Thermal history of Carlin-type gold deposits in Yukon (Canada) as revealed by organic matter geothermometry, clumped isotope data, fluid inclusion microthermometry, and apatite fission-track analyses

Sediment-hosted gold deposits in central Yukon have most of the diagnostic characteristics of Carlin-type gold deposits in Nevada. This study combines organic matter geothermometry with fluid inclusion microthermometry, clumped isotope data (Δ47) for late ore-stage hydrothermal calcite, and apatite fission-track analyses to constrain the thermal evolution of Carlin-type gold zones in central Yukon. The Tmax parameter derived from pyrolysis analyses indicates that organic matter is overmature and records regional temperatures of > 150 °C. Calcite and fluorite associated with the waning stage of mineralization at ca. 74 Ma have mean fluid inclusion homogenization temperatures of 123–173 °C, with an average salinity of 4.8 wt.% NaCl equiv. These temperatures overlap values of 91–162 °C determined from calcite clumped isotope measurements and are similar to data from Carlin-type deposits in Nevada. Fluid mixing is suggested by a variation of the isotopic composition of the fluid in equilibrium with calcite, with the higher temperature end-member having higher δ18OH2O values. In thermal models based on Tmax, fluid inclusion microthermometry, clumped isotope measurements, and apatite fission-track data, a higher temperature event at approximately 220 °C is consistent with pyrobitumen reflectance measurements. This event is either related to maximum tectonic burial prior to mineralization or to the flow of acidic hydrothermal fluids early in the main ore-stage. Each analytical technique used in this study is independent of the others and records part of the thermal and chemical evolution of the Yukon Carlin-type gold deposits and their host rocks.


Introduction
Temperature is one of the major controls on the three components of ore genesis: metal source, transport, and deposition (Ridley 2013), and has been extensively used to distinguish mineralization styles and to better understand ore-forming processes (Lindgren 1933). However, knowledge of the thermal framework of deposits during mineralization is often fragmentary and imprecise, relying on observations and analyses that do not yield unequivocal interpretations. This has proven to be the case for Carlin-type deposits in which Au occurs within thin (generally < 5 µm) arsenian rims on pre-ore pyrite or as sub-micrometer particles. Due to the fine-grained nature of Carlin-type Au ore, it is difficult to establish the relationships between gold and specific minerals or organic compounds that could provide temperature constraints (Cline et al. 2005). The temperature of formation for Carlin-type gold deposits in some of the major deposits in Nevada has primarily been established through the study of fluid inclusions hosted by jasperoids (fine-grained hydrothermal quartz). However, the temperature of hydrothermal fluids is poorly constrained for sediment-hosted gold deposits in central Yukon, which have characteristics of Carlin-type deposits (Pinet et al. 2022a). Bridging this knowledge gap should contribute to improve genetic models and document the variability within the broad spectrum of Carlin-type deposits.
Fluid inclusion microthermometry provides the most direct and, in some cases, the most accurate technique for estimating temperature of mineralization (Wilkinson 2001;Fall and Bodnar 2018). The technique is well established but can present challenges such as interpreting the potential significance of a spread in homogenization temperatures (T h ). An alternative is provided by clumped isotope thermometry, which is based on the temperature dependence of the abundance of bonds between two rare heavy isotopes ( 13 C and 18 O) in calcite (Eiler 2013). The direct determination of temperature from calcite clumped isotopes is a significant advantage compared to interpretations based on calcite/water O-isotope equilibrium, because temperature is calculated from a single carbonate phase (Eiler 2007). Methods based on the maturation of organic matter, such as the Tmax parameter, which is derived from pyrolysis analyses, and organic matter reflectance, are frequently used in petroleum geology (Tissot and Welte 1984), but their application to ore deposits is less common mainly due to the relatively low temperature range over which they can be precisely applied (except for Raman spectroscopy; Beyssac et al. 2002). Moreover, an additional limiting factor is that organic matter paleothermometers only characterize the maximum temperature experienced by a sample (Hartkopf-Fröder et al. 2015), which may have been reached before, during, or after mineralization. Interpretation of low-temperature thermochronology results can also provide insights on thermal evolution at temperatures below ~ 230 °C (zircon fission-track, ZFT), ~ 120 °C (apatite fission-track, AFT), and ~ 70 °C (apatite U-Th/He). As ZFT, AFT, and U-Th/ He ages rarely correspond to the age of a thermal event, modeling is required to exploit the full potential of these methods (Ketcham 2005;Malusà and Fitzgerald 2018).
The aim of this study is to investigate the fluid source(s) and formation temperature of late-ore stage hydrothermal minerals (calcite and fluorite) in the Carlin-type gold deposits in Yukon. This is accomplished using clumped isotope (Δ 47 ) and fluid inclusion microthermometric data supported by published U-Pb LA-ICP-MS calcite dating of late orestage hydrothermal calcite veins (Pinet et al. 2022b). Organic matter maturation data and apatite fission-track analyses make it possible to integrate temperature estimates of late ore-stage hydrothermal fluids into broader spatial and temporal frameworks. This study also provides the opportunity to discuss the merits of each method to estimate temperature.

Carlin-type mineralization in Yukon
Sediment-hosted gold deposits in the Rackla belt of central Yukon (Canada) were discovered in 2010 and have been the focus of several recent studies including a review on the regional to prospect scales (Tucker 2015;Steiner et al. 2018;Pinet et al. 2020aPinet et al. , 2022a, detailed descriptions of mineralized zones (Tucker 2015;Beaton 2015;Pinet and Sack 2019;Pinet et al. 2020bPinet et al. , 2022a, whole rock and stable isotope geochemistry (Tucker 2015;Beaton 2015;Steiner and Hickey 2019;Pinet et al. 2022a;Steiner et al. 2022), U-Pb dating of late ore-stage hydrothermal calcite (Pinet et al. 2022b), and a detailed characterization of both organic matter (Pinet et al. 2022c) and gold-bearing pyrite .
Gold deposits in the Rackla belt share most of the diagnostic features of Carlin-type Au mineralization in Nevada (Table 1) and are among the few examples of Carlin-type Au mineralization documented outside the south-western USA (Arehart et al. 2013;Tucker et al. 2018;Pinet et al. 2022a). Main characteristics of Yukon's deposits are summarized in Table 1, and only a brief summary is provided below. These deposits are located within parauchthonous strata deposited on the North American continental margin east of the Tintina Fault (Fig. 1A). The Rackla belt marks the boundary between the Ogilvie Platform and the Selwyn Basin, with the latter representing a region of deep water off-shelf sedimentation that persisted from the Neoproterozoic to the Middle Devonian (Fig. 1B;Gabrielse 1967;Gordey and Anderson 1993). Gold mineralization is mainly hosted by Neoproterozoic limestone (with few mineralized intervals in siltstone) deposited in slope to base of slope settings (Nadaleen and Gametrail formations belonging to the Windermere Supergroup; Moynihan et al. 2019) and in mid-Paleozoic calcareous siltstones. In the Conrad and Sunrise/Osiris/Ibis deposits, hosted by Neoproterozoic rocks, (Fig. 1C), mineralized zones are roughly concordant with bedding and occur in mixed stratigraphic/structural traps (including anticlinal hinge zones) sealed by less permeable units. Mineralized zones are mainly hosted by carbonate floatstone and grainstone intervals interpreted to have formed through debris flow and in highly fractured sedimentary units. Most fractures and veins predate mineralization, but some are syn-to late mineralization (Tucker et al. 2018;Pinet and Sack 2019;Pinet et al. 2020a, b;Steiner et al. 2022). In contrast, the Anubis mineralized zone differs from deposits hosted by Neoproterozoic rocks located to the east (Fig. 1B) by the age and lithology of the host rocks (Paleozoic siltstones) and by greater structural control on mineralization (Pinet et al. 2022a).
The paragenetic sequence includes an early stage characterized by the dissolution of carbonates (decalcification), an intermediate stage associated with gold-bearing disseminated pyrite and fine-grained quartz (silicification), Table 1 Main characteristics of Carlin-type deposits in Nevada (summarized from Hofstra 1994;Cline et al. 2005;Muntean 2018) and Yukon. The grey fill indicates characteristics that are discussed in the current study Nevada Yukon Regional setting Deposits occur within the rifted cratonic margin of North America along long-lived, deep-seated basement structures Timmons et al. (2001); Grauch et al. (2003) Deposits occur at the boundary between the Olgivie platform and the Selwyn Basin along the long-lived deep-seated Dawson fault Pinet et al. (2020a) Age 42-34 Ma; > 300 Ma younger than host-rocks Arehart et al. (2003); Ressel and Henry (2006) ~ 74 Ma; > 270 Ma younger than host-rocks Pinet et al. (2022b);Tucker (2015) Structural style Pre-mineralization shallowly-dipping thrust, syn-to post-mineralization normal faults Rhys et al. (2015) Moderately to steeply dipping faults, tight folds Moynihan (2016); Steiner and Hickey (2019) Relationship with magmatism Clear at both the regional/district and deposit scales Ressel and Henry (2006) Absent at the district scale. Only 2 known syn-mineralization dykes at deposit scale Tucker (2015) Host rocks Thin bedded carbonaceous sedimentary rocks and debris flow deposits (slope setting) ± magmatic rocks Cook (2015) Thin bedded carbonaceous sedimentary rocks and debris flow deposits (slope setting) ± siltstone Tucker (2015); Beaton (2015); Pinet et al. (2022a) Organic matter maturation (host rock) Generally overmature; Au deposition generally followed thermal maturation Hofstra (1994) Overmature; Au deposition followed thermal maturation This study; Pinet et al. (2022c) Ore body geometry Large spectrum of geometries controlled by shallowly-dipping sedimentary units and/or high-angle faults; deposits are commonly located in the lower plate of a preore regional thrust (impermeable barrier) Rhys et al. (2015) Geometry and a late stage characterized by the deposition of blocky calcite, realgar, orpiment, fluorite, and rare stibnite (Tucker 2015;Pinet et al. 2022a). Argillitization is local and generally difficult to document (Tucker 2015;Pinet et al. 2022a). Gold occurs within thin arsenian pyrite rims on pre-ore pyrite or as micron-scale particles in decalcified matrix material . Whole rock chemical analyses indicate that Au is associated with pathfinder elements such as Tl, As, Hg, and Sb, and that base metal and Ag contents are low (Tucker 2015;Beaton 2015;Pinet et al. 2022a). The timing of mineralization is ca. 74 Ma based on the combination of U-Pb dating of zircon from a strongly altered and Au-mineralized dike (Tucker 2015) and of late-ore stage hydrothermal calcite associated with realgar (Pinet et al. 2022b).

Samples for pyrolysis and reflectance analyses
The data initially generated to determine the maximum temperature experienced by the samples consists of 34 programed pyrolysis (Rock-Eval 6™ and HAWK) analyses including 5 duplicates) corresponding to surface samples outside mineral deposits (n = 5) and to representative core samples from the Conrad (n = 5 from 3 boreholes) and Osiris/Sunrise (n = 16 from 2 boreholes) deposits and from the Anubis mineralized zone (n = 8 from 2 boreholes). Two samples among those with the highest total organic carbon (TOC) content (1.36 and 1.31 wt.%) were selected for organic matter characterization and random reflectance measurements. Sample RKA-17-27 from the Osiris/Sunrise deposit is located approximately 55 m from a mineralized interval. Sample RKA-17-70 from the Anubis zone is located within a mineralized interval yielding 8.91 g/t Au over 3.05 m (unpublished ATAC Resources database). Subsequently, a 374-m-long borehole (OS-12-130; Fig. 1D) representative of the Conrad deposit was selected for a detailed study, with 98 HAWK pyrolysis analyses (including 7 duplicates) and reflectance measurements on 12 samples from low-and high-grade gold zones. The core from this borehole is stored at the Yukon Geological Survey (YGS) core facility in Whitehorse. The main characteristics of mineralized intervals and host rocks in borehole OS-12-130 are summarized in Fig. 2. Criteria used to pick samples for both programed pyrolysis and reflectance analyses include fine grain size (shale to silt) and dark color (medium/dark grey to black).
A detailed description of the types of organic matter in the studied samples is presented in Pinet et al. (2022c), and salient points follow. Pyrobitumen is either disseminated in the rock, fills irregularly shaped pores (1-50 µm) and fractures, or is concentrated along stylolites (Pinet et al. 2022c). Pore-filling (± fracture-filling) pyrobitumen is predominant and has resulted from the thermal cracking of liquid hydrocarbon emplaced before thermal maximum and subsequently transformed to pyrobitumen through thermal cracking. In the organic-rich shale intervals, pyrobitumen is indigenous and occurs as thin laminae with shapes that resemble precursor algae transformed to pyrobitumen, or is disseminated in the shale matrix.

Samples for fluid inclusions and clumped isotope analyses
Due to the fine-grained nature of quartz associated with gold in main ore-stage mineralization (jasperoid of Lovering 1972), fluid inclusions suitable for microthermometry were not identified in any samples from the Rackla deposits. However, fluid inclusions suitable for microthermometric analysis are hosted by late ore-stage hydrothermal calcite and fluorite.
Thirteen calcite samples from the Conrad deposit (n = 7), Osiris/Sunrise deposits (n = 5), and the Anubis mineralized zone (n = 1) were selected from drill cores for either fluid inclusion analysis (n = 5), clumped isotope measurements (n = 5), or both (n = 3).The location and main characteristics of these samples are summarized in ESM1, Table S1. All samples were selected based on their location within or proximal (< 10 m true distance) to Au-rich intervals and the presence of realgar (ESM1 , Table S1). Realgar (± orpiment and fluorite) is spatially associated with Au-bearing intervals Pinet et al. 2020b) and is part of the late ore-stage mineral assemblage (calcite-realgar ± orpiment, ± fluorite; Tucker et al. 2018;Pinet et al. 2022a, b). Cathodoluminescence (CL) imaging indicates that blocky calcite has a largely uniform orange luminescence, excluding samples RKA-17-88, RKA-17-92, and RKA-18-21 that also include dull-brown zones. Occasional brighter orange microfractures are present in some samples (ESM1 ,  Table S1). To avoid mixing of different phases, sub-sampling for clumped isotope analyses was performed using a microdrill in calcite zones showing homogeneous CL responses.

Samples for apatite fission-track (AFT) analyses
Fission-track analyses were conducted on samples located 0.9 and 4.3 km from Carlin-type gold-deposits, north of the Nadaleen Fault (Fig. 1). These samples were part of a regional dataset collected by the Yukon Geological Survey (Pinet and Brake 2018).

Rock-Eval and HAWK programed pyrolysis
A total of 132 bulk samples were analyzed through programed temperature heating by Rock-Eval 6™ (n = 6) and HAWK (n = 126) pyrolysis to document the TOC content and the maturation level (Lafargue et al. 1998). Rock-Eval 6™ and HAWK programed pyrolysis consists of programed heating of a small amount of powdered rock (~ 70 mg) to determine the quantity of free hydrocarbons present in the sample (S1 peak) and the amount of hydrocarbons and oxygen-containing compounds (CO 2 ) that are produced during the thermal cracking of the insoluble organic matter (kerogen) in a sample (S2 and S3CO 2 peaks respectively). The pyrolysis stage, conducted in an N 2 environment, begins with a 3-min iso-temperature step at 300 °C to release free hydrocarbons in the samples (S1, mg HC/g rock). Then, the temperature is increased at a rate of 25 °C/min up to 650 °C. This results in thermal cracking and the release of hydrocarbons and the oxygen contained in pyrolizable kerogen (S2 in mg HC/g rock and S3 in mg CO 2 /g rock). Samples are then automatically transferred to the oxidation oven and heated from 300 to 850 °C with a heating rate of 20 °C/min to measure the residual inert organic carbon (S4, mg CO, and CO 2 /g rock and residual carbon, RC, wt.%) and a portion of the mineral carbon (MinC, wt.%). Total organic carbon (TOC, wt.%) is quantified as the sum of the total quantity of organic matter released during pyrolysis (pyrolizable carbon, PC, wt.%) and the oxidation step (RC, wt. %). The oxygen index (OI) is calculated by normalizing the quantity of pyrolizable CO 2 (S3) to total organic carbon (S3/TOC × 100) and is proportional to the elemental O/C ratio of the kerogen. The hydrogen index (HI) corresponds to the ratio S2/TOC × 100 and is proportional to H/C (Lafargue et al. 1998).
During programed pyrolysis, the thermal cracking of kerogen (S2 peak) reaches a maximum at temperature Tmax (°C), which is a commonly used thermal indicator. The Tmax value becomes erratic below a certain level of maturation (~ 1.5% vitrinite reflectance, Ro), because the kerogen has lost almost all generative potential and the S2 peak is too small to determine an accurate Tmax. Therefore, in overmature samples with low S2, the Tmax cannot be used as a reliable thermal maturity proxy.

Pyrobitumen reflectance
Pyrobitumen reflectance (% BRo) analyses were conducted on 14 polished blocks with a cold-setting epoxy-resin mixture. Measurements were carried out under oil immersion following the Standard Test Method D7708 from the American Society for Testing and Material (ASTM 2014). The standard reference used for reflectance measurement was yttrium-aluminum-garnet with a standard reflectance of 0.906% under oil immersion.

Fluid inclusion measurements
Eight doubly polished 100-mm-thick wafers of calcite were prepared for fluid inclusion petrography and microthermometry (ESM1 , Table S1). Samples were cut using a liquid-cooled diamond rotary saw to avoid the re-equilibration of fluid inclusions (Goldstein 2003).
Fluid inclusion microthermometric measurements were carried out using a Linkam THMG 600 heating-freezing stage at the Geological Survey of Canada, Calgary. Calibration with precision of ± 1 °C at 375 °C and ± 0.2 °C at -56.6 °C was conducted using synthetic H 2 O and CO 2 fluid inclusion standards. Heating experiments were completed prior to conducting cooling experiments to reduce the risk of stretching or decrepitating the fluid inclusions. To avoid fluid inclusion stretching during heating, we measured the T h of all visible fluid inclusions in the same sample once. Fluid salinities were calculated from final ice melting temperatures (T m ) following Bodnar (1993).

Clumped isotope (Δ 47 ) measurements and data processing
Eight samples were analyzed at the Delta lab (Geological Survey of Canada, Quebec) using a Thermo Fisher MAT 253 equipped with seven shielded Faraday cups (m/z = 44-49) and an additional half mass detector at m/z = 47.5 for live background monitoring. The preparation of carbonate samples was performed with an online KIEL IV carbonate device modified according to Schmid and Bernasconi (2010). Each sample was first acidified for 500 s with three drops of 104% phosphoric acid prepared following the method of Burman et al. (2005). Evolved CO 2 and H 2 O were trapped continuously at LN 2 temperature followed by a 60-s pumping of non-condensable gases. The CO 2 was then released at -100 °C and cryopumped through a Porapak trap surrounded by Ag wool, maintained at -14 °C, into a micro-volume at LN 2 temperature. The purified CO 2 was then released into the ion source of the mass spectrometer through an inert capillary. Data for δ 18 O, δ 13 C, and δ 47 were acquired simultaneously using the long integration dual inlet (LIDI) method (Müller et al. 2017). Each measurement comprised 70 cycles of 10-s integrations of the unknown gas followed by a similar integration process for the working gas (δ 13 C = -4.5‰ V-PDB, δ 18 O = 15.9‰ V-SMOW). Each sample was analysed at least 14 times, and standards were interspersed in a 50/50 proportion in order to achieve a precision of < 0.01‰. The data acquired for this study has been archived in the EarthChem, ClumpDB database under the following https:// doi. org/ 10. 26022/ IEDA/ 112497.
Raw beam intensities were imported in Easotope (John and Bowen 2016) for calculation of the Δ 47 . The IUPAC parameters (Brand et al. 2010) were used to correct for 17 O interferences as suggested by Schauer et al. (2016) and Daëron et al. (2016). We adopted a fully carbonatebased calibration scheme using the carbonate standards (ETH1-4) developed at ETH-Zurich (Meckler et al. 2014) and used the recently accepted InterCarb values (Bernasconi et al. 2021). ETH-1 and ETH-2 were used to correct for the non-linearity effect in the ion source, whereas ETH-1, ETH-2, and ETH-3 were used to construct the empirical transfer function (ETF) for translation of the measurement to the CO 2 equilibrium scale (CDES70C). The calibration was made before and after each analysis of each unknown using a running window of 56 standards to correct for slight drift in ion source stability. Negative background corrections were performed using the pressure baseline correction (Bernasconi et al. 2013;He et al. 2012) after every 46 analyses. This was achieved by using high-voltage scans made between 5 and 25 V intensities (m/z = 44) in 5-V increments to produce a correction function that can be applied to the decaying intensity inherent to the LIDI method of measurement. A composite T-Δ 47 produced from the calibration of Jautzy et al. (2020) and recalculated using the InterCarb ETH values (Bernasconi et al. 2021) and all the calcite temperature anchors from Anderson et al. (2021) (all projected back to 70 °C using the acid reaction correction of Petersen et al. 2019) was applied to convert Δ 47 values into temperatures.

Apatite fission track analyses
Apatite grains for AFT dating were extracted from samples following conventional separation methods. Apatite fission track ages were determined by Apatite to Zircon Inc. from spontaneous fission track densities in individual apatite grains and relative uranium concentrations using the LA-ICP-MS method (see analytical procedure in Donelick et al. 2005). Electron microprobe micro-analysis data were collected for the same age/length grains that have undergone AFT analysis. These apatite elemental composition data provide kinetic information for use in the thermal history modeling.

Rock-Eval 6 TM and HAWK programed pyrolysis
Results of programed pyrolysis analysis are presented in Electronic Supplementary Material (ESM1, Table S2). The TOC content ranges from 0.25 to 3.18 wt.% with a mean value of 1.05 wt.%. Considering the high degree of maturation of these samples (see below), the original organic matter content was probably higher than that measured today (Jarvie et al. 2007) and comparable to some Phanerozoic hydrocarbon-prone basins (Tissot and Welte 1984).
All surface and core samples analyzed through programed pyrolysis have S1 (free hydrocarbon) and S2 (hydrocarbon potential/kerogen content) values close to 0 (average = 0.04 ± 0.04 mg HC/g for both S1 and S2). This indicates that there is no free hydrocarbon and hydrocarbon potential remaining, and that samples are overmature (maximum temperature experienced by the sample > 150 °C). Due to the very low S2 content, the Tmax (i.e., thermal maturity proxy) varies from 289 to 668 °C, which is not reliable. The mean hydrogen index (HI) and oxygen index (OI) of the samples are 3.8 ± 4.0 mg HC/g TOC and 47 ± 32 mg CO 2 /g TOC, respectively (ESM1, Table S2). All these values are typical of overmature samples.

Pyrobitumen reflectance
Pyrobitumen reflectance (BRo) measurements provide an indicator of the thermal maturity of the samples. A total of 12 samples from borehole OS-12-130 record mean BRo values of 2.35-2.88% ( Fig. 2D; Table 2), which indicates high thermal maturity, further confirming the programed pyrolysis data indicative of overmature samples. The lowest BRo (2.35%) value corresponds to a high-grade sample (13.35 ppm Au over 2.21 m) at 224 m (Fig. 2). The average pyrobitumen reflectance (% BRo) value for all samples from borehole OS-12-130 is 2.61%, which corresponds to 2.77% equivalent vitrinite reflectance (VRo eqv ) using the conversion equation of Landis and Castaño (1995). A maximum paleo-temperature of ca. 220 °C is estimated based on the relationship between Ro and temperature as proposed by Barker and Pawlewicz (1994) for "slow" (regional) heating rates. Slightly higher temperature estimates would be associated with "fast" (hydrothermal) heating rates.
The mean Bro data for the samples from the Sunrise deposit (Neoproterozoic laminated pyritic shale with mainly dispersed pyrobitumen in the matrix) and from the Anubis zone (dark gray Paleozoic siltstone with pore-filling solid pyrobitumen) are 1.8 ± 0.15% and 1.76 ± 0.18%, respectively (Table 2), slightly lower than for borehole OS-12-130 but still indicative of an overmature state.

Fluid inclusions
In calcite crystals, fluid inclusions occur mainly along cleavage planes (Fig. 3A and B) and crystal boundaries as random individual fluid inclusions ( Fig. 3C and D) and as clouds of numerous fluid inclusions without any specific spatial relationship with calcite cleavage planes (Fig. 3E, F, and G). The similar liquid/vapor (LV) ratio of fluid inclusion assemblages suggests that fluid inclusions did not neck-down after the fluid inclusions have reached the LV curve on their retrograde path (Goldstein 2012). Most isolated inclusions do not show clear petrographic evidence for a relationship to growth of the host crystal and cannot be classified as primary or secondary using the criteria provided by Roedder (1984). For this reason, interpretation of the fluid inclusions as primary (n = 85), secondary (n = 17), and pseudo-secondary (n = 12) was done on a part of the set of measured inclusions only (n = 248 in calcite), with others being classified as undetermined. Fluid inclusions in fluorite (sample 18-06) are much more abundant and occur as fluid inclusion trails (Fig. 3H). Because of lack of clear crosscutting relationships and/or clear crystal growth zonation, the identification of the origin of fluid inclusions, especially in fluorite, was not possible. Fluid inclusions are dominantly liquid vapor (LV) in both calcite and fluorite (Fig. 3), with rare one-phase vapor inclusions (Fig. 3E).
Overall, the inclusions are generally small (3-15 µm, average ~ 5 µm; Fig. 3), which made observing the last ice melting temperature very difficult and the salinity of fluids impossible to determine accurately in some cases. Under ultra-violet (UV) light, none of the fluid inclusions fluoresced, indicating that they are aqueous.
The microthermometric results are presented in Table 3. Homogenization temperature (T h ) for all calcite-hosted fluid inclusions range from 77.1 to 254 °C (Fig. 4A), and the average for all inclusions is 150.6 °C. This average is close to the  mean T h for primary (144.3°) and "undermined" (149.7°) fluid inclusions, as well as to the relatively small number of inclusions interpreted as pseudo-secondary (148.7 °C) and secondary (158.4 °C). All samples show unimodal distribution of T h values with two exceptions, samples RKA-17-92 and RKA-17-108 (n = 25 and n = 26, respectively), which exhibit a slightly bimodal distribution. The majority of samples has a similar range of T h values, and mean temperatures between 123.3 and 150.5 °C, except for sample RKA-18-06 from the Conrad deposit that has a notably higher mean T h value of 172.6 °C for calcite (n = 84). However, the mean T h value of fluid inclusion in fluorite for this sample is 154.4 °C (n = 10), which is ~ 18 °C cooler than the mean T h value for calcite. At a first glance, the microthermometry results for the Conrad deposit (n = 5) suggest that hydrothermal fluids were hotter (T h = 169.5 °C) than those of the Osiris/Sunrise deposits (n = 3, T h = 134.7 °C). However, this difference decreases to < 3 °C if the data for sample RKA-18-06 are removed from the calculation.

Clumped isotopes
Clumped isotope results are presented in Table 3 and Fig. 5. Samples of late ore-stage hydrothermal calcite from the Conrad and Osiris/Sunrise deposits have δ 18 O VSMOW and δ 13 C VPD values of 15.52-24.11‰ and − 4.07-1.02‰, respectively. Samples RKA-17-108 (Conrad deposit) and RKA-18-21 (Anubis zone) have the highest δ 18 O VSMOW values of 24.11 and 23.17‰ (Fig. 5A), respectively. However, their δ 13 C VPD values are ~ 0‰ and within the range defined by other samples (Fig. 5A). With respect to geology, sample RKA-17-108 is from the highly fractured hinge zone of the Conrad anticline (Fig. 6) and was analyzed twice with similar results.
The five samples from the Conrad deposit yielded apparent temperatures (Δ 47 ) of 91 ± 5 to 162 + 10/ − 11 °C with the highest temperature recorded by the two deepest samples (RKA-17-65 and RKA-17-105) and the lowest for a shallow sample (RKA-17-104) proximal to the Nadaleen Fault (Fig. 6). Decrease in yielded temperature with altitude possibly indicates the cooling of hydrothermal fluids during upward migration. Sample RKA-17-104 is the closest to a mafic dike (Fig. 6), but it does not record a thermal imprint expected to be associated with dike emplacement, further suggesting that hydrothermal calcite post-dates the dike, in agreement with the strong alteration of the dike (Pinet et al. 2022b). Finally, two samples from the Osiris/ Fig. 4 Histograms of microthermometry results. A Homogenization temperatures (T h ) and B salinity calculated from ice melting using Bodnar (1993). (C) Plot of T h vs salinity for fluid inclusions in calcite. Note that no clear trend is suggested by the data Sunrise deposit yielded apparent temperatures T(Δ 47 ) of 109 ± 7 and 132 ± 7 °C.

Apatite fission track
The two dated samples yielded overlapping AFT pooled ages of 49.2 + 5.6/-5.0 and 44.2 + 9.2/-7.6 Ma (ESM1 ,  Table S3). For both samples, single AFT grain ages form a relatively narrow age population with no grain older than 225 Ma. This indicates that the samples have been completely annealed and experienced temperatures well in excess of 120 °C. Track lengths average 13.61 and 13.83 µm, with relatively narrow unimodal distribution.

Discussion
The different thermal indicators used in this study each record only part of the geological history, yet direct comparison of these datasets requires caution (Table 4). Programed pyrolysis analyses and reflectance measurements provide information on the maximum temperature experienced by a given sample, which may or may not be associated with the Au-mineralizing event. For example, the fact that the lowest Bro value (2.35%) for borehole OS-12-130 corresponds to a high-grade sample (Fig. 2) suggests either that maximum thermal conditions are not related to Au-bearing fluids or that hydrothermal fluid flow was not restricted to mineralized intervals (i.e., that the thermal imprint of fluid flow varied only slightly for scales greater than the borehole length). Apatite fission track analyses are very sensitive to temperatures that characterize the partial annealing zone (~ 60-120 °C), but do not record the history at higher temperatures. Fluid inclusion microthermometry and clumped isotope analyses provide information on late ore-stage hydrothermal calcite and fluorite related to the mineralizing event at ca. 74 Ma. However, these data must be interpreted considering possible effects caused by younger event(s) unrelated to Au mineralization that may have led to the formation of secondary inclusions and to isotopic reordering.

Comparison between fluid inclusion and clumped isotope results
Both fluid inclusion microthermometric and clumped isotope data were generated for three samples from the Conrad deposit. LA-ICP-MS data presented in Pinet et al. (2022b) indicate that these three samples are characterized by a single calcite phase and by a simple systematic on the U-Pb Tera-Wasserburg plots, suggesting that they record a single hydrothermal event at ca. 74 Ma. Fluid inclusion mean T h and clumped isotope data are within analytical error for samples RKA-17-108 and RKA-17-105, but there is a difference of ca. 39 °C for sample RKA-17-65. Post-formation processes are one possible explanation for the lower temperature recorded in fluid inclusions in sample RKA-17-65. Even if analysis of sample RKA-17-65 was carefully monitored, the lower Th value compared to the T(Δ 47 ) value may be due to inclusion necking or leaking after their formation. An alternative interpretation in which the clumped isotope temperature is overestimated due to solid-state diffusion of O and C atoms seems improbable, as there is no evidence that the sample experienced a high-temperature post-crystallization event. Comparison of fluid inclusion and clumped isotope datasets should also consider the difference in volume of the sample analyzed and the time frame of geological events. The clumped isotope data represent an "average" over a large crystal volume (possibly containing multiple generations of fluid inclusions), compared with measurements for individual 3-to 15-µm inclusion. This difference may mask variations in fluid temperature over short time periods in evolving hydrothermal systems. This interpretation is supported by the correspondence in mean T h and T(Δ 47 ) for two samples that have ranges in T h of 118.5-182 °C (sample RKA-17-105) and 86.1-167.3 °C (sample RKA-17-108).

Variability of temperature estimates
The three samples with both fluid inclusion measurements and clumped isotope analyses are characterized by "simple" textures and geochemical zoning based on LA-ICP-MS elemental maps and in situ LA-ICP-MS calcite dating (Table 3). Clumped isotope temperatures T(Δ 47 ) for these samples are 125-162 °C. These T(Δ 47 ) values are consistent with the temperatures of 138 and 144 °C reported by Steiner et al. (2022) for UV-fluorescent veins. Two additional samples (RKA-18-24 and RKA-18-37) with "complex" or heterogeneous texture were also analyzed for clumped isotopes, LA-ICP-MS elemental mapping, and U-Pb dating.
The T(Δ 47 ) of 97 and 109 °C for these two samples is significantly lower compared with temperature estimates for "simple" samples. This discrepancy most likely reflects the effects of a post-crystallization hydrothermal event indicated by younger U-Pb ages. Fluid inclusion T h values on sample RKA-17-95 also provide some insights on homogenization temperature variability. Calcite in this sample occurs in 1) two parallel and well-defined, blocky calcite veins that are 0.8 and 3 cm wide (veins V1 and V2 in Fig. 7) and 2) in smaller and irregularshaped veins (zone A in Fig. 7A). U-Pb spot analyses in zone A yielded robust overlapping ages of 72.2 ± 2.3 Ma age (see Fig. 6 in Pinet et al. 2022b). In contrast, some U-Pb ages in veins V1 and V2 are significantly younger than mineralization due to partly connected, thin (10 s of µm) microfractures that are significantly enriched in U (Pinet et al. 2022b). Fluid inclusion data for zone A record a mean T h value of 134.6 ± 26.2 °C compared with values of 123.7 ± 29.5 and 124.8 ± 21.3 °C for veins V1 and V2, respectively (Fig. 7B). This ca. 10 °C difference in combination with the younger U-Pb ages supports a lower temperature post-crystallization thermal event.

Thermal history
Due to an absence of post-Permian rocks, the Mesozoic-Cenozoic history of the eastern Rackla belt is poorly constrained by geological evidence. However, information can be gained by inverse modeling of AFT data. The two AFT ages are not significantly different from the 47.0 + 8.6/ − 7.3 Ma AFT age (based on an unknown number of grains) and a "single grain pooled age" of 37.4 + 7.4/ − 8.6 Ma reported by Tucker (2015) for two samples collected within Au-bearing intervals from the Conrad deposit.
In this study, inverse models were performed using the HeFTy software (Ketcham 2013) taking into account a regional maximum temperature of > 150 °C based on programed pyrolysis data for outcrop samples collected outside known mineralized zones that document overmature pyrobitumen. Among the parameters used during inverse modeling (see ESM2 for the detailed modeling strategy), the temperature constraint at ca. 74 Ma is clearly the most critical. We consider the lowest clumped isotope temperature (91 °C) to represent a maximum temperature for the host rocks at ca. 74 Ma, because late orestage hydrothermal minerals (calcite, fluorite, and realgar) cannot have formed at temperatures lower than the host rocks.
Inverse modeling of sample 12-SI-058-1indicates that the measured AFT age and fission-track length can be adequately fitted with a "simple" inverse model (Fig. 8). Similar results were obtained for sample 12-SI-054-2 even if apatite ages are more dispersed. Good T-t paths (i.e., with goodness-of-fit values of > 0.75 using Kuiper's statistics) at 74 ± 4 Ma for sample 12-SI-058-1 are limited to temperatures between 91 °C, the lower temperature recorded by late hydrothermal calcite, and 65 °C, even if lower temperatures were allowed by the model (Fig. 8). This temperature range can be interpreted either as being associated with the waning stage of hydrothermal fluid flow or as the "regional" temperature of the host rock at 74 Ma. The envelope including all acceptable paths (i.e., with goodness of fit of > 0.05) is characterized by a marked decrease in temperature at ca. 35 Ma likely related to exhumation. Also, the maximum temperature reached by the sample before 74 Ma spreads the whole range of temperatures allowed by the model between 150 °C (attested by programed pyrolysis data) and 220 ± 10 °C (attested by pyrobitumen reflectance). As AFT data do not record events that occurred at temperatures greater than the partial annealing zone, it is impossible to know whether samples experienced their maximum temperature in association with the building of the Cordillera, dike intrusion, and/or hydrothermal fluid flow.
Our two "preferred" T-t scenarios for Carlin-type gold deposits of central Yukon (Fig. 9) differ with respect to when rocks in borehole OS-12-130 experienced maximum temperatures (~ 220 °C). The maximum temperature in scenario A corresponds to regional maximum tectonic burial, whereas in scenario B it is related to local hydrothermal fluid flow. Whatever the scenario, hydrocarbon migration documented by pyrobitumen filling pores and fractures occurred during the sedimentary and/or tectonic burial of the sedimentary succession at temperature far below 200 °C, i.e., before Au mineralization. This timing agrees with petrographic observations suggesting that migrated hydrocarbon filling pores were matured after their emplacement ( Fig. 10; Pinet et al. 2022c). Therefore, hydrocarbon-bearing-fluids did not transport Au at the deposit scale, an interpretation in agreement with the absence of hydrocarbons in fluid inclusions in the late-mineralization calcite and fluorite.

Origin of fluids
Previous studies demonstrated that late ore-stage hydrothermal calcite has isotopic ratios distinct from unaltered host rocks and pre-ore veins in both the Conrad and Sunrise/Osiris deposits ( Fig. 5A; Tucker 2015; Beaton 2015; Steiner and Hickey 2019; Pinet et al.2022a;Steiner et al. 2022). Late ore-stage hydrothermal calcite is characterized by δ 18 O (11.2-21.3‰ VSMOW) and δ 13 C values (-4.5 to 1.3‰ VPDB) that are either lower (Conrad deposit) or higher (Sunrise and Osiris deposits) compared with surrounding rocks. The fact that samples of late ore-stage hydrothermal calcite have similar δ 13 C values in the higher δ 13 C Conrad and lower δ 13 C Sunrise/Osiris limestones indicates similar fluid sources for each zone and limited carbon isotope exchange between late-ore hydrothermal fluids and the immediate host rocks on the mm-100 m scale (Pinet et al. 2022a;Steiner et al. 2022). Clumped isotope data provide a means to study the evolution of fluids, because their δ 18 O characteristics can be reconstructed using clumped and bulk isotopes of the carbonate that are measured simultaneously. The composition of the fluids in isotopic equilibrium with the vein calcite was calculated using the measured T(Δ 47 ) and δ 18 O values in conjunction with the calcite-water fractionation factors of Kim and O'Neil (1997) and Daëron et al. (2019). These data are presented in Fig. 11 and form a trend that indicates mixing between a higher-temperature end-member characterized by δ 18 O H2O > 6‰ and a lower-temperature end-member with δ 18 O H2O < − 2‰ if using the fractionation of Kim and O'Neil (1997) (> 4 and < − 4‰ using the fractionation of Daëron et al.2019). Sample RKA-17-108 represents a significant departure from this trend. Meteoric water, with δ 18 O H2O values between − 23 and − 20‰ (VSMOW) for the latitude of central Yukon (Clark and Fritz 1997), is a potential candidate for the low temperature end-member. Fig. 9 Preferred T-t scenarios. The maximum temperature corresponds to maximum tectonic burial in scenario A, whereas maximum temperature is related to hydrothermal fluid flow in scenario B However, the salinity of low homogenization temperature (T h ) fluid inclusions is not significantly different from the salinity of high T h (Fig. 4C) inclusions, which is difficult to reconcile with an increased amount of meteoritic water in low-temperature hydrothermal fluids. Brines residing within fractures in tight rocks are another candidate for this low-temperature end-member, because their evolution over geological time scale under low temperature and low water-rock interactions is thought to promote a 18 O-depletion due to various degrees of isotope exchange with silicarich rock, as recently documented in shield settings (Warr et al. 2021). Such brines could explain the presence of relatively high salinity yet 18 O depleted fluids for some of the samples. The origin of the high-temperature end-member fluid is difficult to ascertain as its δ 18 O H2O value is close, or within the fields of magmatic, metamorphic, and, possibly, connate waters.

Comparison with Carlin-type deposits of Nevada
Carlin-type deposits of Nevada and Yukon share several characteristics, and the data give the opportunity to compare the thermal history for both regions (Table 1). In Carlin-type deposits of Nevada, fluid inclusions in jasperoid quartz at Jerritt Canyon (Hofstra 1994), Meikle (Lamb 1995), Carlin (Kuehn and Rose 1995), Turquoise Ridge (Shigehiro 1999), Getchell (Cline and Hofstra 2000), and Betze-Post (Lubben 2004;Lubben et al. 2012) deposits provide relatively consistent results. These studies indicate that ore-stage fluids had low to moderate temperatures (115-240 °C, mostly 160-220 °C), relatively low-salinity (0-10 wt.% NaCl eq, mostly < 6 wt%, but locally as high as 17 wt%), and they contained CO 2 (< 4 mol%) and low to trace amounts of H 2 S and CH 4 (Cline et al. 2005;Muntean 2018). Evidence of boiling is generally lacking, with a possible exception reported by Groff (2019). A decrease in the fluid temperature at near constant salinity is recorded in fluid inclusions for late ore-stage minerals including fluorite (100-160 °C at Getchell, Cline, and Hofstra 2000), realgar (65-120 °C at Turquoise ridge; Groff 2020), and calcite (Carlin mine, mean T h = 141 °C, Kuehn 1989;137 °C, Groff et al. 2002; Getchellmine: T h = 100-140 °C, Cline and Hofstra 2000; Betze mine: mean T h = 166 °C, Groff et al. 2002). The greater relative abundance of realgar in Au-rich zones in the Yukon deposits relative to most of those of Nevada led Pinet et al. (2022a) to hypothesize a slightly lower temperature of formation for the Yukon deposits. However, the temperature estimates for the late ore-stage minerals in deposits of Nevada are not significantly different from those obtained in Yukon (Table 1), suggesting that Carlin-type deposits of Nevada and Yukon formed at similar temperatures.

Conclusion
Several techniques were used to investigate the thermal framework of Au-bearing Carlin-type deposits in central Yukon and better understand the evolution of such class of ore deposit. The Tmax parameter derived from programed pyrolysis analyses indicates that organic matter is overmature and records temperatures of > 150 °C regionally. Temperature of crystallization of the late-ore mineral assemblage associated with the waning stage of the mineralization event at ca. 74 Ma has been investigated through fluid inclusion analyses and clumped isotope measurements. Among the three samples with both fluid inclusion and clumped isotope analyses, two yielded similar temperature estimates, whereas one has a significantly lower fluid inclusion temperature. Data also indicate that samples previously identified as Fig. 10 Organic petrography photomicrographs taken by incident white light with oil immersion × 50 objective. A Accumulation of fused fine crystalline pyrite (pyr) in the host rock. Pores within the host rock filled with pyrobitumen (pyr); borehole OS-12-130; 237.2 m. B Boundary between organic-rich shale and host rock with pyrobitumen that accumulated mainly at the boundary (borehole 190.7 m) "complex" or texturally and geochemically heterogeneous have lower temperature estimates, suggesting that calcite locally records post-mineralization processes. Mean fluid inclusion T h in eight samples range from 123.3 to 172.6 °C, with a mean calculated salinity of 4.8% NaCl equiv. These values are comparable to clumped isotope temperatures obtained for eight calcite samples ranging from 91 to 162 °C. Calculation of the composition of the fluid in equilibrium with calcite during crystallization supports mixing of two fluids of variable temperature and isotopic signature, the higher temperature end-member being characterized by higher δ 18 O H2O values. In thermal scenarios based on Tmax, fluid inclusion analysis, clumped isotope measurements, and apatite fission track modeling, the ca. 220 °C temperature estimated through pyrobitumen reflectance measurements is either related to maximum tectonic burial prior to gold mineralization or to the early stages of the Au-bearing hydrothermal event.
This study demonstrates the benefit of using different analytical techniques, as each method brings some information on the temperature evolution of Au-mineralized zones and/ or their host rocks. Although underutilized, organic matter geothermometers (programed pyrolysis data and reflectance measurements) provide robust and relatively inexpensive temperature constraints both on the regional and deposit scales for deposits hosted in sedimentary rocks. Comparison of fluid inclusions and clumped isotopes (Δ 47 ) results in various settings and, on different scales, will likely generate new insights on hydrothermal fluid temperature variations in space and time. Moreover, clumped isotopes (Δ 47 ) in lower temperature and younger Carlin-type deposits are shown to yield accurate temperature estimates compared to Archean orogenic gold deposits where reordering of isotopic bounds over time has obliterated the higher temperature of formation (Quesnel et al. 2022). This work is part of the Natural Resources Canada's Targeted Geoscience Initiative and is Geological Survey of Canada contribution number 20220218 and Yukon Geological Survey contribution number 061.
Author contribution All authors contributed to the study conception and design. Sample collection was done by Nicolas Pinet and Patrick Sack. Sample preparation and analysis were performed by Omid Haeri-Ardakani and Josué Jautzy. The first draft of the manuscript was written by Nicolas Pinet, and all authors commented on previous versions of the manuscripts. All authors read and approved the final version of the manuscript.
Funding Open access funding provided by Natural Resources Canada.

Conflict of interest The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not Fig. 11 Plot of δ 18 O H2O vs T(Δ 47 ). At the top of the plot are δ. 18 values for various types of water (Clark and Fritz 1997;Bierlein and Crowe 2000) permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.