1. Introduction
Carbonatites are complex igneous rocks that frequently host economic deposits of critical metals essential for advanced technologies, especially Niobium (Nb) and rare earth elements (REE) (Mitchell, Reference Mitchell2005; Kamenetsky et al. Reference Kamenetsky, Doroshkevich, Elliott and Zaitsev2021; Yaxley et al. Reference Yaxley, Anenburg, Tappe, Decree and Guzmics2022). Nb, for instance, is a highly strategic metal primarily sourced from carbonatites. Nb is extracted from only three carbonatite locations worldwide, with around 90% of global Nb production coming from a single source, the Araxa mine in Brazil (Mitchell, Reference Mitchell2015; Schulz et al. Reference Schulz, Piatak, Papp, DeYoung, Seal and Bradley2017; Williams-Jones & Vasyukova, Reference Williams-Jones and Vasyukova2023). To understand mineralization drivers in carbonatite systems and, thus, help reduce exploration risk, it is essential to establish their geological timeline and geodynamic context (Hou et al. Reference Hou, Liu, Tian, Yang and Xie2015; Williams-Jones & Vasyukova, Reference Williams-Jones and Vasyukova2023; Tappe et al. Reference Tappe, Stracke and Yaxley2024). Geochronological and geochemical techniques provide key insights into the origins of carbonatitic melts, their emplacement histories, and post-intrusion alteration systems (Pirajno et al. Reference Pirajno, González-Álvarez, Chen, Kyser, Simonetti, Leduc and leGras2014; Downes et al. Reference Downes, Dunkley, Fletcher, McNaughton, Rasmussen, Jaques, Verrall and Sweetapple2016; Yang et al. Reference Yang, Lai, Pirajno, Liu, Mingxing and Sun2017). However, only 64% of known carbonatites have been reliably dated (Humphreys-Williams & Zahirovic, Reference Humphreys-Williams and Zahirovic2021), attesting to the significant challenges in dating the age of primary magmatic crystallization in these rocks. These difficulties arise from several factors, such as secondary alteration processes that overprint original mineral assemblages, the potential cryptic presence of xenocrysts, and the frequent scarcity of zircon for U–Pb geochronology. Consequently, these challenges often result in ambiguity in the geological interpretation of carbonatite dates (Amelin & Zaitsev, Reference Amelin and Zaitsev2002; Millonig et al. Reference Millonig, Gerdes and Groat2013; Decrée et al. Reference Decrée, Boulvais, Cobert, Baele, Midende, Gardien, Tack, Nimpagaritse and Demaiffe2015; Yang et al. Reference Yang, Lai, Pirajno, Liu, Mingxing and Sun2017; Slezak & Spandler, Reference Slezak and Spandler2019).
An approach to overcome ambiguity on temporal constraints on lithologies with a complex and/or poly-phased geological history is the integration of multiple isotope systems and minerals, as each mineral–isotope pair has different propensities to growth, resetting or recrystallization under different physico-chemical conditions (Walsh et al. Reference Walsh, Raimondo, Kelsey, Hand, Pfitzner and Clark2013; Slezak & Spandler, Reference Slezak and Spandler2019; Olierook et al. Reference Olierook, Fougerouse, Doucet, Liu, Rayner, Danišík, Condon, McInnes, Jaques and Evans2023). Hence, multi-method investigations enable a more holistic temporal framework by potentially tracking processes corresponding to different ambient conditions (e.g. redox state, temperature). In addition to conventional geochronometers, such as U–Pb (Montero et al. Reference Montero, Haissen, Mouttaqi, Molina, Errami, Sadki, Cambeses and Bea2016; Ghobadi et al. Reference Ghobadi, Gerdes, Kogarko, Hoefer and Brey2018) and Ar–Ar (Madeira et al. Reference Madeira, Mata, Mourão, Da Brum Silveira, Martins, Ramalho and Hoffmann2010), which have been successfully used to date carbonatites, recent advancements in in situ geochronology have expanded the available toolkit. The use of collision and reaction cells in laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) now allows for routine in situ Rb–Sr and Lu–Hf dating (Olierook et al. Reference Olierook, Rankenburg, Ulrich, Kirkland, Evans, Brown, McInnes, Prent, Gillespie, McDonald and Darragh2020; Simpson et al. Reference Simpson, Gilbert, Tamblyn, Hand, Spandler, Gillespie, Nixon and Glorie2021; Rösel & Zack, Reference Rösel and Zack2022; Simpson et al. Reference Simpson, Glorie, Hand, Spandler, Gilbert and Cave2022; Giuliani et al. Reference Giuliani, Oesch, Guillong and Howarth2024; Glorie et al. Reference Glorie, Hand, Mulder, Simpson, Emo, Kamber, Fernie, Nixon and Gilbert2024). Such in situ approaches are particularly advantageous for texturally complex lithologies like carbonatites, as they integrate petrographic context with isotope systematics, enabling spatially-resolved targeting of specific mineral domains within their petrographic context (Vance et al. Reference Vance, Müller and Villa2003; Chew et al. Reference Chew, Drost, Marsh and Petrus2021; Zametzer et al. Reference Zametzer, Kirkland, Barham, Hartnady, Bath and Rankenburg2022; Alfing et al. Reference Alfing, Johnson, Kaempf, Brown, Szilas, Rankenburg and Clark2024; Kutzschbach & Glodny, Reference Kutzschbach and Glodny2024). Beyond age constraints, complementary isotope systems such as Rb–Sr, Sm–Nd and Lu–Hf can further refine information about the source of carbonatitic melts, via initial isotopic ratios (Lee et al. Reference Lee, Lee, Do Hur, Kim, Moutte and Balaganskaya2006; Tappe et al. Reference Tappe, Foley, Stracke, Romer, Kjarsgaard, Heaman and Joyce2007; Hou et al. Reference Hou, Liu, Tian, Yang and Xie2015; Oliveira et al. Reference Oliveira, Brod, Cordeiro, Dantas and Mancini2017; Tappe et al. Reference Tappe, Stracke, van Acken, Strauss and Luguet2020; Yaxley et al. Reference Yaxley, Anenburg, Tappe, Decree and Guzmics2022). For instance, economic REE enrichment in carbonatites has been linked to Sr and Nd isotopic signatures of sediment recycling (Hou et al. Reference Hou, Liu, Tian, Yang and Xie2015; Hou et al. Reference Hou, Xu, Zhang, Zheng, Wang, Liu, Miao, Gao, Zhao, Griffin and O’Reilly2023). Similarly, Nb mineralization is likely controlled by mantle source composition, as carbonated, metasomatically enriched components may contain up to three times more Nb than a primitive mantle source (Williams-Jones & Vasyukova, Reference Williams-Jones and Vasyukova2023). Additional insights into the exhumation, cooling and possible alteration history of carbonatites can be gained through low-temperature geochronometers, such as (U–Th)/He thermochronology, which has proven effective for constraining thermal events at shallow crustal levels (Wu et al. Reference Wu, Liu, Wang, Zeng, Yang, Tian and Zhang2017; Baughman & Flowers, Reference Baughman and Flowers2018).
This study presents multi-method geochronological and isotopic data for the recently discovered mineralized (Nb, REE, P) carbonatites in the Western Australian portion of the northern Aileron Province in central Australia (WA1 Resources, 2022a; Fig. 1). The discovery has intensified exploration interest in this area and subsequent drilling campaigns have since confirmed the mineralized nature of the region’s carbonatites (WA1 Resources, 2022b; Encounter Resources, 2023; Encounter Resources, 2024), including an total (indicated and inferred) mineral resource estimate for the Luni carbonatite intrusion of 220 × 106 t at 1.0% Nb2O5 (WA1 Resources, 2025). These discoveries suggest that the Aileron Province has the potential to rank among the world’s most significant Nb resources (Schulz et al. Reference Schulz, Piatak, Papp, DeYoung, Seal and Bradley2017). However, given that carbonatites in this region were identified recently, only preliminary information exists on their timing of emplacement, mineralization history, geodynamic setting and post-emplacement evolution (Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024). To investigate these aspects, we analysed a range of mineral-isotope pairs, including zircon for U–Pb and Lu–Hf, biotite for Rb–Sr and apatite for Sr, Lu–Hf, Sm–Nd and (U–Th)/He. By integrating mineral textures and regional context, we test the capability of each pair to capture distinct geological processes in carbonatite systems and interpret the significance of the age and isotope constraints within the broader geodynamic framework that enabled this mineralization.
Figure 1. (a) Reduced-to-pole magnetic image of the western portion of the Aileron and Warumpi provinces, highlighting the high magnetic intensity belt in the northern Aileron Province some 70–120 km north of the Central Australian Suture. Regions with low magnetic intensity correspond to areas buried by the Centralian Superbasin (Amadeus and Murrabba basins), Canning Basin and regolith. Locations of two geochronology samples from the Rapide Granite in the Northern Territory are indicated (Kinny, Reference Kinny2002; Kositcin et al. Reference Kositcin, Beyer and Whelan2014). Inset abbreviations: Mu – Musgrave Province; NAC – North Australian Craton; MOPLP – Mirning Ocean–Percival Lakes Province; WAC – West Australian Craton. The black dashed box in the inset marks the main map area. The area labelled “Amadeus Basin” to the north of the Central Australian Suture is along strike to the west of what has been mapped as the Ngalia Basin in the Northern Territory (Edgoose, Reference Edgoose, Ahmad and Munson2013). (b) Enlargement of the black dashed rectangle in (a), showing the magnetic image with interpreted basement geology and drillcore sample locations (source: https://geoview.dmp.wa.gov.au/geoview/). Figure adapted from Kelsey et al. (Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024).
2. Geological background
The Luni and Crean carbonatite intrusions studied in this work are hosted in the Palaeoproterozoic to early Mesoproterozoic (ca. 1860 to 1530 Ma) Aileron Province that is predominantly east–west-trending and situated in central Australia (Fig. 1). The Aileron Province forms the southern part of the North Australian Craton (NAC) and is bounded by the Lasseter Shear Zone to the west and the ca. 1690 to 1600 Ma Warumpi Province to the south (Collins, Reference Collins1995; Hollis et al. Reference Hollis, Kirkland, Spaggiari, Tyler, Haines, Wingate, Belousova and Murphy2013). It remains unclear whether the Warumpi Province is endemic to the Aileron or not, zircon Hf isotope data imply it may reflect a rifted and reattached component (Scrimgeour et al. Reference Scrimgeour, Kinny, Close and Edgoose2005; Hollis et al. Reference Hollis, Kirkland, Spaggiari, Tyler, Haines, Wingate, Belousova and Murphy2013; March et al. Reference March, Hand, Morrissey and Kelsey2024). The northern Aileron Province was subjected to peak upper amphibolite and granulite facies metamorphism, dated at ca. 1594 to 1575 Ma within 20 km of the study area (Kelsey et al. Reference Kelsey, Korhonen, Romano and Spaggiari2022; Wingate et al. Reference Wingate, Lu, Fielding, Kelsey and Spaggiari2022; Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024).
Calcite carbonatite and aillikite lamprophyre intruded Palaeoproterozoic to early Mesoproterozoic granitic, metasedimentary and metagabbronoritic gneisses in the northern Aileron Province during the Neoproterozoic (Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024). In addition, from about 60 km south of our study area, Sudholz et al. (Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023) reported a perovskite U–Pb age of 806 ± 22 Ma from aillikite-type ultramafic lamprophyres from diatremes emplaced into Neoproterozoic sedimentary rocks of the Centralian Superbasin that unconformably overlies the Aileron Province, in what is called the Webb Province. Sudholz et al. (Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023) also document minor occurrences of carbonatite. Aillikite lamprophyres and calcite carbonatite intersected in drillcore and investigated by Kelsey et al. (Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024) yielded a range of dates from different chronometers and samples. From lamprophyre, zircon U–Pb dates ranged between 2674 and 684 Ma; and apatite U–Pb yielded a single date of 863 ± 19 Ma. From carbonatite, apatite Lu–Hf yielded an isochron date of 689 ± 47 Ma. Together, these data were interpreted to suggest a maximum crystallization age of ca. 690 Ma for lamprophyre and carbonatite, and that multiple magmatic pulses were likely (Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024). However, the broad range of dates hinders an unambiguous interpretation of the geological meaning of these temporal constraints (see Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024 for discussion). Regardless, assuming a crystallization age of 800 Ma for comparative purposes, lamprophyres studied by Kelsey et al. (Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024) are isotopically similar to those studied by Sudholz et al. (Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023) and exhibit positive εNd(t) values between +3.3 and +4.6 with εHf(t) values between +5.6 to +7.2, while 87Sr/86Sr(i) are between 0.70359 and 0.70494 (Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024). Assuming the same crystallization age of 800 Ma for carbonatite, they have positive εNd(t) between +2.3 and +3.4 and εHf(t) ranging from +3.4 to +7.7, and 87Sr/86Sr(i) between 0.70326 and 0.70674 (Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024). These similar initial isotope ratios for the lamprophyres and carbonatites imply an isotopically consistent metasomatized peridotite mantle source with limited crustal contamination (Sudholz et al. Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023; Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024). Other, more distal carbonatite occurrences in the Aileron Province include the 731 ± 0.2 Ma Mud Tank carbonatite (Gain et al. Reference Gain, Gréau, Henry, Belousova, Dainis, Griffin and O’Reilly2019) and the ca. 1525 Ma Nolans Bore carbonatite (Anenburg et al. Reference Anenburg, Mavrogenes and Bennett2020). The latter shows 87Sr/86Sr(i) of ca. 0.7054 and εNd(t) of ca. –4 for unaltered minerals (cf. Huston et al. Reference Huston, Maas, Cross, Hussey, Mernagh, Fraser and Champion2016; Anenburg et al. Reference Anenburg, Mavrogenes and Bennett2020), i.e., distinctly different in their isotope compositions compared to data presented by Sudholz et al. (Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023) and Kelsey et al. (Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024). Mud Tank apatite grains have 87Sr/86Sr(i) of ca. 0.70301 and εNd(t) of ca. +0.25 (Yang et al. Reference Yang, Wu, Yang, Chew, Xie, Chu, Zhang and Huang2014) that are more similar to the Sr and Nd isotope data from lamprophyres and carbonatites from our study area (Sudholz et al. Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023; Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024). Taken together, the distinct carbonatite ages and varying isotope signatures indicate multiple, spatially and temporally separated episodes of carbonatite emplacement into the Aileron Province and overlying Centralian Superbasin. The study area was also affected by the ca. 610 to 530 Ma Petermann Orogeny (Wade et al. Reference Wade, Barovich, Hand, Scrimgeour and Close2006; Walsh et al. Reference Walsh, Raimondo, Kelsey, Hand, Pfitzner and Clark2013; Ribeiro et al. Reference Ribeiro, Kirkland, Kelsey, Reddy, Hartnady, Faleiros, Rankenburg, Liebmann, Korhonen and Clark2023) and the ca. 450–300 Ma Alice Springs Orogeny (Buick et al. Reference Buick, Storkey and Williams2008; Piazolo et al. Reference Piazolo, Daczko, Silva and Raimondo2020).
3. Materials and methods
3.a. Sample selection
All recently discovered carbonatite occurrences in the Aileron Province in Western Australia are located in the subsurface, restricting sampling to drillcore material. Understanding of carbonatite genesis in this region is still in its infancy (Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024), with uncertainties in age relationships between different occurrences and possibly also variability within individual intrusions. To address these uncertainties, we focused on selecting the most promising sections within available drillcores that appeared visually suitable for in situ geochronology analyses. Consequently, sampling was not intended to be representative but rather targeted distinct domains with potentially diverse mineralogy to maximize the likelihood of finding minerals suitable for dating. In total, ten samples of carbonatite were selected from recent drilling campaigns, provided by WA1 Resources and Encounter Resources. For the Luni intrusion (WA1 Resources), eight samples were studied from three drillcores (LUDD23001, LUDD23013, LUDD23020), whereas for the Crean intrusion (Encounter Resources), two samples were studied from one drillcore (EAL007) (Table 1). Although several analysed samples fall outside the strict definition of carbonatite (>50% carbonate minerals), they are genetically related to the carbonatite intrusions. For simplicity, and given that our sampling was not representative in a modal sense, we collectively refer to them as carbonatites.
Table 1. Collar locations and depths of samples from this study
3.b. Sample preparation and imaging
Parts of these drillcores that appeared mineralogically or texturally diverse were selected and cut using a precision saw to prepare polished thin sections, with thin sections prepared at Minerex Services (Esperance, Australia). The remaining material underwent high-voltage pulsed electrical fragmentation using a SelFrag, followed by heavy mineral separation using a heavy liquid (LST FastFloat) with a density of 2.9 g cm–3, and magnetic separation using a Frantz isodynamic separator, all at the John de Laeter Centre (JdLC), Curtin University, Perth, Australia. The non-magnetic mineral separates were bulk mounted on double-sided tape, embedded in epoxy resin, ground to approximately half-grain thickness, and polished.
Automated phase identification for all mineral mounts and thin sections was conducted using a Tescan Integrated Mineral Analyzer (TIMA) at the JdLC (Supplementary Text S1). Additional cathodoluminescence (CL) imaging of zircon and apatite was performed using a Tescan Clara field emission scanning electron microscopy (FE-SEM) at the JdLC.
3.c. Geochronology and isotope geochemistry
Isotope data were collected at the JdLC, Curtin University. Not all techniques were applied to each sample, but analytical targets were selected to gain broad spatial coverage and to incorporate a range of textural and mineralogical features (Tables 2 and 3). A detailed description of each method is provided in Supplementary Material S1, with a summary provided in the following text. Analyses focused on three geochronometers: zircon, biotite and apatite.
Table 2. Results of automated phase identification (in vol%) of thin sections analysed via energy dispersive X-ray spectroscopy using the Tescan Integrated Mineral Analyzer (TIMA). Note that the results of thin sections may not fully represent the mineral assemblage of the sampled interval
Table 3. Summary of geochronology and isotope geochemistry results of this study. ϵHf and ϵNd calculated at t = 830 Ma. All uncertainties are provided as 2 standard error of the mean
Zircon U–Pb geochronology was conducted for six samples using LA-ICP-MS on polished mineral mounts (Supplementary Text S2). Additionally, split-stream LA-ICP-MS analyses were conducted on two mineral mounts and one thin section to obtain complementary zircon Hf isotope data (Supplementary Text S2). In situ Rb–Sr geochronology for biotite was performed on four samples in thin section using LA-ICP-MS with a reaction cell to resolve isobaric interferences (Supplementary Text S3). Apatite grains were analysed in situ with LA-ICP-MS using a reaction cell for temporal constraints via the Lu–Hf system (Supplementary Text S4) and using multicollector-ICP-MS (LA-MC-ICP-MS) for Sm–Nd chronometric and isotopic composition constraints (Supplementary Text S5). The Sr isotopic composition of apatite was measured using thermal ionization mass spectrometry (TIMS) to determine initial 87Sr/86Sr values (Supplementary Text S6). Whereas apatite (U–Th)/He was used to deduce the cooling history of the carbonatite and assess potential recent alteration processes (Supplementary Text S7). Additionally, we selected apatite grains from two samples with distinct mineral assemblage (CWA001 and CWA006) for detailed trace element analysis via laser ablation mapping (Supplementary Text S8).
4. Results
4.a. Mineralogy
Results of automated phase identification are shown in Table 2, and false colour mineral maps of all thin sections are provided in Supplementary Material S1 (Supplementary Figures S1–S10). Optical petrography (Fig. 2) and automated phase identification reveal a highly diverse mineralogy, consistent with the approach to sample selection. While detailed sample descriptions based on optical petrography presenting characteristic features of the samples are provided in Supplementary Text S9, a brief summary is given below and followed by descriptions of mineral textures of target minerals.
Figure 2. Plane-polarized light images of characteristic textures and fabrics observed in optical petrography from samples CWA001 (a), CWA006 (b), CWA008 (c) and CWA010 (d), illustrating the high intra- and inter-sample variability observed in the sample sets investigated herein. Detailed petrographic sample descriptions are provided in the Supplementary Text S9, and false colour mineral maps of all thin sections are provided in the Supplementary Material S1 (Supplementary Figures S1–S10). Ank – ankerite; Act – actinolite; Bt – biotite; Ap – apatite; Bt – biotite; Cal – calcite; Gp – gypsum; Kfs – K-feldspar; Pcl – pyrochlore; Zrn – zircon.
Samples CWA001, CWA002 and CWA004 display ankerite as their main carbonate mineral and contain variable amounts of biotite and apatite. Sample CWA001 contains distinct apatite–pyrochlore veins bordered by biotite-rich margins (Fig. 2a) surrounded by K-feldspar. Sample CWA002 is characterized by euhedral biotite sheets (up to 5 mm) that display a distinct core-and-rim zoning, whereas sample CWA004 features abundant (∼79%) biotite that possibly reflects replacement of K-feldspar and is intersected by ankerite–magnetite veins. Samples CWA003, CWA005 and CWA006 are characterized by a monomineralic matrix dominated by gypsum, associated with variable amounts of biotite, apatite and minor celestine. Sample CWA003 notably exhibits mineralogically distinct domains enriched in biotite or apatite. Sample CWA005 is distinguished by coarse-grained biotite and K-feldspar phenocrysts in gypsum, while CWA006 contains dispersed apatite, pyrochlore and magnetite within a gypsum matrix (Fig. 2b). Samples CWA007 and CWA008 share a distinctive texture comprising lens-shaped aggregates of dynamically recrystallized apatite (with internal fabric that is [sub]parallel to the long axis of the lenses) set in polygonally textured ankerite matrices (Fig. 2c). Samples CWA009 and CWA010 contain calcite-dominated matrices with notable occurrences of aegirine, magnetite, apatite and sodic amphibole (Fig. 2d). Sample CWA009 features large aegirine crystals partially replaced by amphibole and biotite, while sample CWA010 is marked by magnetite crystals exhibiting ilmenite exsolution textures and rims of pyrrhotite.
4.b. Mineral textures
4.b.1. Zircon
Zircon grains are predominantly anhedral, up to several 100 μm long and abundantly fractured. Their CL response is highly variable within and also between samples (Fig. 3). Samples CWA006, CWA009 and CWA010 contain a small proportion (∼10%) of grains that range from well-preserved oscillatory zoned patterns (Fig. 3a–f) to relict magmatic zoning (Fig. 3g, h). The remaining 90% of grains within these three samples, and all grains from CWA001, CWA004 and CWA005, show predominantly chaotic patchy domains within grains (Fig. 3i–l) that largely lack oscillatory zoning. Therefore, almost all zircon grains are characterized by distinct zones that are indicative of secondary processes (e.g. resorption, recrystallization, fluid ingress). Some zircon grains are mantled by euhedral pyrochlore outgrowths (Fig. 3d), whereas others contain inclusions of pyrochlore (Fig. 3l). One particularly large zircon grain (>1 mm, Fig. 3a) occurs within the thin section of sample CWA010. This zircon grain contains apatite inclusions, has a seam of external apatite crystals, and is surrounded by a biotite-bearing calcite matrix.
Figure 3. Cathodoluminescence images of selected zircon grains from various samples, illustrating different degrees of primary structure modification. The zircon grain in figure part (a) is imaged in thin section, while images (b–l) show grains in mineral mounts. Notably, figure parts (d) and (l) display overgrowths and pyrochlore inclusions, respectively, suggesting a syngenetic relationship between zircon and pyrochlore formation. Dashed grey lines in (i) and (j) mark grain boundaries where the cathodoluminescence response matches that of the background epoxy resin. Pcl – Pyrochlore.
4.b.2. Biotite
Biotite typically occurs in different textures in the carbonatite samples. Grains are euhedral to subhedral and, under plane-polarized light, range in colour from translucent to brown (Fig. 4a), pale-brown (Fig. 4b), zoned lighter and darker red–brown (Fig. 4c) and homogenously dark red–brown (Fig. 4d). Parallel lamellae (cleavages) show variable spacing densities, ranging from <10 µm to as wide as ∼50 µm (Fig. 4). Mottled textures (Fig. 4a, b) and grain-scale deformation (Fig. 4a, c) are also present.
Figure 4. Plane-polarized light images of representative biotite textures from samples CWA004 (a), CWA005 (b), CWA009 (c) and CWA010 (d), displaying variations in colour, microstructures and mineral inclusions. Note that biotite in figure panels (b–d) develops euhedral to subhedral sheets with well-preserved oscillatory zoning (white arrows in (d)) indicative of magmatic crystallization, whereas biotite with abundant fine-grained pyrochlore inclusions in (a) replaces former K-feldspar, indicating formation via secondary (metasomatic/hydrothermal) processes. Bt – biotite; Kfs – K-feldspar; Pcl – pyrochlore.
4.b.3. Apatite
Apatite is a major constituent in most samples (Table 2), typically occurring as coarse-grained aggregates or large individual grains, commonly reaching several 100 μm and rarely exceeding 1 mm (Fig. 5a). Many grains exhibit cracks, and CL responses are variable. In samples containing apatite grain aggregates (e.g. CWA001), a grey CL response commonly marks apatite inter-grain boundaries, with thin dark grey veins extending from these boundaries into individual apatite crystals (Fig. 5a). Within these apatite aggregates, pyrochlore crystals are commonly observed, and the surrounding apatite appears to conform to the shape of the pyrochlore inclusions (Fig. 5a). Notably, broad zoning and thin, bright CL veins are visible in several grains (Fig. 5f, g), with the abundance and distribution of these veins varying across samples. Certain samples, such as CWA005 and CWA006, display complex, patchy, chaotic patterns in the central domains of the grains, appearing as bright circular features within a darker core structure (Fig. 5b, d).
Figure 5. (a–g) Cathodoluminescence images of apatite grains highlighting complex apatite internal microstructures and zoning within various samples. In this case, all greyscale features visible are apatite, while other minerals surrounding apatite appear black, i.e., are not cathodoluminescence-active (e.g. pyrochlore in figure part (a)). (h–i) Trace element maps of two apatite grains from sample CWA001 and CWA006 confirm complex chemical zoning of apatite grains. (j) Apatite chondrite-normalized (McDonough & Sun, Reference McDonough and Sun1995) REE profiles from the two grains shown in figure part (h) and (i). The abbreviation BDL refers to elements below detection level during trace element mapping via LA-ICP-MS. Colour scales are shown in element concentration (ppm). Ap – Apatite; Pcl – Pyrochlore.
Trace element maps shed light on textures observed in CL images, indicating they reflect the distribution of REE and trace elements (e.g. Sr and Y) in individual grains. Apatite from sample CWA001 notably displays a complex chemical zoning, which can be subdivided into three zones based on the light REE (LREE), Sr and Y content (Fig. 5h). For instance, zone 1 resembles a pristine apatite core with higher Sr (10,794 ppm) and LREEs (4817 ppm) compared to zones 2 (8045 ppm Sr; 1176 ppm ∑LREE) and 3 (5946 ppm Sr; 550 ppm ∑LREE), with the latter being the most depleted in LREEs but with higher heavy REE and Y (HREY) (751 ppm ∑HREY) content compared to the other zones (521 and 576 ppm ∑HREY in zone 1 and 2; Fig. 5h). Despite presenting a more homogeneous Mn and Sr composition, apatite from sample CWA006 also displays multiple zones defined by the variability of Y and REEs, with zone 1 (core) showing an overall higher LREE-HREY content that gradually decreases toward zones 2 and 3 (mantle domains) (Fig. 5i). We note the presence of a narrower rim (zone 4) with higher REE content compared to the other zones.
Chondrite-normalized REE profiles indicate that both apatite cores (zone 1 from CWA001 and CWA006; Fig. 5h, i) are very similar in terms of REE content and LREE-HREY fractionation (Fig. 5j). The apatite LREE variability from CWA001 is reflected in the chondrite-normalized shapes. Zone 1 from CWA001 displays a steep, fractionated profile with (La/Sm)CN of 3.54, whilst zones 2 and 3 show convex profiles with significantly lower (La/Sm)CN of 0.74 and 0.36, respectively, highlighting the LREE depletion compared to zone 1. Apatite from CWA006 displays similar steep, fractionated chondrite-normalized profiles, although with distinct total REE content varying from 4871 ppm (zone 1) to 2687 ppm (zone 3) with a pronounced increase in the rim (zone 4; 6085 ppm). A tabulated dataset with apatite trace element data for the individual areas is provided in Table S1 within the Supplementary Material S2.
4.c. Zircon U–Pb geochronology and Lu–Hf isotope geochemistry
In total, 457 U–Pb analyses of 438 zircon crystals (one analysis [n] per crystal, except for the large grain in sample CWA010 with n = 19 in Fig. 3a) were collected across six samples, with 97 of these from three samples also having complementary Lu–Hf isotopes. The majority of the zircon U–Pb analyses reveal highly scattered and largely discordant data (Fig. S11). Despite the complicated dataset, a significant portion of analyses from two samples (CWA006, CWA010) show clear clustering in U–Pb concordia space (Fig. S11D, F).
Sample CWA006 reveals one concordant cluster of data (defined as a concordia distance between 1.5 and –1.5) and another group that scatters along a trend consistent with loss of radiogenic Pb and some minor to moderate gain of common Pb in some crystals (Fig. S11E). The U and Th contents of zircon analyses of sample CWA006 vary significantly, ranging from very low (U = below detection limit [∼2 ppb], Th = 0.6 ppm) to high (U = 624 ppm, Th = 3079 ppm) concentrations. By restricting the near-concordant data cluster to analyses that display moderate to high U concentrations (>5 ppm), a U–Pb concordia date of 819 ± 2 Ma (mean square weighted deviation [MSWD] = 0.25, p = 0.62, n = 21/60) was derived for sample CWA006 from the Luni intrusion (Fig. S11F).
Sample CWA010 shows a broad cluster of near-concordant data at around 830 Ma for zircon grains analysed in the mineral mount (Fig. S11G). Nonetheless, calculating a reliable weighted mean or concordia age is not warranted due to the dispersed nature of the data, suggesting that processes other than radiogenic decay control much of the isotopic signature of these data. Conversely, the large zircon grain in sample CWA010 (Fig. 3a) was analysed by 19 spot ablations during the split-stream LA-ICP-MS session and provided more consistent results. Using a concordia distance filter of –5 to +5 and discarding one analysis that most likely reflects radiogenic Pb loss (analysis CWA010_TS – 9, Table S2), a relatively imprecise concordia age of 813 ± 16 Ma was calculated (MSWD = 1.2, p = 0.27, n = 8/19) for the single crystal (Fig. S11H).
The remaining four samples (CWA001, CWA004, CWA005, CWA009) show extreme scatter on a Tera–Wasserburg U–Pb plot (Fig. S11). Given an absence of a clear trend or concordant data cluster, no robust age information regarding the crystallization of these zircon grains can be extracted. Furthermore, many of the analyses in these four samples exhibited erratic ablation behaviour, indicative of high compositional variability within grains and/or many inclusions.
The complementary Hf isotopic data on the same analytical volume as used for U–Pb geochronology for samples CWA004, CWA006 and CWA010 yielded imprecise Hf isotopic data due to unusually low abundance of Hf. Compared to the reference materials with precision at about 1 ϵHf units, precision for the unknown zircon analyses was typically between 2 and 4 ϵHf (Fig. S12, Table 3). Despite the imprecise data, the Hf isotopic data from the three analysed samples define two distinct, supra-chondritic populations, with no difference between Hf isotopic data for concordant and discordant analyses when back-calculated to 830 Ma (the interpreted age of crystallization, see discussion). Samples CWA004 and CWA010 yielded weighted mean ϵHf(830 Ma) of +3.5 ± 0.4 (MSWD = 0.78, p = 0.74) and +6.6 ± 0.9 (MSWD = 0.42, p = 0.98), respectively. Sample CWA006 yielded a spread of Hf isotopic data with ϵHf(830 Ma) from +1.7 ± 2.8 to +9.7 ± 3.7, with individual values encompassing, within uncertainty, both supra-chrondritic populations from the other two samples. The complete dataset for U–Pb geochronology and the split-stream (U–Pb, Lu–Hf) sessions is provided in Tables S2 and S3, respectively.
4.d. Biotite Rb–Sr geochronology and apatite Sr isotope geochemistry
Biotite Rb–Sr analyses from four samples (CWA004, CWA005, CWA009 and CWA010) document relatively coherent age components as highlighted for both anchored and free-fitted linear regressions (Supplementary Table S4; Fig. S13). Given that biotite and apatite exhibit appearances implying they are likely co-genetic in samples CWA005, CWA009 and CWA010, and because high Sr concentrations in apatite (e.g. ∼7000 ppm for CWA009 and CWA010) render disturbance of the Sr isotopic composition during later overprint less problematic, the biotite Rb–Sr age precision may be improved by anchoring to the apatite 87Sr/86Sr ratio from the same samples (which should approximate the biotite’s initial 87Sr/86Sr(i)). Measured apatite 87Sr/86Sr ratios are consistently unradiogenic (i.e. low 87Sr/86Sr) and comparable across the different samples (Table 3), yielding 0.70296 ± 0.00006 (CWA005), 0.70284 ± 0.00007 (CWA008), 0.70316 ± 0.00013 (CWA009) and 0.70303 ± 0.00016 (CWA010; Fig. S14). No apatite 87Sr/86Sr data was obtained for sample CWA004, which lacks a sufficient number of apatite crystals. Furthermore, textures in CWA004 do not support co-genetic biotite-apatite crystallization. The mean value of the other four samples (0.70294 ± 0.00011, Supplementary Fig. S14) was assumed for sample CWA004. When anchored to apatite initial 87Sr/86Sr ratios, inverse biotite Rb–Sr isochrons yield statistically robust (i.e. p > 0.05) dates of 831 ± 3 Ma (CWA010), 819 ± 3 Ma (CWA005), 810 ± 4 Ma (CWA009) and 796 ± 9 Ma (CWA004) (Fig. S13, Table 3). Applying no anchor to sample CWA004 (where co-genetic apatite-biotite is not evident), a free-fitted regression yields a biotite Rb–Sr date of 804 ± 17 Ma. The complete dataset for biotite Rb–Sr geochronology, including dates for free-fitted regressions and conventional isochrons, as well as apatite Sr isotope geochemistry is provided in Tables S4–S6.
4.e. Apatite Sm–Nd and Lu–Hf geochronology and isotope geochemistry
Apatite Sm–Nd isotopes from six samples define isochrons but exhibit a large variability of parent–daughter ratios, resulting in different levels of age precision. The most precise temporal constraint is retrieved from sample CWA001, which yielded an apatite Sm–Nd isochron date of 848 ± 36 Ma (MSWD = 1.2, p = 0.09, n = 67). Conversely, other dates are comparably imprecise with 2 standard errors ranging from 9 to 80% (Fig. S15, Table 3). ϵNd(t) values calculated using the initial 143Nd/144Nd(i) isochron intercepts range from +4.5 to +6.7 and yielded a weighted mean ϵNd(t) of +4.8 ± 0.4 (MSWD = 1.1, p = 0.34, n = 6).
Apatite Lu–Hf isotopes from six samples define mixing lines between radiogenic and common-Hf components in inverse isochron space (Fig. S16). The isochrons were anchored to an initial 177Hf/176Hf of 3.55 ± 0.05 (Spencer et al. Reference Spencer, Kirkland, Roberts, Evans and Liebmann2020; Simpson et al. Reference Simpson, Gilbert, Tamblyn, Hand, Spandler, Gillespie, Nixon and Glorie2021), spanning the entire range of initial 177Hf/176Hf ratios of the terrestrial reservoir (Spencer et al. Reference Spencer, Kirkland, Roberts, Evans and Liebmann2020). Apatite from samples CWA001, CWA005 and CWA010 yields the oldest inverse Lu–Hf isochron dates of 722 ± 17 Ma (MSWD = 1.3, p = 0.10, n = 38), 709 ± 18 Ma (MSWD = 1.2, p = 0.17, n = 44) and 715 ± 23 Ma (MSWD = 1.9, p ≈ 0, n = 66), respectively. Conversely, apatite grains from samples CWA006, CWA008 and CWA009 yield younger inverse Lu–Hf isochron dates of 657 ± 21 Ma (MSWD = 1.3, p = 0.10, n = 38), 653 ± 22 Ma (MSWD = 1.2, p = 0.16, n = 47) and 686 ± 31 (MSWD = 1.5, p = 0.01, n = 65), respectively (Fig. S16, Table 3). The complete dataset for biotite Sm–Nd and Lu–Hf geochronology is provided in Tables S7 and S8, respectively.
4.f. Apatite (U–Th)/He thermochronology
Apatite (U–Th)/He data collected for sample CWA008 shows very low U concentrations (median = 0.1 ppm) but yields consistent (though slightly overdispersed) He dates with a weighted mean of 249 ± 13 Ma (2SE, MSWD = 3.8, n = 9/10) (Table 3). The complete dataset for apatite (U–Th)/He thermochronology is given in Table S9.
5. Discussion
5.a. Tonian carbonatite emplacement and implications for Nb mineralization
The dates obtained in this study from different geochronometers range from 831 ± 3 Ma (biotite Rb–Sr) to 249 ± 13 Ma (apatite [U–Th]/He) (Fig. 6). The oldest dates retrieved are from biotite Rb–Sr, zircon U–Pb and apatite Sm–Nd (although imprecise), spanning ca. 830–800 Ma (Fig. 6). These dates are most likely related to the emplacement processes of the carbonatite system, as they are recorded in a diverse range of carbonatitic lithologies across different drillholes, are distinct from the ages of the host rocks (Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024) and overlap the recently published 806 ± 22 Ma perovskite U–Pb age from aillikite lamprophyre located ∼60 km south of the Luni and Crean intrusions (Sudholz et al. Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023). A slightly older apatite U–Pb age of 863 ± 19 Ma from lamprophyre in drillcore EAL002 (Caird target), located about 5 km south of drillcore EAL007, also exists (Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024; Fig. 1). When considered together, these dates can be interpreted to represent (i) distinct episodes of magmatic activity, with multiple pulses of carbonatitic and lamprophyric magmas over a period of up to 60 Myr, (ii) variable impacts of post-emplacement modification, or (iii) a combination thereof.
Figure 6. Summary of temporal constraints for carbonatites that intruded into the Aileron Province (this study, left side), with a comparison to broadly coeval ages of lithologies in Australia interpreted in the context of extensional events during the breakup of the Rodinia Supercontinent. All uncertainties are 2SE. Note that apatite Sm–Nd dates with uncertainties >20% ELR.2510034 (2SE) are not shown. C. – Central; W. – Western; N. – Northern; S. – Southern; P. – Province. References are indicated by bracketed numbers: 1 – Kelsey et al. (Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024); 2 – Wingate et al. (Reference Wingate, Fielding, Kelsey, Turnbull and Lu2024); 3 – Sudholz et al. (Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023); 4 – Zhao and McCulloch (Reference Zhao and McCulloch1993); 5 – Glikson et al. (Reference Glikson, Stewart, Ballhaus, Clarke, Feeken, Leven, Shearaton and Sun1996); 6 – Gain et al. (Reference Gain, Gréau, Henry, Belousova, Dainis, Griffin and O’Reilly2019); 7 – Slezak and Spandler (Reference Slezak and Spandler2019); 8 – Olierook et al. (Reference Olierook, Agangi, Plavsa, Reddy, Yao, Clark, Occhipinti and Kylander-Clark2019); 9 – Pidgeon et al. (Reference Pidgeon, Smith and Fanning1989); 10 – Downes et al. (Reference Downes, Griffin and Griffin2007); 11 – Downes et al. (Reference Downes, Jaques, Talavera, Griffin, Gain, Evans, Taylor and Verrall2023); 12 – Huang et al. (Reference Huang, Kamenetsky, McPhie, Ehrig, Meffre, Maas, Thompson, Kamenetsky, Chambefort, Apukhtina and Hu2015); 13 – Wingate et al. (Reference Wingate, Campbell, Compston and Gibson1998).
Repeated pulses of emplacement are characteristic of many carbonatite magmatic systems (Decrée et al. Reference Decrée, Savolainen, Mercadier, Debaille, Höhn, Frimmel and Baele2020). While zircon U–Pb dates from both Luni and Crean are within uncertainty (819 ± 2 Ma and 813 ± 16 Ma, respectively) of each other, the biotite Rb–Sr dates show greater spread and less overlap between the intrusions. The 831 ± 3 Ma biotite Rb–Sr date from sample CWA010 at the Crean target predates the zircon U–Pb date and the oldest biotite Rb–Sr date of 819 ± 3 Ma (CWA005) from the Luni deposit, but also the younger biotite Rb–Sr date of 810 ± 4 Ma (CWA009) within the same intrusion. Thus, in addition to possible inter-intrusion age differences suggesting Crean might be ca. 10 Myr older than Luni, significant intra-intrusion discrepancies, as documented by the Rb–Sr dates (Table 3), may support the interpretation that the Aileron Province was intruded by carbonatites (and other low-volume mantle melts) over a prolonged period.
A challenge in interpreting the geological meaning of these dates is the strong textural evidence for hydrothermal alteration (Figs. 3–5). This alteration could either reflect syn-emplacement metasomatic processes (e.g. fenitization) of individual carbonatite pulses, in which case the dates constrain both emplacement and alteration, or late-stage post-magmatic fluid activity, in which case the dates might not represent a prolonged period of carbonatite emplacement. For instance, the texture of biotite across different samples suggests varying degrees of modification, and both deposits each display two distinct biotite dates. Especially, sample CWA010 (Crean), with a biotite Rb–Sr age of 831 ± 3 Ma, appears to preserve euhedral sheets of pristine (magmatic) biotite with oscillatory zoning, showing no signs of post-crystallization modification (Fig. 4d). In contrast, biotite in samples CWA005 (819 ± 3 Ma, Luni) and CWA009 (810 ± 4 Ma, Crean) display increasing evidence of post-crystallization modification, such as deformation microstructures, irregular intra-grain colour variability and alteration along grain boundaries and cleavage planes. Notably, biotite yielding the youngest Rb–Sr date (796 ± 9 Ma) in sample CWA004 (Luni) forms as monomineralic aggregates (∼80 vol.%) replacing former K-feldspar, which we interpret to reflect secondary formation via hydrothermal fluids. As such, biotite from sample CWA010 likely preserves the minimum magmatic crystallization age.
Evidence for modification is also observed in zircon (Fig. 3), but the overlap between zircon U–Pb dates and the oldest biotite Rb–Sr dates from Luni and Crean implies that despite zircon’s internal fractures, elevated common Pb, and trends toward higher 238U/206Pb ratios (indicating radiogenic Pb loss), a magmatic or xenocrystic component may persist. Specifically, zircon ages of 813 ± 16 Ma at Crean and 819 ± 2 Ma at Luni are consistent with the oldest biotite dates of 831 ± 3 Ma and 819 ± 3 Ma, respectively. The age equivalence at each deposit suggests that while zircon has undergone post-crystallization modification, some primary age information remains, which raises the possibility that the Crean and Luni carbonatite intrusions are not exactly the same age. However, the present data density and the potential for resetting of the zircon U–Pb system do not resolve whether these intrusions were emplaced separately at ca. 831 Ma and ca. 819 Ma or if the younger Luni dates reflect a later minor magmatic pulse or post-emplacement modification of a ca. 831 Ma intrusion.
The older 863 ± 19 Ma apatite U–Pb date from a proximal lamprophyre (Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024) predates the ca. 830–820 Ma magmatic episode documented in biotite and zircon, and may suggest an earlier onset of low-volume melt generation. However, the relatively small age difference (∼10 Ma, considering uncertainties) makes it difficult to confidently distinguish this single lamprophyre date from the tightly clustered dates presented here based on simple theoretical considerations. For example, the U–Pb apatite data define a discordia trend in 2D Tera-Wasserburg space, representing a mixing line between common and radiogenic Pb components. This trend includes a tail extending toward high common Pb values and a cluster of analyses with elevated common Pb. These clustered analyses exert an anchoring effect, strongly influencing the lower intercept age. By means of example, excluding certain data points (e.g. those with U < 1 ppm), which retain a single high common Pb analysis (reverse discordant), yields an age within uncertainty of the other putative emplacement ages. A more robust interpretation may be achieved by incorporating the 204Pb/206Pb ratio (Ludwig, Reference Ludwig1998). However, this approach is technically hampered by the instrumentation used (LA-ICP-MS; Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024) as it usually requires either multicollector or TIMS analysis, due to the low mass abundance of 204Pb and carrier gas Hg interference. Therefore, it currently appears most parsimonious to interpret the apparently slightly older apatite U–Pb date as consistent with a 830–820 Ma magmatic emplacement, which best constrains initiation of Tonian low-volume mantle melting in the Aileron Province.
Furthermore, all zircon and biotite ages presented here are significantly older than the ∼690 Ma zircon U–Pb and apatite Lu–Hf components identified by Kelsey et al. (Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024) in lamprophyre and carbonatite samples, respectively. Zircon internal structures, truncations observed in CL imagery, grain size, shape and individual spot analyses suggest that all zircon and apatite were xenocrystic, leading Kelsey et al. (Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024) to infer that magmatic emplacement of both lamprophyre and carbonatite occurred after ca. 690 Ma. While Lu–Hf apatite systematics are likely disturbed (as discussed in Section 5.c) and hence are not suitable to constrain magmatic emplacement, zircon grains analysed in this study interestingly do not display a clear correlation between their textural disturbance and the chances of recording a possibly meaningful geochronological component. Specifically, sample CWA006 exhibits a cluster of near-concordant data (Fig. S11E, F) but shows only faint relicts of relict magmatic zoning (Fig. 3g–i). Conversely, data obtained from apparently less disturbed grains (e.g. sample CWA009, Fig. 3e, f) result in considerably more scatter in U–Pb space (Fig. S11). While textural observations offer limited insight into whether the ca. 820 Ma zircon dates are magmatic or result from Pb loss, the large (>1 mm) zircon from Crean is likely not xenocrystic due to its unusual grain size, typical of carbonatitic zircon and its identical date compared to Luni’s zircon date. Additionally, biotite’s shape, texture and abundance imply it is unlikely to be xenocrystic, and its dates, along with the more cryptic zircon U–Pb data, suggest that magmatic emplacement of carbonatite (and possibly lamprophyre) occurred no later than ca. 830–820 Ma.
Irrespective of whether there is one or multiple pulses of carbonate emplacement, these findings highlight an approximate period of carbonatite magmatic emplacement between 830 and 820 Ma, which likely constrains the timing of primary Nb mineralization. Most Nb in the sampled sections is hosted in euhedral pyrochlore crystals (see Supplementary Material S2 and Supplementary Figures S1–S10), suggesting primary igneous crystallization of pyrochlore, with limited remobilization of Nb in the deeper sections of the drillcores, which sampled the relatively unaltered part of the carbonatite intrusions (WA1 Resources, 2024). Furthermore, the occurrence of pyrochlore as both inclusions within and outgrowths on zircon (e.g. Fig. 3d, l) supports broadly coeval crystallization of pyrochlore and zircon.
5.b. Emplacement of carbonatites into the Aileron Province during the Rodinia Supercontinent breakup
The mantle beneath the Aileron Province is potentially metal-rich, owing to metasomatic refertilization by silicate–carbonate melts (Sudholz et al. Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023). High-temperature fertile melts are expected to be present at both the lithosphere–asthenosphere boundary and deep within the subcontinental lithospheric mantle (Sudholz et al. Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023). Whereas our data do not provide direct insights into how deep source melting occurred, the combined Sr–Nd–Hf (Fig. 7) data are broadly consistent with those from the Mount Webb aillikite lamprophyres located ∼60 km south of the carbonatite occurrences documented here and believed to be sourced from re-melting of subcontinental lithospheric mantle that became metasomatized by primary asthenospheric carbonate-rich melts (Fig. 7; Sudholz et al. Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023). The combined Sr–Nd–Hf data from the carbonatites are more enriched than typical depleted asthenospheric mantle (e.g. mid-ocean ridge basalts) but slightly less enriched than the aillikites. Relatively unradiogenic Sr together with radiogenic Nd and Hf (Fig. 7), and the absence of any clear xenocrystic zircon population (cf. Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024), suggest that recycled crustal material exerted only a limited influence on both the carbonatite melt source and its ascent. This limited influence may be attributed in part to the low viscosity of the carbonate melt and its immiscibility with the siliceous wall rocks. Although minor heterogeneity in the metasomatic source, such as variations in vein-to-wall-rock ratios or the composition of metasomes, may also contribute to the slight Sr–Nd differences between the samples studied herein and those from previously published proximal lamprophyres and carbonatites (Sudholz et al. Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023; Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024), it is important to note that those earlier data are based on whole-rock analyses, which are more susceptible to alteration and resetting (Hammerli et al. Reference Hammerli, Kemp and Whitehouse2019; Bruand et al. Reference Bruand, Storey, Fowler, Dhuime and Doucelance2023). Moreover, the observed trend toward higher 87Sr/86Sr ratios and lower ϵNd values could reflect some assimilation of crustal material through entrainment that is probably more likely to affect whole-rock data and, without sufficient equilibration, unlikely to be discernible at the mineral scale. However, negligible impact from sediment recycling is also consistent with only a moderate REE enrichment at Luni and Crean, since REE enrichment in carbonatite deposits is typically linked to fertilization by fluids derived from recycled sedimentary material (Hou et al. Reference Hou, Liu, Tian, Yang and Xie2015; Hou et al. Reference Hou, Xu, Zhang, Zheng, Wang, Liu, Miao, Gao, Zhao, Griffin and O’Reilly2023). Additionally, the similarity of isotopic compositions between the Luni and Crean intrusions indicates that the carbonatite melts were likely derived from the same source area, branching into multiple conduits at upper crustal levels, similar to observations in other carbonatite systems (e.g. Bayan Obo; Xue et al. Reference Xue, Zhang, Chen, Wu, Wang, Di, Xu, Zhao, Li, Zhao and Mitchell2024).
Figure 7. (a) Lu–Hf and (b) Sr–Nd compositions of the carbonatites in the Aileron Province region compared to the global compilation of carbonatite (Yaxley et al. Reference Yaxley, Anenburg, Tappe, Decree and Guzmics2022), the global compilation of basalts (Stracke, Reference Stracke2012) and other localities discussed in this work. BSE – Bulk Silicate Earth; CHUR – CHondritic Uniform Reservoir; DM – Depleted Mantle; EM – Enriched Mantle; HIMU, high-238U/204Pb end-member; PREMA – PREvalent MAntle.
The geodynamic setting at ca. 830 to 800 Ma provides clues to the mechanisms by which the carbonatite melt was transported from its mantle source to the surface (Fig. 8). By 900 Ma, the Rodinia Supercontinent had assembled and incorporated more than 70% of its continental blocks (Li et al. Reference Li, Bogdanova, Collins, Davidson, de Waele, Ernst, Fitzsimons, Fuck, Gladkochub and Jacobs2008; Merdith et al. Reference Merdith, Williams, Collins, Tetley, Mulder, Blades, Young, Armistead, Cannon and Zahirovic2021; Li et al. Reference Li, Liu and Ernst2023). At this time, southern and northern Australia were joined, with Australia proximal to Laurentia, East Antarctica and the North China and Kalahari cratons. Recent palaeomagnetically constrained, full plate reconstructions indicate that continental breakup did not occur until 800–780 Ma (Fig. 8), whereby Australia began to break away from Laurentia (Merdith et al. Reference Merdith, Williams, Collins, Tetley, Mulder, Blades, Young, Armistead, Cannon and Zahirovic2021; Li et al. Reference Li, Liu and Ernst2023). However, the period between ca. 870 Ma and 800 Ma was associated with significant extensional events, induced by anticlockwise rotation of the core of the Rodinia Supercontinent (including northern Australia; Merdith et al. Reference Merdith, Williams, Collins, Tetley, Mulder, Blades, Young, Armistead, Cannon and Zahirovic2021) and perhaps true polar wander (Li et al. Reference Li, Evans and Zhang2004). Extension accelerated from ca. 830 Ma onwards, as evidenced by voluminous mafic intrusions and volcanism, including the ca. 830–825 Ma Gairdner–Willouran large igneous province in Australia (Wingate et al. Reference Wingate, Campbell, Compston and Gibson1998; Wang et al. Reference Wang, Li, Li, Liu and Yang2010) and the ca. 800 Ma Rushinga–Bukoba large igneous province in the Kalahari Craton (Hanson, Reference Hanson2003; Johnson et al. Reference Johnson, Rivers and de Waele2005). Although it is uncertain whether some other magmatically active blocks were part of the Rodinia Supercontinent, ca. 825 to 800 Ma bimodal magmatism also occurred in the South China (Li et al. Reference Li, Li, Zhou, Liu and Kinny2002; Zhou et al. Reference Zhou, Kennedy, Sun, Malpas and Lesher2002; Li et al. Reference Li, Evans and Zhang2004; Li et al. Reference Li, Su, Chung, Li, Liu, Song and Liu2005) and Tarim blocks (Liou et al. Reference Liou, Graham, Maruyama and Zhang1996; Chen et al. Reference Chen, Xu, Zhan and Li2004; Zhang et al. Reference Zhang, Li, Li, Lu, Ye and Li2007). Finally, low-volume ca. 830 to 800 Ma volatile-rich mantle-derived products are common across Australia (Fielding & Jaques, Reference Fielding and Jaques1986; Pidgeon et al. Reference Pidgeon, Smith and Fanning1989; Downes et al. Reference Downes, Griffin and Griffin2007; Downes et al. Reference Downes, Jaques, Talavera, Griffin, Gain, Evans, Taylor and Verrall2023; Sudholz et al. Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023). The combined magmatic events and sedimentary deposition point to rift-related extension from 830–800 Ma, eventually culminating in the breakup of the Rodinia Supercontinent. Extension facilitated (re)opening of transcrustal structural discontinuities that could have enabled ascent of carbonatitic melt from the lithospheric mantle, consistent with occurrences of carbonatites proximal to faults (Encounter Resources, 2024). Like many other mantle-derived, volatile-rich products (e.g. kimberlites, lamproites and lamprophyres), extension appears as a leading cause for rapidly transporting carbonate-rich mantle-derived melt to higher crustal levels (Gernon et al. Reference Gernon, Jones, Brune, Hincks, Palmer, Schumacher, Primiceri, Field, Griffin and O’Reilly2023; Olierook et al. Reference Olierook, Fougerouse, Doucet, Liu, Rayner, Danišík, Condon, McInnes, Jaques and Evans2023; Sudholz et al. Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023). In central Australia, rifting, and potentially elevated mantle potential temperature from nearby plume activity, likely have driven decompression of volatile-bearing, metasomatized peridotite in the deep lithospheric mantle, allowing it to cross its volatile-saturated solidus and generate low-volume, incompatible-element-rich melts. These same processes may explain both carbonatite formation at Luni and Crean and lamprophyre magmatism (Webb Province; Sudholz et al. Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023), given their temporal and geochemical similarities (Figs. 6 and 7).
Figure 8. (a–c) Palaeogeographic reconstruction during the time of carbonatite emplacement that coincides with the breakup of the Rodinia Supercontinent. Plate reconstructions made using resolved topologies, rotation poles and continent boundaries from Merdith et al. (Reference Merdith, Williams, Collins, Tetley, Mulder, Blades, Young, Armistead, Cannon and Zahirovic2021) with GPlates v2.5 (Müller et al. Reference Müller, Cannon, Qin, Watson, Gurnis, Williams, Pfaffelmoser, Seton, Russell and Zahirovic2018), projected orthographically with a longitudinal and latitudinal meridian of 120° and 10°N. P. – Province.
Finally, the locations of carbonatite intrusions in the Aileron Province are proximal to the southwestern edge of the North Australian Craton, and the Mirning Ocean–Percival Lakes Province (MOPLP; Kirkland et al. Reference Kirkland, Smithies, Spaggiari, Wingate, Gromard, Clark, Gardiner and Belousova2017; Lu et al. Reference Lu, Wingate, Smithies, Gessner, Johnson, Kemp, Kelsey, Haines, Martin and Martin2022) that represents a continuous domain of juvenile mid-Proterozoic (ca. 1900–1300 Ma) lithosphere. The MOPLP is mostly deeply buried beneath the Centralian Superbasin and Palaeozoic–Mesozoic Canning Basin but stretches at least as far south as the Proterozoic segment between the Gawler and Yilgarn cratons (Kirkland et al. Reference Kirkland, Smithies, Spaggiari, Wingate, Gromard, Clark, Gardiner and Belousova2017) and separates the West from North Australian Craton (Lu et al. Reference Lu, Wingate, Smithies, Gessner, Johnson, Kemp, Kelsey, Haines, Martin and Martin2022). The eastern boundary of the North Australian Craton is probably delineated by the north-northeast-trending Lasseter Shear Zone (Kelsey et al. Reference Kelsey, Korhonen, Romano and Spaggiari2022; Martin et al. Reference Martin, Murdie, de Kelsey, de Gromard, Thomas, Cutten, Zhan, Lu, Haines and Brett2022), located about 60 km west of the sampled carbonatite (Fig. 1A). This transcrustal corridor is evidenced in deep passive and active seismic and magnetotelluric records, which show a strong gradient at the asthenosphere–lithosphere (Doublier et al. Reference Doublier, Kennett, Fomin, Costelloe, Moro, Kohanpour, Calvert, Huston, Champion and Southby2020; Gorbatov et al. Reference Gorbatov, Hejrani, Zhang, Medlin, Costelloe, Bugden, Kennett, Am Reading Rawlinson and Dentith2020; Duan et al. Reference Duan, Kyi and Jiang2021; Sudholz et al. Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023) and perhaps less strongly at the mantle–crust boundaries (i.e. the Moho; Kennett et al. Reference Kennett, Salmon and Saygin2011; Kennett et al. Reference Kennett, Gorbatov, Yuan, Agrawal, Murdie, Doublier, Eakin, Miller, Zhao and Czarnota2023). Seismic gradients indicate juxtaposed blocks of different velocities, and boundaries between such blocks are typically rheologically weaker and prone to re-activation (Sykes, Reference Sykes1978; Busch et al. Reference Busch, Mezger and van der Pluijm1997; Murphy et al. Reference Murphy, DeLucia, Marshak, Ravat and Bedrosian2024). The carbonatites studied herein, as well as other Australian carbonatites and Neoproterozoic ultramafic mantle-derived products, such as the Mad Gap Yards alnöite (Downes et al. Reference Downes, Jaques, Talavera, Griffin, Gain, Evans, Taylor and Verrall2023), the Bow Hill dyke (Fielding & Jaques, Reference Fielding and Jaques1986) and the Webb aillikite province (Sudholz et al. Reference Sudholz, Reddicliffe, Jaques, Yaxley, Haynes, Gorbatov, Czarnota, Frigo, Maas and Knowles2023), are situated proximal to major crustal boundaries. Thus, pre-existing rheological weaknesses beneath and proximal to the Aileron Province allowed for the emplacement of carbonatites and ultramafic, mantle-derived products during Tonian extension. Similar controls likely operated at the global scale, as broadly coeval, widespread occurrences of carbonatite and alkaline igneous rocks are well correlated with rifting associated with, and ultimately the breakup of, the Rodinia Supercontinent (Ernst & Bell, Reference Ernst and Bell2010). Notable examples occur in former Rodinia fragments, including the Canadian Cordillera (Millonig et al. Reference Millonig, Gerdes and Groat2012), the Aillik Bay region of Labrador (Tappe et al. Reference Tappe, Foley, Jenner, Heaman, Kjarsgaard, Romer, Stracke, Joyce and Hoefs2006), the Sarfartoq Alkaline Field in Greenland (Tappe et al. Reference Tappe, Pearson, Nowell, Nielsen, Milstead and Muehlenbachs2011), the Arbarastakh Alkaline-Carbonatite Complex and the Beloziminsky Alkaline-Ultrabasic-Carbonatite Massif in Russia (Ashchepkov et al. Reference Ashchepkov, Zhmodik, Belyanin, Kiseleva, Medvedev, Travin, Yudin, Karmanov and Downes2020; Doroshkevich et al. Reference Doroshkevich, Prokopyev, Kruk, Sharygin, Izbrodin, Starikova, Ponomarchuk, Izokh and Nugumanova2022), the Vinoren Aillikite Field in Norway (Zozulya et al. Reference Zozulya, Kullerud, Ribacki, Altenberger, Sudo and Savchenko2020), and the Upper Ruvubu alkaline plutonic complex along the East African Rift (Midende et al. Reference Midende, Boulvais, Tack, Melcher, Gerdes, Dewaele, Demaiffe and Decrée2014). Collectively, Neoproterozoic rifting and the eventual breakup of the Rodinia Supercontinent appear to mark a global “bloom” of carbonatite and related magmatism, consistent with their frequent occurrences in extensional continental regimes (Humphreys-Williams & Zahirovic, Reference Humphreys-Williams and Zahirovic2021).
5.c. Post-Tonian history of carbonatites
The 720–650 Ma apatite Lu–Hf dates are approximately 100 Myr younger than the putative emplacement ages of ca. 830–820 Ma, are similar to a maximum emplacement age interpreted by Kelsey et al. (Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024), and younger than the apatite Sm–Nd dates that are broadly consistent with the timing of magmatic crystallization (Fig. 6). This offset, observed within co-genetic minerals from the same sample (e.g. CWA006) and within the same mineral (i.e. Lu–Hf versus Sm–Nd in sample CWA001), precludes distinct episodes of carbonatite emplacement as the cause but hints at apatite crystallization coeval to magmatic biotite and zircon crystallization at ca. 830–820 Ma. Moreover, the closure temperature of Lu–Hf in apatite (∼650–750 °C; Glorie et al. Reference Glorie, Hand, Mulder, Simpson, Emo, Kamber, Fernie, Nixon and Gilbert2024) exceeds that of Rb–Sr in biotite (300–400 °C; Del Moro et al. Reference Del Moro, Puxeddu, Di Brozolo and Villa1982; Jenkin, Reference Jenkin1997), which suggests that neither the cooling history of the carbonatite nor thermal-overprint-driven dry volume diffusion can be invoked to explain the apparent age discrepancy.
Apatite trace element maps from samples CWA001 (Fig. 5h) and CWA006 (Fig. 5i) display preserved cores with higher REE, Mn and Sr content compared to their respective mantle, likely reflecting the primary composition that is surrounded by modified zones with different degrees of trace element depletion. We note that apatite from CWA001 shows a typical LREE depletion trend from zone 1 (core, ∑LREE = 4817 ppm) to zone 2 and 3 (mantle, 1176 and 550 ppm) often associated with fluid-mediated alteration (Henrichs et al. Reference Henrichs, O’Sullivan, Chew, Mark, Babechuk, McKenna and Emo2018; Henrichs et al. Reference Henrichs, Chew, O’Sullivan, Mark, McKenna and Guyett2019; Ribeiro et al. Reference Ribeiro, Lagoeiro, Faleiros, Hunter, Queiroga, Raveggi, Cawood, Finch and Campanha2020), which could reflect the likely late-stage pyrochlore crystallization amidst the apatite texture (Fig. 5a). Despite showing smaller degrees of intra-grain REE depletion (from zone 1 to zone 3), apatite from CWA006 does not display such LREE behaviour likely due to the overall presence of primary euhedral, disseminated pyrochlore that could maintain the REE budget in the sample. Therefore, the observed trace element variations in apatite might reflect post-crystallization hydrothermal alteration directly impacting the REE distribution with potential consequences for Lu–Hf and Sm–Nd geochronometers. Apatite, which is relatively labile to fluid interaction (Harlov & Förster, Reference Harlov and Förster2003; Henrichs et al. Reference Henrichs, Chew, O’Sullivan, Mark, McKenna and Guyett2019), may have experienced selective modification in contact with fluids associated with the carbonatite system itself, or during subsequent independent geological events, or a combination of both. Textural observations support fluid-mediated modification, including predominantly anhedral apatite shapes suggesting resorption, CL imaging and trace element maps showing evidence of both primary igneous growth (preserved cores with higher REE, Mn and Sr content), and pervasive modified domains with distinct degrees of REE depletion (Fig. 5). Thus, these textures indicate at least one post-emplacement process that affected the apatite isotopic composition.
A key control on element mobility is not only the presence of fluids, but also the fluid composition (Harlov & Förster, Reference Harlov and Förster2003). The fluid’s role in preferentially mobilizing Hf over Lu may derive from the different geochemical properties of those elements. High-field-strength elements such as Hf form complexes with halogen-, carbonate-, or sulphate-rich fluids commonly associated with carbonatite systems (Williams-Jones & Migdisov, Reference Williams-Jones, Migdisov, Kelley and Golden2014). By contrast, Lu, Sm and Nd, as REEs (REE3+ cations), do not form such strong complexes and may remain incorporated in apatite (Watson & Green, Reference Watson and Green1981; Fleet & Pan, Reference Fleet and Pan1995). This discrepancy could potentially explain the alteration of Lu–Hf isotopic composition while Sm–Nd isotopes in apatite remained unchanged, suggesting fluids capable of mobilizing Hf without dissolving the apatite itself.
The geological control on the Lu–Hf age offset remains unclear, and it may reflect a secondary process unrelated to the carbonatite system and instead be linked to the broader post-carbonatite-emplacement geological history of the Aileron Province. Such a driver could be fluid activity during deposition of Super-sequence 2 in the Centralian Superbasin (from 720 to 650 Ma; Munson et al. Reference Munson, Kruse and Ahmad2013). A regional geological control is consistent with similar apatite Lu–Hf dates from farther afield, including 689 ± 47 Ma from calcite carbonatite in drillcore EAL001ext (Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024). Moreover, a regional geological event that explains the apatite Lu–Hf dates could also be invoked to explain the cryptic ca. 692 Ma zircon U–Pb population observed in three grains from the lamprophyre dyke in drillcore EAL005. This population has been interpreted as xenocrystic cargo, implying a lamprophyre emplacement after 692 Ma (Kelsey et al. Reference Kelsey, Fielding, Wingate, Smithies, Turnbull, Ribeiro, Maas, Kirkland, Goemann, Romano and Dröllner2024). However, if the ca. 692 Ma signal instead resulted from ancient Pb loss, then the more prominent ca. 799 ± 7 Ma population could represent the magmatic crystallization age, which is more consistent with the findings of this study. However, the consistency of the ca. 692 Ma zircon population is not quite in favour of this hypothesis, and ultimately, the geological processes(es) responsible for the apparent resetting of the Lu–Hf apatite system remain ambiguous.
Irrespective of the cause of the resetting of the Lu–Hf isotope system in apatite, the present findings have important implications for the application of apatite Lu–Hf geochronology. Several studies have emphasized the robustness of Lu–Hf in apatite since such ages (retrieved from non-carbonatitic rocks) were regularly found to be identical (within uncertainty) with other high-temperature geochronometers (e.g. zircon), consistent with a relatively high closure temperature of the Lu–Hf system in apatite (Simpson et al. Reference Simpson, Gilbert, Tamblyn, Hand, Spandler, Gillespie, Nixon and Glorie2021; Gillespie et al. Reference Gillespie, Kirkland, Kinny, Simpson, Glorie and Rankenburg2022; Glorie et al. Reference Glorie, Hand, Mulder, Simpson, Emo, Kamber, Fernie, Nixon and Gilbert2024). Yet, age discrepancies between zircon and apatite, and between apatite Lu–Hf and Sm–Nd, within carbonatites studied here indicate that the apatite Lu–Hf system (at least when hosted in carbonatites) is prone to fluid-mediated modifications. Hence, in the absence of independent age constraints, detailed grain-scale textural investigations of apatite grains are important to avoid erroneous geological interpretations of apatite Lu–Hf dates. Likewise, placing such textural work within a detailed mineral paragenetic sequence (e.g. Cangelosi et al. Reference Cangelosi, Smith, Banks and Yardley2020) based on petrographic studies of well-explored carbonatite systems (i.e. unlike the newly discovered deposits examined here) will help link specific mineral dates to discrete stages of the rock’s history. Regardless, the presented CL imaging and trace element mapping appear to be useful tools to resolve primary (preserved cores) and secondary textures that could indicate disturbance of the Lu–Hf systematics (Fig. 5). The isotopic signatures documented here further suggest that mechanisms exist that may decouple the Lu–Hf and Sm–Nd systems in apatite (cf. Gillespie et al. Reference Gillespie, Kirkland, Kinny, Simpson, Glorie and Rankenburg2022), highlighting the need for further studies to identify the underlying drivers.
The apatite (U–Th)/He age of 249 ± 13 Ma is broadly consistent with existing apatite fission track data from (meta)granitoids approximately ≥400 km farther east in the Aileron Province that have central ages that range between 322 ± 15 Ma and 183 ± 8 Ma (Nixon et al. Reference Nixon, Glorie, Fernie, Hand, van Vries Leeuwen, Collins, Hasterok and Fraser2022). The relatively coherent, albeit slightly overdispersed, thermochronological data may suggest a Permo-Triassic thermal overprint, possibly related to intracontinental fault re-activation along pre-existing structures related to far-field effects from the Tasman Orogen in eastern Australia, as previously argued for the eastern Aileron Province (Nixon et al. Reference Nixon, Glorie, Fernie, Hand, van Vries Leeuwen, Collins, Hasterok and Fraser2022). However, it may also, or instead, correspond to relatively rapid cooling during denudation of overlying rock during Permian glaciation as discussed for other regions in Australia (e.g. Morón et al. Reference Morón, Kohn, Beucher, Mackintosh, Cawood, Moresi and Gallagher2020). More generally, the (U–Th)/He dates from carbonatite are broadly consistent with the broader regional low-temperature evolution of a range of lithologies in central Australia (Glorie et al. Reference Glorie, Agostino, Dutch, Pawley, Hall, Danišík, Evans and Collins2017; Quentin de Gromard et al. Reference de Gromard, Kirkland, Howard, Wingate, Jourdan, McInnes, Danišík, Evans, McDonald and Smithies2019). Therefore, the low-temperature history of the carbonatites as recorded in the apatite (U–Th)/He systematics implies a cooling history in tandem with the basement of the region.
6. Conclusions
Multi-method geochronology and isotope geochemistry provide a comprehensive framework for understanding the emplacement and evolution of carbonatites intruded into the Aileron Province. Biotite and zircon ages presented in this study define an active Tonian carbonatite system between ca. 830 and 800 Ma. Due to abundant textural evidence for modification in the various target minerals (zircon, biotite, apatite) through hydrothermal alteration, it is argued that the magmatic emplacement period and the related Nb mineralization occurred between ca. 830 and 820 Ma. This period was associated with and/or followed by hydrothermal alteration that pervasively overprinted original magmatic components. The combined Sr–Nd–Hf isotope data presented here indicate that the carbonatites were derived from a metasomatically overprinted, depleted mantle source, consistent with isotopic characteristics of nearby lamprophyres, supporting a shared source region, potentially in the subcontinental lithospheric mantle. The consistent isotopic signatures between two recently discovered Nb-mineralized intrusions, Luni and Crean, further indicate that the melts were derived from the same mantle source and later ascended through multiple conduits at upper crustal levels. The emplacement broadly coincided with the breakup of the Rodinia Supercontinent, when extensional tectonics facilitated melt ascent, possibly through reactivated transcrustal structures along and relatively proximal to major crustal and craton boundaries. A later fluid-mediated modification stage, evidenced by apatite Lu–Hf dates (720–650 Ma) as well as textural and chemical investigations of apatite, indicates post-emplacement alteration, perhaps unrelated to primary carbonatite activity. These findings highlight the importance of integrating multiple geochronological methods and isotopic systems to resolve the complex, multi-phase (and perhaps multi-pulse) histories of carbonatites. Such approaches allow for distinguishing emplacement ages from later alteration events, offering critical insights into the geodynamic and mineralization history of these rocks.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0016756825100204
Acknowledgements
This research was funded by WA1 Resources, Encounter Resources and the Timescales of Mineral Systems Group at Curtin University. We thank Noreen Evans and Brad McDonald for helping with LA-ICP-MS analysis, Anusha Shantha Kumara for helping with sample processing, Sarah Sherlock for editorial handling and two anonymous reviewers for their constructive comments that have improved this work. DEK, IOHF, RET and RHS publish with the permission of the executive director, Geological Survey of Western Australia. Electron microscope instrumentation was supported by ARC LE190100176 and LE140100150 at the John de Laeter Centre (JdLC), Curtin University. LA-ICP-MS in the GeoHistory Facility, JdLC, was supported by AuScope, the National Collaborative Research Infrastructure Strategy, and ARC LIEF LE150100013.
Open access
