The process mineralogy of leaching sandstone-hosted uranium-vanadium ores

Highlights

• This paper examines U-V process mineralogy in sandstone-hosted deposits.

• Leach results show very high U but lower V recoveries.

• Locking inhibits U recovery; V-phyllosilicate insolubility inhibits V recovery.

• Vanadium in black shale or stone coal deposits has similar geometallurgy.

Abstract

In the United States, sandstone-hosted ore deposits of the Paradox Basin (Colorado Plateau) are major resources of uranium and vanadium, two metals important to green energy among other applications. Despite historic and current mining interest, and their significance as major domestic resources of critical elements, the geometallurgy of these deposits has received little study. This article documents the geometallurgy and process mineralogy of the U-V ores and identifies the principal barriers to optimal recovery by acid leaching.

Most of the metals occur as pitchblende (mixed uranium oxide-silicate), V-hydroxides, V-bearing phyllosilicates, and diverse vanadates of U, Pb, Cu, and other metals. Commercial extraction is by two-stage heated tank leaching with H₂SO₄ and NaClO₃, yielding high U but lower V recovery (70–75% in the industrial operation). Laboratory leaching experiments coupled with comparisons of head and residue mineralogy indicate that the unrecovered U consists of micron-scale pitchblende grains locked within quartz and other insoluble minerals. The principal cause of suboptimal V recovery is the V-phyllosilicates, which show variable but generally poor solubility at room temperatures. An ancillary cause is locking of a small amount of fine-grained V-hydroxide and pitchblende by authigenic quartz and V-phyllosilicates. Comparison with other global V resources suggests that variable solubility of V-phyllosilicate ore minerals may also diminish recovery from more common ore deposit types, such as V hosted in black shales or stone coal, particularly in heap leaching of low-grade ores at coarse grain sizes.

Introduction

Uranium, an actinide, and vanadium, a first-row transition metal, are both major mineral resources in the green energy transition. Both are considered strategic or critical elements in the US: U for its nuclear potential and V as an ingredient in superalloy steels (Kelley et al., 2017; Fortier et al., 2018). More recently, V has also garnered research interest for developing novel batteries capable of storing the huge amounts of energy required for grid-scale implementation of renewable energy.

Supply is a pressing problem for both metals. Recently the USA has imported most of its U from Kazakhstan, Russia, and Canada, and most of its V from China and South Africa (US Geological Survey, 2021). Rising demand for green energy and concerns over strategic metal supply security are turning attention to available domestic resources. Some V can be produced from spent oil-refining catalysts, power plant waste products, and steelmaking slags (US Geological Survey, 2021). These yield no U and are not sufficient to meet the demand for V, necessitating primary production.

The principal known resources of U and V in the USA are found in the Paradox Basin, a roughly circular, uplifted basin on the Colorado Plateau (Fig. 1). Among the best-known and largest historical producers were the mines in the Uravan Mineral Belt, which stretches across eastern Utah and western Colorado. The only recent to current producers are mines in the La Sal Creek district in Uravan, which produced 29 million pounds of V₂O₅ and an uncertain but large amount of U₃O₈ between the early 1900 s and 1980 (Kovschak and Nylund, 1981). Intermittent production since 2006 has totaled > 8 M lbs V₂O₅, and the district still contains 21.5 million lbs of measured and indicated reserves and resources (Peters Geosciences, 2014).

The deposits in the La Sal district are mainly tabular orebodies occurring in quartz-dominated sandstones of the Jurassic Salt Wash (lower Morrison) formation. Detailed geology of the district is reviewed by Fischer, 1942, Carter and Gualtieri, 1965, and the ores are closely analogous to those of the Slick Rock district to the southeast, described in detail by Shawe (2011). A more recent study by Barton et al. (2018b) summarizes the petrography of other Uravan deposits, as well as the historical research into their metallogenesis.

Ores in the Salt Wash belong mainly to the tabular subtype of sandstone hosted uranium-vanadium deposits, although roll-front processes overprint and redistribute metals in some areas (Burrows, 2010). Vanadium, and its common associate uranium, are most soluble in their oxidized V(V) and U(VI) forms and precipitate mainly by reduction. The most likely reductant in the Salt Wash deposits is H₂S, which occurs at high concentrations in Paradox Basin petroleum plays and is one of the few geologically common species with enough reducing power to precipitate montroseite ((V,Fe)OOH) (Wanty and Goldhaber, 1992). Montroseite is one of the principal ore minerals, having precipitated along with pitchblende (a mix of uraninite [UO₂] and coffinite [USiO₃·nH₂O]) early in the sequence of ore deposition (Barton et al., 2018a, Barton et al., 2018b). During or after the ore stage, some V was dissolved from montroseite and/or V in solution reacted with authigenic quartz to form vanadian phyllosilicates such as roscoelite (K(V,Al,Mg)₂AlSi₃O₁₀(OH)₂), V-illite (K_0.65(V,Al)₂(Al,Si)₄O₁₀(OH)₂), and V-smectite or V-chlorite ((V,Fe,Mg,Al)₆(Si,Al)₄O₁₀(OH)₈), which occur as intergranular cements and fringes around quartz grains. The last stage in the ore mineral paragenesis was supergene redistribution of metals to form vanadates and other high-valent V(V) minerals (Barton et al., 2018b). In detail, the sequence of events and resulting mineralogy and mineral textures vary considerably within deposits and individual orebodies. Table 1 gives the common minerals in the La Sal district U-V ores. Accompanying gangue is mainly quartz (SiO₂), but includes potassic feldspar (KAlSi₃O₈), calcite (CaCO₃), pyrite (FeS₂), and assorted minor metallic phases such as hematite (Fe₂O₃), anatase (TiO₂), ferroselite (FeSe₂), clausthalite (PbSe), chalcopyrite (CuFeS₂), galena (PbS), and jarosite (KAl₃(SO₄)₂(OH)₆).

Except for a few uranyl vanadates, virtually all of the U at La Sal occurs in pitchblende. In terms of volume, the phyllosilicate minerals roscoelite, illite and chlorite are the most common V-bearing phases, with the (hydr)oxides corvusite, montroseite, duttonite, and hewettite also occurring at multiple localities and hosting most of the recoverable V overall (Weeks and Thompson, 1954; Weeks et al., 1959; Carter and Gualtieri, 1965). Most of these minerals deviate significantly from the ideal compositions in Table 1. The uraninite and coffinite that nominally comprise pitchblende in reality contain up to 15 mol % U(VI) rather than pure U(IV) (Finch and Murakami, 1999). Similarly, Wanty et al. (1990) found that the oxidation state of both V and Fe in natural hydroxides such as montroseite is commonly mixed and highly variable with the average V(III)/V_total being 0.66 and the average Fe(III)/Fe_total being 0.62. Mixtures of V(III) and V(IV) in hydroxides may represent direct hydrolysis and precipitation of dissolved V(IV) in groundwater (Wanty et al., 1990), solid state oxidation and dehydration of montroseite to form paramontroseite (VO₂) (Evans and Mrose, 1955; Forbes and Dubessy, 1988), and/or in-situ oxidation and hydration of montroseite to form duttonite ((V,Fe)O(OH)₂) (Thompson et al., 1957). A simplified Eh-pH diagram for U-V systems is shown in Fig. 2.

Virtually anything oxidizing will dissolve U from uraninite and coffinite, and the presence of a carbonate, sulfate, organic, or other ligand stabilizes U in solution (Lunt et al., 2007, Bowell et al., 2011). Thus, hydrometallurgical methods have long been the preferred approach for U extraction. They have been recently reviewed by Schnell, 2014, Bhargava et al., 2015, and the electrochemical and kinetic details are provided by Nicol et al. (1975) among others. Briefly, leaching systems for U are always oxidizing, typically with NaClO₃, Fe³⁺ ion, or MnO₂ (Lunt et al., 2007). Acidic systems such as H₂SO₄ are the most common, but alkaline (NH₄)₂CO₃ is preferred for in-situ recovery or for leaching U in carbonate-rich deposits. A reaction for uraninite leaching in sulfuric acid with sodium chlorate is:

3UO₂ + 3H₂SO₄ + NaClO₃ ↔ 3(UO₂)SO₄ + NaCl + 3H₂O.

Recoveries by acid leaching are high even when the ore is pitchblende rather than pure uraninite. The major geometallurgical issues identified for U leaching in sandstone-hosted ores are slower dissolution rates due to cationic impurities in uraninite (Ram et al., 2013); preg-robbing by smectite, other clays, phosphates, and organic carbon (Lunt et al., 2007; Youlton and Kinnaird, 2003; Pownceby and Johnson, 2014); and acid consumption by carbonates, which raises pH and decreases dissolution rates (Eligwe et al., 1982; Youlton and Kinnaird, 2003).

By comparison, V is much more difficult to leach effectively and the reasons for its leaching behavior are virtually unknown. Prior to the 1950s, most U-V ores were treated by salt roasting rather than leaching in order to extract maximum V. In salt roasting, V ores are heated in an oxidizing atmosphere to 750-840 °C in the presence of NaCl (Burwell, 1961). This converts V into soluble sodium vanadates (e.g., sodium orthovanadate, Na₃VO₄) and separates it from accompanying metals (Burwell, 1961). The roasted ores are then leached in water or acid, yielding typical recoveries of 83–90 % (Burwell, 1961). By the late 1950 s salt roasting of U-V ores had fallen out of favor, since recovery of U was prioritized over V during the Cold War and leaching was cheaper for that purpose (Gupta and Krishnamurthy, 1992). Today the majority of sandstone-hosted U-V ores are crushed and directly leached in an agitated, heated H₂SO₄ – NaClO₃ tank (Gupta and Krishnamurthy, 1992). The mixed pregnant leach solution is sent for U solvent extraction first, and the raffinate from this step goes to V solvent extraction before being recycled through the leaching system. An overall reaction for direct leaching of an idealized montroseite under those conditions is:

NaClO₃ + 6VOOH + 12H⁺ ↔NaCl + 6VO²⁺ + 9H₂O.

No reaction has yet been proposed for V leaching from phyllosilicates, but a plausible reaction for an ideal roscoelite is:

6KVAl[AlSi₃]O₁₀(OH)₂ + NaClO₃ + 18H⁺ + 0.5O₂ = 3Al₄Si₄O₁₀(OH)₂ + 6VO²⁺ + 6K⁺ + 12H₂O + NaCl + 6SiO₂.

This is hypothetical but would be consistent with the incongruent degeneration of micas toward kaolinite compositions as observed in the oxidized acid leaching of copper (Baum, 1999). A study by Tavakoli et al. (2014) examined the kinetics of synthetic vanadium oxide (V₂O₅) leaching in sulfuric acid and found that VO₂ dissolves quickly but that its solubility is relatively low. This leads to lower recovery at high solid–liquid ratios, exacerbated by decreases in solubility with increasing pH and increasing temperature.

Geometallurgical issues in V leaching are not well understood, because most V is extracted as a byproduct of steelmaking by non-hydrometallurgical methods. However, it is clear from unpublished historical and present research that V leaching yields far lower recoveries than U leaching in sediment-hosted ore types. Hazen Research performed metallurgical tests on low-grade U- and V-bearing tailings from Naturita, Colorado (Hazen Research Inc., 1976). The head grade of the test samples was 0.32 % V₂O₅, with the dominant ore minerals consisting of tyuyamunite and roscoelite (Hazen Research Inc., 1976). Recoveries of V were < 40 % from agitated sulfuric acid leaching and sodium carbonate/carbonic acid leaching. The most successful method was agglomeration of the tailings with sulfuric acid, followed by an overnight cure, and then percolation of dilute sulfuric acid through the agglomerate at a rate of 0.01–0.02 gpm/ft² (Hazen Research Inc., 1976). The reasons for the poor recoveries were not documented. Recoveries at the White Mesa Mill, operating since 1980 on La Sal ores, are higher, at around 70–75 %, but still considerably lower than the U recovery (96 %) despite V grades generally being much higher than U grades (Peters Geosciences, 2014). The mill has operated since 1980, crushing and grinding ores to −600 μm followed by two-stage tank leaching in heated sulfuric acid with a sodium chlorate oxidant (Baker and Sparling, 1981).

More recent published studies on V leaching are generally lacking due to the unique nature of the deposits and the relative scarcity of active mining. The Colorado Plateau is one of very few provinces worldwide in which sandstone U deposits contain significant V, so most research on tabular U deposit geometallurgy omits discussion of V. Both U and V were produced in large volumes from the Colorado Plateau during most of the 20th century, but mining in the region has virtually shut down since the early 1980s. Research dwindled along with production, so studies on the geometallurgy of V mainly shifted to slags, titanomagnetite deposits, stone coal, and other feedstocks (e.g. Zheng et al., 2019a, Zheng et al., 2019b; Gilligan and Nikoloski, 2020). Since such ores are refractory to leaching, extraction is by salt roasting, which has thus become the main focus of geometallurgical research on V in recent years (Peng, 2019). Thus, most references to the geometallurgy of V in leaching date from the middle to late 20th century. A renewal of interest in the Colorado Plateau U-V deposits, however, is underway, and may trigger additional commercial and research activity (Mills and Jordan, 2021).

This study aims to examine the geometallurgy of sandstone-hosted U-V ores in acid leaching systems. A particular objective is to identify the causes of suboptimal V recovery during leaching, through leaching experiments coupled with comparisons of head and tail mineralogy. The implications are twofold. Firstly, this study will add an important new direction to the literature on the process mineralogy of sandstone-hosted U ores. This has been a topic of significant research, but almost all of it concerns V-poor deposits and the behavior of V minerals in these leaching systems is thus almost completely unknown (Youlton and Kinnaird, 2003; Youlton, 2014). Secondly, this study will shed light on geometallurgical problems in the hydrometallurgy of V, a subject of increasing interest in settings including black shale or stone coal. These low-grade but large V resources, like La Sal, host a considerable fraction of their V as phyllosilicates. This study can therefore help elucidate the likely geometallurgical problems in this type of unconventional V resource (Li et al., 2009, Li et al., 2010).