Geochemistry of Quartz and Fluid Inclusions Associated with Gold, Tungsten, and Antimony in the Stibnite-Yellow Pine District, Idaho Open Access

Deposits in the historic Stibnite-Yellow Pine mining district, Valley County, east-central Idaho (Figs. 1, 2) were mined for gold, silver, antimony, and mercury, and the district was the largest domestic tungsten producer from 1941 to 1945 (~5,800 t W; Huss et al., 2014; Vikre et al., 2025a). This district is an important domestic source of antimony with a past production of 40,011 metric tonnes (t) Sb, along with 985,602 troy ounces Au, 208,855 troy ounces Ag, 6,790 t W, and ~1.3 million fluid ounces Hg, with contained resources (measured + indicated) of 93,388 t Sb, 6.034 million ounces (Moz) Au, and 8.814 Moz Ag (Cookro et al., 1988; Karl et al., 2020; Zimmerman et al., 2021).

In the Stibnite-Yellow Pine district, ore occurs where the N-trending Meadow Creek and West End faults splay to the east or intersect E-NE–trending faults. Host rocks consist of the Late Cretaceous Atlanta lobe of the Idaho batholith, comprising biotite granodiorite and two-mica granite cut by small stocks and dikes of leucogranite, as well as metamorphosed Neoproterozoic-lower Paleozoic schist, marble, quartzite, and conglomerate rocks (Bailey, 1934; Cole and Bailey, 1948; Lewis, 1984; Smitherman, 1985; Cookro et al., 1988; Stewart et al., 2016; Zimmerman et al., 2021; Lewis et al., 2025; Figs. 1, 2). The host rocks are variably silicified, K-metasomatized (adularia and sericite), sulfidized, argillized, or carbonatized (Lewis, 1984; Gillerman et al., 2025). The mineral paragenesis in the Stibnite-Yellow Pine district has been described previously by Lewis (1984), Gammons (1988), and Cookro et al., (1988) and more recently by Vikre et al. (2025a, b), Konyshev (2020), Konyshev et al. (2025), and Gillerman et al. (2019, 2025). Five hydrothermal events were geochronologically defined by Gillerman et al. (2025): event 1 [86–75 Ma] early milky quartz (± calcite), muscovite, and pyrite with minor molybdenite and other base metal sulfides; event 2 [70–56 Ma] veinlets and disseminations of quartz, sericite, calcite, auriferous pyrite, and arsenopyrite; event 3 [57–56 Ma] vein and disseminated potassium feldspar, quartz, dolomite, calcite, scheelite, and minor pyrite and arsenopyrite; event 4 [52–51 Ma] Au-Ag–bearing veins that contain adularia; and event 5 [47 Ma] quartz, stibnite, and calcite, with minor Ag-bearing sulfosalts in breccia cement and veins and distal zones of cinnabar and realgar at the Fern and Hermes mines (Schrader and Ross, 1926; Gillerman et al., 2025, and references therein). The events overprint one another, with the exception of cinnabar and realgar of event 5 occurring in distal zones from the gold, silver, tungsten, and antimony. Event 4 is more prevalent in the West End deposit, hosted in the metamorphosed Neoproterozoic-lower Paleozoic rocks.

The purposes of this paper are to (1) characterize the paragenesis and chemistry of quartz associated with the gold, tungsten, and antimony hydrothermal events hosted in the intrusive rocks in the district, (2) characterize fluid inclusions in quartz associated with each of these events, (3) characterize fluid inclusions in scheelite, (4) characterize fluid inclusions in stibnite, and (5) discuss the pressure-temperature-composition (P-T-X) conditions of the ore-forming fluids in relation to the geochronology and mineral paragenesis described in accompanying papers of this special issue (Vikre et al., 2025a, b; Gillerman et al., 2025; Konyshev et al., 2025). We will discuss whether the deposits formed from multiple overprinting hydrothermal events or from a progressive hydrothermal Au-W-Sb event that was later overprinted by Ag-Au related to Challis magmatism. Understanding the P-T-X of the ore-forming fluids is key to development of deposit models, which are needed for exploration and discovery of similar Au, W, and Sb deposits regionally and in analogous settings elsewhere.

For this study, six drill holes were studied: four from the Yellow Pine intrusion-hosted deposits, one from the Hangar Flat intrusion-hosted deposits, and one from the West End metasedimentary rock-hosted deposits. Approximately 360 samples were collected from these six cores, along with ~20 surface samples (Figs. 1, 2; see Marsh et al., 2023). One surface sample was also collected from an Hg-Sb deposit north of the Stibnite-Yellow Pine district (Antimony Rainbow; Fig. 1). Optical microscopy aided by cathodoluminescence (CL) imaging, electron microprobe analysis (EMPA), and laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) trace element analyses were used to examine crosscutting quartz types and their associated fluid inclusion assemblages (FIAs). Further characterization was carried out by microthermometry, laser Raman spectroscopy, and mass spectrometry.

Previous studies describe the texture, paragenesis, and geochemistry of hydrothermal minerals and associated FIAs in the Stibnite-Yellow Pine and adjacent Big Creek districts (Lewis, 1984; Cookro et al., 1988; Gammons, 1988; Konyshev, 2020; Konyshev et al., 2025; Fig. 1). In the Stibnite-Yellow Pine district, Lewis (1984) observed the progression of events from early gold followed by stibnite and mercury. From stable isotope data, he assumed the Stibnite-Yellow Pine district deposits developed from an early gold event at ~325°C, followed by stibnite at 180°C, and ending with the late Hg event at ~100°C. Cookro et al. (1988) described Late Cretaceous to Paleocene ages for the occurrences, with auriferous pyrite, scheelite, and stibnite occurring in overlapping stages of deposition. They report preliminary fluid inclusion observations of early, milky quartz with wispy networks of overprinting fluid inclusion populations characteristic of a deep plutonic or metamorphic environment, whereas later quartz formed by recrystallization of earlier quartz and filled open space with cockscomb, cryptocrystalline, and chalcedonic quartz that contains fluid inclusions similar to those encountered in shallow epithermal environments, suggesting temperatures of <200°C. Gammons (1988) described the mineral deposits in the Big Creek mining district, ~28 km north of the Stibnite-Yellow Pine district, as base metal-rich and Au-Ag-W-Sb mineral occurrences distributed along splays of the N-trending Independence fault (Fig. 1). The Big Creek occurrences have consistent Late Cretaceous to Paleocene ages, though Au occurs more commonly as electrum here rather than in pyrite and arsenopyrite in the Stibnite-Yellow Pine district occurrences. He described coarse, milky, deformed quartz veins with fine-grained muscovite associated with pyrite and arsenopyrite and elevated gold concentrations with no suitable fluid inclusions to measure. These veins were overprinted by vuggy quartz-stibnite veins that have two-phase H₂O-NaCl ± CO₂ or three-phase H₂O-NaCl-CO₂ fluid inclusions in quartz with average salinities and homogenization temperatures (Th_total) of ~3.4 wt % NaCl equivalent and 275°C. At the Independence deposit, Gammons (1988) describes small fluid inclusions in scheelite with filling temperatures >200°C. Konyshev (2020) describe the quartz events at the West End deposit in the Stibnite-Yellow Pine district, hosted in the Neoproterozoic-Ordovician metasedimentary rocks, as forming in four phases: (1) deep early veins formed in a ductile setting, with wavy walls, milky white quartz, brassy pyrite, minor molybdenite, and associated propylitic alteration; (2) main-stage planar veins with quartz, sericite, auriferous pyrite, and arsenopyrite cut by late calcite and dolomite; (3) epithermal vein with fine grained quartz, pink calcite, and (4) late stibnite, ± white calcite, pyrite, arsenopyrite, and electrum with narrow alteration halos of illite, smectite, montmorillonite, and minor adularia in all metamorphic units except the marble; and late-stage veins of calcite, silica, dolomite, and pyrite, ± scheelite, ± stibnite (see Konyshev, 2020, table 1.2).

Cathodoluminescence imagery reveals the sequence of growth, recrystallization, and healing of fractures in quartz and other minerals, which makes it possible to determine the entrapment history of FIAs observed within a sample by optical petrography (Goldstein, 2003). The scanning electron microscope (SEM)-CL and an electron microprobe equipped with a CL detector were used to image and map areas of interest in samples (Goldstein, 2003). The homogenization temperatures and compositions of FIAs were determined by microthermometry; active gas (e.g., H₂O, CO₂, CH₄, and N₂) estimates were obtained from laser Raman analysis (Burke, 2001); and noble gases (e.g., He, Ne, Ar) were measured from fluid inclusion bulk extractions (Landis and Hofstra, 2012; Hunt, 2015). For inclusions with low gas contents, salinities and pressures at homogenization were calculated using HOKIEflincs (Steele-MacInnis et al., 2012). For inclusions with detectable CO₂, compositions, densities, and pressures at homogenization were calculated using the Steele-MacInnis (2018) model for the H₂O-NaCl-CO₂ system. Ratios of CO₂ to CH₄ were determined from Raman data using equations from Burke (2001), and amount-of-substance fractions were calculated using the equations of state from Duan et al. (1992) and Krumgalz et al. (1996). Major and trace element compositions of quartz were determined by LA-ICP-MS (Longerich et al., 1996) and electron microprobe (ISO 22489, 2016). The Al and Ti values obtained by EMPA on the Audétat et al. (2015) quartz, run as an internal quality control reference material, were within recommended values. The concentrations of Li, Al, Ti, and Sb obtained by LA-ICP-MS on the NIST 612 glass reference material, run as an external quality control (QC), were consistently within 10% of the recommended values reported on GeoRem (http://georem.mpch-mainz.gwdg.de/). The calculated value of As in NIST 610 on one of the three experiments was outside this 10% range by 10 to 20% for three of the six measurements of that experiment. Detailed descriptions of sample collection and microanalytical techniques are provided in the Appendix.

Quartz textures and fluid inclusion petrography: The sequence of hydrothermal events, mineral paragenesis, and fluid inclusion populations and their properties are summarized in Figure 3. Well-preserved primary fluid inclusions (Roedder, 1984) are rare in the Stibnite-Yellow Pine district samples. Where preserved, primary fluid inclusions were trapped during the waning stages of hydrothermal activity. Our investigation therefore relied heavily on CL characterization of quartz coupled with petrography of secondary inclusions (Roedder, 1984) and re-equilibrated inclusions to discriminate successive hydrothermal events and deformation represented in quartz. Reequilibrated inclusions are inclusions that exhibit evidence of postentrapment deformation such as decrepitation textures or migration shadows (Diamond et al., 2010; Tarantola et al., 2010; Lambrecht and Diamond, 2014). West End deposit samples, hosted in the Neoproterozoic to Ordovician metasedimentary rocks, are inclusion poor. Where present, they are small wispy inclusions with inconsistent liquid to vapor proportions. Inclusions in quartz and calcite are very small and often decrepitated. No measurable FIAs were identified. The following inclusion observations are from samples from the Yellow Pine and Hangar Flat deposits hosted in intrusive rocks.

All primary CL characteristics in Qi are reset to homogeneous light gray luminescence in SEM-CL (Fig. 4A) and dull purple in EMPA-CL (Fig. 5A, E), indicating that disseminated hydrothermal ore deposition and alteration modified Qi in the Atlanta lobe of the Idaho batholith. Fluid inclusions occur along secondary planes and in randomly oriented clusters throughout Qi. The secondary inclusions contained in Qi were trapped during subsequent hydrothermal events and correlate with microveinlets of hydrothermal quartz (Q0, Q1, Q2, and Q3; Fig. 3) that have diagnostic CL colors and textures that are described below.

The Q0 milky hydrothermal quartz veins are early in the paragenesis. They contain an abundance of small, decrepitated inclusions that were not measured in this study. These veins are the deep early veins observed by Konyshev (2020) and are host to the molybdenite used for Re-Os analysis, yielding the age of 86 Ma for event 1(Konyshev, 2020; Gillerman et al., 2025).

The hydrothermal quartz (Q1) associated with the main stage of Au deposition, event 2 of Gillerman et al. (2025), appears dark gray to black in SEM-CL images (Fig. 4A, C) and black to yellow-blue in EMPA-CL (Fig. 5D). The Q1 occurs in veins as milky anhedral quartz and as fine-grained mottled dissemination through wall rock (Fig. 4A) In both settings, Q1 is associated with sericite and zoned auriferous pyrite. It fills secondary planes cutting Qi that can be followed into areas of quartz-sericite-pyrite alteration, where igneous feldspar and biotite grains are replaced by fine-grained potassium feldspar, sericite, rutile, and auriferous pyrite and arsenopyrite. The cloudy appearance of the milky veins in hand sample is the result of abundant and ubiquitous fluid inclusions throughout the Q1 crystals. The Q1 veins are highly recrystallized, and most inclusions were naturally decrepitated. Aqueous carbonic inclusions observed in Q1 with CO₂ contents greater than ~65 mol % may be neonates (i.e., satellite inclusions which form around the relic inclusion after deformation; Tarantola et al., 2010) formed by decrepitation of earlier aqueous carbonic inclusions in Q1. Alternatively, inclusions of this composition may be generated via preferential removal of small amounts of H₂O from aqueous carbonic inclusions by diffusion into the host quartz (Bakker, 2009). Rare planes of secondary FIAs trapped within Q1 milky quartz veins are preserved in pockets adjacent to sericite and pyrite (Fig. 4B, C). These FIAs are sparse, three-phase aqueous-carbonic fluid inclusions with equant to negative crystal shapes that consist of liquid H₂O, liquid CO₂, and a CO₂ vapor phase, with CO₂ phases together accounting for approximately 40 to 60 vol % of the inclusion (Fig. 6A).

The Q2 hydrothermal quartz, associated with event 3 of Gillerman et al. (2025), occurs as large, variably zoned crystals and recrystallized domains in Qi and Q1 that appear dark to light gray in SEM-CL (Fig. 4A) and dull blue in EMPA-CL (Fig. 5B, C, E). The Q2 is present as coarse euhedral crystals filling veins (Figs. 4D, E, G, 5B, C, E) and as fine aggregates in the wall rock (Fig. 4G). It is associated with sulfide-K–feldspar-carbonate-scheelite mineralization (Fig. 6E). The habit and well-developed zoning in Q2 suggest that this stage of quartz formed by open-space filling. Early growth zones and crystal cores are typically defined by an abundance of decrepitated inclusions (Fig. 4E-H). Some early inclusion decrepitate zones are replaced by dark SEM-CL/bright red EMPA-CL patches, which, in places, contain planes of pyrite ± arsenopyrite mineral inclusions (Figs. 4D, G; 5B, C, E). Intermediate growth zones are disrupted by a dark CL band that encapsulates these earlier features and, in some areas, shows impingement on earlier growth zones (Fig. 4G). Later Q2 growth is typically clear and inclusion poor and shows sector zoning (Fig. 4D, E). The Qi and Q1 are cut by Q2 dark-gray CL quartz that contains abundant secondary assemblages of two-phase, liquid-rich fluid inclusions that are equant to irregular in shape and consist of liquid H₂O with ~10 to 20 vol % vapor (Fig. 6B). Primary FIAs are extremely rare but are identified in crystals of pristine Q2. Primary FIAs observed in Q2 are equant in shape and liquid rich and contain 10 to 15 vol % vapor (Fig. 6C). Dark CL patches track some FIAs in Q2 from the location of entrapment along the growth zone or fracture (Fig. 4D), which is evidence of postentrapment morphological ripening of the inclusions. Lambrecht and Diamond (2014) describe such ripening as the change in shape and size and the migration of an inclusion through the host crystal, all while conserving volume and overall composition. Zones of abundant naturally decrepitated inclusions formed by reequilibration of earlier primary inclusions lack the dark CL signature, as the inclusions ruptured before morphological ripening could occur.

The Q3 hydrothermal quartz, associated with event 5 of Gillerman et al. (2025), is bright white in SEM-CL (Fig. 4A, D, G) and bright blue-yellow to white in EMPA-CL (Fig. 5B, C, E). It fringes Q2, has euhedral terminations, and occurs in veinlets that cut across earlier generations of quartz. In some zoned Q2 crystals, early cores are overprinted by dense networks of irregular fracture zones cemented with Q3 (Fig. 4G) that hosts numerous wispy naturally decrepitated inclusions (Fig. 4H). The Q3 is commonly associated with stibnite in breccia matrices, veins, and disseminations. It forms aggregates of small, sector-zoned crystals that project into open space. The Q3 overgrowths contain fine-grained muscovite inclusions aligned with primary growth zones but are devoid of fluid inclusions. In older generations of quartz, stringers of Q3 contain needle-shaped stibnite mineral inclusions and secondary planes of irregular-shaped, liquid-rich inclusion assemblages with 0 to 10 vol % vapor (Fig. 6D). The low and variable vol % of vapor and the irregular shape of this inclusion population are indicative of formation at low temperatures (<200°C; Bodnar et al., 1985). Bright Q3 is overgrown by a latest fringe of inclusion-poor quartz with dull red CL. This stage of quartz is volumetrically insignificant in the system and is only encountered in a handful of samples with banded carbonates crosscut by Q3 quartz-stibnite veins with late, low-temperature textures.

Annealing of decrepitated FIAs results in healing of the early primary inclusion zones and changes in the CL response of the host quartz (Lambrecht and Diamond, 2014). The dark gray CL patches in Q2 (Fig. 4D) and the bright CL response (Fig. 4G) of the overprinting and infiltrating Q3 are examples of annealing after decrepitation. The observed shading (SEM-CL) and color (EMPA-CL) difference between the quartz types and annealing or dissolution features are further distinguished by EMPA-CL spectral intensity versus energy (Fig. 5E). The spectra are from areas selected based on petrography of the EMPA-CL images representing each of the quartz phases.

The latest stage of hydrothermal quartz (Q4) occurs as narrow overgrowth on euhedral Q3 and infill of shattered cores of Q3 appearing bright red in EMPA-CL (Fig. 5A). Fluid inclusions in Q4 have not been measured, but they resemble those found in Q3. This quartz may represent the epithermal quartz observed by Konyshev (2020) and Konyshev et al. (2025) at West End and is present in samples with adularia.

Scheelite textures and fluid inclusion petrography: Scheelite occurs in veins, vugs, and disseminations and as subangular breccia clasts (Wintzer, 2019; Wintzer et al., 2025). The majority of the scheelite observed is highly deformed as pseudomorphs filled with fine-grained recrystallized scheelite. Scheelite is overprinted, crosscut, or brecciated by stibnite, and less frequently is intergrown with stibnite (Wintzer, 2019; Wintzer et al., 2025). Fluid inclusions hosted in scheelite are all secondary in nature, occurring as wispy planes of inclusion decrepitates and rarely as small (<5-μm), two-phase, liquid-rich inclusions (Fig. 6F). Rare primary zones are marked by abundant dark inclusion decrepitates (Fig. 6E). The inclusions observed in scheelite were either decrepitated or too small to measure by microthermometry or Raman spectrometry. In relation to quartz, scheelite is observed overprinting Qi and Q1, intergrown with Q2 (Fig. 6E), and overprinted by Q3.

Stibnite textures and fluid inclusion petrography: Most fluid inclusion work is limited to transparent gangue minerals that are interpreted to be paragenetically related to ore formation. Near infrared (NIR) light microscopy enabled us to study fluid inclusions in optically opaque stibnite from the West End, Hangar Flat, and Yellow Pine deposits. The opacity of a mineral in visible light is related to its band gap energy. The band gap is the amount of energy required to excite an electron from the valence band (outermost electron orbital) to the conduction band (band of electron orbitals that excited electrons move to from the valence band). Semiconductors, such as sulfide minerals, absorb all radiation with energies greater than or equal to their band gap, making them opaque in visible light (Campbell et al., 1984). There are a few minerals (i.e., enargite, sphalerite, hematite) with which an infrared light microscope and camera setup may overcome this roadblock by illuminating minerals with band gap energies ≥1 eV (Campbell et al., 1984). Stibnite has a band gap energy up to 1.78 eV (Filip et al., 2013), which makes it possible to use low-wavelength NIR radiation to observe FIAs.

Stibnite in the Yellow Pine deposit occurs in fracture-filling veins and breccias that are a few mm to several cm wide (Fig. 7A-F). In quick plates and thick-section samples from the Yellow Pine, Hangar Flat, and West End deposits, stibnite cuts and overgrows disseminated pyrite and arsenopyrite and infills the center of quartz veins (Fig. 8; see Konyshev, 2020; Konyshev et al., 2025; Vikre et al., 2025a, b). In reflected light, stibnite is massive to acicular with strong bireflectance. In the near infrared, stibnite is transparent to opaque; grains are acicular or massive and exhibit pressure lamellae or sugary brecciation (Fig. 9). Transparent, unfractured, and undeformed stibnite contains five FIA types. These include primary FIAs of two-phase, liquid-rich inclusions with equant to negative crystal shapes, approximately 10 vol % vapor, and either elongate (type S1a; Fig. 6G) or stubby (type S1b; Fig. 6H) morphologies depending on the angle of observation relative to the crystallographic axis. Stibnite hosts multiple types of secondary FIAs including two-phase, vapor-rich (type S2; Fig. 6I) inclusions with >85 vol % vapor, single-phase liquid inclusions (type S3; Fig. 5J), and two-phase, liquid-rich inclusions (type S4; Fig. 6K).

Spots for LA-ICP-MS analysis were selected for the igneous quarts through Q3 quartz types defined by petrography and CL imagery in intrusion-hosted samples. The Qi and Q1 have higher Ti but lower Li, Al, and Sb than Q2 and Q3 (Fig. 10A, B; Marsh et al., 2023). Average concentrations in Qi and Q1 are 1 ppm Li, 51 ppm Al, 26 ppm Ti, 2 ppm As, and 3 ppm Sb. For Q2 and Q3, average concentrations are 119 ppm Li, 1,508 ppm Al, 8 ppm Ti, 12 ppm As, and 160 ppm Sb. There is not a significant difference in trace element composition of Qi and Q1. Q3 has elevated Sb relative to Q2 (Fig. 10B, C). There is a positive correlation between Al and Li in Q2 and Q3 (Fig. 10A).

Although the analytical spot size is much smaller (<2 μm), overall the EMPA analyses agree with the trace element contents measured in Qi through Q1 versus Q2 through Q3 by LA-ICP-MS (Marsh et al., 2023). Titanium values are elevated in Qi and Q1, and Al is elevated in Q2 and Q3. For Qi and Q1, the average concentrations are 167 and 880 ppm for Al and 21 and 12 ppm for Ti, respectively. For Q2 and Q3, the average concentrations are 2,330 and 1,331 ppm for Al. There was only one value each above the 11-ppm detection limit for Ti in Q2 and Q3 at 15 and 12 ppm, respectively.

The aqueous-carbonic fluid inclusions observed exhibit a range of behaviors on heating that indicate trapping occurred across a wide range of pressure-temperature (P-T) conditions (Table 1; Fig. 11; Microtherm_YP.csv in Marsh et al., 2023). To avoid complications with aqueous-carbonic inclusions affected by postentrapment modification, only FIAs with consistent phase ratios, showing minimal signs of necking or decrepitation were analyzed. Of the inclusions selected, those with homogenization data falling more than 30°C outside the average of the FIA were treated as showing evidence of stretching or leakage and were excluded from interpretation. Most of the CO₂-rich FIAs hosted in secondary planes of Q1 cut Qi. Most of these FIAs decrepitated during heating between 229° and 325°C (Table 1; Marsh et al., 2023). A few CO₂-rich inclusions homogenize to vapor between 336° and 387°C (n = 3) and have maximum salinities based on clathrate melting between 5.1 and 7.0 wt % NaCl equiv (n = 9) and CO₂ concentrations of 30 to 35 mol % (n = 9; Steele-MacInnis, 2018; Marsh et al., 2023). Other FIAs homogenize to the liquid phase between 281° and 363°C and have clathrate melting temperatures that yield salinities of 3.2 to 9.0 wt % NaCl equiv and CO₂ contents of 27 to 36 mol % (n = 6 for the temperature of total homogenization [Th_total], n = 13 for X_NaCl-X_CO2; Steele-MacInnis, 2018; Marsh et al., 2023). Several inclusions in the range of 363° to 387°C (n = 3) exceed the Th_total of their FIAs by >30°C and are considered to have stretched during heating; these data are not used to estimate trapping temperatures. A rare aqueous-carbonic FIA in a milky Q1 vein homogenized to the vapor phase at 303° to 308°C, with CO₂ vapor phase homogenization temperatures and clathrate melting temperatures indicative of CO₂ concentrations of 22 to 23 mol % and salinities of 4.5 to 5.0 wt % NaCl equiv (n = 2; Steele-MacInnis, 2018; Marsh et al., 2023).

Liquid-rich Q2 FIAs in microveinlets that cut Qi contain variable amounts of CO₂; some contain enough CO₂ to form clathrate on freezing, whereas others do not. The CO₂-bearing inclusions have ice melting temperatures (Tm_ice) of –3.7° to –3.4°C and clathrate melt temperatures (Tm_clath) of 8.4° to 8.8°C with homogenization to the liquid phase ranging from 183° to 212°C (Table 1; n = 13 for Tm_ice; n = 15 for Th_total; n = 13 for Tm_clath; Steele-MacInnis, 2018; Marsh et al., 2023). These clathrate melting temperatures fall outside the modeled H₂O-NaCl-CO₂ system of Steele-MacInnis (2018) and reflect low concentrations of CH₄, as discussed in later sections. To compensate for the effect of CH₄ on the clathrate melting temperature, an approximation of the composition of these Q2 inclusions was achieved by increasing the ice melting temperature entered in Steele-MacInnis (2018) by 0.2° to 0.3°C, within error of the value measured for each assemblage, yielding a composition of 2.5 to 2.9 wt % NaCl equiv and 3 mol % CO₂. At Hangar Flat, primary liquid-rich inclusions hosted in dark zones within Q2 homogenize between 189° and 204°C and have salinites of 5 wt % NaCl equiv, based on Tm_clath, and 2 mol % CO₂ (n = 10 Tm_ice, n = 3 Tm_clath, n = 9 Th_total; Steele-MacInnis, 2018; Marsh et al., 2023).

The Q2 cutting Qi hosts secondary liquid-rich FIAs that have salinities of 6.0 to 6.3 wt % NaCl and homogenize from 175° to 204°C (Table 1; n = 13). Well-preserved Q2 secondary liquid-rich inclusions hosted in milky Q1 veins also lack CO₂ and homogenize to liquid from 174° to 237°C. They have maximum salinities, based on melting temperatures of ice, of 5.0 to 6.7 wt % NaCl equiv (n = 24 for Tm_ice; n = 22 for Th_total; Steele-MacInnis et al., 2012; Marsh et al., 2023).

Primary fluid inclusions are only rarely identified in zoned Q2 due to prolific postentrapment deformation. Primary liquid-rich inclusions homogenized to liquid from 217.5° to 222°C, with salinities of 6.0 wt % NaCl equiv based on Tm_ice (Table 1; n = 4; Steele-MacInnis et al., 2012; Marsh et al., 2023). Secondary two-phase liquid-rich inclusions in zoned Q2 have Th_total of 200° to 227°C and salinities of 8.0 wt% NaCl equiv based on Tm_ice (Table 1; n = 5; Steele-MacInnis et al., 2012; Marsh et al., 2023). Abundant planes of inclusion decrepitates are identified in Q2 (Fig. 4D, H). Secondary liquid-rich FIAs measured within Q2 lack CO₂, homogenized across a wide temperature range between 180° and 233°C, and have salinities of 7 to 9 wt % NaCl equiv (n = 6; Steele-MacInnis et al., 2012; Marsh et al., 2023).

Primary fluid inclusions could not be identified in bright CL Q3, and the secondary, liquid-rich inclusions present are mostly too small for microthermometric measurement. A rare assemblage of these secondary inclusions, with consistent liquid-vapor proportions, homogenized between 156° and 161°C, with salinity of 1.9 wt % NaCl equiv (n =5; Steele-MacInnis et al., 2012; Marsh et al., 2023). All Q3 FIAs examined are liquid rich, showing no evidence of clathrate formation on freezing.

Microthermometry of the type S1a (n = 18) and S1b (n = 9) inclusions in the stibnite yielded salinities between 4.2 and 10.6 wt % equiv NaCl and Th_total of 150° to 189°C (Table 1; Marsh et al., 2023). For type S2 (n = 5) inclusions, salinities of 6.3 to 9.7 wt % NaCl equiv and inclusion decrepitation occurred between 144° and 195°C. The extended range in salinity compared to the paragenetically related Q3-hosted FIAs may be due to the influence of NIR radiation on the sample during the measurement, because the cycling technique recommend by Casanova et al. (2018) was not used. Crushing-stage experiments on stibnite found no evidence of high-density gas, nor was clathrate observed during cooling and heating experiments. In the type S3 single-phase inclusions, no phase changes were observed during cooling to –196°C, which suggests they may have leaked. Type S4 (n = 4) inclusions consist of dilute water with Th_total = 105°C.

Raman spectroscopy is a useful tool to help determine the composition of hydrothermal fluids trapped in fluid inclusions. Burke (2001) tables 1 and 2 highlight Raman active and inactive content common in fluid inclusions and their corresponding Raman shifts (cm^–1). Raman peaks at 1,284 and 1,387 cm^–1 in aqueous-carbonic fluid inclusions associated with early Q1 that cuts Qi confirm the presence of CO₂ and the absence of CH₄ in Q1 fluids (Burke, 2001, and references therein; Raman_gasratios_YP.csv in Marsh et al., 2023). Raman peaks at 1,284, 1,387, and 2,916 cm^–1 in vapor bubbles of secondary two-phase aqueous carbonic inclusions in dark CL Q2 contain 95.5 mol % CO₂ and about 4.4 mol % CH₄. Dissolved gases were not detected in the liquid phase. Vapor bubbles of primary aqueous-carbonic inclusions trapped in early Q2 also contain CO₂ and CH₄, but with less CH₄, ranging from 1.9 to 3.5% (n = 3). Low levels of CH₄, from 4.6 to 5.5% (n = 2), are also present in vapor bubbles of aqueous-carbonic inclusions in late Q2. These gases were not detected in vapor bubbles of liquid-rich inclusions in Q3.

Fluid inclusion extracts consist largely of H₂O, CO₂, N₂, and CH₄ (Table 2; Fig. 12A). Although the measured gas fractions cannot be used to calculate pressures (because some water is lost by adsorption onto vacuum inlet surfaces during analysis; Landis and Hofstra, 2012), the relative differences are consistent with a substantial decrease in gas content and pressure through the paragenetic sequence. For example, early quartz and calcite have the highest gas contents followed by quartz + pyrite, stibnite, and then late calcite with the lowest gas content (Table 2). This result is consistent with the decrease in trapping pressure evident from microthermometric analyses of FIAs in Qi through Q3 and stibnite (see “Discussion”). The ternary diagram (Fig. 12B) shows that extracts from earlier event minerals generally contain more N₂ and CH₄ than those from later event minerals.

Noble gas isotope analyses show that ⁴⁰Ar is more abundant than ⁴He, which is more abundant than ²⁰Ne (Fig. 12C). Although ³He/⁴He and ²⁰Ne/²²Ne ratios could generally be measured, with a few exceptions ³⁸Ar, ³⁶Ar, and ²¹Ne isotopes were below detection, which precluded measurement of ²⁰Ne/²¹Ne, ⁴⁰Ar/³⁶Ar, and ³⁸Ar/³⁶Ar ratios (Table 2). We, therefore, chose to display the ³He, ⁴He, and ²⁰Ne isotope results on the ²⁰Ne/⁴He versus R/R_A diagram of Prinzhofer (2013; Fig. 12D), which allows the proportion of air (air-saturated water), crust, and mantle (mid-ocean ridge basalt [MORB]) gas in the extracts to be distinguished. The measured ²⁰Ne/²²Ne ratios have a wide range from 10.03 to 14.35 (Table 2; Fig. 12E).

Our fluid inclusion study indicates that the composition of hydrothermal fluids evolved through the paragenetic sequence from early CO₂-H₂O-NaCl in Q1 related to gold deposition to H₂O-NaCl ± CO₂ ± CH₄ in Q2 related to W mineralization, to H₂O-NaCl in Q3 associated with Sb mineralization. The temperature of homogenization decreases through the paragenetic sequence from 358° to 156°C. These temperatures represent minimum trapping estimates of the fluid inclusions, as no immiscible pairs of FIAs were observed (Roedder, 1984).

The minimum pressure calculations from H₂O-NaCl-CO₂ FIAs in Qi and Q1 related to event 2 gold at 70 to 56 Ma of Gillerman et al. (2025), reflect a moderately deep crustal environment. Preserved secondary aqueous-carbonic FIAs that crosscut Qi and are paragenetically associated with auriferous pyrite and sericite (Fig. 4A) have Th_totals 281° to 358°C, contain 27 to 36 mol % CO₂, and have salinities of 3 to 9 wt % NaCl equivalent, bulk densities of 0.77 to 0.84 g/cm³, and calculated homogenization pressures (Ph_total) of 1.3 to 1.7 kbar (Table 1). Assuming a density of 2.65 g/cm³ for the overlying rock column, these minimum pressure estimates correspond to a minimum paleodepth of ~5 to 6 km under lithostatic pressure conditions.

Experimental studies of the brittle-ductile transition temperatures in different rock types show that the approximate transition temperature ccurs at 300°C in wet granite and at 320°C in wet quartz at intraplate stresses of 15 MPa and strain rates of 10 to 14 sec^–1, which is a typical P-T environment for many orogenic gold deposits (Simpson, 2001). Higher stresses of 50 MPa produce lower transition temperatures of 280° and 300°C, respectively (Simpson, 2001). For rocks that experienced equivalent strain rates, Stӧckert et al. (1999) determined that the brittle-ductile transition of quartz approximately coincides with the K-Ar closure temperature of biotite at 310° ± 30°C. This temperature range of 280° to 300°C has been used as the lower range of the brittle-ductile transition by Diamond and Tarantola (2015) to interpret cooling and decompression trends in quartz-hosted inclusions from orogenic gold deposits. For the Stibnite-Yellow Pine district, 310° ± 30°C is a reasonable range for the brittle-ductile transition in the quartz veins and overprinted phenocrysts, as the microthermometry data together with the texture of quartz and fluid inclusions show a shift from lithostatic to hydrostatic conditions. Many of the inclusions hosted in the Q1 veins show reequilibration textures that were produced through decompression and rupture as the internal pressure of the fluid inclusions exceeded the confining pressure of the host quartz (Roedder and Bodnar, 1980; Pecher, 1981; Bodnar et al., 1989; Sterner and Bodnar, 1989; Vityk and Bodnar, 1995; Diamond et al., 2010; Fig. 4A-C). Quartz of this type is cloudy with abundant reequilibrated fluid inclusions that were overpressured and decrepitated during fluctuations from high to low pressure. Widespread observation of these textures indicates that the formation of Q1 quartz preceded, or was coincident with, pressure changes within the system. Kontak and Tuba (2017) noted decrepitation textures formed during pressure cycling of early primary inclusions in orogenic gold deposits. The rare pristine CO₂-rich inclusions measured in Q1 are thought to represent the latest, coolest pulses of this mineralization stage as the system transitioned to the hydrostatic pressure domain—conditions during initial formation of the Q1 veins were likely hotter and higher pressure. These aqueous-carbonic inclusions with ThCO₂ that homogenize by vapor expansion have lower bulk densities of 0.58 g/cm³ at 22 mol % CO₂ and a minimum trapping pressure of 438 bar at Th_total of 303° to 308°C, corresponding to an average of 5 km paleodepth under hydrostatic conditions.

This minimum paleodepth range of 5 to 6 km estimated by fluid inclusion microthermometric data is within the depth estimates of 4 to 9 km for gold-bearing quartz veins in the Buffalo Hump district 70 km to the northwest (Lund et al., 1986). Fayon et al. (2017) modeled the emplacement and exhumation of the Idaho batholith based on cooling ages of minerals with different closure temperatures; they modeled the batholith to be between 7 and 10 km depth by 70 Ma and around 5 km depth by 43 Ma (see figure 7 of Fayon et al., 2017). These depth estimates are in good agreement with those calculated for secondary aqueous-carbonic inclusions trapped after emplacement of the batholith and deposition of early milky quartz veins in association with sericite and auriferous pyrite at circa 70 to 56 Ma (Gillerman et al., 2025; Table 1; Fig. 3) at the minimum depth of 5 to 6 km.

The minimum pressure calculations from H₂O-NaCl-CO₂ ± CH₄ FIAs in Q2 related to event 3 scheelite at 57 to 56 Ma of Gillerman et al. (2025) range between 1 and 6 km (Table 1). Naturally decrepitated primary FIAs occur within Q2 cores along growth zones and are overgrown by Q2 with primary low-CO₂, liquid-rich FIAs (Fig. 4D-F). The former suggest that a low-CO₂ liquid phase was also present at lithostatic conditions, while the latter FIAs represent the first fluids trapped after the system transitioned from lithostatic to hydrostatic conditions. These low-CO₂ aqueous inclusions contain minor amounts of CH₄, revealed by Raman, and homogenize to liquid from 189° to 204°C (n = 9). Fluid composition for this assemblage was calculated using equations of state of Bakker (1999) and Bowers and Helgesen (1983) using the program BULK (version 01/03) giving 0.01 to 0.02 X_CO2, and 0.00003 to 0.0001 X_CH4. The calculated compositions and densities are in good agreement with those calculated for the same Q2 inclusions using Steele-MacInnis (2018; 5 wt % NaCl equiv; 0.02 X_CO2) and suggest that the CH₄ fraction of the vapor bubble is negligible, allowing that the density of the inclusion is obtained from CO₂, NaCl, and H₂O. The Tm_ice, Tm_clath, volume fraction at room temperature, and Th_total for these inclusions in the H₂O-NaCl-CO₂ system (Duan et al., 1995; Anderko and Pitzer, 1993) have Ph_total of 72 to 376 bar and fluid composition of 2.2 mol % CO₂ and 4.7 to 5.1 wt % NaCl equiv (Table 1; Steele-MacInnis, 2018). These Ph_totals correspond to a minimum paleodepth range of 1 to 4 km at hydrostatic pressure conditions (Table 1), assuming a density of 1 g/cm³ for the overlying water column. Similar liquid-rich, two-phase inclusions with minor CO₂ and CH₄ cut Qi and homogenize between 183° and 212°C. With composition approximated at 2.5 to 2.9 wt % NaCl equiv and 3.2 to 3.3 mol % CO₂, the minimum trapping pressures range from 486 to 602 bar (Table 1; Marsh et al., 2023). Assuming hydrostatic pressure, the estimated paleodepth would range between 5 and 6 km. The 1- to 6-km depth range for event 3 (W) is in the range of the minimum depths estimated for event 2, with a lower average of 4 km. This broad range is due to the steep isochore calculated for these assemblages (solid green in Fig. 11E), such that 5°C equates to about 100 bar pressure difference.

Similar fluid inclusion reequilibration textures have been recognized in CL by Taylor et al. (2021) in samples from Grass Valley, California, and form when early FIAs undergo fluctuations between lithostatic and hydrostatic pressure to the point of decrepitation. Decrepitation textures observed in scheelite indicated a similar fluctuation between lithostatic and hydrostatic pressure as observed in the quartz. Most scheelite examined in this study is highly recrystallized, with rare coarse-grained crystals intergrown or being overprinted by stibnite, suggesting that scheelite was precipitated prior to or during the transition from lithostatic to hydrostatic pressure conditions.

Pressures and temperatures derived from the secondary H₂O-NaCl FIAs in Qi, Q1, Q2, and Q3—interpreted to be related to events 3 (W) and 5 (Sb) of Gillerman et al. (2025)—require temperature correction in order to inform any paleodepth discussion, as they were likely trapped in the single-phase field of the H₂O-NaCl system rather than at their minimum temperatures along the L + V curve. Without a temperature correction to estimate their formation conditions, they yield calculated minimum pressures indicative of near-surface conditions (Table 1). Secondary low- to moderate-salinity, liquid-rich inclusions hosted in dark CL quartz cutting Qi yield minimum pressures of 9 to 16 bar in the system H₂O-NaCl based on homogenization between 175° to 204°C (Table 1; Fig. 3). Secondary, liquid-rich inclusions hosted in Q1 veins homogenize from 174° to 237°C and yield salinities of 5 to 7 wt % NaCl equiv and minimum trapping pressures of 10 to 30 bar (Table 1).

Stibnite-hosted FIAs have Th_totals ranging from 150° to 189°C and minimum pressures of 5 to 12 bar, while secondary inclusions in Q3 have Th_totals and pressures of 156° to 161°C and 5 to 6 bar (Table 1). The euhedral texture of the Q3, scheelite, and stibnite crystals from the Stibnite-Yellow Pine district samples suggests formation by open-space filling at hydrostatic pressure. Paleodepths inferred from these Ph_totals in most cases yield unrealistically shallow depths of less than 100 m in the hydrostatic regime.

The lack of vapor-rich inclusions late in the paragenesis supports the need for a pressure correction into the single-phase field of the H₂O-NaCl system. The apparent uninterrupted growth of quartz between Q2 and Q3 (Figs. 5E, 10D) allows that they formed from a single evolving hydrothermal fluid, which may imply that any pressure corrections applied to Q3 event 5 (Sb) should not exceed the lowest pressures calculated for CO₂-bearing inclusions hosted in Q2. The irregular shapes and inconsistent liquid-vapor ratios of inclusions hosted in Q3 are also likely to reflect formation at temperatures <200°C (Bodnar et al., 1985). This would suggest deposition of Q3 and stibnite (event 5) occurred shortly after Q2 (W; event 3) at average paleodepths of 4 km. This is in good agreement with prior studies, which inferred ~5 km of paleodepth by 45 Ma (Fayon et al., 2017). A pressure correction to 400 bar for the H₂O-NaCl inclusions in Q2, Q3, and stibnite shifts the entrapment temperatures to 197° to 252°, 179° to 184°, and 171° to 211°C, respectively. Absent the geochronology, which suggests that event 3 (W, 57 Ma) and event 5 (Sb, 47 Ma) occur 10 m.y. apart as overprinting hydrothermal events (Gillerman et al., 2025), optical and CL microscopy and fluid inclusion temperature, pressure, and composition (T, P, and X) indicate this is a single evolving hydrothermal system.

The range of compositions, densities, and CO₂ contents measured in H₂O-NaCl-CO₂ fluid inclusions within Q1 (~20–40 mol % CO₂) is consistent with the fluid compositions that would be produced via phase separation of a ~7 to 10 mol % CO₂ parent fluid (Fig. 11C; Li et al., 2020). However, we observed no cotrapped, low-CO₂ (2–6 mol % CO₂) assemblage that would be expected from such as phase separation. We consider this CO₂-rich fluid inclusion population to represent a rare, preserved fluid from the end of event 1, considering the abundant evidence of decrepitation of the majority of the fluid inclusions trapped in the early veins and phenocrysts. We assume that they were trapped in the two-phase field of the H₂O-CO₂-NaCl system between ~1.4 and 1.8 kbar at lithostatic pressure and ~280° to 350°C. This interpretation restricts the true trapping temperatures of this fluid stage to conditions within the two-phase field, with Th_total for each inclusion occurring along a bubble point curve for that specific composition rather than at the solvus. Carbon dioxide concentration may have important influence on transport and deposition during the gold stage of mineralization in the Stibnite-Yellow Pine district. While low concentrations of CO₂ may act as a buffering agent to optimize the transport of Au with reduced sulfur (Phillips and Evans, 2004), progressive evolution of a 10 mol % CO₂ parent fluid by phase separation and periodic buildup, overpressure, and escape of a ~30 mol % CO₂ vapor could result in episodic disequilibrium that plays a role in metal deposition, as increased CO₂ can reduce the solubility of Au in a hydrothermal fluid (Kokh et al., 2017; Monecke et al., 2019).

Petrographic observations and microthermometric measurements indicate that CO₂ concentrations in the hydrothermal fluid declined from ~37 to 22 mol % CO₂ during the formation of Q1, and that subsequent fluids trapped in Q2 quartz immediately following the transition from lithostatic to hydrostatic pressures contained only 2 mol % CO₂. All later fluids, including those identified in later Q2, contained no measurable CO₂. As the hydrothermal events progressed, the chemistry of the fluids represented in the inclusions evolved. Textures observed in the quartz indicate that decompression, CO₂ content, and quartz strength influenced which fluid stages were preserved in the inclusion record as the hydrothermal system evolved. The pressure differential between the ~1.6-kbar aqueous-carbonic inclusions trapped in quartz phenocrysts at lithostatic pressure and the ~0.4 kbar determined for vein-hosted inclusions formed at hydrostatic pressure is substantial, amounting to a drop of >1.2 kbar. While the quartz phenocrysts may be protected from the dynamic conditions affecting the veins, quartz-hosted fluid inclusions in veins cannot withstand such a large pressure differential at higher temperatures (Roedder and Bodnar, 1980; Pecher, 1981; Bodnar et al., 1989; Sterner and Bodnar, 1989; Lacazette, 1990; Goldstein and Reynolds, 1994; Vityk and Bodnar, 1995; Diamond et al., 2010). We propose that the lobate dark CL bands in late Q2 (Fig. 4D, G) are the result of quartz etching/dissolution caused by the periodic supersaturation of CO₂ as high-pressure fluids escaped into the hydrostatic domain (Monecke et al., 2019; Li et al., 2020). When opened to the hydrostatic environment, the CO₂ vapor phase likely escapes from the system, much like the immiscible vapor phase in a porphyry deposit (fig. 16 of Landtwing et al., 2010). Similar loss of CO₂ in a decompression setting is mentioned by Hagemann and Lüders (2003) at the Wiluna Au-Sb deposit. Release of the CO₂-rich vapor resulted in minor etching of the Q2 and formation of carbonate veinlets, leaving behind a dominantly low-salinity, low-CO₂ liquid.

Trace element chemistry helps distinguish the quartz generations in the Stibnite-Yellow Pine district. Titanium concentrations measured by LA-ICP-MS and EMPA, along with calculated or assumed pressures, were applied to Ti-in-quartz (TitaniQ) geothermometers. For Qi overprinted by Q1, using depths of emplacement of the Atlanta lobe estimated from Fayon et al. (2017) of between 7 and 12 km (1,800–3,000 bar), temperatures calculated ranged from 347° to 645°C using Huang and Audétat (2012) and 271° to 506°C using Thomas et al. (2010; assuming an activity of 1 for TiO₂ based on the presence of igneous rutile). This temperature is low for Qi and is likely due to subsequent hydrothermal alteration as evidenced by the dark gray CL response that overprints original Qi growth zonation. Temperatures calculated for Q1 using Huang and Audétat (2012), with a pressure of 500 bar based on microthermometry (Table 1), are ~50°C greater than or equal to Th_totals for fluid inclusion in Q1. For Q2 and Q3, using Huang and Audétat (2012) and a pressure of 150 to 600 bar, calculated temperatures are 150°C greater than Th_totals on fluid inclusions in Q2 and Q3, which is likely because the later hydrothermal quartz was not in equilibrium with any titanium phase, producing elevated temperatures in error (Haung and Audétat, 2012). The elevated Al, Li, Sb, and As in Q2 and Q3 is a common signature for hydrothermal quartz (Rusk et al., 2008; Jourdan et al., 2009; Rusk, 2012; Tanner et al., 2013; Breiter et al. 2013, 2020; Rottier and Casanova, 2020; Gao et al., 2022). The noted increase in Sb content in Q3 supports the paragenesis of stibnite and Q3 quartz (Fig. 10C). Based on discrimination plots by Rusk (2012), the trace element signature of Qi resembles that of high-temperature porphyry quartz, as expected for igneous phenocrysts and crosscutting microfractures, with elevated Ti and relatively low Al. Q1 sits mostly in the orogenic Au field, with lower Ti and slightly higher Al than Qi. The trace element signatures of Q2 and Q3 follow the trend of orogenic quartz of Rusk (2012), with high concentrations of Al and low Ti contents (Fig. 13A). The compilation of trace element data in quartz from hydrothermal and magmatic quartz by Rottier and Casanova (2020) shows a similar trend in lower Ti and increasing Al from lower to higher temperature environments, as evident in our data. However, their compiled data does not follow the trend of Rusk (2012) in distinguishing deposit types by Ti versus Al trends (see fig. 11 of Rottier and Casanova, 2020). Figure 13B, an Sb/Ti versus Al/Ti correlation diagram of data from this study overlaying compiled data by Rottier and Casanova (2020), shows Qi and Q1 values in and above the fields for early veins in porphyry systems associated with potassic alteration; Q2 and Q3 trend with and slightly above the late, low-temperature vein quartz associated with base metal mineralization in porphyry systems and quartz from porphyry related epithermal deposits.

Active (e.g., H₂O, CO₂, CH₄, N₂) and noble gas abundance and noble gas isotopic compositions are summarized in Table 2 and displayed on Figure 12. Fluid inclusion extracts (Ne and He) have noble gas isotope compositions that plot in three groups (Fig. 12D). The Qi and early calcite veins (Fig. 3) form a cluster with R/R_A values near 0.2 that is reasonable for volatiles derived from S-type, crustally contaminated, Late Cretaceous intrusions (Gaschnig et al., 2010, 2011; Fayon et al., 2017) and their contact metamorphic aureoles. The bulk analysis results for Q1 and auriferous pyrite plot along the air-crust mixing line, with volatiles likely derived from metasedimentary rock and air-saturated meteoric groundwater. Volatiles in Q2, Q3, and stibnite have ²⁰Ne/⁴He and R/R_A values that plot in a trend between the air-crust and air-mantle mixing lines (Fig. 12D), showing a shift from a crustal signature to an overprinting mantle signature.

Late calcite extends from air toward the mantle corner, with volatiles derived from less crustally contaminated early Paleocene to Eocene intrusions and air-saturated meteoric groundwater. It should be noted that no pure mineral separates of Q2 could be produced due to overprinting by Q3 and stibnite. Consequently, the composition of Q2 volatiles associated with scheelite could not be distinguished.

Although the ²⁰Ne/²²Ne ratios do not correlate with R/R_A, they provide additional evidence of mixing between air-saturated meteoric groundwater with a ²⁰Ne/²²Ne ratio of 9.80 and mantle-derived neon with much higher ²⁰Ne/²²Ne ratios (Fig. 12E). For example, the extract from stibnite with the highest ²⁰Ne/²²Ne ratio rivals those from MORBs and ocean island basalts (OIBs; Hilton and Porcelli, 2003). The absence of extracts with ²⁰Ne/²²Ne ratios <9.80 indicates that there was little or no nucleogenic production of ²²Ne from ¹⁹F, which suggests that the U, Th, and F contents of minerals in the Stibnite metasedimentary rocks, Idaho batholith, and Challis intrusions were relatively low.

To evaluate how the major volatile, CO₂, relates to helium isotopes, CO₂/³He ratios were plotted against R/R_A ratios (Fig. 12F). All the CO₂/³He ratios are much greater than air and plot in a triangular field. The left side of the triangle lies between a mantle end member with low CO₂/³He ratios and a crustal end member with high CO₂/³He ratios. The right side extends from a postulated magmatic end of the triangle to very high CO₂/³He ratios, which may indicate decarbonation of marble in the metasedimentary rocks.

To evaluate how the isotopic compositions of carbonate minerals relate to those of helium, CO₂/³He ratios were plotted against δ¹³C values (Fig. 12G). Although only four samples were analyzed by both methods, pre-ore calcite and calcite associated with Q1 plot in an intermediate position between limestone and organic sediment, whereas late dolomite is shifted toward mantle (MORB or subcontinental lithospheric mantle [SCLM]). This result suggests that most of the variation in the δ¹³C values of hydrothermal carbonates from the Stibnite-Yellow Pine district samples, from –4.5 to –10.4‰ (median –8.4‰; Lewis, 1984), is controlled by the proportion of carbon derived from marine limestone and organic sediment in metasedimentary rocks. The highest δ¹³C value approaches the median δ¹³C composition of marble in the Stibnite metasedimentary rocks of −2.9‰ (Lewis, 1984). The lowest values may reflect oxidation of organic carbon or methane derived from carbonaceous sediments.

As discussed in Landis and Hofstra (2012), it is important to realize that extracts are a mixture of the primary and secondary populations of fluid inclusions present in each mineral separate. Furthermore, mineral separates consisting of two or more minerals, or multiple phases of the same mineral, will yield more complex mixtures of different fluid compositions and ages. Thus, it is important to interpret the results in the context of the paragenetic sequence and fluid inclusion types in the study area. It is also important to consider the results in the context of the geologic framework. Any result that differs from air-saturated meteoric groundwater contributes some information about fluids from deeper levels in the system. The noble gas data reveals three signatures, from crustal to mantle through a mixture of the two. Pre-ore Qi has R/R_A values that reflect the composition of the crustal melt-derived Atlanta lobe of the Idaho batholith. The Au (Q1; event 2) indicates fluids derived from metasedimentary rocks. The W (Q2; event 3) and Sb (Q3; event 5) contain He and Ne that extend between crustal and mantle sources. The late carbonate has a mantle source (Figs. 3, 12).

The sequence of mineralization and character of hydrothermal fluids in the Stibnite-Yellow Pine district have many features in common with other global Au-W and/or Sb deposits and occurrences (Table 3). There is a general trend of H₂O-NaCl-CO₂ ± CH₄ to H₂O-NaCl, low- to moderate-salinity fluids that decrease in temperature from Au to W to Sb events, though not all the studies reviewed contain Au, W, and Sb, nor do they have identical P-T-X characteristics. A consistent observation is a decrease in temperature coincident with Sb deposition. At temperatures of 150° to 350°C, at acidic to neutral pH Sb₂S₂(OH)^o and Sb(OH)₃ are dominant species for Sb transport; the solubility of Sb decreases with temperature (Krupp, 1988; Spycher and Reed, 1989; Munoz et al., 1992; Zotov et al., 2003).

The late-stage temperatures and shallow depths recorded for the Stibnite-Yellow Pine samples are typical of the environment that hosts many epithermal and epizonal orogenic deposits. Late-stage drusy quartz is observed in fault-related veins from the Denggezhuang gold deposit in Jiaodong, China (see fig. 11C, D from Li et al., 2020). Stibnite is a common mineral in epizonal orogenic deposits with economic amounts of gold, which typically form at 3- to 6-km depths (i.e., Donlin Creek, Goldfarb et al., 2004). From these characteristics, it is possible that the Stibnite-Yellow Pine district deposits are similar to many orogenic gold deposits and district (Groves et al., 1998; Bierlein and Crowe, 2000; Ridley and Diamond, 2000; Mernagh, 2001; Goldfarb et al., 2005). But the age constraints of Gillerman et al. (2025) suggest that the fluid history of the district is protracted and complicated, with metals transported into the same trap areas five times over 40 million years. The long period of ore formation with multiple hydrothermal events may also be a common factor with other Au-W-Sb deposits. Ores in the Vendée antimony district, France, formed in two stages—Variscan (~370–299 Ma) east-west tension gashes and shear zones developed during sinistral movement and followed by Autunian (299–282 Ma) north-south veins originating during dextral north-south movement (Bailly et al., 2000). The Wiluna deposit, Australia, has a reported age of 2623 ± 12 Ma (muscovite ⁴⁰Ar/³⁹Ar) that is considered to be the formation age and is consistent with the regional deposit ages, yet there are three younger ages also reported (⁴⁰Ar/³⁹Ar apparent age of 2600 Ma; muscovite ⁴⁰Ar/³⁹Ar plateau age of 2565 ± 12 Ma; muscovite ⁴⁰Ar/³⁹Ar plateau age of 2,563 ± 12 Ma). These outlying three ages are proposed to either be reset or represent a later hydrothermal event (Kent and Hagemenn, 1996). The Pontgibaud deposit, France, formed in three stages: Fe-As, Zn-Pb-Cu, and Sb with reported ages from 313 ± 5 to 297 ± 9 Ma (Marcoux and Picot, 1985; Marcoux et al., 1988; Munoz et al., 1992). Bril et al. (1991) reports K/Ar data for associated alteration minerals from various Au-W-Sb-As/Pb-Ag veins in the Massif Central of France, with ages spanning 100 m.y.

Geochronology indicates that gold, tungsten, and antimony formed in successive hydrothermal events at 70 to 61, 57 to 56, and 47 to 46 Ma (Gillerman et al., 2025). Hydrothermal event 2 (70–56 Ma) resulted in the deposition of disseminated gold-bearing arsenian pyrite and arsenopyrite associated with fine-grained muscovite-K-feldspar-quartz (Q1) alteration and formed from aqueous-carbonic fluids with crustal helium derived from metasedimentary country rocks. These features resemble those of orogenic gold deposits hosted in granitic intrusions and metasedimentary rocks elsewhere in the world. Hydrothermal event 3 at ca. 57 Ma resulted in the deposition of scheelite with associated minor arsenian pyrite and arsenopyrite, K-feldspar, and quartz (Q2). These were deposited during a transition from lithostatic to hydrostatic pressure. Although uncertain, we surmise that the tungsten was sourced either from rutile breakdown (e.g., Cave et al., 2017) during latest Paleocene metamorphism of the country rocks (Wintzer, 2019) or from magmatic fluids derived from concealed Paleocene intrusions similar to those in the Bitterroot lobe of the Idaho batholith (Gillerman et al., 2025). The scheelite (scheelite 3 of Wintzer, 2019, figure 3.3H) related to stibnite dated at 47 Ma suggests the antimony event is a younger event depositing epithermal stibnite ± cinnabar mineralization associated with euhedral quartz (Q3) and pink calcite that contain helium and neon potentially derived from Challis magmatism at shallow depths. If Q3 formed as an overgrowth after Q2 rather than continuous growth, this would lend textural support for the formation of Q3 and stibnite during a separate, later event. This interpretation would place the formation of Q3 and stibnite at a relatively shallow depth of less than 2 km following a period of rapid exhumation.

Without the geochronologic constraints, our fluid inclusion data, quartz textures, and trace element chemistry could be interpreted as a single evolving hydrothermal system. Quartz textures show apparent continuous growth between Q2 (W) and Q3 + stibnite (Figs. 5E, 10D); liquid-vapor ratios and inclusion morphologies are irregular in Q3, and inclusion P-T-X measurements reflect a continuous trend of fluid cooling, decompression, and degassing across all phases studied. Textural evidence of scheelite recrystallization along with enrichment and depletion trends in the trace element chemistry between scheelite dated at 57 versus 47 Ma (Wintzer, 2019) may suggest remobilization or fractionation due to alteration, perhaps influenced by the onset of Challis magmatism (50–43 Ma). The influence of the Challis magmatism is observed in the Fayon et al. (2017) U/Th-helium study, where they note a resetting of the helium in zircon from the Yellow Pine area. Samples from lower elevations cooled at 40 Ma, but others from higher elevation cooled by 60 Ma. They attribute this cooling difference to the lower elevation samples being reset by Challis magmatism. Additionally, mosaic, flamboyant, and chalcedonic quartz textures associated with mineralization in epithermal deposits are largely absent (except at West End as described by Konyshev et al., 2025), and dendritic textures known for late-stage ore minerals (i.e., Gissler et al., 2024) were not observed in the samples examined from Yellow Pine and Hanger Flat. While bulk isotopic signatures for drusy quartz + stibnite contain mantle He suggestive of a magmatic source, similar RA values have been measured in fluids sampled from Poncha Hot Springs, Colorado, some 320 km from any coeval magmatic activity, indicating that deep fluid circulation along structures can impart a similar signature (Easley et al., 2011). These observations permit that the Au-W-Sb mineralization may have formed within a compressed timeline from an evolving fluid, with additional Au-(Ag) introduced by later epithermal fluids that partially recrystallized scheelite and stibnite, yielding the Eocene scheelite ages.

1. Hydrothermal fluids evolved from H₂O-CO₂-NaCl to H₂O-CO₂-NaCl ± CH₄ to H₂O-NaCl ± CO₂-CH₄, moderate to low salinity, moderate to high density, and a decrease in temperature through the Au, W, and Sb episodes of mineralization.

2. Trace elements in quartz ranged from higher temperature, Ti-bearing quartz in Qi and Q1 to a lower temperature Li- and Al-bearing quartz in Q2 and Q3. The Sb enrichment in CL bright Q3 confirms that it is paragenetically related to stibnite.

3. Active and noble gas isotope results suggest that the Au event formed from fluids derived from metasedimentary rocks. Paragenetically later minerals associated with the W and Sb events contain He and Ne that extend between crustal and mantle sources; late carbonate has a mantle source.

4. The Stibnite-Yellow Pine district is similar to other Au-W-Sb occurrences around the world in fluid chemistry, evidence of pressure fluctuation, stibnite late in the paragenesis, a decrease in temperature as a means of Sb deposition, and potential prolonged timelines.

5. During Au-As mineralization, gradual cooling from 400° to 300°C accompanied decompression of approximately ~1.5 kbar across the brittle-ductile transition at temperatures of ~300°C (Fig. 11E; Table 1). Subsequent fluids associated with scheelite and stibnite precipitation were lower temperature, from 237° down to 166°C. Overall, the depth of formation may have decreased from 5 to 6 km during early cooling as fluids transitioned from the lithostatic to the hydrostatic conditions to as shallow as ~4 km by the Eocene, consistent with regional uplift and erosion.

6. Geochronologic studies suggest that the Stibnite-Yellow Pine district experienced a series of superimposed hydrothermal events over 37 m.y. These include events from (1) a deep magmatic hydrothermal environment associated with intrusion Idaho batholith with minor molybdenum mineralization, (2) a deep metamorphic hydrothermal environment with significant Au-As mineralization, (3) a metamorphic or magmatic hydrothermal environment with critical W mineralization, and (4 and 5) an environment with magmatic volatiles producing epithermal Au-Ag at West End, critical Sb at Yellow Pine and Hangar Flat, and the historic Hg mineralization at Fern and Hermes mines. Haowever, absent the geochronology, the texture and chemistry of fluid inclusions and minerals could be interpreted as a single evolving hydrothermal system, as has frequently been done in other mining districts with similar mineral assemblages.

A special thanks to Midas Gold (now Perpetua Resources) for access to the Stibnite-Yellow Pine district properties and for samples. Thank you to Leslie Logan, Sarah Balah, and Lexi Wallenberg for field and laboratory assistance and to Jim Reynolds, Richard Goldfarb, and Matt Steele-MacInnis for critical discussion about this manuscript. We are grateful to Matt Steele-MacInnis, Steve Box, and George Case for thorough review of this manuscript. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Erin Marsh is currently a Supervisory Research Geologist with the U.S. Geological Survey. She holds an MS in geology from the University of Colorado Boulder and has over 25 years of experience in economic geology and geochemistry. She has conducted field and laboratory projects on a variety of deposit types from reduced intrusion related and orogenic gold to sedimentary exhalative Mississippi Valley type, iron oxide appetite–iron oxide copper-gold, and W-skarn. Much of her current work is focused on surficial systems processes including the formation of Ni-Co laterites and provenance of placer gold deposits.