New metamorphic constraints on the Nova-Bollinger Ni–Cu deposit, Fraser Zone, Western Australia

Abstract

Metamorphic investigations were conducted using mineral chemistry, phase equilibria modelling and conventional thermobarometry on four metamorphic country rock samples from the Nova-Bollinger Ni–Cu deposit, Western Australia. New P–T constraints obtained from phase equilibria modelling of garnet-bearing granulites and cordierite–sillimanite-bearing metasedimentary rocks indicate that the Nova-Bollinger deposit formed at mid-crustal depths (0.75–0.76 GPa and 880–920 °C), along high thermal gradients around 1200 °C/GPa. Peak metamorphic conditions are consistent with those obtained from the Nova-Bollinger intrusive rocks, and with metagabbros from the southern Fraser Zone, but are slightly elevated relative to metapelites examined in the southern Fraser Zone. The peak metamorphic conditions at Nova-Bollinger reflect the combined effects of regional-scale high-T conditions during orogenesis superimposed by contact metamorphism in a thermal aureole adjacent to the Nova-Bollinger mafic–ultramafic magmas. Thermal metamorphism may have been further enhanced by local heat transfer from intruding sulfide liquid, which pervasively infiltrated country rocks on scales of tens to hundreds of metres beneath the orebodies. This additional heat source may be locally significant, despite being volumetrically limited. Retrograde P–T conditions were determined by garnet–biotite conventional thermobarometry of garnet-core–matrix biotite pairs (0.29–0.45 GPa and 550–640 °C). These P–T conditions suggest that the Nova-Bollinger country rocks cooled along a near-isobaric cooling path and reflect the effects of thermal decay following emplacement of voluminous magmas into the Fraser Zone during Stage I Albany–Fraser Orogeny.

KEY POINTS

1. We present the first quantitative P–T constraints for host rocks of the intrusion-hosted Nova-Bollinger Ni–Cu deposit in the Fraser Zone, Western Australia.

2. Peak metamorphism at Nova-Bollinger occurred at 0.75–0.76 GPa and 880–920 °C (1200 °C/GPa).

3. GBAQ-GB thermobarometry records re-equilibration conditions at 0.29–0.45 GPa and 550–640 °C, indicating near-isobaric retrograde cooling.

4. Infiltration of sulfide liquids into host country rocks may provide a volumetrically low but significant contribution of heat to high-T metamorphism during formation of orthomagmatic deposits in mid-crustal terranes.

Keywords:

• Albany–Fraser Orogen
• high thermal gradient metamorphism
• Nova-Bollinger
• P–T modelling
• garnet–biotite thermometry
• orthomagmatic
• granulite-facies metamorphism
• thermal aureole
• Fraser Zone

Introduction

World-class orthomagmatic deposits are major contributors to the global supply of base-metal resources, including nickel, copper and the platinum group elements, which are crucial for a carbon-free energy future. These deposits have typically been associated with magmatic systems emplaced within the mid–upper crust, as part of large igneous provinces (e.g. Barnes et al., 2016; Begg et al., 2010; Naldrett, 2004; Pirajno et al., 2009). In contrast, deposits related to orogenic settings have traditionally been considered as less economically viable for major magmatic Ni–Cu reserves (Naldrett, 2004, 2010). However, recent discoveries have uncovered an increasing number of orthomagmatic Ni–Cu deposits within orogenic belts, commonly within high-grade metamorphic rocks. Examples include the Selebi-Phikwe deposits in the Limpopo Belt, Botswana (Maier et al., 2008), Xiarihamu deposit in the Tibetan Plateau, West China (Li et al., 2015), various occurrences in the Ivrea Zone in Italy (Chong et al., 2021; Fiorentini et al., 2018; Locmelis et al., 2016), Savannah (formerly Sally Malay) in the Kimberley region of Western Australia (Le Vaillant et al., 2020), the Limoeiro deposit in Brazil (Mota-e-Silva et al., 2013) and the Nova-Bollinger Ni–Cu deposit in Western Australia (Barnes, Taranovic, Miller, et al., 2020; Bennett et al., 2014; Maier et al., 2016; Taranovic et al., 2021, 2022).

Many of the magmatic intrusions associated with Ni–Cu bearing orthomagmatic occurrences are hosted within amphibolite–granulite-facies rocks in convergent settings (Table 1), indicating that the mid–lower crust may be more prospective for Ni–Cu mineralisation than previously thought (Bagas et al., 2016; Fiorentini et al., 2018; Latypov et al., 2024; Locmelis et al., 2016). In many cases, however, the metamorphic conditions under which the intrusions were emplaced and crystallised, and therefore the petrogenetic models for the formation of the associated deposits, have either been poorly constrained or so far not been investigated. As a result, models of exploration developed for much shallower intrusions may be less effective in high-grade terranes, as the drivers for sulfide saturation (Mavrogenes & O’Neill, 1999) and magmatic remobilisation (Stone et al., 2004) are reliant on robust P–T constraints.

Table 1. Summary table of P–T conditions and relevant ages of peak metamorphism and emplacement of representative Ni–S occurrences in orogenic settings.

Occurrence	Unit	Temperatures	Pressures	Peak metamorphic ages	Emplacement ages	Method
1. Savannah deposit, Halls Creek Orogen, Western Australia	Tickalara Metamorphics (a)	700–800°C^a	0.3–0.5 GPa^a	1867–1845 Ma^a	1845–1840 Ma (j)	Phase equilibria modelling
2. Nova-Bollinger deposit, Fraser Zone, Western Australia	Snowys Dam Formation (b)	830–860°C^a	0.7–0.9 GPa^a	1334–1293 Ma^a	1300 Ma (k)	Phase equilibria modelling
	Fraser Gabbro (c)	900–950 °C^b	≤0.7 GPa^b	1315–1296 Ma^b
3. Valmaggia, Ivrea-Verbano Zone, Italy	Kinzigite Formation (d)	650–950°C^a	0.35–1.2 GPa^a	280 Ma^a	249.1 Ma (l)	Conventional thermobarometry
	Mafic Complex (e)	800–1000 °C^b	0.3–1.0 GPa^b	290–250 Ma^b
4. Voisey’s Bay, Canada	Tasiuyak Paragneiss (f)	700–750°C^a	0.3–0.6 GPa^a	1860 Ma^a	1350–1322 Ma (m)	Forward equilibria modelling
5. Limoeiro deposit, Rio Capibaribe Terrane, Brazil	Vertentes Complex (g)	700–850°C^a	0.3–0.5 GPa^a	1718 Ma^b	650–550 Ma? (n)	Forward equilibria modelling
6. Xiarihamu Ni–Cu deposit, Eastern Kunlun Orogen, Qinghai, China	East Kunlun eclogites (h)	590–650°C^b	>1.6 GPa^b	428–436 Ma^b	422–423 Ma (o)	Conventional thermobarometry
7. Selebi-Phikwe, Limpopo Belt, Botswana	Central Zone, Limpopo Belt (i)	750–890°C^a	0.76–1.33 GPa^a	2600–2650 Ma^a	Unconstrained	Conventional thermobarometry

aMetamorphic conditions and ages obtained through metapelite analyses.

bMetamorphic conditions and ages obtained through meta-igneous analyses.

Sources for peak metamorphic conditions and ages: (a) Bodorkos et al. (1999), (b) Clark et al. (2014), (c) Glasson et al. (2019), Kirkland et al. (2016), (d) Redler et al. (2012), (e) Kunz and White (2019), Peressini et al. (2007), (f) Soto et al. (1999), Mitchell et al. (2014), (g) Mota-e-Silva et al. (2013, 2015), França et al. (2019), (h) Meng et al. (2013), (i) Maier et al. (2008). Sources for emplacement ages: (j) Le Vaillant et al. (2020), (k) Taranovic et al. (2022), (l) Locmelis et al. (2016), (m) Tettelaar and Indares (2007), (n) not constrained yet; estimated older than 650–550 Ma (Mota-e-Silva et al., 2013), (o) Li et al. (2020).

To further develop the relationship between magmatic sulfide targets and their associated high-grade terranes, we focus on the Nova-Bollinger Ni–Cu deposit, which is located within the south-central part of the Fraser Zone in the Albany–Fraser Orogen, Western Australia (AFO; Figure 1). This deposit contains extensive Ni–Cu sulfides that are hosted primarily within two petrogenetically related orebodies—Nova and Bollinger. These orebodies yield an estimated combined resource of 13.1 Mt Ni at 2.0 wt%, 0.8 wt% Cu and 0.1 wt% Co (Independence Group, 2019). The investigated rocks include garnet-bearing metagabbro and cordierite–sillimanite-bearing metapelites obtained from drill-core samples obtained close to Nova-Bollinger deposit.

Figure 1. Regional geological map of the eastern Albany–Fraser Zone displaying locations of Nova-Bollinger and other Ni–Cu deposits/prospects. Previous metamorphic investigations in the southern Fraser Zone at Wyralinu Hill, Gnamma Hill and Mt Malcolm in Clark et al. (2014) and Glasson et al. (2019) are shown. Map adapted and modified after Taranovic et al. (2021).

This study integrates phase equilibria modelling, thermobarometry and mineral chemistry to: (1) determine the P–T conditions experienced by the metamorphic host rocks at the Nova-Bollinger deposit, (2) constrain their retrograde cooling trajectories, and (3) refine insights into the relationship between high-T metamorphism and orthomagmatic mineralising processes. The results of this study also help to refine tectonic models proposed for the formation of the Nova-Bollinger deposit adding four new P–T constraints to the Albany–Fraser Orogen, expanding the envelope of quantitative P–T data further northwards into the Fraser Zone.

Geological background

Albany–Fraser Orogen

The Albany–Fraser Orogen (AFO) is a ∼1200 km-long, northeast–southwest to east–west-trending Paleo–Mesoproterozoic orogen that formed as a result of variable reworking of the south-southeastern Yilgarn Craton between 1815 and 1140 Ma (Kirkland et al., 2015; Nelson et al., 1995; Smithies et al., 2015; Spaggiari et al., 2011, 2015; Tucker et al., 2018). It is bounded by the Madura Province to the east (Kirkland et al., 2017; Spaggiari et al., 2018) and the Pinjarra Orogen to the west (Markwitz et al., 2017; Myers, 1990), and it wraps around the south-southeast margin of the Archean Yilgarn Craton (Figure 1).

The main tectonic units that make up the AFO are the Northern Foreland and the Kepa Kurl Booya Province (Figure 1). The Northern Foreland is the reworked southern and southeastern margin of the Yilgarn Craton; it contains ca 2720–2620 Ma orthogneisses that were deformed during the Mesoproterozoic (Cassidy et al., 2006; Kirkland et al., 2014; Spaggiari et al., 2009, 2015). The Kepa Kurl Booya Province occurs adjacent to the Northern Foreland and represents the crystalline basement of the AFO (Figure 1). It is subdivided into the Tropicana, Biranup, Nornalup and Fraser zones (Figure 1). The Tropicana Zone is located along the eastern margin of the Yilgarn Craton (Figure 1b) and is composed mainly of 2720 Ma Archean orthogneisses that were subsequently overprinted by ca 2718–2554 Ma granulite-facies metamorphism (Kirkland et al., 2014; Occhipinti et al., 2014). Magmatic and sedimentary rocks associated with Paleoproterozoic (ca 1810–1625 Ma) extension and reworking of the Archean Yilgarn margin were emplaced and deposited within the Biranup Zone (Kirkland et al., 2011; Spaggiari et al., 2011, 2015). Extension occurred in three pulses and was coeval with voluminous felsic magmatism and the formation of a widespread volcano-sedimentary basin known as the Barren Basin (Kirkland et al., 2011; 2015; Spaggiari et al., 2013, 2015, 2018, 2020). The Nornalup Zone occurs to the south-southeast of the Biranup Zone (Figure 1b) and comprises Paleoproterozoic granite basement rocks that are partly overlain by ca 1560–1310 Ma Arid Basin units and extensively intruded by ca 1330–1140 Ma felsic rocks (Kirkland et al., 2011; Myers et al., 1996; Spaggiari et al., 2011, 2015). Sedimentary protoliths to the Arid Basin were primarily deposited in a passive margin setting, east of the Nornalup Zone and include the Gwynne Creek Gneiss and Snowys Dam Formation (Figure 1c; Spaggiari et al., 2015). Accretion and obduction of a Mesoproterozoic oceanic arc onto the eastern AFO between 1389 and 1330 Ma, causing the closure of this passive margin and transition to a convergent margin setting (Spaggiari et al., 2015, 2018).

The Biranup and the Nornalup zones were heavily intruded by Mesoproterozoic magmatic rocks of the Recherche Supersuite (ca 1330–1280 Ma) and Esperance Supersuite (ca 1200–1140 Ma) during the Albany–Fraser Orogeny (Kirkland et al., 2011, 2014; Smithies et al., 2015; Spaggiari et al., 2011). The Albany–Fraser Orogeny occurred in two stages between ca 1330 and 1260 Ma (Stage I), and between ca 1225 and 1140 Ma (Stage II). Stage I was triggered by oceanic arc accretion and involved coeval emplacement of the calc-alkaline felsic Recherche Supersuite and mafic magmas of the Fraser Gabbro, and moderate-pressure high-thermal gradient metamorphism (Bodorkos & Clark, 2004; Clark et al., 2000, 2014, Glasson et al., 2019; Smithies et al., 2015). Stage I magmatism and metamorphism are restricted to the easternmost part of the AFO, in the Fraser Zone, the eastern Nornalup Zone and once-contiguous terranes in East Antarctica (Windmill Islands; Morrissey, Hand, et al., 2017; Tucker et al., 2018). Stage II of the Albany–Fraser Orogeny involved long-lived high to ultrahigh thermal gradient metamorphism and voluminous felsic magmatism across the entire AFO (Clark et al., 2000; Kirkland et al., 2014, 2016; Payne et al., 2021; Smithies et al., 2015; Spaggiari, Kirkland, Smithies, Occhipinti et al., 2014; Spaggiari et al., 2015). Stage II metamorphism and magmatism are interpreted to reflect either intracratonic orogenesis or the delayed response of final collision between the West Australian Craton and the South Australian Craton (Clark et al., 2000; Kirkland et al., 2017; Smithies et al., 2013; Spaggiari et al., 2011; Spaggiari, Kirkland, Smithies, Occhipinti et al., 2014; Waddell et al., 2015).

Fraser Zone

The Fraser Zone is located in the eastern AFO (Figure 1b). It is a 425 km-long and 50 km-wide tectonic zone that trends northeast and has a distinct gravity and magnetic signature marked by a Mesoproterozoic supracrustal sequence and the presence of voluminous dense mafic–ultramafic rocks (Maier et al., 2016). Most of the Fraser Zone is overlain by Paleozoic sedimentary sequences, which limit clear exposure of the Proterozoic basement rocks (Kirkland et al., 2017; Spaggiari et al., 2020). Magmas in the Fraser Zone represent juvenile melts that have incorporated variable amounts of crust, including reworked parts of the Yilgarn Craton (Kirkland et al., 2016; Maier et al., 2016; Smithies et al., 2013, 2015; Taranovic et al., 2022). These magmatic rocks in the Fraser Zone were emplaced between ca 1330 and 1260 Ma and comprise large regional sheeted sill complexes of gabbro and granite of the Fraser Gabbro and Recherche Supersuite hosted within quartz–feldspar-rich sequences of the Snowys Dam Formation (Smithies et al., 2013, 2015; Spaggiari et al., 2014, 2015). Deformation and granulite-facies metamorphism occurred between 1304 and 1267 Ma (Clark et al., 2014; Glasson et al., 2019). This close temporal relationship supports the notion that magmatism was the main thermal driver of high-grade metamorphism in the Fraser Zone (Clark et al., 2014).

To date, metamorphic P–T constraints in the Fraser Zone have been confined to its southernmost areas owing to the overlying Paleozoic sedimentary sequences dominating much of the Fraser Zone. Garnet–sillimanite-bearing metapelites in the southern Fraser Zone were interpreted by Clark et al. (2014) to have been metamorphosed at 0.7–0.9 GPa and 850 °C between 1292 and 1267 Ma. Garnet-absent metagabbros from the southern Fraser Zone were metamorphosed at 1293 Ma at similar peak conditions (900–950 °C and 0.7 GPa; Glasson et al., 2019).

Limited evidence for Stage II tectonism is recorded in the Fraser Zone (1225–1140 Ma; Clark et al., 2014; Kirkland et al., 2011, 2014; Spaggiari et al., 2009; Spaggiari, Kirkland, Smithies, Occhipinti et al., 2014). Recent U–Pb titanite geochronology and ⁴⁰Ar/³⁹Ar thermochronology have shown that the Fraser Zone experienced moderately high temperatures at mid-crustal levels during Stage II of the Albany–Fraser Orogeny. However, these conditions were not conducive to new metamorphic zircon growth (Kirkland et al., 2014, 2016; Scibiorski et al., 2016).

Several models have been proposed for coeval magmatism and metamorphism in the Fraser Zone during Stage I of the Albany–Fraser Orogeny. One model suggests that tectonism was triggered by the accretion of an oceanic arc onto the eastern margin of the AFO and the subsequent development of a back-arc basin (Clark et al., 2014; Glasson et al., 2019; Kirkland et al., 2011; Kuper et al., 2024; Morrissey, Hand, et al., 2017; Morrissey, Payne, et al., 2017; Spaggiari et al., 2018, 2020). An alternative model proposes that oceanic arc accretion onto the eastern margin of the AFO caused crustal thickening followed by orogenic collapse, extension and slab delamination, to asthenospheric upwelling and the voluminous intrusion of calc-alkaline and mafic magmas (Maier et al., 2016; Smithies et al., 2013, 2015; Spaggiari et al., 2015).

Nova-Bollinger Ni–Cu deposit

The discovery of Nova-Bollinger in an otherwise greenfields terrane suggested that the Fraser Zone in the AFO needed to be reconsidered as a viable terrane for magmatic Ni–Cu exploration (Bennett et al., 2014; Maier et al., 2016). Since this initial discovery, many other Ni–Cu prospects have been identified in the Fraser Zone including Silver Knight, Mawson and Octagonal (Figure 1b). The Nova-Bollinger Ni–Cu intrusive complex is located in the south-central Fraser Zone, approximately 70 km east-northeast of Mt Malcolm (Figure 1c) and is composed of multiple intrusive bodies spread across a 4 km² area (Taranovic et al., 2021). These mafic–ultramafic bodies, known informally as the Fraser Gabbro, are thought to have been emplaced between 1305 and 1290 Ma (Maier et al., 2016; Smithies et al., 2013; Spaggiari et al., 2015). The Nova-Bollinger intrusive rocks have specifically been dated to 1304 ± 22 Ma (Morrison et al., 2022). Of these intrusions, two main mafic–ultramafic bodies associated with sulfide mineralisation have been studied: the bowl-shaped upper intrusion, which is an interlayered mafic–ultramafic cumulate sequence; and the lower intrusion, which comprises massive mafic–ultramafic orthocumulate rocks that host the bulk of the sulfide mineralisation (Figure 2; Barnes, Taranovic, Miller, et al., 2020; Maier et al., 2016; Smithies et al., 2013; Taranovic et al., 2021). Both intrusions are derived from the same parental magma source and believed to have been emplaced initially as a bifurcated sill (Taranovic et al., 2021). The lower limb developed as a chonolith into the lower intrusion, while the upper continued to inflate and expand with ongoing magma input and developed as a lopolith into the upper intrusion (Barnes, Taranovic, Miller, et al., 2020; Taranovic et al., 2021).

Figure 2. Schematic cross-section of the Nova-Bollinger intrusive complex, showing the upper and lower intrusion, and the Nova, Bollinger and C5 orebodies. Schematic drill-core sample positions for SFRD0130 and NBU1372 are illustrated in yellow stars. Adapted from Taranovic et al. (2022).

The upper intrusion is at least 450 m thick and is subdivided into an upper layered series and a homogeneous lower basal series. The layered series comprises feldspathic lherzolite with variable thickness (20–100 m) and displays sharp interlayering with adjacent norite and gabbronorite (Taranovic et al., 2021). The lower basal series grades upwards from orthocumulate gabbros to mesocumulate peridotite and was simultaneous with the development of the connected lower intrusion (Barnes, Taranovic, Miller, et al., 2020; Taranovic et al., 2021). The lower intrusion is a chonolith that is at least 1.5 km long, up to 100 m thick and 300–500 m wide. It comprises orthocumulate feldspathic lherzolite to metagabbro. Recent apatite age data on Nova-Bollinger lower intrusion cumulate rocks indicate emplacement ages at ca 1304 ± 22 Ma (Morrison et al., 2022). Orthopyroxene–spinel–hornblende symplectites are found along olivine–plagioclase crystal boundaries and are interpreted as reaction coronas. These are observed in both the upper and lower intrusions (Torres-Rodriguez et al., 2021). The lower intrusion contains abundant disseminated and net-textured sulfide mineralisation with variable thicknesses (Figure 2).

The bulk of sulfide mineralisation forms two main orebodies, Nova and Bollinger, which are found within and directly beneath the lower intrusion. Both orebodies have been petrogenetically and isotopically linked to the Nova-Bollinger intrusion (Barnes, Taranovic, Miller, et al., 2020; Taranovic et al., 2021, 2022). The massive, semi-massive and brecciated components of these orebodies occur within tightly folded sequences of garnet-rich metasedimentary rocks, metagabbroic gneisses and minor marble layers that belong to the Snowys Dam Formation (Maier et al., 2016; Taranovic et al., 2021). The emplacement of these sulfides was synchronous with partial melting of immediately adjacent country rocks, producing leucosomes that were intimately associated with sulfide veins to depths of up to 100 m beneath the host intrusions (Barnes, Taranovic, Miller, et al., 2020). Similar occurrences of vein-type massive sulfide mineralisation invading into host-country rocks are noted at Kalatongke in the Central Asian Orogenic Belt (Mao et al., 2022) and in many other intrusion-hosted deposits such as Savannah and Voisey’s Bay (Barnes et al., 2018).

Petrographic studies on the Nova-Bollinger intrusive rocks and the sulfide ore textures indicate that the parent magmas and associated sulfide liquids were emplaced at the same time as peak metamorphism in the Fraser Zone (Barnes, Taranovic, Miller, et al., 2020; Barnes, Taranovic, Schoneveld, et al., 2020; Taranovic et al., 2021). Phase equilibria modelling and two-pyroxene thermometry indicate that complex spinel–pyroxene–amphibole reaction coronas and symplectites in the Nova-Bollinger olivine cumulates formed at moderate pressures (0.76–0.96 GPa) as the intrusive rocks cooled slowly from 1005–1055 °C to 850–900 °C (Torres-Rodriguez et al., 2021). The latter temperatures were inferred to represent the ambient peak metamorphic conditions of the Fraser Zone, owing to similarities with P–T constraints obtained from rocks located approximately 80 km south of the Nova-Bollinger intrusions, in the southern Fraser Zone (Clark et al., 2014; Glasson et al., 2019). However, metamorphic rocks at the Nova-Bollinger deposit are mineralogically different from metamorphic rocks from the southern Fraser Zone. To provide deposit-specific P–T conditions for the Nova-Bollinger Ni–Cu deposit, this study constrains the peak metamorphic conditions using metapelitic and metagabbroic rocks obtained proximal to the orebodies.

Analytical methods

Sample processing and petrographic analysis

Four samples were obtained from Nova-Bollinger for phase equilibria modelling. These samples are representative of the metapelite and metagabbro rocks that host sulfide mineralisation and come from two drill cores: NBU1372 (Bollinger) and SFDR0130 (Nova). The metagabbro specimens come from rocks that are structurally beneath the sulfide mineralisation from both drill cores. The metapelite samples come from rocks between semi-massive sulfide mineralisation and the base of the lower intrusion. All four samples are interpreted as restitic granulites. They contain a small proportion of discontinuous quartz–feldspar domains visible in hand-sample and/or thin-section, and we interpret these features to reflect evidence for partial melting and the former presence of leucosomes in these rocks (Figure 3a–d). The sample locations, petrographic descriptions and calculated peak metamorphic conditions are summarised in Table 2.

Figure 3. Close-up of drill-core samples used in this study. (a) JC21NB7; (b) JC21NB10; (c) JC21NB33; and (d) JC21NB37. Melt inclusions are highlighted in dashed lines.

Table 2. Summary of metamorphic data.

Sample	Drill core	Orebody	Rock type	Interpreted peak assemblage	Peak P–T conditions (Geo-PS)	Peak P–T conditions (Thermocalc)
JC21NB7	NBU1372 (150.8 m)	Bollinger	Cordierite–garnet–sillimanite gneiss	bi–g–sill–crd–pl–ksp–ilm–q–ru–liq	0.70–0.72 GPa; 870–880°C	0.70–0.72 GPa; 860–870°C
JC21NB10	NBU1372 (182.8 m)	Bollinger	Garnet–two-pyroxene–amphibole granulite	g–cpx–opx–pl–amp–ilm ± L	0.60–0.94 GPa; 850–990°C	0.58–0.94 GPa; 840–990°C
JC21NB33	SFRD1030 (209.7 m)	Nova	Cordierite–garnet–sillimanite gneiss	bi–g–sill–crd–ksp–pl–ilm–q–ru–liq	0.71–0.73 GPa; 860–870°C	0.71–0.73 GPa; 860–870°C
JC21NB37	SFRD1030 (238.8 m)	Nova	Garnet two-pyroxene granulite	g–cpx–opx–ilm–pl–L	0.50–1.12 GPa; 830–1120⁰C	0.58–1.10 GPa; 830–1100°C

g, garnet; sill, sillimanite; cd, cordierite; pl, plagioclase feldspar; ksp, potassium feldspar; amp, amphibole; cpx, clinopyroxene; opx, orthopyroxene; ilm, ilmenite; ru, rutile; q, quartz; liq/L, liquid (liquid and L are differentiated in their compositions, metapelite and metagabbro, respectively).

Mineral abbreviations adapted from Holland and Powell (1998).

Whole-rock geochemistry

Whole-rock geochemistry was carried out on samples JC21NB7, JC21NB10, JC21NB33 and JC21NB37 to provide bulk compositions for phase equilibria modelling. These four samples were petrographically analysed and found to be barren host rocks, distal from any significant sulfide mineralisation. Whole-rock analyses were carried out using X-ray fluorescence (XRF), laser-ablation-inductively coupled-mass spectrometry (LA-ICP-MS) and volumetric FeO titration at Bureau Veritas in Perth, Australia. Samples were dried, crushed and subsequently pulverised in a vibrating disc pulveriser. Major and trace elements were analysed using a 66:34 flux with 4% LiNO₃ added to form a glass bead. The accuracy and precision for major-element concentrations determined by XRF were <1%. Loss of ignition (LOI) was determined by a robotic TGA system with furnace temperatures between 110 °C and 1000 °C. The whole-rock geochemical data are summarised in Table 3. Detection limits for major elements and trace elements were between 0.001 and 0.01 wt% and between 0.01 and 1 ppm, respectively. Precision and accuracy were monitored by in-house reference materials provided by Bureau Veritas. Precision for major elements is better than 1% of the reported values, and precision for trace elements is better than 10% of the reported values.

Table 3. Whole-rock geochemistry used in this study.

	Metapelite (Bollinger)	Metapelite (Nova)	Metagabbro (Bollinger)	Metagabbro (Nova)
Major elements (wt%)	JC21NB7	JC21NB33	JC21NB10	JC21NB37
SiO2	61.29	70.94	48.34	46.86
Al2O3	17.30	11.38	19.94	20.11
Fe2O3^a	9.21	8.25	13.21	13.85
MnO	0.05	0.08	0.22	0.21
MgO	4.63	4.43	4.65	4.61
CaO	1.08	0.44	9.82	12.45
Na2O	0.28	0.37	2.98	1.50
K2O	1.89	1.34	0.28	0.10
P2O5	0.06	0.01	0.02	0.01
TiO2	0.89	0.54	0.52	0.51
LOI	2.51	1.56	0.01	0.30
Total	99.18	99.30	99.99	100.21

^aFe₂O₃ = FeO + Fe₂O₃.

Major-element mineral chemistry

Electron probe microanalysis (EPMA) was undertaken at the Centre for Microscopy, Characterisation and Analysis (CMCA), at the University of Western Australia, using a JEOL 8350 F electron microprobe. A beam current of 20 nA and accelerating voltage of 15 kV were used for all point analyses. Silicate and oxide analysis was performed over several runs targeting specific silicate minerals with accelerating voltages of 15–20 nA, 15–20 kV and unfocused beam sizes of 0–10 μm. The amphibole formulae were calculated from the EPMA results using the method of X. Li et al. (2020a), assuming a Li-free amphibole, and followed the classification of Hawthorne et al. (2012). Biotite analyses were calculated from EPMA results using the method of X. Li et al. (2020b). For a full description of methodology, analytical conditions and standards used, refer to Supplemental data Appendix B.

Phase equilibria modelling

Phase equilibria modelling was undertaken for samples JC21NB7, JC21NB10, JC21NB33 and JC21NB37. Two models were calculated for each sample using software Thermocalc v. 3.40 (Holland & Powell, 2011) and GeoPS v. 3.3 (Xiang & Connolly, 2022). Both forward modelling approaches use the same internally consistent and updated thermodynamic dataset ds62 of Holland and Powell (2011). All models were calculated for the geologically realistic system NCKFMASHTO (Na₂O–CaO–K₂O–FeO–MgO–Al₂O₃–SiO₂–H₂O–TiO₂–Fe₂O₃). The activity–composition relationships (a–x models) and melt models of Green et al. (2016) were used for the mafic granulite samples (JC21NB10 and JC21NB37), and the a–x models of White et al. (2014) were used for the metasedimentary samples (JC21NB7 and JC21NB33). Other phases considered in the modelling were biotite (White et al., 2007), spinel–magnetite (White et al., 2002), orthopyroxene, cordierite, biotite and garnet (White et al., 2014) and ilmenite–hematite (White et al., 2000).

Minimal mineral reaction microstructures were observed in these samples, such as the minor growth of retrograde biotite, partial replacement of cordierite by pinite and fibrolite (sample JC21NB7 and JC21NB33) and amphibole overgrowing clinopyroxene and plagioclase in sample JC21NB10 (refer to Results). These samples are all lithologically and texturally homogeneous at the hand-sample and thin-section scale. Accordingly, the bulk compositions used for phase equilibria modelling were based on the whole-rock geochemical compositions of these samples (Table 3), recalculated to molar oxide percent.

The H₂O contents used for phase equilibria modelling were based on the abundance of hydrous phases (i.e. cordierite, amphibole and biotite) in the observed mineral assemblage in each sample and an estimate of the H₂O content of each phase from their measured EPMA composition, or a conservative estimate of the typical H₂O content of these minerals in granulites (Rigby & Droop, 2011). The Fe₂O₃ (as ‘O’) vs FeO content used in each bulk composition was based on the abundance of Fe³⁺-bearing minerals in thin-section and their stoichiometrically re-casted EPMA compositions (Droop, 1987). By this approach, all four samples were modelled under anhydrous and reduced conditions. The sensitivity of the chosen H₂O and Fe contents on the P–T stability of the interpreted peak mineral assemblage of each sample was investigated with a series of T–M_H2O and T–M_O models (refer to Supplemental data Appendices B and C).

The T–M_H2O models show changes in phase equilibria topology and the position of the solidus over a range of H₂O contents from an ‘anhydrous’ composition (M = 0, 0.01 wt% H₂O) to a relatively hydrous composition (M = 1), which is proxied by the LOI obtained from whole-rock geochemistry, or a nominally high H₂O content. The calculated T–M o models [M o = Fe₂O₃/(FeO + Fe₂O₃)] show how the pseudosection topology changes over the compositional range M o = 0 (Fe = 99% FeO; i.e. a reduced composition) to M o = 0.75 (Fe = 75% FeO, i.e. a relatively oxidised composition).

Mineral compositional isopleths and modal abundances were calculated in GeoPS v. 3.3 (Xiang & Connolly, 2022). The modelled isopleths were compared with measured EPMA compositions of relevant peak phases and their estimated abundance in each sample. These data were used to further constrain the peak P–T conditions experienced by the host rocks at the Nova-Bollinger deposit.

Thermobarometry

Metamorphic conditions for metapelite samples JC21NB7 and JC21NB33 were estimated using the garnet–biotite–­alumino-silicate–quartz (GBAQ) geobarometer (Wu, 2017) and garnet–biotite (GB) geothermometer (Holdaway, 2000). The GBAQ method was coupled with the GB thermometer to obtain both P and T conditions iteratively. The GBAQ was chosen over the garnet–aluminosilicate–silica (quartz)–plagioclase (GASP; Holdaway, 2001) barometer owing to the absence of measurable plagioclase in metapelite sample JC21NB33. Representative compositions of garnet–biotite pairs were analysed by EPMA (see above). Biotite was analysed from several different microstructural locations: (1) biotite crystals that occur directly adjacent to and in contact with garnet porphyroblasts, (2) foliated and non-foliated biotite occurring in the matrix where it was not in contact with garnet, and (3) inclusions of biotite inside garnet porphyroblasts. The ferric iron content of all the garnet and biotite analyses used for thermobarometry was assumed to be 3 mol% and 11.6 mol%, respectively. These values were based on calculated values obtained from metapelites set out in Holdaway et al. (1997). Additional details are provided in Supplemental data Appendix B.

Results

Petrography

Metapelite sample JC21NB7 (cordierite–garnet–sillimanite gneiss)

Sample JC21NB7 is from drill-hole NBU1372 (150.8 m) and is located 8 m structurally above massive sulfide mineralisation at the Bollinger orebody (Figure 2). This sample contains garnet, cordierite, sillimanite, plagioclase, K-feldspar, biotite, quartz, ilmenite with accessory rutile, and zircon, monazite, apatite and pyrrhotite (<1 vol% each). The gneissic foliation is characterised by interlayered centimetre-wide quartz–feldspar-rich domains and garnet–cordierite–sillimanite-rich layers (Figures 3 and 4a).

Figure 4.  Representative photomicrographs of collected samples. (a–d) Metapelite sample JC21NB7: (a) gneissic domains of q + feldspar and sill₁ + crd + g in JC21NB7; (b) pinnite + sill₂ (fibrolite) + ilm bundles along rims and penetrations in biotite and cordierite; (c) foliated biotite and blocky sill₁ in cordierite-rich matrix; and (d) randomly oriented biotite + ilmenite in quartz + garnet-rich matrix. (e–h) Metapelite sample JC21NB33: (e) gneissic domains of qz + feldspar and sill₁ + cd + g-rich domains marked by dashed yellow line, bi + g are strongly foliated; (f) foliated bi + ilm around garnet, sill₂ + pinn are found along rims of biotite; (g) sill₁ rimmed by foliated biotite, rims of both sill₁ and bi are flanked by sill₂ + pinn bundles, some minor random orientation in bi is observed here; and (h) 20× magnification showing sill₂ + pinn bundles wrapped around and penetrating into crd. (j–l) Sample JC21NB10 showing granular matrix of pl + cpx + opx + parg with coarse-grained poikilitic garnet. (m) Sample JC21NB37 showing coarse–very coarse-grained cpx + opx + g in pl-rich matrix, minor pyrrhotite is present in all samples. Abbreviations (from Holland & Powell, 1998): ap, apatite; bi, biotite; cpx, clinopyroxene; crd, cordierite; g, garnet; ilm, ilmenite; ksp, K-feldspar; opx, orthopyroxene; parg, pargasite; pinn, pinnite; pl, plagioclase; po, pyrrhotite; q, quartz; sill₁, sillimanite; sill₂, fibrolite.

Coarse-grained garnet (1–2.5 mm) contains abundant inclusions of all other minerals in this sample. These inclusions are typically rounded and 0.1–0.4 mm, except for inclusions of biotite grains that retain internal foliation direction along their basal cleavages, which are elongate and aligned with the foliation external to the garnet. Coarse-grained (0.5–3.0 mm), elongate and blocky sillimanite crystals occur in the matrix (Figure 4b) and are commonly surrounded by and included in cordierite, biotite, K-feldspar or quartz. These sillimanite grains contain rounded inclusions of quartz, rutile and biotite (Figure 4a). Cordierite is moderately abundant (8 vol%) and occurs as moderately coarse-grained (0.2–1 mm) and subhedral to elongated crystals that are aligned with foliated biotite and sillimanite in the matrix. In many instances, the cordierite rims are extensively replaced by randomly oriented and fine-grained pinite and fibrous sillimanite (fibrolite; Figure 4b) and biotite. Biotite is moderately abundant (∼8 vol%) and commonly occurs as moderately coarse grains that are aligned with the foliation (Figure 4c). A minor amount of biotite occurs in sillimanite–pinnite aggregates around cordierite and as randomly oriented and anhedral grains in the matrix (Figure 4d). Ilmenite occurs as both fine, rounded grains and elongated crystals in biotite. Rutile is rare, occurring as sub-rounded inclusions (0.2 mm) in garnet and sillimanite, and in the foliated matrix. Pyrrhotite is the primary sulfide phase in this sample and occurs as rounded, fine-grained blebs alongside ilmenite and cordierite, mostly within the garnet–cordierite–sillimanite domains.

The interpreted peak mineral assemblage in sample JC21NB7 is interpreted to contain biotite, garnet, cordierite, sillimanite, plagioclase, K-feldspar, ilmenite, quartz, rutile and melt. Biotite and sillimanite have varied microstructural relationships with the other mineral phases in this sample. Inclusions of foliated biotite, cordierite and sillimanite within garnet, and the coarse-grained and foliated nature of biotite and sillimanite grains in the matrix support our interpretation that biotite, cordierite and sillimanite formed part of the prograde and peak mineral assemblage in this sample. However, the random orientation and anhedral nature of other biotite grains, and their location at the edges of matrix phases, and the replacement of cordierite by pinite–fibrolite-biotite aggregates, are also indicative of post-peak biotite and sillimanite growth.

Metapelite sample JC21NB33 (cordierite–garnet–sillimanite gneiss)

Sample JC21NB33 was obtained from drill-hole SFRD0130 (209 m depth) and is located 90 m structurally above massive sulfide mineralisation at the Nova orebody (Figure 3). This sample contains garnet, cordierite, sillimanite, K-feldspar, plagioclase, biotite, ilmenite and quartz. Accessory phases include rutile, zircon, monazite, apatite and pyrrhotite. This sample has a distinct gneissic foliation that is defined by alternating centimetre-scale quartz–feldspar-rich domains and garnet–cordierite–sillimanite-rich domains (Figure 4e).

Garnet occurs as abundant coarse to very-coarse-grained (up to 4 mm diameter) porphyroblasts within garnet–cordierite–sillimanite-rich domains, and it is typically surrounded by foliated biotite and contains rounded biotite inclusions (Figure 4e, f). Garnet also contains inclusions of cordierite, pyrrhotite, ilmenite, K-feldspar, sillimanite, quartz, biotite and rare rutile. Sillimanite occurs in the matrix as foliated and coarse (up to 0.4 mm long) crystals (Figure 4g). Cordierite is moderately abundant, occurs as rounded to slightly elongate grains in the matrix and occurs adjacent to garnet, foliated biotite and coarse sillimanite, ilmenite, quartz and feldspar. Similarly to sample JC21NB7 (see above), cordierite rims are extensively replaced by a second generation of fine-grained sillimanite that is intergrown with pinite and minor biotite and ilmenite (Figure 4h). Biotite occurs in several textural contexts: (a) as strongly foliated coarse grains (average 0.5 mm length; Figure 4e, f), (b) as fine–medium-grained crystals (0.1–0.3 mm) in random orientation across both quartz–feldspar-rich domains and garnet–cordierite–sillimanite-rich domains (Figure 4f), and (c) as very fine (<0.1 mm) assemblages of pinnite, with ilmenite and fibrolite along cordierite grain boundaries (Figure 4f–h). Biotite also occurs as inclusions in garnet, cordierite and quartz. The dominant feldspar in sample JC21NB33 is K-feldspar. It occurs in contact with all other mineral phases and within both the quartz–feldspar domains and garnet–sillimanite–cordierite-rich domains. Plagioclase is present but rare. It occurs in the garnet–sillimanite–cordierite domains as anhedral grains that are intergrown with quartz and as grains in the matrix that share irregular grain boundaries with biotite and sillimanite. In the quartz–feldspar domains, plagioclase occurs as minor elongate ribbons adjacent to coarse quartz. Ilmenite occurs as fine-grained, blocky, tabular and elongate crystals (up to 0.2 mm), commonly adjacent to biotite. Pyrrhotite occurs as very fine-grained (up to 0.2 mm), round, disseminated grains of pyrrhotite throughout the matrix and as inclusions in garnet and biotite. Rutile is rare, occurring as small anhedral crystals (up to 0.1 mm) in the matrix and as inclusions in garnet.

The interpreted peak mineral assemblage for sample JC21NB33 is biotite, garnet, cordierite, sillimanite, K-feldspar, plagioclase, ilmenite, rutile, quartz and melt. While most biotite grains in this sample are coarse and foliated, the randomly oriented, interstitial and anhedral nature of some biotite grains and the occurrence fine-grained biotite in pinite–sillimanite aggregates also suggest that biotite forms part of the post-peak mineral assemblage. Sillimanite occurs as both coarse-grained idioblastic crystals in the matrix and dense fine-grained fibrolite aggregates around cordierite rims (Figure 4g, h); this textural observation is also observed in sample JC21NB7 and likewise suggests that sillimanite forms part of both the peak and post-peak mineral assemblages, with the latter growing at the expense of cordierite.

Metagabbro sample JC21NB10 (garnet–orthopyroxene–clinopyroxene–hornblende granulite)

JC21NB10 comes from drill-hole NBU1372 (182.8 m depth) and is located approximately 20 m structurally beneath the Bollinger orebody (Figure 2). This sample is granoblastic to weakly foliated. The sample contains abundant porphyroblasts of coarse-grained, subhedral garnet (up to 2.5 mm in diameter) within a granular matrix of plagioclase, clinopyroxene, orthopyroxene, pargasite and ilmenite (Figure 4j). Clinopyroxene is present as pale green, anhedral, medium to coarse-grained crystals (0.4–0.8 mm) and fine-grained blebby inclusions in garnet. Orthopyroxene occurs in the granular matrix as fine- to medium-grained (0.1–0.6 mm) crystals that are commonly in contact with clinopyroxene, pargasite and plagioclase. Pargasite occurs in the matrix and as inclusions in the other mineral phases (Figure 4k). Plagioclase is abundant, largely in the matrix as moderately coarse-grained, subhedral to anhedral grains (0.5–1.5 mm; Figure 4j). Ilmenite is a minor phase in this sample and occurs as fine-grained, subhedral to elongate crystals within and adjacent to all other phases. Neither quartz nor K-feldspar was directly observed in this sample.

The interpreted peak mineral assemblage for JC21NB10 is garnet, pargasite, clinopyroxene, orthopyroxene, plagioclase, ilmenite and melt. Unlike the other samples in this study, there is limited evidence for the occurrence of a former melt phase in this sample; few small (∼5 mm diameter) plagioclase-rich domains with diffuse boundaries are observed to crosscut the weak fabric in this rock (Figure 3c). We interpret such domains to represent localised evidence for low-volume partial melting. Alternatively, the rock underwent partial melting followed by near-complete melt loss, leaving a refractory composition and only rare, trapped remnants of this former melt phase.

Metagabbro sample JC21NB37 (garnet–orthopyroxene–clinopyroxene granulite)

Sample JC21NB37 comes from drill-hole SFRD0130 (283.8 m depth) and is located approximately 12 m structurally above massive sulfide mineralisation in the Nova orebody (Figure 2). This sample contains garnet, plagioclase, clinopyroxene, orthopyroxene, ilmenite and accessory pyrrhotite. The rock is granoblastic and coarse-grained with a very weak foliation defined by elongate pyroxene and plagioclase. Garnet forms large poikiloblastic crystals (up to 3 mm in diameter) and contains subhedral to rounded inclusions of plagioclase, clinopyroxene and orthopyroxene (Figure 4m). Pale green clinopyroxene and light brown-pink orthopyroxene crystals are moderately coarse-grained (up to 1.5 mm) and subhedral, and commonly occur adjacent to one another in the matrix and as fine–medium-grained inclusions in garnet (Figure 4m). Plagioclase is abundant within matrix and as minor inclusions in all other phases. Ilmenite is ubiquitous, fine-grained (up to 0.2 mm in size) and tabular. Neither quartz nor K-feldspar was directly observed in this sample. The interpreted peak mineral assemblage for JC21NB37 is garnet, clinopyroxene, orthopyroxene, plagioclase, ilmenite and melt. Melt is inferred to form part of the peak mineral assemblage because plagioclase-rich segregations with diffuse boundaries are identified in hand-sample (Figure 3d).

Major-element mineral chemistry

The major-element mineral chemistry for samples JC21NB7, JC21NB10, JC21NB33 and JC21NB37 is summarised below. Representative EPMA data and calculated compositional ratios for each sample are provided in Supplemental data Appendix A.

Garnet

Representative garnet traverses display broadly flat element profiles from core to rim across all samples (Figure 5). Andradite (Fe³⁺) and uvarovite (Cr³⁺) components in the analysed garnets are nominally low (<0.01). Garnets from both metapelite samples are broadly enriched in both pyrope [X_Pyr = Fe²⁺/(Fe²⁺ + Ca²⁺ + Mg²⁺ + Mn²⁺)] at 0.31–0.51 and almandine [X_alm = Fe²⁺/(Fe²⁺ + Ca²⁺ + Mg²⁺ + Mn²⁺)] at 0.43–0.63. These garnets are also typically poor in grossular [X_grs = Ca²⁺/(Fe²⁺ + Ca²⁺ + Mg²⁺ + Mn²+)] and spessartine [X_sps = Mn/(Fe²⁺ + Ca²⁺ + Mg²⁺ + Mn²+)] with respective contents of 0.03–0.06 and 0.01–0.02. Garnets in both metagabbro samples are almandine-rich (0.55–0.63). Pyrope and grossular display broadly similar enrichments (0.13–0.21 and 0.20–0.23, respectively), while spessartine contents are nominally low (0.01–0.03; Figure 5). All garnets display characteristic retrograde diffusion patterns marked by almandine enrichment and corresponding depletions in pyrope and grossular (Tuccillo et al., 1990). Spessartine content remains nominally low across core–rim transects in all garnets.

Figure 5. Quaternary element profiles across representative garnet in metamorphic samples in: (a) metapelite sample JC21NB7; (b) metagabbro sample JC21NB10; (c) metapelite sample JC21NB33; and (d) metagabbro sample JC21NB37. Red squares, almandine; blue triangles, grossular; green circles, pyrope; yellow diamonds, spessartine.

Clinopyroxene

Clinopyroxene occurs in both metagabbro samples JC21NB10 and JC21NB37 as very fine to fine-grained anhedral inclusions in garnet (0.02–0.50 mm) and as subhedral grains within the matrix (0.10–1.20 mm; Figure 3f–h). These crystals from both metagabbro samples are compositionally unzoned. Included clinopyroxene and matrix clinopyroxene grains both display enriched Mg components [X_En = Mg/(Fe²⁺ + Mg + Ca)], with X_En values of 0.21–0.35 and 0.23–0.34, respectively. Wollastonite contents [X_Wo = Ca/(Fe²⁺ + Mg + Ca)] are 0.43–0.5 and 0.47–0.50, respectively, and all grains contain low jadeite content [j(cpx) = 0.01–0.05 p.f.u.]. Matrix clinopyroxene is relatively more enriched in Al₂O₃ and depleted in TiO₂ (Al₂O₃ = 1.73–4.58 wt%, TiO₂ = 0.26–0.52 wt%), relative to clinopyroxene inclusions in garnet (Al₂O₃ = 1.16–3.70 wt%, TiO₂ = 0.17–0.41 wt%).

Orthopyroxene

Orthopyroxene is observed in metagabbro samples JC21NB10 and JC21NB37, and has little to no core–rim compositional variation, with similar X_Fe and Al₂O₃ values across both samples (X_Fe = 0.54–0.59; Al₂O₃ = 1.50–1.83 wt%).

Amphibole

Amphibole is present in sample JC21NB10 as fine to moderately coarse rounded inclusions within garnet and as a fine–coarse subhedral matrix phase (Figure 4j). The majority of the amphibole can be classified as pargasite, irrespective of the microstructural location of the grain and their grain boundary relationships with the surrounding phases (Supplemental data Appendix A). The pargasite included in garnet has slightly lower X_Fe (0.14–0.17) and TiO₂ contents (1.4–2.1 wt%) than those within the matrix (X_Fe = 0.15–0.20; TiO₂ = 1.4–2.6 wt%). Both occurrences of amphibole have consistent alkali contents (Na + K = 0.69–0.76 and 0.67–0.81 p.f.u., respectively).

Biotite

Biotite is present in metapelite samples JC21NB7 and JC21NB33, and occurs in three distinct microstructural contexts: as (a) biotite inclusions in garnet; (b) coarse foliated crystals aligned with gneissic foliation; and (c) both non-foliated randomly oriented biotite occurring as interstitial minerals in the matrix assemblage.

In JC21NB7, biotite inclusions in garnet and biotite in the matrix that occurs in contact with garnet display relatively high X_Mg (0.75–0.82) contents compared with those not in contact with garnet (0.71–0.76), indicating extensive retrograde Fe–Mg exchange between garnet-associated biotite and their adjacent garnet hosts. There were no significant differences in X_Mg values between foliated (0.71–0.76) and non-foliated (0.72–0.77) matrix biotite in JC21NB7. The Ti content of biotite in JC21NB7 is relatively enriched, regardless of microstructural location; biotite inclusions and biotite in contact with garnet have Ti contents of 0.18–0.22 p.f.u. Foliated and non-foliated biotite in the matrix that is not in contact with garnet have Ti contents of 0.18–0.23 and 0.18–0.25 p.f.u., respectively. Biotite in JC21NB7 displays slightly elevated F contents of 0.27–0.78 wt% but negligible Cl (<0.01–0.03 wt%).

In sample JC21NB33, the highest X_Mg values (0.74–0.83) come from non-foliated matrix biotite, while X_Mg values of foliated matrix biotite and biotite inclusions are 0.71–0.80 and 0.74–0.82, respectively. The Ti contents of biotite across all microstructural contexts are similar, with biotite inclusions having the highest Ti contents (0.11–0.23 p.f.u.) and foliated and non-foliated biotite having similar ranges (0.11–0.18 and 0.11–0.18 p.f.u., respectively). Biotite in JC21NB33 only contains up to 0.29 wt% F and negligible Cl (<0.01–0.03 wt%).

Cordierite

Cordierite is present in sample JC21NB7 and JC21NB33, is compositionally homogeneous from core to rim and is relatively magnesian-rich (MgO = 10.7–11.8 wt%), with low X_Fe contents of 0.06–0.17.

Feldspar

Plagioclase occurs in JC21NB7 and is andesine in composition [X_An = Ca/(Ca + Na + K) = 0.44–0.49]. Plagioclase in the metagabbro samples varies in composition. Plagioclase in sample JC21NB10 is anorthite-rich (X_An = 0.48–0.58), relative to JC21NB37 (X_An = 0.69–0.78). K-feldspar is present in all metapelites (JC21NB7 and JC21NB33) and has X_Orth [X_Orth = K/(Ca + Na + K)] values of 0.75–0.89.

Ilmenite and magnetite

Ilmenite is ubiquitous and internally homogeneous in all samples. Ilmenite contains 49.1–51.2 wt% TiO₂, with calculated FeO and Fe₂O₃ contents of 1.63–6.26 wt% and 42.4–45.1 wt%, respectively. MnO contents range from 0.12 to 0.59 wt%.

Magnetite only occurs in sample JC21NB10 as rare, fine-grained exsolved blebs next to ilmenite, and it is not considered to form part of the peak mineral assemblage. The magnetite contains 0.67–1.4 wt% Al₂O₃, corresponding to X_Al values [X_Al = Al/(Al + Fe³⁺ + 2Ti)] of 0.02. TiO₂ and V₂O₅ contents are 0.6–3.7 wt% and 0.9–1.8 wt%, respectively.

Phase equilibria modelling

Phase equilibria modelling was done on two metapelite (JC21NB7 and JC21NB33) and two metagabbro (JC21NB10 and JC21NB37) samples to constrain the peak metamorphic conditions experienced by host rocks at the Nova-Bollinger deposit. The calculated P–T models for each sample are shown in Figure 6, and a Venn diagram summarising the P–T stability fields of all samples is provided in Figure 7. The modelled rocks are considered to have undergone partial melting and melt loss, and as such are inferred to preserve a residual, anhydrous composition. These modelled residual whole-rock compositions allow for inferences on the peak and retrograde P–T evolution of the rock, assuming that the last melt-loss event was at or prior to the rock attaining peak P–T conditions. The peak P–T conditions for each sample have been further constrained by contouring the metamorphic models for mineral composition and abundance, and by comparing these modelled compositions with EPMA data from representative minerals in each sample (see Supplemental data Appendix C).

Figure 6.  Pesudosections calculated through GeoPS on metapelite samples (a) JC21NB7, and (b) JC21NB33, and metagabbro samples (c) JC21NB10 and (d) JC21NB37. Peak metamorphic assemblages are marked in bold red lines; the solidus in each sample is marked by yellow dashed lines, white stars indicate further constraints through modal and compositional isopleths, white arrows indicate inferred cooling paths. amp, amphibole; bi, biotite; crd, cordierite; g, garnet; ilm, ilmenite; ksp, K-feldspar; opx, orthopyroxene; cpx, clinopyroxene, ky, kyanite; sill, sillimanite; pl, plagioclase; mt, magnetite; sp, spinel; liq, liquid; ru, rutile; q, quartz.

Figure 7. P–T Venn diagram of peak metamorphic assemblage domains calculated for Nova-Bollinger country rocks. Solidus lines of each sample are coloured and dashed according to their respective samples. The grey arrow indicates inferred the cooling path of metapelites. Peak metamorphic conditions calculated from previous studies in the southern Fraser Zone are coloured in grey. Error bars of 0.1 GPa and 50 °C prescribed in Palin et al. (2016) are shown on peak fields.

The recent release of new phase modelling software GeoPS (Xiang & Connolly, 2022), which employs Gibbs free energy minimisation and incorporates up-to-date thermodynamic datasets and the a–x models of Green et al. (2016) and White et al. (2014), provides an opportunity to compare P–T calculations with those obtained using the same dataset and a–x models in existing phase equilibria modelling software Thermocalc v.3.40 (Holland & Powell, 2011). The reader is referred to Supplemental data Appendix C for a detailed discussion between both programs.

Calculated Thermocalc and GeoPS P–T models are broadly similar for each sample (Table 2; compare Figure 6 and Figures S1–S4, Supplemental data Appendix B). Their pseudosection topologies have near-identical mineral parageneses, and the P–T realms of these mineral stability fields are similar. Only minor derivations were observed between the GeoPS and Thermocalc models for the stability of spinel and magnetite at low pressures (<0.5 GPa) in sample JC21NB10. The P–T constraints on the interpreted peak mineral assemblage in GeoPS and Thermocalc models from each sample are remarkably similar; the P–T range of the peak field differs by only ∼10 °C and ∼0.02 GPa between the GeoPS and Thermocalc models (Table 2; Supplemental data Appendix C). Given the consistency between the two programs, and for clarity, the P–T results are discussed in the context of the GeoPS models below. See Supplemental data Appendix C for further information on the operating parameters and model outputs from Thermocalc.

Sample JC21NB7 (cordierite–garnet–sillimanite gneiss)

The interpreted peak mineral assemblage for sample JC21NB7 is biotite, garnet, cordierite, sillimanite, plagioclase, K-feldspar, ilmenite, rutile, quartz and melt (Figure 6a). This assemblage is stable over a narrow P–T range, 0.70–0.72 GPa and 870–880 °C (Figure 6a, assemblage 18). Biotite occurs in several microstructural contexts in this sample (see above), and so it is somewhat ambiguous as to whether biotite forms part of the peak mineral assemblage. Biotite is interpreted to form part of the prograde mineral assemblage owing to biotite inclusions inside garnet porphyroblasts (Figure 4d). It is interpreted to form part of the post-peak mineral assemblage because it is also commonly randomly oriented and anhedral, occurs interstitial to or crosscuts other matrix phases (Figure 4d) and occurs in fine-grained sillimanite–pinite ± ilmenite intergrowths that replace the cordierite rims (Figure 4b). However, many of the matrix biotite grains in this sample are also coarse-grained and foliated (Figure 4c), and for this reason we choose to include biotite in the peak mineral assemblage. In the case that biotite does not form part of the peak assemblage, then the peak field in Figure 6a expands to slightly higher temperatures (870–940 °C at 0.68–0.72 GPa, assemblage 4 in Figure 6a). Regardless, peak temperatures occurred at pressures below 0.72 GPa and above 0.60 GPa, owing to the coexistence of cordierite and sillimanite in the peak mineral assemblage (Figure 6a). The solidus is modelled to occur at ∼830 °C.

Calculated modal abundance isopleths, compositional isopleths for X_alm and X_grs in garnet, and X_Fe in cordierite and biotite (Supplemental data Appendix C), broadly coincide with the modelled P–T realm of the peak mineral assemblage in this sample, the measured EPMA compositions of these minerals (Supplemental data Appendix A) and the observed mineral abundance in thin-section (Table 3). Modelled melt modal proportions (≤5–7 vol%, Supplemental data Appendix C) are also consistent with the observed abundance of quartz–feldspar domains in hand-sample and thin-section, which we interpret as former leucosomes.

Sample JC21NB33 (cordierite–garnet–sillimanite gneiss)

In sample JC21NB33, the peak metamorphic assemblage is interpreted as biotite, garnet, cordierite, ilmenite, rutile, K-feldspar, plagioclase, quartz, sillimanite and melt. This assemblage is stable over a very narrow P–T range, 0.71–0.72 GPa and 850–870 °C (Figure 6b, assemblage 18). Similarly to metapelite JC21NB7, biotite occurs in many microstructural contexts in this sample. We include biotite in the peak mineral assemblage of this sample because many grains in the matrix are coarse-grained and foliated (Figure 4e), but biotite is also abundant as anhedral and randomly oriented grains, and as dense fine-grained intergrowths with fibrolite and pinite at cordierite grain boundaries (Figure 4g, h). Should our interpretation be re-evaluated and biotite excluded from the peak mineral assemblage, then the peak field in Figure 6b likely expands to slightly higher temperatures, to the adjacent biotite-absent, plagioclase-absent field (870–930 °C, assemblage 14). Plagioclase is observed in sample JC21NB7 but is rare. However, the model in Figure 6b does not calculate a field in which biotite and plagioclase stably coexist with garnet, cordierite and sillimanite. The pressure range of the peak mineral stability field is constrained by the modelled absence of ilmenite above 0.73 GPa, and rutile below 0.71 GPa, and the observed coexistence of these two minerals in thin-section. A pressure constraint is also provided by the observed and modelled coexistence of cordierite and sillimanite over a similarly narrow pressure range (∼0.70–0.72 GPa; Figure 6b). The solidus is modelled to occur at ∼810 °C.

Similarly to sample JC21NB7, calculated compositional isopleths for X_Fe, X_alm and X_grs in garnet, and X_Fe in cordierite and biotite (Supplemental data Appendix C), and modal proportion isopleths for these minerals, are in agreement with the modelled P–T realm of the peak mineral assemblage in this sample, the measured EPMA compositions of these minerals (Supplemental data Appendix A), and the observed mineral abundance in thin-section (Table 3). Modelled melt modal proportions (≤5 vol%, Supplemental data Appendix C) are also consistent with the observed abundance of quartz–feldspar domains in hand-sample and thin-section.

Sample JC21NB10 (garnet–two-pyroxene–amphibole granulite)

In sample JC21NB10, the interpreted peak mineral assemblage is garnet, orthopyroxene, clinopyroxene, plagioclase, ilmenite and amphibole, and melt. This assemblage is stable over a broad P–T range, from 0.85–0.95 GPa at 850–900 °C to 0.6–0.8 GPa at 900–980 °C (Figure 6c). The lower temperature constraint on the peak field is provided by the position of the solidus (∼850–880 °C). The maximum temperature is constrained by the modelled absence of amphibole at temperatures >980 °C. The pressure is constrained by the modelled stability of spinel and magnetite <0.65–0.75 GPa, and the modelled absence of orthopyroxene >0.9 GPa.

In hand-sample, very small plagioclase-rich domains with diffuse boundaries are observed (Figure 3c), and we interpret these features to represent the former presence of trapped partial melt pockets. The modelled modal proportions of melt in this sample are very low, barely above 0 vol% in the lower-temperature realm of the peak field where it is adjacent to the solidus (Supplemental data Appendix C). This result is consistent with our observation that JC21NB10 only contains very localised and small felsic segregations. The modelled modal proportions of amphibole, garnet, orthopyroxene and clinopyroxene in the peak field are also broadly in agreement with the estimated abundances of these phases in the sample (Supplemental data Appendix C, Figure S5). These modal proportions overlap within the peak field at 0.78–0.82 GPa and 850–900 °C, providing a tighter constraint on the peak P–T conditions. The measured X_Fe content of clinopyroxene, as determined by EPMA, falls with the peak field, but the measured X_Fe, X_alm and X_grs garnet and X_Fe orthopyroxene contents occur to lower pressures and temperatures (Supplemental data Appendix C, Figure S5); while the garnets in this sample are texturally in equilibrium with the other mineral phases, it is possible that they preserve chemical evidence for post-peak major-element diffusion.

Sample JC21NB37 (garnet–two-pyroxene granulite)

The interpreted peak mineral assemblage of sample JC21NB37 is garnet, orthopyroxene, clinopyroxene, ilmenite, plagioclase and melt. This mineral assemblage is stable over a large P–T range that extends in a triangular shape from 820–930 °C at 0.5–0.7 GPa to 1.12 GPa at 1120 °C (Figure 6d). The peak field is bound to lower pressures by the modelled presence of amphibole and the position of the solidus (over ∼810–920 °C). The upper temperature and pressure constraints on the peak field are provided by the modelled absence of garnet and presence of quartz, respectively. This sample contains garnet, but it does not contain amphibole or quartz. Ilmenite is stable across most of the pseudosection except at the highest pressures and lowest temperatures, where rutile is stable.

Calculated compositional isopleths for X_Fe, X_alm and X_grs in garnet, and X_Fe in clinopyroxene and orthopyroxene (Supplemental data Appendix B), are in broad agreement with the modelled P–T realm of the peak mineral assemblage, and the measured EPMA compositions of these minerals (Supplemental data Appendix A). The observed abundance of garnet, orthopyroxene and clinopyroxene in thin-section (Table 3) also overlaps with the position of the peak field. Further, the modelled melt modal proportions in the peak field (3 vol%, Supplemental data Appendix C, Figure S7) are consistent with our observations of a low number of plagioclase-rich domains in hand-sample, which we interpret to represent a former melt phase. These modal and compositional isopleths share a region of common overlap in the peak field (0.70–0.74 GPa at 860–900 °C; Supplemental data Appendix C), providing a tighter constraint on peak P–T conditions for this sample.

Geothermobarometry

GBAQ-GB thermobarometry was applied to two metapelite samples (JC21NB7 and JC21NB37). The thermobarometry results yielded P–T constraints of 0.29–0.45 GPa and 550–640 °C (Figure 8). Biotite inclusions paired with adjacent garnet core compositions yielded similar P–T estimates in both JC21NB7 and JC21NB33 (0.31–0.38 GPa and 568–608 °C, and 0.33–0.40 GPa and 576–614 °C, respectively). Results from garnet rims and adjacent matrix biotite in JC21NB7 display the lowest P–T range (0.29–0.33 GPa and 560–580 °C), which may be indicative of extensive Fe–Mg exchange between these matrix phases. The highest P–T conditions were obtained from garnet rims paired with foliated biotite in the matrix from sample JC21NB33 (0.34–0.41 GPa and 566–615 °C; Figure 8).

Figure 8. P–T conditions obtained through GBAQ-GB thermobarometry on metapelite samples JC21NB7 (unfilled shapes) and JC21NB33 (filled shapes). The kyanite–sillimanite–andalusite isograds are marked in bold black lines.

Discussion

Here we assess the usefulness of applying conventional thermobarometry and phase equilibria modelling to high-grade metamorphic rocks at the Nova-Bollinger deposit. We provide new insights into: (1) the P–T conditions and evolution of the Nova-Bollinger deposit, (2) the cooling trajectory of the Nova-Bollinger host rocks, (3) the contribution of magmatic sulfide liquid to drive thermal metamorphism, and (4) the role of high-grade metamorphism in understanding orthomagmatic deposits.

New constraints on peak metamorphism from the Nova-Bollinger host rocks

Phase equilibria modelling of the four Nova-Bollinger metamorphic samples predicts peak stability fields that cover a wide range of pressures and temperatures, 0.62–0.88 GPa and 850–980 °C (Figure 7). The metagabbro samples have a large region of overlap between their predicted peak stability fields (0.6–0.85 GPa at 910–980 °C). Conversely, the peak stability fields calculated for the metapelite samples occur in close proximity (at ∼0.7 GPa and 860 °C), but they do not directly overlap with one another; these fields are very narrow, owing to the diversity of mineral phases in the peak assemblage, including biotite, cordierite, rutile and sillimanite (Figure 7). Overlapping modal and compositional isopleths of garnet and pyroxene in the metagabbros further constrain the peak conditions of these samples to ∼0.8 GPa at 880 °C, and 0.75–0.8 GPa at 880–920 °C, for JC21NB10 and JC21NB37, respectively (Figures 6 and 7). Modal and compositional isopleths for garnet and cordierite in metapelite samples JC21NB7 and JC21NB33 also overlap with the location of their calculated peak stability fields (Figures 6 and 7). Taking into account the random errors associated with phase equilibria modelling methods (±0.1 GPa and ± 50 °C; Palin et al., 2016), we interpret the peak metamorphic conditions recorded by the Nova-Bollinger country rocks to be approximately 0.75–0.76 GPa and 870–920 °C (Figures 7 and 9a).

The peak P–T conditions obtained in this study are calculated to be of a slightly higher temperature, but lower pressure, than those obtained from metapelites in the southern Fraser Zone (Clark et al., 2014). The metapelites in this study contain abundant coexisting coarse-grained sillimanite and cordierite, and accessory rutile (e.g. Figure 4c). These phases are strongly pressure-dependent and only coexist over a narrow pressure range (∼0.7 GPa; Figure 6a, b). Metapelites sampled and used for P–T calculations from other localities in the southern Fraser Zone are inferred to have formed at higher pressures than the samples in this study, owing to the absence of cordierite, which was predicted to be stable only below ∼0.7 GPa (Clark et al., 2014). Cordierite is typically associated with low–moderate-pressure, high-T metamorphism of noritic and psammitic rocks within the thermal aureoles of magmatic heat sources, such as those found in the Nova Venécia Complex, Brazil (Wisniowski et al., 2021). The modelled conditions are also consistent with conventional two-pyroxene thermometry and modelled pseudosection peak metamorphic estimates obtained from the Nova-Bollinger intrusion rocks between 0.76 and 0.96 GPa and between 850 and 1055 °C (Torres-Rodriguez et al., 2021). Biotite-bearing metapelite country rocks that occur close to a thermal anomaly (within ∼1 km distance), at mid-crustal depths, are likely to undergo dehydration and recrystallisation, resulting in the formation of fine-grained sillimanite and cordierite-rich assemblages (e.g. Wisniowski et al., 2021). The metamorphic rocks analysed in this study were obtained within 100 m depths beneath the lower intrusion. The coexistence of sillimanite, cordierite and biotite in the Nova-Bollinger metapelites supports earlier interpretations that these rocks were metamorphosed in close proximity to the contact aureole of syn-metamorphic Nova-Bollinger mafic–ultramafic magmas (estimated to 1 km by Barnes, Taranovic, Miller, et al., 2020). Conversely, the absence of cordierite in rocks from the southern Fraser Zone, and their slightly cooler peak temperatures, as recorded by Clark et al. (2014), suggests that these rocks may have equilibrated distal to any syn-metamorphic magmatism. This was a conclusion also reached by Glasson et al. (2019).

Both of the Nova-Bollinger metagabbros in this study preserve peak metamorphic assemblages that contain garnet, orthopyroxene, clinopyroxene and plagioclase. Pargasitic amphibole is also present in JC21NB10. In contrast, metagabbros from the southern Fraser Zone are rich in amphibole and biotite, and do not contain garnet (Glasson et al., 2019). The P–T constraints of Glasson et al. (2019) indicate that the hydrous, garnet-absent metagabbros in the southern Fraser Zone metamorphosed at pressures <0.7 GPa, and temperatures of ∼950–1010 °C. While the P–T constraints of Glasson et al. (2019) overlap slightly with the calculated peak stability fields of the metagabbros investigated in this study, the modal and compositional isopleths from these metagabbros suggest that they were metamorphosed at slightly higher pressures and lower temperatures (Figure 7). However, the magmatic precursors to both Nova-Bollinger and south Fraser metagabbros are likely to have experienced pressures greater than 0.7 GPa during their emplacement from depth (Glasson et al., 2019). The relatively higher pressures calculated here indicate that the Nova-Bollinger rocks and its intrusions were emplaced within garnet stability fields and at slightly greater depths than those in the south Fraser Zone. Additionally, it has been suggested that fluids from pelites and psammites adjacent to the metagabbros were released as a result of partial melt crystallisation during high-T metamorphism (Glasson et al., 2019). These available fluids may have driven extensive crystallisation of hydrous phases in the adjacent metagabbros (Glasson et al., 2019). However, as shown by sample JC21NB37, the presence of amphibole in the Fraser Zone metagabbros is not ubiquitous. The crystallisation of amphibole in the Nova-Bollinger metagabbros may therefore not be purely metasomatic. The bulk geochemical compositions of the Nova-Bollinger metagabbros show that they have a strongly aluminous composition (Table 3), nearly twice as high as the aluminium content of igneous and meta-igneous rocks from the southern Fraser Zone (Glasson et al., 2019). Within the Fraser Zone, geochemical and isotopic evidence of gabbroic rocks readily records the hybridisation and contamination of mantle-derived melts with sedimentary and psammitic components of the Snowys Dam Formation during their emplacement (Maier et al., 2016; Smithies et al., 2013; Taranovic et al., 2022). Contamination of the magmatic precursors to the Nova-Bollinger metagabbros in this study likely occurred prior to metamorphism, allowing for aluminium-rich hybridised melts to form garnet-rich, amphibole-bearing gabbroic rocks at slightly higher pressures and lower temperatures to less aluminium-rich gabbroic equivalents in other parts of the Fraser Zone (see also Maier et al., 2016; Smithies et al., 2013).

Cooling pathways of the Nova-Bollinger country rocks

Garnets in the Nova-Bollinger metapelites are compositionally homogeneous, with only minor compositional differences towards their rims that likely reflect retrograde cooling equilibration (Figure 5a–d). Intracrystalline volume diffusion in garnet in granulite-facies metamorphic terranes has been known to cause flattening of compositional profiles (Spear, 1991; Spear & Florence, 1992; Tirone & Ganguly, 2010). This process effectively removes information about the peak metamorphic conditions, which was originally recorded by the garnet cores. Consequently, any P–T calculations derived by pairing garnet cores and matrix biotite do not inform on peak or re-equilibration conditions (Spear, 1991). Some granulite-facies garnets display reverse compositional zoning involving an increase in Mn and a decreased in Mg at the garnet rims. This chemical profile is typically the result of partial re-equilibration with coexisting phases in response to changes in P–T conditions during retrograde cooling. In these cases, garnet rim compositions can provide information about the P–T conditions at which diffusion effectively ceased in the garnet grains (Ikeda, 2004; Spear, 1991; Tirone & Ganguly, 2010). Thermobarometric results from both metapelite samples in this study are similar, and similar P–T constraints were obtained from all garnet–biotite pairs, regardless of the microstructural context of the biotite (0.29–0.45 GPa, 550–640 °C; Figure 8). These P–T results include those obtained from biotite inclusions paired against garnet cores. Therefore, while biotite and garnet preserve textural evidence of the equilibrium mineral assemblage during peak metamorphism, the prograde to peak metamorphic history of the rock has not been preserved by the chemistry of the armoured biotite inclusions in the garnet cores.

In the Nova-Bollinger metapelite country rocks, anhedral, fine-grained and randomly oriented biotite is intergrown with fibrolite around cordierite. This textural relationship suggests that the rocks cooled along a P–T trajectory that involved increasing modal proportions of biotite and sillimanite and decreasing modes of cordierite. Assuming that isochemical conditions were maintained, the metapelites cooled along a near-isobaric to slightly down-pressure P–T path from peak conditions at 0.76 GPa and 880–920 °C, down to at least ∼700 °C Figures 6a, b and 9a). Previous studies have also proposed that retrograde cooling progressed near isobarically (0.6–0.7 GPa) to ∼750 °C (Clark et al., 1999, Clark, 1995; 2014; Glasson et al., 2019), between 1301 and 1293 Ma (Figure 9c–f; Clark et al., 1999). This interpretation is based in part on inferred retrograde P–T constraints from spinel–pyroxene coronae separating plagioclase and olivine in metagabbro in the southern Fraser Zone (0.6–0.7 GPa at 750 °C; Figure 9e; Clark et al., 1999; Clark, 1995).

Figure 9. Summary P–T diagram of geothermobarometry results and peak metamorphic conditions from Nova-Bollinger and other related locales. Isobaric and slightly decompressive cooling paths are shown in blue and red arrows, respectively; the purple arrow denotes inferred the prograde path of Snowys Dam Formation in Kuper et al. (2024). P–T constraints from previous and current studies are shown and obtained from: (a) this study; (b) thermobarometric results obtained from GBAQ-GB from this study; (c) metapelites from the southern Fraser Zone (Clark et al., 2014); (d) garnet-absent igneous and metagabbro rocks from the southern Fraser Zone (Glasson et al., 2019); (e) post-peak isobaric cooling conditions determined in Clark et al. (1999); (f) inferred Stage II Albany–Fraser Orogeny P–T conditions of metapelite between Fraser and Biranup zones (Kirkland et al., 2016); (g) initial formation of symplectic coronae textures at Nova-Bollinger (Torres-Rodriguez et al., 2021); (h) two-pyroxene thermometry recording peak P–T conditions from Nova-Bollinger cumulates (Torres-Rodriguez et al., 2021). Geothermal gradients are displayed in dashed lines; sillimanite-out field based on JC21NB7 and JC21NB33 pseudosections are outlined in a stippled line; kyanite–sillimanite–andalusite isograds are shown and labelled as thick bold black lines.

Microstructural observations in thin-section, and phase equilibria modelling, suggest that the Nova-Bollinger metapelites cooled along broadly isobaric paths at ∼0.6–0.7 GPa to at least ∼700 °C. Isobaric cooling paths in orogenic settings are typically the result of either near-isothermal decompression resulting from gravitational instability and orogenic collapses (Vanderhaeghe, 2012; Whitney et al., 2004). High-grade cordierite-bearing metapelites in the Ivrea Zone record the initial onset of peak P–T conditions, followed by isothermal decompression and subsequent isobaric cooling during collapse of the Variscan Orogeny (Guergouz et al., 2018). Alternatively, isobaric cooling paths may also follow from initial compression and thermal perturbation resulting in slight crustal thickening and large temperature fluxes followed by thermal decay at depth (Vanderhaeghe, 2012). While the prograde thermal history of the Nova-Bollinger rocks remains unconstrained here, recent zircon U–Pb geochronology results from the Snowys Dam Formation have been interpreted to record prograde crystallisation consistent with crustal thickening and basin formation in an extensional environment (Figure 9; Kuper et al., 2024). This indicates that pressures greater than those recorded for peak temperatures are unlikely to have been attained prior to peak metamorphism. Subsequent magmatism derived from aesthenospheric upwelling related to slab delamination, following collision and accretion of the Loongana Arc on the eastern AFO, likely drove crustal thickening and increased temperatures (Kirkland et al., 2016; Kuper et al., 2024; Smithies et al., 2013; Spaggiari et al., 2015, 2018, 2020). Cooling and solidification of these magmas from peak P–T conditions resulted in isobaric thermal decay at depth in the Nova-Bollinger deposit and the wider Fraser Zone (Figure 9; Clark et al., 2014; Glasson et al., 2019; Kirkland et al., 2016; Kuper et al., 2024).

Garnet–biotite thermobarometry on the same samples suggests that the rocks subsequently cooled down-pressure, along a slightly decompressive P–T path, to 0.29–0.45 GPa at 550–640 °C (Figure 9b). Without age data specifically tied to the growth of garnet and biotite in the samples investigated in this study, their rate of cooling remains unconstrained. Apatite thermochronology has been used to suggest that rocks at the Nova-Bollinger deposit cooled at a rate of at least 15 °C/My, from peak conditions at ca 1304 Ma (∼850 °C), to 1284 Ma (Morrison et al., 2022). Assuming that the rocks in this study also experienced peak metamorphism at ca 1304 Ma, their P–T constraints obtained from phase equilibria modelling and thermobarometry are also consistent with this cooling rate. Other studies have suggested that the Fraser Zone cooled to 695–725 °C and 0.65–0.85 GPa by 1205 ± 16 Ma on the basis of phase equilibrium models and U–Pb systematics in titanite from rocks obtained near bounding major faults (Figure 9f; Kirkland et al., 2016). Similarly, biotite and hornblende thermochronometry has suggested that the central and southwestern Fraser Zone cooled to temperatures between 550 and 350 °C between ca 1217 and 1157 Ma (Scibiorski et al., 2016). The variable cooling rates from the southwestern and central regions of the Fraser Zone suggest that internal faulting within the Fraser Zone may have occurred (Morrison et al., 2022; Scibiorski et al., 2016). Recent evidence based on heterogeneous cooling rates obtained within the Fraser Zone has proposed that its internal architecture may be more heterogeneous than previously thought, owing to varied degrees of faulting and exhumation post peak metamorphism that may be reflected in the rocks analysed here (Morrison et al., 2022; Scibiorski et al., 2016).

The interpretation that the Nova-Bollinger system experienced rapid uplift following peak metamorphic conditions (ca 20 Ma; 0.75–0.76 GPa and 880–920 °C; Figure 9a) to shallower conditions (0.29–0.45 GPa at 550–640 °C; Figure 9b) is not incompatible with the work of previous studies that show evidence for slow-cooling textures in sulfide-mineralised mafic–ultramafic rocks in this part of the Fraser Zone (Barnes, Taranovic, Miller, et al., 2020; Barnes, Taranovic, Schoneveld, et al., 2020; Morrison et al., 2022; Taranovic et al., 2021). Cooling and solidification of the silicate and sulfide components of the Nova-Bollinger intrusion are estimated to have occurred over timescales of several tens of thousands of years to millions of years (Barnes, Taranovic, Miller, et al., 2020). In contrast, rapid cooling and uplift of the Nova-Bollinger system to mid–upper crustal levels are inferred to have occurred on the scale of tens of million years (ca 20 Ma), based on the thermochronometric data of Morrison et al. (2022). The latter process is significantly slower than known timescales of silicate solidification (10–100 ka), and within the upper limits estimated for sulfide crystallisation and the formation of sulfide pentlandite–chalcopyrite loop textures that are characteristic of the Nova-Bollinger ores (Barnes, Taranovic, Schoneveld, et al., 2020; Barnes & Robertson, 2019). In summary, the timescales of both solidification of the Nova-Bollinger mafic–ultramafic system and its sulfide ores are compatible with the inferred timescales of cooling and uplift.

Metamorphism as a tool to understand magmatic processes

During Stage I of the Albany–Fraser Orogeny, high-T metamorphism in the Fraser Zone was accompanied by voluminous mafic and felsic calc-alkaline magmatism (Clark, 1995; Clark et al., 1999; 2014; Glasson et al., 2019; Kirkland et al., 2016; Maier et al., 2016; Smithies et al., 2013; Spaggiari et al., 2013; Taranovic et al., 2021, 2022). The general consensus is that high-T metamorphism was facilitated at a regional scale by crustal thickening and subsequent mantle heating, although it is unclear whether this occurred as a result of orogenic collapse, lithosphere delamination and asthenosphere upwelling (e.g. Smithies et al., 2015; Spaggiari et al., 2015), or mantle heating in a back-arc basin (e.g. Glasson et al., 2019; Kuper et al., 2024; Morrissey, Payne, et al., 2017). Recent work suggests that at the Nova-Bollinger deposit, parental magmas to the syn-metamorphic mafic–ultramafic intrusions also enhanced granulite-facies metamorphism by providing an additional, more localised, heat source (Barnes, Taranovic, Miller, et al., 2020; Maier et al., 2016).

Previous metamorphic profiles of the Nova-Bollinger intrusion were based on P–T estimates obtained from metapelites elsewhere in the Fraser Zone that may have been metamorphosed within the orogenic system relatively distal from a magmatic heat source (Barnes, Taranovic, Miller, et al., 2020; Clark et al., 2014; Glasson et al., 2019; Taranovic et al., 2021). Conversely, the peak P–T constraints from our study suggest that that the Nova-Bollinger host rocks reached slightly higher temperatures than previously thought (up to 920 °C) and that peak metamorphism occurred close to the inferred P–T conditions of the Nova-Bollinger intrusion when it was first emplaced in the mid–lower crust (0.76–0.96 GPa at ∼1035 °C; Figure 9g, h; Torres-Rodriguez et al., 2021).

The broadly coeval timing of emplacement of the Nova-Bollinger intrusion (ca 1304 ± 22 Ma; Morrison et al., 2022) and the onset of granulite-facies metamorphism and deformation in the Fraser Zone (ca 1304 ± 7 Ma; Spaggiari et al., 2015) provides strong evidence for a causal relationship between magmatism and high-T metamorphism at the Nova-Bollinger deposit (Clark et al., 2014; Glasson et al., 2019; Morrison et al., 2022; Torres-Rodriguez et al., 2021). It is plausible that the Nova-Bollinger host rocks record more elevated peak temperatures and higher thermal gradients than the metapelites sampled by Clark et al. (2014) from the southern Fraser Zone because they are the metamorphic expression of a superimposed thermal regime. It is proposed that their P–T conditions may reflect the combined effects of regional scale high-T conditions during orogenesis (for reasons discussed above), and contact metamorphism in a thermal aureole adjacent to the Nova-Bollinger mafic–ultramafic magmas (Barnes, Taranovic, Miller, et al., 2020; Clark et al., 2014; Glasson et al., 2019; Kirkland et al., 2016; Maier et al., 2016; Taranovic et al., 2021).

The samples analysed in this study were collected proximal to mineralised sulfide ores hosted in country rocks (12–20 m away from massive sulfide in relevant drill cores). While the presence of sulfide mineralisation is limited to pyrrhotite blebs in all samples, we highlight that sulfide liquids infiltrating country rocks could also impart a volumetrically small but significant heat source over short timescales (see above), acting as a third localised thermal driver of granulite-facies conditions at the Nova-Bollinger deposit. Sulfide liquids are known to have significantly higher densities and heat capacities than their silicate magma counterparts (Barnes, Taranovic, Miller, et al., 2020; Kress et al., 2008). At the Nova-Bollinger deposit, the mass of the sulfide component in the Nova-Bollinger orebodies has been estimated to approximately 20% of the mass of the lower intrusion (Taranovic et al., 2022). This component includes significant sulfide infiltration ores that extend up to 100 m beneath the lower intrusion and into the host country rocks (Barnes, Taranovic, Miller, et al., 2020). Post peak-metamorphism, the Fraser Zone is thought to have been variably exhumed to upper–mid-crustal depths. Prior to this, abundant pentlandite–chalcopyrite ‘loop’ textures in the Nova-Bollinger ores (Barnes, Taranovic, Miller, et al., 2020; Barnes, Taranovic, Schoneveld, et al., 2020), the flat compositional profiles in garnet and metamorphic reaction textures in the surrounding host rocks (Barnes, Taranovic, Miller, et al., 2020; Taranovic et al., 2021) indicated that the rocks cooled relatively slowly, eventually converging onto prevailing peak metamorphic conditions (Barnes, Taranovic, Miller, et al., 2020).

Heat conduction from the syn-metamorphic solidified intrusions and crystallising sulfide liquids would have been sustained, prolonging high-T conditions in the surrounding host rocks (Barnes, Taranovic, Miller, et al., 2020; Barnes & Robertson, 2019; Taranovic et al., 2021). It is likely that while the Nova-Bollinger intrusive magmas remain the primary contributor of enhanced heat in the surrounding country rocks, localised thermal contributions by infiltrating sulfide may still be substantial in expanding the thermal envelope of the system. Despite the relatively smaller volume sulfide has in comparison with its host intrusion, the protracted solidification and subsequent conductive cooling of sulfide mineralisation within host country rocks may represent a smaller but longer-lived agent to be considered in understanding the thermal histories of the Nova-Bollinger deposit.

Role of high-grade metamorphic rocks in orthomagmatic deposit formation

The presence of high-grade metamorphic rocks, particularly those within the mid–lower crust, does not immediately imply the presence of viable orthomagmatic deposits and require a combination of favourable factors (Barnes, 2023; Latypov et al., 2024). However, the lithological composition of mid–lower crustal settings can have inherent advantages over those at shallower depths. In particular, the lithological composition of mid–lower crustal settings may influence sulfur saturation through contamination and reduction in sulfur contents (Jesus et al., 2020). As depth increases, there is a reduced need for external sulfur to achieve sulfide saturation owing to the inverse relationship between sulfur solubility and pressure (Mavrogenes & O’Neill, 1999), especially for magmas that were already enriched in S.

Thickened crustal profiles, typically generated during syn–late orogenesis, may serve to facilitate high-pressure fractionation as upwelling magmas encounter and stall against and within crystalline high-grade metamorphic rocks at depth (Barnes & Lightfoot, 2005; Jesus et al., 2020; Latypov et al., 2024). High-grade metamorphic rocks may also act as sulfur contributors that can be assimilated during magma transport (Ripley & Li, 2013). Hybridisation and contamination of mantle-derived melts by granulite-facies country rocks in the Fraser Zone, including those belonging to the Nova-Bollinger intrusion, indicate that high-grade metamorphic terranes may contribute sulfur to ascending magmas (Barnes, Taranovic, Miller, et al., 2020; Maier et al., 2016; Smithies et al., 2013; Taranovic et al., 2021, 2022). Additionally, the role of volatile C in sulfide liquid transportation within the mid–lower crust is increasingly recognised (Blanks et al., 2020; Cherdantseva et al., 2024; Yao & Mungall, 2022). The rocks investigated in this study do not cover metamorphosed carbonate rocks. However, their close proximity to and possible local assimilation by Nova-Bollinger magmas have been reported in Taranovic et al. (2022). The presence of high-pressure marbles, sulfur-containing metamorphic rocks and the dynamics of sulfur solubility in high-pressure conditions imply that high-grade metamorphic terranes can provide suitable conditions and material for the formation and preservation of orthomagmatic sulfide deposits, especially those found in collisional settings (Begg et al., 2010; Jesus et al., 2020; Latypov et al., 2024; Ripley & Li, 2013).

Orthomagmatic Ni–Cu–Co deposits have typically been associated with large igneous provinces and mantle plumes (Naldrett, 2004). However, the increasing recognition of orthomagmatic deposits within orogenic settings indicates that while plume-related magmatism is important, it is not essential, in the formation of base-metal sulfide within the crust (cf. Ezad et al., 2024). The close association of high-grade metamorphic rocks typically hosting these deposits suggests that prospectivity within these terranes can be viable, as reported in Savannah, WA (Le Vaillant et al., 2020) and the Nova-Bollinger deposit (Taranovic et al., 2021). Discoveries of orthomagmatic ores in such terranes may thus reflect a function of uplift, erosion and exposure that are dependent on modern characterisation of these terranes rather than the absence of actual mineralised deposits within the mid–lower crust (Latypov et al., 2024). This is made more apparent by the increasing recognition and understanding of deep crustal systematics, ore formation and trans-lithospheric metal transfer (Barnes, 2023; Fiorentini et al., 2018; Holwell et al., 2022; Latypov et al., 2024; Locmelis et al., 2021). As such, an understanding of the thermotectonic history of the host rocks can have significant importance for exploration.

Conclusions

The use of phase equilibria modelling methods has been shown to be useful in determining peak P–T conditions in the Nova-Bollinger host rocks where P–T conditions were calculated at approximately 0.76 GPa and 880–920 °C. These P–T conditions are consistent with previous P–T estimates obtained within the Nova-Bollinger intrusion (Torres-Rodriguez et al., 2021) and with igneous and metamorphosed garnet-absent gabbro studies conducted in the southern Fraser Zone (Glasson et al., 2019). However, the calculated peak conditions are relatively higher than those calculated for metapelites in Clark et al. (2014): it is argued that the higher thermal conditions in Nova-Bollinger country rocks are related to its residence within the thermal aureole of the Nova-Bollinger intrusion and were superimposed to the granulite-facies metamorphic conditions experienced in the wider Fraser Zone. Additionally, it is proposed that sulfide liquid infiltration into host rocks may represent a small but significant localised heat contributor, extending the thermal aureole of the intrusion.

A broadly isobaric cooling path to at least 700 °C following peak metamorphism is inferred for the Nova-Bollinger metapelites based on microstructural observations and modelled phase equilibria assemblages. The inferred isobaric cooling path documented here contrasts with isobaric cooling paths that are commonly observed in orogenic collapses, such as the Ivrea Zone in Italy during the late-Variscan orogeny, but is consistent with isobaric cooling paths inferred within the wider Fraser Zone, which resulted from thermal decay at depth following voluminous magmatism in Stage I Albany–Fraser Orogeny. Conventional GBAQ-GB thermobarometry conducted on garnet cores–matrix biotite pairs in metapelites yields conditions of 0.29–0.45 GPa at 550–640 °C but remains unconstrained in age, potentially reflecting subsequent late-stage segmented faulting within the Fraser Zone.

Acknowledgements

We thank Steve Barnes and an anonymous reviewer for providing constructive comments on an earlier version of this manuscript. This project benefitted from a collaborative research agreement between the University of Western Australia and IGO Ltd, which provided generous financial and in-kind support including access to drill-core samples and underground mine operations at the Nova-Bollinger Ni–Cu deposit. Malcolm Roberts, CMCA UWA, is thanked for assistance with EPMA setup, calibrations and data processing. The first author, Joshua Chong, is the recipient of a SIRF scholarship from the University of Western Australia.

Disclosure statement

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials. Additional data that support the findings of this study are also available from the corresponding author, Chong, J, upon reasonable request.