The origins of many supergiant ore deposits remain unresolved because the factors responsible for such extreme metal enrichments are not understood. One factor of critical importance is the timing of mineralization. However, timing information is commonly confounded by the difficulty of dating ore minerals. The world's largest uranium resource at Olympic Dam, South Australia, is exceptional because the high abundance of U allows U-Pb dating of ore minerals. The Olympic Dam U(-Cu-Au-Ag) ore deposit is hosted in ca. 1.59 Ga rocks, and the consensus has been that the supergiant deposit formed at the same time. We argue that, in fact, two stages of mineralization were involved. Paired in situ U-Pb and trace element analyses of texturally distinct uraninite populations show that the supergiant size and highest-U-grade zones are the result of U addition at 0.7–0.5 Ga, at least one billion years after initial formation. This conclusion is supported by a remarkable clustering of thousands of radiogenic 207Pb/206Pb model ages of Cu sulfide grains at this time. Upgrading of the original ca. 1.59 Ga U deposit to its present size at 0.7–0.5 Ga may have resulted from perturbation of regional fluid flow triggered by global climatic (deglaciation) and tectonic (breakup of Rodinia) events.
Since the discovery of the supergiant Olympic Dam Cu-U-Au-Ag deposit (South Australia) in 1975, neither its size (e.g., 11.1 Gt of ore, including 2.6 Mt of U3O8; BHP, 2020) nor the diversity of metals and minerals have been adequately explained. Although Olympic Dam is regarded as a type example of an iron oxide–copper–gold deposit (IOCG; Hitzman et al., 1992), fundamental questions regarding metal and fluid sources and age(s) of these sources and related mineralization have remained unanswered, possibly because current thinking links ore formation to a single tectonic-magmatic event, the emplacement of the Gawler silicic large igneous province (LIP) at ca. 1.59 Ga (Johnson and Cross, 1995; Allen et al., 2008; McPhie et al., 2011b; Ciobanu et al., 2013; Cherry et al., 2018; Courtney-Davies et al., 2020).
Most of the U at Olympic Dam is present in uraninite, brannerite, and coffinite, whereas the main Cu minerals are chalcopyrite, bornite, and chalcocite (Ehrig et al., 2018, 2012). The U minerals and Cu sulfides are fine grained, disseminated, and closely associated with abundant hematite. These minerals occur within the Olympic Dam Breccia Complex (ODBC; Reeve et al., 1990), which has an area of ~6 km × 3 km (Fig. 1) and thickness in the range 500 m to >1000 m. The ODBC occurs within the undeformed A-type Roxby Downs Granite (1593.87 ± 0.21 Ma; Cherry et al., 2018). The age and contact relationships of the Roxby Downs Granite suggest that it intruded the overlying, broadly comagmatic Gawler Range Volcanics (1594.73 ± 0.30 Ma; Cherry et al., 2018).
The most common clast type in the breccia complex is Roxby Downs Granite. The texture, contact relationships, distribution, and non-stratified character of breccias in the ODBC are consistent with subsurface fragmentation of already solid granite involving a combination of tectonic and hydrothermal processes (Oreskes and Einaudi, 1990; Reeve et al., 1990; McPhie et al., 2011a). Large domains in the center of the ODBC also include clasts of the Gawler Range Volcanics and younger bedded clastic facies (1590.97 ± 0.58 Ma; Cherry et al., 2018). Granitoid detritus in the bedded clastic facies was probably derived from the Roxby Downs Granite, requiring partial exhumation of the granite before ca. 1591 Ma (Cherry et al., 2018), most likely during the ~3 m.y. gap between emplacement of the granite and deposition of the bedded clastic facies.
Some uraninite in the ODBC was formed ca. 1.59 Ga (e.g., Ciobanu et al., 2013; Macmillan et al., 2016b; Apukhtina et al., 2017), but younger ages have also been recorded (Trueman et al., 1988; Johnson, 1993; Macmillan et al., 2016b). In addition, resource estimates for U and total Pb suggest that the U mineralization in its present form may be larger than it was at 1.59 Ga: U and Pb distributions do not show consistent spatial correlations (Fig. 1), and the U/Pb ratio (by weight) is 4.5, higher than the ratio of 3.8 expected if all U was in place at 1.59 Ga, producing radiogenic Pb in a closed system. Some of the Pb is common and/or thorogenic in origin, which increases the discrepancy. Notably, the highest grades of the U ore have U/Pb ratios >10 (Fig. 2), indicating pronounced Pb deficits in parts of the ODBC (Fig. 1). This relationship suggests that either a large fraction of uranogenic Pb was lost or more U was added long after initial U mineralization at 1.59 Ga. Lead loss would result in complementary radiogenic Pb enrichment elsewhere in the district, but no such Pb repositories have been identified to date. The timing of U deposition is thus critical because current ore-genesis and exploration models do not recognize the possibility of post–1.59 Ga U addition.
TWO TYPES AND AGES OF URANINITE
Uraninite, coffinite, and brannerite contain >85% of the U present and are disseminated in sulfide and gangue minerals (Ehrig et al., 2012); the remaining 15% of U is hosted in other minerals, notably hematite (Oreskes and Einaudi, 1990; Ciobanu et al., 2013). The major textural types of uraninite are (1) euhedral grains <30 μm in size (“class 1 primary uraninite”; Macmillan et al., 2016b) and (2) subhedral to round grains (<30 to ~100 μm), which form larger aggregates and fill veinlets as much as 1 mm wide (Figs. 2 and 3B). The latter are prominent in high-grade ore zones and are equivalent to “class 4 massive uraninite” (Macmillan et al., 2016b).
Uranium-lead dating of fine-grained euhedral uraninite yielded ages ca. 1.59 Ga, such as the 1588 ± 4 Ma suite shown in Figure 3A, consistent with other U-Pb ages for fine-grained euhedral uraninite associated with early U mineralization (Macmillan et al., 2016b; Apukhtina et al., 2017). The euhedral uraninite grains have high total rare earth element (REE) contents, relatively unfractionated REE patterns with low Ce/Lu and pronounced La and Eu depletions, and low Y/Ho (Fig. 3B). In contrast, texturally distinct non-euhedral uraninite grains have variably preserved U-Pb ages near 0.5 Ga (532 ± 7 and 474 ± 4 Ma; Fig. 3A), and are characterized by lower total REE contents and REE patterns with very low La/Sm, pronounced peaks at Sm, and a lack of Eu anomalies (Fig. 3B).
LATE NEOPROTEROZOIC–CAMBRIAN URANIUM ADDITION AT OLYMPIC DAM
Evidence from Age and Composition of Uraninite
Since the start of mining at Olympic Dam, assaying of diamond drill core and subsequent resource modeling indicated a deficit of Pb compared to the amount expected in a U deposit formed at 1.59 Ga (“The apparent inconsistency between the pre-mid Proterozoic age of the Olympic Dam copper- uranium- gold mineralization and its overall low lead content…”; Trueman, 1986, p. 2). Our data suggest that mismatched U and Pb abundances in the deposit are best accounted for by at least part of the U having been added late, as late as 1 b.y. after initial formation at ca. 1.59 Ga. The highest U ore grades (>2000 ppm U) are associated with U/Pb ≥ 10, equivalent to “chemical” U-Pb ages <0.7 Ga (Fig. 2). Furthermore, the lack of abundant fission fragments and relatively low inferred neutron fluence in the U ores are more easily reconciled with a Neoproterozoic rather than a Mesoproterozoic U age (Kirchenbaur et al., 2016). These observations are consistent with the presence of the texturally distinct generation of ca. 0.5 Ga uraninite described here (Fig. 3). Uraninite of this age is locally found with partially preserved yet significantly altered remnants of the older (1.59 Ga) euhedral uraninite (Macmillan et al., 2016b), and it dominates the highest U ore grades in the deposit, corroborating the younger “chemical” U-Pb ages (Fig. 2). REE patterns in this uraninite generation, having abundance maxima around Sm-Gd and lacking Eu depletions, are distinct from those in euhedral ca. 1.59 Ga uraninite (Fig. 3B) and resemble REE signatures typical of low-temperature uraninite in the large, high-grade Proterozoic unconformity-related U deposits of Canada and northern Australia (Fryer and Taylor, 1987; Frimmel et al., 2014).
Evidence from Pb Isotopes in Cu Sulfides
Independent evidence for a major U mineralizing event in the late Neoproterozoic to Cambrian is recorded in Pb isotope compositions of hydrothermal Cu sulfides at Olympic Dam. Lead isotope data acquired by laser-ablation inductively coupled plasma mass spectrometry in thousands of chalcopyrite, bornite, and chalcocite grains from across the deposit define trends that record mixing of common Pb with radiogenic Pb characterized by 207Pb/206Pb* in the range 0.07–0.06 (Fig. 4A). U concentrations in these sulfides, including those with highly radiogenic Pb, vary widely, but the vast majority have low U/Pb ratios, implying that radiogenic Pb is “unsupported” (i.e., did not evolve within the low-U sulfide carrier minerals) and inherited from U minerals, a common observation in old U deposits (Gulson and Mizon, 1980; Kister et al., 2004). The timing of production and release of highly radiogenic Pb in U minerals and its capture as “unsupported” radiogenic Pb in Cu sulfides can be constrained using simple modeling with the variables t1 (time of formation of the U mineral), t2 (time of release of radiogenic Pb from the U mineral, i.e., capture in sulfides), and the radiogenic 207Pb/206Pb inferred from isotopic analyses of the Cu sulfides (see also Item S3 in the Supplemental Material1). Varying two of these parameters predicts the third, thus providing model ages that can be compared with other evidence. Unsupported radiogenic Pb with 207Pb/206Pb* of 0.07–0.06 can be produced in U minerals formed in the period 0.9–0.6 Ga if Pb release occurred in the recent geological past (t2 = 0; Fig. 4B). If Pb release occurred earlier, the parental U minerals cannot be older than ca. 0.7 Ga. Destabilization and alteration of U minerals, known from many U deposits (Fayek et al., 1997; Martz et al., 2019), is well documented at Olympic Dam (Macmillan et al., 2016a, 2016b). The remarkably homogeneous and low 207Pb/206Pb* in the Cu sulfide minerals thus implies a major period of U mineral formation at 0.7–0.5 Ga, broadly consistent with the ca. 0.5 Ga ages of non-euhedral uraninite (Fig. 3A). This uraninite population, with its distinct texture and trace element composition (Fig. 3B), could be a remnant or late phase of the U mineralization event(s) preserved in the Pb isotope records of the Cu sulfide minerals. The presence of young ( 1.59 Ga) radiogenic Pb in the sulfides also implies widespread modification of precursor Cu sulfide minerals and perhaps new sulfide mineral growth concomitant with this stage of U mineralization. Renewed hydrothermal activity post–1 Ga is also recorded in other minerals at Olympic Dam (Apukhtina et al., 2020; Maas et al., 2020).
Late Neoproterozoic–Cambrian U at Olympic Dam may be linked to global tectonic and climatic events. The ODBC was first exposed prior to deposition of the Mesoproterozoic Pandurra Formation (Cherry et al., 2017) and again prior to deposition of ~350 m of flat-lying Cryogenian and younger sedimentary formations (Drexel et al., 1993). These sedimentary formations were deposited under periglacial conditions during the Marinoan glaciation (Tonkin and Creelman, 1990), part of the global late Neoproterozoic glaciation (Hoffman et al., 2017). The later-stage U at Olympic Dam, broadly constrained here to the period 0.7–0.5 Ga, thus overlapped with Cryogenian glaciation and/or deglaciation and the associated rise in atmospheric oxygen (Lyons et al., 2014). This period also overlaps with the final breakup of the Rodinia supercontinent and early amalgamation of Gondwana (Veevers, 2004; Li et al., 2008). The younger (0.7–0.5 Ga) U mineralization at Olympic Dam may thus have been a result of enhanced mobility of oxidized U in surficial and basinal fluids combined with near-contemporaneous exhumation and shallow burial of the ODBC.
Our study challenges the existing paradigm of U mineralization at Olympic Dam being a single event the same age as the ca. 1.59 Ga Gawler silicic LIP host rocks. Rather, the existence of texturally and chemically distinct generations of uraninite and evidence from sulfide Pb isotope compositions indicate that U at Olympic Dam is the result of at least two major stages of U deposition, the first at ca. 1.59 Ga and the second broadly constrained to the period 0.7–0.5 Ga. Supergiant U mineralization at Olympic Dam is thus the result of staged accumulation of U in two episodes a billion years apart.
This study was supported by the Australian Research Council Linkage project (grant LP130100438) and BHP Olympic Dam. Jesse Clark and James Taylor are thanked for assistance with the geological map. We are grateful to Liam Courtney-Davies, Kevin Ansdell, and Julio Almeida for insightful reviews and for sharing their knowledge of uranium deposits, and to Marc Norman for editorial handling.