Thank you for taking the time to review our manuscript and for your positive and constructive comments. Please find our responses below.

Orientation effects:

All the micas analysed are massive crystals, mounted perpendicular to the c-axis (ie ablating parallel to the cleavage planes). We will add this detail to the methodology for clarity. On our laser system we have found this orientation to produce much more consistent ablation compared to mounting parallel to the c-axis (ie ablating through the cleavage planes). We have also observed stronger orientation effects compared to Larson et al. 2025. All our mica reference materials, and samples where possible, are mounted perpendicular to cleavage. I believe this is the most common practice in many labs and recommended by Larson et al 2025. Mounting samples flat to avoid polishing is a good suggestion to minimise potential Ca contamination, however we have found this orientation produces more spot to spot variability and inconsistent ablation. This is especially true for samples with weak cleavage (such as our MDC phlogopite) which tends to be ‘flaky’ when mounted flat. With regards to the observed scatter in Kola between sessions for K-Ca, all ablations were in the same massive crystal orientated perpendicular to cleavage, so unfortunately can’t be attributed to crystal orientation. Future work is planned to assess orientation effects in K-feldspars.

Fluence & ablation characteristics:

I agree that laser repetition rate and fluence effect ablation properties and downhole fractionation (DHF), due to the interaction between particle size and crater depth with time. To minimise DHF effects we opted for shallower craters (smaller depth to diameter ratio) using 100 um, 5 Hz and a ‘medium’ fluence of 3.5 J/cm2. Using 5 Hz and 3.5 J/cm2 is standard practice in our lab and produces reproducible RbSr dates for micas. We investigated fluence effects prior to publication of Redaa et al. 2021, and found that higher fluence increased fractionation effects between minerals and Mica-Mg pressed powder, which at the time was used as the primary RM. On our laser system using 10 Hz also increases DHF and hence matrix offsets between minerals and glasses. This is a significant consideration with the current data processing workflow, using a glass to correct for DHF and then a mineral to correct for additional matrix offsets for Rb-Sr and K-Ca ratios. In this study we used a mica to correct feldspars, which produced accurate results under these ablation conditions, but matrix effects may become significant when using a smaller laser spot size and/or faster repetition rates.

The first 10s of the signals were cropped primarily to reduce effects of Ca surface contamination rather than signal stability, which rises fairly sharply. While not all samples/sessions were affected by Ca contamination, a consistent approach was adopted for cropping signals. As suggested, I will add some examples signal intervals to the Supplementary Material for micas and feldspars, including 1) good, flat signal with a short stabilisation time, 2) examples of surface contamination, and 3) another example with heterogeneous common Ca within the mineral during ablation (common for feldspars).

PA Factor accuracy:

I would consider PA Factors to be one of the long-standing limitations with single collector ICP-MS analyses. The measured PA factors in the Mass Hunter software are accurate for what the ICP-MS instruments are designed for – specifically solution-based concentration analyses within +/- 5% accuracy. However, for laser ablation isotopic analysis the inherent uncertainty in the PA Factor measurements can be significant leading to bias when switching between modes. As noted by Zack & Hogmalm 2016 the uncertainties are exacerbated in reaction mode as the kinetic energy differences between on-mass and mass shifted ions effects the detector response, hence a PA factor measured for 103Rh would not be accurate for 87Sr16O. In this study we measure PA factors for each session using N2O/SH6 reaction gases to get the most accurate PA Factors (eg for 85Rb and 40Ca), however inaccuracies still remain.

On our ICP-MS/MS we have observed differences in drift between pulse and analog signals during long sessions – as detail for LuHf in Glorie et al., Chem. Geo., 2024.  This drift is acknowledged and can be accounted for in the Mass Hunter software using dynamic PA Factors which can reanalyse the PA Factors on the fly during a run – however, this feature is only available in solution/spectrum mode. Unfortunately, for LA-ICP-MS time resolved data collection, the system is limited to only measuring the PA factor at the start of the run, hence additional considerations are required.

In addition, the recent publication by Petts et al, GGR, 2026 details non-linearity in analog mode. While the causes of this are unknown and we are yet to investigate this in detail on our ICP-MS system, it does highlight that matching cps rates is certainly a beneficial approach.

Matching the detector mode between samples and RMs has been standard practice for isotope analyses for many years (eg. Chew et al 2018, https://doi.org/10.1111/ggr.12257). Measuring 238U in analog and 206Pb in Pulse, can certainly produce accurate analyses as long as these isotopes are measured in the same mode in the RM’s and samples (eg. Chang et al 2006, https://doi.org/10.1029/2005GC001100). Similarly, in Redaa et al. 2021, we measured Rb in analog and Sr in pulse in all materials tested. Inaccuracies can occur if 238U/85Rb/40Ca is measured in Pulse in the RMs but Analog in the samples (or vice versa).

Every measurement has some level on uncertainty (including the PA Factors) and whether this is significant/noticeable depends on the isotope system being analyse. For some samples uncertainty from signal variability (inclusions or low sensitivity counting statistics) may overshadow any PA Factor uncertainty. However, when analysing RMs and materials with very smooth, reproducible signals even a small offset in the PA Factor can create significant bias. As an extreme example, analysing S isotopes requiring per mil precision the PA Factor offset can be very significant - seen as a step in the isotope ratios when an analyses goes from analog to pulse during an ablation.

To minimise complications from all these factors we aim to analyse samples and RMs in the same detector mode wherever possible for all our isotope systems, to avoid additional uncertainty or bias from any off-line correction. This was the reason for making the second SNM glass with a better match of concentration (and P/A mode) as the samples.

Figure 6a: bimodal K-Ca data for MDC. 

The analyses in this example were located on different regions of the MDC crystal, where location correlates with the bimodal distribution. They were located along parallel cleavage planes ~1 mm apart. The common Ca concentration is different between the two areas (natural trace element heterogeneity) but they are age homogeneous. What is notable about these spots in that the Ca varies but the common Sr doesn’t (Fig 6a versus 6b). With common Ca and Sr uncorrelated this suggests these elements are located in different sites within the mica structure with implications for element diffusivity. I will clarify the cause of the bimodal population in the figure caption and add additional discussion for clarity to L350-359.

Minor edits

L162, L440, Fig 5 legend. Thank you for picking up on these typos. They will be corrected in the final manuscript.

- We appreciate the reviewer’s thorough examination of the data and supplementary tables. They have highlighted some of the limitations with the earliest data sets collected (Tables S4 and S6) and articulated the reasons why we subsequently refined the method with a new SNM glass. Reviewer’s comments are included below and our responses in bold and proceeded by a hyphen.

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).

- We did our initial testing with SNBM but realised there were some limitations and complications with PA factors and subsequent need for corrections. This was the reason for making the SNM glass with a better concentration match to the samples. We will make this clearer in the text which session used which glasses, and add which glass was used for each session in Table S1.

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.

- Agree entirely with the limitation relating to the data in Table S4 (SF₆ + H₂, first session). The reviewer has highlighted the PA issues with the SNBM glass, and subsequent requirement for correction of ⁴⁴Ca/⁴⁰Ca using Kola. Unfortunately, this was only apparent after collecting the data, and we also had questions as to the validity of this approach which was the reason for refining the method with the second SNM glass. We repeated the analysis of SF₆ + H₂ and N₂O low H₂ gas combinations to verify this original data (repeat sessions: Tables S5 and S7 respectively). While there is a potential for the initial to be incorrect for Hogsbo due to unaccounted for drift in the Analog detector for the data in Table S4, the same dates were measured for Hogsbo in sessions S5 within uncertainty, when using SNM glass (without the ⁴⁴Ca/⁴⁰Ca issue mentioned by the reviewer). While the absolute ages are slightly higher for Hogsbo for some sessions, they are the same as the RbSr dates within uncertainty. Despite the limitations we have included this initial data after validation with repeat and more robust analyses, rather than choosing not to include it. We can add additional discussion of this in the text.

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).

- The cause of the excluded data points above the line were primarily due to spikes in the data. These spikes were significantly reduced in later sessions after repolishing and cleaning the samples. The data from the two earliest sessions Table S4 and S6 were worst affected but the spikes. Again, this was a learning point taken from these initial data. Both these sessions were repeated (Tables S5 and S7), obtaining cleaner data and comparable ages within uncertainty, validating the initial data treatment.  We can add additional discussion in the text on why some data points were not used in the calculation and include some representative laser signals in the Sup Mat illustrating the spikes and Ca surface contamination observed in the initial sessions.

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.

- We will aim to make the Introduction more concise and move the paragraph as suggested.

Line 34, and elsewhere throughout the manuscript: Spaces should be included after the semi-colons in the reference lists.

- We will correct this and other reference formatting issues as highlighted by the reviewer below (EndNote legacy formatting).

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.

- We will review the grammatical use of commas throughout

Line 100, and elsewhere throughout the manuscript: There are a lot of instances where double parenthesis are used with references. This should be avoided.

- Will be corrected

Lines 109 and 110: The reference format varies between “Amelin and Zaitsev” to “Amelin & Zaitsev”. Make the formatting consistent throughout the text.

- Will be corrected

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.

- We have simply used the terminology used in the original references but agree that they are the same thing. We will make them consistent in this manuscript.

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.

- Will be corrected

- 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.

- We will reword for clarity

Lines 282-287: The issue of not incorporating detector drift in correcting the ratios should be discussed with the results later on.

- We will add additional discussion with the data

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.

- The correction factor is derived from the standard age equation, comparing the measured and expected dates of the matrix reference material. We can state it explicitly in the text for clarity.

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).

- They were all polished epoxy blocks. After observing initial Ca contamination, we repolished the blocks using a brand-new lap and distilled water, this significantly reduced the contamination. Also using an ultrasonic bath and distilled water helped remove previously ablated material from craters, leading to fewer spikes in subsequent analyses. The cleaning is currently explained ~L445, but we can include it in the methodology for clarity.

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.

- Approximate common Ca concentrations are included in Table 1, we can add common Sr concentration also for comparison. Using MDC as the example, it has higher but consistent common Sr (~56 ppm) compared to common Ca (<1 to ~30, with occasional 40-50 ppm). We can reword the discussion in terms of K/Ca and Rb/Sr variability as suggested. The Ca is clearly the most heterogeneous – reflected in the range of ppm and the spread along the isochron. This implies that Ca is located differently to Sr within the structure of the mica, however the exact mechanism isn’t able to be determined from this data set. This can be highlighted for future investigation.

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.

- I think the ‘non-linearity’ issues at low cps as reported in Petts et al 2026, can be considered as a counting statistics and data processing issues, as opposed to a ‘detector’ issue as such. When there are a significant number of measured count rates close to the minimum, normal mean of ratio calculations can be skewed, this is exacerbated when there are zeros in the data. Also, the calculation of uncertainty can be underestimated in some data processing software packages, and various computational methods have been proposed to achieve more realistic uncertainties. To minimise these issues, we used a relatively long counting time on ⁴⁴Ca to improve the statistics at low count rates: 100 ms, 1 count = 10 cps. For Kola with one of the lowest common Ca concentrations, it gave ~100-500 cps. In addition, during data processing the signals were cropped to avoid any zeros in the 44Ca signals as calculation of uncertainty breaks down when zeros are present in the denominator of a ratio. We can add cps data to the Supplementary Tables.

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.

- There were different affects from Ca contamination: spikes of remobilised ablated material (predominantly affecting the lower cps isotope ⁴⁴Ca) and surface contamination on the samples which affected the ⁴⁰Ca signal more (due to ⁴⁰Ca being the most abundant isotope). We note the seeming contradiction as written and will add additional clarification, and include example signals in the Supplementary Material.

Line 457: I do not think that the abbreviate “EM” was defined prior in the text.

- Will correct 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.

- Will add the cps data

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.

- Will clean this up.

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 were two repeat sessions for SF6 + H2 (Tables S4 & S5) and N2O low H2 (Tables S6 & S7), which are shown in Figs 7 & 9. We will add an additional Supplementary table with a data overview for clarity with additional details for each session, including the SNBM/SNM glasses used and reaction gas combinations.

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.

- This is correct. Will include a header to clarify these columns