Single Spot Rb‐Sr Isochron Dating of Biotite by LA‐MC‐ICP‐MS/MS

Laser ablation tandem mass spectrometry is a burgeoning field for in situ Rb‐Sr geochronology. Here, we determined simultaneous isotope ratios of 87Sr/86Sr and 87Rb/86Sr in metamorphic biotite from western Maine, using an ESL™ imageGEO™193 excimer laser ablation system coupled to a Thermo Scientific™ Neoma™ MC‐ICP‐MS/MS. Measurements were made on Faraday cups with Rb+ at mass 87; Sr isotopes were reacted with SF6 gas and measured as SrF+ at masses 103–107. Twenty‐two laser spots in biotite from a single sample yield a "traditional" Rb‐Sr isochron date of 289 ± 6 Ma. Time‐resolved signals reveal significant zoning in 87Sr/86Sr and 87Rb/86Sr within single spot analyses, which were used to construct single spot isochrons. Individual laser spots contain multiple isochronous subpopulations; some spots contain up to three distinct Rb‐Sr isochrons that are decoupled from variations in Rb/Sr. Thirty‐five isochron dates were determined using this "sub‐spot" approach, with 87Sr/86Sr intercepts that systematically vary with Rb‐Sr date; two‐point isochrons were calculated for individual integrations (n = 780) based on these variable intercepts. Both methods yield age peaks at 303, 270 and 240 Ma. These data suggest that the Rb‐Sr system has the potential to record multiple heating, cooling or fluid‐alteration events spanning ~ 100 My within small domains in single biotite crystals.

The 87 Rb- 86 Sr system is a pillar of isotope geochronology that has been utilised to address numerous questions within the geosciences (e.g., Veizer 1989, Turchyn and DePaolo 2019, Fantle et al. 2020) as well as other fields such as biology and anthropology, where it serves as an effective environmental tracer (e.g., Bartelink and Chesson 2019, Coelho et al. 2017, Holt et al. 2021).Because multiple generations of Rb-bearing minerals (primarily micas and feldspar) are often observed in deformed and metamorphosed rocks, the Rb-Sr system has significant potential to address outstanding tectonic problems, particularly to quantify material flow paths (P-T-t) and tie them to their underlying tectonic mechanisms.The advent of tandem mass spectrometry with collision/reaction cell technology presents an exciting alternative to bulk crushing and wet chemistry for determining Rb-Sr isochron dates by enabling the inline separation of isobaric interferences (e.g., 87 Sr from 87 Rb) and thus facilitating laser-based, in situ Rb-Sr approaches (e.g., Zack and Hogmalm 2016, Hogmalm et al. 2017, Gorojovsky and Alard 2020, Redaa et al. 2021, R€ osel and Zack 2022, Scheiblhofer et al. 2022).Most recently, the addition of pre-cell mass filters for multi-collector collision cell ICP-MS, in the form of a quadrupole (e.g., Proteus, Bevan et al. 2021) or Wien-style filters (e.g., prototype Vienna instrument, Craig et al. 2021, Thermo  Scientific TM Neoma, Dauphas et al. 2022) allows for simultaneous collection of Rb and mass-shifted Sr isotopes.
The isochron diagram (or Nicolaysen diagram) is the fundamental tool for determining dates in many geochronological systems (e.g., Rb-Sr, Sm-Nd, Re-Os, Lu-Hf, U-Pb; Nicolaysen 1961).One basic tenet of the isochron calculation is that all of the uncertainty of the isochron fitand thus the uncertainty on the sample datecan be attributed to the analytical uncertainties of the measurements (i.e., one assumes that every sample along an isochron is isochronous).The conventional approach to laser-based Rb-Sr isochron dating is to average the integrations within a single laser spot analysis (e.g., Zack and Hogmalm 2016, Hogmalm et al. 2017, Tillberg et al. 2017, 2020, Gorojovsky and Alard 2020).Recently, R€ osel and Zack (2022) proposed single spot Rb-Sr isochron dating using an assumed, lithology-based initial 87 Sr/ 86 Sr intercept.However, with the improved precision and simultaneous measurement of Rb and Sr isotopes on the Neoma MC-ICP-MS/MS, it is now possible to examine sub-spot integrations within individual laser analyses and to determine the robustness of this assumption for in situ Rb-Sr data.
Here we present single spot biotite Rb-Sr isochrons measured on the Neoma MC-ICP-MS/MS that reveal complex cooling information spanning ~100 million years within a single sample and even within single spots.The resulting single spot 'multichrons' demonstrate that is possible for biotite to contain multiple date domains within a single, 100-lm laser ablation analysis.Our data further suggest that the uncertainties on individual data points on many mica Rb-Sr isochrons may hold additional age information that is hidden by the conventional approach to laser spot analyses.The high-precision simultaneous isotope measurement approach enables determination of initial 87 Sr/ 86 Sr based on the fit of individual isochrons within single laser spots, and because of the large range of Rb/Sr ratios in micas such as biotite, these minerals open up a new microsampling frontier for isochron-based geochronology.

Sample description
Biotite grains in garnet-bearing schist Ra-D72 from the Mooselookmeguntic contact aureole, western Maine, USA, were analysed in this study (see Figure S1 for simplified geological map).Pelitic and psammitic rocks of the Central Maine belt of the Northern Appalachians underwent regional metamorphism (M2) during the Acadian orogeny ca.405-400 Ma (Smith and Barreiro 1990, Solar   1. et al. 1998et al. , Johnson et al. 2003)).The Mooselookmeguntic pluton, a two-mica granite, intruded the regionally metamorphosed rocks ca.370 Ma (Smith and Barreiro 1990, Solar et al. 1998, Tomascak et al. 2005).Thermal metamorphism associated with the intrusion (M3) created a contact aureole with well-defined zones from lower garnet to upper sillimanite within ~600 m of the contact.The M3 thermal metamorphic event is associated with pseudomorphism of M2 garnet and staurolite porphyroblasts by chlorite, muscovite and/or biotite, depending on the location within the aureole (e.g., Guidotti 1970, Guidotti et al. 1996, Guidotti and Johnson 2002).Sample Ra-D72 is a pelitic schist from the garnet zone of the Mooselookmeguntic pluton.The matrix is composed primarily of white mica, quartz and plagioclase (~50-100 lm grain size), with porphyroblasts of garnet and biotite.Large biotite porphyroblasts (1-3 mm) that are the subject of this study are present throughout the matrix; these biotite grains grew and were subsequently deformed during regional metamorphism of the Acadian orogeny (M2).
Previous constraints on the thermal history of plutons from western Maine include 40 Ar/ 39 Ar analyses of hornblende, feldspars and micas from a variety of igneous and metamorphic rocks.Lux and Guidotti (1985) reported 40 Ar/ 39 Ar cooling spectra from hornblende from Mooselookmeguntic and the nearby Songo pluton with anomalously old apparent ages in the low-T initial increments of the step heating analyses, with high-T gas release plateau dates of 359, 330 and 305-310 Ma. De Yoreo et al. (1989b) reported one hornblende 40 Ar/ 39 Ar plateau date of 372 AE 8 Ma from the Mooselookmeguntic pluton, which is indistinguishable from the U-Pb zircon crystallisation age (Tomascak et al. 2005); incidentally, of the previously dated samples, this sample is geographically located closest to sample Ra-D72 presented in this study.Four other hornblende 40 Ar/ 39 Ar dates from the Mooselookmeguntic and Rumford plutons and Ammonoosuc Formation range from 328-322 Ma, which have been attributed to a Carboniferous metamorphic event (De Yoreo et al. 1989b).Muscovite 40 Ar/ 39 Ar plateau dates yield wide ranges from the Mooselookmeguntic pluton (238-304 Ma), Phillips pluton (272-371 Ma) and Sebago pluton (240)(241)(242)(243)(244)(245)(246).Biotite 40 Ar/ 39 Ar plateau dates also yield a wide range from Mooselookmeguntic (255-293 Ma), Phillips (288-353 Ma) and Sebago (229-245 Ma) (Lux and Guidotti 1985, De Yoreo et al. 1989a, 1989b).

Methods
Laser ablation MC-ICP-MS/MS analyses were performed in the application laboratory at the Thermo Fisher Scientific TM factory in Bremen, Germany, using an ESL TM imageGEO TM 193 excimer laser ablation system equipped with a TwoVol3 TM (TV3) ablation chamber.The analytical cup of the TV3 was connected to a Thermo Scientific TM Neoma TM MC-ICP-MS/MS (referred to as Neoma MS/MS in the following; Figure 1) via ~2 m of 1 mm ID tubing.A rock chip of sample Ra-D72 was prepared as a 2.54 cm epoxy mount cut perpendicular to the foliation, ground flat using silicon carbide paper, and polished using diamond suspensions (9-1 lm) and a 0.3 lm alumina slurry.Biotite was ablated using a 100 lm round spot at 12 Hz repetition rate with a beam energy density of 3.5 J cm -2 .Each analysis consisted of 20 s of background collection during laser warm-up, 40 s of ablation and 20 s of washout.The MicaMg nanoparticle pressed powder pellet (Garbe-Sch-€ onberg et al. 2014) was analysed as calibration material before and after every ten to twenty unknowns.Biotite from the LaPosta pluton was run as a quality control material and gave an Rb-Sr isochron date of 90 AE 4 Ma (Figure S2), which is within uncertainty of the multi-mineral Rb-Sr isochron date of 93.8 AE 2.5 Ma (Walawender et al. 1990).Biotite GA1550 was also dated as a quality control material at 101 AE 10 Ma (Figure S3), consistent with previously published Rb-Sr isochron ages of 100.3 AE 1.9 Ma (Williams et al. 1982) and 100.4AE 0.7 Ma (Li et al. 2008; ages recalculated by R€ osel and Zack 2022 using the decay constant of Villa et al. 2015).
The Neoma MS/MS and imageGEO were controlled using the Qtegra TM ISDS v2.17 and ActiveView2 TM software.Complete analytical settings for the imageGEO and Neoma MS/MS can be found in Table S1.The TV3 ablation cell used a He flow rate of 1 l min -1 and a N 2 flow rate of 2 ml min -1 to transport the ablated material to the ICP.Prior to the ICP, an additional 0.86 l min -1 of Ar was added to the cell outflow from the sample gas port of the Neoma MS/MS.
The pre-filter settings of the Neoma MS/MS were identical to those reported in Dauphas et al. (2022).The center mass of the transmission window was selected to be 87 Sr using a Wien electric field of 320 V.The set Wien magnetic field of 50% (equivalent to 0.315 T) and pre-filter slit position of 60% restricted the transmitted range to approximately 82-94 m/z.Restricting the transmitted range was essential for reducing molecular interferences such as 32  and 88 SrF + respectively in low resolution (DM/M = 2100) measured on plateau (Weyer et al. 2003).
The Sr + within the transmitted mass range was fluorinated to SrF + in the collision/reaction cell by the addition of 0.075 ml min -1 SF 6 .Conditions within the collision/reaction cell were tuned for maximum Sr sensitivity. 87Rb, which does not react with SF 6 , was measured on mass.
The Jet Interface was used with the X-skimmer cone and Jet sampler cone to maximise sensitivity.The RF power was 1300 W and the torch sampling depth 7.00 mm.This experiment took advantage of some significant improvements that were not possible on the pre-commercial version of the Neoma MS/MS used in Dauphas et al. (2022).All previously reported in situ Rb-Sr dating experiments by MC-ICP-MS/MS (Bevan et al. 2021, Craig et al. 2021and Dauphas et al. 2022) have had insufficient detector array dispersion to achieve static collection of 85 Rb + to 88 SrF + .As such the MC-ICP-MS/MS had to be operated in dynamic peak jumping mode, introducing spectral skew (van Malderen et al. 2018) and significant reduction in dwell time from the expected 100% of static multi-collection.However, the commercial Neoma MS/MS used here is capable of statically collecting 85 Rb -88 SrF (Table 1), although we used 87 Rb -88 SrF in this study.Subsequent work has shown no difference in isochrons within uncertainty between collecting 85 Rb and 87 Rb.Integration times for all isotopes were 1 s.Sensitivity was significantly higher when utilising full static collection mode; the 88 SrF + signal intensity of 8.7 9 10 7 cps (or 1.39 V) for the NIST SRM 610 reference glass measured in this study is significantly greater than the 4.7 9 10 7 cps (or 0.75 V) previously achieved in peak jumping mode.This increase in measured signal intensity occurred despite the amount of material ablated (100 lm round spot, 12 Hz repetition rate, beam energy density of 3.5 J cm -2 ) being reduced by approximately 50% (previously 100 lm round spot, 20 Hz repetition rate, beam energy density of 5.0 J cm -2 ).
Due to the expected low Sr mass fraction within the biotite unknowns, the best possible precision in 87 Sr/ 86 Sr was achieved by collecting both 86 SF + and 87 SrF + on Faraday detectors connected to 10 13 Ω amplifiers.These amplifiers saturate at 6.24 9 10 7 cps (1 V, 10 11 Ω scale), but this value was never approached for any ablation of sample or reference material.All other nuclides were collected on Faraday collectors connected to 10 11 Ω amplifiers, which saturate at signal levels two orders of magnitude higher than the 10 13 Ω amplifiers.Time resolved analyses were processed using a new Rb-Sr data reduction scheme (DRS) in iolite 4 (Paton et al. 2011) (see Appendix S1).Data from the Neoma MS/MS analyses were time-synchronised using laser log files.Background selection consisted of eight seconds, sampled from ten to two seconds before the start of each ablation.Due to the known issues of correcting down-hole inter-element fractionation of Rb from Sr in micas and the MicaMg nanopellet (e.g., Redaa et al. 2021), no down-hole fractionation correction was applied, and all spot analyses were cropped to exclude the first three seconds of ablation.The down-hole fractionation of Rb from Sr during ablation is still an issue; however, by cropping the most extreme fractionation at the beginning of each ablation, these effects were significantly minimised.This method is considered fit-for-purpose for this study, as we reproduce the dates of the above reference micas within the analytical uncertainty of the analyses.An isochron for a single laser analysis of LaPosta biotite is shown in Figure S4, which demonstrates that the date for a sample of known age does not change down-hole.All reported 87 Sr/ 86 Sr ratios were corrected for instrumental mass bias assuming the exponential law (Avanzinelli et al. 2005) The iolite 4 Neoma Rb-Sr DRS functions as follows: (1) The DRS begins by baseline-subtracting all input channels and optionally applying a mask to background intervals to avoid large variations in ratio channels.(3) A mass fractionation factor is then determined from the mass-shifted 88 Sr/ 86 Sr ratio, using the Russell equation (Russell et al. 1978).Mass fractionation corrected ratios for 87 Sr/ 86 Sr, 87 Rb/ 86 Sr and 84 Sr/ 86 Sr are then calculated.(4) These mass fractionation corrected ratios are then normalised to the accepted value of the calibration material results, which are interpolated using a smoothed cubic spline.Additionally, the raw (un-mass fractionation corrected) 87 Sr/ 86 Sr and 87 Rb/ 86 Sr ratios are similarly normalised to the accepted value of the primary calibration material to monitor the effect of the mass fractionation correction.
A relative standard error for 87 Sr/ 86 Sr and 87 Rb/ 86 Sr was assigned for each individual 1-second integration by combination of Poisson counting statistics on the count rates (Cottle et al. 2009) with the contribution from Johnson-Nyquist noise calculated for each associated amplifier (Koornneef et al. 2014, Kimura et al. 2016), assuming they are the only sources of random noise (Saji et al. 2016, Coath et al. 2017).Rb-Sr isochron dates were determined using IsoplotR (Vermeesch 2018) using model 1 (maximum likelihood) for traditional Rb-Sr spot dates, and models 1 and 3 (maximum likelihood with overdispersion) for single spot dates.
Trace element maps of a characteristic biotite grain in sample Ra-D72 were acquired by LA-ICP-MS at the University of Maine using the methods described in Cruz-Uribe et al. (2021) and Walters et al. (2022).Instrumentation consisted of an ESL TM NWR193 UC 193 nm excimer laser ablation system equipped with a TwoVol3 fast washout ablation cell, coupled to an Agilent Technologies TM 8900 ICP-MS/MS.The analytical cup of the TwoVol3 ablation cell was connected via ~0.5 m of 1 mm ID flexible tubing to an ESI dual concentric injector torch (DCI1).A 3 lm 9 3 lm square spot was rastered in parallel lines across the sample surface using a scan speed of 90 lm s -1 , a repetition rate of 179 Hz and a beam energy density of 2.5 J cm -2 .Two lines were ablated on basalt reference glass GSE-1G from the USGS before and after map acquisition.Time-resolved signals were processed using the Trace Elements DRS, and trace element mass fractions were determined semiquantitatively relative to the GSE-1G glass reference material.Trace element maps were produced using CellSpace in iolite 4 (Paton et al. 2011, Paul et al. 2012, Petrus et al. 2017, Woodhead et al. 2007).

Results and discussion
Laser spot means and individual integrations for twentytwo analyses of 87 Rb/ 86 Sr and 87 Sr/ 86 Sr in biotite can be found in online supporting information Tables S2-S3.Individual integration values for 87 Rb/ 86 Sr range from 248-10275, and 87 Sr/ 86 Sr values range from 1.72-38.8,which provides significant spread in the determined Rb-Sr isochrons.An isochron composed of fully integrated spot data was determined using the entire dataset and is shown in Figure 2. Thirty-five single spot Rb-Sr isochrons (extracted from individual laser spot analyses, in some cases sub-spots) were determined and are given in Table 2 and shown in Figures 3-6; these calculations are described in more detail below, as are two-point Rb-Sr isochron dates calculated for each individual integration.Single spot isochrons and the two-point individual integration isochrons yield dates of 204-324 Ma.
The traditional approach to laser-based Rb-Sr isochron dating, in which individual laser spots are fully integrated and plotted on an isochron diagram to determine a date, gives a date of 289 AE 6 Ma for twenty-two laser spot analyses of biotite grains in sample Ra-D72 (Figure 2a).In this approach, time-resolved signals are trimmed to include only the most homogenous parts of a given spot analysis, such that the uncertainty of each spot is minimised and results in smaller uncertainties on the mean isotopic results.However, analysis of the biotite grains in this study reveal that each spot analysis is not homogeneous, and the untrimmed, fully integrated signals for any given spot result in many spots with large uncertainties due to the wide range of 87 Rb/ 86 Sr and 87 Sr/ 86 Sr (Figure 2b).Due to the precision and simultaneous measurement of Rb and Sr isotopes on the Neoma MS/MS in this study, it is possible to examine each integration cycle within a single spot analysis (for instance, arrows in Figure 2b  individual integrations to construct "sub-spot" isochrons from each laser spot reveals complicated but resolvable date information.For example, the zoning present in the time resolved signals for Rb and Sr in laser spot 18 (Figure 3a) results in domains in each spot with distinct 87 Rb/ 86 Sr (grey open squares, Figure 3b) and 87 Sr/ 86 Sr.In spot 18, there are two zones that are isochronous within analytical uncertainties, with the first 14 integrations having lower 87 Rb/ 86 Sr relative to the later integrations.A single isochron date of 307 AE 4 Ma was determined for spot 18 (Figure 3c).The intercept defined by the isochron was used to calculate two-point isochron dates for each integration, which are plotted for reference in Figure 3 (b, filled circles).
Eight laser spots, including spot 18, have zoning in 87 Rb/ 86 Sr that is sufficiently broad to enable the determination of single spot isochron dates.Four such examples are shown in Figure 4. Three clusters of dates emerge from these single spot isochrons: a younger population ca.240 Ma (blue isochrons in Figure 4), a population ca.270 Ma (orange isochron in Figure 4), and a series of older dates from 300-325 Ma (maroon isochrons in Figures 3-4).Four of the ablated spots did not have sufficient spread in 87 Rb/ 86 Sr to calculate isochron dates and are not included in Table 2 (spots 12, 15, 17, 20).
Eleven laser spots ablated multiple date zones, thereby enabling multiple isochrons to be fitted from a single spot (Figure 4c, Figure 5a-c).Subsets of each laser spot that were statistically isochronous were selected and isochrons were fitted to these integrations.For example, two Rb-Sr isochron dates were determined for spot 5 (Figure 5a-c): integrations 1-18 gave a Rb-Sr date of 239 AE 2 Ma (Figure 5c) and integrations 19-35 gave a date of 295 AE 5 Ma (Figure 5c).The dates that emerge from these 'multichrons' are consistent with the three populations identified in the single spot isochrons (ca.240, 270, 300 Ma), but with a slightly broader series of older dates from 290-325 Ma.We also note that in this particular sample, those spots with multiple age populations consistently progress down-hole from younger to older ages, but emphasise that other biotite grains from other samples measured in the same and other sessions using the same instrumental set-upincluding those that display similar multiple-isochron behaviours within single spotsexhibit older-to-younger behaviours (Cruz-Uribe et al. in preparation).We infer that this relationship arises from how the three grains in Ra-D72 were mounted, i.e., preferentially exposing inclusion-poor rims rather than inclusion-rich cores, such that many spots sample rim-tocore transects.Some single spot (and sub-spot) isochrons have sufficient spread in 87 Rb/ 86 Sr to calculate an isochron date, but have very high 87 Rb/ 86 Sr and no data close to the lower intercept.These spots yield dates that are consistent with the three date populations found within the sample, but with large uncertainties on the isochron fit (e.g., spot 10,  S5).This approach is commonly utilised in U-Pb isochron calculations (e.g., Garber et al. 2017) and minimises the age uncertainty while being faithful to the calculated age.However, in this study we used an initial 87 Sr/ 86 Sr ratio that was calculated from the isochron fitted to each independent sub-spot population, and make no assumptions about intercepts based on rock composition (e.g., R€ osel and Zack 2022).
For the eighteen spot analyses with sufficient spread to calculate single and sub-spot isochron dates, two-point Rb-Sr dates were calculated for each integration projected from the corresponding 87 Sr/ 86 Sr intercepts.The histogram of all single integration dates for sample Ra-D72 is shown in grey in Figure 7. Kernel density estimates (KDEs) were generated using the ksdensity function in MATLAB R2016a using these calculated 87 Rb/ 86 Sr dates for each integration.The solid green KDE shown in Figure 7 represents the calculated optimal smoothing bandwidth for the KDE calculation (10.53 My), whereas the dashed purple and dotted blue lines represent half (5.27 My) and double (21.06 My) this value, respectively.The KDE analysis reveals three distinct dates recorded in this sample at 303 Ma, 270 Ma and 239 Ma.Analytical uncertainties for each calculated Rb-Sr date are typically 10 My or less, suggesting that the presence of three date groupings in the sample is a statistically robust result.
All the integrations within sample Ra-D72 are plotted on a single isochron diagram in Figure 8.This composite isochron diagram results in a geologically reasonable apparent lower 87 Sr/ 86 Sr intercept of 0.7197 for the entire sample.However, investigation of each spot analysis and the ( 87 Sr/ 86 Sr) 0 = 0.13 ± 0.17  The histogram is on a slightly different vertical scale.
fitting of individual isochrons with intercepts based on the isochron fits reveals that the initial 87 Sr/ 86 Sr is variable between isochrons.Figure 9 shows the initial 87 Sr/ 86 Sr plotted versus date for each of the isochrons fitted in this study, with the weighted mean for the three date groupings shown as black open circles with error bars representing one standard deviation for each grouping.In this sample the initial 87 Sr/ 86 Sr systematically increases with decreasing date (Figure 9), though there is significant scatter in the range of initial 87 Sr/ 86 Sr ratios due to extrapolation from high 87 Rb/ 86 Sr to the intercept during isochron fitting.An increase in initial 87 Sr/ 86 Sr with decreasing Rb-Sr date is consistent with the model proposed by Davies et al. (2018), during which partial to complete resetting of the Rb-Sr system drives initial 87 Sr/ 86 Sr towards higher values as isochrons are flattened, and then subsequently steepened with continued radiogenic ingrowth of Sr. Multiple resetting events thus lead to complexities in the initial 87 Sr/ 86 Sr.
We suggest that using the apparent lower intercept for entire population (Figure 8) to calculate single-integration or single-spot Rb-Sr dates is unrepresentative and introduces a potential observational bias, because this approach does not account for the variability in initial 87 Sr/ 86 Sr values for different dates in this sample.For instance, if an anchored 87 Sr/ 86 Sr intercept of 0.7197 is applied to the sample and subsequent KDE, the same three age populations are still present, though the dates are slightly compressed and translated due to the assumed lower intercept for younger dates (Figure S6).As an example, calculated Rb-Sr dates for individual integrations in spot 5 (Figure 5) using the apparent lower 87 Sr/ 86 Sr intercept of 0.7197 from the composite isochron vary by up to 25 My compared with using intercepts determined by the "sub-spot" isochron approach (Figure 5c).On the other hand, the composite isochron could be a potentially useful screening tool to get a sense for the broad range of dates present within a series of analyses.The integrations plotted on the composite isochron clearly encompass isochrons that span tens of millions of years, which would be lost in the traditional isochron approach (Figure 2).

Tectonic implications
The three distinct date groupings that are identified in sample 270,239 Ma) demonstrate the utility of single spot Rb-Sr isochron geochronology.The Rb-Sr dates are 70-120 million years younger than the contact metamorphism associated with the intrusion of the Mooselookmeguntic pluton at 370 Ma, which suggests that the Rb-Sr system is recording multiple events associated with cooling, heating and/or fluid-assisted processes.Unlike bulk Rb-Sr measurements, multiple isochrons spanning tens of Ma can be observed within a single laser spot, with more than 100 million years of geological information preserved in a single sample.From the time resolved signals (Figures 3 and 5) we see that Rb-Sr date is decoupled from 87 Rb/ 86 Sr (Figures 3b and   ).In order to better understand how Rb and Sr were distributed within the biotite grains in this sample, trace element maps of a characteristic grain in this sample were determined by LA-ICP-MS (Figure 10).Textural sector zoning of inclusions is apparent in some biotite grains (e.g., Camilleri 2009); the trace element maps show that Rb, Sr, Ba and Rb/Sr (elemental ratio) zoning mimic the pattern of the textural sector zones.The wide range of Rb/ Sr ratios is consistent with the wide range of 87 Rb/ 86 Sr measured in biotite within individual laser spots and broadly within the biotite population.Future analyses, including Rb/Sr date mapping by LA-MC-ICP-MS/MS, are needed to further elucidate the distribution of different age populations within single biotite crystals.
Finally, these new Rb-Sr data demonstrate three clear age peaks that overlap with previous, regional 40 Ar/ 39 Ar dates from hornblende (305-372 Ma), muscovite (238-371 Ma) and biotite (229-353 Ma) (see Sample description, above).Using the fully integrated isochron age from sample Ra-D72 (289 AE 6 Ma) in combination with the 40 Ar/ 39 Ar dates would nominally suggest a protracted period of slow cooling of these rocks in the middle crust following magmatic activity from ca. 370-325 Ma (Mooselookmeguntic and Sebago plutons).However, the presence of three discrete age populations in single biotite analytical spots suggests a more punctuated thermal and/or fluid record preserved in biotite.A remaining question is whether or not these dates represent discrete geological events that directly caused punctuated geochemical changes (i.e., partial to complete resetting), or alternatively that the biotite (and thus the host rocks) responded in a punctuated fashion to a slow background cooling process nominally implied by the 40 Ar/ 39 Ar dates.Answering these questions requires additional, regional scale analyses of the Rb-Sr system across western Maine, as well as other orogens globally.

Conclusions and outlook
In situ single spot Rb-Sr isochron dating of micas by LA-MC-ICP-MS/MS represents a new analytical and geological horizon for the Rb-Sr system.Single spot Rb-Sr isochrons in biotite from western Maine, USA, record three distinct events that are obscured by the traditional spot dating approach.The punctuated nature of the events recorded by the Rb-Sr system is somewhat at odds with the broad, long-lived cooling histories that are typically recorded by the Ar/Ar system.With their perfect cleavage and propensity to hold ions in interlayer sites, micas are not typically thought of as minerals that record multiple geological events, and are considered to be sensitive to resetting by thermal and fluidmediated processes.Here we show that very distinct isotope ratio and trace element zones are preserved in biotite, marking a departure from the conventional view of Rb and Sr retention in mica minerals.The prospects for single spot isochron dating combined with isotope ratio mapping by LA-MC-ICP-MS/MS open up a new frontier for mica geochronology.We additionally note that the increase in initial 87 Sr/ 86 Sr intercept vs. calculated isochron date observed in the sample studied here suggests that the characterisation of initial 87 Sr/ 86 Sr may not be straightforward in all types of samples, and should be determined as part of the isochron calculation when possible.The precision of the single spot multi-collector method using the Neoma MS/MS permits this determination of initial 87 Sr/ 86 Sr due to the synchronous measurement of all isotopes, enabling single spot isochrons from individual integrations independent of sample composition.
We emphasise the remarkable amount of information and the potential for recovering igneous and metamorphic histories from single spots in single Rb-bearing phases.Many previous Rb-Sr efforts, particularly for bulk-rock-scale analyses, have necessarily assumed equilibrium between coexisting phases in order to recover Rb-Sr isochrons.The advantage of our approach is the ability to sample and obtain meaningful geological ages for \ 100-lm scale zones in micas, which are common in a wide variety of rock types that may not have other dateable phases; modified configuration of the multi-collector array to include ion counters on 86 SrF + and 87 SrF + will enable laser sampling at smaller spatial scales and/or in lower Sr grains, as is common in many mica minerals.We also see very significant potential for simultaneous collection of the LIBS (laser induced breakdown spectroscopy) spectra in order to determine major and trace element mass fractions during time-resolved analyses.The high sensitivity of LIBS for first column elements suggests this will be a promising technique for mica analyses in particular.Table S2.Mass bias corrected LA-MC-ICP-MS/MS data for analyses.Table S3.Mass bias corrected LA-MC-ICP-MS/MS data for individual integrations from twenty-two laser spot analyses.This material is available from: http://onlinelibrary.wiley.com/doi/10.1111/ggr.12518/abstract(This link will take you to the article abstract).

Figure 1 .
Figure 1.Simplified ion path of Rb/Sr through the Neoma MC-ICP-MS/MS.The ions extracted from the plasma are dispersed on a plane of mass-to-charge (m/z) around m/z 87 by a Wien filter, and a selectable slit is used to remove interference-generating ions such as 40 Ar + or 105 Pd + .The dispersed ions are then recollimated by a second Wien filter prior to entrance to a collision/reaction cell containing SF 6 .Sr + ions react to form SrF + .The Neoma MC-ICP-MS/ MS is then used to separate and direct the Rb + and SrF + ions to the Faraday collector array.Note that for simplicity only a select few Faraday cups are shown in this schematic diagram.The full cup configuration is given inTable 1.

Figure 2 .
Figure 2. Rb-Sr isochrons of biotite from sample Ra-D72 determined by LA-MC-ICP-MS/MS using the traditional spot analysis approach.Each data point represents one laser spot analysis.In (a) signals were trimmed to the most homogenous parts of each selection; in (b) each data point includes all integrations for each spot.Range bars represent 2s.

(Figure 6 .
Figure 6.(a-b) Example laser spots with multiple isochrons, or 'multichrons', which span the range of dates in sample Ra-D72.

Figure 7 .
Figure 7. Histogram and KDEs for individual date integrations for all single spot isochrons identified in sample Ra-D72.Kernel density estimates are shown for three bandwidths determined using the ksdensity function in MATLAB: the solid green KDE represents the calculated optimal smoothing bandwidth for the KDE calculation (10.53 My).Bandwidths at half (5.27 My, dashed purple line) and double (21.06 My, dotted blue line) the optimal calculated bandwidth are shown for reference.All three KDEs are on the same vertical scale.

Figure
Figure 8. Composite Rb-Sr isochron for all integrations

Figure 9 .
Figure 8. Composite Rb-Sr isochron for all integrations (n = 780) from twenty-two laser spot analyses of biotite in sample Ra-D72.An initial 87 Sr/ 86 Sr ratio of 0.7197 was calculated using the maximum likelihood model in IsotplotR.Isochrons at 320 and 230 Ma are shown for reference. 5b

Figure 10 .
Figure 10.Trace element maps of a characteristic biotite grain in sample Ra-D72 determined by LA-ICP-MS using a 3 lm 3 3 lm spot (see details in Methods).(a) Rb, (b) Sr, (c) Ba, (d) Rb/Sr.Map scales in a-c are lg g -1 .

Figure S1 .
Figure S1.Simplified geological map of the Mooselookmeguntic contact aureole and surrounding plutons.

Figure
Figure S4.(a) Single spot isochron for one laser analysis of the LaPosta biotite.(b) Time-resolved signal 87 Rb/ 86 Sr and two point dates for the spot shown in (a).Two-point dates calculated using the isochron intercept in (a).

Figure S5 .
Figure S5.Rb-Sr isochrons for integrations 18-35 from laser spot 10.(a) Isochron and fit based solely on measured analyses.(b) Isochron and fit with initial 87 Sr/ 86 Sr anchored to intercept determined in (a).

Figure S6 .
Figure S6.Histogram and KDEs for individual date integrations for all single spot isochrons identified in sample Ra-D72.

Figure S7 .
Figure S7.Reflected light photomicrographs of sample Ra-D72.(a) Overview and (b, c, d) close ups showing spot locations of laser analyses.

Table 2 .
Single spot Rb-Sr isochron dates The 87 Rb counts are then determined by subtracting any Sr on m/z = 87, using the observed mass-shifted 87 Sr/ 88 Sr ratio and the non-shifted m/z = 88 background subtracted counts.Raw (uncorrected for mass fractionation) ratios for 87 Sr/ 86 Sr, 87 Rb/ 86 Sr, 87 Rb/ 87 Sr, 88 Sr/ 86 Sr and 84 Sr/ 86 Sr are then calculated.
Table 2).For example, integrations 5-17 of spot 10 give an isochron fit of 276 AE 17 Ma; the uncertainty of this calculated date is partly due to the extrapolation from 87 Rb/ 86 Sr values [ 3400 back to the intercept.If we pin the initial 87 Sr/ 86 Sr at the intercept determined by the isochron (0.19), then the calculated date is 276 AE 4 Ma (Figure