Graphitization of Neoproterozoic sedimentary marbles in The Aird, Scottish Highlands

The controls of carbon removal from a protolith carbonate are dependent on oxidation- or reduction-related carbonate mineral destabilization, and the mixing of C–O–H fluids during amphibolite- to granulite-facies metamorphism associated with slab descent in subduction zones (Ague 2000; Galvez et al. 2013; Ague and Nicolescu 2014), orogenic episodes (Rumble et al. 1986; Evans et al. 2002) or igneous-intrusive shearing (Satish-Kumar 2005). Carbonate-derived graphitization can occur through two processes: decarbonation and devolatilization. The former varies from carbonate reduction ± mixing with mantle-derived fluids (Rumble et al. 1986) that yields isotopically heavy graphite (Satish-Kumar et al. 2011; Luque et al. 2012; Parnell et al. 2021); and the latter through mixing of carbon-saturated aqueous fluids derived from carbonate and sedimentary protoliths, which usually yields lighter carbon isotopic compositions of graphite (Evans et al. 2002; Satish-Kumar 2005; Luque et al. 2012). Decarbonation-related graphitization in the absence of igneous fluids usually occurs through carbonate reduction and subsequent reactions of a carbon-saturated C–O–H fluid (Galvez et al. 2013) between authigenic minerals, calcite and/or dolomite with quartz or other silicates (Luque et al. 2012). Devolatilization-related graphitization occurs from the oxidation and dissolution of carbonate minerals (i.e. calcite and dolomite), leading to volatile reactions with low oxygen fugacity fluids derived from organic carbon and/or silicate minerals (e.g. micas, garnets and amphiboles) in metasedimentary rocks, usually within interlayered or conformable sequences (e.g. Rumble et al. 1986; Evans et al. 2002; Satish-Kumar 2005).

Graphite can occur in a range of geological environments, from the contact or regional metamorphism of kerogen in carbonaceous sediments (Wada et al. 1994; Wright et al. 2012) to fluid-phase deposition from carbonate or igneous fluids (Ferraris et al. 2004; Luque et al. 2012; Galvez et al. 2013). The mineralization of graphite from a fluid phase can be influenced by minerals which have layer-silicate surface properties at the nanoscale that enhance admixture with carbon (Moro et al. 2017). Phlogopite mica is one such mineral, occurring commonly within ultramafic and carbon-rich magmas, as well as in marbles. The association of phlogopite with reduced carbon in marbles (Peck et al. 2006) can make it challenging to distinguish them from the same association in mantle-derived carbonatitic intrusions (Le Bas et al. 2002; Ferraris et al. 2004; Humphreys-Williams and Zahirovic 2021). Graphite–phlogopite mixtures in mantle-derived ultramafic and carbonatite rocks suggest that phlogopite can act as a preferred template for carbon, and have been interpreted to indicate a mantle reservoir of carbon (Ferraris et al. 2004; Rajesh et al. 2009; Ashchepkov et al. 2015; Kelemen and Manning 2015; Galvez et al. 2020; Humphreys-Williams and Zahirovic 2021). Graphite–phlogopite admixtures have been found in mantle-origin Iherzolite (Ferraris et al. 2004), peridotite xenolith mixtures (Kaeser et al. 2007), kimberlite pipes (Kostrovitsky et al. 2001), carbonatitic fluids described in peridotite (Naemura et al. 2009) and carbon–phlogopite assemblages in the mantle wedge of subduction zones (Kelemen and Manning 2015).

Graphite–phlogopite assemblages have been known to occur in metamorphosed carbonate-metasedimentary rocks. Some of these graphite–phlogopite occurrences occur in gemstone marbles, notably the ruby-bearing marble–sillimanite gneiss belts from Afghanistan to China (Yang et al. 2019) and Vietnam (Hauzenberger et al. 2005), ruby/corundum-bearing marble/calc-silicates in British Columbia (Dzikowski et al. 2014) and sapphire/corundum-bearing marble/calc-silicate on Baffin Island (Belley et al. 2017). Graphite–phlogopite assemblages are widely documented in mixed carbonate–silicate deposits, including the Ampandrandava metacarbonate interlayered with graphitic-sillimanite schist in Sri Lanka (Satish-Kumar et al. 2021) and granulite-facies marble assemblages in the Skallen region of East Antarctica (Satish-Kumar and Wada 2000). An example of graphite-bearing marbles comparable with those in The Aird forms part of a suite of granulite-facies Mesoproterozoic marbles in the New Jersey Highlands (Peck et al. 2006). Graphite–phlogopite nodules occur in pyritic, calcite marble in Lime Quest, New Jersey, where nodules up to 1 cm in size were identified and extracted by Jaszczak (1997). Graphite–phlogopite nodules shown in Jaszczak (1997) highlighted varied graphite–phlogopite crystal morphologies of basal pinacoid faces of millimetre-scale tabular ± overgrowth crystals, hexagonal crystals and spherical crystal clusters of graphite minerals. Graphite also occurs nearby in marble at Ogdensburg, New Jersey (Palache 1941; Jaszczak 1994). Palache (1941) and Jaszczak (1997) identified graphite–phlogopite with graphite overgrowths, and spherical graphite attached to phlogopite crystal surfaces.

Graphite–phlogopite assemblages, their crystal atomic structure and crystal composition were analysed using different forms of scanning electron microscopy (SEM), including electron dispersive spectroscopy (Jaszczak 1997), X-ray diffraction (XRD) using field emission scanning electron microscopy (FE-SEM) (Satish-Kumar et al. 2011) or electron microprobe analysis (Ferraris et al. 2004). Hydrocarbon–phlogopite-bearing fluid inclusions were also imaged using confocal microscopy dimensions at the nanometre to micrometre scale (Moro et al. 2017). Carbon isotopic compositions, SEM microphotography and Raman spectroscopy of nodular graphite–phlogopite constrained temperature-dependent calcite–graphite geothermometry, graphite morphology and crystallinity in studies by Satish-Kumar et al. (2011), who distinguished two distinct graphite populations: the highly crystalline, platy graphite, mineralized in isotopic equilibrium with carbonate-derived CO₂ fluids; and phlogopite-enclosed disordered graphite with distinctively lower $δ$¹³C compositions. Graphite–phlogopite ordering is discussed in Ferraris et al. (2004), who also utilized Raman spectrometry alongside high-resolution transmission electron microscopy (HRTEM) to indicate co-precipitation and topotactic intergrowth of laminae graphite and phlogopite, connected by layers of weak van der Waals bonds.

The Lewisian Foreland Terrane, Northern Highlands Terrane and Grampian Terrane are three distinct geological regions in the Scottish Highlands that comprise Proterozoic successions. The Northern Highlands Terrane comprises polymetamorphosed, late Mesoproterozoic–Neoproterozoic (1000–870 Ma) fluvial–marine metasandstone and metapelite successions cut by Caledonian intrusive rocks (Strachan et al. 2010; Cawood et al. 2015; Krabbendam et al. 2022). The Moine Thrust (Fig. 1) separates the Northern Highlands Terrane from Archean–Paleoproterozoic gneisses, Neoproterozoic Torridon sandstones and later Paleozoic rocks in the Lewisian Foreland Terrane. The southern boundary of the Northern Highlands Terrane is the Great Glen Fault, which separates it from the Neoproterozoic–Lower Paleozoic metasediments and Caledonian intrusive rocks in the Grampian Terrane (Chew and Strachan 2014; Krabbendam et al. 2022).

The Wester Ross Supergroup in the Northern Highlands Terrane is confined to the Moine Nappe between the Caledonian or older Moine and Sgurr Beag thrusts, and is comprised entirely of the Morar Group of metasedimentary lithologies that record Renlandian prograde amphibolite-facies metamorphism at c. 950–940 Ma (Bird et al. 2018). The structurally overlying Loch Ness Supergroup (Krabbendam et al. 2022) records the Sgurr Beag Nappe-hosted Glenfinnan and Loch Eil groups, that were deposited prior to the 840–725 Ma Knoydartian Orogeny (Mako et al. 2021; Krabbendam et al. 2022). Metasedimentary rocks of The Aird and Glen Urquhart are positioned in the hinterland of the Sgurr Beag Nappe. Mako et al. (2021) described polymetamorphosed accessory phases in Glen Urquhart, and the isotopic resetting of monazite and xenotime in 750 Ma quartz-feldspathic metasedimentary rocks by later reprecipitation at 600 and 425 Ma.

The Northern Highlands Terrane hosts inliers of Archean–Paleoproterozoic basement, comprising eclogites (Bird et al. 2023), mafic–intermediate orthogneisses and reworked amphibolite-facies 2.5–1.8 Ga supracrustal rocks (Strachan et al. 2020b); the latter described by Rock et al. (1987) as comprising amphibolite–granulite-facies silicic gneisses, amphibolites, calc-silicates and marbles. The Caledonian Orogeny reworked numerous stratigraphic successions in the Northern Highlands Terrane through greenschist–amphibolite-facies regional metamorphism and tight folding during the 470–460 Ma Grampian event (i.e. Grampian I) (Strachan et al. 2010; Mako et al. 2021; Krabbendam et al. 2022). A further compressional event in the west and far north of the Northern Highlands Terrane, termed Grampian II, occurred at c. 450 Ma (Bird et al. 2013), followed by plutonism at Glen Dessary (Goodenough et al. 2011; Milne et al. 2023). Orogenesis related to Grampian II in the Glenfinnan Group (Bird et al. 2013) and fault-bound plutonism (Milne et al. 2023) is documented to the west of the Great Glen Shear Zone. However, this remains underexplored in Glenfinnan Group rocks to the north of the Great Glen Shear Zone (Bird et al. 2013). The Scandian event between c. 437 and c. 415 Ma (Oliver et al. 2008; Strachan et al. 2010) led to widespread upright refolding of earlier structures, and from c. 432 Ma onwards was accompanied by widespread intrusion of granitoids (Milne et al. 2023). Some of these bodies in Glen Scaddle and the Assynt Alkaline Complex record Scandian ductile deformation (Strachan and Evans 2008; Goodenough et al. 2011), whilst others are unaffected (e.g. Cluanie: Milne et al. 2023).

Regional transpression throughout northern Scotland and granite emplacement through transpressive and transtensive faults (i.e. Newer Granites) occurred following the Scandian Orogeny at c. 430 Ma (Dewey and Strachan 2003; Strachan et al. 2010, 2020a; MacRae et al. 2023; Milne et al. 2023). The intrusion of the Abriachan Granite has recently been constrained to 420.1 ± 3.4 Ma (MacRae et al. 2023), placing a further upper limit on widespread regional fenitization and alkaline carbonatitic mineralization (Deans et al. 1971; Garson et al. 1984; Heptinstall et al. 2023). Sinistral displacement on the Great Glen Fault system dates back to at least c. 428 Ma where it is recorded in shear zones associated with the emplacement of Clunes tonalite (Stewart et al. 2001). Sinistral motion is considered to have persisted until c. 400 Ma, when syntectonic granitoid emplacement is recorded at Rosemarkie (Mendum and Noble 2010). Within the overall context of sinistral fault motion, there was a switch from regional transpression to transtension at the end of the Scandian Orogeny at c. 415 Ma (Dewey and Strachan 2003; Strachan et al. 2020a). Late Caledonian successor basins were infilled in the Devonian by Old Red Sandstone successions to the north and the east of the Northern Highland Terrane (Parnell 1985; Elmore et al. 2006). The Great Glen Fault experienced multiple phases of largely dextral reactivation during basin inversion and extensional events, including during the later Devonian, Carboniferous–Permian, Jurassic–Cretaceous and early Cenozoic (Underhill and Brodie 1993; Le Breton et al. 2013; Dichiarante et al. 2016, 2020; Tamas et al. 2023).

The Aird marble was quarried at four main sites in the Glenfinnan Group: at Rebeg, South Clunes, Blairnahenachrie and Kirkton (Jolly and Cameron 1880; Wallace 1885, 1887; Peach et al. 1913; Horne and Hinxman 1914). Rock et al. (1984) described the South Clunes, Rebeg and Kirkton quarries as consisting predominantly of marble, which is interlayered with pelitic schists at Kirkton. The marbles are known to be cross-cut by pegmatites, and exhibit brecciated fabric, hematite veins and fenitization (metasomatic alteration by alkaline fluids). Fenitization was described by Garson et al. (1984), particularly veins bearing the blue amphibole magnesioriebeckite (previously termed crocidolite because of the colour). Fenitization has been interpreted as an alteration of country rock by alkaline and carbonatite fluids (Elliott et al. 2018) in the region of the Great Glen Fault (Deans et al. 1971; Garson et al. 1984). Marble in The Aird is today exposed as isolated outcrops and fluvial boulders in the Moniack Burn derived from the quarries at Rebeg and South Clunes.

Limestones are poorly documented in the Wester Ross Supergroup and Loch Ness Supergroup successions of the former Moine Supergroup (Rock et al. 1986b). Although they are barely exposed now, old records indicate that the marbles of The Aird were formerly more widely reported and mined. The New Statistical Account for Scotland (Fraser 1845) records ‘numerous beds of primary granular limestone’ (p. 492), and Wallace (1885) reported that quarrying was ‘very successfully carried on for a good many years’, ‘kilns … on an unusually large scale for this district’ and ‘great quantities were shipped, and for this purpose a canal was cut by the late proprietor’ (p. 170). These descriptions do not accord with the assumed lack of limestone in the Neoproterozoic.

Marbles in The Aird stratigraphically overlie other metasedimentary rocks of Neoproterozoic age, laid down around c. 870 Ma, and may have undergone prior metamorphism in Proterozoic times. However, the influence of the Caledonian Orogeny in The Aird is indicated by zircon hosted in intrusive pegmatite (Cutts et al. 2010), which has yielded a U–Pb age of 463 ± 4 Ma. This age is considered to correspond to a time of folding and fabric development in the Loch Ness Supergroup during the Grampian Orogeny. Monazite hosted in graphite–kyanite schist at Glen Urquhart gave a similar U–Pb date of 473 ± 2 Ma (Mako et al. 2021). Carbonatitic veining in the Great Glen region was previously documented by Garson et al. (1984) and has recently been investigated for trace metal-bearing minerals in The Aird carbonatitic veins (Heptinstall et al. 2023). The carbonatite veins described in The Aird may have been intruded coeval with the c. 429–425 Ma emplacement of the ‘Newer Granites’ (Rogers and Dunning 1991; Milne et al. 2023) and mafic–peralkaline Moine Thrust granitoids (Goodenough et al. 2011; Walters et al. 2013), the emplacement of the c. 420 Ma Abriachan alkaline granite (MacRae et al. 2023), or later lamprophyre dyke swarms that were intruded up to c. 410 Ma elsewhere in the Great Glen Shear Zone (Searle 2022). This study will assess carbon and sulfur isotopic compositions of graphite and accessory sulfide minerals to investigate the paragenesis and post-Grampian I/II modification of graphite–phlogopite minerals during the Caledonian Orogeny.

Graphite, phlogopite and pyrite-bearing samples in The Aird were collected by the authors, based in the Geology and Geophysics Department at the University of Aberdeen, from outcrops at the Kirkton and Rebeg quarries, an exposure at Blairnahenachrie, and from a stream section cutting marbles in Moniack Burn (Fig. 1). Key specimens highlight graphite mineralization and Caledonian Orogen paragenesis (Fig. 2). The Moniack Burn stream section is adjacent to outcrops of highly weathered black marble with cross-cutting carbonatitic veining at Rebeg Quarry. A single c. 3–4 kg block of graphitic marble was sampled from the poorly exposed bedrock at Glen Urquhart. We do recognize, however, that a single sample at Blairnahenachrie and Glen Urquhart can have representative limitations.

High-resolution image and compositional analyses were performed at the University of Aberdeen ACEMAC Facility using a Zeiss Gemini field emission gun scanning electron microscope (FEG-SEM) on polished blocks of the marble. Samples were carbon coated and analysed at 20 kV, with a working distance of 10.5 mm. Marble samples were examined through electron dispersive spectrometry (EDS) of graphite-bearing nodules (Fig. 3), and graphite entrained within pyrite veins (Fig. 4) was analysed using Oxford Instruments EDS X-ray analysis to ascertain compositional data on the phlogopite minerals, which assumes 22 oxygen equivalents (Table 1). The standards used were a mixture of natural minerals, metal oxides and pure metals, with calibrated by per the factory specification. Stoichiometry was determined by fixed oxygen contents of specific minerals, which has a compositional error of c. 0.01 wt%, with atomic unit calculations assuming 11 oxygens per formula in order to be comparable to the phlogopite empirical formula.

Optical and polarized light microscopy of thin sections was carried out at the University of Aberdeen on a Leica DM750 P microscope, and the photography used a Leica ICC50 E high-resolution imager.

Raman analyses were performed on graphite before and after calcite dissolution by HCl to check the effect of induced disorder on the sample surfaces. Raman analyses were first performed on unpolished cut surfaces and then SEM observations were made. This approach was necessary in order to avoid Raman spectroscopy after polishing. Micro-Raman spectroscopy of graphite samples was carried out at the University of Aberdeen using a Renishaw inVia reflex Raman spectrometer, with a backscattering geometry in the range of 700–3200 cm⁻¹, a 2400 l mm⁻¹ spectrometer grating and a CCD detector (Fig. 5). Microscopic observations were carried out with a ×100 optical power objective with a numerical aperture (NA) of 0.90. The slit opening was 65 µm with a CCD area of approximately 10 pixels (80% of the total signal height hitting the CCD chip) and the confocal hole was 200 $μ$m. A 514.5 nm diode laser was used for excitation with an output of 50 mW. Optical filters (10%) were used to adjust the power of the laser to less than 5 mW. Parameters for the Raman spectra of carbonaceous material were calculated after deconvolution using LabSpec software (Schito et al. 2017). Palaeotemperatures (Table 2) were calculated using the conversion of the R2 ratio according to Beyssac et al. (2002).

Sulfur isotope analysis took place at the Scottish Universities Environmental Research Centre (SUERC), pyrite samples from The Aird were combusted with excess Cu₂O at 1075°C in order to liberate the SO₂ gas under vacuum conditions. Liberated SO₂ gases were analysed on a VG Isotech SIRA II mass spectrometer, with standard corrections applied to raw $δ$⁶⁶SO₂ values to produce true $δ$³⁴S. The standards employed were the international standard NBS-123 and IAEA-S-3, and SUERC standard CP-1. For comparison, pyrite sampled from the carbonatite body at Loch Urigill in the Loch Borralan Complex, during sampling for Heptinstall et al. (2023), was also measured (Table 3).

Stable carbon isotope analysis was conducted on graphite samples digested in 10% HCl overnight to remove trace carbonate. Samples were analysed by the standard closed-tube combustion method by reaction in a vacuum with 2 g of wire-form CuO at 800°C overnight. Data are reported in per mil (‰) using the $δ$ notation v. Vienna Pee Dee Belemnite (V-PDB). Repeat analysis of the SUERC laboratory standard gave $δ$¹³C reproducibility of around ±0.2‰ (1s) (Fig. 6). The graphite samples were collected by dissolving calcite-rich layered marble, which included graphite–phlogopite-enriched layers, entrained graphite in calcite and poorly defined layers of graphite–Mg chlorite ± replacive muscovite (Table 2). However, the technique did permit distinctions between different graphite forms for isotopic analysis. Crystal textural characteristics of graphite minerals recorded in optical, polarized and EDS analysis include mineral assemblage arrangement, and graphite crystal size, shape and presence of twinning or overgrowths.

Fluid-inclusion studies were performed at the University of Aberdeen on doubly polished wafers using a Linkam THMS-600 heating–freezing stage mounted on a Nikon Labophot transmission light microscope. The instrument is equipped with a range of objective lenses, including a ×100 lens, and was calibrated against synthetic H₂O (374.1 and 0.0°C) and CO₂ (−56.6°C) standards (Synthetic Fluid Inclusion Reference Set, Bubbles Inc., USA). Approximate pressure corrections were made following the approach of Goldstein and Reynolds (1994). Homogenization temperatures were measured by observing the first appearance of a vapour bubble on reheating the homogenized fluid (Table 2).

Marble exposures in The Aird included a c. 4 × 3 m outcrop at Kirkton (NH 606452), a c. 3 × 2.5 m outcrop at Rebeg Quarry (NH 562422), displaced boulders of marble up to 0.8 m in width at Blairnahenachrie (NH 597415) and Moniack Burn riverbed samples up to 1.5 × 0.8 × 0.6 m in size (NH 558425). The Kirkton exposure comprises black graphitic marble, interlayered with schist, with veins of hematite and pegmatite. The Rebeg Quarry consists of highly weathered and fissile black marble, with hematite and carbonatite veining. The marble boulders at Blairnahenachrie include a single large block of white, granular marble that showed no discernible layers or veining. Outside of The Aird, samples were collected from centimetre-scale pyrite and carbonatitic veins sampled from psammite in Learnie Quarry (NH 736614). Pyrite and graphite were sampled from marble in the nearby, poorly exposed, Glen Urquhart inlier (NH 484316).

Moniack Burn riverbed debouched boulders are likely to have been transported from the Rebeg and South Clunes quarries further upstream. Moniack Burn comprises approximately millimetre- to centimetre-scale layers of calcite, dolomite, micaceous minerals, graphitic and pyrite (Fig. 2a). Graphite-rich horizons usually occur in layered calcite marble and calcite–dolomite marble that was later deformed into tight folding. Layered marble samples are cross-cut by both calcite and sulfide veins, which usually followed earlier calcite veins or penetrated along layers in the marble. These calcite and sulfide veins usually occur as c. 1–8 cm-long and <1 cm-wide veins in layered marbles. Sulfide veins assimilated graphite from The Aird marbles. Boulder specimens also included basement schist comprising hydrothermal breccia ‘veins’ of comminuted country rock matrix with larger xenoliths of quartzite, schist and macroscopic quartz (also described in Heptinstall et al. 2023), all cross-cut by calcite veining (Fig. 2b). Debouched tightly folded marble is cross-cut by undeformed hematite veins (Fig. 2c), which are also present in in situ exposures of metasedimentary and marble host rocks in Rebeg Quarry and the Moniack Burn riverbed.

Hematite occurs especially in brecciated marble, where it coats the margins of calcite veining, and forms zones of diffuse impregnation in the country rock. Hematite occurs as both deformed veins and impregnations (Fig. 2d), and sub-millimetre width planar veins of hematite (Fig. 2c). However, it is unclear if hematite veining and impregnations occurred during post-Scandian-associated Great Glen Fault deformation, or much younger hematite veining events.

Pegmatites cross-cut outcrops of marble at Kirkton and Rebeg, and boulders in Moniack Burn, ranging from <1 cm to >10 cm in width, are shown to cross-cut marbles at Rosemarkie. Minor brecciation has altered the pegmatite bodies.

Graphite-rich horizons occur in boulders of layered calcite and calcite–dolomite marble in Moniack Burn, and as less distinct horizons and single crystals at Kirkton and Blairnahenachrie. Graphite horizons comprise numerous nodules, commonly with phlogopite (Fig. 2a). Graphitic nodules typically have dimensions of 50 µm up to several hundred micrometres wide (Figs 2a and 3), and can comprise several coalesced graphite crystals. Nodules occur in discrete layers up to several centimetres in length.

Graphite commonly occurs within nodules in association with phlogopite at most localities (Fig. 3a, c, e and f), and also occurs with Mg-chlorite (Fig. 3b, c), pyrite veins (The Aird) (Fig. 4) and kyanite (Glen Urquhart). Smaller quantities of the minerals calcite (Fig. 3b), sphene (Fig. 3c), quartz (Fig. 3f) and replacive muscovite (Fig. 3b) are also present in graphite-hosted nodules. Graphite occurs as lath-shaped (Fig. 3a, c and f) or equidimensional (Fig. 3b–e) crystals. Graphite crystals removed from partly dissolved marble samples exhibited a range of bulbous and platy forms; however, crystallography analysis was not carried out as it is beyond the scope of this work. The graphite does not show overgrowths, and the different graphite forms occur independent of each other, often in layers of the same sample. Collectively, graphite–phlogopite horizons are up to 0.5 cm in width (Fig. 2a).

Graphite crystals commonly occur in lath-shaped assemblages (Fig. 3a); in irregular aggregate assemblages with other silicate and calcite minerals (Fig. 3b); in subordinate assemblages around one or more central silicate crystals (Fig. 3c), where several small graphite crystals may occur at the periphery of phlogopite and Mg-chlorite nodules (Fig. 3c); or as solitary grains in calcite (Fig. 3d). Graphite–phlogopite nodules are the most abundant of the different forms of graphite-bearing nodule identified in marble-hosted rocks of The Aird. Graphite usually occurs with phlogopite but there is much phlogopite with no associated graphite. Graphite also occurs within dark laminae in the marble composed of a meshwork of sub-millimetre quartz and feldspar crystals (Fig. 2a). In places, graphite crystals define the layers and they allow folding to be discerned (Fig. 2c). Graphite-associated phlogopite measured in rocks from The Aird have fluorine (FO) contents of up to 2.2 wt% (Kirkton), 1.93 wt% (Blairnahenachrie) and 2.03 wt% (Moniack Burn) (Table 1); with an approximate phlogopite stoichiometry of K_{0.655–0.937}Mg_{2.077–2.658}Al_{1.289–1.491}Si_{2.847–2.926}O₁₀(F_{0.717–1.186},OH).

Sulfides occur in two settings: as distinct minerals or sub-millimetre scale masses within the layered marble, and as pyrite-rich veins that also host barite and Pb-minerals. Pyrite-rich veins followed calcite veins along layers of the marble country rock (Fig. 2a), and as veins in Learnie Quarry psammite. Graphite inclusions occur within sulfide veins (Fig. 4a) and outside the limits of sulfide veins in layered marble (Fig. 4b). Graphite crystals are typically <100 µm in size, both within (Fig. 4a) and outside of pyrite veins (Fig. 4b).

Calcite veining occurs as mostly linear (Fig. 2a) and, less often, as irregular (Fig. 2b) veins. Calcite veining varies in scale, from c. 2–3 cm in width and several centimetres in length to thin (<1 cm) and 3–5 cm-long calcite veins that are shown to comminute the schist matrix (Fig. 2b). Micrometre-scale calcite veins group together into a mass several hundred micrometres wide or splinter off from veins approximately millimetres in width and centimetres in length, that were described in Heptinstall et al. (2023) as enriched in Sr, Sc, V, Y, Nb, Ba and REE.

Moniack quartz crystals within breccia zones contained primary two-phase (liquid and vapour) aqueous fluid inclusions up to 10 µm in size (Fig. 2b), from which a mean homogenization temperature of 122.5°C (SD = 9.9°C; n = 26) was measured. Petrographical study suggests that quartz mineralization was coeval with brecciation. The homogenization temperature probably equates to at least 150°C after pressure correction. The quartz-bearing fluid inclusions did not contain CO₂, and so are unlikely to be related to calcite dissolution during metamorphism or calcite veining. They are crosscut by later veins of calcite and smeared out by low-displacement (a few millimetres) small faults (Fig. 2c). Graphite also occurs in some calcite veins filling cross-cutting brittle fractures up to millimetres in width. Tight folding in the marble affecting the calcite–dolomite–graphitic horizons (Fig. 2c) and the early hematite (Fig. 2d) suggests that ductile deformation occurred following graphitization and the intrusion of oxic fluids.

The Moniack Burn (see Fig. A1a in Appendix A), Kirkton (Fig. A1b) and Blairnahenachrie (Fig. A1c) thin sections host tabular to prismatic, phlogopite minerals that occur as light brown crystals in visible light and anisotropic at c. 90° extinction angles. Graphite shown in preferential association with phlogopite in a c. 1 mm-wide graphitic horizon, occurs as mostly tabular, lath-shaped to equidimensional crystals (Fig. A1a, b). Graphite crystal morphology varies from exhibiting flat edges and apparent basal-pinacoid faces (especially in lath-shaped assemblages) to irregular crystal edges. Monoclinic, <0.5 mm Mg-chlorite crystals occur as light green, prismatic crystals with anomalous (blue–green–yellow) birefringence colours and perfect cleavage parallel to prismatic habit, and comparably higher relief than phlogopite crystals (Fig. A1a, b). Irregular tabular to prismatic colourless calcite crystals vary in size from <0.1 mm up to c. 0.5 mm in width and exhibit amorphous (Fig. A1c) to rhombohedral cleavage planes (Fig. A1d), with multiple twinning and ‘twinkling’ through rotation under polarized light. Calcite in Rosemarkie also exhibits <1 mm clusters of c. 0.1 mm calcite and Mg-chlorite crystals surrounded by coarser (i.e. c. 0.5 mm) calcite crystals (Fig. A1d).

Twelve samples of graphite from marble in The Aird yielded carbon isotope compositions (⁠$δ$¹³C ‰) in the range −1.6 to 0.4‰ (Table 2; see also Table B1 in Appendix B). These values differ from a carbon isotope composition from graphite in Glen Urquhart (−3.4‰) (Table 2) and a carbonatite vein carbon isotope composition in Moniack Burn marble (c. −5.7‰). The majority of carbon isotopic compositions of Moniack Burn marble are within the range of −1.9 to 2.6‰ (Table B1). Eight samples of pyrite (Table 3) from The Aird yielded sulfur isotope $δ$³⁴S compositions of −16.6 to −15.5‰ CDT (n = 5) from calcite veins in Moniack Burn, a $δ$³⁴S composition of +7.6‰ from a single pyrite vein sample at Kirkton and a $δ$³⁴S composition of +0.7‰ (n = 2) from pyrite masses in marble at Blairnahenachrie. Pyrite veins in Rosemarkie (Learnie Quarry) psammite and pyrite hosted in coarse calcite at Glen Urquhart had $δ$³⁴S compositions of +7.3‰ (n = 2) and +6.1‰ (n = 2), respectively (Table 3).

Graphite micro-Raman spectroscopy (Fig. 5) measurements were analysed from several localities in The Aird (Moniack Burn, Kirkton and Blairnahenachrie) and also from Glen Urquhart. Standard parameters for comparing samples and assessing thermal maturation are recorded in Table 2. Raman spectroscopy of samples from The Aird all show a dominant graphite spectrum G-peak but additionally show a small disorder peak. The graphite from Glen Urquhart lacks a disorder peak and is fully graphitized. All graphite spectra additionally show an asymmetrical S1 peak (Fig. 5) in the second-order region. The asymmetry of the S1 band reflects the transition from two- to three-dimensional carbonaceous material that generally occurs at temperatures higher than 500°C (Wopenka and Pasteris 1993; Satish-Kumar 2005; Ferrari et al. 2006). The spectra for The Aird all yielded palaeotemperatures between 583 and 590°C, and higher than 600°C where pure graphite is present (Glen Urquhart) (Table 2).

Micro-Raman spectroscopic thermometry can be prone to error, including the resolution of the Raman thermometry technique used, which decreases as the limit of 600°C is approached. We cannot rule out edge effects or surface effects; however, fully ordered spectra outnumber disordered spectra. No polished surfaces were analysed, as these may be sources of error. Further, in order to confirm fully ordered graphite, analyses were performed on graphite extracted after HCl dissolution and measured on different points of the graphite surface in order to limit edge effects.

Despite very different environments of formation, carbonatites and marbles may appear similar and can be confused with each other, particularly in polymetamorphosed terranes (Le Bas et al. 2002; Wu 2008), but can be distinguished by isotope and trace element compositions. A carbon isotopic composition $δ$¹³C range of −1.6 to 0.4‰ (Table 2) in The Aird graphite is consistent with carbon isotopic compositions of calcite (⁠$δ$¹³C, −1.9 to 2.6‰) in The Aird marble (Fig. 6; see Table B1) and carbonate carbon isotope compositions typical of the Proterozoic (Shields and Veizer 2002), and broadly in line with marble-hosted graphite–phlogopite assemblages elsewhere in the world (Table 4). In contrast, the carbon isotopic composition of marble in Glen Urquhart is close to the $δ$¹³C range of mantle-derived carbonatites, with isotopic compositions of −3.4‰. However, we recognize the uncertainty in carbon isotopic compositions due to limited graphite and calcite sampling at Moniack Burn and other marbles in The Aird (Table B1), especially due to an absence of calcite-enclosed graphite and other carbon isotopic compositions (i.e. folded marble $δ$¹³C, 4.6‰).

The carbon isotopic compositional analysis technique used in this study did not distinguish between calcite-enclosed graphite, graphite nodules with phlogopite and graphite nodules with other micas or Mg-chlorite. The Aird graphite carbon isotopic compositions are outside the $δ$¹³C range of −10 to −5‰, for derivation of graphite from the mantle (Luque et al. 2012), or the $δ$¹³C range of −8 to −4‰ for carbonatite calcite, as reported by Broom-Fendley et al. (2017), or calcite vein (i.e. carbonatite) documented in this study with a tight range of carbon isotopic compositions (⁠$δ$¹³C, −5.7 to −5.6‰). Carbon isotope compositions of graphite and marble calcite are similar, although a disequilibrium in isotopic compositions between marble calcite (⁠$δ$¹³C, 1 to 4.6‰) and phlogopite nodule-derived graphite crystals (⁠$δ$¹³C, −1.6 to 0.4‰) is comparable to findings by Satish-Kumar et al. (2011). However, unlike Satish-Kumar et al. (2011), there is no indication of a second population of carbon isotopic compositions, which limits the interpretation of carbon isotopic compositions of graphite in this study.

The association between graphite and phlogopite at The Aird appears to be due to the preferential precipitation of graphite on/around phlogopite. Both are phases formed during metamorphism, derived from carbon and silicates in the precursor sediments. Preferential precipitation is indicated by:

1. Graphite occurs consistently with phlogopite, while there are also many phlogopite-bearing layers with no graphite.

2. In some nodular structures, a core rich in phlogopite is surrounded by several smaller crystals of graphite.

The composition of phlogopite can help to distinguish between a mantle or carbonate origin, specifically the F content (Table 1). The fluorine content of some phlogopite grains in Moniack Burn marbles is up to 2.2 wt%, which is typical of compatible behaviour in minerals of a magmatic origin (Dooley and Patino-Douce 1996; Edgar et al. 1996; Sun et al. 2022). However, most phlogopites in The Aird have compositions of <2 wt% F, which is not diagnostic, or especially indicative, of a magmatic origin (Sun et al. 2022). Elsewhere, isochemical fluorine reactions known to occur between fluorine-rich humite group minerals (i.e. norbergite, chondrodite, humite and clinohumite) and phyllosilicates such as phlogopite and amphiboles are known to occur in high-grade marbles, as in the granulite Ambasamudram marble (Satish-Kumar and Niimi 1998). Similar reactions between fluorine-rich humites and phlogopite in granulite-facies marbles could positively skew F contents in marble-derived phlogopite. Norbergite and chrondrodite F reactions with amphiboles and phlogopite minerals are also described in the Franklin marble (Kearns et al. 1980). However, F-rich humite group minerals may occur in the absence of phlogopite in marbles (Boureghda et al. 2023).

The lath-shaped to equidimensional graphite crystals identified in The Aird marble are most strongly associated with phlogopite (Fig. 3a, c, e and f), in addition to an association with accessory talc (Fig. 3a) or Mg-chlorite (Fig. 3b, c; see also Fig. A1a, b), or, less commonly, calcite (Fig. 3d). Phlogopite occurs as usually well-formed, prismatic crystals with perfect cleavage (Fig. A1a, b). Graphite in The Aird occurs as consistently tabular, lath-shaped to equidimensional crystals within lath-shaped, aggregate or subordinate nodular assemblages, usually with prismatic phlogopite. When compared to ‘well-formed’ tabular, hexagonal and spherical graphite crystals in the Franklin Marble, that comprise basal pinacoid faces typical of flake graphite (Jaszczak 1994, 1997), the graphite crystals in The Aird appear to have a flake-type morphology. The analytical methods in this study cannot constrain in sufficient detail whether graphite crystals in The Aird exhibit basal pinacoid faces that are typical of flake-type graphite in granulite-facies marbles (Jaszczak 1997; Satish-Kumar et al. 2011). A more extensive investigation of graphite crystallography in The Aird could add to graphite morphology characterization.

Graphite nodules have similar forms throughout The Aird, commonly as lath-shaped (Fig. 3a), aggregate (Fig. 3b) or subordinate (Fig. 3c) nodular forms, usually in preferential association with phlogopite. Well-formed, accessory minerals, Mg-chlorite (Fig. 3b), quartz (Fig. 3f) and calcite (Fig. 3b), are inliers within aggregate nodules or in graphite crystals, possibly indicative of a carbonate-silicate protolith. Aggregate forms of graphitic nodules rarely are replaced by muscovite (Fig. 3b). It is unclear if the muscovite is retrograde but the poorly formed crystal edges are suggestive of retrograde metamorphism. There is a similar association between graphite and phyllosilicate minerals elsewhere, most notably phlogopite in carbonates, many from high-temperature, granulite-facies ruby-hosted marbles (Yang et al. 2019), ruby/corundum-bearing marble (Dzikowski et al. 2014) and the sillimanite-bearing Franklin marble (Peck et al. 2006).

The preferential association of phlogopite with graphite implies that phlogopite provided the appropriate surface conditions for nucleation and growth of graphite crystals, regardless of the origin of the graphite. Graphite formation by carbonate reduction is discussed by Galvez et al. (2013), who described graphitization from a CO₂-rich fluid by fluid–rock C–O–H hydration reactions between quartz and calcite, generating graphite at temperatures as low as 430°C, and accessory products composed of Ca/Mg-silicates depending on the carbonate protolith composition. Graphite is also documented to precipitate in saturated closed-system C–O–H fluid inclusions (Cesare 1995), which can be correlated depending on the fluid temperature and fluid chemistry, the latter from the fractionation of H₂O which decreases the fluid density, as the binary H₂O–CO₂ or H₂O–CH₄ fluids cross the graphite saturation surface. Carbonates could alternatively have undergone devolatilization through CO₂ and H₂O volatile exchange between prograde carbonates (i.e. calcite and dolomite) and dehydrated silicate (e.g. micaceous) minerals in associated pelitic rocks, depending on the pressure and temperature gradients, which can drive calcite dissolution, CO₂ fluid release and degassing (Ague 2000; Ague and Nicolescu 2014).

Graphitization associated with carbonate–pelite devolatilization has previously been documented in the mixing of high oxygen fugacity, carbonate-derived CO₂ and low oxygen fugacity metasedimentary-derived CH₄ aqueous fluids, under mid-to high-grade metamorphic facies (e.g. Evans et al. 2002; Satish-Kumar 2005), sometimes mineralizing graphite from highly mobile, fracture-bound, carbon-saturated hydrothermal fluids (Rumble et al. 1986; Luque et al. 2012). Devolatilization mixing of carbonate-derived CO₂ and metasedimentary-derived CH₄ fluids would be likely to deposit graphite that is isotopically lighter than decarbonated carbonate or mantle-derived graphite (Luque et al. 2012). Graphite in The Aird is assumed to have formed from decarbonation reactions in an impure carbonate–silicate protolith, forming extensive assemblages of graphite–phlogopite nodules and minor graphite associations with accessory minerals, Mg-chlorite and talc. Devolatilization-associated graphitization in The Aird is unlikely, due to possibly limited volatile exchange with carbon-poor Loch Ness Supergroup pelitic rocks; a lack of fracture-bound, isotopically light graphite; and graphite-associated accessory mineralogies that are consistent with decarbonation reactions between a calcite and quartz-enriched protolith (e.g. Galvez et al. 2013). However, the fluid–rock characteristics that influence the CO₂–phlogopite preferential association remain underexplored in decarbonated marbles, especially in the absence of reduced organic carbon or mantle-derived fluids (e.g. Satish-Kumar et al. 2011; Galvez et al. 2013).

The original c. 870 Ma limestone would have undergone two or three orogenic episodes prior to the Great Glen Fault-associated deformation (i.e. post-Scandian) (Strachan et al. 2010; Mako et al. 2021; Krabbendam et al. 2022). It is also plausible that isotopic compositions recorded in graphite and sulfide minerals may have been reset by post-Grampian I/II dissolution and reprecipitation. Pressure and temperature constraints during the Knoydartian Orogeny for Lower psammite, Morar pelite and Upper psammite rocks in the Morar Group, in the Wester Ross Supergroup (Krabbendam et al. 2022), are suggested by Cutts et al. (2009) to have reached c. 7.5 kbar and 650°C, respectively. This was prior to or during garnet growth, associated with an early extensional regime, followed by crustal thickening. Monazite–xenotime thermometry of the Loch Eil Group and Glenfinnan Group rocks in the Naver Nappe are recognized by Mako et al. (2019) to have experienced decompression heating from 8–9 kbar and 600°C to 6–7 kbar and 700°C, with peak metamorphism occurring at c. 425 Ma. Different temperature constraints occur in monazite and xenotime elsewhere in the Sgurr Beag Nappe (Mako et al. 2021), which could have had the effect of forming two or more distinct trace element compositions and geochronologies.

The presence of well-formed Mg-chlorite and talc as accessory associated minerals, rather than as later replacive assemblages, is suggestive of decarbonation reactions between protolith carbonate and silicate minerals. The origin of graphite is also constrained by the low graphite contents in surrounding Glenfinnan Group metasedimentary rocks. However, the pelitic rocks cannot be ruled out as an external source of CO₂ fluids. Graphitization in The Aird is likely to have formed from a decarbonated CO₂-rich fluid, given the tight carbon isotopic range of graphite minerals and the apparent absence of carbon isotopic zoning. Graphitization possibly occurred during a single phase, although this may not correlate to peak metamorphism, given a noticeable disequilibrium in graphitic carbon isotopic compositions compared to marble calcite isotopic compositions.

The Raman spectra indicate almost complete ordering of the graphite at The Aird. The slight disorder, although not observed in the graphite from Glen Urquhart, is not sufficiently marked to be certain that they have experienced different degrees of metamorphism. Peak metamorphic temperatures in The Aird of c. 580–600°C, inferred from Raman spectroscopy, is an appropriate temperature range for amphibolite-facies metamorphism (Luque et al. 2012). However, there is a decrease in temperature resolution as the limit of 600°C is approached. Raman spectroscopy geothermometry measurements taken from The Aird marble and Loch Ness Supergroup pelitic rocks suggests the rocks underwent mid–upper amphibolite-facies metamorphism during the Grampian Orogeny, with predicted similar peak temperatures of c. 600°C (Strachan et al. 2010; Chew and Strachan 2014). The main assumption of Raman geothermometry in this study is that the ordering of graphite is a function of metamorphic temperatures, and that disorder in graphite corresponds to temperatures generally lower than 650°C. Our data indicate that after graphite precipitation in The Aird, the highest metamorphic temperatures were recorded in the Glen Urquart graphite. No interpretation is possible of temperatures above the limit of full graphitization, nor can we precisely assess the metamorphic degree. However, we can give some information about different ranks reached after graphite precipitation.

Homogenization temperatures established through two-phase (liquid and vapour) aqueous fluid inclusions in breccia-hosted quartz crystals suggest mean homogenization temperatures of 122.5°C (c. 150°C after pressure correction), far below the temperature of graphite indicated by Raman spectroscopy. No positive identification of carbon-derived fluids is indicated by aqueous fluid inclusions hosted in breccia-hosted quartz crystals, which together with significantly cooler thermometry, below those measured by Galvez et al. (2013), suggests that fluid inclusions are likely to be unrelated to regional metamorphism-derived graphitization in The Aird. Fluid-inclusion thermometry suggests that brecciation is likely to have occurred during post-Scandian shearing.

The sulfur isotope data indicate several populations of pyrite. The vein-hosted pyrite consists of a light population from Moniack Burn (⁠$δ$³⁴S, −16.6 to −15.5‰) and a moderately heavy (⁠$δ$³⁴S, 6–8‰) population evident at Kirkton, and also Glen Urquhart and Rosemarkie (Table 3). The diagenetic pyrite in Blairnahenachrie carbonate has a near-zero composition compared to the vein pyrite. The moderately heavy pyrite sampled from Kirkton, Glen Urquhart and Rosemarkie is comparable to pyrite mineralized in the Caledonian Strontian Granite 100 km to the SW, also along the Great Glen Fault (Laouar et al. 1990). The heavier measurements here are also similar to values from pyrite in the ‘Moine’ (Lowry et al. 2005), which may suggest that pyrite veining along the Great Glen Fault was mobilized from elsewhere in the Loch Ness Supergroup.

The vein pyrite from Moniack Burn has a composition that is light enough to suggest biogenic processing of sulfur, which may have remobilized into the veins from a layered source. A succession in the close vicinity of The Aird with evidence of biogenic pyrite is the c. 400 Ma Lower Old Red Sandstone at Strathpeffer (Parnell 1985). Neither vein set at Moniack Burn and Kirkton matches the composition of the mantle-derived pyrite at Loch Urigill in the Loch Borralan Alkaline Complex, which is within the $δ$³⁴S −6 to +3‰ CDT range of most carbonatite-hosted pyrite (Farrell et al. 2010; Gomide et al. 2013). The sulfide vein fluids at The Aird are therefore unlikely to be mantle-derived, and are more likely to be associated with shallow (crustal) fluid–rock interactions with Glenfinnan metasedimentary rocks. Graphite entrained in pyrite veins is also likely to have been remobilized from the marble by sulfide fluids, given the similar graphite grain morphologies in both the pyrite vein and the marble host rock.

Graphitization is assumed to have occurred in a single episode of metamorphism during the Grampian Orogeny, and is likely to have occurred at a similar time to ductile deformation, at peak metamorphism temperatures of c. 580–600°C that are typical of mid–upper amphibolite-facies metamorphism (Luque et al. 2012). Graphitization associated with the Scandian Orogeny appears to have insufficient evidence of ductile shearing that was sufficient for greenschist–amphibolite-facies metamorphism, given the petrographical evidence of brittle deformation features and aqueous fluids identified in quartz grains within hydraulic breccia that do not contain any CO₂. So it is not possible to confirm directly that graphitization occurred from a CO₂-rich fluid from an external source. The graphite carbon isotopic composition strongly supports a sedimentary carbonate source, as discussed above, as there is insufficient mineralogical and carbon isotopic compositional evidence of graphitization associated with mantle-derived carbonatite magmatism.

The pyrite in the matrix of the marble consists of coalesced crystals, interpreted as having formed during burial diagenesis, and which are unrelated to later sulfide veining. The graphite and phlogopite were products of subsequent metamorphism. Graphitization-associated metamorphism of the Loch Ness Supergroup, especially in the Glenfinnan Group, as in The Aird, may have occurred in the Grampian I event between c. 470 and 460 Ma (Strachan et al. 2010; Mako et al. 2021). However, overprinting of Knoydartian fabric (Cutts et al. 2010) or evidence of Grampian II orogenesis (Bird et al. 2013; Milne et al. 2023) is limited. Future Re–Os geochronology investigations of pyrite in The Aird may further constrain pre-Caledonian and Grampian I/II orogenic reworking of The Aird marble.

Fluid inclusions within breccia-hosted quartz crystals and folded hematite veins in The Aird marbles imply that the rocks in The Aird underwent post-Scandian hydraulic brecciation and hematite deposition transitional to the alkaline and carbonatitic mineralization reported in Heptinstall et al. (2023). Fluid inclusions in authigenic quartz within breccia suggest a temperature for brittle deformation of about 150°C (Table 5). Brecciation and hematite veining in The Aird may correlate with the partitioning of transpressive strain and sinistral displacements described along the Great Glen Fault that post-dates Scandian ductile thrusting (Strachan et al. 2010, 2020a, b). The timing of Scandian to post-Scandian sinistral motion on the Great Glen Fault is constrained to have begun as early as c. 428 Ma, marked by emplacement of the Clunes tonalite (Stewart et al. 2001), and continued until at least c. 399 Ma, marked by syndeformational emplacement of granite veins at Rosemarkie (Mendum and Noble 2010). Recent U–Pb zircon dating of the Abriachan Granite by MacRae et al. (2023) argues for emplacement at 420.1 ± 3.4 Ma. Consequently, shearing, calcite veining, pyrite veining and especially fenitization recorded in the Abriachan Granite (Garson et al. 1984) and The Aird (Heptinstall et al. 2023) may have occurred between c. 420 and c. 399 Ma. While there is some suggestion of earlier hematite veining, based on petrographical relationships, the majority is post-Devonian.

Hematite veining is found through the region west of the Great Glen Fault at Loch Ness, and in Badenoch Group metasediments and Devonian sediments, although these examples are unfolded (Elmore et al. 2006). Palaeomagnetic studies of the Devonian sediments immediately west of Loch Ness in the Mealfaurvonie Outlier indicate that the hematite is of Carboniferous–Permian age (Elmore et al. 2006). Widespread reddening by hematite from above at this time reflects uplift and basin inversion related to the Hercynian Orogeny (Dichiarante et al. 2020). Recent radiometric dating (e.g. Mendum and Noble 2010; MacRae et al. 2023), carbonatitic-related mineralization (Garson et al. 1984; Heptinstall et al. 2023) and petrographical findings of post-orogenic modification by brittle deformation in this study do not all correlate to pre-existing Great Glen Fault kinematics (e.g. Stewart et al. 2001; Strachan et al. 2010, 2020a). We propose that the findings in this study may be further constrained through radiometric dating and by assigning brecciation, fenitization, veining of crustal fluids and shearing textures to sinistral or dextral transpressional or transtensional deformation regimes of the Great Glen Shear Zone, and stratigraphic correlation to sinistral or dextral movement of the Great Glen Fault.

Given these constraints, the paragenesis can be summarized from the rock descriptions (Table 5) as:

1. diagenesis, including pyritization;

2. metamorphism in Grampian I, including precipitation of graphite and phlogopite; however the phlogopite surface conditions attributed to preferential precipitation of graphite from CO₂ fluids remain underexplored in decarbonated marbles;

3. late Scandian hydraulic brecciation of rocks with deposition of hematite (after 428 Ma), transitional to;

4. fenitization (e.g. Heptinstall et al. 2023) accompanied by calcite veining, pyrite veining and graphite remobilization from c. 420 to 399 Ma;

5. local shearing and brittle faulting associated with the Great Glen Fault in the Loch Ness region, from c. 399 Ma; and

6. further hematite mineralization, possibly coeval with Carboniferous–Permian hematite mineralization in the Mealfaurvonie Outlier.

Previous observations have linked the decarbonation of protolith carbonate during amphibolite-facies regional metamorphism to the Grampian event (Rock et al. 1986b; Cutts et al. 2010; Strachan et al. 2010; Goodenough et al. 2011). Descriptions of Loch Ness Supergroup graphitic marbles in The Aird, and marbles elsewhere in the Northern Highlands Terrane including the Lewisian-age Scardroy Inlier and distinct Neoproterozoic Glen Urquhart marble, indicate that graphite is more widespread in northern Scotland than previously identified. Graphite crystals and their nodular assemblages occur in several forms, including graphite–phlogopite horizons up to 0.5 cm in width. In these horizons, graphite nodules are usually associated with phlogopite, talc or Mg-chlorite.

The rocks west of Loch Ness exhibit tight folding, brecciation and shearing, orientated particularly parallel to the fault. The Devonian sediments are also deformed by the tight folding (Mykura and Owens 1983). However, the tight folding in The Aird marble is much more ductile than in the Devonian, suggesting ductile deformation occurred prior to the Devonian, when there was further brecciation and shearing (Stewart et al. 1999; Mendum and Noble 2010). The deformation along the Great Glen Fault represents repeated episodes of strike-slip movement and brittle faulting, and there is no requirement to conflate different folding events (Stewart et al. 1999; Strachan et al. 2010).

Ductile deformation occurred at a similar time to peak metamorphism and graphitization between 475 and 460 Ma (Cutts et al. 2010; Mendum and Noble 2010). This is supported by dating of a pegmatite in The Aird at 463 ± 4 Ma by Cutts et al. (2010) and a monazite age of 473 ± 2 Ma in Glen Urquhart (Mako et al. 2021). However, we cannot rule out that ductile deformation documented in The Aird is associated with later (Scandian) movement of the Moine Thrust, at c. 430 Ma (Goodenough et al. 2011). The period of brecciation, fenitization, calcite veining and pyritization with graphite remobilization also occurred during the Caledonian Orogeny but before Devonian sedimentation (Garson et al. 1984; Heptinstall et al. 2023).

The graphite–phlogopite association is recorded in Proterozoic marble at The Aird, in the Northern Highland Terrane, Scotland. The occurrence supports previous suggestions that graphite and phlogopite show a preferred relationship. We have shown:

1. The carbon at The Aird is derived from decarbonation of protolith carbonate, with graphitization derived from a CO₂-rich fluid that had similar carbon isotopic compositions to the host marble, and the isotopic composition range of sedimentary carbonates in the Proterozoic.

2. Graphite and phlogopite formed during amphibolite metamorphism of Proterozoic limestone at temperatures of nearly 600°C, at or prior to c. 470 Ma (c. Grampian I). Well-formed minerals Mg-chlorite and talc are also likely to have been mineralized coeval to amphibolite-facies metamorphism and graphitization, given the few retrograde textures (i.e. replacive muscovite). Graphites in The Aird are unlikely to be synchronous with breccia-hosted, quartz-derived fluid inclusions of indeterminate composition that suggest fluid temperatures of c. 150°C.

3. Three distinct sulfide populations are present in The Aird marbles. Two sulfur isotopic compositions are attributed to veins that post-date graphitization: a light population suggestive of a biogenic source and which is likely to have been derived from crustal fluid–rock interactions, and a heavy population suggestive of a mantle origin. Near-zero sulfur isotopes are likely to have been derived from a diagenetic source in the marble protolith.

4. The marbles of The Aird underwent extensive post-Scandian hydraulic brecciation and hematite deposition after c. 428 Ma: that is, transitional to calcite veining, pyrite veining and fenitization by mantle-derived alkaline and carbonatitic fluids, between c. 420 and c. 399 Ma. The graphite in the marble predates cross-cutting veins of calcite attributed to carbonatite-related fluids following the Scandian Orogen. Any graphite in calcite or sulfide veins is probably incorporated from the country rocks.

5. Subsequent uplift, shearing (i.e. folded hematite veins), late brittle faulting and undeformed brecciation is likely to be associated with the Great Glen Fault, from c. 399 Ma, as described in Glenfinnan Group rocks along the Great Glen Shear Zone. Further hematite deposition by <100°C oxidizing fluids occurred later, possibly during the late Carboniferous–early Permian.

We are grateful to J. Still, J. Johnston and J. Bowie for skilled technical support. We are grateful for previous reviews by Eimear Deady and an anonymous reviewer during an earlier proof, and to the reviewers of this paper for their constructive suggestions.

EAH: conceptualization (equal), data curation (equal), formal analysis (lead), investigation (lead), methodology (equal), project administration (equal), software (equal), validation (lead), visualization (equal), writing – original draft (lead), writing – review & editing (equal); JP: conceptualization (equal), funding acquisition (lead), methodology (equal), project administration (equal), resources (equal), supervision (lead), visualization (equal), writing – original draft (supporting), writing – review & editing (equal); AS: data curation (equal), investigation (supporting), methodology (equal), software (equal); AJB: data curation (equal), formal analysis (supporting); JGTA: data curation (equal), formal analysis (supporting), methodology (equal); DM: formal analysis (supporting).

This work was funded by the Natural Environmental Research Council (grant No. NE/M010953/1 awarded, in partial support, to J.P.). The isotope facility at SUERC is supported by NERC.

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.

All data generated or analysed during this study are included in this published article (and, if present, its supplementary information files).

Thin-section photographs of the marble are shown in Figure A1.

Graphite isotopic compositions for samples taken from Moniack Burn are shown in Table B1.