Molybdenite Reference Materials for In Situ LA‐ICP‐MS/MS Re‐Os Geochronology

Re‐Os isotope‐dilution geochronology has been widely used to date the timing of molybdenite, pyrite and chalcopyrite formation across a variety of geological settings. However, in situ methods have been impeded by the isobaric interference of 187Re on 187Os. In situ Re‐Os geochronology using LA‐ICP‐MS/MS has been shown to be a useful technique to chemically separate Os from Re, as Os reacts with CH4 to create higher‐mass reaction products, which can then be measured with minimised interference of 187Re. However, application of the method requires matrix‐matched primary reference materials, e.g., age‐homogenous molybdenite amenable to laser ablation. Here, we characterise and present two new molybdenite mineral reference materials for in situ Re‐Os geochronology by LA‐ICP‐MS/MS, verified by ID‐TIMS Re‐Os measurements. We also present case studies from molybdenite samples with varying Re mass fractions and Re‐Os age mapping. The method provides accurate and precise age data, with excellent precision for high Re samples. The benefits of the LA‐ICP‐MS/MS approach include: (1) simple sample preparation, (2) rapid data acquisition, (3) targeting of specific textural domains including growth zones and (4) the ability to simultaneously collect trace elements used to link the timing and conditions of ore‐formation.

Re-Os geochronology is an important tool for understanding the timing of metalliferous ore formation and the depositional ages of organic-rich sediments (e.g., Hirt 1963, Stein et al. 2001, Selby and Creaser 2004).Rhenium is a strongly chalcophile element that can be concentrated in hydrothermally-derived sulfides such as pyrite (FeS 2 ), chalcopyrite (CuFeS 2 ) and, most notably, molybdenite (MoS 2 , Stein et al. 2002, Barton et al. 2019).Dating molybdenite with the Re-Os isotope system is demonstrably robust, as molybdenite generally does not incorporate significant Os during crystallisation, and can remain unaffected by later high-grade metamorphism, deformation and hydrothermal fluid flow events (Stein et al. 2001, Bingen and Stein 2003, Barra et al. 2017).Consequently, a large number of studies have used Re-Os geochronology to constrain the timing of mineralisation in a variety of ore deposits and petroleum systems (e.g., Stein et al. 2001, Bingen and Stein 2003, Barra et al. 2017, Sai et al. 2020).These studies relied on isotope dilution thermal ionisation mass spectrometry (ID-TIMS) based measurements on chemically separated Re and Os aliquots (e.g., Selby and Creaser 2001a, b, Selby et al. 2009, Barra et al. 2005, Hou et al. 2006, Stein 2014, Feely et al. 2020).
Attempts to measure Re-Os isotopes via laser ablationinductively coupled plasma-mass spectrometry (LA-ICP-MS) have proven difficult, principally because of the large ( [ 90%) corrections necessary to deconvolve the isobaric interference of 187 Re on 187 Os.Additional to this problem, authors have suggested that the Re-Os system in molybdenite is susceptible to resetting, where parent and daughter isotopes exhibit contrasting mobilities leading to age heterogeneities (e.g., Ko sler et al. 2003, Stein et al. 2003, Selby and Creaser 2004, Zimmerman et al. 2022).This apparent isotopic decoupling behaviour can result in large age variations within crystals, perhaps detectable by in situ LA-ICP-MS.The isotope dilution method circumvents this behaviour because large grains are crushed and homogenised with the assumption that at some scale, the analysed sample is isotopically closed (Selby and Creaser 2004).However, a disadvantage of grain-scale homogenisation is the destruction of any isotopic record related to growth zoning, recrystallisation and sub-grain development during later events.
LA-ICP-MS/MS has proved a fruitful avenue to chemically distinguish isobaric interferences between Re and Os with the use of methane as a reaction gas (H akansson 2019(H akansson , Hogmalm et al. 2019).In the LA-ICP-MS/MS set-up, a reaction cell (also referred to as the "collision/reaction cell") is housed between two quadrupole mass analysers.The first quadrupole acts as a mass filter, allowing only one preselected isotope (or m/z ratio) to enter the reaction cell at a time, which then reacts with the gas to form product ions with higher mass.The second quadrupole is set to the mass of the product ion, enabling chemical separation of isobaric interferences.This technique is used to measure isotope ratios from beta decay isotope systems, where the daughter isotope reacts efficiently with the reaction gas, but the parent does not, for example 87 Rb-87 Sr (Zack and Hogmalm 2016, Hogmalm et al. 2017, Redaa et al. 2021) and 176 Lu-176 Hf (Simpson et al. 2021, 2022, Tamblyn et al. 2022).The set-up is also useful for the removal of interferences in alpha decay systems, e.g., the removal of common lead in the U-Pb system (Gilbert and Glorie 2020).It has been demonstrated that for the Re-Os system, 187 Os reacts with CH 4 to produce OsCH 2 + with a mass of 201 (i.e., +14 mass-shift), while 187 Re is less reactive (e.g., Hogmalm et al. 2019).Hogmalm et al. (2019) used a molybdenite mineral nano-powder as the primary reference material, which has the potential limitation that its ablation characteristics differ from natural molybdenite mineral grains (e.g., Redaa et al. 2021), and precludes evaluation of within-grain Re-Os isotopic disequilibrium in reference materials (Hogmalm et al. 2019).In situ Re-Os isotopic measurement via LA-ICP-MS/MS may benefit from the development of matrix-matched reference material crystals with sufficient Re and Os mass fractions, well-characterised and homogenous Re-Os isotope ratios, and sufficiently large crystals for easy mounting and ablation.
We have characterised two natural molybdenite mineral samples from Queensland, Australia and Quebec, Canada, by both LA-ICP-MS/MS and ID-TIMS, and demonstrate they can be reliably used as primary reference materials (RMs) for in situ Re-Os geochronology.Using these primary RMs, we explore the advantages and limitations of the in situ Re-Os method by presenting Re-Os age data, including Re-Os age maps, for molybdenite samples with variable Re mass fractions (ranging between 0.1 and 5100 lg g -1 ).We demonstrate the uncertainties on the in situ Re-Os weighted mean ages from Re-rich molybdenite can approach those of ID-TIMS analysis.The main advantages of the in situ method are: (1) the speed of data acquisition, (2) the spatial and textural resolution of Re-Os sampling within mineral grains, and (3) the ability to collect trace element data simultaneously that provides petrogenetic information that can be directly tied to the age data.

Samples Merlin deposit, Queensland, Australia
The Merlin deposit is located within the Cloncurry district of the Mt.Isa Inlier, Queensland, Australia.It is the highest-grade Mo-Re deposit known in the world and is hosted by a metasedimentary rock package comprising phyllite, carbonaceous slate and calc-silicate rocks (Brown et al. 2010, Babo et al. 2017).Mo-Re mineralisation occurs in two styles, (1) infill of matrix-supported hydrothermal breccias and (2) stylolitic veins and disseminations (Babo et al. 2017) (Babo et al. 2017).
Two molybdenite samples with pre-existing Re-Os ID-TIMS ages (Babo et al. 2017) were chosen for this study: sample MDQ0252 taken at 466.25 m depth from drill core MDQ0252 and sample MDQ0221 taken at and 473.8 m depth from drill core MDQ0221 (Table 1).Sample MDQ0252 is a partially stylolitic molybdenite vein, 5 mm in diameter, and was used as a primary reference material in this study.Sample MDQ0221 consists of smaller (0.5 mm in diameter) molybdenite veins.An additional molybdenite sample (MDQ0120) was chosen for age and trace element mapping.

Moly Hill, Quebec, Canada
The Moly Hill deposit is located in La Motte within the Abitibi belt, Quebec, Canada.Molybdenite is hosted in quartz veins/pegmatites within the Preissac pluton, a late-Archaean muscovite monzogranite intrusion (Sabina 2003).Crystallisation of the Preissac pluton has been dated to 2681-2660 Ma by U-Pb and Sm-Nd geochronology of monazite and titanite (Ducharme et al. 1997).Molybdenite from the pluton was dated by ID-TIMS Re-Os geochronology at 2766-2526 Ma (range of three ages) and 2750 AE 27 Ma, recalculated from Suzuki et al. (1993) and Birck et al. (1997) by Ko sler et al. (2003) with the updated Re decay constant of 1.6668 AE 0.0034 9 10 -11 (Table 1).One molybdenite sample (named Q-MolyHill) hosted with coarse-grained quartz was obtained from a commercial mineral specimen supplier.The sample consists of a single euhedral freestanding crystal of 1.25 cm in diameter.

Jinka Moly, Northern Territory, Australia
The Molyhil skarn deposit is located in the Jinka area of the Aileron Province of central Australia, ~225 km to the northeast of Alice Springs (Barraclough 1979).The Mo-W mineralisation occurs at the contact between altered metacarbonates (Deep Bore Metamorphics) and granite (Marshall Granite), where the intruding granite provided the hydrothermal fluids for skarn formation.U-Pb zircon age data indicates the granite crystallised between ~1732-1720 Ma (Kositcin et al. 2018).Previous molybdenite ID-TIMS Re-Os analysis gave an age of 1720 AE 8 Ma (Table 1; Huston et al. 2021) and Ar-Ar step heating of hornblende from the skarn gave an age of 1701.6 AE 4.8 Ma (Reno and Fraser 2021).
A sample of molybdenite was obtained from the Molyhil deposit.In order to clearly distinguish this sample from Moly Hill in Quebec, the sample is named NT-Jinka.Molybdenite from Jinka forms grains up to 2.5 cm in length and together with chalcopyrite overprints a magnetite-quartz-apatiteamphibole skarn assemblage.

Moonta-Wallaroo, South Australia, Australia
The Moonta-Wallaroo district is located in the southern Olympic Cu-Au province in the eastern Gawler Craton (South Australia) and is characterised by iron oxide Cu-Au mineralisation (Ruano et al. 2002, Bockmann et al. 2022).Molybdenite occurs as a minor phase in Cu-Au-bearing veins within metasediments and felsic magmatic rocks of the Wallaroo Group (Skirrow et al. 2007) A sample (M15581) of pure coarse-grained ( [ 3 cm) molybdenite from the Yelta Mine locality (Table 1) was obtained from the Tate Memorial Museum at the University of Adelaide.Yelta is located within the same geophysical domain as the younger of the two ID-TIMS molybdenite ages.

Sample preparation
All samples were mounted as rock blocks or mineral grains in 25-mm epoxy disks, and polished using 3-lm diamond paste.Care was taken to polish each molybdenite sample on a separate piece of sandpaper or cloth, to avoid smearing and cross-contamination of the soft molybdenite grains.Grains were optically examined under reflected light for inclusions.

In situ LA-ICP-MS/MS spot analyses
The Re-Os measurements were undertaken using a RESOlution-LR 193 nm excimer laser ablation system (Applied Spectra), with a S155 sample chamber (Laurin Technic), coupled to an Agilent 8900 ICP-MS/MS, housed at Adelaide Microscopy, the University of Adelaide.A squid mixing device (Laurin Technic) was used to smooth the signals from the laser ablation system.Measurements were conducted across three sessions, with an initial session to determine optimal gas flow rates for efficient reaction of Os.CH 4 was used as the reaction gas, after Hogmalm et al. (2019, 99.999% purity), in the first session at a flow rate of 0.223 ml min -1 , and in the second two sessions at a flow rate of 0.388 ml min -1 (20 and 35% on Mass Flow Controller 4, respectively).In the first session, He was added at a flow rate of 7 ml min -1 to test the effect on sensitivity and reaction rates.A reaction product ion scan was run to confirm mass 201 ( 187 Os 12 C 2 H 2 + ) as the optimal reacted mass of 187 Os (Hogmalm et al. 2019), referred to herein as 187+14 Os.Samples that are characterised for Re mass fractions (synthetic glass NIST SRM 610 and sulfide RM STDGL3, Jochum et al. 2011, Danyushevsky et al. 2011), and Os (synthetic sulfide reference material NiS3, Gilbert et al. 2013), were analysed to optimise reaction rates in the reaction cell for 187+14 Os, while supressing reaction rates for interference 187+14 Re.N 2 was added (at 3.5 ml min -1 ) to the carrier gas after the sample chamber to increase sensitivity (Hu et al. 2008).The ICP-MS/MS was first tuned in the absence of the CH 4 reaction gas (no-gas mode) to minimise oxide interferences and to optimise N 2 and Ar gas flow rates for a robust plasma, and then tuned with CH 4 reaction gas to maximise sensitivity for the Os reaction products using the CH 4 reaction gas mixture.
Molybdenite crystals were ablated with a fluence of 3.0 J cm -2 at a repetition rate of 5 Hz, and with laser spot diameters of 30, 43, 67, 100 and 120 lm.The following isotopes for geochronological calculations were measured, with mass shifts in brackets: 185 Re, (185+14) Re, 187 Os+Re, (187+14) Os (which contains some reacted 187 Re), 189 Os and (189+14) Os (Table 2).A number of additional masses were measured to monitor inclusions and trace elements in the molybdenite (Table 2).Analyses included 30 s of background followed by 30 to 40 s of ablation.Dwell times for the measured masses are in Table 2.
All LA-ICP-MS/MS spot analysis data processing was completed using the software LADR (Norris and Danyushevsky 2018), to correct for gas background, monitor for down-hole fractionation and calibrate for instrument mass bias and drift.We note that 187 Re/ 187 Os down-hole fractionation in molybdenite was not detected in this study, regardless of the spot size used (e.g., Figure 1).Due to the potential for unreacted 187 Os and 187 Re at mass 187, 185 Re was measured as a proxy and 187 Re calculated assuming natural abundance.In our observations, ~1.3-2.5% of the 187 Re reacts with CH 4 and, therefore, an interference correction was carried out by calculating the amount of 187+14 Re present in each analysis from the measured 185+14 Re counts by assuming natural abundance and subtracting this from the raw 187+14 Os counts.The interference correction is carried out prior to calculating the average ratio for each analysis and normalisation to the primary reference material.The session run with 0.223 ml min -1 CH 4 and 7 ml min -1 He has a Re reaction rates of 1.3%, and the sessions run with 0.388 ml min -1 CH 4 have Re reaction rates between 2 and 2.5% (online supporting information Figure S1).Therefore, He may have the effect of reducing the Re interference, with the caveat that it also reduces sensitivity on the reaction products.The mass fractionation for interference corrections was corrected by measuring 187 Re/ 185 Re in NIST610, which contains negligible Os (Jochum et al. 2011, Hogmalm et al. 2019)

LA-ICP-MS/MS mapping methods
LA-ICP-MS/MS maps were collected to verify homogeneity of the 187 Re/ 187 Os ratio in the primary reference material and to explore the potential applications of isotopic ratio mapping.Rhenium, Os and trace elements were acquired simultaneously using the dwell times listed in Table 2. Laser and mass spectrometer settings were the same as those for spot analyses, except that the squid mixing device was removed to improve response time and spatial resolution, and the raster spot size and scan speed were 20 lm and 10 lm s -1 respectively.The raw (187+14) Os counts per second (cps) data were corrected for the 187 Re interference, using the measured (185+14) Re as described above.These data were then imported to Iolite3 to correct for instrument drift and mass bias (Paton et al. 2011), to produce corrected 187 Re/ (187+14) Os ratios.Subsequently Re-Os ages were calculated using the 187 Re decay constant of 1.6668 AE 0.0034 9 10 -11 .Age maps and trace element maps were then generated using Laser Map Explorer (inhouse software, Hasterok et al. in prep.).The maps were filtered (Table 3) to identify the molybdenite grains and remove contamination from epoxy at the edges and holes in the grain.For both maps, noise reduction was applied using and containing ~450 lg g -1 Re. (187+ 14) Os corrected refers to the counts of (187+ 14) Os corrected for the 187 Re which also reacts with CH 4 .The molybdenite contains variable 185 Re but has a homogenous 187 Re/ 187 Os ratio with no detectable down-hole fractionation.
Common Os (measured as 189 Os) was not detected.
three filters: (1) 95 Mo cps \ 0.5 9 10 8 (to ensure only moly data is shown), ( 2) 27 Al cps [ 1 9 10 7 (to remove significant inclusions), and (3) Re-Os age [ 3500 Ma (to remove extreme noise; [ 29 the age of the mineral).The age filter removed \ 0.1% and \ 0.4% for MDQ0252 and MDQ0120, respectively.The Mo filter mostly removes empty zones around the margins of the map, the Al filter removed epoxy contamination around the holes and the date filter removed date estimates more than three standard deviations from the mean.

ID-TIMS analyses
The two primary reference material candidates (MDQ0252 and Q-MolyHill) were further characterised for Re-Os isotopic analysis using ID-TIMS at the John de Laeter Centre (JdLC), Curtin University, Western Australia.
The isotope dilution technique was used in conjunction with the Carius tube method, to digest the sample material and equilibrate the sample-spike (Shirey and Walker 1995).To measure the abundance of 187 Re and 187 Os in molybdenite by isotope dilution technique, we used a mixed 185 Re-188 Os-190 Os double spike, following the approach described in Markey et al. (2003Markey et al. ( , 2007)).For this study, the JdLC spike was designed with a 188 Os/ 190 Os ratio near unity to provide for a more precise determination of the fractionation correction, following the University of Alberta Radiogenic Isotope Facility (RIF) methodology.The 187 Os-190 Os abundance in the double spike was calibrated against the ammonium hexachloro-osmate AB2, which was prepared and calibrated at RIF (Selby and Creaser 2001).For the initial calibration of our spike, we simultaneously determined the isotopic composition and the concentration of the double spike, by measuring separately double spike and the spikestandard mixture.The 185 Re abundance in mixed spike solution was calibrated against a gravimetric Re standard solution made from 99.999% Re metal and shows a reproducibility better than 0.4% 2s (n = 5).
Multiple aliquots of each molybdenite sample were weighed (~10 to 60 mg) and transferred directly into cleaned Carius tubes.A carefully weighed amount of aforementioned mixed double spike was then added to each Carius tube.Inverse aqua regia (1 ml concentrated, PTFE-distilled hydrochloric acid and 3 ml concentrated, PTFEdistilled purged nitric acid) was subsequently added to the Carius tubes and frozen to prevent loss of volatile Os.Once each tube had been sealed, the samples were digested overnight in an oven at 220 °C.After complete digestion, Os was separated from the inverse aqua regia sample solution by solvent extraction (Cohen and Waters 1996), with final purification of the Os achieved via micro-distillation (Brick et al. 1997).Rhenium separation was achieved using anion exchange chromatography.
The purified Os was loaded onto platinum filaments with a NaOH-Ba(OH) 2 activator solution and the Os isotopic measurements were obtained using negative TIMS on a Thermo Scientific Triton TM .Osmium isotope determinations were conducted either using a secondary electron multiplier detector in peak jumping mode, or on Faraday collectors in static mode.Long-term 'instrument reproducibility' on the Triton TM was monitored through the measurement of the AB2 Os reference material (University of Alberta) during each measurement session, yielding a 187 Os/ 188 Os isotope ratio of 0.10682 AE 0.00018 (n = 5, 2s), consistent with the 0.10684 AE 0.00004 187 Os/ 188 Os isotope ratio reported by Selby and Creaser (2003).The Re isotopic compositions and mass fractions were determined in solution via ICP-MS on a Thermo Scientific ELEMENT XR TM .The mean of multiple measurements of a natural Re standard solution returned an 185 Re/ 187 Re isotopic ratio of 0.59285 AE 0.00042 (n = 106, 2s), with the uncertainty being just outside the natural 185 Re/ 187 Re isotopic ratio of 0.5974 (Gramlich et al. 1973).Subsequently, a minor fractionation correction was applied.
Error propagation on individual measurements includes uncertainties associated with the calibration of the spike, mass spectrometer measurements, and blank corrections.Due to the high Re and 187 Os mass fractions for these molybdenite samples, analytical blanks are essentially insignificant for age calculations.Procedural blank values for Os were 0.4 AE 0.1 pg (n = 2) and 1.7 AE 0.4 pg for Re.The 187 Os/ 188 Os ratios for the blank were 0.38 AE 0.08.
The 'Henderson molybdenite' reference material (Henderson RM) was used to monitor inter-laboratory agreement and clean laboratory chemical purification procedures.Three aliquots ranging from weights of ~30-60 mg were processed using the identical procedure described above

Results
In situ LA-ICP-MS/MS spot analyses All obtained data have been pooled and presented from across three measurement sessions (online supporting information Table S1).The presented uncertainties are 2 standard error of the means and calculated ages include the uncertainty on the 187 Re decay constant.For dates where the MSWD is [ 1, i.e., greater than expected for a single population of N, the presented uncertainties include overdispersion (Vermeesch 2018).The primary reference material for the Re-Os isotopic measurements was molybdenite sample MDQ0252 from Merlin.Analyses of MDQ0252 show variable Re and Os mass fractions, but consistent 187 Os/ 185 Re ratios and therefore ages (e.g., Figures 1 and 2a).Rhenium mass fractions range between 65 and 5100 lg g -1 .The weighted mean age of the LA-ICP-MS/MS ages for MDQ0252 (i.e., as the primary reference material) is 1518 AE 4 Ma, (Figure 2a, n = 71, MSWD = 0.87), which excludes one analysis as an outlier.The small uncertainty on the weighted mean age is due to limited dispersion in the single-point data.
A separate molybdenite sample from Merlin (MDQ0221), with Re mass fractions between 60 and 1300 lg g -1 and an ID-TIMS age of 1523 AE 6 Ma (Babo et al. 2017), was run as a secondary reference material.Across the three sessions MDQ0221 returned a weighted mean Re-Os age of 1516 AE 6 Ma (Figure 2b, n = 70, MSWD = 0.77), with two analyses excluded as outliers.
Molybdenite from Moly Hill in Quebec (Q-MolyHill) was measured across two measurement sessions and returned an in situ Re-Os age of 2620 AE 10 Ma (n = 62, MSWD = 0.54), with Re mass fractions of 5-60 lg g -1 .Two analyses were excluded from the age calculation as outliers (Figure 2c).NT-Jinka molybdenite was measured in one session, and the calculated ages show more significant dispersion compared with other samples (Figure 2d).The weighted mean of the ages gives 1710 AE 23 Ma (n = 45, MSWD = 2.88), with five analyses excluded as outliers (as statistically calculated by the program IsoplotR, Vermeesch et al. 2018).The dispersed ages tail out to ca. 1300 and ca.2100 Ma (Figure 2d).Rhenium mass fractions in the molybdenite are low, and range between 0.1-15 lg g -1 .
Molybdenite from the Moonta-Wallaroo district (M15581) was analysed in one session and gives a weighted mean age of 1575 AE 12 Ma (Figure 2e, n = 20, MSWD = 0.48).The Re mass fraction ranges between 50 and 320 lg g -1 .

LA-ICP-MS/MS mapping results
LA-ICP-MS/MS mapping of two samples of molybdenite from Merlin is presented in Figure 3. Sample MDQ0252 is largely homogenous in Re-Os age, despite the strong Re zoning in the molybdenite.Minor apparent age differences largely occur at holes or cracks in the crystal, and/or areas with very low Re.Hence, when such areas are avoided, the crystal can be regarded as homogenous for the purpose of laser ablation dating.A vein of molybdenite from another sample in the Merlin deposit (MDQ0120) shows minor age zoning that correlates with enrichment in trace elements, but low Re mass fractions.

ID-TIMS
Results of ID-TIMS analyses of samples MDQ0252 and Q-Molyhill are presented in Table 4 and Figure 4. Six aliquots of MDQ0252 were analysed, and show dispersed 187 Os/ 187 Re ratios and ages, ranging from 1501 AE 6 Ma to 1527 AE 11 Ma (Figure 4a).The dispersion in age correlates to Re mass fractions, with higher Re aliquots (ranging from 1043 to 1617 lg g -1 ) producing progressively younger ages (Figure 4c).When a weighted mean is calculated excluding one very Re rich ( [ 1300 lg g -1 ) outlier (as determined by the program IsoplotR, Vermeesch et al. 2018), the 187 Os/ 187 Re ratio is 0.025649 AE 0.000041 or AE 0.000105 if 'overdispersion' is considered (MSWD = 3.3, Vermeesch et al. 2018), corresponding to a weighted mean of the Re-Os ages of 1520 AE 4 Ma, including the uncertainty on the Re decay constant (Figure 4, n = 5, MSWD = 1.3).This 187 Re/ 187 Os ratio was used as the primary RM for LA-ICP-MS/MS analyses.
Five aliquots of Q-MolyHill yield slightly dispersed 187 Os/ 187 Re ratios that are also inversely correlated with Re mass fraction.However, the variation in Re mass fraction is relatively low (between 44 to 51 lg g -1 ) and a statistically acceptable weighed mean 187 Os/ 187 Re ratio of 0.044699 AE 0.00006 or 0.00017 if 'overdispersion' is considered (Figure 4d, MSWD = 3.9) can be calculated based on all five aliquots.The weighted mean of the resulting ages is 2623 AE 6 Ma, including the uncertainty on the Re decay constant (n = 5, MSWD = 1.3, Figure 4d).

Discussion
Assessment of MDQ0252 and Q-MolyHill as Re-Os reference materials Molybdenite samples MDQ0252 from the Merlin deposit and Q-MolyHill from the Preissac pluton in Quebec were selected as candidates for reference material development.
Using the Re-Os ID-TIMS ratio for MDQ0252 as the primary RM, LA-ICP-MS/MS ages obtained for Q-MolyHill are identical with the TIMS age within measurement uncertainty (Figure 5, Table 1).Hogmalm et al. (2019) use a mean calibration factor to calculate ages for natural molybdenite, which compares the analysed Re-Os ratios with the expected age from pressed pellets ([ 187+14 Os/ 185+14 Re]/age).Our approach involved a direct calibration of the Re-Os ratios to the MDQ0252 molybdenite crystals during data reduction, and therefore did not require an external correction factor to be applied during the calculation of LA-ICP-MS/MS ages.
The obtained ID-TIMS age for sample MDQ0252 is identical to the published age presented by Babo et al. (2017), suggesting it could be used as a reliable reference value.However, the results of our ID-TIMS analysis 4 0 1 of MDQ0252 show that 187 Os/ 187 Re ratios from single aliquots are not homogenous, and hence form a spread in calculated ages between 1501 AE 8 to 1527 AE 13 Ma.
Most data, however, produce ca.1520 Ma ages (Figure 4c).The younger ages are also directly correlated with higher Re mass fractions (Figure 4c), suggesting that  there may have been a recrystallisation or precipitation of higher-Re molybdenite at ca. 1500 Ma.This is consistent with molybdenite geochronology and textural interpretations of the Merlin deposit, which record the formation of minor molybdenite veins and disseminations at ca. 1500 Ma (Babo et al. 2017).While this event is interpreted to only be minor in the deposit, it appears to have had a detectable effect on the molybdenite veins from sample MDQ0252.Components of the younger ca.1500 Ma ages may also be present in the LA-ICP-MS/MS data, however these are difficult to identify as the uncertainty on each individual data point is too large for such assessment (Figure 2a).We suggest, therefore, that MDQ0252 can be used as a reliable reference material for in situ Re-Os dating when Re mass fractions are monitored and high Re outliers are excluded.Following these recommendations, we have used MDQ0252 as the primary reference material for all LA-ICP-MS/MS analyses in this study.
The molybdenite from Moly Hill in Quebec may also be considered as a potential primary reference material for future Re-Os analysis.In situ LA-ICP-MS/MS analyses and ID-TIMS analyses both show that the molybdenite is homogenous in 187 Os/ 187 Re ratios and has reasonable but not excessively high Re mass fractions (between 5 and 60 lg g -1 ).The latter is important as very high Re mass fractions using a large spot size can trigger the detector in the ICP-MS/MS to shift from pulse to analogue mode, which means that the analysed reference material can no longer correct for unknown analyses in pulse mode, due to difficulties in determining accurate pulse-to-analogue correction factors in the ICP-MS/MS.When sample MDQ0252 is analysed using large spot sizes ( [ 100 lm) the Re can be measured in analogue mode, as it contains very high Re mass fractions (up to ~5100 lg g -1 ).Importantly, the reference materials developed in this study are natural molybdenite crystals and not nano-powders.Hence, there were no matrix effects during ablation between the reference material and the unknown molybdenite samples.

Re-Os case studies
In this section, we demonstrate the reliability of the presented approach for LA-ICP-MS/MS Re-Os geochronology using case studies of several molybdenite samples with known ages (Figure 5).A second molybdenite sample from the Merlin deposit (MDQ0221) returned an in situ age of 1516 AE 6 Ma (n = 70, MSWD = 0.77), within uncertainty to those from a previous ID-TIMS study (1523 AE 6 Ma, Babo et al. 2017; Figure 5c).Molybdenite from the Jinka area in the Northern Territory also returned an in situ age of 1710 AE 20 Ma (n = 45, MSWD = 2.88), which is with within uncertainty of Re-Os ID-TIMS (1720 AE 9 Ma) and hornblende Ar-Ar ages (1701.6AE 4.8 Ma) from the same deposit (Cross 2009, Reno and Fraser 2021; Figure 5), despite low Re mass fractions in the analysed grains (0.1 to 15 lg g -1 ).The scatter in the individual ages, as seen in the weighted mean of the ages with an MSWD of 2.8, could be due to low Re and therefore low 187 Os analyses, which have low signal to noise ratios and are susceptible to scatter caused by the 187 Re interference correction (Figure 2d, Figure S2).The scatter could also be due to isotopic disequilibrium within the analysed molybdenite crystals.Molybdenite from Yelta mine in the Moonta-Wallaroo district returned similar ages within uncertainty (1575 AE 12 Ma, n = 20, MSWD = 0.48) to Re-Os ID-TIMS ages from other molybdenite deposits within the area (1599 AE 6 Ma and 1574 AE 6 Ma, Skirrow et al. 2007; Figure 5e).
No strong evidence for significant age variation (i.e., differing extents of within-grain Re and Os mobility) between Re and Os was observed outside uncertainty in this study in molybdenite.Many studies suggesting that decoupling may create scatter in the data from microbeam methods, and that whole homogenised molybdenite grains or large fractions should only be considered for isotopic analysis (Ko sler et al. 2003, Stein et al. 2003, Selby and Creaser 2004, Zimmerman et al. 2022).However this approach is only applicable if Re or Os are not lost from the molybdenite crystal.Considering the small sampling size in each analysis and the precise spatial resolution, LA-ICP-MS/MS is the most likely method to detect microscale decoupling of Re from its daughter isotope.The NT-Jinka sample in particular may show evidence of some isotopic disequilibrium or resetting, as there are several outliers, and the weighted mean population has a high MSWD of 2.88.This may warrant further in situ Re-Os analyses and textural and geochemical investigation into the molybdenite from Jinka.We also note that Re and Os resetting may occur within the measurement precision of the method and therefore may not be detected.LA-ICP-MS/MS may provide a fruitful method for investigating case studies of Re and Os decoupling in the future to better understand the phenomenon, as has been done in many other geochronometers by microbeam methods.
Age mapping of two molybdenite samples from the Merlin Deposit shows the potential use of in situ LA-ICP-MS/MS mapping with multi-mineral calibration for correlating Re-Os ages with trace elements, to investigate age zoning or unravel the petrogenetic history of molybdenite grains (Figure 3; e.g., Markmann et al. 2024).Molybdenite from MDQ0252 shows similar heterogeneity in Re mass fraction as calculated from single spot analyses; with grains containing up to 5100 lg g -1 Re (Figure 3a).The age map contains residual measurement artefacts, generated from the sequential measurement of isotopes across grain boundaries and regions of contrasting Re mass fractions, but generally shows that the molybdenite is homogenously ca.1500 Ma despite the variation in Re zoning (Figure 3b).Re-Os isotopic mapping of a molybdenite vein from another sample from the Merlin deposit (MDQ0120) shows potential variable age zoning (Figure 3b).Grains or domains of the molybdenite with high Re mass fractions (up to 1600 lg g -1 ) are correlated with low Cu, Ni and Ag and generally are inversely correlated with the Re-Os ages (Figure S1).Based on these two examples, simultaneous Re-Os age and trace element mapping using the in situ method seems promising and may be a useful future application, especially in mineral systems with complex multi-stage histories.

Strengths and limitations of the Re-Os LA-ICP-MS/MS method
The relationship between Re mass fraction, corresponding radiogenic 187 Os mass fraction, age, and single spot or weighted mean age uncertainty is presented in Figure 6.The lowest uncertainty obtained on individual analyses in this study is ~3% (Figure 6b).However, all calculated weighted mean ages for molybdenite analysed by the LA-ICP-MS/MS method return age uncertainties under 1.4% (n = 20-70), even for low Re samples (\ 10 lg g -1 , Table 1, Figure 5).The precise in situ Re-Os ages can be explained by the agehomogenous nature of the molybdenite grains analysed, forming single age populations (Figure 2), and the absence of common Os in any analyses.The limits on age or Re mass fractions to produce useful dates using the applied method are difficult to constrain, as the detection limit of 187+14 Os is difficult to estimate due to the lack of a primary reference material with a homogenous Os mass fraction, and will depend on the spot size utilised during ablation.The median 187+14 Re detection limit across the three measurement sessions is ~0.005 lg g -1 , considering the lower isotopic abundance and higher ionisation potential of 187+14 Os, the detection limit of 187+14 Os is probably much higher than this value.However, the lowest 187+14 Os detectable by this technique may still have large uncertainties due to low signal to noise ratios, and therefore analyses may still be imprecise.For 187 Os mass fractions between ~0.5 and 0.1 lg g -1 , uncertainties on individual analyses increase to over 5%, although weighted mean ages calculated from this sample are still only 1.4% (Figure 6).
The molybdenite dataset of Barton et al. (2020) contains average Re values and ages for 3089 molybdenite samples from over 700 ore-bearing hydrothermal systems.Calculation of 187 Os from this database allows us to estimate the proportion of molybdenite that may return meaningful ages using the in situ method.If a very conservative 'minimum 187 Os' mass fraction required to calculate a precise weighted mean age (\ 1.4%) of ~0.01 lg g -1 is considered, then 81% of molybdenite in the database could produce precise Re-Os ages using the in situ method, provided the data are not over dispersed (Figure 6).However, we note that the database may be biased to higher-Re samples, due to sampling bias and by the detection limit for Re from the analytical technique utilised.Nevertheless, low Re (\ 10 lg g -1 ) molybdenite of Precambrian age may still produce precise and accurate in situ Re-Os weighted mean ages.A limitation of this method is recognised for young ages (\ 150 Ma), where low Re (\ ~20 ppm) samples may not be able to provide precise ages via this technique.For a full discussion of Re contents in molybdenite in different style of deposits, the reader is referred to Barton et al. (2020).
Another potential limitation of the in situ method is the inability to resolve temporally similar age populations within samples.The significant uncertainties on individual analyses may not allow separate age populations to be discerned without textural or compositional evidence.For example, ID-TIMS analysis suggests that Merlin sample MDQ0252 records two molybdenite crystallisation events, at ca. 1520 and 1500 Ma (Figure 4).We did not see significant scatter in our in situ measurements of Merlin sample MDQ0252, which would suggest that we analysed these two age populations.If molybdenite age populations are within ~3% of each other in other samples, this detail may not be statistically resolvable using the LA-ICP-MS/MS method.However, with other lines of evidence, such as those obtained from textural or compositional (i.e., trace element) analysis, it may be possible to discern multiple ages in in situ molybdenite analyses.
Strengths of this technique include the ability to target specific molybdenite grains and textures in situ, as well as to collect trace element data simultaneously to determine the composition and to assist in understanding the petrogenesis of the analysed sulfides.The in situ approach also provides an opportunity to remove measurement intervals of inclusions in time-resolved LA-ICP-MS/MS data, ensuring that only the target mineral is incorporated into Re-Os isotopic calculations.Small grains (30 lm) can be analysed in this method, allowing for analysis of fine grained or dispersed molybdenite, which may not be feasible to analyse using traditional methods.Additionally, simultaneous age and trace element mapping may provide significant advances in our understanding of molybdenite growth and/or Re and Os isotopic disequilibrium (e.g., Figure 3, Figure S1).Future avenues of method development may include the investigation of Re and Os decoupling within molybdenite grains and Re or Os nanoparticle formation (Barra et al. 2017), trace element and age mapping of molybdenite, and the dating of Re-poor minerals such as pyrite, chalcopyrite and cobaltite (e.g., Saintilan et al. 2017, Li et al. 2022).
Among the important advantages of the in situ approach are the relatively simple sample preparation requirements and the rapid and cost effective data acquisition.These aspects, together with the high spatial resolution, open opportunities to efficiently complete large geochronology campaigns on ore systems that very commonly have complex multi-stage mineralisation events, such as Precambrian ore deposits where protracted geological histories are common (e.g., Haroldson et al. 2020, Schutesky and de Oliveira 2020, Maas et al. 2022).The methodology offers additional benefits to the minerals industry where rapid acquisition of geochronological data on ore formation is needed to influence investment decisions on exploration and resource development.

Conclusions
In situ LA-ICP-MS/MS measurement of Re-Os provides accurate and precise age data for molybdenite.Two natural molybdenites, MDQ0252 from the Merlin deposit in Queensland and Q-MolyHill from the Preissac pluton in Quebec, have been developed as potential reference materials for in situ analysis.These samples contain homogenous Re-Os isotope ratios and have Re mass fractions amenable to LA-ICP-MS/MS analysis.The spatial resolution afforded by the in situ approach allows targeting of textural domains, growth zones, and small or scare grains.Additional benefits include the simple sample preparation, rapid and cost effective data acquisition and capability for simultaneous collection of trace elements, all of which facilitate improved understanding of the timing and nature of ore-forming or fluid events in sulfide-bearing systems.Other future directions include the ability to trace element and age map sulfides, to provide a possible avenue to investigate age and trace element variability in individual sulfides, as well as the opportunity to investigate Re and Os isotopic disequilibrium in molybdenite.

Figure 1 .
Figure 1.Example LA-ICP-MS/MS ablation signal from molybdenite sample MDQ0252 from the Merlin deposit, prior to processing in LADR (Norris and Danyushevsky 2018), analysed with a 43 lm spot size,

Figure 2 .
Figure 2. Results of LA-ICP-MS/MS of molybdenite.Molybdenite samples are presented as weighted means at they contain no common Os, as such individual spot ages can be calculated.Uncoloured analyses were not included in the weighted mean age calculations, as they were determined as outliers by the program IsoplotR (Vermeesch 2018).All uncertainties are 2s.(a) Sample MDQ0252, (b) Sample MDQ221, (c) Sample Q-MolyHill, (d) Sample NT-Jinka and (e) Sample M15581.

Figure 3 .
Figure 3. Results of molybdenite trace element and Re-Os age mapping.(a) Re map of MDQ0252.(b) Age map of MDQ0252.(c) Re map of MDQ0120.(d) Age map of MDQ0120.

Figure 4 .
Figure 4. Plots of measurement results from ID-TIMS (all uncertainties are 2s).Where a second uncertainty is reported, it incorporates over dispersion in the data, due to a large MSWD, as calculated by IsoplotR (Vermeesch 2018). 187Os/ 187 Re ratios and calculated ages are shown for comparison, as well as results from Babo et al. (2017).The uncoloured analysis in (a) sample MDQ0252, was determined as an outlier by IsoplotR (Vermeesch 2018).(b) contains data for sample Q-MolyHill.(c) and (d) are coloured for Re mass fraction, age uncertainties include the propagated uncertainty on the Re decay constant.

Figure 5 .
Figure 5. Summary LA-ICP-MS/MS, ID-TIMS and literature ages from samples in this study.LA-ICP-MS/MS and ID-TIMS ages from this study are within uncertainty of each other (Table 1).(a) Sample MDQ0252, (b) Sample MDQ221, (c) Sample Q-MolyHill, (d) Sample NT-Jinka and (e) Sample M15581.

Figure 6 .
Figure 6.(a) Relationship between mass fraction, radiogenic 187 Os mass fraction, age, and (b) individual (spot) measurement uncertainty (2s, including propagated uncertainty on the decay constant of 187 Re) and weighted mean age uncertainty.The 187 Os mass fraction of the weighted mean ages is an average.For low Re and therefore low radiogenic 187 Os molybdenite, individual measurement uncertainties become significant ( [ 5%).However, when low Re and therefore low radiogenic 187 Os (e.g., NT-Jinka) data is pooled in weighted means, the uncertainties on calculated ages are still excellent.For comparison, 187 Os mass fractions are calculated from Re content and age from the molybdenite database of Barton et al. (2020).Note that these estimations are relevant for samples with no common Os.
. The matrix infill molybdenite gives a Re-Os ID-TIMS age of 1535 AE 6 Ma (Table 1; Duncan et al. 2013, Babo et al. 2017), however vein and disseminated molybdenite taken from drill core give Re-Os ID-TIMS ages of 1521 AE 6 Ma to 1529 AE 6 Ma (Babo et al. 2017).Minor Mo-Re mineralisation hosted in carbonate veins gives a Re-Os ID-TIMS age of 1503 AE 5 Ma (Babo et al. 2017).Duncan et al. (2013) also report one older Re-Os age (1559 AE 5 Ma) from the deposit, but the petrological context and geological significance of this molybdenite age is not clear . Re-Os ID-TIMS geochronology of molybdenite from geophysically contrasting areas in the Moonta-Wallaroo give ages of 1599 AE 6 Ma and 1574 AE 6 Ma (Table 1; Skirrow et al. 2007).

Table 2 .
Dwell times for measured masses (Babo et al. 2017)on Re mass fraction).Work is on-going with other reaction gases to further suppress Re reaction rates.Common Os was monitored on masses 189 Os and 189+14 Os, although was not detected in significant mass fractions in any measured molybdenite in this study.Ages were calculated in IsoplotR (Vermeesch 2018) from 187 Re/ 187+14 Os using the 187 Re decay constant on 1.6668 AE 0.0034 9 10 -11 , the uncertainty on the decay constant is propagated into age uncertainties(Ko sler  et al. 2003).Prior to this study, there were no natural Re-Os mineral grains available as primary RMs for laser-ablation Re-Os geochronology.We have characterised the molybdenite sample MDQ0252 (see further information below) with a previously measured 187 Os / 187 Re ratio of 0.0256 AE 0.0001024(Babo et al. 2017)to use as a primary reference material.The 187 Os / 187 Re ratio calculated in this study from ID-TIMS and used for all isotopic calibrations was 0.025649 AE 0.000105, refer to the Results section for further details.For trace elements, the reference materials used were the synthetic sulfide reference material NiS3 (Gilbert et al. 2013 bias is calculated as 0.988.As Re mass fractions are high in molybdenite, this interference correction can be significant (up to 75%) for low Re or young samples, currently limiting the applicability of the method to high Re samples ( [ 10 lg g -1 , depending on age), or old samples, which have accumulated sufficient radiogenic Os ( [ ca.(Gilbert et al. 2013) and the synthetic glass reference material NIST SRM 610(Jochum et al. 2011).Molybdenum was used as the internal standard element for molybdenite, at 59.94% m/m.As the reference material used was matrixmatched to the molybdenite samples, no further correction factors were required.

Table 3 .
Filter parameters for age maps and associated statistical estimates

Table 4 .
Measurement results for molybdenite aliquots by ID-TIMS