Response of Rb-Sr system in biotite during contact metamorphism in the aureole of the Makhavinekh Lake Pluton, Labrador

McFarlane, Christopher R.M.

doi:10.5194/egusphere-2025-6320

Preprints

https://doi.org/10.5194/egusphere-2025-6320

Preprints

31 Dec 2025

| 31 Dec 2025

Response of Rb-Sr system in biotite during contact metamorphism in the aureole of the Makhavinekh Lake Pluton, Labrador

Christopher R.M. McFarlane

Abstract. High-temperature contact metamorphism in the aureole of the 1322 Ma Makhavinekh Lake Pluton, Labrador, led to progressive consumption of 1850 Ma garnet formed during upper-amphibolite facies regional metamorphism that produced migmatitic paragneiss (Tasiyuak Gneiss). Biotite Rb-Sr isotope measurements were carried out in situ by laser ablation MS/MS ICP-MS allowing biotite in a variety of textural settings to be characterized. This natural laboratory provides important constraints on the nature of Rb-Sr closure temperature (Tc) as a function of textural setting in high-grade metamorphic rocks. Intact biotite inclusions armoured in garnet preserved in the outer aureole (>4km from the contact) display a range of Rb-Sr isochron ages between ~1850 Ma and ~1322 Ma consistent with a zone of partial retention of Sr in biotite. Isotopic resetting in the outer aureole was controlled by microfractures in garnet that provided short-circuit diffusion pathways for redistribution of radiogenic Sr into plagioclase-bearing contact metamorphic assemblages; biotite inclusions isolated from microfractures retained 1850 Ma Rb-Sr isochron ages. Biotite grains falling along a ~1322 Ma isochron attest to efficient intra- and intercrystalline Sr diffusion at T ≥ 500 °C on timescales of ≥ 5 Myr. Samples in the central part of the contact aureole (3.7 to 1.1 km from the contact) contain partly resorbed biotite surrounded by contact metamorphic Opx + Crd coronal assemblages in addition to armoured inclusions in relict garnet. These display similar Rb-Sr behaviour to outer aureole samples with the exception that ~1322 Ma biotite domains display higher Rb/Sr due to more extreme loss of Sr. In the inner aureole, where garnet has been completely consumed by contact metamorphic assemblages, a new generation of biotite neoblasts occurs in textural equilibrium with Opx + Crd. This biotite preserves Rb-Sr ages ≤1322 Ma with initial ⁸⁷Sr/⁸⁶Sr best interpreted as a mixture of radiogenic Sr accumulated in regional biotite and whole-rock Sr liberated from low-Rb/Sr regional metamorphic garnet, apatite, and plagioclase. This study reveals how the exact textural setting of biotite in high-grade metamorphic rocks influences the preservation of Rb-Sr ages and demonstrates that there is no universal closure temperature for biotite Rb-Sr. It also reveals that in situ Rb-Sr dating of granulite or UHT rocks might provide robust chronometric data if grains isolated from intergranular diffusion are systematically evaluated to reveal zones of partial retention.

Received: 18 Dec 2025 – Discussion started: 31 Dec 2025

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

Preprint (PDF, 3622 KB)

Supplement (98 KB)

Download & links

Christopher R.M. McFarlane

Status: final response (author comments only)

RC1:
'Comment on egusphere-2025-6320', Kyle Larson, 05 Feb 2026
The work submitted is exactly the kind of study that needs to be carried out as the broader community continues to apply in situ beta-decay geochronometers to address geological problems. I find the work to be of high quality, from the experimental setup through to discussion.
As with any study, there are a few changes that could be made that would further improve the work and ensure consistency in interpretation.
The authors describe mineral assemblages and textures in detail, yet they are supported by a single mXRF map and 4 ppl images. It would be appreciated if the descriptions could be better illustrated by additional images either in the main text or as supplemental files so the reader can see what is being described. Directly parallel to this is better referencing of all supplied figures within the text.

Some of the main points of the paper are based on the interpretation of model Rb-Sr dates; however, detailed information about how the dates were calculated is not provided. The existing information in the Methods section is vague and does not articulate specifically how uncertainties were propagated. Given the central role these dates play in the overall narrative, their calculation needs to be well documented.

Finally, in the finalization of the Rb/Sr data, the data from the mica in the study are normalized against a biotite reference material of known age to compensate for a systematic offset between the primary reference material (pressed nanopowder) and actual mica grains. It is specifically noted that no additional uncertainty was added to the unknowns during the final normalization step. It is later noted in the discussion that the biotite reference material returned a date of 452 ± 7 Ma. That uncertainty is significant and, if converted to uncertainty on the slope of the 452 Ma isochron, it would equal ~ 0.777%. Given that all unknown mica are normalized relative to that 452 ± 7 Ma result, it is important to propagate the uncertainty associated with that normalization through to the final dates. Doing so has direct implications for interpretation. For example, an isochron date of 1338.4 ± 6.5 is noted as being older than the expected 1322 Ma date of pluton intrusion (source of head for the contact metamorphism), indicating something within the Rb-Sr systematics. If the uncertainty on the matrix normalization is propagated to the slope on the isochron in quadrature, however, the uncertainty increases to 21.60 Ma, which overlaps the expected 1322 Ma date. A similar interpretation is made about the 1332.2 ± 8.4 Ma date, which, once the uncertainty on the matrix normalization is propagated through, has an uncertainty of 22.16 Ma, which also overlaps with the 1322 Ma expected date. Perhaps there are other, more appropriate ways of propagating the uncertainty from the isochron used to matrix normalize the mica data, but regardless of how it is done, it should be incorporated into the study so that nuanced interpretations can be made.

I want to again emphasize the need for this kind of systematic study for these developing methodologies and my appreciation to the author for taking the time to do it.
Citation: https://doi.org/10.5194/egusphere-2025-6320-RC1
- AC1:
  'Reply on RC1', Christopher McFarlane, 07 Feb 2026
  
  Many thanks for the constructive comments.
  1. On the question of additional figures and better referencing to figures in the text: I was reluctant to add too many figures in order to maintain a concise presentation. However I can see the value of an additional figure of BSE images that might reveal more clearly some of the textures encountered in these rocks. This is easily done if the editor feels this would improve the manuscript. I will also double-check that figures are better alluded to in the text.
  2. A statement about the calculation of model ages can be added to the methods sections. This is simply equivalent to a two-point isochron between the measured Rb-Sr ratios and an assumed initial 87Sr/86Sr. For all analyses an initial value of 0.73 is assumed. This initial of is probably inappropriate for biotite neoblasts in the inner aureole but it makes the discussion a bit easier to approach. The errors assigned to these model ages are proportion to the error on the measured 87Rb/86Sr. There were no data with sufficiently low 86Sr to assume that all 87Sr was radiogenic.
  3. I fully appreciate the issue of propagating error for this type of Rb-Sr analysis. I feel this is currently one of the outstanding issues that needs to be addressed in the community (and I appreciate KL's efforts to lead this charge). But I think it is more nuanced than simply adding an excess error based on the secondary standards. For example, the largish error on the isochron age for the Loch Borralan is primarily a result of the narrow spread of 87Rb/86Sr; this reaches a maximum of about 75. In contrast, the biotite in the MLP aureole ranges up to 87Rb/86Sr of >2000. So there is a significant mis-match between the material I used for correction of Rb/Sr mass bias (i.e. Loch Borralan) and the unknown targets. Propagating the error from the former unto unknowns could arguably add inappropriate excessive error; this could potentially obscure real (and subtle) variability. So I consider Loch Borralan phlogopite as an standard to ensure accuracy of the results (by correcting for bias). In reality, I think a standard used to assess propagated error should have the same spread of Rb/Sr as the unknowns. Unfortunately such a material doesn't currently exist. For example, if Loch Borralan contained grains with Rb/Sr between 500 and 1500 the error on the isochron age could easily be reduced to <1%. In that case, error on the standard could be more usefully propagated to the unknowns. So I am a bit reluctant to add significant excess error to the MLP isochrons using the error measured on Loch Borralan. It is easily to incorporate language to the effect described above. It is also easy to cite the long-term reproducibility of MicaMg-NP and/or NIST610 analyzed during each session. I am open to suggestions on this matter.
  
  Citation: https://doi.org/10.5194/egusphere-2025-6320-AC1
  - RC2: 'Reply on AC1', Kyle Larson, 13 Feb 2026
    
    I appreciate the author's comments on uncertainty propagation and the potential complications. Accounting for the long-term reproducibility would be a positive step forward in this regard, and it may also be worth looking into modelling the potential effect the uncertainty on the matrix correction material age might have for specimens of different ages and Rb/Sr. That all being said, I do not foresee such changes forcing any substantial revisions to the main conclusions of the study.
    
    Citation: https://doi.org/10.5194/egusphere-2025-6320-RC2
    
    AC2: 'Reply on RC2', Christopher McFarlane, 20 Feb 2026
    
    Yes, I believe there is scope to investigate modelling how the age and Rb/Sr spread of primary and secondary standards affects propagated errors on unknowns; this should be fruitful to explore in the future. In the revised manuscript the final report ages will include additional error reflecting dispersion of the primary reference material as well as an additional 1% error on isochron ages to reflect the long-term reproducibility of the secondary standards used.
    
    Citation: https://doi.org/10.5194/egusphere-2025-6320-AC2
RC3:
'Comment on egusphere-2025-6320', Anonymous Referee #2, 19 Feb 2026

Most of the comments that I have are pretty minor. I really like this case study, and I think the partial resetting is very clearly shown. The only, and I mean the only thing that I think makes this not publishable in its current form is the lack of secondary reference materials. This needs to be addressed.
First question – how many secondary reference materials were analyzed, and did their ages overlap with known ages? There is reference to a single in-house biotite material from Scotland, but no information is given about it.
Second question – if the material from Scotland, which has not been shown to be a reference material, since it isn’t reported to have a known age, is the only secondary material analyzed, then how can we trust the ages? It appears that the method of Giuliani et al. (2024) is applied (please cite here) in order to do an isochron-based correction on the 87Rb/86Sr ratios. This is a great approach, but then it has to be applied to at least one (if not more) secondary reference micas to show that the correction works properly. This is not demonstrated. So, there is no traceability for the geochronology. The dates that are reported are absolutely fine relative to one another, and I think the consistency with the SHRIMP dating suggests that there is no significant bias, but without independently showing that the ages of secondary micas can be determined, then the foundation of the geochronology is shaky. I think the details of the age correction need to spelled out (and I agree with the previous comment about propagating uncertainties), and secondary reference biotite grains of known age (not in-house materials) need to be dated to determine the validity of the method.
The other thing that I think would really help strengthen this study overall would be to include some of the trace elements in the analysis. In my lab we have shown without a doubt that Rb does not react with SF6 – the author also points this out. You can save counting time by not measuring 104RbF and instead count for 5 ms on Li, Cs, Ba, and a number of other elements. Figure 9 is good, but it would also be interesting to see what these other trace elements are doing during this. We have observed some really interesting patterns between TE and partial resetting of micas in the Alps, and Cs and Ba are particularly telling. Obviously this can’t be added after the fact, but it might be useful to note for the future. See work by Kunz and others on TE in micas.
Other comments:
Italicize in situ throughout, as is the standard for Latin words
Line 10 and throughout: I suggest that LA-ICP-MS/MS is the accepted abbreviation for this type of tandem mass spectrometry. That is what we have tried to establish in GGR. It allows for the addition of MC on the beginning, if needed, and refers to any MS/MS machine. Also, the ICP is always before the MS/MS anyway, so I would argue that it is more physically correct as well.
In the abstract, the author goes back and forth between past and present tense. I suggest keeping things present tense except for methods that were performed. For example, line 16 should be retain, not retained, as these biotites display in the previous sentence.
Introduction – might be helpful to add a little bit of the debate about the experimental constraints here too (Hofmann and Giletti 1970 vs Hammouda and Cherniak 2000), and the work of Villa 2016. I know that the experimental constraints are a mess, but it’s worth noting that as an important part of why the natural laboratory provides better constraints for real-world diffusion in micas
Line 51 – this might be a good place to mention that high precision is achieved by LA-MC-ICP-MS/MS (Bevan et al. 2021, Craig et al. 2021, Dauphas et al. 2022, Cruz-Uribe et al. 2023, Huang et al. 2025). So the precision on an individual spot by quadrupole is lacking, but the in situ Rb-Sr method has the potential for extremely high precision. I think it is also worth citing a little bit more of the in situ Rb-Sr literature here, particularly studies such at Kutzschbach et al. 2024 (mapping), Giuliani et al 2024 (data processing), Stijn Glorie, etc. Also, since this manuscript is about single spot model dates, it is important to cite Rösel and Zack (2022) and Cruz-Uribe et al. (2023) in this context.
Line 60 – I think this is a great test case. We’ve been hunting for years for the best way to isolate the thermal component, and this seems to be as good as it gets
Figure 3 – all of the minerals are labeled except the biotite
Line 201 – question about P/A factors: I think we’ve all been having this discussion in the 8900 community for a while. As far as I can tell in talking with folks in the factory in Japan, it is not actually possible to tune the P/A factors in mass-shift mode. After years I still haven’t been able to force my machine to do this. I think we all have different ways of dealing with it, and I just wondered if you had a magical trick to make it work! I’ve found that the precision of the P/A tune (which is done without gas, despite trying to trick it into using gas) is not good enough to have linearity in the isotope ratios. I don’t think it’s an issue with anything you’re doing, I’m really just curious if you’ve found a way around the factory problems.
Lines 205-220 please refer to these as reference materials as opposed to standards. Thank you from the IAG and ISO.
Line 215 – we have also noticed a pretty big difference in ablation between plagioclase and glass, particularly NIST glass, using a ns laser. Have you thought about using BCR-2G or BHVO-2G for the Sr isotopes in plag?
Additional methods comments – I assume you are using 85 to calculate 87Rb, as the reaction is very efficient but not complete. Please give these details, and cite Meija for the IUPAC 85/87 ratio. Please also give ICP-MS dwell times for each isotope
Section 5.1 – these are great. I think the approach is very similar to Cruz-Uribe et al. (2023) who used KDE plots to examine the distribution of single integration dates across multiple spots in sample. I think it is important to cite that method here. I realize these are single spots instead of single integrations, but I think it is a similar enough approach that it warrants a citation. I like the statistical approach that you are taking though, and I am glad to see these types of approaches being used to interrogate geochronologic data. This is really important for the Rb-Sr system.
Figure 6 – it would be helpful to have the label of inner aureole on the figure itself – that was helpful on the previous two figures. The same would be helpful for Figure 8 (melt film pseudomorphs at contact)
Lines 350-355. I think it’s also worth mentioning here that the chemical potential gradient that is the driver for diffusion might induce back-diffusion of Sr into biotite if the Sr is more compatible in biotite than in garnet (or result in basically no chemical potential gradient). I don’t know the relative compatibilities of Sr in grt and bt off the top of my head, but in measuring both, my instinct tells me that the Sr is perfectly happy in bt, and so there might not be a driver for diffusion out of the biotite. Then it doesn’t matter what the diffusion rates are, if there is nothing actually driving diffusion. Now, biotite in the matrix would be a completely different story, because now we have potential fluids within grain boundaries, minerals like plag, etc, that certainly have higher distribution coefficients for Sr than biotite would.
Lines 390-405. This is very interesting. Cruz-Uribe et al. (2023) address this concept as well, and I think a citation of Davies et al. (2018) would not go amiss here.

Citation: https://doi.org/10.5194/egusphere-2025-6320-RC3
- AC3: 'Reply on RC3', Christopher McFarlane, 27 Feb 2026
  
  I’ll try to answer the important reviewer questions 1 and 2 at the same time. First, yes, it is absolutely the case that we analyze numerous in-house micas from well-characterized settings with each run. At the time of collecting the data for this study (that I did for Goldschmidt 2023 in Lyon) I did not yet have any grains of La Posta to use as a matrix-correction reference material (MCRM in Stijn Glorie’s nomenclature). However, I did have a colleague (A. Vezinet) at ISTERRE Grenoble analyze the Loch Borralan phlogopite as part of a batch of other possible reference materials. When La Posta is used as MCRM the Loch Borralan returns an Rb-Sr isochron of 416 ± 15 Ma and SrI = 0.7051±0.0026. So I can refer to Loch Borralan phlogopite as a in-house material traceable to La Posta. I have recalculated all mass-bias corrected ages in the MLP to this value using NIST610 as primary standard. This brings the method in line with more recent studies and recommendations. This has only a slight effect on the results and does not affect the discussion or conclusions. The new figures and calculated age are very subtly different (you'd need a magnifying glass to see the changes in the figures). Secondarily, the MLP aureole provides its own check on the accuracy of the analytical approach and DRS since the ages in the aureole are already tightly constrained; this of course is the main reason for studying these micas. As mentioned in previous response to reviewer 1, additional error from dispersion of primary standard (now using NIST610) and an additional 1% error added to cover secondary material reproducibility is incorporated. However I think it is important to include a sentence or two highlighting the challenges that remain for full error propagation based on secondary material with limited ranges of Rb/Sr relative to unknowns.
  With respect to adding additional trace elements to the mass list - yes in total In agreement. Our current methodology spends more detector time analyzing a broader range of minor and trace elements including Ti in addition to Li, Cs, Ba. Future studies published from our lab will include this.
  Other minor details will be cleaned up. These include switching to LA ICP-MS/MS; italicizing in situ (pending editorial approval); fixing verb tense switching in abstract; more carefully describing reference materials vs. in-house materials.
  I am hesitant to launch into a review of experimental vs. empirical studies in this study as I would prefer to do the topic more justice in future works. But the point is well taken
  Line 51 and forward; this section will be recrafted to highlight the potential for high-precision single-spots with reference to the LA MC-ICP-MS/MS literature as well as additional references to some key LA ICP-MS/MS papers where robust Rb-Sr isochrons are reported.
  Figure 3 labeling of Bt will be fixed
  With respect to P/A factors in mass-shift mode, I wonder if this was a software glitch in the earlier version of MassHunter? As far as our methodology goes MassHunter is perfectly happy to derive P/A factors in mass-shift mode; it also provides a timestamp of when they were last updated. There are a few cryptic boxes to check and settings to apply, but the software looks to be doing what it is told to do. We have never encountered top-hat time-series signals that indicate non-linearity across the 1.3Mcps cross-over (which sometimes might happens with 107SrF if you hit an apatite inclusion). This is also a reason I like to examine MBC 88Srs/86Srs since non linearity would be visible in this ratio (mass-shifted 88Sr and 86Sr are typically in A and P mode respectively on NIST610 so any non-linear behaved would totally mess with the final ratio).
  Line215 - comment about 87Sr/86Sr in plagioclase. The difference in ablation between matrices will be magnified as a function of ablation conditions. We have tuned Sr isotope measurement settings to minimize bias between NIST61x and feldspars in general. We observe that both plag and kfs require higher fluence and lower repetition rates compared to micas (so we don’t do these in the same run). Residual bias is easily checked by comparing the raw vs. corrected ratios and noting only subtle differences in the 3^rd decimal place. A lot of effort has been spent to get these ablation conditions correct for each standard-sample combination.
  A reference to the canonical 87Rb/85Rb value has been added to the methods section
  Section 5.1 Linearized Probability Plots vs. KDE. I personally don't recall seeing Linearized probably plots being used to examine Rb-Sr model ages. I get the point that KDE are similar. I can certainly make reference to Cruz-Uribe et al. (2023) using KDE to examine heterogeneous model age distributions.
  Figures 6, 7 8 will be edited to make sure they are labeled appropriately
  Lines 350-355. This is a fair comment and it might take a couple of extra sentences to communicate this argument. Looking at Kds in GERM database suggests no particularly affinity for Sr in either garnet or biotite (although solid-melt Kds for biotite in felsic melts are significantly higher than for garnet). So yes, I could also argue that despite the pre-existing chemical gradient there may be no driving force for Sr diffusion - this would be on top of the low diffusivity of Sr in garnet.
  Lines 390-405. Yes, regret that I should have added reference to Davies etal (2018) to reference that these concepts are not new.
  
  Citation: https://doi.org/10.5194/egusphere-2025-6320-AC3

Christopher R.M. McFarlane

Supplement

https://doi.org/10.5194/egusphere-2025-6320-supplement

Christopher R.M. McFarlane

Viewed

Total article views: 383 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
247	93	43	383	27	235	113

HTML: 247
PDF: 93
XML: 43
Total: 383
Supplement: 27
BibTeX: 235
EndNote: 113

Views and downloads (calculated since 31 Dec 2025)

Month	HTML	PDF	XML	Total
Dec 2025	24	3	1	28
Jan 2026	137	54	15	206
Feb 2026	86	36	27	149

Cumulative views and downloads (calculated since 31 Dec 2025)

Month	HTML	PDF	XML	Total
Dec 2025	24	3	1	28
Jan 2026	137	54	15	206
Feb 2026	86	36	27	149

Viewed (geographical distribution)

Total article views: 359 (including HTML, PDF, and XML) Thereof 359 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 28 Feb 2026

Short summary

The mineral biotite is a common rock-forming mineral and notable for its ability to incorporate Rb into its structure. Microanalytical tools such a laser ablation and tandem mass spectrometry, allows Rb-Sr biotite dating with context preserved. This study reveals Sr mobility that is controlled by location in a rock. Reconstructing the timing of geological events using biotite Rb-Sr geochronology is, therefore, contingent on grain-scale evaluation of biotite textures.


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