Laser ablation (in situ) Lu-Hf geochronology of epidote group minerals

Abstract

Epidote group minerals, including allanite, clinozoisite and epidote are common in a range of metamorphic, igneous and hydrothermal systems, and are stable across a wide range of pressure–temperature (P–T) conditions. These minerals can incorporate substantial amounts of rare earth elements (REEs) during their crystallisation, making them potential candidates for Lu–Hf geochronology to provide age constraints on various geological processes. Here we report on a first exploration into the feasibility of in situ Lu–Hf geochronology for epidote group minerals from various geological settings and compare the results with age constraints from other geochronometers. Magmatic allanite samples from pegmatites and monzogranites in the Greenland anorthosite complex, Coompana Province and Qingling Orogen provided dates consistent with magmatic events spanning from c. 2660 to 1171 Ma. In the Qingling pegmatites, a younger phase of hydrothermal allanite was dated at c. 215 Ma, consistent with the timing of regional REE mineralisation. Allanite from the Yambah Shear Zone, Strangways Metamorphic Complex, yielded Lu–Hf age of c. 430 Ma. It predates the garnet and apatite growth at c. 380 Ma, suggesting the Lu–Hf system can be preserved in allanite during prograde amphibolite-facies metamorphism. Additionally, Lu–Hf dates for hydrothermal clinozoisite and epidote are consistent with the timing of hydrothermal alteration and mineralisation in a range of settings, demonstrating the utility of the technique for mineral exploration. Despite the current lack of matrix-matched reference materials, the successful application of laser ablation Lu–Hf geochronology to epidote group minerals offers valuable geochronological insights into various geological processes that can be difficult to access through other geochronometers.

Introduction

Epidote group minerals e.g., epidote, clinozoisite and allanite, exhibit stability across an extensive range of P–T conditions (Franz and Liebscher 2004; Grapes and Hoskin 2004). Their structural formula is represented as A₂M₃(SiO₄)(Si₂O₇)(O, F)(OH), or in a simplified form, A₂M₃Si₃O₁₁(O, F)(OH), where A = Ca, Sr, Pb²⁺, Mn²⁺, Th, REE³⁺, and U, and M = Al, Fe³⁺, Fe²⁺, Mn³⁺, Mn²⁺, Mg, Cr³⁺, and V³⁺ (Deer et al. 1986). Epidote and clinozoisite are the most prevalent members of the epidote group minerals and often form continuous solid solution series of Ca₂Al₃Si₃O₁₁(O, F)(OH)–Ca₂Fe₃Si₃O₁₁(O, F)(OH), facilitated by the substitution of Fe for Al (Franz and Liebscher 2004). Allanite is the REE-rich member of the epidote group, with trivalent REE occupying the A sites and charge-balanced by a divalent cation (Fe²⁺, Mn²⁺, Mg) substituting for a trivalent one in the M sites, forming the idealized formula of CaREEAl₂FeSi₃O₁₁O(OH) (Gieré and Sorensen 2004). Epidote group minerals are commonly associated with metamorphic rocks, and also frequently occur in igneous rocks and hydrothermal mineral deposits (Bird and Spieler 2004; Enami et al. 2004; Grapes and Hoskin 2004; Schmidt and Poli 2004), and thus hold great geological and economic importance. For instance, epidote is a representative mineral in the propylitic alteration zone associated with porphyry systems and the calcic alteration zone in iron-oxide copper gold (IOCG) systems (Groves et al. 2010; Sillitoe 2010). Its distribution and trace element composition can serve as valuable indicators for mineral exploration (e.g., Cooke et al. 2014; Pacey et al. 2020; Schlegel et al. 2022). Epidote, clinozoisite and allanite can remain stable during zeolite- to amphibolite-facies metamorphism in contact and regional metamorphic terranes (Franz and Liebscher 2004; Gieré and Sorensen 2004; Grapes and Hoskin 2004), providing crucial constraints on the P–T conditions and geological evolution history (Janots et al. 2008, 2009; Gabudianu Radulescu et al. 2009; Spear 2010; Janots and Rubatto 2014; Boston et al. 2017). Furthermore, allanite can occur in abundance in certain carbonatites and pegmatites, forming economic mineralisation that can be mined for critical REEs and uranium (e.g., Li and Zhou 2018; Wei et al. 2023).

In recent years, there have been significant advancements in epidote group minerals geochronology. Among these minerals, allanite stands out as it contains the highest concentrations of Th and U within the epidote group minerals (typically 0.05–3 wt% Th and 10–3000 ppm U; Gieré and Sorensen 2004 and references therein). It has proven to be a successful U–Th–Pb geochronometer, with applications using techniques such as isotope dilution-thermal ionization mass spectrometry (ID-TIMS; e.g., Barth et al. 1994), secondary ion mass spectrometry (SIMS; e.g., Catlos et al. 2000), and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS; e.g., Gregory et al. 2007, 2012; El Korh 2014; Smye et al. 2014). Similar to allanite, U and Th are sometimes incorporated into the structure of epidote-clinozoisite, yielding meaningful dates that have been determined through stepwise leaching Pb-Pb dating (Buick et al. 1999), TIMS (Oberli et al. 2004) and LA-ICP-MS (Peverelli et al. 2021). However, the application of U–Th–Pb dating to epidote group minerals has often encountered challenges due to the low concentrations of U-Th, high levels of common (i.e. initial) Pb, excess ²⁰⁶Pb derived from ²³⁰Th, and absence of matrix-matched reference materials (Gregory et al. 2007; Smye et al. 2014; Peverelli et al. 2021). These limitations have constrained the broader utilization and accuracy of the epidote-clinozoisite-allanite U–Th–Pb chronometer.

In addition to Th and U, most epidote group minerals can incorporate variable amounts of REEs during crystallisation, particularly in allanite (Frei et al. 2004; Gieré and Sorensen 2004). These REE-rich epidote group minerals have great potential as Lu–Hf geochronometers, and theoretically, meaningful geochronological data can be obtained using the recently developed in situ Lu–Hf method (Simpson et al. 2021). The collision/reaction cell technology is capable of separating ¹⁷⁶Lu from ¹⁷⁶Hf and resolving isobaric interferences, and thus in situ Lu–Hf geochronology opens a new window for direct and rapid dating of multiple minerals (e.g., garnet, apatite, xenotime, fluorite and calcite) within petrographic context. It has been proven that precise Lu–Hf geochronological constraints can be obtained from minerals with low concentrations of Lu (several ppm to sub-ppm) (Simpson et al. 2021; Glorie et al. 2022a, b, 2023a, b). Lutetium–Hf down-hole fractionation is absent during the laser ablation, and minerals with similar ablation characteristics (e.g., garnet, apatite and xenotime) give similar age off-sets (Simpson et al. 2021). As a result, the correction of Lu–Hf isotope ratios is less dependent on the matrix-matched reference materials. For instance, accurate fluorite Lu–Hf dates were obtained by Glorie et al. (2023b) using a calcite reference material to correct the ¹⁷⁶Lu/¹⁷⁷Hf isotope ratios given their similar ablation characteristics. Furthermore, Lu–Hf and U–Pb isotopic systems in the same mineral may display differences in robustness to thermal disturbance (e.g., apatite and fluorite; Glorie et al. 2023a, b), and thus Lu–Hf geochronology of epidote group minerals may record a different geological history that complements U–Th–Pb geochronology. However, the feasibility and validity of utilizing epidote group minerals for Lu–Hf geochronology have yet to be established. In this study, we present the first Lu–Hf dating of epidote group minerals, including allanite, epidote and clinozoisite, sampled from diverse geological environments, and compare the results with age constraints from other geochronometers. The Lu–Hf ages were determined using LA-ICP-MS/MS, following a protocol akin to that employed for garnet Lu–Hf dating (Simpson et al. 2021; Brown et al. 2022; Tamblyn et al. 2022). Our findings demonstrate that in situ epidote group minerals Lu–Hf geochronology can indeed yield meaningful ages for low- to intermediate-grade metamorphism, magmatism, mineralisation, and hydrothermal alteration, and can reveal historical information not captured by other dating minerals.

Geological background and sample description

Sampling strategy

Precambrian rocks are the priority targets for the purpose of this research as they have greater radiogenic Hf accumulation and thus potential higher ¹⁷⁶Hf/¹⁷⁷Hf ratios (Simpson et al. 2021). Rocks from different geological settings are selected to cover various epidote group minerals. Clinozoisite sample is from the largest and best-preserved Archean anorthosite complex in SW Greenland, which has been reported to contain clinozoisite from hydrothermal alteration (Polat et al. 2009, 2010). Magmatic allanite samples include the previously studied allanite U–Pb reference material LE40010 from SW Greenland (Smye et al. 2014; Yang et al. 2022), and allanite from Paleoproterozoic pegmatite-hosted REEs deposit/prospect in central China (Zheng et al. 2020; Zhang et al. 2022, 2023; Wei et al. 2023). These allanite samples also have published U–Pb geochronology that can be used for comparison. Metamorphic allanite is sampled from the extensively investigated Strangways Metamorphic Complex, Central Australia (Hand and Sandiford 1999; Mawby et al. 1999; Möller et al. 1999; Scrimgeour and Raith, 2001; Raimondo et al. 2014; Fournier et al. 2016), where allanite occurs as abundant accessory phase throughout the matrix and inclusions in garnet porphyroblasts of metamafic rocks (Bendall 2000). Epidote samples are from Xikuangshan IOCG deposit, SW China (Zhou et al. 2014) and drill holes with porphyry-style alteration in Coompana Province, Australia (Jagodzinski et al. 2019). Timing of their mineralisation and alteration are poorly constrained and thus epidote Lu-Hf system may provide direct geochronological constraints. One additional magmatic allanite sample comes from the monzogranite with porphyry-style alteration in Coompana Province. Detailed geological settings and sample description are introduced below and summarized in Table S1.

Anorthosite complex, SW Greenland

The Archean craton in SW Greenland represents a section of the North Atlantic Craton (NAC, Fig. 1A), consisting of six arc-generated blocks from north to south, named Maniitsoq, Fiskefjord, Sermilik, Bjørnesund, Kvanefjord, and Ivittuut (Windley and Garde 2009). The suturing of these crustal blocks is suggested to have formed by accretion and collision controlled by a series of SE-dipping subduction zones associated with granulite-facies and upper amphibolite-facies metamorphism (Dziggel et al. 2014; Dyck et al. 2015). The final assembly time of the NAC was marked by the injection of the c. 2.72 Ga “Qaarusuk dyke” (Baadsgaard and McGregor 1981), the c. 2.55 Ga Qôrqut Granite Complex (Næraa et al. 2014) and the more recently identified c. 2.56 Ga granulite-facies metamorphism (Dyck et al. 2015). During the Paleoproterozoic, numerous dyke swarms of various orientations and ages have emplaced across the NAC, predominately around c. 2.5 Ga, 2.37 Ga, 2.22 Ga and 2.05–2.03 Ga, respectively (Nilsson et al. 2013).

Fig. 1

(A) Simplified geological map of the Naajat Kuuat and Fiskenæsset anorthosite complexes from SW Greenland (modified after Hoffmann et al. 2012). (B and C) Microphotographs of sample 207824, clinozoisite- and chlorite- bearing veins intersect the plagioclase and garnet grains or develop along their boundaries, reflected light. (D) Microphotograph of allanite shard of sample LE40010, reflected light, portion with low Lu/Ce ratios are circled by dashed line. Mineral abbreviations are from Whitney and Evan (2010)

Anorthosite layered complexes are widespread in the Archean Craton of SW Greenland while only the Fiskenæsset Complex in the Bjørnesund block and the Naajat Kuuat Complex in the Sermilik block have been intensively investigated (Fig. 1A; Myers 1985; Polat et al. 2009, 2010; Hoffmann et al. 2012 and references therein). The Archean anorthosite complexes comprise thick layers of anorthosite, leucogabbro, gabbro, and ultramafic rocks. Their primary stratigraphic units, from bottom to top, include Lower Gabbro, Ultramafic, Lower Leucogabbro, Middle Gabbro, Upper Leucogabbro, Anorthosite, and Upper Gabbro (Myers 1985). Rocks from the Fiskenæsset Complex have provided a Sm–Nd errorchron age of 2973 ± 28 Ma (MSWD = 33) and regression of Pb isotope data defines a date of 2945 ± 36 Ma (MSWD = 44), representing the emplacement of Fiskenæsset Complex in an intra-oceanic island arc (Polat et al. 2010). Similar age of c. 2985 Ma was obtained by Lu–Hf or Sm–Nd isochron regression lines from the Naajat Kuuat Complex (Hoffmann et al. 2012). Subsequently, the Fiskenæsset complex was intruded by tonalite, trondhjemite, and granodiorite (TTG) sheets, which have now metamorphosed to orthogneisses, at c. 2950 Ma (Myers 1985; Polat et al. 2009, 2010). TTGs sampled around the Naajat Kuuat Complex were dated by U–Pb zircon geochronology at c. 2800 Ma (Hoffmann et al. 2012). The metamorphosed anorthosite complexes and adjacent orthogneisses are intruded by muscovite granite sheet at 2660 ± 20 Ma (Pidgeon and Kalsbeek 1978) and mafic dyke swarms around c. 2.37 Ga and c. 2.05–2.03 Ga (Nilsson et al. 2013).

Sample 207824 is a garnet anorthosite from the northernmost part of the Sermilik block (position 64°12ʹ41.76ʺ N; 50°22ʹ6.24ʺ W), and thus located close to the Naajat Kuuat Anorthosite Complex and the c. 2.55 Ga Qôrqut Granite Complex (Fig. 1A; Hoffmann et al. 2012; Næraa et al. 2014). The sample closely resembles the garnet anorthosite described by Polat et al. (2009, 2010). It is composed of plagioclase (∼ 95%), garnet and biotite, with irregular clinozoisite-chlorite veins (Fig. 1B and C). Clinozoisite-bearing veins intersect the plagioclase and garnet grains or develop along their boundaries. These veins are typically colourless, but easily discernible under reflected light (RL). Clinozoisite is interpreted to be formed by hydrothermal alteration of plagioclase (Polat et al. 2010), but there is currently no age constraint available for the hydrothermal alteration. The presence of garnet in these rocks has been attributed to the potential contamination of anorthositic magma by basaltic flows, a hypothesis yet to be verified (Windley and Smith 1974).

Sample LE40010 is a dark-brown allanite shard (Fig. 1D) separated from a megacryst (∼ 2 cm long and ∼ 1 cm wide). This megacryst is hosted within a quartz pegmatite from the Fiskenæsset anorthosite complex, and was collected by B.F. Windley in 1950. Detailed petrology of sample LE40010 has been described by Smye et al. (2014). Allanite LE40010 has undergone post-crystallisation hydrothermal alteration, exhibiting heterogeneous zonation in backscattered electron (BSE) images and cracks containing inclusions of thorite and uraninite. This sample was initially analysed using ID-TIMS and yielded a weighted mean ²⁰⁷Pb/²³⁵U age of 2646 ± 94 Ma and a ²⁰⁶Pb/²³⁸U age of 2735 ± 226 Ma (Smye et al. 2014). More recently, a relatively homogeneous shard of LE40010 was analysed in situ by LA-(MC)-ICP-MS, producing a lower intercept age of 2613 ± 43 Ma on a Tera-Wasserburg plot and a ²⁰⁷Pb-corrected ²⁰⁶Pb/²³⁸U age of 2654 ± 63 Ma (Yang et al. 2022).

Huayangchuan uranium-polymetallic deposit, Qinling Orogen, China

The Qinling Orogen, Central China, consists of four distinct tectonic units from north to south, including the southern margin of North China Craton (SNCC), North Qinling, South Qinling, and the northern margin of the Yangtze Block (Chen and Santosh 2014). The Huayangchuan U-polymetallic deposit, along with the adjacent Huangjiagou pegmatite, is situated in the Xiaoqinling district within the SNCC (Fig. 2A). Huayangchuan and Huangjiagou are nestled between the Mesozoic Laoniushan and Huanshan granitic batholiths (Fig. 2B). The predominant stratigraphy in the Xiaoqinling district comprises the Taihua Supergroup, intruded by multiphase magmatism including the Archean Wengchapu TTG gneiss, Proterozoic monzogranite, granite porphyry and pegmatite dikes, Triassic carbonatite dikes, and Jurassic-Cretaceous granitoids. Detailed geological descriptions of the Huayangchuan deposit and Huangjiagou pegmatite can be found in Zheng et al. (2020), Zhang et al. (2022, 2023), and Wei et al. (2023).

Fig. 2

(A) Simplified geological map of China, showing the location of Huayangchuan, Huangjiagou and Kangdian region. (B) Simplified geological map of the Huayangchuan uranium-polymetallic deposit and Huangjiagou pegmatite, modified from Zheng et al. (2020). (C) Simplified geological map of Kangdian IOCG Province showing location of the Xikuangshan deposit, modified after Su et al. (2021)

Pegmatite-hosted REE mineralisation marks the initial mineralisation phase at Huayangchuan (Zheng et al. 2020). Allanite is the dominant REE ore mineral within the pegmatite. It can be up to centimetres in size and easily identified as dark-brown elongated grains with reddish radiation-damaged rims. Sample HYC is allanite separated from the Proterozoic pegmatite at Huayangchuan, which petrology has been described by Zheng et al. (2020). It displays heterogeneous BSE brightness and cracks crosscutting the grains (Fig. 3A), implying the effect of post-crystallisation hydrothermal alteration. The Proterozoic pegmatite comprises coarse-grained (up to several centimeters) K-feldspar (60–70%), quartz (30–35%) and minor biotite, with accessory minerals of allanite, apatite, titanite and zircon (Zheng et al. 2020). The pegmatite has been dated to be 1807 ± 14 Ma, 1826 ± 8 and 1829 ± 11 Ma by zircon U–Pb geochronology (Yang et al. 2019; Li et al. 2023). Similar pegmatites across the Xiaoqinling district are interpreted as products of the partial melting of the metamorphic basement during the late Paleoproterozoic (∼ 1.8 Ga; Zhao et al. 2015). Major mineralisation at Huayangchuan is hosted in Triassic carbonatite dykes (0.01–1 m) that commonly crosscut the early Archean gneiss, TTG suite, and the Proterozoic pegmatite. Its mineralisation time is constrained by titanite and monazite U–Pb geochronology to be 212.3 ± 6.4 Ma and 211.8 ± 5.0 Ma, respectively (Wei et al. 2023).

Fig. 3

(A) BSE image for allanite from Huayangchuan. HYC allanite grains display heterogeneous brightness and cracks. (B) Microphotograph of the type I allanite hosted within pegmatite, reflected light. Pegmatite consists of K-feldspar, quartz and biotite, with accessory minerals of allanite, titanite and apatite. (C and D) BSE images of the type I allanite. Allanite is hosted in pegmatite together with zircon, titanite and apatite. Allanite grains consist of homogenously bright core and dark rim. (E) Microphotograph of the type II allanite, reflected light. Allanite forms aggregates along the calcite-quartz veins and intergrows with apatite and monazite. (F) BSE image of the type II allanite, monazite and apatite occur as inclusions in allanite aggregates

Huangjiagou pegmatites are discovered in the west of the Huayangchuan ore district (Fig. 2B). They contain granitic mineral assemblages (quartz, K-feldspar, and biotite; Fig. 3B) and a complex accessory mineral assemblage including allanite, zircon, thorite, titanite, betafite, uraninite, brannerite, ilmenite, magnetite, pyrite, galena, calcite, monazite and apatite (Zhang et al. 2022, 2023; Fig. 3B–F). Zircons from Huangjiagou pegmatites yield a weighted mean ²⁰⁷Pb/²⁰⁶Pb age of 1826 ± 11 Ma and magmatic allanite yields a lower intercept age of 1829 ± 23 Ma on the Tera-Wasserburg plot and a weighted mean ²⁰⁸Pb/²³²Th age of 1840 ± 17 Ma (Zhang et al. 2022). A younger U–Pb age of 1752 ± 32 Ma obtained from allanite is interpreted to represent post-crystallisation high-temperature metamorphism or hydrothermal metasomatism (Zhang et al. 2023). The pegmatites underwent a secondary phase of metasomatism, characterised by epidotization, biotite veins, and celestite-quartz-calcite crystal caves (Zhang et al. 2023). Samples HJG-85, HJG-89 and HJG-90 (Fig. 3B–F) are allanite-bearing pegmatites with varying degrees of metasomatism, and two types of allanite could be identified. Type I allanite is magmatic in origin, euhedral to subhedral, and interlocking with quartz, K-feldspar and biotite (Fig. 3B). Type I allanite coexists with zircon, titanite and apatite (Fig. 3C and D). In BSE images, it displays heterogeneous zonation with a brighter core surrounded by darker rim. Type II allanite is hydrothermal in origin and forms aggregates along the calcite-quartz veins during the later metasomatism (Zhang et al. 2023). Type II allanite is normally anhedral, intergrown with titanite and apatite and monazite (Fig. 3E and F). Type II allanite has homogeneous BSE brightness and thin dark rims along the cracks or grain boundaries, and monazite has uniform BSE brightness. Type II allanite and cogenetic titanite yielded similar U–Pb ages of c. 131 Ma. However, it is worth noting that the type II allanite contains abundant common Pb, and all analyses are clustered in the upper left part of the Tera-Wasserburg plot (Zhang et al. 2023).

Xikuangshan Fe-Cu deposit, Kangdian IOCG province, China

The Kangdian IOCG province is situated on the southwestern margin of the Yangtze Block (Fig. 2A and C; Zhao and Zhou 2011; Zhou et al. 2014). Widespread Paleoproterozoic strata crop out along NS- and NNE-trending faults, including the Dongchuan (1742–1716 Ma), Dahongshan (1711–1675 Ma), and Hekou (1722–1669 Ma) Groups, acting as the major host rocks for Fe–Cu deposits. The sediments consist of weakly metamorphosed siliciclastic, dolomitic and interbedded volcanic rocks, with intrusions of coeval 1736–1655 Ma bimodal magmatism (Zhao et al. 2010; Zhao and Zhou 2011). The Paleoproterozoic strata are covered by Mesoproterozoic Kunyang (1043–995 Ma), Huili (1082–1018 Ma), and Julin Groups (1050–1043 Ma), which developed in an intracontinental rift setting (Greentree et al. 2006; Chen et al. 2014; Zhu et al. 2016). Corresponding bimodal igneous and volcanic rocks (1072–1023 Ma) are widespread in the Kangdian region (Zhu et al. 2016; Chen et al. 2018). Arc-related gabbros and granites were emplaced during c. 870–760 Ma, followed by sedimentation of late Neoproterozoic dolostones of the Dengying and Doushantuo Formations.

All the Kangdian IOCG deposits are hosted in the late Paleoproterozoic strata, including the Dahongshan, Lala, Yinachang, Lanniping and Xikuangshan deposits (Fig. 2C; Zhou et al. 2014). The detailed ore geology, mineralisation styles and hydrothermal alteration zonation are described in Zhao and Zhou (2011) and Zhou et al. (2014). Multiple mineralisation events occurred in the Kangdian IOCG province, mainly at c. 1.7 Ga and c. 1.0 Ga (molybdenite Re–Os; Li et al. 2003; Chen and Zhou 2012; Zhao et al. 2013). The late Paleoproterozoic isotopic ages have been proposed to record the initial Fe-Cu mineralisation events in the region (Zhao and Zhou 2011; Zhao et al. 2013), whereas the Mesoproterozoic ages are interpreted to represent the timing of either hydrothermal remobilisation or an independent mineralisation event (Chen and Zhou 2014; Zhao et al. 2017).

Xikuangshan is a sedimentary rock-hosted stratiform copper deposit, composed of Fe-oxide and Cu-sulfide orebodies hosted in meta-siltstone/meta-tuffaceous rocks of the Yinmin Formation of the Dongchuan Group. The sedimentary rocks of the Yinmin Formation are extensively overprinted by late K-feldspar, sericite, chlorite, quartz, epidote, tourmaline and carbonate. Widespread multi-phase breccias are closely associated with orebodies (Zhou et al. 2014). The geochronological framework of Xikuangshan is poorly constrained at present. Zircon from metatuff layers in the host rock of nearby Yinmin deposit yielded a U–Pb age of 1742 ± 13 Ma (Zhao et al. 2010), and the orebodies are crosscut by dolerite dike of 1701 ± 28 Ma (Zhao et al. 2010, 2013). Chalcopyrite at Xikuangshan yielded Re–Os model ages ranging from c. 1842 Ma to 861 Ma (Li, 2013). Allanite from the nearby Lanniping deposit revealed multiple mineralisations and overprinting at 1728 ± 40 Ma, 1015 ± 66 Ma, and 800 ± 32 Ma (Su et al. 2021). Sample Xikuangshan is a metatuffaceous host rock of the Yinmin Formation. It has experienced extensive hydrothermalism of quartz, chlorite, epidote and tourmaline (Fig. 4). The rock is dominated by quartz and fine-grained chlorite. Epidote grains are light green in handspeciman, euhedral to subhedral, ranging from 10 to 1000 μm. Epidote normally form aggregates or disseminates as fine grains in quartz or chlorite matrix.

Fig. 4

Microphotograph of the analyzed mount of sample Xikuangshan, reflected light. Tuffaceous host rock has experienced extensive hydrothermalism of quartz, chlorite, epidote and tourmaline. Epidote normally forms aggregates or disseminates as fine grains in quartz or chlorite matrix

Yambah shear zone, central Australia

The Arunta Inlier (Fig. 5A and B) is a large metamorphic complex that records the last major tectonic event in central Australia, the Alice Springs Orogeny (450–300 Ma; Hand and Sandiford 1999; Raimondo et al. 2014). The onset of intraplate compressional deformation during the Alice Springs Orogeny (Fig. 5B) is constrained by (1) Sm–Nd isochron dates of 449 ± 10 Ma from upper amphibolite-facies Harts Range Detachment Zone (Mawby et al. 1999), (2) zircon U–Pb dates of 443 ± 6 Ma, recorded in the staurolite-bearing Edwards Creek Shear Zone (Möller et al. 1999), and (3) a monazite U–Pb date of 445 ± 5 Ma and garnet amphibolite Sm–Nd isochron date of 434 ± 6 from the Entire Point Shear Zone mylonites (Scrimgeour and Raith, 2001). During the Devonian–Carboniferous (c. 380–320 Ma), the Strangways Range Complex in the Arunta Inlier experienced diachronous high-strain deformation and fluid infiltration (Fournier et al. 2016).

Fig. 5

(A) Simplified geological map of Australia, showing the location of Arunta Inlier and Coompana Province. (B) Simplified geological map of the Harts and Strangways Metamorphic Complexes in the Arunta Inlier showing the location of Yambah Shear Zone, modified after Fournier et al. (2016). The published geochronological data are from (1) Mawby et al. (1999); (2) Möller et al. (1999); (3) Ballèvre et al. (2000) and (4) Scrimgeour and Raith (2001). (C) Simplified geological map of Coompanan Province, modified after Pawley et al. (2020a)

The Yambah Shear Zone is an amphibolite-facies schist zone transecting adjacent interlayered mafic, felsic and metapelitic granulite-facies rocks of the Strangways Range Complex. Detailed geology of the Yambah Shear Zone is introduced in Bendall (2000) and Fournier et al. (2016). The shear zone is dominated by coarse-grained muscovite-biotite-quartz schists which can also include kyanite, staurolite and/or garnet. The shear zone also contains mafic amphibolites and minor calc-silicate units, composed of hornblende-plagioclase-quartz ± garnet ± staurolite, chlorite-biotite-garnet and diopside-garnet-epidote assemblages respectively. The schists are interpreted to be derived from the adjacent Strangways-aged granulites (c. 1730–1690 Ma; Claoué-Long et al. 2008) as individual granulite layers can be traced into their equivalent schist unit (Bendall 2000). Mafic amphibolites from the Yambah Shear Zone produce a spread of Sm–Nd isochron dates from c. 401 Ma to 379 Ma (Bendall 2000), consistent with the Sm–Nd isochron date of 381 ± 7 Ma from the West Bore Shear Zone (Ballèvre et al. 2000) and the monazite U–Pb date of 387 ± 18 Ma from Edwards Creek Shear Zone (Möller et al. 1999) in the Strangways Metamorphic Complex (Fig. 5B). Thermobarometry and phase equilibria studies indicate the mid amphibolite-facies assemblages in the Yambah Shear Zone developed during clockwise prograde metamorphism at conditions of 6 kbar, 600 °C during the Alice Springs Orogeny (Bendall 2000).

Sample Yambah is a coarse-grained chlorite–garnet–allanite–orthoamphibole–biotite-bearing schist (Fig. 6). Garnet in this rock occurs as euhedral porphyroblasts (to 15 mm), forming about 30% of the rock. Garnet contains inclusions of allanite, ilmenite, quartz and apatite (Fig. 6A and B). Chlorite is the dominant phase in the rock, comprising about 60% of the rock, and defines the strong fabric (Fig. 6C). Allanite (100–500 μm), apatite (100–2000 μm), needle-like ilmenite, and minor monazite and xenotime occur throughout the matrix and along the chlorite-defined foliation. Allanite grains are euhedral and form foliation parallel to the chlorite. They have homogeneous BSE brightness. Locally apatite and monazite form cystic aggregates wrapped by the chlorite matrix. Orthoamphibole occurs as euhedral grains within the chlorite dominated foliation (Fig. 6D), and also locally as inclusions within garnet that define curviplaner inclusion patterns.

Fig. 6

Microphotographs of the sample Yambah, reflected light. (A and B) Coarse-grained garnet is wrapped by chlorite fabrics, and ilmenite and allanite occur as inclusions in garnet porphyroblast. (C) Allanite disseminates in the chlorite fabrics, and apatite and monazite form cystic aggregates wrapped by the chlorite matrix. (D) Minor orthoamphibole is surrounded by the chlorite fabrics, and allanite and apatite are disseminated in the chlorite fabrics

Coompana Province, Australia

The Coompana Province is a completely buried crustal block that occurs at the junction of the West, South and North Australian cratons (Fig. 5A and C). The basement geology of the Coompana Province used to be poorly known due to the lack of outcrops and our understanding of this Province has been greatly enhanced by the recent Coompana Drilling Project, including 8 drill holes (CDP001 to CDP008; Pawley et al. 2018; Wise et al. 2018; Jagodzinski et al. 2019).

The oldest rocks are the c. 1618 Ma Toolgana Supersuite intersected in drill holes CDP001 and CDP006. They are similar in age and geochemistry to the c. 1647–1608 Ma St Peter Suite in the Gawler Craton (Symington et al. 2014; Reid et al., 2019) and are interpreted to represent the subduction and arc magmatism outboard of the Gawler Craton (Kirkland et al. 2017; Hartnady et al. 2020). Geophysical datasets suggest these rocks are common throughout the Coompana Province, defining the NNE-trending grain prominent on aeromagnetic and gravity datasets (Pawley et al. 2020a). Following the cessation of arc-related magmatism at c. 1526 Ma (i.e., Bunburra Suite), the A-type Undawidgi Supersuite developed during intracontinental extension during c. 1505–1487 Ma (Dutch 2018; Wise et al. 2018). From c. 1200–1070 Ma, the Coompana Province underwent two stages of intraplate magmatism, including the 1190–1140 Ma Moodini Supersuite granitoids and the c. 1074 Ma mafic Giants Head Suite (Jagodzinski et al. 2019; Pawley et al. 2020a; Spaggiari et al. 2020). The Moodini Supersuite granitoids are strongly magnetic and form a NE-trending belt of overlapping plutons within a shear-bounded corridor (Pawley et al. 2020a). The Moodini Supersuite magmatism was contemporaneous with widespread crustal anatexis, with metamorphism constrained at c. 1163 Ma (Jagodzinski et al. 2019). The Giants Head Suite includes mafic intrusive and extrusive rocks interpreted to form part of the c. 1075 Ma Warakurna Supersuite (Pawley et al. 2020b).

Sample 4381386 is allanite-bearing monzogranite sampled from the depth of 532.19–532.28 m in drill hole CDP004 (Fig. 7A). The allanite-bearing monzogranite belongs to the Moodini Supersuite and crosscuts the Undawidgi Supersuite, having gradational boundaries with the surrounding rocks. It is reddish brown-grey, fine- to medium-grained and equigranular. It is composed of plagioclase, K-feldspar and quartz, minor biotite and accessory minerals allanite, titanite, apatite and zircon. A single allanite grain (3.5 mm×1 mm) is hosted in plagioclase and contacts with biotite and quartz. The sample has experienced post-crystallisation hydrothermal alteration, with epidote locally replacing plagioclase. Previously, a biotite leuco-monzogranite from 536.5 to 537.38 m and a foliated dark green quartz-syenite from 674.35 to 676.57 m yielded zircon U–Pb dates of 1197 ± 12 Ma and 1174 ± 9 Ma, respectively (Jagodzinski et al. 2019). More recently, an apatite Lu–Hf date was obtained of 1195 ± 35 Ma (Bedoya et al. 2024).

Fig. 7

Microphotographs of sample from Coompana Province, reflected light. (A) Sample 4,381,386 from drill hole CDP004, allanite intergrows with quartz, plagioclase and biotite. Titanite occurs as accessory mineral in the monzogranite. (B) Sample 4,381,382 from drill hole CDP001, epidote occurs in quartz vein crosscutting the altered monzogranite. (C) Sample 4,381,384 from drill hole CDP001, plagioclase is partially altered to epidote in the monzogranite. Titanite occurs as accessory mineral. (D) Samples 4,381,389 from drill hole CDP005, epidote-chlorite veins/veinlets crosscut the granite. (E) Sample 4,381,390 from drill hole CDP005. Granite is crosscut by quartz-epidote veins. Plagioclase grains are partially altered to epidote and biotite is partially altered to chlorite. (F) Sample 4,381,391 from drill hole CDP005, granite is altered by epidote and chlorite. Epidote replaces plagioclase or forms corona outside titanite

Samples 4381382 and 4381384 are sampled from depths of 448.27–448.37 m and 501.58–501.7 m in drill hole CDP001 (Fig. 7B and C). They are medium-grained massive monzogranite of the Moodini Supersuite that intruded into the migmatitic rocks of the Toolgana Supersuite. The monzogranite is composed of K-feldspar, plagioclase, quartz, and biotite with accessory minerals of titanite, apatite, zircon and allanite. Both samples 4381382 and 4381384 have experienced hydrothermal alteration, with the development of epidote-quartz veins crosscutting the monzogranite (Fig. 7B) or epidote replacing plagioclase (Fig. 7C). Anhedral epidote grains are hosted in quartz veins in sample 4381382 (Fig. 7B). Plagioclase grains in sample 4381384 are greenish but dominated by plagioclase rather than epidote, indicating the incomplete epidotization of plagioclase (Fig. 7C). Previously, zircon U–Pb geochronology of monzogranite at 421.70–423.15 m indicates the monzogranite was derived by the partial melting of the c. 1618 Ma Toolgana Supersuite orthogneiss (Dutch 2018), and emplaced at 1141 ± 15 Ma (Jagodzinski et al. 2019). This is confirmed by apatite Lu–Hf and U–Pb dates of 1154 ± 11 Ma and 1135 ± 16 Ma, respectively (Bedoya et al. 2024).

Samples 4381389, 4381390 and 4381391 are coarse-grained granites sampled from depths of 566.96–567.1 m, 637–637.19 m, and 645.97–646.13 m, respectively, in drill hole CDP005 (Fig. 7D–F). The granite samples are reddish, and composed of K-feldspar, plagioclase, quartz, biotite, and accessory minerals of titanite, apatite and zircon. The samples are characterized by extensive hydrothermal alteration of albite, sericite, chlorite, epidote and fluorite (Jagodzinski et al. 2019). Epidote normally replaces plagioclase or occurs in the epidote-chlorite-quartz veins crosscutting the coarse-grained granite. Coarse-grained titanite is subhedral to euhedral and interlocks with plagioclase, quartz and biotite or occurs as inclusions in plagioclase. Titanite is enclosed by epidote-chlorite or crosscut by epidote-quartz veins (Fig. 7E and F), indicating the development of epidote alteration is later than the titanite crystallisation. Granite sampled at 478.14–479.96 m yielded consistent zircon U–Pb, apatite Lu–Hf and apatite U–Pb dates of 1148 ± 7 Ma, 1164 ± 16 Ma and 1149 ± 16 Ma, respectively, representing the timing of magmatic crystallisation (Jagodzinski et al. 2019; Bedoya et al. 2024).

Analytical methods

All samples were mounted in 2.5 cm diameter epoxy mounts, and RL or BSE images were obtained to reveal the petrogenetic context of the analysed epidote group minerals.

Lu–Hf geochronology and trace elements

Lu–Hf geochronology and trace elements analyses of epidote group minerals, garnet and apatite were conducted at Adelaide Microscopy, The University of Adelaide, following a methodology primarily based on Simpson et al. (2021), which we briefly outline here and summarized in Table 1. The Lu–Hf analyses were performed using a RESOlution 193 nm excimer laser ablation system coupled with an Agilent 8900 ICP-MS/MS. NH₃ was employed as the reaction gas, supplied as a mixture of 10% NH₃ in 90% He. Laser beam diameters ranged from 67 to 120 μm for allanite, apatite and garnet, and 173 μm for clinozoisite and epidote. Laser repetition rates of 10 Hz were used to enhance sensitivity. The following isotopes (mass shifts included in brackets) were measured: ²⁷Al, ⁴³Ca, ^(47+66)Ti, ⁵⁷Fe, ⁸⁸Sr, ^(89+83)Y, ^(90+83)Zr, ^(140+15)Ce, ¹⁴⁶Nd, ¹⁴⁷Sm, ¹⁷²Yb, ¹⁷⁵Lu, ^(175+82)Lu, ^(176+82)Hf, ^(178+82)Hf. ¹⁷⁵Lu was measured as a proxy for ¹⁷⁶Lu, and ¹⁷⁸Hf was measured as a proxy for ¹⁷⁷Hf. Calculation of ¹⁷⁶Lu and ¹⁷⁷Hf was conducted assuming stable present-day ¹⁷⁶Lu/¹⁷⁵Lu and ¹⁷⁷Hf/¹⁷⁸Hf ratios, with detailed procedures outlined in Simpson et al. (2021).

Table 1 Analysis and LA-ICP-MSMS tuning parameters for Lu–Hf analyses

LADR (Norris and Danyushevsky 2018) was used as the data reduction software. Background-subtracted isotopic ratios were normalised to NIST-610 glass using the Nebel et al. (2009) ID-MC-ICP-MS isotopic compositions of ¹⁷⁶Lu/¹⁷⁷Hf = 0.1379 ± 0.005 and ¹⁷⁶Hf/¹⁷⁷Hf = 0.282111 ± 0.000009. NIST-610 was analysed every 20–40 unknown samples and further used to correct the isotope ratios for instrument mass bias and drift. Subsequently, a matrix-fractionation correction needs to be conducted on the ¹⁷⁶Lu/¹⁷⁷Hf ratio. Although a reference material with the same matrix as the analysed samples is desirable, it has been shown that the correction factors for materials with similar ablation characteristics (crater depth, ablation rate), analysed with the same analytical conditions, are indistinguishable (Glorie et al. 2023a, b). Here, Högsbo garnet (1029 ± 1.7 Ma; Romer and Smeds 1996) was utilized to correct the matrix-induced fractionation for both the garnets and the epidote group minerals in this study, due to their similar ablation characteristics (Figure S1) and the absence of Lu–Hf downhole fractionation (Simpson et al. 2021, 2022; Glorie et al. 2023b). Reference garnet Black Point BP-1 (concordant U-Pb age of 1745 ± 14 Ma, multi-session Lu-Hf age of 1740 ± 7 Ma; Lane, 2011; Glorie et al. 2024) and reference allanite LE40010 (²⁰⁷Pb-corrected ²⁰⁶Pb/²³⁸U age of 2654 ± 63 Ma; Yang et al. 2022) were used to verify the accuracy of the calibration procedure. Additionally, some apatite samples were dated by the Lu–Hf method in the same rocks as the epidote group minerals. OD-306 apatite (1597 ± 7 Ma; Thompson et al. 2016) was used to correct the matrix-induced fractionation for apatite and reference apatite Bamble-1 (corrected Lu–Hf age: 1097 ± 5 Ma) was monitored for accuracy checks (Glorie et al. 2022a, b). IsoplotR was used to calculate the inverse isochron ages (Vermeesch 2018). The ¹⁷⁶Lu decay constant of 0.00001867 ± 0.00000008 Myr^− 1 (Söderlund et al. 2004) was used for all age calculations. Inverse isochrons were chosen over normal isochrons as they provide a more robust age calculation method for datasets with strong uncertainty correlations (Li and Vermeesch 2021). For samples with insufficient variability in ¹⁷⁷Hf/¹⁷⁶Hf ratios between individual analyses, unanchored isochron regression cannot robustly constrain the Lu–Hf age. Fortunately, the range of terrestrial initial ¹⁷⁷Hf/¹⁷⁶Hf ratios is relatively narrow compared to the Lu/Hf ratios typically observed in epidote group minerals. Hence, following Glorie et al. (2023a, 2024), an anchored regression is used to calculate the Lu–Hf age for samples with low ¹⁷⁷Hf/¹⁷⁶Hf variability. The anchored regression algorithm is detailed in Vermeesch et al. (2024) and the initial ¹⁷⁷Hf/¹⁷⁶Hf anchor used in this work is 3.55 ± 0.05, which covers the entire range of terrestrial possibilities (Spencer et al. 2020). Trace elements were processed using the glass NIST-610 as a standard, all allanite analyses were assumed to have 6.9 wt% Ca and all epidote-clinozoisite analyses were assumed to have 16.6 wt% Ca. Garnet analyses of Yambah were assumed to have 11.1 wt% Al (probe data in Bendall 2000), and garnet analyses of sample 207,824 were assumed to have 12 wt% Al. All apatite analyses were assumed to have 39.4 wt% Ca.

U–Pb geochronology

LA-ICP-MS monazite and titanite U–Pb analyses were collected on an Agilent 8900 ICP-MS coupled with a RESOlution 193 nm excimer laser ablation system at Adelaide Microscopy, The University of Adelaide. The ablation of monazite was performed with a spot size of 13 μm, a frequency of 5 Hz, and a fluence of 4–5 J/cm². The total acquisition time for each analysis was 60 s, including 30 s of background measurement, followed by 30 s of laser ablation. Common lead was not corrected in the age calculations due to unresolvable interference of ²⁰⁴Hg on the ²⁰⁴Pb isotope peak. Mass bias, elemental fractionation, and instrument drift were corrected using the monazite standard MAdel (TIMS normalization data: ²⁰⁷Pb/²⁰⁶Pb age = 492.01 ± 0.77 Ma, ²⁰⁶Pb/²³⁸U age = 517.9 ± 2.6 Ma and ²⁰⁷Pb/²³⁵U age = 513.13 ± 0.20 Ma; updated from Payne et al. (2008) with additional TIMS analyses). Data accuracy was monitored by analyses of in-house monazite standard 222 (SHRIMP ²⁰⁶Pb/²³⁸U age 450.2 ± 3.4 Ma; Maidment 2005) and Ambat (²⁰⁶Pb/²³⁸U age c. 520 Ma). Standards were analysed every 12–15 unknown analyses. Titanite grains were ablated with a frequency of 5 Hz, fluence of 3.5 J/cm² and a spot size of 43 μm. The acquisition time of each analysis was 70 s, comprising 30 s of background measurement and 40 s of ablation. Titanite U–Pb data were corrected for mass bias, elemental fractionation and instrument drift based on the measured isotopic ratios of the primary titanite reference, MKED-1 (TIMS normalization data: ²⁰⁷Pb/²⁰⁶Pb age = 1521.02 ± 0.55 Ma; Spandler et al. 2016). Mt Painter was analysed as a secondary standard to measure data accuracy by standard–sample bracketing every ∼ 15 unknown analyses. Data reduction was conducted in LADR (Norris and Danyushevsky 2018) and age calculations were performed in IsoplotR (Vermeesch 2018).

Sm–Nd geochronology

Mineral separates were prepared via magnetic and heavy liquid separation techniques which were also used to remove impurities from the garnet fraction. Surface contamination on the handpicked mineral separates was removed by an ultrasonic cleaner in 1 M HCl solution. A whole-rock fraction of the matrix surrounding the garnets was crushed and milled. Between 125 and 300 mg and 100–150 mg of sample was used for mineral separates and whole rock, respectively. The mineral separates were milled under ethanol in an agate mortar to a fine grain size. To minimize contamination of mineral fractions by REE-rich inclusions, the milled fractions were leached in hot 1 M HF for 1 h. The leachate was pipetted from the residual solid material and the solid material washed in cold 6 M HCl separately to remove any trace of the leachate fraction. For the whole-rock component, around 150 mg of milled whole rock was dissolved in steel jacketed, high pressure Teflon bombs using a HNO₃–HF acid mixture heated to 200 °C for periods of 5 days. All samples were spiked with a mixed ¹⁴⁷Sm–¹⁵⁰Nd spike prior to dissolution. Nd and Sm isotopic ratios were measured by thermal ionization mass spectrometry (TIMS) on a Finnigan MAT 262 system in static mode at the University of Adelaide. The isotopic ratios were corrected for fractionation to ¹⁴⁶Nd/¹⁴⁴Nd = 0.721903 and to a ¹⁵²Sm/¹⁴⁹Sm ratio of 1.9347. Spiked samples of BCR-1 yielded a ¹⁴³Nd/¹⁴⁴Nd ratio of 0.512598 ± 17 after spike unmixing. The ¹⁴³Nd/¹⁴⁴Nd reproducibility of the internal standard over the course of the study (n = 10) was 0.511602 ± 0.00001. For age calculations Sm/Nd errors were estimated to be ± 0.3%. Age calculation (reported at 95% confidence) are performed in IsoplotR (Vermeesch 2018) based on a decay constant for ¹⁴⁷Sm of (6.524 ± 0.024) × 10^− 12 a^− 1 (Villa et al. 2020). The total procedural Sm and Nd blanks were 100 pg.

Results

All geochronology data is summarized in Table S2–S7, and all ages are reported with 2σ absolute uncertainties.

Anorthosite complex, SW Greenland

Clinozoisite sample 207824 was analysed over two analytical sessions, yielding Lu concentrations of 0.27–28 ppm, ¹⁷⁶Lu/¹⁷⁷Hf ratios between 0.06 and 4.55, and ¹⁷⁷Hf/¹⁷⁶Hf ratios between 2.84 and 3.57. Its Al/(Al + Fe) ratios range from 0.72 to 0.95, confirming the dominance of clinozoisite endmember. Fifty-nine spots have ¹⁷⁶Lu/¹⁷⁷Hf ratios lower than 1, and analyses with ¹⁷⁶Lu/¹⁷⁷Hf ratios higher than 1 normally have higher Lu concentrations of 3.4–28 ppm. All analyses define an inverse isochron Lu–Hf age of 2372 ± 69 Ma (n = 70, MSWD = 1.1; Fig. 8A). Additionally, thirty-eight Lu–Hf analyses of garnet yield an inverse isochron age of 2577 ± 15 Ma (n = 38, MSWD = 2.8; Fig. 8B) anchored by an initial ¹⁷⁷Hf/¹⁷⁶Hf ratio of 3.55 ± 0.05. Garnet has Lu concentrations of 73–171 ppm.

Fig. 8

Lu–Hf inverse isochrons for (A) clinozoisite in sample 207,824, (B) garnet in sample 207,824 and (C) allanite LE40010 from Fiskenæsset anorthosite complex

Allanite Sample LE40010 was analysed over two analytical sessions. Despite all analyses have relatively uniform trace elements concentrations, e.g., Lu contents of 1.4–1.9 ppm and Ce contents of 50,325–68,919 ppm, the resulting ¹⁷⁶Lu/¹⁷⁷Hf ratios (3.08–6.21) and ¹⁷⁷Hf/¹⁷⁶Hf ratios (2.42–3.01) define two clusters (Fig. 8C). Cluster with lower ¹⁷⁶Lu/¹⁷⁷Hf ratios corresponds to allanite portion with lower Lu/Ce ratios (Figs. 1D and 8C), indicating the allanite shard is not compositionally homogeneous. The clustered ¹⁷⁶Lu/¹⁷⁷Hf ratios and ¹⁷⁷Hf/¹⁷⁶Hf ratios yield unanchored inverse isochron ages with uncertainties of 300–500 Ma and thus an initial ¹⁷⁷Hf/¹⁷⁶Hf ratio of 3.55 ± 0.05 was used to anchor the inverse isochron. Thirty-nine analyses, analysed with a 67 μm laser beam size, in session 2 yielded an inverse isochron Lu–Hf age of 2591 ± 135 Ma (n = 39, MSWD = 0.82; Fig. 8C). Using larger laser spots of 100 μm, twenty-seven analyses from session 3 yield smaller uncertainties and define an inverse isochron Lu–Hf age of 2703 ± 116 Ma (n = 27, MSWD = 0.55). All analyses from the two sessions combined yield an inverse isochron Lu–Hf age of 2660 ± 109 Ma (n = 66, MSWD = 0.77; Fig. 8C).

Huayangchuan uranium-polymetallic deposit, Qinling Orogen, China

Allanite sample HYC was analysed by fifty-seven laser spots over two analytical sessions. The allanite separates yielded very dispersed Lu contents of 4.3–25 ppm, ¹⁷⁶Lu/¹⁷⁷Hf ratios of 7.61–22.11 and ¹⁷⁷Hf/¹⁷⁶Hf ratios of 0.80–2.73. There is a positive correlation between Lu contents and final ¹⁷⁶Lu/¹⁷⁷Hf ratios. Despite the different BSE brightness and zonation of allanite grains, all analyses plot along a single isochron (Fig. 9A). Thirty-six analyses from session 2 define an inverse isochron age of 1812 ± 39 Ma (n = 36, MSWD = 2) anchored by an initial ¹⁷⁷Hf/¹⁷⁶Hf ratio of 3.55 ± 0.05. Twenty-one analyses from session 3 define an identical inverse isochron age of 1810 ± 32 Ma (n = 21, MSWD = 1.5) using the same anchor ¹⁷⁷Hf/¹⁷⁶Hf ratio. However, unanchored inverse isochron ages of 1870 ± 39 Ma and 1854 ± 43 Ma are obtained from two sessions, with initial ¹⁷⁶Hf/¹⁷⁷Hf ratios of 0.2663 ± 0.0051 and 0.2712 ± 0.0057, respectively. The initial ¹⁷⁶Hf/¹⁷⁷Hf ratios are lower than the recommended crustal Hf isotope ratios of ∼ 0.280 – ∼0.282 (Spencer et al. 2020; Simpson et al. 2021) and the unanchored Lu–Hf ages are older than the zircon U–Pb ages of the host pegmatite, and thus the anchored inverse isochron ages are preferred.

Fig. 9

(A) Lu–Hf inverse isochron for sample HYC, including data from two laser sessions. (B) Zr vs. Sr plot for two types of allanite at Huangjiagou. (C) Lu–Hf inverse isochron for type I allanite at Huangjiagou. (D) Lu–Hf inverse isochron for type II allanite at Huangjiagou. (E) U–Pb age for monazite inclusions in type II allanite at Huangjiagou, dashed ellipses with potential Pb loss are excluded for age calculation. (F) U–Pb age for titanite at Huangjiagou

Allanite samples HJG-85, HJG-89 and HJG-90 contain two types of allanite, corresponding to magmatic (type I) and hydrothermal (type II) stages, respectively. Two types of allanite have distinct geochemistry such as the mobile Sr and immobile Zr concentrations (Fig. 9B). Type I allanite has Sr contents of 63–1289 ppm and Zr contents of 2.8–7.3 ppm while type II allanite has Sr contents of 1767–3217 ppm and Zr contents of 0.25–1.4 ppm. Two types of allanite have comparable REE contents, including 3.1 × 10³–8.4 × 10⁴ ppm Ce and 0.23–4.3 ppm Lu in type I allanite and 4.9 × 10⁴–9.6 × 10⁴ ppm Ce and 1.0–3.3 ppm Lu in type II allanite. However, two types of allanite are characterized by different Lu/Hf ratios. Type I allanite from Huangjiagou has ¹⁷⁶Lu/¹⁷⁷Hf ratios that vary between 0.004 and 8.75 and ¹⁷⁷Hf/¹⁷⁶Hf ratios between 2.53 and 3.57, and high ¹⁷⁶Lu/¹⁷⁷Hf ratios correspond to high Lu centration. Seventy-eight analyses define an inverse isochron age of 1821 ± 74 Ma (n = 78, MSWD = 0.83; Fig. 9C). Type II allanite in sample HJG-90 has higher ¹⁷⁶Lu/¹⁷⁷Hf ratios that vary between 8.89 and 69.30, and ¹⁷⁷Hf/¹⁷⁶Hf ratios between 2.45 and 3.60. Forty-four analyses define an anchored inverse isochron age of 215 ± 34 Ma using an initial ¹⁷⁷Hf/¹⁷⁶Hf ratio of 3.55 ± 0.05 (n = 44, MSWD = 1; Fig. 9D).

Monazite and apatite aggregates occur as inclusions in type II allanite in sample HJG-90. Monazite grains are 10–500 μm, and are unzoned in BSE images (Fig. 3F). Thirty analyses from monazite are discordant due to the presence of common Pb and Pb loss. Seventeen analyses define an isochron with a lower intercept age of 214.1 ± 4.0 Ma (n = 17, MSWD = 1.4, prob. = 0.15; Fig. 9E), representing the crystallisation time of monazite. Analyses from titanite cogenetic with type I allanite in samples HJG-85 and HJG-89 (Fig. 3B–D) are plotted as a single isochron, and they collectively yield a lower intercept age of 131.5 ± 1.6 Ma on the Tera-Wasserburg plot (Fig. 9F).

Xikuangshan Fe-Cu deposit, Kangdian IOCG province, China

Epidote sample Xikuangshan was analysed over two analytical sessions, and epidote grains with higher ¹⁷⁶Lu/¹⁷⁷Hf ratios were targeted during the second session. Epidote Xikuangshan has Al/(Al + Fe) ratios of 0.24–0.78, Ce contents of 24–965 ppm and Lu contents of 0.17–2.1 ppm. The majority of epidote analyses have Lu contents lower than 1 ppm and yielded large uncertainties for the ¹⁷⁷Hf/¹⁷⁶Hf ratios, leading to a significant uncertainty on the calculated Lu–Hf age. Fifty-two analyses define an inverse isochron age of 1008 ± 103 Ma (n = 52; MSWD = 1.1; Fig. 10).

Fig. 10

Lu–Hf inverse isochron for epidote from Xikuangshan deposit

Yambah shear zone, Central Australia

Allanite, garnet, and apatite were analysed for Lu–Hf geochronology in sample Yambah. Both allanite and apatite are distributed evenly throughout the chlorite foliation in the matrix. In addition allanite occurs as inclusions in garnet, indicating that at least some of the allanite predates the growth of garnet. Two hundred and forty-six allanite analyses have dispersed Lu contents of 1.3–23 ppm, ¹⁷⁶Lu/¹⁷⁷Hf ratios of 0.04–98.96 and ¹⁷⁷Hf/¹⁷⁶Hf ratios of 0.76–3.55, and yield an inverse isochron age of 432 ± 7 Ma (n = 246, MSWD = 1.7; Fig. 11A). The majority of allanite contains Lu contents between 10 and 22 ppm, corresponding to higher ¹⁷⁶Lu/¹⁷⁷Hf ratios.

Fig. 11

Geochronology results from Yambah Shear Zone. Lu–Hf inverse isochrons for (A) allanite, (B) garnet and (C) apatite. (D) Monazite U–Pb age, discordant analyses are marked as dashed ellipses and excluded for age calculation. (E) Sm–Nd isochron for a garnet-hornblende-staurolite-bearing schist interlayered with garnet-chlorite-clinozoisite-bearing schist

Three ∼ 1 cm garnet grains were analysed from sample Yambah. During the analysis, spots from both cores and rims were analysed to cover areas with different Lu concentrations and Lu/Hf ratios. Care was taken to avoid hitting allanite and apatite inclusions. One hundred and thirty-one analyses yield an initial ¹⁷⁷Hf/¹⁷⁶Hf ratio of 3.54 ± 0.01 and an inverse isochron age of 380 ± 4 Ma (n = 131; MSWD = 1.1; Fig. 11B). Analyses with ¹⁷⁶Lu/¹⁷⁷Hf ratios > 50 come from the garnet core, which has Lu concentrations of 19–297 ppm. A similar inverse isochron Lu–Hf age of 373 ± 3 Ma (n = 60, MSWD = 1.3) was obtained for apatite in the matrix of sample Yambah (Fig. 11C). Yambah garnet has Lu contents of 0.02–297 ppm and apatite has Lu contents of 2.6–40 ppm. Six monazite grains hosted in the biotite matrix were analysed and thirty concordant analyses defined two age populations. Three analyses yielded old ²⁰⁶Pb/²³⁸U dates from 414 to 444 Ma, and the rest 27 analyses yielded a weighted mean average ²⁰⁶Pb/²³⁸U age of 370 ± 2 Ma (n = 27, MSWD = 0.65), slightly younger than the garnet and similar to the apatite Lu–Hf ages (Fig. 11D).

Garnet Sm–Nd geochronology was undertaken on a sample from a garnet-hornblende-staurolite-bearing schist interlayered with garnet-chlorite-allanite-bearing schist targeted for Lu-Hf dating. Ideally the same sample would be used for Sm-Nd and Lu-Hf dating, however due to the inability to remove allanite all inclusions from the garnets in the chlorite schist, an allanite-free sample of garnet-hornblende-staurolite schist from an adjacent layer in the shear zone was chosen as a comparison. Individual garnets approximately 2 mm in diameter were cut from the sample and crushed in a mortar and pestle. They were combined with magnetically separated hornblende and staurolite fractions and a whole rock. Together with the leached garnet fraction, the mineral and whole rock fractions combine to produce an age of 384.8 ± 5.2 Ma (Fig. 11E).

Coompana Province, Australia

A single allanite grain in sample 4381386 from drill hole CDP004 was analysed for Lu–Hf geochronology (Fig. 7A). Thirty-one analyses were collected from this grain and they yielded Lu contents of 0.69–16 ppm and an inverse isochron Lu–Hf age of 1171 ± 35 Ma (n = 24; MSWD = 0.84; Fig. 12A). Titanite from the same sample contains varying amounts of common Pb and some evidence for potential Pb loss, i.e., deviation from the regression isochron. Analyses with potential Pb loss were excluded (dashed ellipse in Fig. 12B) and sixteen leftmost analyses on the Tera-Wasserburg plot yielded a U–Pb intercept age of 1170 ± 12 Ma (n = 16; MSWD = 0.21; Fig. 12B).

Fig. 12

Geochronology results from Coompana Province. (A and B) Allanite Lu–Hf and titanite U–Pb ages from drill hole CDP004, dashed ellipses with potential Pb loss are excluded for age calculation. (C and D) Epidote Lu–Hf and titanite U–Pb ages from drill hole CDP001, Lu–Hf data are collected from two laser sessions; (E and F) Epidote Lu–Hf and titanite U–Pb ages from drill hole CDP005

Epidote in samples 4381382 and 4381384 from drill hole CDP001 has low ¹⁷⁶Lu/¹⁷⁷Hf ratios and high ¹⁷⁷Hf/¹⁷⁶Hf ratios, indicative of low Lu and high initial Hf concentrations typically found in epidote. Epidote from drill hole CDP001 has Lu contents of 0.03–18 ppm, predominantly lower than 4 ppm. Fifty-six analyses from session 3 yielded an inverse isochron age of 1086 ± 59 Ma (n = 56; MSWD = 1.1), and grains with higher ¹⁷⁶Lu/¹⁷⁷Hf ratios were analysed again in session 4. One hundred and five analyses from two sessions collectively define an inverse isochron age of 1079 ± 30 Ma (n = 105; MSWD = 1.2; Fig. 12C). Titanite from sample 4381384 yielded a U–Pb lower intercept age of 1196 ± 9 Ma (n = 31; MSWD = 1.7; Fig. 12D).

Compared to the epidote from drill hole CDP001, the epidote in samples 4381389, 4381390 and 4381391 from drill hole CDP005 has lower Lu contents (0.27–19 ppm, predominantly < 2 ppm), higher initial Hf and lower ¹⁷⁶Lu/¹⁷⁷Hf ratios. One hundred and twelve analyses from these three samples yield an inverse isochron age of 1072 ± 60 Ma (n = 112; MSWD = 0.79; Fig. 12E). Titanite from sample 4381389 yielded a U–Pb intercept age of 1175 ± 8 Ma (n = 35; MSWD = 1.3; Fig. 12F).

Discussion

Allanite, clinozoisite and epidote from various geological settings have variable Lu concentrations ranging from 0.03 to 28 ppm, thus yielding significant variability in ¹⁷⁶Lu/¹⁷⁷Hf ratios to define robust isochrons (Figs. 8, 9, 10, 11 and 12). In spite of the presence of common Hf in the analysed epidote group minerals and the lack of matrix matched reference materials, the Lu–Hf geochronology results corrected by Högsbo garnet have provided age constraints on magmatism, REE mineralization, metamorphism and hydrothermal alteration. In the following section, we will interpret these dates and compare Lu-Hf dates with age constraints obtained by other geochronology methods, and discuss the feasibility and geological significance of the Lu–Hf geochronology for the epidote group minerals.

Accurate and reproducible Lu–Hf dates for magmatic and hydrothermal allanite

Mesoproterozoic magmatic allanite in the Coompana Province

Allanite can provide geochronological constraints on a range of geological processes, such as the development of magmatic systems and REE ore deposits. Four magmatic allanite samples were analysed in this study. For the Coompana Province, the allanite Lu–Hf age of 1171 ± 35 Ma, obtained for sample 4381386 in drill hole CDP004 (Lu concentrations of 0.69–16 ppm), agrees within uncertainty with the titanite U–Pb age of 1170 ± 12 Ma in the same sample (Fig. 12A and B). The newly obtained allanite and titanite ages are also consistent with the published zircon U–Pb age of 1174 ± 9 Ma for drill hole CDP004 (Jagodzinski et al. 2019), recording the crystallisation time of the monzogranite. The identical geochronology results obtained for the allanite Lu–Hf, titanite U–Pb and zircon U–Pb systems underscore the viability of in situ Lu–Hf geochronology to date allanite. The results also support the use of a garnet reference material do date allanite, giving confidence to the calibration procedures used in this paper.

Archean magmatic allanite in the Fiskenæsset complex

The oldest allanite sample analysed in this study, LE40010 from a quartz pegmatite intruded in the Fiskenæsset anorthosite complex, yields a multi-session Lu–Hf isochron age of 2660 ± 109 Ma (Fig. 8C), which is consistent with published allanite U–Pb ages by ID-TIMS (2646 ± 94 Ma & 2735 ± 226 Ma; Smye et al. 2014) and LA-(MC)-ICP-MS (2613 ± 43 Ma & 2654 ± 63 Yang et al. 2022), and the intrusion of muscovite granite sheet at 2660 ± 20 Ma in the Fiskenæsset anorthosite complex (Pidgeon and Kalsbeek 1978). The large age uncertainty may be attributed to the low Lu concentration (1.4–1.9 ppm; Fig. 8C) and clustered Lu/Hf ratios in the allanite shard, implying that LE40010 is unsuitable for future Lu-Hf reference material.

Palaeoproterozoic magmatic and triassic hydrothermal allanite in Qingling pegmatites

Magmatic allanite samples obtained from Huayangchuan (HYC) and Huangjiagou (HJG-85, HJG-89 and HJG-90) are hosted in c. 1.81–1.83 Ga pegmatites (Yang et al. 2019; Zhang et al. 2022; Li et al. 2023), and modified by later high-temperature metamorphism and hydrothermal metasomatism (Zhang et al. 2023). For HYC allanite, anchored isochron Lu–Hf ages of 1812 ± 39 Ma and 1810 ± 32 Ma were obtained from two analytical sessions (Fig. 9A). These two ages are the same within uncertainty, demonstrating good reproducibility of the analytical method and are also consistent with published zircon U–Pb ages for the Huayangchuan pegmatites (1807 ± 14 Ma, 1826 ± 8 and 1829 ± 11 Ma; Yang et al. 2019; Li et al. 2023). Type I allanite from Huangjiagou yielded a Lu–Hf inverse isochron age of 1821 ± 74 Ma (Fig. 9C), which is also in agreement with published allanite and zircon U–Pb ages (1829 ± 23 Ma and 1826 ± 11; Zhang et al. 2022). Compared to the HYC allanite, the 1821 ± 74 Ma type I allanite from Huangjiagou has lower Lu concentrations (4.3–25 ppm for HYC vs. 0.23–4.3 ppm for Huangjiagou), inducing both lower ¹⁷⁶Lu/¹⁷⁷Hf ratios and lower age precision (Fig. 9A and C).

In addition, a younger generation of hydrothermal allanite (type II) was recognized at Huangjiagou and yields an inverse isochron Lu–Hf age of 215 ± 34 Ma (Fig. 9D). Despite the low accumulation of radiogenic ¹⁷⁶Hf in the young allanite, resulting in large analytical uncertainties, the allanite Lu–Hf age is notably older than the published allanite U–Pb age (131 ± 16 Ma; Zhang et al. 2023) as well as the titanite U–Pb age obtained for the same rock (131.5 ± 1.6 Ma; Fig. 9F). However, the Lu–Hf age agrees with the U–Pb age (214.1 ± 4.0 Ma) obtained for monazite inclusions within type II allanite (Figs. 3F and 9E). Type II allanite occurs along the calcite-quartz veins (Fig. 3E and F), and the c. 215 Ma ages coincide with Triassic ages widely recorded in carbonatite dykes and associated REE mineralisation in the Qinling Orogen, such as the Huayangchuan, Huanglongpu and Huangshui’an deposits (Wei et al. 2023 and references therein). Despite the presence of monazite inclusions, type II allanite contains similar REE concentrations to type I allanite (Table S2), indicating type II allanite and cogenetic monazite are not formed by recrystallisation or metasomatism of type I allanite. The relatively higher Sr and lower Zr concentrations of type II allanite (Fig. 9B) imply it may be formed by hydrothermal fluids, potentially related to the Triassic carbonatite. Hence, we interpret the Triassic type II allanite Lu-Hf age of c. 215 Ma as the timing of new generation of REE mineralisation overprinting the Huangjiagou pegmatite, while the disparate U–Pb ages obtained for titanite and allanite (Zhang et al. 2023) potentially signify later thermal disturbances during the intrusion of adjacent giant Huashan Granite batholith (Fig. 2B; zircon U-Pb ages of 142.6–132.0 Ma, zircon saturation temperatures of 727 ± 20 °C; Li et al., 2018a, b) due to their lower closure temperatures.

Prograde metamorphism recorded by allanite during Alice Springs Orogeny

Metamorphic geochronology commonly depends on the U–Th–Pb dating of accessory minerals such as zircon, monazite, xenotime and apatite. However, the (re)crystallisation of these phases during metamorphism may be hindered by unfavourable bulk-rock compositions or low metamorphic temperatures, limiting the use of these geochronometers. Allanite is a common constituent in low–medium grade metamorphosed mafic–intermediate igneous rocks, quartzofeldspathic rocks and marls. The formation of solid solutions between allanite and clinozoisite-epidote, and their ability to incorporate high contents of trace elements in their lattice make them stable over a wide range of P–T conditions (Franz and Liebscher 2004; Grapes and Hoskin 2004; Janots et al. 2008; Spear 2010). Hence, allanite geochronology (e.g., U–Th–Pb, Lu–Hf) may provide additional and unique insights into metamorphic histories (e.g., Janots et al. 2009; Gabudianu Radulescu et al. 2009; Janots and Rubatto 2014; Boston et al. 2017). Allanite from the amphibolite-facies Yambah Shear Zone, which records diachronous deformation during the Alice Springs Orogeny (Figs. 5B and 450–300 Ma; Hand and Sandiford 1999; Raimondo et al. 2014), occurrs both as inclusions in Yambah shear zone garnet and as a mineral in the shear fabric (Fig. 6), and its Lu–Hf geochronology and trace elements are used to interpret the evolution of Alice Springs Orogen.

Garnet and apatite from sample Yambah yield Lu–Hf isochron ages of c. 380–370 Ma, and the majority of monazite grains record U–Pb age of 370 Ma (Fig. 11). In comparison Sm–Nd data from garnet-hornblende-staurolite schist from an adjacent layer in the shear zone gives 383.4 ± 5.2 Ma. These c. 380 Ma ages coincide with geochronological constraints from West Bore Shear Zone and Edwards Creek Shear Zone in the Strangways Range (Fig. 5B; Möller et al. 1999; Ballèvre et al. 2000; Bendall 2000), representing the timing of peak metamorphism during the Alice Springs Orogeny (Bendall 2000). Allanite in sample Yambah, however, seems to reveal the early prograde metamorphic history of Alice Springs Orogeny. Allanite occurs within euhedral c. 380 Ma garnet and also within the chlorite-dominated foliation that wraps the garnet porphyroblasts (Fig. 6), suggesting at least part of the allanite crystallised prior to garnet growth. This interpretation is consistent with the high Lu concentrations of allanite (1.3–23 ppm, average 14 ppm) and low Lu concentration of the majority of garnet analyses (Table S2), otherwise the crystallisation of garnet would exhaust the whole rock Lu budget (Rubatto and Hermann 2007). Allanite Lu–Hf analyses produce an inverse isochron age of 432 ± 7 Ma, similar to the older monazite U–Pb dates from the Yambah Shear Zone (Fig. 11D). The associated MSWD value of 1.7, however, implies that they may not represent a statistically single age population. This could be the result of recrystallisation of some of the allanite within the foliation that wraps the garnet, meaning the analysed population of allanite has been isotopically disturbed to some extent. Conceivably the true growth age was ∼ 450 Ma, consistent with the evidence elsewhere in the orogenic belt for initial shear zone formation in the late Orodvician (Fig. 5B; Mawby et al. 1999; Möller et al. 1999; Scrimgeour and Raith, 2001). Nonetheless, the bulk of the allanite is still closer in age to 450 Ma than to the age of the garnet at 380 Ma. This suggests the bulk of the matrix allanite has been inherited into the present shear fabric rather than crystallising within it. It is worth noting that apatite has higher Lu concentrations of 2.6–40 ppm with an average of 16 ppm and contradicts its younger crystallisation age than garnet. This may be attributed to Hf loss or Lu gain in apatite during c. 370 Ma tectonic or hydrothermal event, which may have also modified the monazite (Fig. 11D) but had minor effect on the allanite Lu-Hf system.

The success of allanite Lu–Hf geochronology has enabled an opportunity to obtain time constraints of major silicate phases exhibiting known P–T stability, thereby facilitating the integration of petrological, microstructural, and geochronological data. Considering the estimated peak P–T conditions of the Yambah Shear Zones at 6 kbar, 600 °C around c. 380 Ma (Bendall 2000), the retention of older allanite Lu–Hf dates suggests that its Lu–Hf system is not easily reset by amphibolite-facies prograde metamorphism. This interpretation seems consistent with the preservation of Palaeoproterozoic and Triassic ages in the Huangjiagou allanite.

Hydrothermal alteration revealed by clinozoisite and epidote

Palaeoproterozoic hydrothermal clinozoisite in the Greenland anorthosite complex

In addition to being major rock-forming minerals in magmatic and metamorphic rocks, epidote and clinozoisite are common in hydrothermal systems, occurring in veins, segregations and cavities. With the new in situ Lu–Hf dating technique, it is possible to directly date the metasomatism and fluid-rock interaction involving the formation of epidote.

Sample 207824 from the Archean anorthosite complex contains clinozoisite veins crosscutting garnet porphyroblasts and plagioclase (Fig. 2B and C). The garnet was suggested to have formed during the emplacement of the c. 3.0 Ga anorthosite complex (Windley and Smith 1974; Polat et al. 2009), however, we obtained a garnet Lu–Hf isochron age of 2577 ± 15 Ma (Fig. 8B), which corresponds to the final assembly time of the North Atlantic Craton (c. 2.56 Ga; Dyck et al. 2015), and which is within error of the emplacement of the Qôrqut Granite Complex just north of the sample location (Næraa et al. 2014). Hence garnet Lu–Hf geochronology has provided direct evidence to the metamorphism of the anorthosite complex. The clinozoisite Lu–Hf age of 2372 ± 69 Ma (Fig. 8A) is younger than the garnet age, consistent with their crosscutting relationship. This clinozoisite-bearing hydrothermal vein has been previously described in the Fiskenæsset Complex (Polat et al. 2009, 2010) but without any age constraints. For the first time, we have constrained the age of hydrothermal alteration in the Archean anorthosite complex, although its genesis and geologic significance remain unclear. Mafic dykes that intruded the Archean anorthosite complexes yielded baddeleyite U–Pb ID-TIMS ages of 2365 ± 2 Ma and 2374 ± 4 Ma (Nilsson et al. 2013). It is thus conceivable the clinozoisite alteration in anorthosite developed at the same time as the brittle faulting and mafic dyke emplacement. Considering plagioclase commonly contains limited HREE and the small volume percent of clinozoisite (< 1%; Fig. 1B and C), Lu in the altered clinozoisite (0.27–28 ppm, average of 4.0 ppm) may have come from the hydrothermal fluids that leached HREE from the surrounding mafic rocks, or from retrograde breakdown of garnet (with Lu concentrations of 73–171 ppm) within the anorthosite.

Late mesoproterozoic hydrothermal epidote in the Coompana Province

Lu–Hf dates of 1079 ± 30 Ma and 1072 ± 60 Ma were obtained for epidote in drill holes CDP001 and CDP005 in the Coompana Province (Fig. 12C and E). These dates contrast with titanite U–Pb ages of 1196–1175 Ma (Fig. 12D and F), indicating hydrothermal alteration postdates the Moodini Supersuite magmatism. Instead, the epidote alteration may have developed during the last major magmatic episode in the Coompana Province, which is the ∼ 1074 Ma intrusion of the mafic Giants Head Suite (Fig. 5C; Wise et al. 2018; Jagodzinski et al. 2019).

Late Mesoproterozoic Hydrothermal epidote in the Xikuangshan deposit

Epidote from the host rock of the Xikuangshan deposit yielded a Lu–Hf age of 1008 ± 103 Ma (Fig. 10), which is consistent with allanite U–Pb age of 1015 ± 66 Ma from the adjacent Lanniping deposit (Su et al. 2021), and c. 1.0 Ga bimodal magmatism (Zhu et al. 2016; Chen et al. 2018) and mineralisation (Chen and Zhou 2012; Li 2013) in the Kangdian IOCG province, albeit with large uncertainty due to the low Lu concentrations and ¹⁷⁶Lu/¹⁷⁷Hf ratios. The c. 1.0 Ga epidote at Xikuangshan may either record hydrothermal remobilisation of c. 1.7 Ga ore minerals in the system or represent a new phase of mineralisation during intracontinental rifting.

Implications for mineral exploration

As one of the most ubiquitous minerals in prophylitic alteration in porphyry ore systems and calcic alteration in IOCG systems (Groves et al. 2010; Sillitoe 2010), epidote occurrences and geochemistry have been widely used as powerful vectoring tools in mineral exploration (Cooke et al. 2014; Pacey et al. 2020; Schlegel et al. 2022). However, similar epidote-bearing assemblages can also be developed in other geologic settings, such as regional low-grade metamorphism or hydrothermal alteration caused by later intrusions. It is vital to precisely constrain the alteration time, to link epidote alteration to regional mineralisation and potential ore-causative intrusions. The epidote Lu–Hf geochronology and recently published epidote U–Pb geochronology (Peverelli et al. 2021) complement each other, considering two isotopic systems may have different parent-to-daughter ratios and different closure temperatures, and thus assist differentiation of prophylitic/calcic alteration in ore systems from similar assemblages developed during non-mineralising processes. In addition, the novel in situ method does not require mineral separation or dissolution, providing the ability to quickly obtain large Lu–Hf geochronology datasets during mineral exploration. In addition to calcite and fluorite Lu–Hf geochronology (Simpson et al. 2022; Glorie et al. 2023a), epidote Lu–Hf dating opens up new opportunities to a wide range of applications for future exploration programs.

Current limitations

In contrast to (felsic) apatite, calcite and fluorite which do not incorporate significant concentrations of common Hf (Simpson et al. 2022; Glorie et al. 2023b), epidote group minerals, including allanite, can incorporate varying amounts of common Hf (Figs. 8, 9, 10, 11 and 12). As a result, the success rate of the Lu–Hf dating approach in epidote group minerals is intrinsically related to both the concentration of Lu and the ingrowth time for radiogenic Hf (Simpson et al. 2021, 2022). Magmatic and metamorphic allanite in low- to medium-grade metamorphic rocks may act as primary REE reservoirs, which often have Lu concentrations high enough for Lu–Hf geochronology. However, caution may be required in the selection of hydrothermal epidote samples, especially when epidote has a high modal proportion in a given rock. Large volumes of epidote would dilute Lu concentrations in a laser target, unless the hydrothermal fluids that conditioned crystallisation of epidote are Lu-rich. Hence, pre-screening of Lu-rich samples is recommended to identify grains or zones with high ¹⁷⁶Lu/¹⁷⁷Hf ratios (Table S2). Targeting such Lu-rich zones can significantly reduce the final age uncertainties (Fig. 12C). The use of large laser beam spot sizes can also significantly improve precision at the expense of lower spatial resolution. In this study, age precisions of 4–10% can be achieved for epidote samples with Lu concentrations as low as 1–2 ppm (Figs. 10 and 12E).

Another current limitation for Lu–Hf geochronology applied to epidote group minerals is the lack of suitable (matrix-matched) reference materials. In this paper, garnet reference materials Högsbo and Black Point were employed to correct and monitor the matrix-induced fractionation. While garnet is not an epidote-group mineral, it appears to have similar ablation characteristics. The agreement observed between allanite and epidote Lu–Hf ages and those obtained through other geochronological methods underscores the suitability of garnet as a reference material for epidote-group Lu–Hf dating. However, detailed characterisation of garnet crystals is required as they may have highly variable Lu concentrations such as the Högsbo garnet (Simpson et al. 2021). Appropriate spots sizes are required to measure the garnet reference materials and samples in the same detector mode (Glorie et al. 2024), i.e., pulse mode based on the ¹⁷⁵Lu count rates (< 1 Mcps) of all analysed epidote group minerals in this study. There is potential for improvement when suitable reference materials can be identified. For instance, Lu-rich allanite from pegmatites, particularly when not subjected to complex post-crystallisation alteration, holds promise as a closely matrix-matched reference material for epidote group minerals.

Conclusion

This study presents the first Lu–Hf dating of epidote group minerals by LA-ICP-MS/MS. Allanite, epidote and clinozoisite from a variety of geological environments yielded meaningful ages consistent with those from other geochronological methods, demonstrating the feasibility of the epidote group minerals Lu–Hf method. Magmatic allanite samples from pegmatite and monzogranite yielded Lu–Hf dates in agreement with the timing of magmatism in the Fiskenæsset complex, Coompana Province and Qingling Orogen. Additionally, Lu–Hf dates for hydrothermal allanite agree with the timing of REE mineralisation in the Qinling Orogen. Metamorphic allanite from the Yambah Shear Zone yielded Lu–Hf ages that are c. 50 Ma older than the c. 380 Ma peak metamorphism of the Alice Springs Orogeny, suggesting preservation of older dates during prograde metamorphism. Allanite Lu–Hf system remains closed during amphibolite-facies metamorphism at conditions of 6 kbar, 600 °C. Hydrothermal clinozoisite and epidote provided time constraints for fluid-rock interactions and hydrothermal mineralisation in a range of settings, demonstrating the utility of the technique for mineral exploration.