'Comment on egusphere-2026-2790', Anonymous Referee #1, 28 Jun 2026 reply
-
RC2: 'Comment on egusphere-2026-2790', Anonymous Referee #2, 04 Jul 2026
reply
Review of “Coupled K-Ca and Rb-Sr dating by LA-ICP-MS/MS – reaction gas optimisation and geological applications”
Overall, this research is very timely and definitely needed to advance K/Ca geochronology. The manuscript itself is generally well-written and logically organized with informative figures. I think the work has been done well, however, there are some potential issues with data acquisition/processing that I think need to be addressed more. The authors do allude to them in the text, but they need to be discussed more. If the data is affected by the issues, then it is invalid to present them without proper discussion and transparency of the data. Below are my moderate and also minor comments for the manuscript.
Comments:
It is not clear which glasses were used for which analytical sessions. Using only SNBM brings forth potential issues of P/A corrections of the 40Ca/44Ca (or its reciprocal), as the authors have mentioned in the text. Using SNM is a more straightforward and (likely) accurate approach as it does not require the P/A correction to be applied. Thus, SNM would be the ideal choice for all data acquisition, but it is apparent that it was not used in all analytical sessions (as it is not reported in all Supplementary Data tables). It needs to be stated which analytical sessions used which glasses, because the need for the P/A correction is only SNBM was used brings forth potential issues resulting in inaccurate corrections, and inaccurate data (which I expand on in my next comment).
I do not think it is specified how the samples were organized throughout an analytical run. Only after looking at the Supplementary Data I see that MDC and SNBM were cycled throughout blocks of the other samples (I am using Table S4 data as the example here). Since only SNBM glass is listed, it implies that the 44Ca/40Ca ratios were corrected using the Kola phlogopite intercept for this particular analytical run. Kola phlogopite was the first sample analyzed in the session. So, using the Kola phlogopite intercept to correct 44Ca/40Ca is not appropriate for the entire run because it will not account for instrument drift throughout the run, only capturing the pulse/analog offset for 40Ca in the initial portion of the run. If I plot the data for Kola unanchored, and compare it to the plotted unanchored data for Högsbo (i.e., the last sample run in that session), they have very different intercepts (39.5 vs. 48.2 for 40Ca/44Ca). The date for Högsbo may, therefore, be incorrect as its 44Ca/40Ca would be incorrectly normalized. It seems this organization of reference materials/samples is similar in the other sessions, with Kola being first and Högsbo being last. If this is the reason why the Högsbo dates are consistently too old (except for N2O + high H2), then is this data valid to include (I am just picking on Högsbo here, but it is a relevant question for the other datasets too). I know these potential issues are alluded to in Section 3.4, but it needs to be emphasized again when discussing the results.
There are a lot of isochrons with excluded datapoints for K/Ca when digging into the Supplementary Data. The reader is not made aware of this in the main text. What reason(s) justify the exclusion of these datapoints? The excluded result are consistently plotting above the isochron regressions, so it appears to be systematic. This is especially important to explain for MDC, which is used to matrix-normalize the rest of the data. Using Table S4 data as the example again, if I do not exclude the datapoints for MDC that have been excluded in your data processing, I receive a date of c. 316.5 Ma for MDC. Using that date to matrix-normalize Högsbo, I receive an anchored isochron of c. 1068 Ma, which is significantly older than the reported age. However, if I do not anchor the Högsbo isochron, I receive c. 1032 Ma, which fits very nicely with its reported age. Yet, the initial 40Ca/44Ca is too high at 48.2 (which again, points towards a slightly incorrect normalization of the initial ratio).
Line-by-line:
Introduction: Generally, I find the Introduction a bit long. There is a lot of information provided that reads as more of a general review rather than introducing the specific work that is performed by the authors. If it could be made more concise, I think it would be beneficial to set up the remainder of the manuscript. One paragraph in particular caught my attention, which starts at line 86. This information does not flow with the above and below paragraphs, and may be better suited to a section where the data processing/presentation procedures are described.
Line 34, and elsewhere throughout the manuscript: Spaces should be included after the semi-colons in the reference lists.
Line 83, and elsewhere throughout the manuscript: There are a lot of sentences where commas are omitted where they should be included, typically when they are started with a conjunctive adverb or adverbial phrase. In this example, the comma is not included after “In this study…” E.g., line 108 included a comma after “deposit” where it does not belong. Please, review the text for these minor grammar issues.
Line 100, and elsewhere throughout the manuscript: There are a lot of instances where double parenthesis are used with references. This should be avoided.
Lines 109 and 110: The reference format varies between “Amelin and Zaitsev” to “Amelin & Zaitsev”. Make the formatting consistent throughout the text.
Lines 110, 111, 112: What is the difference between “discordant, lower intercept 206Pb/238U” and “U-Pb isochron”? They both sound like lower-intercepts using Tera-Wasserburg plotting.
Line 119, and elsewhere: I have noticed in a few instances throughout the text that references built into the sentence are still placed in parenthesis “…as described in (Mortimer et al., 1987),…” It looks like the references were automatically placed in-text, and this may be an artefact.
Line 173: What is meant by the production of “40Ar16O interference”? It produces 40Ar16O, but that is not an interference in itself, it can create an interference with another m/z.
Lines 282-287: The issue of not incorporating detector drift in correcting the ratios should be discussed with the results later on.
Lines 296-297: Describe how the correction was performed. It is only apparent when looking at the formula in the Supplementary Data tables how this was done.
Lines 326-327: How were the samples prepared? It is not stated in the text. I am assuming in epoxy mounts. If they were polished using water (even distilled water), it may allow Ca to contaminate the materials. We have found that polishing the mounts without any form of liquid (i.e., using 7000-grit dry sandpaper) significantly eliminates contamination of the micas in the epoxy mount and only ~2 clean shots are necessary prior to analysis (there is still sometimes common Ca signal, but it is usually just the first few second and then the signal it stable).
Lines 351-355: I do not understand why the difference in the ratios supports difference sites being occupied by Ca and Sr. The micas should (ideally) crystallize with uniform 40Ca/44Ca and 87Sr/86Sr, according to the initial values. I think it is easier to view this from the standpoint of 40K/40Ca and 87Rb/87Sr values. The former would be variable for the case in Figure 6, and the latter would be relatively homogeneous. If we assume that both K and Rb are evenly distributed in the mica structure (at least at the size of the ablation volume), then it means Ca was unevenly distributed, and Sr was evenly distributed upon mica crystallization. If Ca is only in the I-site, but Sr is potentially in both I- and M-sites, why would that cause Ca to be heterogeneous and Sr to be relatively homogeneous in the mica structure? It would be useful to know the difference in Ca and Sr content of MDC, as I would assume it has higher common Ca content than common Sr content overall. Maybe a heterogeneous distribution is more apparent for Ca when it is higher in concentration in the mica than Sr.
Lines 393-396: It sounds like a potential detector nonlinearity issue. This has recently been demonstrated by Petts et al. (2026; GGR) and we have also noted this in our lab for Sr ratios when 86Sr CPS are low. It could be an issue for K/Ca and I agree with the last statement of these lines. The issue may only become apparent when the reference material 44Ca CPS is not matched to the unknowns. It would be useful to see the CPS data for the analytical runs in the Supplementary Data. This has implications for using Kola as the 44Ca/40Ca normalization (i.e., Lines 284-287). If there is a nonlinearity issue, then the calculated intercept for Kola will not match the other materials that have higher Ca content.
Line 448-450: As stated, 44Ca contamination would lead to younger dates, but Kola and G17560 were discussed as having dates that are considered too old. These statements are contradictory.
Line 457: I do not think that the abbreviate “EM” was defined prior in the text.
Supplementary Data:
It would be useful in the Supplementary Tables to have the raw CPS data for the various m/z analyzed in the runs. This data is necessary to determine m/z that were recorded in pulse versus analog detector modes.
The tables are also a bit disorganized. There are isochron plots copied over the data cells in some of the tabs, as well as isochron plots pasted over other isochron plots.
I do not understand how the Supplementary Data tables line up with the figures in the main text. For example, only one set of Högsbo, Kola, Robin, etc. data is listed in Table S4 “SF6 + H2”, but in Figures 7 and 9, there are two dates shown for each.
There are ‘x’ marked in the columns labelled “to MDC”. I assume these ‘x’ indicate rejected results, but it is not apparent in the current organization of the data tables.
Citation: https://doi.org/10.5194/egusphere-2026-2790-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 109 | 40 | 9 | 158 | 17 | 8 | 9 |
- HTML: 109
- PDF: 40
- XML: 9
- Total: 158
- Supplement: 17
- BibTeX: 8
- EndNote: 9
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General comments:
This is a very well-crafted manuscript that is a pleasure to read. It lays out the fundamental working principles and practical guidelines for simultaneous measurement of Rb-Sr and K-Ca dates using LA MS/MC-ICP-MS. There is ample detail to all the necessary steps to guarantee accurate and precise data collection. The only thing that is missing is a true secondary reference material for K-Ca that has been independently measured e.g. by isotope dilution. Nonetheless, the use of well-characterized natural and synthetic materials clearly demonstrates the applicability of this technique. There remain, of course, plenty of obstacles that the authors summarize towards the end of the manuscript. The present work provides an important foundation for future studies.
In addition to a few minor technical edits (see below) there are a couple of suggestions I can offer:
Line 162 typo: [c]lean
Line 284-285: Are the authors suggesting here that the built-in P/A calibration factors are inaccurate on the Agilent 8900? I have read other papers that describe need for ‘offline’ P/A correction and evidence for non-linearity of detector. I find this challenging to understand. P/A factor can be established in Masshunter in mass-shift mode. What is the evidence that the Masshunter P/A routine yields inaccurate results? No such discussion exists for U-Pb dating for example despite 207 and 238 commonly in P vs. A modes respectively.
Line 440: Recent work by (Olierook et al., 2026) – fix reference.
Figure 5 legend. Top-most entry read ‘G17560_pln’. What is ‘pln’? This material is described as lepidolite.
Figure 6a: the K-Ca data is strongly bimodal. Would it be worth commenting on this in the text? It might be an important observation. Does this mean that some spots encountered high common-Pb domains whereas other encountered more radiogenic Ca domains. Would this be a primary feature or solid-state unmixing? Other options?