Source-to-Sink Evolution of Volcano-Sedimentary Li-B Deposits at Rhyolite Ridge, Southwestern Nevada Open Access

Efforts to mitigate climate change through decarbonization and the development of low-carbon-emitting energy technologies has led to a global surge in the demand for Li, a critical component of rechargeable Li-ion batteries used in electric vehicles and renewable-energy storage systems (e.g., Bibienne et al., 2020; Tabelin et al., 2021; U.S. Geological Survey, 2023). Although most global Li production has been derived from hard rock (pegmatites and granites) and brine sources (e.g., Kesler et al., 2012; Munk et al., 2016; Gourcerol et al., 2019; Bowell et al., 2020), claystone-hosted Li deposits associated with volcano-sedimentary systems have become an emerging resource target. Recent resource assessments indicate that similar Li claystone prospects at McDermitt caldera along the Nevada-Oregon border (USA; Fig. 1A) now represent one of the largest known Li resources on Earth (e.g., Benson et al., 2023).

The formation of volcano-sedimentary Li deposits depends on three main factors: (1) a fertile source of Li, typically rhyolitic volcanic rocks, (2) meteoric and/or hydrothermal fluids to extract and transport Li from source rocks, and (3) a hydrologically closed basin where Li can concentrate through evaporation of lake waters to remain dissolved in brines or bound in diagenetic clay minerals (e.g., Munk et al., 2016; Castor and Henry, 2020). Caldera settings like McDermitt provide ideal conditions for generating significant volcano-sedimentary Li deposits due to their large (>800 km²) closed basins, voluminous intracaldera rhyolitic tuff, and postcaldera sedimentation, effusive magmatism, and hydrothermal circulation (Benson et al., 2017a, b; Henry et al., 2017; Castor and Henry, 2020; Benson et al., 2023). Volcano-sedimentary Li deposits also occur in settings unrelated to caldera systems, yet they are far less studied and understood in terms of Li genesis. Examples include the sizeable Jadar deposit in Serbia (Putzolu et al., in press), Lake Mead in southern Nevada (Brenner-Tourtelot and Glanzman, 1978; Landsem et al., 2023), and more recently discovered prospects in Sonora, Mexico (Verley and Vidal, 2013), the Mojave Desert, California (Gagnon et al., in press), and near Clayton Valley in southwestern Nevada (Reynolds and Chafetz, 2020; this study). Compared with volcanic caldera (e.g., Henry et al., 2017) or maar crater systems (Thompson et al., in press) that develop over relatively short spatial and temporal scales, volcano-sedimentary Li systems may develop over vastly larger and variable scales. Specific investigations into the interplay between tectonic, magmatic, and hydrologic processes, in both space and time, are therefore essential to understanding source-to-sink Li systems in noncaldera or tectonic sedimentary basins.

Clayton Valley is one of the richest Li districts in the United States and hosts numerous exploratory volcano-sedimentary Li prospects and the Silver Peak mine, which produces Li from brines and is currently the only active commercial-scale Li mining operation in the nation (U.S. Geological Survey, 2023) (Fig. 1B). Only 20 km west in the northern Silver Peak Range, economically important volcano-sedimentary deposits containing both Li and B at Rhyolite Ridge are currently being evaluated by Ioneer Ltd. and poised for development as early as 2027 (Ioneer, 2024). Lithium and B mineralization occurs as stratiform claystone deposits within the upper Miocene to lower Pliocene Cave Spring formation. The primary ore section is ~25 to 40 m thick with a lower zone composed of marl with abundant searlesite [NaBSi₂O₅(OH)₂], and an upper zone in which Li and B are bound in smectite or mixed illite-smectite claystone (Reynolds and Chafetz, 2020). The total mineral resource in the South Basin prospect (i.e., Cave Spring basin) was recently estimated to be 351 Mt of ore at average grades of 1,739 ppm Li and 6,379 ppm B (middle and lower members of the Cave Spring formation), which equate to ~3.25 Mt of lithium carbonate equivalent (LCE) and ~12.8 Mt of boric acid equivalent planned for extraction over the 26-year initial mine life (Ioneer, 2024).

Despite the identification of substantial proven volcano-sedimentary Li-B resources at Rhyolite Ridge, exploration and development efforts in this area and throughout the wider Clayton Valley region face challenges due to fundamental uncertainties about the mineralization system. First, the original source(s) of Li and B in the Cave Spring formation have not been investigated and remain unknown. Second, the tectono-magmatic setting of the Li-B deposits and their possible relationship to a proposed caldera is unclear. Previous studies have characterized the underlying Silver Peak volcanic center as a collapsed caldera filled with a thick sequence of slightly alkaline tuffs and lavas dated at ca. 6 to 4 Ma (Albers and Stewart, 1972; Robinson, 1972; Robinson et al., 1976); however, this interpretation lacks substantiating evidence. Importantly, concentrations and possible enrichment in Li, B, and other critical trace elements in these volcanic rocks have not been evaluated in prior studies. Furthermore, the relative influences of magmatic-hydrothermal fluids versus closed hydrologic system diagenesis on the transport and enrichment process remain unexplored. Addressing these uncertainties is essential for accurately assessing the resource potential and guiding future exploration strategies in the Clayton Valley region, as well as contributing to the broader understanding of Li and B deposits in similar geologic settings.

Here, we integrate new insights from our recently published, 1:24,000-scale geologic map (Ogilvie et al., 2023), volcanic petrography, major and trace element geochemistry, and U-Pb geochronology to establish the space-time pattern of magmatism and sedimentation and to evaluate the Li-B enrichment in lacustrine strata of the Cave Spring formation from source to sink. Field relationships, along with a compilation of new and published geochemical and geochronologic data, support a newly proposed regional correlation of the Li-enriched Rhyolite Ridge tuff that challenges the existence of its hypothetical source caldera in the Silver Peak Range (Albers and Stewart, 1972; Robinson, 1972). Collectively, these findings serve as a critical step in establishing linkages between local and regional magmatism, basin formation, and volcano-sedimentary Li-B enrichment in the Rhyolite Ridge area and across the greater Clayton Valley region.

The Silver Peak Range and neighboring Clayton Valley are situated within the southern Walker Lane belt (Fig. 1A), a zone of active dextral transtension in western Nevada that accommodates ~20% (10–11 mm yr^–1) of cumulative Pacific-North America plate motion (Dokka and Travis, 1990; Bennett et al., 2003; Faulds and Henry, 2008; Lifton et al., 2013). This region lies immediately south of the Mina deflection (Fig. 1B), a complex system of ENE-striking left-lateral faults that represents a major right step in the Walker Lane belt and transfers strain from the NW-striking dextral Fish Lake Valley-Death Valley fault zone to various NW-striking dextral faults of the central Walker Lane belt (Faulds and Henry, 2008). Active strike-slip and high-angle normal faulting are well documented from geologic mapping, geodetic, and paleoseismic studies (Reheis, 1991; Reheis and Sawyer, 1997; Angster et al., 2019) and was vividly expressed during the M6.5 Monte Cristo Range earthquake on May 15, 2020, which occurred ~28 km north of Rhyolite Ridge and represents the largest recorded seismic event in Nevada in the past 70 years (Hammond et al., 2021; Koehler et al., 2021).

The greater Clayton Valley region lies along the western fringe of cratonic North America (Fig. 1A) that endured a prolonged deformation history involving widespread late Paleozoic to Mesozoic terrane accretion, crustal shortening and thickening, and surface uplift, followed by widespread magmatism, extensional collapse, and dismemberment since early Miocene time (Albers and Stewart, 1972; Dickinson, 2006; Colgan and Henry, 2009). The first major extensional episode occurred during middle to late Miocene time (ca. 15–8 Ma) and involved top-to-the-northwest slip on the Silver Peak-Lone Mountain low-angle detachment fault system (Fig. 1B). Significant northwest-southeast extension generated the supradetachment Esmeralda basin across the region and much of Esmeralda County (Albers and Stewart, 1972; Stewart and Diamond, 1990; Oldow et al., 1994; Diamond and Ingersoll, 2002).

Low-temperature thermochronology indicates that basin formation was contemporaneous with ca. 12 to 7 Ma cooling and erosional exhumation of the White Mountains (Stockli et al., 2003) and the Lone Mountain, Weepah Hills, and Mineral Ridge metamorphic core complexes (Oldow et al., 1994; Burrus, 2013) (Fig. 1B). A pronounced and regionally extensive angular unconformity between the Silver Peak (or Esmeralda) formation and overlying ca. 6 to 4 Ma rocks of the Silver Peak volcanic center (Robinson, 1972) suggests that detachment slip ceased by ca. 8 to 6 Ma (Stewart and Diamond, 1990; Diamond and Ingersoll, 2002), and was followed by a shift to dominantly local lacustrine sedimentation in more isolated basins (Reheis and Sawyer, 1997; Oldow et al., 2009; Reynolds and Chafetz, 2020; Ogilvie, 2023). The modern structural grain is controlled by N- and NE-striking normal faults that crosscut older structures and show demonstrable late Quaternary slip and geodetic strain, especially on the southeast margins of northern Fish Lake, and the Big Smoky and Clayton valleys (Reheis and Sawyer, 1997; Lifton et al., 2013) (Fig. 1B).

Here we provide a concise overview of the tectonostratigraphy of the northern Silver Peak Range (Figs. 2, 3). For detailed descriptions of all lithologic units in the study area, refer to the 1:24,000-scale geologic map sheet and report by Ogilvie et al. (2023).

The oldest rocks in the study area are Neoproterozoic tectonites in the lower plate of the Silver Peak detachment fault that represent an exhumed metamorphic core complex along Mineral Ridge (Figs. 1B, 2). These lower-plate rocks belong to the regionally recognized Precambrian Wyman Formation and are composed of mostly phyllitic metaclastic and metacarbonate sedimentary rocks (Albers and Stewart, 1972; Diamond and Ingersoll, 2002). The structurally highest portions of the Wyman Formation atop Mineral Ridge are strongly metamorphosed to calc-silicate schist, marble, and mylonitic gneiss near the trace of the Silver Peak detachment fault (Fig. 1B).

Much of the Silver Peak Range consists of highly attenuated Paleozoic strata and unconformably overlying Neogene volcano-sedimentary rocks in the upper plate of the detachment fault system (Fig. 2). Paleozoic strata consist of weakly metamorphosed marine shale and carbonates of the Cambrian Campito, Poleta, Harkless, Mule Spring, and Emigrant Formations and the early Ordovician Palmetto Formation (Robinson et al., 1976; Diamond and Ingersoll, 2002), which represent the base of the former passive margin sequence of southwestern Laurentia. Because all Cambrian-Ordovician units have experienced significant vertical attenuation due to structural omission along low-angle normal faults, the contacts between them are referred to as “attenuation faults” (Diamond and Ingersoll, 2002).

Both structural attenuation of Neoproterozoic-Ordovician strata and antiformal doming of the Mineral Ridge core complex (Fig. 1B) were directly associated with early Eocene emplacement of the peraluminous Mineral Ridge pluton, which is present in the shallow subsurface at Mineral Ridge and likely throughout the eastern Silver Peak Range (Diamond and Ingersoll, 2002).

Paleozoic units are unconformably overlain by the lower Miocene Icehouse Canyon assemblage (Oldow et al., 2009), which consists of regionally extensive ignimbrites and local andesitic volcanic centers associated with the Ancestral Cascades volcanic arc and dated in the study area at ca. 23 to 21 Ma (K-Ar; Robinson et al., 1968). The Icehouse Canyon assemblage is conformably overlain by fluvio-lacustrine clastic strata and volcanic rocks of late Tertiary age called the Coyote Hole group. The group was formerly considered a part of the broader Esmeralda formation (e.g., Albers and Stewart, 1972; Stewart and Diamond, 1990; Diamond and Ingersoll, 2002), which is now largely abandoned and subdivided (Oldow et al., 2009). The Coyote Hole group is composed of four members: the Silver Peak formation, Rhyolite Ridge tuff, Argentite Canyon formation, and “Cave Springs” [sic] formation (Oldow et al., 2009).

The Silver Peak formation (informal; after Oldow et al., 2009) is composed of clastic sedimentary rocks of middle to upper Miocene age. It is interpreted as a supradetachment basin-fill sequence related to contemporaneous slip on the Silver Peak-Lone Mountain detachment fault (Stewart and Diamond, 1990; Diamond and Ingersoll, 2002). Several widespread and correlative tuffs in the Silver Peak formation have been dated at ca. 13 to 11 Ma in the Volcanic Hills, Weepah Hills, and southern end of Fish Lake Valley using K-Ar (Robinson et al., 1968) and zircon (U-Th)/He methods (Burrus, 2013). These dates and stratigraphic relationships suggest that deposition of the Silver Peak formation occurred in a regional and likely contiguous basin, which was initiated by ca. 16 to 15 Ma and continued until ca. 9 to 8 Ma (Stewart and Diamond, 1990; Diamond and Ingersoll, 2002). A widely recognized and regionally extensive angular unconformity separates the Silver Peak formation from all overlying volcano-sedimentary strata and suggests significant tectonic and/or erosional denudation during latest Miocene time (Albers and Stewart, 1972; Robinson et al., 1976; Stewart and Diamond, 1990; Diamond and Ingersoll, 2002; Oldow et al., 2009; Ogilvie et al., 2023).

Unconformably overlying the Silver Peak formation is a variably thick sequence of volcanic rocks originally referred to as the “Silver Peak volcanic center” (Fig. 1B; Robinson, 1972) and more recently as the Rhyolite Ridge and Argentite Canyon formations (nomenclature introduced by Oldow et al., 2009) (Figs. 3, 4). The Rhyolite Ridge formation consists of a composite sequence of nonwelded, lithic-rich rhyolite tuffs and lavas exposed at Rhyolite Ridge and throughout the Silver Peak Range (Fig. 4A-E). Legacy K-Ar and high-precision ⁴⁰Ar/³⁹Ar geochronology constrain the age of this unit to ~6.1 to 6.0 Ma (Robinson et al., 1968; Oldow et al., 2009). The Rhyolite Ridge formation is overlain by the Argentite Canyon formation, which consists of porphyritic latite or trachytic lavas and a distinctive welded ash flow tuff dated by ⁴⁰Ar/³⁹Ar and U-Pb geochronology at 5.87 to 5.76 Ma (Figs. 3, 4F-H) (Oldow et al., 2009; this study). The Argentite Canyon formation erupted locally in the southwestern part of the map area, as shown by dike-sill networks that clearly feed the extrusive latite lavas capping the surrounding ridges of Wild Horse Canyon (Figs. 2, 4H). The source of the tuffs in the Rhyolite Ridge formation is not as clear. Both the Rhyolite Ridge and the Argentite Canyon formations have been interpreted as products of a hypothetical and concealed “Silver Peak caldera” just south of the Rhyolite Ridge study area in the central Silver Peak Range (Robinson, 1972; Stewart et al., 1974; Robinson et al., 1976). Additional field, petrographic, and geochemical data for these units are presented in subsequent sections and in Ogilvie et al. (2023).

The Rhyolite Ridge and Argentite Canyon formations are overlain by Li-B–enriched sedimentary strata of the upper Miocene to lower Pliocene Cave Spring formation. The Cave Spring formation represents the uppermost part of the Coyote Hole group and is described in more detail in the following section. The Coyote Hole group is overlain in turn by the Fish Lake Valley assemblage (Oldow et al., 2009)—a heterolithic sequence of poorly lithified claystone, siltstone, sandstone, and sparse interbedded basaltic lavas in the northwestern part of the Silver Peak Range and study area (Figs. 2, 3). This unit is dated at ca. 3.76 to younger than 0.76 Ma based on dating and correlation of interbedded lavas and tuffs, including the Bishop Tuff (Reheis and Sawyer, 1997; Oldow et al., 2009). Though not mapped in direct depositional contact with the older Cave Spring formation, the Fish Lake Valley assemblage likely represents the continuation of fluvio-lacustrine sedimentation during Pliocene-Pleistocene time in extensional basins to the west of Cave Spring basin, including modern northern Fish Lake Valley. The Fish Lake Valley assemblage itself also contains dispersed Li and B mineralization in lacustrine siltstones, distributed over the area herein named “Fish Lake Hills basin” (Fig. 2) and considered as a future probable resource (Ioneer’s North Basin deposit).

The upper Miocene to lower Pliocene Cave Spring formation depositionally overlies volcanic rocks of the Argentite Canyon and Rhyolite Ridge formations (Fig. 3). It is predominantly composed of lacustrine strata that host the economically significant volcano-sedimentary Li and B mineralization in the study area (Oldow et al., 2009; Reynolds and Chafetz, 2020; Ogilvie et al., 2023). The Cave Spring formation consists of a conformable sequence of interbedded lacustrine claystone, marl, limestone, volcaniclastic rocks, and tuffs, with anomalously high Li content bound in marl-illite-smectite clays, and B mostly within the sodium-borosilicate searlesite (Fig. 5) (Reynolds and Chafetz, 2020; Chafetz, 2023).

Previous studies initially referred to the Cave Spring formation as “Sedimentary Unit 4” (Robinson et al., 1976) before it was assigned the name “Cave Springs Formation” [sic] by Oldow et al. (2009), who described it as being composed of detritus derived from the tuff of Rhyolite Ridge (Trt) and older Paleozoic units. Reynolds and Chafetz (2020) provided the first detailed sedimentologic and stratigraphic analyses and reported the exceptionally high Li and B contents within this unit. They also proposed that the name be revised to the “Cave Spring formation” based on the singular name “Cave Spring” indicated on existing U.S. Geological Survey topographic base maps. Hence, we abandon the previous term “Cave Springs Formation,” and adopt the informal name, “Cave Spring formation,” as used by Reynolds and Chafetz (2020) and Ogilvie et al. (2023). We recommend using “Cave Spring formation” in future studies.

The Cave Spring formation reaches up to 460 m thick and occurs in a single modified half-graben basin west of Rhyolite Ridge that we here name the “Cave Spring basin” (Figs. 1, 2). Ogilvie et al. (2023) subdivided the Cave Spring formation into three members, listed in ascending order as follows:

1. A lower heterolithic member (Tcsl) composed of limestone, marl, sandstone, siltstone, pyroclastic tuff (“gritstone”), and coarse diamictites with tuff and lava blocks derived from the Rhyolite Ridge and Argentite Canyon formations (Fig. 5A, B) (Albers and Stewart, 1972; Ogilvie et al., 2023);

2. A middle member (Tcsm) composed of well-stratified gray marl, claystone, siltstone, and carbonate that contains stratiform Li-B ore (Fig. 5C-F) (Reynolds and Chafetz, 2020);

3. An upper member (Tcsu) composed of uniformly gray, thin-bedded siltstone, sandstone, and thin ash fall tuffs.

Each of the members is separated by distinct, porous, coarse pumice-lapilli pyroclastic beds (“gritstones”), which display original gray color but are commonly brown or orange and leached due to oxidation by infiltrating fluids (Fig. 5D).

Ore zones occur in two principal stratigraphic intervals. The lower interval, in Tcsl, is composed of laminated to well-bedded marl and carbonate with variable thickness and grades (Fig. 5B). The upper interval, in Tcsm, contains an ~15 m-thick zone of Li-rich smectite with finely interlaminated carbonate (Fig. 5D, F). This upper zone overlies a relatively indurated searlesite zone (~20 m thick), which is the primary development target due to its high B values and anomalous Li in intricately bedded marls (Fig. 5C, E) containing mixed layer clays (Chafetz, 2023). The physical competence and relatively low smectite and carbonate content of the searlesite zone make it particularly amenable to processing because the searlesite matrix leaches readily with minimal acid consumption. Although Oldow et al. (2009) portrayed the Cave Spring formation as underlying the Argentite Canyon formation, field relationships, drill hole logs, and the presence of Argentite Canyon latite clasts in lower lake beds clearly demonstrate that these sediments overlie and thus postdate the latite volcanic rocks (Reynolds and Chafetz, 2020). The age of the formation is constrained between ca. 5.8 and 4.7 Ma based on stratigraphic relationships and ⁴⁰Ar/³⁹Ar and U-Pb dates of interbedded tuffs (Fig. 3; Chafetz, 2023; this study).

Contrasting tectonic models have been proposed for the Silver Peak Range that imply different structural settings during late Miocene to Pliocene deposition of the Li-B–enriched Cave Spring formation. The thickness and seemingly restricted distribution of the underlying Silver Peak volcanic center led early workers to postulate the existence of an ~10-km-diameter collapsed caldera or a caldera-like structure concealed beneath the central part of the range (Albers and Stewart, 1972; Robinson, 1972; Stewart et al., 1974; Robinson et al., 1976). Although they did not address postcaldera basin development explicitly, the proximity of the northern margin of this hypothetical caldera implies either an intracaldera lake or peripheral/moat basin setting for the coeval Cave Spring formation, analogous to the McDermitt caldera. In contrast, more recent mapping and field observations have led to the interpretation that the Cave Spring formation was deposited in an actively deforming transtensional basin in the upper plate of the Silver Peak-Lone Mountain detachment fault that was subsequently inverted under dextral transpressional strain (Oldow et al., 1994, 2009). For both models, most of the hypothesized primary structures (i.e., caldera ring-faults, NE-striking reverse faults, NW-striking dextral faults) are speculative and either have not been documented or have been misinterpreted. Alternatively, detailed structural, stratigraphic, and geophysical investigations in the Silver Peak Range and Weepah Hills suggest that the Cave Spring formation and other Pliocene-Pleistocene strata were likely deposited in extensional basins associated with high-angle normal faults that entirely postdate slip on the Silver Peak detachment (Stewart and Diamond, 1990; Reheis and Sawyer, 1997; Diamond and Ingersoll, 2002; Ng, 2018; Ogilive, 2023; Ogilvie et al., 2023). Although beyond the scope of this paper, a comprehensive structural analysis of the northern Silver Peak Range is the focus of a separate manuscript that is in preparation.

Geologic mapping of the Rhyolite Ridge area was conducted at a scale of 1:24,000 between September 2021 and November 2022 in conjunction with multispectral satellite imagery and a 1-m lidar-derived digital surface model. Field mapping and data collection were executed on Apple iPad Pro tablets using Touch GIS software, which provides an intuitive platform to record GPS locations, geologic linework, notes, structural measurements, and photographs within the portable GIS system (Darin and Wilson, 2020). Mapping efforts were enhanced by the acquisition of new high-resolution (1 m/pixel) lidar topography and airborne electromagnetic data sets provided by the U.S. Geological Survey and U.S. Department of Energy Geoscience Data Acquisition for Western Nevada project (U.S. Geological Survey, 2020b). The geologic map of the study area was recently published at 1:24,000 scale by the Nevada Bureau of Mines and Geology (Ogilvie et al., 2023); Figure 2 presents a slightly simplified version of this map that emphasizes the Neogene geology.

Uranium-lead zircon geochronology of three rhyolitic tuff samples was conducted by high-resolution laser ablation-inductively coupled plasma-mass spectrometry (HR-LA-ICP-MS) at the University of Arizona LaserChron Center (USA) following the methods of Gehrels et al. (2008) and Gehrels and Pecha (2014). After standard procedures of crushing, sieving, and density and magnetic separations, zircon grains were mounted on an epoxy disc with primary zircon standards FC-1, Sri Lanka, and R33; they were sanded down to a depth of ~20 μm, polished, cleaned, and imaged prior to isotopic analysis using a Hitachi 3400N SEM and a Gatan CL2 detector system (www.geoarizonasem.org). Cathodoluminescence (CL) images are provided in Appendix Figure A1.

Analyses of individual zircon grains and standards were conducted using a ThermoFisher Element2 single-collector, HR-LA-ICP-MS. Zircon ablation utilized a Photon Machines Analyte G2 excimer laser equipped with a HelEx ablation cell using a spot diameter of 25 μm, a laser energy density of ~5 J/cm², a repetition rate of 8 Hz, and an ablation time of 10 s, which resulted in ablation pits ~12 μm deep; sensitivity with these settings is approximately ~5,000 cps/ppm. Each analysis consisted of 5 s on peaks with the laser off (for backgrounds), 10 s with the laser firing (for peak intensities), and a 20-s delay to purge the previous sample and save files. The ablated material was then carried in helium gas into the ICP-MS plasma source, which sequences rapidly through the measurement of U, Th, and Pb isotopes.

Concordia and weighted mean age plots were generated using IsoplotR (Vermeesch, 2018), and all errors are reported at ±2σ and include only internal analytical uncertainties. Weighted mean ages were calculated using the youngest coherent population of individual zircon dates in each sample; older grain dates are interpreted as preeruptive or antecedent and excluded. In cases where the youngest population of dates was overdispersed (mean square of weighted deviates [MSWD] >> 1), the largest subset of this group was used for the weighted mean age estimate. Dates that do not overlap concordia within 2σ limits of analytical uncertainty were also excluded from weighted mean calculations, following the recommendations of Spencer et al. (2016). Raw U-Pb data are provided in Appendix Table A1.

Fresh, whole-rock samples were processed directly in the field at each site to avoid the risk of cross-contamination while in the laboratory. Most of the tuff samples contain pumice and lithic fragments; although not all lithics and pumice were removed from the geochemical samples, care was taken to exclude large lithic fragments that might introduce bias into the bulk analysis of all components (i.e., matrix, glass shards, phenocrysts, lithic fragments, and pumice).

For this study, we present whole-rock major and trace element geochemical data from a total of 42 samples in the study area from the Rhyolite Ridge and Argentite Canyon formations. Data from 35 of these samples were analyzed by the U.S. Geological Society contract lab, AGAT Laboratories, and are compiled from Ogilvie et al. (2023); the full, tabulated data set is also available in the regularly updated nationwide compilation of geochemical data generated by the U.S. Geological Society Earth Mapping Resource Initiative (U.S. Geological Survey, 2021). The remaining seven samples were analyzed by ALS Laboratory Group to complete the lithogeochemical data set published by Ogilvie et al. (2023). The analytical methodologies of each lab are summarized below.

Raw geochemical data for all samples discussed here are provided in Appendix Table A2. Other data including micro-X-ray fluorescence (μXRF) maps, cross-polarized transmitted-light (TL) photomicrographs, and optical cathodoluminescence (CL) images of representative thin sections from the Rhyolite Ridge and Argentite Canyon formations are provided in Appendix Figures A2 and A3.

Thirty-five samples compiled here were analyzed by AGAT Laboratories in Mississauga, Canada. Major elements were determined by wavelength-dispersive X-ray fluorescence spectroscopy (WDXRF). Each sample was fused with lithium metaborate/lithium tetraborate flux and the resultant glass disk is introduced into the WDXRF and irradiated by an X-ray tube. This method also provides a gravimetric measure of loss on ignition (LOI). Trace elements were analyzed for all samples by fusing them at 750°C with sodium peroxide and then dissolving the fusion cake in dilute nitric acid. The resulting solutions were then analyzed for a total of 60 elements, including Li and B, by inductively coupled plasma-optical emission spectroscopy (ICP-OES) and inductively coupled plasma-mass spectrometry (ICP-MS).

Seven whole-rock samples (MD22RR-75, RRIO22-49, RRIO22-55A, RRIO22-55B, RRIO22-56, RRIO22-60A, and RRIO22-61) were prepared and analyzed by ALS Laboratory Group in Reno, Nevada. Concentrations of all major elements and 42 trace elements were analyzed using in-house method codes ME-ICP06, ME-MS81, and ME-4ACD81. For each sample, 0.2 g was added to a lithium metaborate flux (0.90 g), mixed well, and fused in a furnace at 1,000°C, then cooled and dissolved in a 100-mL solution of 4% nitric acid and 2% hydrochloric acid. Because some sulfides and metal oxides may only partially decompose or volatilize in a lithium-borate fusion, base metals including Li were instead decomposed by a four-acid digestion; B was not able to be measured due to its volatility. The resulting solutions were then analyzed for major elements and base metals by inductively coupled plasma-atomic emission spectroscopy (ICP-AES). Results were determined in conjunction with a LOI at 1,000°C and corrected for spectral interelement interferences. Trace elements were analyzed via ICP-MS.

Tuff of Montezuma Range: Sample RRIO22-64 was collected from a widespread and undated rhyolite ash flow tuff in the Montezuma Range on the east side of Clayton Valley (Fig. 1B). Price et al. (2000) noted anomalously high Li contents (up to 192 ppm) in this ash flow tuff unit, which they referred to as the “rhyolite (or tuff) of Montezuma Range.” U-Pb zircon analysis of 34 grains from this tuff yields a weighted mean age of 6.05 ± 0.08 Ma (MSWD = 0.11) based on the youngest 12 grains which form a distinct age mode (Fig. 6A). This age is statistically identical to three ⁴⁰Ar/³⁹Ar dates determined for the Rhyolite Ridge tuff that span 6.09 ± 0.02 to 6.03 ± 0.03 Ma (Oldow et al., 2009).

Argentite Canyon tuff: Sample MD22RR-128 was collected from the welded Argentite Canyon tuff just east of Rhyolite Ridge (sensu stricto) in the hanging wall of the North Spring fault (Fig. 2). Uranium-lead zircon analysis of 28 grains from this sample yields a youngest mode of overdispersed dates that range from 6.63 to 4.99 Ma (Fig. 6B). Considering stratigraphic relationships with the overlying Cave Spring formation and published dates of 6.10 ± 0.60 to 5.85 ± 0.03 Ma for this tuff (Robinson et al., 1968; Oldow et al., 2009), we use the largest mode from this group (n = 9) to estimate a weighted mean age of 5.76 Ma ± 0.18 Ma (MSWD = 0.20) for this unit. Zircon with slightly younger dates (n = 8) probably reflect imprecision owing to their young age and unusually low U concentrations (30–81 ppm), correspondingly low ²⁰⁶Pb and ²⁰⁷Pb, and difficulties in the correction for nonradiogenic 204_Pb.

Upper Cave Spring formation: Sample RRIO21-27 was collected from a 1- to 2-m-thick, quartz-phyric pumice lapilli tuff within the upper part of Cave Spring formation at the northern end of Cave Spring basin (Fig. 2). The tuff is stratigraphically located in the middle part of the upper member of the Cave Spring formation (Ogilvie et al., 2023; Tcsu), and ~130 m below the eroded top of the formation. U-Pb zircon analysis of 31 grains from this tuff yields three pre-Miocene zircon interpreted as xenocrysts (ca. 24, 26, and 167 Ma), and a youngest mode of 16 overdispersed dates that range from 5.23 to 4.15 Ma (Fig. 6C). Based on the largest subset of this group (n = 12), we estimate a weighted mean age of 4.73 ± 0.08 Ma (MSWD = 0.88) for this tuff and for deposition of the upper member of the Cave Spring formation. This date is consistent with a ⁴⁰Ar/³⁹Ar date of 5.074 ± 0.004 Ma from an underlying tuff bed near the base of this member (Chafetz, 2023).

Petrographic study of the Rhyolite Ridge and Argentite Canyon formations was performed to characterize the mineralogical and textural variations and the type and intensity of alteration. Field characteristics and rock descriptions are based on Ogilvie et al. (2023) and incorporated here for context.

The Argentite Canyon formation is composed mostly of porphyritic latite lavas connected to subvertical dikes, and subordinately of welded tuff of nearly identical age. Effusive lavas and shallow intrusions (Tal) contain up to 40 vol % megacrystic K-feldspar and plagioclase up to 40 mm long and minor (<5 vol %) biotite and amphibole phenocrysts disseminated in a dark gray, aphanitic groundmass composed of K-feldspar and quartz (Figs. 4G, 7A-C). Samples show variable degrees of alteration characterized by albitization of K-feldspar, and sericitization of plagioclase (App. Figs. A2, A3). At microscopic scale, feldspar phenocrysts commonly show complex mineral zonation with primary Ca-rich cores rimmed by Na-rich mantles of magmatic origin, as revealed by CL images (App. Fig. A3). A late secondary rim of albite, characterized by a porous texture in optical microscopy and brownish color in CL imagery, is dominant in most altered samples and indicates Na mobility during posteruptive cooling (Fig. 7C; App. Fig. A3).

The Argentite Canyon tuff (Tat) varies from subwelded to moderately welded with a thin (<1 m) discontinuous basal vitrophyre locally. It commonly displays a distinct eutaxitic foliation and diagnostic glassy, black fiamme containing white plagioclase phenocrysts (Fig. 4F). The tuff contains abundant mm-sized volcanic lithic fragments and abundant phenocrysts of anhedral plagioclase and K-feldspar (10–15 vol %), subhedral biotite (5–10 vol %), and accessory amphibole, clinopyroxene, and quartz disseminated in a maroon to pale-orange aphanitic groundmass (Fig. 7D-F). Altered tuff samples show weak carbonate alteration that is easily visible on photomicrographs and μXRF maps (App. Fig. A2).

The Rhyolite Ridge formation includes the Rhyolite Ridge tuff (Trt) and subordinate interbedded aphanitic rhyolitic lavas, volcanic breccias, and tuff breccias (Trlb) (Fig. 4A). Rhyolitic lavas are commonly flow-banded and locally display primary internal flow folds that are asymmetric to isoclinal and formed during lava emplacement (Fig. 4B). In other areas, grayish-green outcrops of perlitic rhyolite are more common and characterized by amorphous perlite or volcanic glass spherules up to 10 cm wide, and locally supported by a matrix of light green ashy tuff with minor chloritic alteration (Fig. 7G-I). These rhyolitic lavas are interbedded with rhyolitic tuffs and tuff breccias, especially in the lower part of the unit. Clasts within the breccias are angular pieces of similar flow-banded rhyolite lava that locally constitute up to ~70% of the rock volume and may represent vent-derived rock fragments.

The majority of the Rhyolite Ridge formation is composed of the Rhyolite Ridge tuff, which consists of a series of nonwelded and lithic-rich rhyolite ash flow tuffs that occur as at least three separate flow units each up to ~120 m thick (Fig. 4A). These tuffs form resistant cliffs that are white to pale beige, where fresh, and weather to pink, orange, or dark brown (Fig. 4B-E). The tuffs contain anhedral to subhedral phenocrystic quartz (~10 vol %), plagioclase (5 vol %), sanidine (5 vol %), and accessory subhedral to euhedral biotite (Fig. 7J-L), as well as unwelded white to peach-colored pumice lapilli that rarely exceed 30 mm in diameter (Fig. 4D). Lithic fragments (20 vol %) within the tuff are almost exclusively composed of angular red to black aphanitic rhyolite lava or porphyritic rhyolite tuff that typically range in size from 2 to 25 mm (Figs. 4D, 7L). Altered Rhyolite Ridge tuff samples are characterized mostly by carbonate alteration forming veins and disseminations (App. Figs. A2, A3).

Whole-rock geochemical analyses were performed on most units in the study area, though with a particular emphasis on volcanic rocks from the Rhyolite Ridge and Argentite Canyon formations that immediately underlie the Li-B–enriched sediments of the Cave Spring formation (Fig. 3). Although most samples were collected from within the map area, several were collected from adjacent areas in northern Fish Lake Valley and Clayton Valley to investigate regional stratigraphic correlations of ash flow tuff units (Fig. 1B), as discussed in subsequent sections. A summary of the salient geochemical features of these volcanic rocks is presented below, including published data from Ogilvie et al. (2023) and new data from this work (App. Table A2).

Based on their major and trace element compositions, we distinguish altered samples from their least altered precursors. The distinction is made based on the use of the alteration box plot (Large et al., 2001) and corroborated with petrographic observations by optical microscopy, CL imaging and μXRF mapping on thin sections. In the alteration box plot (Fig. 8A), a first group of samples from Rhyolite Ridge and Argentite Canyon formations falls in the least altered rhyolite field with consistent values of alteration index (AI = 41–71) and chlorite-carbonate-pyrite index (CCPI = 23–40). A second group of samples shows decreasing values of CCPI (5–38) while increasing AI (44–89) toward the K-feldspar and sericite/illite mineral nodes. Alteration indices therefore indicate that most volcanic rocks have experienced variable degrees of mostly mild hydrothermal alteration, primarily sericitization (as secondary fine-grained muscovite/illite assemblage) and K-feldspathization (possibly as secondary adularia).

Interbedded tuffs and lavas from the least altered Rhyolite Ridge and Argentite Canyon formations have metaluminous to peraluminous compositions (A/CNK = 0.6–1.2) with Rhyolite Ridge tuff and lavas showing a more peraluminous affinity relative to Argentite Canyon rocks (Fig. 8B). Most altered samples have variable values of A/NK (1.0–1.6) relative to the precursors, which reflects the elevated K contents due to sericitization and K-feldspathization. When plotted on the total alkali versus silica (TAS) diagram (Le Bas et al., 1986), least altered samples from Rhyolite Ridge and Argentite Canyon formations are classified as rhyolites to trachydacites and trachyandesites (Fig. 8C). Least altered Rhyolite Ridge tuffs and lavas have intermediate contents of SiO₂ (61–71 wt %) and Al₂O₃ (14–17 wt %), calc-alkaline compositions with moderate contents of Na₂O (2.5–4.3 wt %) and K₂O (3.7–7.9 wt %), and relatively low contents of mafic elements such as Fe₂O₃ (2.1–4.8 wt %), MgO (0.5–1.4 wt %), and TiO₂ (0.3–0.8 wt %). In comparison, Argentite Canyon tuff and lavas are less evolved as indicated by lower contents of SiO₂ (57–61 wt %) and Al₂O₃ (14–18 wt %) and higher values of Fe₂O₃ (2.8–4.9 wt %), MgO (0.7–2.2 wt %), and TiO₂ (0.5–0.9 wt %). Altered samples are characterized by higher values of SiO₂ and K₂O (Fig. 8D) that likely reflect mild hydrothermal alteration, including silicification and K-metasomatism.

Both formations display similar rare earth element (REE) patterns characterized by strong enrichment in light REE (>10–100 times chondritic values), neutral or flat heavy REE (La/Yb = 13–46), and variable negative Eu anomalies (Eu/Eu* = 0.5–0.9), all of which are typical of the average upper continental crust (UCC) (Fig. 9A-B). Most altered samples are characterized by depletions in REE, reflecting feldspar-destructive alteration, except for one Argentite Canyon sample that shows a distinct REE pattern with high LREE values. Though altered samples from the Argentite Canyon and Rhyolite Ridge formations show variable depletions in some large ion lithophile elements (LILEs) like Ba and Sr, least altered samples have similar concentrations of other LILEs, high field strength elements (HFSEs) like Zr, Hf and Nb, and REE, with comparable values to typical UCC (Fig. 9C, D).

Samples from the Rhyolite Ridge and Argentite Canyon formations show relatively high, yet variable concentrations of both Li and B, without evident geochemical correlation. Whereas most samples contain 25 to 100 ppm Li (mean = 73 ppm), some contain exceptionally high Li concentrations up to 358 ppm (Fig. 10A, B). These anomalous values are nearly an order of magnitude higher than typical Li concentrations in the UCC (Li = 24 ppm; Rudnick and Gao, 2014) and in rhyolites and granites (Li = 40–60 ppm; Macdonald et al., 1992; Price et al., 2000; Hofstra et al., 2013). Boron concentrations in the Rhyolite Ridge and Argentite Canyon formations are even more variable, ranging from around 25 to 100 ppm (mean = 160 ppm), and up to 817 ppm (Fig. 10C-D), and are also nearly an order of magnitude higher than typical B concentrations in the UCC (B = 17 ppm; Rudnick and Gao, 2014) and in rhyolites and granites (B = 3–11 ppm; Trumbull and Slack, 2018). The least altered samples of Rhyolite Ridge and Argentite Canyon formations have average contents of 61 and 147 ppm for Li, and 102 and 193 ppm for B, respectively, which correspond to enrichment factors of 3 to 6 for Li and 6 to 11 for B, relative to the average UCC. There is no apparent correlation between Li and B, nor with other elements or alteration indices.

Immobile elements in least altered samples of Rhyolite Ridge and Argentite Canyon formations show overlapping concentrations (Fig. 11), notably for TiO₂ (0.27–0.97 wt %), Nb (13–26 ppm), Zr (184–437 ppm), and Rb (90–257 ppm). Altered samples are characterized by lower values of TiO₂ (0.08–0.68 wt %), compared to the least altered ones; concentrations of Nb (8.1–42 ppm), Zr (71–664 ppm), and Rb (145–442 ppm) are similar and independent of alteration. The variation of Ti content may reflect the partial alteration of biotite phenocrysts into chlorite, which does incorporate Ti in its crystal structure. Apart from a rough positive correlation between Zr and TiO₂, there is no correlation between the Nb, Zr, and Rb contents in volcanic rocks from the Rhyolite Ridge and Argentite Canyon formations. This erratic distribution may represent intraformational heterogeneity in the studied samples.

Over distances of kilometers, strata of the Cave Spring formation are rather consistent laterally in thickness, lithology, and geochemistry, and members exhibit distinctive signatures (Reynolds and Chafetz, 2020).

Geochemical data from two representative drill holes (SBHC-17 and SBH-103, located about 880 m apart) that penetrate the Li-B–enriched Cave Spring formation sedimentary rocks have been examined to characterize the trends for most major and trace elements (Fig. 12). The middle member of the Cave Spring formation (Tcsm), which hosts the principal Li-B ore zone, is characterized by a notable enrichment in Ca, Mg, Na, K, Rb, Cs, Be, Sb, Nb, Mo, W, Tl, and corresponding depletions in Al, Fe, Sc, Ti, V, Cr, Mn, Co, Ni, Cu, Zn, Y, Zr, P, Ba, S, Bi, Th, and U, compared to the lower and upper members of this unit (Tcsl and Tcsu). In both drill holes, the highest Li values are encountered mainly in the upper part of Tcsm (SBHC-17: 133–180 m; SBH-103: 75–108 m) and some of Tcsl (SBHC-17: 220–225 m; SBH-103: 126–149 m). In contrast, elevated B values are mostly found in a restricted horizon of Tcsm (SBHC-17: 145–177 m; SBH-103: 88–109 m), which correlates with the searlesite-bearing layers, although some B-rich horizons are also found at greater depths in Tcsl (SBH-103: 140–160 m). Besides vertical variations in the distribution of Li and B, the geochemical logs also highlight similar trends for Mg, Na, Rb, Sr, and Cs (Fig. 12). Principal component analysis indicates that Li is strongly positively correlated with some LILE like Sr, Mg, Ca, whereas B is strongly correlated with Na, thus reflecting distinct mineralogical associations (i.e., clays vs. searlesite); other trace elements show some correlations with Li or B, including K, Rb, Cs, Ba, W, Mo, Sb, Tl, Hg, and Ge (App. Fig. A4).

When normalized to the UCC, the Cave Spring formation is strongly enriched in Li and B (up to 100–1,000 times UCC, respectively), as well as in some LILEs (e.g., Cs, Sr, Rb), HFSEs (e.g., U), and specialty metals (e.g., W, Mo, As, Sb, Tl) (Fig. 13A-C; App. Fig. A5). In detail, we note that the lower and middle members are more enriched in Li-B on average than the upper member by nearly an order of magnitude, with similar enrichment trends in some specialty metals (Mo, As, Sb, Tl). Conversely, the Cave Spring formation is depleted in transition metals (e.g., Cu, Zn, Co, Ni, Sc) and some HFSEs (e.g., Zr, Hf, Nb, Ta) relative to UCC (App. Fig. A5). For comparison, we also plotted the UCC-normalized trace element compositions of the least altered samples from Rhyolite Ridge, Argentite Canyon, and Silver Peak formations underlying the Cave Spring formation. The Rhyolite Ridge and Argentite Canyon formations show similar geochemical patterns marked by enrichment in Li, B, Cs, Mo, As, Sb, and Tl, and depletion in transition metals relative to UCC (Fig. 13D-E). In contrast, the Silver Peak formation shows a relatively flat pattern with moderate enrichment in specialty metals and depletion in transition metals (Fig. 13F). The similarity between the UCC-normalized spectra of Rhyolite Ridge and Argentite Canyon formations and those of the Cave Spring formation suggest that the enriched elements (Li, B, Cs, Sb, Mo, As, Tl) in the claystones and carbonates were derived from the surrounding and underlying volcanic host rocks.

Various studies have recognized the important role of Li-enriched rhyolitic rocks as an original source for Li brine systems in Clayton Valley, as well as the efficacy of leaching and Li depletion in rhyolites by hydrothermal and/or meteoric waters (e.g., Davis et al., 1986; Price et al., 2000; Hofstra et al., 2013; Coffey et al., 2021). Stable isotopic studies of rhyolitic glass from the Yellowstone-Snake River Plain volcanic province show no obvious correlation between deposit age and Li depletion, which suggests that hydration occurs soon after rhyolite emplacement while the rock is cooling (Ellis et al., 2022). Thus, we envision that most Li depletion occurred shortly after the ca. 6.0 to 5.8 Ma emplacement of the Rhyolite Ridge tuff and Argentite Canyon formation.

Based on their successive ages and the presence of coarse breccias and clasts of latite and rhyolite tuff in the basal part of the lower Cave Spring formation (Tcsl), the Argentite Canyon latites and Rhyolite Ridge tuff were exposed at surface during the early stages of sedimentation in the Cave Spring basin. The similar ages and anomalously high Li and B contents observed in both the least altered samples of the Rhyolite Ridge tuff (up to 153 ppm Li and 355 ppm B; avg = 61 ppm Li and 160 ppm B) and the Argentite Canyon formation (up to 358 ppm Li and 388 ppm B; avg = 147 ppm Li and 193 ppm B) support the interpretation that they were the primary sources of Li and B within the Cave Spring formation. Interestingly, we note the lack of correlation between alteration degree and Li-B contents in both the Rhyolite Ridge and Argentite Canyon formations (Fig. 10), which suggests that Li and B were rapidly leached during cooling and weathering of the volcanic rocks.

Trace element geochemistry of the lacustrine strata also suggest a volcanic association with the Rhyolite Ridge and Argentite Canyon formations, as shown by the similarity in UCC-normalized spider diagrams (Fig. 13). Particularly, the Cave Spring formation is characterized by strong enrichment in incompatible elements like Cs, W, Mo, S, As, Sb, Se, and Tl, ranging between 6 and 30 times the UCC values on average, which is a strong indication for a volcanic origin (App. Fig. A5). Additionally, geochemical logs show spiky patterns with depth that correlate with high concentrations of incompatible trace elements (e.g., Rb, Sb, Ge, W, Tl). This variable pattern further supports a volcanic origin associated with ash flow tuff deposits that have anomalous metal contents.

Enrichment factors for Li and B in the Cave Spring formation are 30 and 165 times the UCC values on average, respectively, and can reach up to 50 and 325 times the UCC values on average in the middle Cave Spring formation hosting the main Li-B ore zone (App. Fig. A5). Furthermore, geochemical logs and multivariate statistical analysis indicate that Li is strongly correlated with Sr, Mg, and Ca, whereas B is strongly correlated with Na (Fig. 12; App. Fig. A4). Although this geochemical decoupling reflects the distinct mineralogical association of Li- and B-bearing phases (i.e., clays vs. searlesite) in the Cave Spring formation, the fact that Li- and B-rich layers are only partly correlated stratigraphically throughout the Cave Spring basin suggests different Li and B enrichment processes during deposition of the sedimentary sequence.

The close spatial and temporal relationship with Li-B–rich volcanic rocks and the high concentrations of lithophile incompatible elements in the lacustrine strata refute significant hydrothermal enrichment of Li and B in the Cave Spring formation. Although there is no evidence of pervasive alteration or veining in the latter, it is noteworthy that stratiform silica layers are common in the lowest 50 m of sediments (i.e., basal Tcsl) of the Cave Spring formation and close to the underlying volcanic basement. Thus, hydrothermal fluids released during the degassing and cooling of the Rhyolite Ridge and/or Argentite Canyon volcanic rocks might have circulated near the surface and led to the remobilization of Li and B during early sedimentation in the basin, similar to the hydrothermal enrichment of illite-bearing claystones within McDermitt caldera (Benson et al., 2023). Our compilation of published and new geochronological data obtained in this work indicates that the underlying Argentite Canyon tuff was emplaced at 5.76 ± 0.18 Ma, prior to deposition of the middle Cave Spring strata at 5.15 ± 0.01 Ma (Figs. 3, 6B). Although these dates record a hiatus of up to ca. 400 to 800 k.y. between these units, we cannot rule out the possibility of active geothermal systems during the inception of the Cave Spring basin, considering the duration of magmatic-hydrothermal systems, the lack of absolute dates for the lowermost lacustrine strata, and the presence of intercalated pyroclastic tuff (“gritstone”) in the lower Cave Spring formation, which has not yet been dated.

Lithium was transported into the developing Cave Spring basin from ca. 5.8 to 4.7 Ma and supplied by supergene weathering of underlying volcanic rocks and possibly aided by early hydrothermal fluid circulation. Intraformational pyroclastic flows and reworked ash fall tuffs (Fig. 5A) may have served as local, supplementary sources within the basin that contributed to syndepositional or early diagenetic Li-B enrichment. Hydrothermal activity did not secondarily affect the Cave Spring strata, because the beds lack significant alteration or veining and the mineralization is clearly stratiform. Lithium was precipitated contemporaneously in lake bottom muds and concentrated by authigenesis in finely laminated claystones (Chafetz, 2023), like the closed hydrologic system diagenesis model proposed by Castor and Henry (2020) in the McDermitt caldera.

Early studies in the region postulated the presence of a caldera centered in the Silver Peak Range and concealed beneath a thick sequence of tuffs and lavas of the Rhyolite Ridge and Argentite Canyon formations, or the “Silver Peak volcanic center” (Fig. 1B) (Robinson et al., 1968; Albers and Stewart, 1972; Robinson, 1972; Stewart et al., 1974; Robinson et al., 1976). Notably, Robinson (1972) estimated that the Rhyolite Ridge tuff reaches its maximum composite thickness of ~450 to 550 m near the margin of the inferred Silver Peak caldera and that it gradually thins outward in all directions. Likewise, early studies believed that the Rhyolite Ridge tuff had a relatively restricted distribution in only the Silver Peak Range, noting that there was no evidence that it, or other volcanic units, were enriched in Li (Albers and Stewart, 1972). Although speculative, the existence of such a caldera source for the Rhyolite Ridge tuff in the central Silver Peak Range would imply that volcano-sedimentary Li-B deposits of the Cave Spring formation were deposited in an intracaldera, or peripheral (based on their inferred caldera margin; see Fig. 2) setting. Here, we evaluate this hypothesis to clarify the genesis and volcano-tectonic setting of these resources.

During our field work, we found no substantiating evidence of typical caldera-related structures or facies in the Rhyolite Ridge area (e.g., caldera ring fault or fracture systems, excessively thick and homogeneous intracaldera tuffs, megabreccias, or rheomorphism), which are well expressed at known calderas like McDermitt (Fig. 1A) (Benson et al., 2017a,b; Henry et al.,2017; Castor and Henry, 2020). First, Robinson (1972, p. 1696) suggested that “latite flows were erupted from vents located along the ring fracture zone,” of the Silver Peak caldera and that “the only exposure of the original [caldera] wall is in a deep canyon cut through the ring fracture zone on the northwest side of the caldera.” Although a precise location was not provided, this location must refer to the south end of Wild Horse Canyon (Fig. 2), where our detailed geologic mapping reveals that hypabyssal latite (Tal) intrudes directly through Silver Peak strata (Fig. 4H). The location of this intrusion between the NW-dipping Wild Horse and Beacon Hill faults (Fig. 2) suggests that it was emplaced in the shallow crust prior to initiation of these faults that would have served as conduits for ascending magma. Critically, the consistent northwest dip direction of these and other faults to the north and west (Fig. 2) are opposite of the expected southeast dip of ring faults along the inferred northern caldera margin (Robinson, 1972; Robinson et al., 1976), further refuting the notion of a caldera structure at this location. In addition, mapping and drilling in South Basin reveal no structural or facies boundary indicative of a caldera margin or moat deposition. Based on the lack of direct evidence for the purported “Silver Peak caldera” within or near the study area, we prefer to invoke the more appropriate “Silver Peak volcanic center” (Robinson, 1972) for only the locally derived rhyolite and latite lavas and domes in the Rhyolite Ridge and Argentite Canyon formations.

We propose a stratigraphic correlation of the tuff of the Montezuma Range on the east side of Clayton Valley, with the Rhyolite Ridge tuff in the Silver Peak Range and northern Fish Lake Valley (Fig. 14). This correlation is based on the remarkable similarities among these ash flow tuffs, especially the following features of the the Montezuma tuff: (1) it is a composite sequence of interbedded nonwelded lithic-rich rhyolite tuffs and lesser tuffaceous sandstones at least 350 to 500 m thick (Albers and Stewart, 1972; Davis and Vine, 1979; Price et al., 2000); (2) it has a comparable lithic composition, bulk geochemistry, and Li content (Fig. 14B); and (3) it has a U-Pb age of ca. 6.05 ± 0.08 Ma (Fig. 6A), identical to multiple ⁴⁰Ar/³⁹Ar ages of 6.03 ± 0.03 and 6.09 ± 0.03 Ma at Rhyolite Ridge (Oldow et al., 2009).

These characteristics are also shared by other tuff exposures across the greater Clayton Valley region. Field relationships, published geologic maps, and subsurface data suggest that the Rhyolite Ridge tuff and its correlatives gradually thicken southeastward from ~150 m to more than 500 m along an ~80 km transect from the northern White Mountains to the southern Montezuma Range (Fig. 14A). Strikingly similar mineralogical and geochemical characteristics in correlative tuffs in the southeast Volcanic Hills and Montezuma Range support this correlation (Fig. 14B). The Rhyolite Ridge tuff has also been identified in the VRS-1 petroleum exploration well in northern Fish Lake Valley where it is ~300 m thick (Ng, 2018), and in exploration wells beneath the younger sedimentary basin fill of Clayton Valley, where it has a statistically indistinguishable ⁴⁰Ar/³⁹Ar date of 6.134 ± 0.095 Ma (Gagnon et al., 2023), but its total thickness is unknown (Fig. 14A). Notably, a correlative tuff located ~11 km northeast of Rhyolite Ridge in the Weepah Hills is substantially thinner (0–30 m; Stewart and Diamond, 1990; Burrus, 2013) and has a statistically identical date of 6.0 ± 0.9 Ma (Burrus, 2013; AHe date from their unit “Tat” in their Alum mine formation).

The spatial patterns in the extent and thickness suggest that the Rhyolite Ridge tuff was deposited in a NW-SE–trending paleovalley at ca. 6 Ma. The northeast margin of this paleovalley seems to coincide with the parallel and topographically high structural culmination (or antiformal dome) of Mineral Ridge, which would explain the near absence of this tuff beyond the Silver Peak Range. Although the location of the source vent for the Rhyolite Ridge tuff is unknown, the extraordinary thickness of the tuff on the east side of Clayton Valley suggests that it may have originated from a vent located in the vicinity of the Montezuma Range (Fig. 14A), an idea originally postulated by Davis and Vine (1979).

This regional stratigraphic correlation and the above constraints on the tuff’s thickness and distribution allow us to estimate the minimum dense rock equivalent (DRE) volume for the Rhyolite Ridge tuff. Assuming an initial area of ~1,300 to 2,100 km² and an average thickness of 250 to 300 m (Fig. 14), a bulk tuff density of 2.4 to 2.6 g/cm³, and a DRE (i.e., rhyolite) density of 2.7 g/cm³, we estimate the minimum DRE volume of the Rhyolite Ridge tuff to be between 289 and 607 km³. Given the uncertainty in the full extent of the tuff, this range of ca. 300 to 600 km³ represents a conservative estimate. This minimum erupted volume for the Rhyolite Ridge tuff is slightly lower yet comparable to conservative volume estimates for the 0.767 Ma Bishop tuff (DRE <600–650 km³; Hildreth and Wilson, 2007) and the 16.4 Ma McDermitt tuff (DRE >600–1,100 km³; Henry et al., 2017).

Additional geologic mapping, petrologic, geochronologic, and paleotransport investigations are needed to test our proposed regional correlation of the Rhyolite Ridge tuff and to establish its source location, initial extent, and volume with greater precision. This is especially important given the growing number of Li prospects that are collocated with occurrences of the tuff (Figs. 1B, 14A), and to enhance ongoing exploration efforts for undiscovered Li brine and volcano-sedimentary Li deposits across the greater Clayton Valley region.

Despite receiving comparatively more attention, the source of Li-rich brines in neighboring Clayton Valley is also uncertain. Davis et al. (1986) suggested that the Li in Clayton Valley may have been derived from weathering of extensive exposures of undated rhyolitic tuffs, lavas, and obsidian at Clayton Ridge and in the Montezuma Range (Fig. 1B), which is supported by their unusually high Li concentrations of up to 228 ppm (Price et al., 2000). Similar Li content is observed in groundwater and subsurface volcanic and sedimentary samples within Clayton Valley, where leaching experiments and basin modeling suggest that the Li in brines is sourced from leaching of subsurface tuffaceous rocks and/or Li-enriched sedimentary rocks by hydrothermal fluids (Fig. 15A) (Jochens and Munk, 2011; Hofstra et al., 2013; Munk et al., 2016; Coffey et al., 2021). Geochemical studies and regional correlations of volcanic strata between Rhyolite Ridge and Clayton Valley are thus a crucial step toward evaluating potential Li sources and shared characteristics of the robust regional Li system.

Our proposed correlation of the Rhyolite Ridge tuff across Clayton Valley and its widespread distribution across the Silver Peak Range and in the subsurface strengthens the case for its significance as a local and regional Li source for the prolific Li-brine system in Clayton Valley, and undiscovered resources throughout the greater Clayton Valley region. Further research leveraging radiogenic (Sr, Nd) or stable (Li, O) isotopes in combination with other geochemical tracers and new geochronological constraints from various occurrences of the Rhyolite Ridge tuff could allow comprehensive testing of this hypothesis and enhance our understanding of Li distribution and mobility within the region.

Our findings underscore the potential of extensional tectonic basins as key exploration targets for volcano-sedimentary Li(-B) resources, a structural setting that has historically been less studied compared to more traditional volcanic settings. Unlike the large volcano-sedimentary Li deposits found in volcanic settings like McDermitt caldera (Fig. 15B) (Benson et al., 2017b; Henry et al., 2017; Castor and Henry, 2020) or maar crater basins in northwest Arizona (Thompson et al., in press), volcano-sedimentary Li-B deposits of the Cave Spring formation developed in a purely extensional tectonic environment, with a common Li source in the regionally extensive Rhyolite Ridge tuff (Fig. 15A). Although the volcano-tectonic settings of Rhyolite Ridge and McDermitt caldera share some similarities such as significant sediment accommodation, elevated heat flow, crustal melting, and contemporaneous magmatism, there are several key differences between them. First, an important distinction is the contribution of mantle-derived melts associated with the Yellowstone hotspot at McDermitt, as indicated by volumetrically important peralkaline magmatism and postcaldera resurgence of icelandite volcanism (Benson et al., 2017a,b; Henry et al., 2017); volcanic rocks at Rhyolite Ridge have a strongly crustal source signature by comparison. Second, the strong enrichment in B in lacustrine strata at Rhyolite Ridge is unique and unseen at McDermitt caldera, where the claystones are only enriched in Li. Finally, Rhyolite Ridge lacks the extensive epithermal Hg, U, and Ga-Sb mineralization that is prevalent along postcaldera ring faults at McDermitt (Castor and Henry, 2000; Henry et al., 2017; Benson et al., 2023). Instead, Li-B mineralization at Rhyolite Ridge appears to be primarily associated with supergene weathering of the Rhyolite Ridge tuff and/or locally sourced latite lavas of the Argentite Canyon formation and closed hydrologic system diagenesis driven by significant tectonic subsidence associated with regional extension, with limited hydrothermal enrichment.

These results suggest that, like caldera systems, extensional tectonic settings can indeed provide all the critical conditions necessary for forming economically significant Li resources, especially in arid climates where extensive Li-enriched ignimbrites, like the Rhyolite Ridge tuff, are present (Fig. 15). Moreover, the strongly crustal source within cratonic North America (Fig. 1A) is responsible for the economically significant enrichment in B at Rhyolite Ridge. Consequently, this alternative structural framework and tectonic setting reveals promising opportunities for Li-B exploration not only within the Rhyolite Ridge and Clayton Valley regions, but also in similar extensional settings worldwide.

Detailed geologic mapping, geochemical, and geochronologic data offer critical insights into the source-to-sink evolution of volcano-sedimentary Li-B deposits of the Cave Spring formation at Rhyolite Ridge, Nevada. Unusually high Li and B concentrations in peraluminous to metaluminous rhyolitic and trachytic tuffs and lavas of the Rhyolite Ridge and Argentite Canyon formations are substantially higher than average upper continental crust, though with no apparent correlation between the two elements. Hydrothermal alteration of volcanic rocks is spatially variable and characterized by weak sericitization, K-feldspathization, and carbonate disseminations, as well as depleted REE patterns in more altered samples. Drill hole geochemical data in the overlying Cave Spring formation reveal that mineralization is stratiform, heterogeneous, and marked by depletion in transition metals and significant enrichment in Li, B, and several other incompatible trace elements (e.g., Cs, Sb, Mo, As, Tl) that reach up to 30 times higher than typical UCC. Enrichment of Li and B in the Cave Spring formation is stratigraphically controlled and thus particularly influenced by contemporaneous supergene weathering of the surrounding volcanic rocks of the Rhyolite Ridge and Argentite Canyon formations, potentially enhanced by early and limited hydrothermal fluid circulation within the basin.

Field observations and geologic mapping challenge the previously hypothesized Silver Peak caldera structures in the central Silver Peak Range. Instead, our new U-Pb geochronology, petrographic analysis, and geochemical data support a regional stratigraphic correlation of the Rhyolite Ridge tuff with similar tuffs along an ~80 km northwest-southeast transect from the northern White Mountains and Fish Lake Valley, through the Silver Peak Range and Clayton Valley, to the Montezuma Range. We estimate a minimum DRE volume of ~300 to 600 km³ for the Rhyolite Ridge tuff and interpret that it was sourced from an unknown vent located in the Montezuma Range, where we report a date of 6.05 ± 0.08 Ma (U-Pb zircon) for a correlative rhyolite ash flow tuff; the distribution and thickness trends suggest that it was deposited in a NW-SE–trending paleovalley.

Given its unusually high Li content and extensive distribution, the Rhyolite Ridge tuff is a critical proximal source of Li in the prolific Clayton Valley brine system and associated volcano-sedimentary Li(-B) deposits. Further research is needed to test this correlation and pinpoint the precise source location of the tuff. Nevertheless, these findings significantly enhance the understanding of Li(-B) resources in the region and suggest promising exploration opportunities throughout greater Clayton Valley and across similar extensional and rift-related settings. The unusually high Li content and widespread extent of the Rhyolite Ridge tuff underscore its importance as the probable source of Li in the Clayton Valley brine system and in related exploratory prospects across the region.

This research was supported by the U.S. Geological Survey’s Earth Mapping Resources Initiative (Earth MRI) and National Cooperative Geologic Mapping Program under U.S. Geological Survey award number G21AC10365 to Darin and Harlaux and a Geological Society of America graduate student research grant to Ogilvie; additional support for analytical work was provided by the University of Nevada, Reno, College of Sciences. We thank Wes Johns, Rachel Micander, Domink Vlaha, and Mary Hannah Giddens for crucial assistance in the field, and Ioneer for logistical support and for generously sharing their data to enhance this study and our understanding of the local geology. We are grateful to Robin Armstrong and Guest Editor Thomas Benson for their constructive comments, which helped to improve the quality and focus of the manuscript.

Mike Darin is a structural geologist with more than 15 years of experience examining the tectonic evolution of plate boundaries, including the Gulf of California-San Andreas transform-rift system, the Anatolian orogen, and the Cascadia subduction zone. He earned his bachelor’s degree from the University of Colorado at Boulder and a master of science degree in geology from the University of Oregon, after which he spent two years in the oil and gas industry on the deepwater Gulf of Mexico exploration team at ConocoPhillips Co. He went on to earn a Ph.D. in geology at Northern Arizona University and spent several years on the faculty at the University of Nevada, Reno, with the Nevada Bureau of Mines and Geology. His research explores the links between tectonics, sedimentation, magmatism, and mineralization through geologic mapping, structure, stratigraphy, and geo/thermochronology. He currently works as a regional geologist with the Oregon Department of Geology and Mineral Industries (DOGAMI), where he leads the geologic mapping program for the state, continuing his research into the genesis of economic sediment-hosted lithium deposits at McDermitt caldera.