Volcano-Sedimentary Deposits: Overview of an Emerging Type of Lithium Resource

Lithium (Li) is part of the critical raw materials group and as such appears as a ubiquitous entry in the critical mineral lists recently published by different countries and by autonomous intergovernmental institutions (Austrade, 2020; International Energy Agency, 2021; Lusty et al., 2021; European Commission, 2023; U.S. Geological Survey, 2024). The increase of the Li demand is currently driven by its importance to develop low-CO₂ emission technologies, as decarbonization of modern society is largely dependent on the commercial-scale production and global proliferation of electric vehicles and power generation units, which are predominantly powered by rechargeable Li-ion batteries (Goodenough and Park, 2013; Kim et al., 2019).

Current Li production is restricted to Australia, Brazil, Argentina, Chile, and China with a subordinate production coming from Portugal, United States, and Zimbabwe (Iyer and Kelly, 2023; U.S. Geological Survey, 2024). In these districts Li is produced from salars and lithium-cesium-tantalum (LCT) pegmatites (e.g., Bowell et al., 2020).

In salar deposits, saline brines form in continental restricted basins associated with volcanic deposits where high evaporation rates concentrate Li (e.g., Munk et al., 2016, in press). Major salar operations in South America are found in the so-called “Lithium Triangle” between Chile, Argentina, and Bolivia, but Li brine deposits are also found in China (Tibetan Plateau) and in the western United States (Clayton Valley). In LCT pegmatites, also described as “hard-rock” ores, Li production is restricted to the Li pyroxene mineral spodumene because of established calcination and leaching processing methods currently used by industry (Zhao et al., 2023). This type of ore is generally high grade, since the main ore mineral, spodumene, contains around 3.2 to 3.7 wt % Li—a metal grade significantly higher than other Li-bearing silicate minerals (Dessemond et al., 2019). However, the physical features of LCT pegmatites, characterized by low overall mineralized volumes and by irregularly distributed orebodies, still pose challenges to both exploration and economic exploitation of this ore type (Müller et al., 2023).

The Li market has been dominated by salar deposits until 2018 largely owing to the low costs of production initially focused on evaporative-based processing. However, direct lithium extraction (DLE) is now being considered as a route to reduce water use and production of large volumes of coprecipitated waste (Vera et al., 2023). More recently there has been an increase in Li exploration and production focused on pegmatite ores (Gardiner et al., 2024, and references therein). This shift in the Li market has been triggered by the development of new technological solutions for Li-ion batteries, which are based on variable composition of cathodes and thus on a variable preferred type of Li feedstock (i.e., Li₂CO₃ vs. LiOH). Lithium production from brines was driven by the low-cost production of Li₂CO₃ from an evaporated brine, which is the main feedstock material for conventional lithium-cobalt-oxide (LCO) batteries (Mohr et al., 2012; Bradley et al., 2017). The recent development of longer-range and higher-energy nickel-cobalt-aluminum (NCA) and nickel-manganese-cobalt (NMC) Li-ion batteries, which prefer LiOH as the input feedstock, has accelerated Li exploration toward pegmatites, since LiOH is more economical to produce from spodumene compared to salar brine (Gardiner et al., 2024, and references therein).

Despite such recent developments, to support the expected long-term growth of the Li-ion batteries market an eightfold increase of the current global Li production will be required by 2040 (Benchmark Mineral Intelligence, 2023; U.S. Geological Survey, 2024). This implies that regardless of feedstock, the discovery of more deposits of diverse types is paramount to securing a sufficient future Li supply. In addition to salar and LCT pegmatite deposits, new potential deposit types include low-temperature oil-field brines, rare metal granites (RMGs), geothermal brines, and volcano-sedimentary (VS) deposits (Kesler et al., 2012; Bowell et al., 2020; Romer and Pichavant, 2020; Sanjuan et al., 2022), the latter being the subject of this paper. The McDermitt caldera-hosted VS system that straddles Nevada and Oregon is shown to contain measured and indicated resources of up to 4.49 million metric tonnes (Mt) Li hosted in clay minerals, making it the largest known accumulation of Li on the planet (Castor and Henry, 2020; Benson et al., 2023). Further significant tonnages of respectively 0.26 and 0.70 Mt Li have been reported at Rhyolite Ridge (Nevada, US) and Jadar (Serbia), and each of these deposits is reported to be amenable to acid leaching technologies for processing the Li ores (Ioneer, 2020; Rio Tinto, 2022; Lithium Americas, 2024).

Volcano-sedimentary deposits are therefore an emerging Li ore type. In VS deposits Li is contained in phyllosilicates, borosilicates, and phosphates, which form via the alteration of Li-fertile volcanic to plutonic material in the critical zone environment (Bowell et al., 2020). However, critical features such as complex Li mineralogy and deportment, as well as presence of orebodies with low and variable grades, result in major uncertainties concerning the potential development of effective processing solutions. In the last decades, this hampered investment in the exploration and development of VS ores in favor of more traditional Li sources such as salars and spodumene pegmatites. However, the rising demand for Li and the need to target alternative Li sources increased the interest in this unconventional ore type, with major investments made in the western United States, eastern Europe and elsewhere in the world. A series of recent academic studies focused on some features of VS-style Li mineralization from well-known Li-endowed districts such as the western United States. For example, Benson et al. (2017a) examined the mechanisms of Li enrichment in rhyolite magmas associated with the McDermitt caldera, which are the precursor of the Li enrichment in the associated VS-style mineralization. For the same district, Castor and Henry (2020) and Benson et al. (2023) proposed genetic models to explain minerogenetic processes accounting for Li enrichment in the VS ores. Other studies (i.e., Coffey et al., 2021; Munk et al., 2016; Gagnon et al., 2023) explored the paleoclimatic control and the peculiar association of Li clays and Li brines in Clayton Valley (Nevada, US), which is the only Li producer in the United States to date from a traditional brine process. The increasing interest in VS deposits also led a reanalysis of similar mineralization styles that have been traditionally targeted for other commodity types, such as borates. An example is the borate district in western Anatolia (Turkey), which according to the recent study by Kadir et al. (2023) has high potential for VS-style Li ores; its counterpart in western Balkans, whereby the exploration program aimed to target borates and led to the discovery of jadarite, an Li borosilicate that formed in a VS setting (i.e., Stanley et al., 2007; Putzolu et al., in press). Based on the most recent studies, VS deposits share some key characteristics but display a high diversity of ore assemblages, tectonic setting, and nature of the ultimate source of ore metals. However, no existing study provides a general overview of the features of this emergent Li ore type. In this paper we review the features of VS systems by considering the published literature and by presenting new observations and data. Our goal is to provide a deposit model that can be used for understanding the defining features of VS deposits and the potential variations to be expected in these ore systems. We will summarize the features of the tectonic settings, associated igneous rocks, hosting basins, ore deposition processes, and mineralogical and geochemical footprint in VS deposits and VS-endowed districts. This summary allows us to provide an updated and consistent classification scheme of VS deposits for future exploration.

Interest in VS deposits as a potential source for Li is relatively recent, with significant exploration and related academic projects only having been developed in the last six to seven years. Nevertheless, Li-bearing VS deposits have been known in the western United States since the late 1950s (Ames et al., 1958; Tourtelot and Brenner-Tourtelot, 1977; Brenner-Tourtelot and Glanzman, 1978; Vine and Dooley, 1980; Asher-Bolinder, 1991). Investment in this deposit type was limited in the past because of the unfeasible application of high-temperature roasting methods developed for spodumene deposits. However, recent advances in more economically viable sulfuric acid leaching methods (e.g., Roth et al., 2022) led to a resurgence of the interest of the industry in the exploration and exploitation of VS deposits. Inconsistent descriptions of VS deposits in the literature have led to uncertainties about their nomenclature and about their key features and how they might be interpreted. For examples, pioneering studies from Ames et al. (1958) and Asher-Bolinder (1991) described VS ores as “Li-bentonite deposits” and as “Li-smectites of closed basins,” respectively. A similar nomenclature has been reported in more recent studies, with key examples being “Li-rich claystone deposits” (Henry et al., 2017) and “Li-clay deposits” (Bradley et al., 2017; Bowell et al., 2020). We believe that this nomenclature oversimplifies some of the defining features of this ore type, such as the following:

1. Variability of ore mineralogy: Known VS deposits display a pronounced diversity of ore mineralogy, which despite often consisting of smectite clays, may also include other phyllosilicates (illite, mica, and chlorite groups), as well as borosilicates and phosphates.

2. Association between lithium and boron: In several instances Li and B are shown to behave similarly from the magmatic, to the hydrothermal, and eventually to the surface environment. Therefore, several VS deposits also contain significant borate resources, which in districts like Turkey and the western United States represent the only commodity type that is currently produced from the VS ores.

3. Complex pathways to ore formation: Nomenclature in the literature proposes that the Li enrichment in a surficial environment is solely due to sedimentary and low-temperature processes accounting for the Li leaching from source rocks and consequent precipitation in secondary minerals. In this paper, we will discuss evidence that supports the link to both Li-fertile volcanic and plutonic rocks and the enrichment of Li in the system due to elevated heat flow and high-temperature processes.

We propose that this ore type should hereafter be referred to as “volcano-sedimentary Li(B) deposits” or “volcano-sedimentary Li(B) systems,” which is a nomenclature that has been adopted by recent studies (i.e., Benson et al., 2023; Kadir et al., 2023; Putzolu et al., 2023a, b). However, the nomenclature “volcano-sedimentary Li deposits” can be also locally adopted to describe VS deposits that do not contain economic borate resources (e.g., Thacker Pass, McDermitt, and Nevada North deposits). According to this mineral system-based nomenclature, VS ores include a series of Li(B)-rich deposits hosted by mixed volcanic and sedimentary successions that are deposited in lacustrine (endorheic) basins and developed nearby or within Li(B)-fertile igneous provinces (Fig. 1).

Known VS ores display a suite of defining characteristics, which can be used to build a deposit model and form the basis for geologic guidelines to inform mineral exploration (Fig. 1). These include the following:

1. Basin hydrology: VS deposits are hosted in (paleo)lacustrine systems that develop under hydrologically closed regimes. These conditions prevent loss of mobile ore metals, such as Li and B, which in the surficial environment can be easily transported by fluids. The hydrology of VS-hosting basins ranks among the most critical features that preserve the fertility of VS deposits, as it minimizes the leaching of mobile ore metals by fluid outflow and prevents the dilution process of dissolved Li⁺ and B³⁺ cations by hampering inflow of large volumes of meteoric water.

2. Basin architecture: This feature plays a primary role in defining the hydrological conditions explained above. Volcano-sedimentary deposits are in structurally restricted basins, which generally form either by tectonic or volcanic subsidence. These include extensional basins, such as semigraben and pull-apart basins, and collapsed caldera structures. These types of basins are commonly found in back-arc and intracratonic settings.

3. Climate: Despite specific basin hydrology and architecture being the key features to form a favorable geologic trap for Li, they only account for the preservation of the ore fertility in a VS system. A VS system needs to experience additional processes to concentrate Li to reach an ore-grade level. The development of playa lake depositional environments under arid to semiarid conditions is critical for Li enrichment via evaporation processes.

4. Active volcanic and/or magmatic processes: This is a ubiquitous feature of VS ores. Evidence discussed in this paper supports a strong spatial and temporal association with felsic plutonic and volcanic provinces, both in terms of input of volcanic material and of increased thermal gradients via geothermal to hydrothermal processes. Active magmatism at the time of development of a VS ore plays a primary role in sourcing Li through leaching of intrabasinal volcaniclastic rocks and via transport in higher-temperature fluids.

5. Heat flow: This feature is linked to the presence of active magmatic processes. The onset of high heat flow in VS ores plays a significant role in enhancing the Li distribution in the orebodies via the following processes: increase of the kinetics of alteration of Li-rich source rocks (e.g., volcanic glass-rich volcaniclastic and pyroclastic rocks); onset of thermally driven circulation of Li-bearing fluids; onset of thermally driven evaporation of lacustrine waters and formation of Li-rich brines and of gel-like media via thermal distillation; and alteration and upgrade of early formed Li minerals by late diagenetic or late hydrothermal mineralizing processes.

6. Host-rock reactivity: Basin infill in VS deposits largely includes extrusive igneous rock types such as ignimbrites and ash falls. These rock types often contain high amounts of volcanic glass and pumice or the relicts of such rock types with high surface areas. This amorphous vitric component under specific chemical pH-Eh conditions is highly reactive and easily alters to form mineralogically complex assemblages that fix Li from the mineralizing fluids.

7. Age: Most known VS deposits occur in young geologic terranes (i.e., Tertiary or younger). Only a few older examples of VS deposits have been recorded, with ages dating back to the Mesozoic, Palaeozoic, and Paleoproterozoic. The lack of old examples of VS ores suggests that preservation of this ore type is likely to be hampered by their shallow environment of formation, leading to the easy mechanical erosion of the orebodies, and by the narrow pressure-temperature-composition (P-T-x) stability field of ore-bearing phases such as phyllosilicates and borosilicates.

It should be noted that features 1 to 5 of VS ores match with features observed for salar deposits (Munk et al., 2016; Bradley et al., 2017). However, the presence of active mineralizing processes and the preservation in the geologic record, respectively features 6 and 7, represent the main key diversions of VS deposits from salar systems, which will be further discussed in this paper. While the above key characteristics are observed in all VS Li deposits, other features such as ore assemblages, grades, and volumes of the mineralized facies vary significantly (App. 1). Most VS deposits, such as those in western United States (i.e., Thacker Pass, McDermitt, Rhyolite Ridge) and elsewhere (e.g., Shavazsay deposit, Uzbekistan) show an Li enrichment associated with clay minerals. This subclass of VS ores has between 0.14 and 0.28% Li, with the highest concentrations observed in the Thacker Pass, Sonora, Nevada North, and Shavazsay deposits (Fig. 2; App. 1). Ore tonnages in clay-rich VS deposits vary between 0.06 and 3.78 Mt Li. Clay-dominated VS deposits are characterized by ranges of Li grades and tonnages similar to those of Li mica mineralizations in RMG deposits (App. 1; Fig. 2). However, some VS deposits fall outside the tonnage-grade field measured in clay-type Li deposits and RMGs. Clay-dominated VS deposits with associated Li brine resources (e.g., Zeus and Cypress projects from the Clayton Valley district, US) have tonnages within the range of clay-rich VS ores and RMGs (i.e., 0.24–0.34 Mt Li); however, these are characterized by systematic lower grades of around 0.09 to 0.11% Li (App. 1; Fig. 2). Another notable example is the Jadar VS deposit (Serbia), where the Li ore is largely hosted by the borosilicate jadarite containing 3.25 wt % Li. This unique mineralogy results in average deposit grades of 0.83% Li, shifting the Jadar deposit into the field of spodumene LCT pegmatites that represent the high-grade end members of the known geologic sources of Li (App. 1; Fig. 2).

The major global VS districts are found in (Fig. 3) (1) North America, which mostly includes VS deposits from the Basin and Range (i.e., western US and northern Mexico), (2) South America, which includes deposits from Central Andes, (3) Tethyan district, which includes deposits from western Balkans, western Anatolia, central Iberia, and southern Alps, and (4) Asia, which includes deposits from China and Uzbekistan. Here we will illustrate the overall features of the tectonic environment of VS deposits from the Tethyan (i.e., western Balkan and western Anatolia) and North America (i.e., Basin and Range; western US) regions, which are the mineral districts with the most comprehensive descriptions available in the literature.

Volcano-sedimentary Li(B) deposits in the Balkans are developed within the Sava-Vardar zone, a SE-trending belt that marks the suture between the African and Eurasia plates that can be traced to join the Izmir-Ankara-Erzincan zone in western Anatolia (Schmid et al., 2008, 2020). A simplified crustal architecture of western Balkans at the Oligocene-Miocene transition is shown in Figure 4A. The main VS systems in western Balkans are the Jadar, Valjevo, Pranjani, Rekovac, Jarandol-Piskanja, and Prajani deposits. These basins show Li enrichments in clays and locally in the borosilicate jadarite and host significant borate resources (Borojević Šoštarić and Brenko, 2022; Putzolu et al., in press). The VS deposits in the Balkan region developed as endorheic lakes during the Miocene postcollisional extension across the Sava-Vardar zone (Borojević Šoštarić and Brenko, 2022). A spatial and temporal relationship exists between the basin development and felsic plutons. Plutonic and subordinate volcaniclastic rocks in the Sava-Vardar zone are of Oligocene to Miocene age and are related to postcollisional evolution of the Dinaride-Hellenide orogen belt (Borojević Šoštarić and Brenko, 2022, and references therein). Oligocene I-type granites (around 32 Ma) originated during a late-collisional stage and from melting of the lower crust (amphibolites), whereas Miocene S-type granites (cooling around 18–16 Ma) developed during a later stage of back-arc extension and exhumation of metamorphic core complexes (Koroneos et al., 2011; Stojadinović et al., 2017; Löwe et al., 2021; Borojević Šoštarić and Brenko, 2022). During this episode of crustal thinning, upwelling of asthenospheric material caused the melting of the upper crust and the emplacement of granitoids and pegmatites with a pronounced crustal affiliation (Koroneos et al., 2011). Miocene extension reactivated ancient thrust faults that formed during the collisional stage and locally triggered uplift and low standing of the base level (Marović et al., 2007; Balling et al., 2021; Löwe et al., 2021) that led to a series of high-salinity closed lakes (so-called Western Balkan Lake system; Obradović and Vasić, 2007). At this switch from collision to extension, extensional basins developed synchronous to generation of Li(B)-fertile felsic magmas with a crustal affiliation (Borojević Šoštarić and Brenko, 2022, and references therein). These processes resulted in a favorable combination of Li(B) input into the lacustrine environment via physical (e.g., deposition and/or reworking of extrusive igneous rocks) and/or chemical (i.e., transport by fluids) transfer.

Volcano-sedimentary deposits in western Anatolia account for around 70% of the known global reserve of borates (U.S. Geological Survey, 2024). However, recent studies have highlighted that these B-rich VS deposits also have a largely untested potential to host Li in the form of clays (Helvacı et al., 2004; Kadir et al., 2023). Much of the late Mesozoic and Cenozoic tectonic evolution of western Anatolia is related to the Izmir-Ankara-Erzincan zone, which represents the southeastern continuation of the Sava-Vardar zone from the Balkan VS systems. The main VS systems in western Anatolia are the Bigadiç, Kestelek, Sultançayir, Emet, and Kirka deposits (Helvacı et al., 2004), which are spread within the Izmir-Ankara-Erzincan zone. Like in the Balkan region, the development of VS-hosting basins in western Anatolia is synchronous with a Miocene hyperextension stage, as well as with lithospheric thinning and igneous activity, which are linked to rollback of the Hellenic slab (McKenzie, 1972; Royden, 1993). The Bigadiç, Sultançayır, and Kestelek basins developed along the İzmir-Balıkesir transfer zone, whereas the Selendi and Emet basins formed adjacent to the Menderes metamorphic core complex (Helvacı, 2019). Magmatic activity during postcollisional extension occurred during the Late Oligocene-Early Miocene and Late Miocene-Early Pliocene, with the first stage likely responsible for the Li(B) enrichment in the Anatolian VS district (Helvacı, 2019, and references therein). Late Oligocene-Early Miocene magmatism led to the emplacement of basaltic andesite to rhyolites with a calcalkaline to shoshonitic affinity (Bingöl et al., 1982; Güleç, 1991; Aldanmaz et al., 2000), whereas the Late Miocene-Early Pliocene stage led to the emplacement of sodic and more basic igneous rocks (Güleç, 1991; Aldanmaz et al., 2000). Geologic evidence supports that Li(B)-fertile magmatism during the Miocene was contemporaneous with exhumation of the Menderes metamorphic core complex and with the development of extensional basins that acted as final repositories for the erupted felsic material. The geochemical and isotopic study of local igneous rocks (Palmer et al., 2019; Lefebvre-Desanois et al., 2023) suggests that the Li(B)-fertile magmatism resulted from melting of a mantle source, which was metasomatized by crustal material likely detached from the subducting slab.

Most of the VS deposits from North America occur in the Basin and Range province in the western United States. The Basin and Range developed during Cenozoic intracontinental extension that had started in the Late Eocene (~40 Ma) following the Sevier-Laramide phases of crustal shortening and mountain building (~155–60 Ma) (Dickinson, 2002). The subduction of the eastward-dipping Farallon slab then evolved to slab rollback and detachment around 16 Ma (Camp et al., 2015; Fig. 4B), with resultant crustal extension and topographic collapse of the Nevadaplano high-elevation plateau (DeCelles, 2004). The crustal stretching gave rise to exhumation of deep crustal rocks as metamorphic core complexes and to the formation of the current Basin and Range topography (Armstrong, 1982; Dickinson, 2004). The Fallon slab rollback triggered widespread bimodal volcanism from the resulting influx of hot asthenosphere underneath the overriding North American plate (e.g., Dickinson, 2004; Camp et al., 2015) (Fig. 4B).

Most of the known VS deposits in the Basin and Range occur as lacustrine sediments deposited in extensional basins (e.g., Cave Spring Formation-Rhyolite Ridge, Clayton Valley, Sonora, Horse Spring Formation; App. 1). The one notable exception is the ~16.3 Ma McDermitt caldera, which developed in the northern limb of the Basin and Range along the continental margin close to the western edge of the North American craton marked by the ⁸⁷Sr/⁸⁶Sr 0.706 line (Kistler and Peterman, 1973; Henry et al., 2017) (Fig. 4B). The McDermitt caldera, along with coeval Columbia River Basalt Group flood lavas, are considered to mark the impingement of the Yellowstone plume head under the edge of the North American craton (e.g., Coble and Mahood, 2012; Camp et al., 2015: Henry et al., 2017). Eruption of large volumes of fractionated Li-rich silicic magma at 16.39 Ma (i.e., McDermitt tuff) generated the collapse of the McDermitt caldera and formation of intracaldera Li-rich volcano-sedimentary units (e.g., Benson et al., 2017a; Henry et al., 2017).

A significant proportion of the igneous rocks associated with VS deposits are volcanic glass-rich fragmental volcanic rocks of rhyolitic composition composing pyroclastic, volcaniclastic, and reworked epiclastic units. For example, distal extrusive units (e.g., ashfall and ash-flow deposits) have been documented in the Clayton Valley deposit (Coffey et al., 2021), whereas more proximal fragmental material is ubiquitous in VS deposits from the McDermitt caldera (Benson et al., 2017b) and from the Rhyolite Ridge basin (Darin et al., in press). At Jadar, despite a much lower volume of intrabasinal igneous material, the development of the hosting basin is contemporaneous with the emplacement of two-mica granite of the adjacent Mount Cer intrusive complex (Putzolu et al., in press). In this section we present perspectives about the following subjects.

The temporal and spatial relationship between VS ores and highly fractionated igneous rocks, together with the incompatible behavior of Li in silicate melt systems (Brenan et al., 1998; Chen et al., 2020), suggests that evolved rhyolitic magmatic systems are the likely primary source of Li(B) in the VS mineralizations. However, there is a paucity of whole-rock Li and B data in the academic literature relating to the unmineralized hosting units and, where present, these compositional data should be interpreted in light of the likely high mobility of Li⁺ and B³⁺ in surface environments (e.g., Gaillardet and Lemarchand, 2018; von Strandmann et al., 2020). Limited examples of whole-rock Li values from unmineralized units in the McDermitt caldera (Ingraffia, 2020), Rhyolite Ridge (Price et al., 2000, U.S. Geological Survey, 2021; Darin et al., in press), and the Mount Cer two-mica granite (Borojević Šoštarić and Brenko, 2022) range from 40 to 480, 19 to 358, and 30 to 97 ppm, respectively. These values are similar to those reported by Hofstra et al. (2013), who demonstrated that Li concentrations greater than 100 ppm, or around three times the average upper crustal value (35 ppm ± 11; Teng et al., 2004), can be considered as enriched whole-rock values. Anomalously Li rich (>1,000 ppm) volcanic rocks are present at peraluminous S-type magmatic centers in the Andes (e.g., Tocomar and Macusani; Pichavant et al., 1988), although such rocks are rare and small volume in comparison to volcanism associated with most VS deposits.

Because of the limited number of Li data available in the literature, other more routinely analyzed large-ion lithophile elements (LILEs) or fluorophile elements (Burt et al., 1982), such as Rb and Cs, might be considered as a proxy for Li (Fig. 5). The unmineralized volcanic rocks from McDermitt show a range of Rb and Cs values of 80 to 352 and 0.5 to 629 ppm, respectively (Benson et al., 2017b; Henry et al., 2017). Igneous rocks from Rhyolite Ridge report Rb and Cs values of 90 to 442 and 1.5 to 312 ppm, respectively (Darin et al., in press), whereas the two-mica granite from Jadar has 157 to 336 ppm Rb and 5.8 to 25.5 ppm Cs (Borojević Šoštarić and Brenko, 2022). The measured Rb and Cs concentrations from these VS deposits are three times higher than the average values of the upper continental crust (i.e., Rb = 82 ppm, Cs = 4.9 ppm; Rudnick and Gao, 2003). These values are also in the range of the Rb and Cs concentrations documented in topaz rhyolites and peraluminous granites displaying an RMG affiliation (Fig. 5), as also observed by Burt et al. (1982) and Hofstra et al. (2013). Although felsic units associated with VS deposits are significantly enriched in Li and LILEs, these values are still significantly lower than the concentrations reported from the mineralized units. However, it must be noted that the Li concentrations in the considered igneous rocks are from fragmental extrusive and/or late intrusive units, which have undergone volatile phases exsolution as part of the eruptive/emplacement process. Therefore, their compositions might not reflect the original Li and LILE footprint at the time of emplacement. In this context, analysis of melt inclusions provides a window into the preeruptive endowment of Li, LILEs, and other key ligands of igneous rocks from VS deposits. Hofstra et al. (2013) demonstrated that numerous rhyolitic centers in the western United States have Li concentrations in melt inclusions several orders of magnitude higher than those measured via whole-rock geochemistry. Melt inclusions from the ignimbrites of the McDermitt caldera (Benson et al., 2017a) display Li vapor-corrected mean values of 1,436 ppm ± 205, which fall in the same order of magnitude of the Li grades of the VS mineralization in the Thacker Pass deposit (Castor and Henry, 2020; Benson et al., 2023). The study of Benson et al. (2017a) also reports melt inclusions with high F values (avg values = 5,159 ppm ± 400 F, maximum values >10,000 ppm F), which are in line with the recurrent geochemical affinity between F^– and Li⁺ in highly fractionated felsic igneous rocks. Experimental works of Webster et al. (1989) and Foustoukos and Seyfried (2007) demonstrated that during the development of large silicic volcanic centers, the abrupt drop in lithostatic pressure due to eruption enhances the preferential Li⁺ and F^– partitioning from the silicate melt into the magmatic volatile phase. This partitioning behavior offers two potential end-member pathways for the transfer of elevated Li and F into the VS mineralizing system: (1) condensation of Li-F complexes onto the surfaces of fragmented volcanic glass shards generated during the eruptive process (Vlastelic et al., 2011; Hofstra et al., 2013), which later become available for remobilization within the intracaldera lacustrine environment, or (2) Li⁺ and F^– influx into the hosting lacustrine sediments via hydrothermal fluids from continued magmatism within resurgent calderas (Benson et al., 2023).

The association of different styles of mineral deposit with specific granitoid magma types has been established across a range of ore systems (Guilbert, 1986). These associations are considered as proxies of potential melt source regions and of possible tectonic environments favorable for the formation of specific deposit types. In the same way, similarities exist between the magma types associated with VS and those responsible for formation of topaz rhyolites and RMGs (Burt et al., 1982; Hofstra et al., 2013), which have been variably attributed to S-type and A-type magmatism. However, there are some caveats to note in these assumptions, such as (1) postemplacement modification of whole-rock geochemistry due to weathering; (2) for many VS deposits, the associated igneous units include extrusive and highly fractionated pyroclastic rocks, and uncertainty may exist when comparing these rock types with intrusive granitoids (Frost et al., 2016); and (3) the specificity of differentiating between A-type and S-type magmas based on geochemistry alone (Barbarin, 1999; Bonin, 2007).

The whole-rock geochemical footprint of extrusive and intrusive igneous rocks from a range of VS districts displays several commonalities beyond the dominance of LILEs and high-SiO₂ contents (Figs. 5, 6A, B). Based on the classification framework of Frost et al. (2001, 2016) and Frost and Frost (2011), which is based on major element compositions (i.e., Fe index [Fe*], modified alkali-lime index [MALI], and aluminum saturation index [ASI]), we consider that the majority of igneous rocks in VS districts are as follows: (1) ferroan, namely those that have undergone Fe enrichment (Fig. 6A), although caution should be taken because of the relatively low concentrations of both Fe and Mg in these rocks, which would potentially increase analytical uncertainty, (2) alkalic to alkalicalcic in composition (Fig. 6B), and (3) peralkaline to strongly peraluminous (ASI > 1.1), forming a distinctive array parallel to the peralkalinity line (Shand, 1942) (Fig. 6C). Notably, the considered data set shows a geochemical decoupling from a conventional metaluminous trend to the predominant peralkaline to strongly peraluminous trends (Fig. 6C).

Trace element data provide additional insights; in the discrimination diagrams of Pearce et al. (1984), igneous rocks from VS districts plot around the triple point of syncollisional, volcanic arc, and within-plate granitoids (Fig. 7A), whereas conversely, in the 10⁴ × Ga/Al vs. Zr discrimination plot (Whalen et al., 1987) these are distributed from the I-type/S-type field to predominantly plot within in the A-type field (Fig. 7B). However, it is notable that the Mount Cer data occupy a trajectory driven by low Zr values compared to McDermitt. This distribution is more distinct in the [Zr + Nb + Ce + Y] vs. [(Na₂O + K₂O)/CaO] plot (Whalen et al., 1987), where igneous rocks from Mount Cer and Rhyolite Ridge fall in the fractionated granite field (Fig. 7C)—a footprint that is in line with the observed peralkaline to strongly peraluminous distribution noted above. Where the compositional data broadly satisfy the A-type criteria of Whalen et al. (1987), these igneous rocks clearly show similarities with A-type suites generated in a postcollisional setting (Eby, 1992). These trace element footprints compare well with the observed major element characteristics and point to the A-type/S-type magmas or more specifically peralkaline and alkaline granitoid (PAG) and muscovite-bearing peraluminous granitoid (MPG) classifications (Barbarin, 1999; Ballouard et al., 2020). This distinction becomes important due to the implicit nature of the melt source assigned by the ISA-type classification.

The convergence of the geochemistry at signatures indicative of MPGs or PAGs suggests that a commonality may exist in the nature of the melt source. Comparison of major and trace element proxies (Fig. 8A, B) shows that VS systems are predominantly associated with source regions enriched in (meta)pelitic material or have assimilated significant volumes of such material on ascent to their final depth of emplacement or eruption. This implies that the melt source region is likely to be rich in muscovite/biotite/phlogopite, which are the dominant mineral repositories for Li, Rb, Cs, and F in the continental crust and in enriched sections of the upper mantle (Romer and Kroner, 2016; Simons et al., 2016; Ballouard et al., 2023; Beard et al., 2023). This is also supported by observed Nb-Ta fractionation behavior (Fig. 8C) that suggests that magmas associated with VS deposits require partial melting mediated by muscovite to biotite breakdown, followed by effective melt extraction (Ballouard et al., 2020). This is achievable by the hybridization between felsic melts sourced from residual pelitic rocks and mafic melts derived from the lithospheric mantle. Such a mixed-source hypothesis provides an explanation for the observed continuum between PAG and MPG characteristics. Additionally, this hybridization may account for the mixed Sm-Nd isotope profiles described at the McDermitt/High Rock caldera complexes in the western United States (Coble and Mahood, 2016) and at Mount Cer in the western Balkans (Koroneos et al., 2011).

Based on the compilation of literature data, the magmatic systems associated with VS deposits have a mixed signature across the classic A-type and S-type classifications or more specifically PAG and MPG class magmas. These signatures could be generated in postcollisional to back-arc settings that would allow the introduction of primitive PAG-type magmas into mica-rich lower to midcrustal rocks.

Most VS Li deposits occur in historically arid to semiarid regions (e.g., Basin and Range district, western United States and northern Mexico) containing topographically closed basins. Previous studies have linked evaporative processes to Li enrichment and clay mineral precipitation in (paleo)lacustrine settings (Vine, 1975; Asher-Bolinder and Erickson, 1982). In these examples, Li is mobilized from the alteration of volcaniclastic rocks and transported via groundwater to a shallow lacustrine basin, then evaporative concentration of the lake water facilitates the formation of Mg(Li) clays. In this section we will provide examples of climatic mechanisms that control the Li enrichment in VS ores.

In Clayton Valley (Nevada, US), Miocene to Pleistocene lake sediments contain newly formed clay minerals such as mixed layer smectite-illite and Mg(Li) smectite, which originated in a low-temperature environment (Davis, 1981; Morissette, 2012). The initial hypothesis of Ibarra et al. (2018) and Menemenlis et al. (2022) that Pliocene and early Pleistocene warmwet conditions drove enhanced weathering and high evaporative rates, along with Li enrichment in Clayton Valley, was recently validated by Gagnon et al. (2023). These authors have reported a correlation between Li grade of mineralized units and the increase of the δ¹⁸O of the associated lacustrine carbonates as a response to the onset of arid climatic conditions. Similarly, near Hector (California, US), Li concentrations are shown to positively covary with a fractionation of heavy oxygen (¹⁸O) in lacustrine smectites and carbonates from the Miocene Barstow Formation (Gagnon et al., in press). Compilations of plant assemblages, stable isotopic data, and sedimentary evidence indicate that during the Middle Miocene, western North America was warmer and wetter and dominated by a mixed forest flora (Pound et al., 2012; Steinthorsdottir et al., 2021; Kukla et al., 2022)—a setting also supported by recent climate model simulations (Burls et al., 2021). As such, during the genesis of the high-grade McDermitt caldera Li deposit, a warm climate and evaporative concentration of paleolake waters during the warm-wet Middle Miocene Climatic Optimum (Kasbohm and Schoene, 2018; Steinthorsdottir et al., 2021) likely drove the formation of Li-rich smectite.

The early Miocene flora in Serbia shows typical subtropical associations, which lasted until the Middle Miocene Climate Optimum (Zachos et al., 2001). Obradović et al. (1997) studied the Miocene sediments of the Valjevo basin where flora typical for evergreen rainforests was found. However, some changes in the mixed mesophytic flora assemblage included records of vegetation suggestive of dry periods and related evaporitic conditions supporting a climatic seasonality. By the Late Miocene, as the Pannonian basin developed, the climate in the Balkan region evolved toward much drier Mediterranean-type conditions, although the presence of the large water bodies of Paratethys and the Pannonian influenced climate, making it locally milder and more humid (Ivanov et al., 2011). Andrić-Tomašević et al. (2021) identified Early to Middle Miocene drying stages inferred from the sediment record in the Miocene Pranjani basin, proposing that the drying stages were linked to the interplay between climate, uplift, and subsidence cycles. At Jadar, the uppermost part of the mineralized sequence includes a thick gypsum section that formed during this period of aridity in the basin. Drying periods drove the formation of the Li-rich smectites (Putzolu et al., in press), similar to the models proposed for VS basins in western United States.

Volcano-sedimentary Li(B) deposits are hosted by basins developed by tectonic or volcanic subsidence, which respectively include extensional and caldera-hosted basins. Extensional basins consist of graben to half-graben, as well as transtensional (pull-apart) basins, with typical examples located in the Basin and Range region and in the western Balkans (Fig. 9A). In most VS-hosting basins formed via tectonic processes, extension is aided by high-angle faults that account for a minor lateral offset, therefore resulting in orebodies with a limited lateral variation in terms of sedimentary facies and potentially of grade distribution (Fig. 9A, B). An exception to this model is the Jadar deposit, which can be classified as a less common subtype of the extensional basins. Here, the volcano-sedimentary sequence is hosted by a buried basin that experienced a greater rate of extension. At Jadar, high extension rates were aided by a low-angle detachment originated through reactivation of older thrust structures in the pre-Miocene basement. Higher extension rates in the Jadar basin were accommodated by listric faulting and resulted in a greater lateral offset and in a significant subsidence of the basin that was overlain by marine sediments (Fig. 9A, C), which accounted for the effective preservation of the system via burial. On the other hand, intense extension also caused a significant uplift and erosion of the surrounding country rocks that resulted in the input of large volumes of exotic nonlacustrine detritus and in the development of thick intrabasinal epiclastic units (Putzolu et al., in press). Furthermore, the extensional processes through low-angle faults originated a greater lateral offset of the volcano-sedimentary units, thus resulting in a complex distribution of mineralized zones (Fig. 9A, C).

Caldera-hosted basins are a much rarer category, with the best examples being the VS deposits in the McDermitt caldera (i.e., Thacker Pass and McDermitt), Kırka-Phrigian (Turkey), and Shavazsay (Uzbekistan) (Helvacı et al., 2021; Benson et al., 2023; Dolgopolova et al., 2023). Caldera-type basins form during eruption of high volumes of felsic volcaniclastic to pyroclastic rocks and concomitant collapse of the magma chamber roof. The major structures in caldera-hosted systems include peripheral ring faults originated during multiple collapse events, as well as interior radial faults linked to later caldera resurgence (Cole et al., 2005) (Fig. 9D). In this setting, lacustrine sedimentation and Limineralizing processes can be partially synchronous to the events of caldera resurgence. The resulting radial faulting can contribute to a significant segmentation of the orebodies in discrete faulted domains and can also serve as conduits for ascending higher-temperature mineralizing fluids (Fig. 9D). For example, lacustrine sediments in the Thacker Pass deposit show fault block rotation and tilting of the sedimentary layering that are due to the offset along ring faults originated by the late doming of the caldera during resurgence (Ingraffia, 2020).

Lithium reserves in VS deposits are predominantly hosted in phyllosilicates, yet more unconventional ore minerals such as borosilicates and phosphates can also be enriched in Li (App. 3). Volcano-sedimentary deposits can contain Li-free borates and borosilicates, which are themselves valuable mineral commodities and can add value to a prospect. In this section we will describe the mineralogy, textures, and minerogenesis of Li minerals. Key features of boron minerals in VS deposits will be also provided.

Trioctahedral smectite-group clays generally account for the most frequent style of Li mineralization in VS deposits. Smectite clays represent a low-grade mineral, with Li concentrations of 0.5 wt % Li reported from the pioneering work of Foshag and Woodford (1936). Recent studies reported Li concentrations around 0.4 to 0.5 wt % Li for smectites from Jadar (Putzolu et al., in press) and Thacker Pass (Benson et al., 2023) deposits and up to around 0.7 wt % Li for smectites from Rhyolite Ridge (Benson et al., 2023, and references therein). Hectorite and its F-bearing counterparts are the well-known Li(F)-rich end members of the Mg smectite group and have been traditionally reported in VS deposits to describe clay-rich interceptions with high Li(F) concentrations. However, smectites form intimately intergrown aggregates and are rarely compositionally pure; therefore, their identification based on field observations, even if corroborated with geochemical assay, is often not confirmed by laboratory-based characterization by X-ray diffraction (XRD) and electron beam analyses. Several studies have confirmed that a subordinate Li fraction can be enriched in other Mg-rich smectites such as saponite and stevensite and even in Mg-poor species as nontronite and montmorillonite (Tardy et al., 1972; Kadir et al., 2023; Putzolu et al., 2023a). Lithium smectites from the McDermitt caldera, Jadar, and western Anatolia deposits form the better-documented occurrences of this ore type (Castor and Henry, 2020; Benson et al., 2023; Kadir et al., 2023; Putzolu et al., 2023a, b, in press; Emproto et al., in press). In the McDermitt caldera (i.e., Thacker Pass and McDermitt deposits), the Li claystones occur as varve-like laminites that contain Li clays interbedded with carbonates, altered ash, and silica layers (Fig. 10A). In this ore type, Li smectite forms late infills of porosity within newly formed silica and calcite (Fig. 10B). Li-rich claystones often show soft-sediment deformation textures, such as displacive boundaries developed via growth of carbonates nodules and/or silica pods (Fig. 10C) and by segregation of secondary silica layers (Fig. 10D). Epigenetic textural features, such as erosive and mud-crack networks, are often superimposed onto the Li smectite horizons (Fig. 10E). In clay-type ores, Li smectites display a close association with secondary silica—an assemblage that results in the development of poker chip textures made of smectite-silica intercalations with soft-sediment deformation features such as sedimentary loading and dewatering (Fig. 10D, E). In the Jadar deposit, Li smectites form massive pods and lenses within lacustrine sediments (Fig. 10F). In these units hectorite-like clays either occur as pseudormorphs after relict shard morphologies or as clustered newly formed nodules (Putzolu et al., in press).

Tainiolite is a rare trioctahedral Li-Mg illite with maximum Li concentrations of 1.39 wt % Li (Emproto et al., in press). The best-documented occurrence of tainiolite is in the southern portion of the McDermitt caldera (i.e., Thacker Pass deposit), where it forms a fine-grained pervasive replacement of earlier Li smectites mostly along lithological boundaries and fractures (Benson et al., 2023) (Fig. 10G). Tainiolite has been also reported for older and likely metamorphosed VS deposits, as the Permo-Triassic Li(B) mineralization of Val Tanaro (Italy), where it forms well-developed and pure crystals associated with recrystallized anhydrous borosilicates such as danburite (CaB₂(SiO₄)₂) (Dini et al., 2022).

Cookeite is an Al-rich dioctahedral chlorite with Li contents of approximately 0.9 to 1.4 wt % Li. Although cookeite is not a typical ore mineral in recent and conventional clay-type VS deposits such as those in the Basin and Range district, it represents the main Li repository in ancient aluminiferous Li-rich claystones, where it forms a late and fine-grained replacement of kaolinite-rich soils formed via supergene alteration (i.e., bauxitization) of Li-rich tephra units (Ling et al., 2023). Typical examples of this deposit type are the Permian aluminiferous Li-rich claystones from China (Ling et al., 2023) and the Carboniferous clay districts from Missouri and Pennsylvania (Fig. 3; Keller and Stevens, 1983; Feineman et al., 2020; Rozelle et al., 2021). Li-rich minerals of the true mica group have been documented for a few VS deposits such as at Sonora (Mexico; Verley, 2014) and Shavazsay (Uzbekistan; Golovanov, 2001). These minerals include trioctahedral micas of the zinnwaldite-polylithionite series and have Li concentrations around 2.9 to 4.2 wt %. It is worth mentioning that no relevant information about the textural features of these minerals in VS systems is available in the literature, yet in both mentioned deposits the origin of Li micas appear to predate the secondary assemblages (i.e., smectite clays, zeolites) and it is likely linked to input of Li-rich material inherited from primary igneous units.

Jadarite is the sole example of a Li-rich borosilicate, with Li and B contents of about 3.3 and 14.6 wt %, respectively (Stanley et al., 2007). Jadarite textures are largely controlled by the variation of the host-rock type. In lacustrine shales jadarite occurs as stratabound set of nodules that locally develop enterolithic textures (Fig. 11A, B, respectively), whereas in altered tuff layers it forms pseudomorphs after shard morphologies and/or clusters of spherulites (Fig. 11C) (Putzolu et al., in press).

Lithium-rich phosphates are a rare ore mineral in VS deposits and include lithiophosphate, which has Li concentrations of about 17.2 to 18 wt % Li. The only documented occurrence of lithiophosphate is in the Jadar deposit, where it forms fine-grained (around 100 µm) infills in the carbonate host rock, as well as late coating onto jadarite nodules (Putzolu et al., in press).

Boron mineralogy in VS deposits includes the borosilicate and borate groups. The most common mineral of the former group is searlesite, which is a hydrous Na borosilicate that can represent a major boron repository in most recent VS deposits such as in the North America (e.g., Rhyolite Ridge and Kramer-Boron deposits) and Tethyan (western Balkan and western Anatolia) districts (García-Veigas and Helvacı, 2013; Helvacı, 2019; Reynolds and Chafetz, 2020; Putzolu et al., in press). It is worth mentioning that jadarite is the only mineral species that falls in both Li and B mineral repository groups of VS deposits. Other important borosilicates are reedmergnerite and danburite, which are anhydrous minerals that occur in older and likely metamorphosed B-rich VS deposits, with representative examples in the Junggar basin (China; Guo et al., 2021) and Val Tanaro (Italy; Dini et al., 2022). Borate mineralogy is extensively variable; however, the most recurring and abundant minerals that in VS deposits play a role in developing economic resources are Na- (e.g., borax, ezcurrite, and kernite), Na(Ca)- (e.g., probertite and ulexite), and Ca-rich (e.g., colemanite) species. A more comprehensive description of the mineralogy of borates, including minor to trace and more exotic species, is provided by Helvacı et al. (2012).

Most of the textural types observed for borates and borosilicates are similar to those of evaporitic minerals (Warren, 2016) and are (1) primary precipitates, which occur as isooriented crystals precipitated at the sediment-water boundary (i.e., bottom-growth type;, Fig. 11D, E) or as sets of discrete grains that are deposited via gravitation settling following precipitation from a brine (i.e., cumulate-type; Fig. 11D, F), or (2) secondary precipitates, which form nodules originated via nucleation processes from pore waters seeping within the sedimentary package (Fig. 11D). Formation of this textural type typically triggers a soft-sediment deformation in the host rock (Fig. 11D, G). Other textural types of borates have been recently described in the Jadar Li-B deposit (Putzolu et al., in press), where jadarite and Na borates also occur as (1) direct replacements, which are found in volcaniclastic rock types either as alteration fronts or a spherulitic aggregates, or (2) breccia-type, a borate/borosilicate type that forms late and crystalline infills of fractures within brecciated lacustrine sediments. It is worth noting that wheras these textures have been traditionally described for Li-free boron species such as searlesite and Na to Ca borates, the recent work of Putzolu et al. (in press) has shown that the variation of jadarite textures cover most of the above textural types.

Gangue mineralogy in VS deposits is subdivided in relict phases inherited by the intrabasinal igneous units and in secondary mineral that are synchronous to the Li(B) ore assemblage. Relict primary phases include pyroxene, amphibole, volcanic glass, mica, and Na- and K-feldspar. Except for micas, most primary phases are heavily affected by alteration processes linked to the ore-formation events, thus their identification in drill cores is challenging. The secondary gangue mineral assemblage is largely dominated by aluminosilicates, such as (1) Na to Ca(Na) zeolites, which occur as nodular clusters similar to jadarite (Fig. 12A), (2) secondary feldspars (i.e., mostly K-feldspar) that are commonly observed as the residuum of alteration of volcaniclastic rocks (see Putzolu et al., in press) and occur as massive, fine-grained, and pale-orange units interbedded in the lacustrine sediments (Fig. 11C). Carbonate minerals as dolomite, calcite, and ankerite are ubiquitous and either occur as micritic lacustrine sediments (Fig. 11A) or as late aggregates such as nodules or rosettes (Fig. 12B-D). Several generations of secondary silica have also been documented, such as (1) rims and/or beds closely bounded with the Li(B) ore, which is likely synchronous to the mineralization event (Fig. 10D); (2) late banded veins that crosscut the basement ignimbrites or lavas in the lowermost section of the basins (Fig. 12E); this silica type is likely formed via circulation of geothermal fluids during the emplacement of late lava dikes/sills; and (3) disseminated silica with fine-grained oxide coatings occurring in the upper section of the basins (Fig. 12F); this silica type is likely due to the supergene alteration of the uppermost volcaniclastic units.

Diagenetic alteration of intrabasinal igneous rocks is considered as a key process for the development of the secondary Li(B) mineralization in VS deposits. Variations of textural features of the Li(B) minerals in VS deposits can be explained through the variable involvement of precipitation from solution (neoformation) and ore formation via alteration of precursor intrabasinal igneous units (solid-state transformation). For example, Benson et al. (2023) have observed that Li-rich claystones in the Thacker Pass deposit often form sharp contacts with altered intrabasinal ash and tuff units and are commonly overprinted by newly formed carbonate rosettes within set of fractures. According to these features, a viable process for the Li smectite formation would involve the direct clay precipitation from lake waters with high pH and high silica activity. In this context, precipitation of hectorite-like clays rather than a more conventional Mg clay assemblage is due to the high F^– and Li⁺ in the lake waters sourced either from the leaching of fertile igneous rocks or from fluids exsolved during the cooling of the intracaldera ignimbrites (Ingraffia, 2020; Benson et al., 2023; Emproto et al., in press) (Fig. 13A, B). Other authors have observed that Li(B) ore-formation processes are also due to direct replacement of intrabasinal igneous rocks. For example, Li-rich horizons in the McDermitt caldera, and in the Jadar and Bigadiç deposits, locally preserve relict volcanic glass morphologies and are associated with the development of extensive zeolitization and K-feldspar alteration (Kadir et al., 2023; Putzolu et al., 2023a, in press). The formation of the zeolite-feldspar assemblage from volcanic glass has been traditionally reported for altered Li-free volcaniclastic units in hydrologically restricted basins and is described through the closed-hydrologic system diagenesis (CHSD) model (Langella et al., 2001, and references therein). The CHSD model was first used by Castor and Henry (2020) to describe the mineralogical evolution of clay-type VS deposits and recently adapted by Putzolu et al. (in press) to borosilicate-type Li(B) ores. During CHSD, interaction of volcanic glass-rich rock types with pore waters increases the Na⁺ concentration in the fluids via the following reaction:

Increasing Na⁺ raises the pH of the system, whereas OH^– ions act as a hydrolysis agent and decompose the Si-Al-O bonds of the volcanic glass framework (de′ Gennaro et al., 1988). This process results in the formation of metastable silica gels, which crystallize Mg smectites and Na zeolites as the system evolves from near-neutral toward higher pH and Na⁺-saturated conditions. During its latest development stage, as Mg²⁺, Ca²⁺, Li⁺, and Na⁺ are fractionated in early smectites and zeolites, the residual pore fluids are indirectly enriched in K⁺ and promote the formation of authigenic K-feldspars (Langella et al., 2001, and references therein; Castor and Henry, 2020; Putzolu et al., 2023b, in press). This minerogenetic pathway could explain some of the textures and paragenetic relationships observed in clay- and borosilicate-type VS deposits. For example, in clay-type VS deposits (e.g., McDermitt deposit, Oregon), Li smectites are associated with nodular Na zeolites (i.e., analcime) that locally envelop relict shard morphologies (Fig. 13C, D). Here, paragenetic relationships support that Na zeolites form shortly after smectite, and thus they likely mark a minerogenetic stage when smectite formation is suppressed in favor of sodic alteration linked to zeolitization. Furthermore, evidence for the existence of a gel precursor is given by the presence of nodular silica-zeolite-hectorite aggregates that develop from stratiform secondary silica layers interbedded within altered intrabasinal tuffs (Fig. 13E-H). These nodules show wavy morphologies that likely reflect the plastic movements of the colloids/gels at the sediment-water boundary. Segregation of these silicate gels was likely due to the SiO₂ excess associated with the alteration of rhyolitic glass (~70 wt % SiO₂) to smectite and/or Na zeolites (maximum ~55 wt % SiO₂). It is worth noting that the application of the CHSD model to explain some of the features of VS deposits indicates that hectorite-like ores would form under near-neutral to slightly alkaline pH conditions, shortly before the system is saturated in Na⁺ and thus in the zeolite component. Conversely, the close paragenetic association between Na zeolites and jadarite observed by Putzolu et al. (in press) implies that borosilicate-type ores would need a significantly higher pH of mineralizing fluids. A key implication of the CHSD model is that VS deposits evolve under physicochemical conditions that allow the effective alteration of volcanic glass. As described by Hall (1998), if specific conditions are not met, volcanic glass can be refractory to alteration and thus can be preserved in the geologic record. These conditions can be applied to VS deposits as physicochemical gates that a system should pass through to attain ore fertility and can be outlined as follows:

1. Moderate fluid/rock ratio: Volcanic glass is generally stable in absence of fluids; therefore, the presence of lake and pore water is pivotal to enhance the fluid-rock exchanges and the release of Na⁺, SiO_2(aq), and Li⁺ to form ore-bearing assemblages. This could explain the presence of relatively preserved volcanic glass-rich interceptions at the top of some VS deposits (McDermitt deposit, Oregon; Putzolu et al., 2023a), likely marking the end of the lacustrine cycle of the basins.

2. Presence of a heat source: The rate of alteration of volcanic glass to aluminosilicates has been shown to be correlated with the temperature of the system. High temperatures favor the kinetics of breakdown of Si-O-Al bonds in the volcanic glass framework. The high rate of this process explains textural features in VS deposits that support formation of Li(B) minerals during early diagenesis, which likely plays a primary role in aiding the fixation of highly mobile elements (Li⁺ and Na⁺) before they are leached from the basins. This also implies that distal volcanic glass-rich tephra units deposited away from volcanic centers are unlikely to form significant VS deposits, consistent with what has been demonstrated for Li brine deposits (Ellis et al., 2022).

3. High pH: Volcanic glass is stable under acidic conditions, and it is known that pH of fluids in volcanic terranes can remain relatively low because of inherited volatiles (i.e., HCl, HF, and SO₂) linked to ongoing volcanic activity. Therefore, the increase of pH of the system is pivotal to enhance the minerogenetic processes to form mineralized assemblages. Arguably, the main pH buffering agents in VS deposits are the evolved composition of volcanic glass itself, which can release Na⁺ and K⁺ to the system, and the high HCO^3– provided by exsolving volcanic-magmatic fluids.

Early diagenesis explains most of the features observed in a number of VS deposits, yet these models do not consider that in few deposits the early Li(B) ores are affected by postformation processes. The primary Li(B) assemblage in VS deposits includes mineralogical phases that are stable at low P-T and under near-neutral to high-pH conditions. Therefore, if aging of a VS system is characterized by changes in the P-T-x of the surrounding geologic environment, the primary ore can be subjected to late mineralogical changes that have a key role in the ore distribution. The best-documented postformation processes in VS deposits are mostly related to the increase of the temperature of the system either via circulation of hydrothermal fluids or after burial of the basins. Benson et al. (2023) recently described the role of the late circulation of hydrothermal fluids on the Li ore redistribution in the Thacker Pass deposit. Circulation of hot (~300°C) hydrothermal fluids enriched in Li⁺, K⁺, and F^– occurred after deposition of the earlier Li smectite ore during the stage of caldera resurgence, leading to the epigenetic illitization of smectite and to a 4× upgrade of the initial Li concentrations (Fig. 13A, I). It is worth mentioning that Li illite (tainiolite-like) has been also observed in VS deposits with no evidence of a late hydrothermal overprint via caldera resurgence.

An example is Val Tanaro (Italy), where Li illite is found in Li(B)-rich beds showing recrystallization features and having an anhydrous borosilicate assemblage. Dini et al. (2022) observed that Li illite is found in rocks that have experienced greenschist facies conditions; therefore, despite lacking studies on this, it should be considered that the metamorphism of an Li smectite-rich VS deposit could be another pathway to redistribute Li in higher-grade and more crystalline phyllosilicate assemblages.

Recent work by Putzolu et al. (in press) has shown that burial and/or metamorphism of VS deposits that contain high levels of organic material can trigger an increase of the thermal rank of the system to attain the oil generation window. This process is associated with formation of epigenetic bitumen and with a consequent acidification of pore fluids. The late pH retrogression is shown to have a strong control on the Li(B) residency in the Jadar basin, as it triggers the dissolution of the early jadarite-bearing assemblage and the Li and B remobilization to form late diagenetic lithiophosphate and Na-free borates.

A late fluid circulation during burial diagenesis has been also inferred for the origin of the cookeite-bearing claystones from southwest China (e.g., Pingguo area, Guangxi). In these unconventional ores, the basins initially contained kaolinite, which during low-grade metamorphism triggered by burial was subjected to illitization and chloritization. As observed by Ling et al. (2023) detrital kaolinite likely derived from the bauxitization of Li-fertile felsic volcaniclastic rocks, and thus was characterized by unusual Li concentrations that led to the formation of cookeite rather than more conventional chlorite-group minerals.

Mineralogical zoning in VS deposits can be observed at regional, district, and deposit scales and typically consists of variation in the relative endowment of boron vs. lithium resources.

A regional zonation of the mineralogy of VS deposits is documented in western United States, whereby VS mineralization includes clay-type ores (e.g., McDermitt caldera, Clayton Valley, and Hector) and clay/borate-type ores (e.g., Kramer-Boron and Rhyolite Ridge). Clay-dominated deposits are mostly found along the northern and western boundaries of the Basin and Range province (e.g., Nevada North, McDermitt caldera), whereas a general increase of borate contents associated with clays has been observed in deposits located within the adjacent Walker Lane transtensional system. However, despite this overall regional zoning, VS deposits showing clay-only and clay/borate-style characteristics coexist in the same area, such as in the Clayton Valley area (e.g., Rhyolite Ridge and Clayton Valley) and the central Mojave Desert (e.g., Kramer-Boron and Hector). Additionally, Li enrichment in smectite clays is spatially associated with colemanite mineralization also in the Miocene Barstow Formation near Old Borate (Benson, 2023) in the Calico Mountains (California, US) and the Miocene Horse Spring Formation (Landsem et al., 2023) in the White basin and Anniversary mine areas (Nevada, US), although the paragenetic relationship and sequencing of the Li and B mineralization in these deposits has not yet been established.

A zoning of the Li vs. B mineralogy can be observed on a district scale as in the western Balkans, where the zonation can be tracked along the SE-trending Sava-Vardar zone. Deposits in the northwestern limb of the Sava-Vardar zone show mixed Li and B resources, whereas deposits in the southeastern limb are boron prevailing (Borojević Šoštarić and Brenko, 2022, and references therein). A zonation of the Li mineralogy can be observed based upon the proximity to the associated granitoid plutons. Lithium resources in deposits distal to plutons are mostly in clays (e.g., Valjevo, Pranjiani, Rekovac, and Jarandol), whereas Li is abundant in borosilicates (i.e., jadarite) in lacustrine units of the Jadar deposit, which developed proximal to the Mount Cer granite complex (Borojević Šoštarić and Brenko, 2022; Putzolu et al., in press).

Deposit-scale zoning patterns locally consist in a lateral variation of ore and gangue mineralogy that produces facies heterotopies within the basin. The best examples of lateral zoning are documented for the Jadar deposit, a few VS ores in Turkey (e.g., Göcenoluk, Kırka and Bigadiç), and the western United States (Kramer-Boron) (García-Veigas and Helvacı, 2013; Bowser, 2015; Warren, 2016; Putzolu et al., in press). The lateral zoning in these deposits has been ascribed to stages of increasing degrees of evaporation and pH of mineralizing fluids (Fig. 14A). This process enhances the precipitation of mineralogical assemblages stable under near-neutral pH in the lake marginal facies, which can include Li smectites, Ca zeolites (e.g., clinoptilolite, heulandite) and Ca borates (e.g., colemanite). The marginal units are generally those showing the least texturally destructive alteration facies and that thus locally preserve relict volcaniclastic rock types interlayered with the lacustrine sediments and with ore-bearing minerals. As evaporation increases, the migration of the water-filled portion of the lake toward the center of the system leads to the formation of assemblages stable under higher pH conditions, such as Na borates (e.g., ulexite, borax, kernite), Na borosilicates (e.g., searlesite and jadarite), and Na zeolites (e.g., analcime and natrolite), and eventually of secondary K-feldspar (Fig. 14A). This lateral mineralogical zoning partially reflects the closed-hydrologic system diagenesis (CHSD) model and implies that fractionation of specific elements in the earlier formed assemblage is a first-order control on the chemistry of the residual fluids. In some VS deposits this pattern leads to a spatial decoupling of Li- and B-mineralized facies, as Li smectites would be concentrated in the more marginal facies, whereas the Na borate ore would be more abundant in the lake center sections of the system. An exception to this is the Jadar deposit, where despite the lateral zoning of the Li smectite vs. borate zones, the borate zone in the lake center contains significant Li concentration in the borosilicate jadarite (Putzolu et al., in press). It is worth noting that this model does not consider the following:

1. Increasing evaporation and pH increase in lacustrine systems is seldom a unidirectional and linear process. For example, in the Jadar basin mineralogical assemblages that formed at high pH (i.e., borosilicates) are locally overprinted by lower-pH assemblages (e.g., Ca borates) (Putzolu et al., in press). This pH retrogression process is to be expected, as lacustrine basins can be characterized by domains with a more open hydrological regime, where high evaporation rates and alkaline pH of fluids are not attained owing to episodes of input of less saline meteoric fluids.

2. Some VS deposits are characterized by a vertical mineralogical zoning caused by the upwelling of higher-temperature fluids. An example of the latter scenario is the Thacker Pass deposit in the McDermitt caldera where the circulation of hydrothermal fluids has led to the development of Li illite in the lowermost section of the orebody and a preserved earlier Li smectite at the top of the sequence (Castor and Henry, 2020; Benson et al., 2023). Illitization of Li smectite led to a complex vertical evolution of the Thacker Pass deposit, which is characterized by a transition zone that contains illite-smectite mixed layers, between the upper Li smectite and the lower Li illite zones (Castor and Henry, 2020; Benson et al., 2023; Emproto et al., in press) (Fig. 14B). This transitional facies is likely to be an alteration halo that developed as the illite-forming fluid was progressively depleted in F⁻, K⁺, and Li⁺ while overprinting the smectite-rich horizons. Circulation of the illite-forming fluid also led to a vertical evolution of the gangue mineralogy as it dissolved the early analcime-dolomite association and led to extensive potassic to sodic alteration (i.e., secondary K-feldspar and albite) accompanied by silicification at the bottom of the basin (Fig. 14B). As shown by Castor and Henry (2020) the vertical mineralogical zoning in Thacker Pass deposit is also mirrored by anomalous concentrations of pathfinder elements (e.g., Mo, S, Rb) within the abundant illitized units (Fig. 14B). Similar elemental anomalies have been observed in the Oregon limb of the McDermitt caldera (i.e., McDermitt deposit) and are associated with relatively lower temperature (<200°C) hydrothermal systems (Benson et al., 2023), yet no significant vertical zonation of the Li clay mineralogy has been observed (Putzolu et al., 2023a). As discussed by Benson et al. (2023) this caldera-scale zonation of the Li clay mineralogy is due to a preferential focusing of the fluid flow in the southern limb of the system that was much more proximal to the center of postcollapse resurgence.

Based on the available information on globally significant deposits (Fig. 3; App. 1), VS ores can be classified as follows:

1. Clay-type: This VS-style mineralization is characterized by high volumes of Li phyllosilicates and by a lack of boron minerals, which results in orebodies with high Li/B ratios. Based on Li clay mineralogy, most clay-type VS ores range between smectite and illite end members (Fig. 15A, B; App. 4). Archetypal examples of smectite-rich end members ores are the McDermitt caldera, Clayton Valley, and hectorite deposits of the western United States. However, discrete mineralized units showing smectite-type characteristics are observed in other VS deposits such as Jadar (Serbia), Bigadiç (Turkey), and Rhyolite Ridge (western US). Based on the geochemistry of orebodies, Li smectites from the McDermitt caldera, Rhyolite Ridge, Hector, and Jadar can be classified as pure hectorite ores, whereas those from Bigadiç, Clayton Valley, and other VS occurrences from Western United States (i.e., Horse Spring, Barstow Formation, and Burro Creek) show a mixed footprint pointing out to the occurrence of both hectorite and lower-Li-grade Mg smectites such as saponite and/or stevensite (Fig. 15A, B). Examples of illite-rich end members are the Thacker Pass (McDermitt caldera, western US) and Sonora (Mexico) deposits, where geochemistry of mineralized units tends toward high Li grades following an illitization trend (Fig. 15A, B). A subset of data from Rhyolite Ridge shows a minor offset from the nominal hectorite line, indicating that this VS deposit experienced minor Li upgrade through illitization. (Fig. 15A, B). Lithium chlorite (cookeite)-rich deposits are a less common end member of clay-type VS, with typical examples in southwest China in the Yunnan, Guangxi, Shanzi, and Henan districts, as well as in Missouri and Pennsylvania (US). Noteworthy, some clay-type VS deposits have a more complex mineralogy that includes phyllosilicate mineral types formed under a wide P-T-x range such as high-temperature (i.e., magmatic to hydrothermal) Li-F micas and lower-temperature clays.

2. Clay/borate-type: This VS type has been typically targeted for boron in the form of borates and borosilicates and only recently has been studied for its potential to host economic volumes of Li mineralization. Since this ore type contains large volumes of borates, it can be qualitatively described as a VS mineralization with low Li/B ratios. Lithium in this VS ore type is mostly associated with Mg(Li) smectites. Archetypal examples of clay/borate-type ores are found in the western United States (e.g., Rhyolite Ridge and Kramer-Boron deposits) and in western Anatolia (e.g., Bigadiç deposit).3Borosilicate-type: This VS ore, also called “jadarite-type,” has been only discovered recently and is characterized by high Li and B concentrations and by a mineralogical Li-B coupling in the mineral jadarite. The only documented example of borosilicate-type VS is the Jadar deposit in Serbia (Putzolu et al., in press).

Some minerals in VS deposits can also contain Li-bearing phosphates (e.g., lithiophosphate). However, a phosphatetype VS ore has not been included in this classification as this Li mineral class has only been documented in specific facies of the Jadar deposit, and to the best of our knowledge is unknown in other deposits of the VS class. Furthermore, this classification does not consider that some VS deposits can contain Li-F micas such as polylithionite and zinnwaldite. This type of Li repositories was not considered, as their occurrence in some deposits is likely due to the contribution of relict volcanogenic rock types (i.e., rhyolites, ongonites) and thus cannot be reconciled with the diagenetic-hydrothermal processes accounting for a VS-style mineralization.

The McDermitt and Thacker Pass deposits (McDermitt caldera, western US; App. 5, Fig. A1A, B) best represent the features of clay-type VS ores of the smectite and illite end members. The McDermitt caldera occurs at the Oregon-Nevada state border, covers an area of about 40 × 30 km, and formed around 16.39 ± 0.02 Ma on eruption of the peralkaline, Li- and F-rich McDermitt tuff (Benson et al., 2017a; Henry et al., 2017). The caldera collapse triggered the formation of geomorphological conditions to form ephemeral and hydrologically closed lacustrine basins on top of the McDermitt tuff, resulting in the deposition of thick Li-rich smectitic sediments and intercalated volcaniclastic rocks (Benson et al., 2023). Li-bearing intracaldera sediments are relatively well exposed on the northern limb of the system (i.e., McDermitt deposit), whereas they are covered by alluvium and lava flows in the southernmost portion of the caldera (Thacker Pass deposit). Discovery of Li-rich clays at McDermitt dates back to the 1970s, when local mineral surveys conducted by Chevron and the U.S. Geological Survey targeted hydrothermal uranium ores.

McDermitt: The McDermitt deposit is located in the Oregon portion of the caldera, north of the McDermitt River. The Li clay orebody locally crops out to form pale-tan escarpments. At depth, the succession consists of clay-rich intercepts interbedded with secondary silica horizons. Evidence in drill core and at surface indicates the input of both airfall material as tuffaceous horizons and pyroclastic/volcanoclastic flows that are locally preserved within the orebody. Clay-rich horizons show high Li concentrations, which are locally correlated with anomalies of Mg, Rb, Sr, Mo, and Cs (Putzolu et al., 2023a). Hectorite-like smectite is the main Li-bearing mineral observed in the clay-rich horizons, whereas the main gangue minerals include zeolites (analcime, heulandite, and clinoptilolite), amorphous silica, and calcite (Putzolu et al., 2023a). The clay-bearing succession occurs on top of a welded and hydrothermally altered facies of the McDermitt tuff or locally is crosscut by intermediate lava sills and banded silica veins. Textures indicating pseudomorph replacement of volcanic glass by smectite, zeolite, and K-feldspar have been observed and support that direct replacement the intrabasinal volcaniclastic units can partially account for the Li enrichment process. However, like in the Thacker Pass deposit, clay-rich units locally form sharp contacts with volcaniclastic rocks; therefore, a direct clay precipitation from lake waters also contributed as part of the ore-forming processes.

Thacker Pass: The Thacker Pass deposit occurs in the southernmost portion of the McDermitt caldera. The most recent models for the origin of the Thacker Pass deposits were proposed by Castor and Henry (2020), Benson et al. (2023), and Emproto et al. (in press). Castor and Henry (2020) proposed that the Li clay formation in the Thacker Pass deposit can be explained by the alteration of volcanic glass via closed hydrologic system diagenesis, yet this model cannot easily be reconciled with the formation of the Li illite mineralization in the higher-grade zone of the ore. On the other hand, Benson et al. (2023) and subsequent work by Emproto et al. (in press) proposed a model that involves an initial process of clay precipitation in a shallow lake environment under high pH, which led to the development of a smectite-type Li ore, followed by influx of high-temperature resurgence-related fluids in the vicinity of Thacker Pass that led to illitization of smectite. An in-depth description of the overall features of the stratigraphy of the Thacker Pass deposit is provided by Ingraffia (2020), and the mineralogy of the deposit is further expanded in detail by Emproto et al. (in press). The deposit, from bottom to top, consists of the basal McDermitt tuff, a thin zone of silicified ash, and finely laminated lacustrine sediments with interbedded tephra, volcaniclastic sediments, and rhyolite and basaltic lavas. The bulk of the lacustrine sedimentary section is composed of the relatively low grade (up to 0.4 wt % Li in whole rock) Li-rich Mg smectite that transitions at depth to an Li-rich illite (up to 0.9 wt % Li in whole rock; Benson et al., 2023). Lithium concentrations follow the transition from smectite to mixed-layer clays to illite and then back toward a mixed mineralogy under the most illite- and Li-rich zone (Emproto et al., in press). The upgrade of the Li content in the illite zone correlates with the concentrations of Rb, Cs, Mo, K, F, and As and matches with an increase of the fluorite and sulfides abundances and with a noticeable lack of zeolites. These characteristics have only been observed at Thacker Pass and in the sediments immediately to the north in the Montana mountains (Benson et al., 2023).

Pingguo: A typical example of chlorite-rich clay-type VS deposit is in the Pingguo area (Guangxi region) of southwest China. Here, high Li concentrations (around 0.03–0.5 wt % Li) are found in cookeite from claystones contained in the Heshan Formation. The Heshan Formation includes a basal black bauxite overlain by the Li-rich claystone, which were deposited over karstified middle Permian limestones of the Maokou Formation. Recent models for the formation of this type of ore contrast with those proposed for the formation of other types of clay-rich ores. Ling et al. (2023) has inferred that the Li concentrations in karst networks were associated with kaolinite that formed via bauxitization of bimodal volcanic rocks of the nearby Emeishan large igneous province (263–251 Ma). Following burial, kaolinite from the residual soil units underwent solid-state transformation to cookeite through circulation of high-pH and Al(Li)-rich late diagenetic fluids.

Shavazsay: The Shavazsay deposit is located in the Akhangaran district of the Tashkent region (App. 5, Fig. A2A, B) and best represents the features of clay-type VS endowed of a variable phyllosilicate suite formed under different stages of Li-fertile magmatic, hydrothermal, and diagenetic events. The Shavazsay deposit was discovered in 1969 and was subjected to Li exploration from 1975 to 1995. The Shavazsay deposit is part of the Chatkal ridge, Middle Tienshan continental arc—a geologic terrane that experienced subduction processes until the late Carboniferous. Magmatism in the area is associated with Permian postsubduction collision, which originated volcano-plutonic complexes that include intermediate plutonic rocks and leucogranites and their extrusive equivalents (i.e., rhyolite, trachyrhyolite, and trachyandesite). The volcano-plutonic suite extends over 100 km with several caldera structures and volcano-tectonic depressions. The Chilten caldera is the main host for Li-bearing VS mineralization and includes the Shavazsay deposit and the North Shavaz and Kamyshly occurrences. The caldera complex includes upper Devonian and Carboniferous volcanic and intrusive rocks and is filled by volcano-sedimentary rock types that comprise from bottom to top (1) basal conglomerates, (2) rhyolite lavas, tuffs, and volcaniclastic breccias, (3) Li-rich claystones of the lower Permian Oyasay Complex; these consist of stratiform horizons up to 350 m thick and are characterized by extensive silicification and elevated organic contents (i.e., up to 4% carbonaceous to bituminous matter), and (4) upper basalt unit, which caps the system on a basin scale. This volcano-sedimentary sequence is crosscut by dikes of topaz rhyolites (“ongonite,” Li-F micro-granite) and trachyrhyolites. The Shavazsay VS mineralization originated during a syn- to postcaldera stage and under an extensional regime and was likely synchronous to Li-F–rich magmatism and collapse associated with the caldera development. The ore-bearing horizons of the Shavazsay deposit form two large outcrops (Central Shavaz and Ashibuzuk) at a relatively shallow depth (up to 200 m) and, like the Thacker Pass deposit, are associated with deposit-scale anomalies of Rb, F, B, Mo, and Cs. The Li grade at individual sections varies between 0.14 and 0.37 wt %, exceeding 0.9 to 1.4 wt % Li in mica metasomatites. Lithium in the Shavazsay deposit shows a complex metal deportment due to the presence of various types of host rocks (i.e., volcaniclastic to sedimentary) (Dolgopolova et al., 2023). Preliminary analyses have revealed that Li is contained in both primary Li-F micas, such as polylithionite and phengite, and in secondary phyllosilicates such as illite and smectite clays (tainiolite- and montmorillonite-like, respectively). Whole-rock K-Ar geochronology of polylithionite from metasomatites gave an age of 286 ± 10 Ma (Golovanov, 2001), which ranks the Shavazsay deposit among the oldest and best-preserved VS deposits globally.

Sonora: The Sonora deposit is located in northern Mexico, in the southern portion of the Basin and Range, and includes the La Ventana, Fleur, and El Sauz concessions (App. 5, Fig. A3A-D). The information available on the features of the Sonora deposit is limited to the technical reports of SRK (2015) and Ausenco (2018). The Sonora deposit is hosted by a half-graben basin filled with a Miocene sequence of lacustrine sediments. The orebodies overlie ignimbrites, rhyolitic tuffs, and breccias of the Oligocene to Miocene upper volcanic complex of the Sierra Madre Occidental and are largely capped by Quaternary basaltic flows. The Li mineralization forms lower and upper clay units, which are laterally continuous and separated by an intrabasinal ignimbrite unit. The lower clay unit includes a reworked andesitic tuff at its base, whereas the top of this unit is dominated by lacustrine beds of clays, silica, and carbonates. The lower clay unit includes the following subunits from bottom to the top: (1) the so-called “Hot Spring type section,” which consists of silica nodules dispersed in fine-grained clays, (2) calcareous sandstone with a clay matrix, (3) tuffaceous sequence, which displays an extensive clay alteration affecting feldspar, (4) calcareous sequence, which mostly includes brecciated intervals, and (5) upper tuffaceous sequence. The highest Li grades have been observed in the lower clay units, whereby whole-rock analysis of composite samples yielded maximum Li concentrations of 0.81 wt %. Information on the Li mineralogy is limited to the preliminary XRD analysis performed by Ausenco (2018). The XRD results indicate that the phyllosilicate assemblage is dominated by 10 Å species (i.e., mica and/or illite), with subordinate smectite clays. Gangue mineralogy includes carbonates, Na zeolites, feldspars, quartz, and amorphous phases. The paragenetic/petrographic features of the Sonora deposit are poorly constrained, although the possible occurrence of both low-crystallinity Li illites and true Li-F micas seems to be confirmed by the geochemical footprint of the Sonora mineralization (Fig. 15A, B). However, additional paragenetic and petrographic investigations are required to fully constrain the nature of the Li clay association.

Rhyolite Ridge: The Rhyolite Ridge deposit (App. 5, Fig. A4A-C) is a searlesite-rich VS deposit located in Nevada (western US). The first description of this deposit was provided by Reynolds and Chafetz (2020), whereas additional information on associated volcanic rocks can be found in Darin et al. (in press). The Rhyolite Ridge deposit includes a 300-m-thick sequence (Cave Spring Formation) of Upper Miocene-Lower Pliocene volcano-sedimentary rock types that were deposited in an extensional basin at the border of the Silver Peak volcanic center (Darin et al., in press). Volcano-sedimentary units overlie the Rhyolite Ridge tuff and Argentite Canyon Formation latite. The main Li orebody in the Rhyolite Ridge deposit is a 40- to 75-m-thick lacustrine unit (i.e., unit 5 after Reynolds and Chafetz, 2020) that occurs between airfall units with abundant pumice. The main boron ore within unit 5 is hosted in marls and has grades of 3.5 wt % B and 0.15 to 0.2 wt % Li associated with high concentrations of searlesite and illite-smectite. This boron-rich horizon is overlain by a smectite-rich claystone, with Li concentrations up to 0.25 wt % Li. Reynolds and Chafetz (2020) inferred that a zonation in Li clays mineralogy occurs in the Rhyolite Ridge deposit, with hectorite-like clays dominating the Li-rich claystones and with illite-smectite occurring in the searlesite-rich marl. This is confirmed by Chafetz (2023), who indicate an Li upgrade of the smectite-rich units via illitization and formation of mixed-layer clays. Other features of Rhyolite Ridge include the presence of gangue minerals reported in other VS deposits (e.g., amorphous silica pods, analcime, secondary K-feldspar) and of elemental anomalies (e.g., high Mg, Cs, Rb, and Sr) supporting ore formation in an alkaline lake environment.

Kramer-Boron: The Kramer-Boron deposit (App. 5, Fig. A5A-C) is located in the northwestern Mojave Desert about 1.5 km from the town of Boron (California, western US). The Kramer-Boron deposit has been a major source of borates since 1926, yet no detailed information is available about the Li concentration in this deposit. Rio Tinto has recently announced that battery-grade Li can be obtained from the waste rocks (likely Li claystones) of the boron production (Rio Tinto, 2021). The Kramer-Boron deposit is hosted by the Middle to Upper Miocene Ricardo Formation (Gale, 1946; Barnard and Kistler, 1966). The orebody includes 40- to 120-m-thick lacustrine shales interbedded with volcaniclastic units and overlies lava flows of the Saddleback basalt formation, and it is covered by a fanglomerate units (Gale, 1946). Mineralization mostly include Na borates (borax, kernite), Na(Ca) borates (ulexite), and Ca borates (e.g., colemanite) (Swihart et al., 1996; Gans, 2023), which were deposited in a shallow lake with laminated sediments that potentially host Li clays. Primary borates are present as iso-oriented aggregates formed at the sediment-water interface (primary precipitates) and are interlayered with the likely Li-rich claystones. The primary sedimentary bedding is crosscut by late-diagenetic borate veinlets that formed via boron remobilization during thermal diagenesis at temperatures >65°C (Bowser, 2015). A lateral zonation of the mineralogy and chemistry of borates has been observed, with Na-rich species mostly occurring in the core of the system and Ca borates prevailing in the marginal facies. Fluid modeling work conducted by Swihart et al. (1996) indicates that (1) borates were precipitated from thermal fluids and (2) the lateral evolution of borates is controlled by mixing with more dilute and lower-pH fluids and/or by precipitation under lower temperatures due to the increasing distance from the core of thermal fluid discharge. Additional evidence of circulation of high-temperature (likely hydrothermal) fluids is given by the presence of realgar-orpiment and of other sulfides, which were codeposited with borates. Petrographic work by Bowser (2015) also shows that volcaniclastic units in the Kramer-Boron deposit are altered to a secondary assemblage consisting of secondary aluminosilicates such as feldspars and zeolites, which would suggest that lower-temperature diagenesis could also have played a role in the ore-formation process.

Bigadiç: Volcano-sedimentary systems from western Anatolia (Turkey) are other typical examples of clay/borate type VS systems, with a series of major deposits currently under production for boron (Helvacı et al., 2004). Here, we will summarize the features of the Bigadiç clay/borate VS deposit (App. 5, Fig. A6A, B), which is the best-described and most representative example from this mineral district with studies from Helvacı (1995) and Kadir et al. (2023). The Bigadiç deposit includes Middle to Upper Miocene volcano-sedimentary rocks, which are coeval to volcanism in the area with major active periods around 23 and 23.6 Ma (Erkül et al., 2005a, b). The lacustrine sediments were deposited within fault-bounded basins developed over pre-Miocene metamorphic rocks and are locally capped by alluvium units. The ore-bearing lacustrine units overlie Lower Miocene volcanic to volcaniclastic rocks of the Kocaiskan Formation, which includes basaltic flows as well as rhyolites, dacites, and rhyodacites. It is worth mentioning that in this unit discharge of hot fluids via geothermal springs is still active, with spring water having up to 15 ppm B, 2 ppm Li, and a temperature of 90°C (Helvacı, 1995). Previous studies (Helvacı, 1995; Erkül et al., 2005b; Kadir et al., 2023) have subdivided the orebody in five units, which from top to bottom are (1) upper Li-bearing unit, (2) upper tuff unit, (3) lower Li-bearing unit, (4) lower tuff unit, and (5) lower limestone unit. The intrabasinal tuffs (upper and lower tuff units) include pyroclastic rocks crosscut by subvolcanic units. The upper and lower Li-bearing units consist of lacustrine sediments as shales, claystones, and marls and are also the main host for the borate mineralization. The recent study of Kadir et al. (2023) describes that hectorite is the major Li repository of the system and that diagenetic alteration of volcanic glass under a high-pH environment is the most likely trigger for Li enrichment. The isotopic modeling presented by these authors also indicates that part of Li and most of the B enrichment in the basin can be explained by addition via hydrothermal fluids exsolved from a magma reservoir underplating the lacustrine basin.

Jadar: The Jadar deposit in western Serbia (App. 5, Fig. A7A-C) is the type locality and only known example for the occurrence of jadarite and as such is the sole deposit that can be classified as borosilicate-type, or jadarite-type, Li mineralization. The Jadar deposit was discovered in 2004 as part of an exploration survey aimed at borate mineralization in lacustrine systems (Rio Tinto, pers. commun., 2021). The deposit occurs in the so-called Western Balkan lithium-boron metallogenic zone (Borojević Šoštarić and Brenko, 2022), wherein other VS deposits of the clay/borate-types are also located. The Jadar deposit is hosted by an extensional basin and is spatially associated with the Mount Cer complex, which is a composite granitic laccolith with Oligocene to Miocene ages (i.e., 32 and 18–16 Ma; Löwe et al., 2021) that partially overlaps with the development of the Jadar basin. The orebody occurs at depth (100–720 m) and includes Miocene volcano-sedimentary units as dolomite-rich lacustrine shales, granite breccias, and volcaniclastic rocks, which were deposited over a district-scale unconformity onto pre-Miocene metamorphic rocks. Recent work by Putzolu et al. (in press) has shown that zeolitization and diagenesis of volcanic glass-rich units (i.e., intrabasinal volcaniclastic rocks) can partially explain the formation of jadarite. Isotopic modeling also supports that diagenesis occurred under a high heat-flow regime (up to 95°C), which was likely characterized by mixing of meteoric waters and higher-temperatures fluids. Jadar not only represents the only known example of a principally Li borosilicate-type VS but also includes discrete mineralized facies enriched in Li clay (hectorite-like), Li phosphates, Na borates, and Ca(Na) borates. The diversity of Li(B) mineralization styles observed in the Jadar deposit thus covers almost the whole spectrum of ore types to be expected in VS deposits.

Economic Li enrichment is observed in several types of crustal environments. Lithium-rich RMGs and LCT pegmatites form via dehydration melting of mica-rich (meta)sediments or by remelting of felsic (meta)igneous rocks, followed by a high degree of magmatic-hydrothermal fractionation of incompatible elements (Ballouard et al., 2023; Knoll et al., 2023; Koopmans et al., 2024). These processes occur either during or shortly after orogenic development and generally occur in the lower- to midcrustal environments under high P-T conditions (e.g., Müller et al., 2017). Other Li-endowed systems form at lower P-T conditions in the critical zone environment, whereby meteoric to hydrothermal fluids leach Li from fertile igneous rocks to form either brines or VS-style mineralization. The presence of economic concentrations of Li at different crustal levels and of the Li speciation in mineral to fluid repositories formed in the high P-T magmatic-hydrothermal (e.g., micas and spodumene) to low P-T critical zone (e.g., clays, borosilicates, brines) environments poses some critical questions about the mechanisms regulating the Li cycle and recycling at the crustal scale. We highlight that local association of VS deposits with other types of Li(B) mineralization offers additional perspectives for a better understanding of the crustal Li cycle. Future studies should be focused around the following key subjects:

1. Lithium transfer in the critical zone: Formation of both Li brines in salar deposits and VS Li(B) mineralization occurs in the shallow environment and is aided by similar geologic conditions (i.e., closed basins, evaporative concentrations, and presence of Li-fertile volcanic rocks; e.g., Munk et al., 2016, in press). However, little is known about the aging behavior of salar systems—specifically, how would a salar age in the geologic record? Can a salar be preserved? If so, would the Li initially carried by the brines be trapped in mineral repositories to form a VS-style ore? Additionally, can the hydrothermal to meteoric leaching of a VS deposit aid the Li transfer in a brine? Whereas the possible genetic relationship between brines and VS resources remains unclear, recent academic and industry research has shown that this association may occur. A well-known example is the Clayton Valley district (Nevada, US), where high Li concentrations are observed both in clay-rich VS units (Cypress and Zeus deposits; Munk et al., 2016) and in brines. Furthermore, some salar deposits from central Andes (i.e., Cauchari-Olaroz and Pastos Grandes) are locally associated with borate mineralization similar to that observed in VS deposits (Alonso et al., 1988, 1991; Alonso, 1992). Recent exploration surveys conducted in these salars have shown that the shales hosting the borate mineralization can contain high Li concentrations (>1,000 ppm), which is likely hosted in magnesian smectite clays (Lithium Americas (Argentina) Corp., pers. commun., 2024).

2. Lithium transfer in orogenic belts: Most VS deposits occur in younger terranes (Tertiary to Recent) with rare examples in Precambrian to Paleozoic rocks (Fig. 3; App. 1) (Tanaro, Italy; Junggar and Liaoning, China). The few VS deposits in mature terrains lack hydrous Li smectites and hydrous borates and show anhydrous Li(B) assemblages comprising Li illite, danburite (CaB₂Si₂O8), reedmergnerite (NaBSi₃O₈), and locally tourmaline-rich pegmatoid veins (Peng and Palmer, 1995; Guo et al., 2021; Dini et al., 2022). It follows that metamorphism, and potentially partial melting, will have a strong effect on the primary features of VS and on the Li(B) redistribution in higher P-T phases. Future studies are advised to test the behaviors of Li and B phases in VS ores under high metamorphic grades and also to assess the likelihood that partial melting of a pre-enriched source, such as VS-style ores, might contribute to the Li(B) transfer in orogenic belts.

What follows are two examples of VS-style mineralization associated with brines (i.e., Clayton Valley, western US) and with mature geologic terranes that underwent high-grade metamorphism and melting (Liaoning, northeastern China).

The Clayton Valley deposit (western US; App. 5, Fig. A8A-C) represents the best-documented example of the clay-type VS mineralization associated with an Li brine resource (e.g., Coffey et al., 2021). The origin of brines in Clayton Valley was initially considered to be the result of leaching of surrounding volcanic tuffs and subsequent concentration in the basin through poorly defined mechanisms (Davis et al., 1986; Price et al., 2000). The more recent work of Jochens and Munk (2011) and Coffey et al. (2021) has shown that interaction of Mg(Li) smectites with heated meteoric fluids (temperature >60°C) is the most likely process responsible for the Li redistribution in the brines. The peculiar association between VS-style and brine resources in Clayton Valley is likely due to the current hydrology of the system, being characterized by the presence of an active terminal basin that acts as an effective fluids repository thus enhancing the brine accumulation (Munk et al., 2016, in press).

The borate-rich VS deposits from the Liaoning province (northeastern China; App. 5, Fig. A9A, B) occur in a Paleoproterozoic volcano-sedimentary sequence that lies between two Archean cratons dominated by granitoids and greenstone belts (e.g., Zhang, 1988; Peng and Palmer, 2002). The geodynamic evolution of this area of China was characterized by the presence of a late Archean continental margin (2.55–2.52 Ga), which then was subjected to continental collision and associated granulite facies metamorphism (2.52–2.49 Ga) and extensive postcollisional granitic magmatism (2.47 Ga) (Kroner et al., 1998). Borate-rich VS ores occur at the bottom of the Paleoproterozoic continental margin sequence (Lieryu Formation) and mostly include anhydrous species such as ludwigite (Mg₂FeBO₅) and lack hydrated primary borates (e.g., ulexite, borax) that are dominant in recent VS ores (Peng and Palmer, 1995; Hu et al., 2015). Of note is that borate mineralization is locally crosscut by pegmatite and felsic veins that contain abundant tourmaline crystals (Peng and Palmer, 1995). The prevailing genetic model for the Liaoning deposits includes (Peng and Palmer, 1995; Hu et al., 2015; Peng et al., 2022) the following: (1) formation of a conventional VS-style borate mineralization in an extensional basin, which experienced input of felsic volcanic material and favorable arid climatic conditions, and (2) later prograde metamorphism due to postorogenic granite emplacement, which led to the dehydration of the primary borates to form an anhydrous assemblage. During this stage, local partial melting affected the aluminosilicate footwall rocks and also led to the boron redistribution from the VS units to form tourmaline-rich pegmatoidal bodies and dikes.

Historically, the thixotropic properties of hectorite drove the exploration efforts in clay-type VS deposits rather than their potential as an Li resource. The increase in Li demand, the need for geopolitical secure supply, and the improving processing technologies (e.g., Roth et al., 2022) have placed a renewed need for effective exploration strategies for VS deposits. Current understanding of VS Li(B) systems recognizes three essential co-occurring geologic characteristics from known deposits. These are strong spatial and temporal correlations with (1) extensional tectonics in intracontinental and/or back-arc settings that may be indicated by exhumation of metamorphic core complexes, (2) felsic peralkaline and/or peraluminous magmatism and associated hydrothermal activity, and (3) closed hydrological basins, which have undergone significant evaporation resulting in Li saturation in pore and lake waters to form authigenic smectites. These key features may be identified during detailed pre-field work desktop studies using national geological survey data, academic publications, and reports compliant with the Committee for Mineral Reserves International Reporting Standards (CRIRSCO). A particular attention should be paid to the mappable criteria of favorable soil and rock compositions, the occurrence of indicator minerals (e.g., hectorite, borates, and borosilicates), co-occurring styles of mineralization (e.g., Hg, U), and key geologic features of VS deposits described in this paper.

Once a target area is selected, remote sensing techniques (photometry and spectral analysis) in combination with literature stratigraphic data is invaluable for the identification of outcrops for reconnaissance mapping and sampling. At the initial mapping stage, particular attention should be paid to the structures, lithological facies, textures, and mineralogy. A significant challenge is the identification of Li-bearing clays in the field. Portable analyzers, such as X-ray fluorescence (pXRF) and laser-induced breakdown spectroscopy (pLIBS), can aid in the identification of favorable pathfinder elements (e.g., Cs, Rb, Mg, and K). However, only pLIBS can determine Li concentrations and will require specific calibrations for Li-bearing minerals characteristic of VS ores to avoid false positives (i.e., carbonates) (Hampton and Benson, in press).

Geochemical analysis of rock and soil specimens will require (1) selection of suitable digestion methods capable of near or total digestion of the clay fraction, (2) utilization of Li- and B-free reagents such as multi-acid or Na peroxide methods, and (3) analysis of deleterious elements (U, Hg, and potentially F) for at least a portion of the samples. Analytical finishes such as inductively coupled plasma-optical emission spectrometry (ICP-OES) and inductively coupled plasma-mass spectrometry (ICP-MS) or other methods that can directly determine Li concentrations are required, with attention paid to the selection of suitable standards and blanks. At this stage, XRD analysis can aid to confirm the mineralogical residency of Li (e.g., smectite vs. illite vs. jadarite vs. phosphates), which is key to address potential zonation of the ore mineralogy and to guide early-stage metallurgical test work.

Seismic and gravity studies are useful for determining overall basin geometries, which is essential to design drill programs and ultimately define mineral resources and reserves. Electrical resistivity tomography (ERT) and time-domain electromagnetic (TEM) methods have shown promising results in determining the geometry of the mineralization, with ERT in particular capable of differentiating the less resistive clay horizons from the more resistive volcanic rocks and basement (e.g., ABH Engineering Inc., 2024).

Volcano-sedimentary deposits are defined as shallow Li(B) mineralization formed by the alteration of felsic volcanic rocks in closed lacustrine environments. As described in this review paper, the key characteristics of this deposit type are as follows:

1. Based on the prevailing ore mineralogy, VS deposits are classified as clay type, clay/borate type, and borosilicate type. Variations of Li mineralogy among these different classes are reflected in the total contained Li and the grade distribution. Clay- and clay/borate-type VS mineralizations show comparable grade and tonnage distributions to mica-type rare metal granite deposits. Changes of the Li clay mineralogy from smectite, to illite, and rarely to Li-F micas have the potential to locally increase Li grades in VS deposits. Borosilicate-type (or jadarite-type) Li mineralization represents the high-grade end member of the VS deposit class, with measured Li grades and tonnage comparable to the range of those observed in larger LCT pegmatites.

2. VS Li(B) deposits predominantly occur in extensional continental settings associated with crustal thinning and coeval felsic-dominated igneous activity. In most cases, the hosting extensional basins developed in the hinterland of orogenic belts formed during the postsubduction phase of gravitational belt collapse and/or slab rollback. A much rarer case is VS deposits in intracontinental settings, where the Li mineralization is hosted by collapsed caldera structures (e.g., McDermitt caldera, western US).

3. VS Li(B) deposits show a strong spatial and temporal association with highly fractionated volcaniclastic and pyroclastic rocks, and locally granites and pegmatites, which are considered as the ultimate source of the contained Li and B. These igneous units show strong enrichments in LILEs and in fluxing elements (F, B, P) and have a S- to A-type magma affinities equivalent to muscovite-bearing peraluminous granitoids to peralkaline and alkaline granitoids. The geochemistry of igneous rocks is indicative of variable degrees of hybridization between melts generated from mica-rich protoliths (pelites) and deeper sourced melts with mantle-like signatures. Melt inclusion chemistry from associated extrusive rocks indicates extremely high pre-eruptive melt contents of Li and F. The high volcanic explosivity index of the associated eruptions is indicative of significant magmatic volatile phase exsolution that may result in the deposition of large volumes of Li(F)-rich volcanic ash in the VS-hosting basins.

4. Minerogenetic processes that account for the formation of the Li(B) ore in VS deposits require a high pH and high SiO_2(aq) activity environment, which is typical of shallow lakes that developed in volcanic terranes under arid conditions and intense evaporation. Under these conditions, deposition of the Li ore in the form of smectite, and locally jadarite, is linked to alteration of volcanic glass and formation of metastable silica gels and to direct precipitation from lake to pore waters saturated in SiO₂, Li⁺, and F^–. Part of the Li mineralization can be also formed via circulation of higher-temperature fluids generated by the high heat flows linked to a series of processes such as cooling of intracaldera tuffs, magmatic caldera resurgence, emplacement of subvolcanic intrusions, and underplating of granitic intrusions.

5. Most known VS deposits occur in young terranes or in stable geologic settings and are of Tertiary to Recent age. Young VS deposits show an Li(B) ore assemblage dominated by mineral species stable at high pH and low temperatures (i.e., smectite, borates, hydrous borosilicates). The narrow P-T-x stability fields of these assemblages suggests that any significant change of the surrounding conditions could be detrimental to the preservation of the primary features of the mineralization. This would result in a poor rate of preservation of VS deposits over geologic time. The few examples of VS deposits in older geologic terranes show postformation metamorphic features that have overprinted the primary and highly hydrated ore minerals and led to an ore redistribution in higher crystallinity assemblages that includes illite and anhydrous borosilicates and borates.

The authors are grateful to Rio Tinto, Rio Sava Exploration, Jindalee Ltd, Lithium Americas Corp., Lithium Americas (Argentina) Corp., Albemarle Corporation, Bacanora Lithium Plc, Noram Lithium Corp., and to the Ministry of Mining and Geology of the Republic of Uzbekistan for granting access to mining sites and data sets. The laboratory staff of the Image and Analysis Centre of the Natural History Museum of London is acknowledged for the pivotal support granted during the analytical work. Reimar Seltmann is kindly acknowledged for providing material and insights on the Shavazsay deposit (Uzbekistan). This research received funding from the National Environment Research Council (NERC) Lithium for Future Technology (LiFT) project (NE/V007068/1). This is also a contribution of AD to the project on “Resourcing low-carbon technologies for green economy of Uzbekistan” (REP-24112021/70), funded under the MUNIS project, supported by the World Bank and the Government of the Republic of Uzbekistan. The statements do not necessarily reflect the official position of the World Bank and the Government of the Republic of Uzbekistan. The authors also wish to thank the editors-in-chief of Economic Geology (David Cooke and Larry Meinert) and the Special Issue guest editors (Tom Benson, Adam Simon, and Simon Jowitt), as well as John Reynolds and an anonymous reviewer for providing insightful comments on an early version of this paper.

Francesco Putzolu is a postdoc researcher at the Resourcing the Green Economy Theme of the Natural History Museum (NHM) of London and a specialist in the mineralogy, geochemistry, and origin of unconventional ore systems. Prior to joining the NHM, he obtained his Ph.D. degree at the University of Naples “Federico II” (Italy), where he studied the enrichment of critical metals (Co, rare earth elements, Sc) in supergene systems such as laterites, bauxites, and nonsulfides. His current research interests encompass lithium enrichment in volcano-sedimentary deposits, as well as lithium and boron behavior in magmatic-hydrothermal deposits such as rare metal granites and pegmatites.