Origin of Siderite and Baryte in a Carbonate-Replacement Ag-Pb-Zn-Cu Sulphide Deposit: Walton, Nova Scotia, Canada

Abstract

Siderite and baryte are common non-sulphide phases in sedimentary exhalative (SEDEX) deposits, but their formation remains poorly understood. Siderite is important as an exploration vector in some deposits, whereas baryte is important as a S source in some deposits. The past-producing Walton deposit (Nova Scotia, Canada) consists of two ore types: (1) a sulphide body primarily hosted by sideritised Viséan Macumber Formation limestone (0.41 Mt; head grade of 350 g/t Ag, 4.28% Pb, 1.29% Zn, and 0.52% Cu), and (2) an overlying massive baryte body of predominantly microcrystalline baryte (4.5 Mt of >90% baryte). This study used optical microscopy, SEM-EDS, cathodoluminescence (CL), LA-ICP-MS, and SIMS sulphur isotope analysis of siderite and baryte to elucidate their origin and role in deposit formation. Siderite replaces limestone and contains ≤9 wt. % Mn, is LREE-depleted (PAAS-normalised REEY diagrams), and has low (<20) Y/Ho ratios. Sideritisation occurred due to dissimilatory iron reduction (DIR) that led to the breakdown of Fe-Mn-oxyhydroxides and organic matter, as indicated by light δ¹³C_VPBD values and negative Y anomalies. The baryte body is dominated by a microcrystalline variety that locally develops a radial texture and coarsens to a tabular variety; it also occurs intergrown with, and as veins in, massive sulphides. Based on fluid inclusion data from previous studies, the coarser baryte types grew from a hot (>200 °C) saline (25 wt. % NaCl) fluid containing CO₂-CH₄ and liquid petroleum. Marine sulphate δ³⁴S_VCDT values typical of the Viséan (~15‰) characterise the baryte body and some tabular baryte types, whereas heavier (~20‰) and lighter (~10‰) values typify the remaining tabular types. The variations in tabular baryte relate to distinct zones identified by CL imaging and are attributed to the sulphate-driven anaerobic oxidation of methane (SDAOM) and oxidation of excess H₂S after sulphide precipitation. These results highlight the importance of hydrocarbons (methane and organic matter) in the formation of both the siderite and the baryte at Walton and that DIR and the SDAOM can be important contributing processes in the formation of SEDEX deposits.

1. Introduction

Sedimentary-rock-hosted sulphide deposits commonly contain significant amounts of post-depositional carbonate minerals related to diagenesis or the later influx of hydrothermal fluids. Calcite and dolomite are most common, but siderite is more rarely also a major component in some sedimentary exhalative (SEDEX) or clastic dominated (CD) deposits (e.g., [1]). Siderite is typically assumed to be hydrothermal (e.g., [2]), although some studies argue for precipitation due to fluid mixing of hydrothermal and diagenetic fluids [3]. The controversial origin of siderite in these deposits is important to resolve because it is vital in models of how such deposits form (e.g., [4,5,6]), given that siderite is typically associated with ore-stage sulphides.

Rare earth element + Y (REEY) concentrations in carbonates are useful for constraining the chemical conditions attending their primary formation and during later processes (e.g., [7]). The ability to measure REEY in carbonates using in situ methods such as laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) has revealed the presence of microscale variations in these phases (e.g., [8,9,10,11]) and is used to understand the origin of regional carbonates suspected to be related to hydrothermal systems (e.g., [3,12,13,14,15]). Because REEY patterns and Y/Ho values can differentiate diagenetic and hydrothermal carbonates, they have important implications [16].

Baryte is common in sedimentary-rock-hosted Zn-Pb sulphide deposits, with SEDEX or CD deposits typically containing the most [4]. Naturally occurring baryte can be classified as marine, diagenetic, or hydrothermal [17], with each group’s specific isotopic signature (δ³⁴S, δ¹⁸O) reflecting the origin of the causative fluids [18,19]. The trace element chemistry of baryte has also been used to address its formation (e.g., [20,21,22]). Cathodoluminescence (CL) imaging allows for the identification of textures and features not visible using petrography alone (e.g., [23,24]) and for subsequent targeting of in situ isotopic and trace element analyses. However, CL imaging integrated with in situ isotope analysis of baryte are seemingly underused in the literature, especially together.

This study focusses on the past-producing Walton deposit (baryte ore body: 4.5 Mt of >90% baryte; sulphide ore body: 0.41 Mt with head grade values of 0.52% Cu, 4.28% Pb, 1.29% Zn, and 350 g/t Ag) [25,26] in Nova Scotia (Canada). This region hosts other temporally related ore bodies similar to Mississippi Valley-type (MVT) deposits as recently reviewed by Conliffe et al. [27]. Siderite forms an important pre-ore stage to both baryte and sulphides at Walton, and previous studies considered it as either formed during diagenesis in a shallow burial environment [28] or as the initial stage of the hydrothermal mineralising process [29]. The general features of the baryte ore body were noted by previous authors including Tenny [30] and Boyle [31], with its relationship to the sulphide mineralisation described by Boyle [32] and Sangster et al. [33].

At Walton, the presence of a baryte body overlying the sulphide Zn-Pb ore body is comparable with a number of SEDEX deposits (e.g., the Anarraaq deposit, [34]), and the extensive siderite associated with the sulphide ore body (>10 Mt, e.g., [28]) is also a feature more common to SEDEX deposits than MVT deposits [4,35]. However, the anomalous historical Ag grades (>150 g/t) at Walton are atypical of both SEDEX and MVT deposits. Although Walton has recently been referred to as a SEDEX deposit [36], previous studies have not evaluated it as such (e.g., [29,37,38]). Of importance to this study is the origin of both the siderite and baryte ore bodies and how they may be interrelated, how they may contribute to understanding the sulphide mineralisation, and the possible role of hydrocarbons in the formation of siderite and baryte.

To address the nature and origin of the siderite and baryte at Walton and how they relate to the origin of and implications for the sulphide mineralisation, this study re-examines the mineral paragenesis and related textures of the non-sulphide phases. This re-examination forms the basis of a geochemical study using in situ microanalytical techniques [trace element geochemistry (LA-ICP-MS); sulphur isotope systematics (SIMS)] to augment the understanding of the various carbonate and baryte phases. These data are then used to re-evaluate the origin of these phases at the Walton deposit and can also contribute to the understanding of other similar deposits.

2. Geological Background

2.1. Regional Geology

The Walton deposit is located in the Kennetcook sub-basin of southern Nova Scotia, which overlies siliciclastic basement rocks of the Meguma terrane, the most outboard terrane of the northern Appalachian orogenic collage (Figure 1a,b). This terrane was emplaced against the adjacent Avalon terrane coincident with the closure of the Rheic Ocean around the Late Devonian, with the suture delineated by the E-W trending Cobequid-Chedabucto Fault Zone (CCFZ; Figure 1a) [39]. The Meguma terrane is dominated by metaturbiditic rocks of the Ediacaran-Ordovician Meguma Supergroup [40] and late Devonian meta- to peraluminous granites (ca. 380 Ma; [41]), with lesser amounts of Silurian to Triassic strata.

The Ediacaran-Ordovician Meguma Supergroup consists of a sandstone-dominant unit overlain by a slate-rich unit [40]. Unconformably overlying this are Silurian-Devonian metavolcanic and metasedimentary rocks (i.e., Rockville Notch Group; not shown in Figure 1) [45]. Regional Neoacadian deformation and lower greenschist facies metamorphism of these rocks were followed by the deposition of the late Devonian–early Carboniferous (Tournaisian) Horton Group, which predominantly consists of terrestrial sandstone, shale that is locally sulphidic, and conglomerate (Figure 1c) [46,47]. The Horton Group has an angular unconformity, nonconformity, or fault contact with underlying rocks and is disconformably to unconformably overlain by the Windsor Group [48]. The Horton Group thickens considerably offshore in the large Maritimes Basin depocentre [49] and importantly includes widespread bimodal volcanic rocks at its base (see discussion in [50]).

The Mississippian (Viséan) Windsor Group, which overlies the Horton Group, consists of several carbonate and evaporitic units and commences with a transgressive marine unit, the Macumber Formation (Figure 1d) [51,52]. This formation consists of laminated, bituminous limestone that locally passes laterally to carbonate bioherms on basement paleotopographic highs (Gays River Formation; [46,53]). Both facies host Zn-Pb mineralisation in southern Nova Scotia—the Scotia Mine (formerly Gays River) deposit in dolomitised mound facies and the Walton deposit in sideritic laminated facies (e.g., [54]). The strata overlying the Macumber Formation in the Kennetcook sub-basin include the White Quarry (gypsum and anhydrite) and Stewiacke (halite) formations. The Tennycape Formation overlies the Stewiacke Formation and consists of various colours of sandstone, siltstone, and shale [32].

Other sedimentary-rock-hosted Zn-Pb sulphide deposits near the Windsor Group-Horton Group contact include the Jubilee (Zn-Pb) and Smithfield (Zn-Pb) deposits [27,54,55]. The Gays River and Macumber formations are coeval with the Big Cove Formation in Newfoundland, which also contains MVT-style mineralisation (Figure 1; [27,56,57]).

Overlying the Windsor Group is the late Viséan to Namurian Mabou Group, which contains grey to red mudrocks and small amounts of lacustrine evaporites [43]. The Pennsylvanian Scotch Village Formation of the Cumberland Group unconformably overlies the Windsor Group and consists of sandstone, grey shale, and coal and wood fragments [43]. Waldron et al. [58] noted evidence for the deformation of the Horton, Windsor, and Mabou groups and interpreted the deformation to predate the deposition of the Scotch Village Formation. At Walton, the Horton Group is thrust over the Macumber Formation [43].

2.2. Walton Geology and Mineralisation

The Horton Group contains non-marine red, grey, and black clastic rocks of the Horton Bluff and Cheverie formations [59,60]. In the Walton area, the basal Horton Bluff Formation consists of fluvial and lacustrine shale to conglomerate, whereas the overlying Cheverie Formation consists of quartzitic sandstone, siltstone, and shale [32,33]. Sandstone matrices contain clays, micas, feldspars, and siderite. The Horton Bluff Formation is locally sulphidic and organic-rich (up to 5 wt. % C; e.g., [32]). In the Kennetcook sub-basin, the Horton Group was deformed [44,58] prior to the intrusion of a 315 Ma mafic dyke [61] that truncates planar deformation fabrics in the Horton Bluff Formation.

In the Walton area, the Macumber Formation is laminated, averages 8 m (range of 0–25 m; [29]), and is subdivided into a lower and upper unit [62,63]. The lower unit, interpreted to have formed below the fairweather wave-base, consists of fine- and coarse-grained limestones, whereas the upper unit consists of planar-bedded lime mudstone interbedded with deep-marine sulphates [53]. The Macumber Formation limestone contains an estimated 5% primary porosity [64]. The Pembroke Formation was originally used informally to describe breccias overlying the Macumber Formation, but at the Walton area, it has been used to describe a different breccia lower in the stratigraphy, and with different characteristics [63]. To be consistent with the usage of Lavoie et al. [63] and other studies, only the carbonates corresponding to the breccias at the top of the Macumber Formation will be referred to as the Pembroke breccia here.

As noted above, the Walton deposit is a composite ore deposit with a sulphide ore body (0.41 Mt; head grade values of 0.52% Cu, 4.28% Pb, 1.29% Zn, and 350 g/t Ag) grading into and overlain by a baryte ore body (4.5 Mt of >90% baryte) (Figure 2) [25,26]. The baryte ore body crosscuts massive siderostone, and siderite locally makes up 10% or more of the ore body [32]. The sulphide ore body contains pyrite-marcasite, sphalerite, galena, tennantite, chalcopyrite, and minor (para)rammelsbergite with both the sulphide and sulphate ore bodies containing manganiferous siderite [33]. Aspects of the sulphides, such as their texture, chemistry, and relative abundances, have been described previously [29,32,37]. A comprehensive textural and mineral-chemical study that describes in detail the colloform nature of the ore and textural evidence for zone refining will be provided in a future publication. Additional sulphide mineralisation in the underlying Horton Group sandstone consists of pyrite, chalcopyrite, and tennantite. Sangster et al. [33] summarised the metal distribution in the sulphide ore body and noted that it was zoned with respect to Cu/(Cu+Pb+Zn) (i.e., Cu-rich at the base), and also reported an absence of sphalerite and galena in the Horton Group.

2.3. Nature of Siderite and Baryte from Previous Studies

Burtt [29] noted that although surface samples of the Macumber Formation at the deposit are limestone, samples from the uppermost underground drill holes are siderite with local ankerite (Figure 3). Samples from deeper drill holes are solely siderite, and the nature and position of the transition from limestone to siderostone is unknown [29]. Savard et al. [28] interpreted the siderostone as forming from the replacement of limestone because of features such as layered ooids, brachiopod fragments, and a microbreccia (due to slumping) still being discernible. There is estimated to have been at least 10 Mt of siderite prior to replacement by baryte and sulphides and post-ore karstification [29]. The exact extent of the siderite in the area is unknown, but it may have been more than double that figure [28].

Descriptions of the overlying baryte ore body rely mostly on the observations of several previous workers [30,31,37]. The baryte ore body is now present as a NE-plunging lens that is discordant to the rocks it crosscuts [26]. It is described as predominantly fine-grained with a microcrystalline texture (Figure 3e–g). Other baryte types documented by Boyle and Jambor [37] are (1) highly intergrown flamboyant and radiating fibrous aggregates (Figure 3h); (2) intergrown spherulitic aggregates; (3) intergrown tabular aggregates; and (4) tabular euhedral crystals. These coarse varieties are rare in the baryte ore body [31]. Local bitumen is also present in the baryte ore body (Figure 3f). The origin of the baryte ore body has traditionally been attributed to the replacement of either the Macumber Formation or the overlying evaporites [32].

3. Previous Isotopic and Fluid Inclusion Studies

Previous isotopic studies at Walton presented bulk analyses of δ³⁴S for sulphates (baryte and anhydrite) and sulphides [65], δ¹⁸O and δ¹³C for carbonate cements [28,66], and δ¹⁸O, δ¹³C, and Sr for siderite [28]. In general, these data were interpreted to suggest the following: (1) the replacement of gypsum and anhydrite by baryte; (2) the thermochemical sulphate reduction of sulphate carried by heated saline brines; and (3) the ¹⁸O enrichment of the fluids that produced carbonates distinct from such phases precipitated from contemporaneous (Viséan) seawater.

There is one extensive fluid inclusion study of Walton [38] that followed a more limited investigation [67]. These data, mostly for the later stage tabular baryte versus the early finely crystalline baryte (see below), indicated homogenisation temperatures for aqueous and liquid petroleum inclusions ranging from 100 to 300 °C, with the dominant fluid salinity (for the former) of 20–25 wt. % eq. NaCl with a substantial CaCl₂ component. Importantly, a few measurements in sphalerite [38] also indicated similar high temperatures (220 °C; n = 3) and high salinities. These data are similar to those for the temporally related MVT settings in the Maritimes Basin (summary in [27]).

4. Materials and Methods

4.1. Sampling and Petrography

Samples primarily from the 850 L mining level and deeper (Figure 2b) were collected from the drill core housed at the Stellarton Core Library (Stellarton, NS, Canada). Additionally, a small number of dump samples (surface samples of previously mined material) and outcrop samples from Walton were collected (Figure 2a and Figure 3). Archived samples (drill core, outcrop, and dump samples) from the study of Kontak and Sangster [38] were also used. The dump samples consist of massive baryte inferred to be mostly from the baryte ore body and massive sulphides (mostly pyrite and galena). It is important to note that, for this study, none of the samples contained any of the various Ag sulphides described in previous studies. Unaltered and unmineralised Macumber Formation limestone samples from Cheverie (~30 km from the Walton deposit; Figure 1b,d) were also sampled.

Transmitted and reflected light petrography conducted on ~60 polished thin sections used an Olympus BX-51 petrographic microscope at Laurentian University (Sudbury, ON, Canada) to determine paragenesis, document textures of the carbonates and barytes, and to locate samples suitable for subsequent microanalytical work. A subset of thin sections was also examined under ultraviolet (UV) light.

4.2. Scanning Electron Microscopy (SEM) and Energy Dispersive Spectroscopy (EDS)

Following petrography, a suite of carbonate (calcite, dolomite, ankerite, and siderite), sulphide (pyrite, chalcopyrite, sphalerite, and tennantite), and baryte phases were characterised using backscattered electron (BSE) imaging, energy dispersive spectroscopy (EDS) spot analysis, and cathodoluminescence (CL) imaging. Selected polished thin sections were carbon-coated and analysed using a Tescan Vega 3 scanning electron microscope (SEM), with CL and BSE imaging capabilities, coupled to a Bruker EDS detector and software at Laurentian University (Sudbury, ON, Canada). The analyses were used to confirm paragenesis, identify carbonate phases, and determine compositional and textural differences in the mineral phases. The operating conditions were an accelerating voltage of 10–15 kV, beam absorbed current of 1.0 nA–400 nA, varying beam size <5 µm, and counting rate typically >40,000 counts per second. Lower accelerating voltage and beam current were used during CL imaging to improve the quality of the images.

4.3. Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS)

Trace and rare-earth element (REE) analyses of carbonates and baryte were carried out in the same polished thin sections used for petrography and SEM-EDS. Data were acquired at the Mineral Exploration Research Centre–Isotope Geochemistry Laboratory (MERC-IGL), Laurentian University (Sudbury, ON, Canada), using a Photon Machines Analyte G2 ArF excimer laser, with a 193 nm wavelength and two-volume HelEx II cell coupled to a Thermo Scientific iCap triple quadrupole ICP-MS. The external standards used were NIST 610 and NIST 612 glasses [68], GSE1 and GSD-1G glasses [69], and an FeS1 compressed puck. Standards were analysed after every ~10 spots and at the beginning and end of each run.

The internal standards were dependent on the mineral analysed (e.g., Ca standard for calcite). Beam sizes of 15 µm (carbonates) and 50 µm (carbonates and baryte) were used, with the larger spot sizes used to enhance the signal of the REEs. The operating conditions for spot analyses of carbonates were a laser pulse frequency of 7 Hz, fluence of 3 J/cm², 525 mL/min He (cup), 100 mL/min He (cell), 625 mL/min Ar, and 6 mL/min N₂. For baryte, lower fluences of 1.5 and 2 J/cm² were used with all other conditions the same as with the carbonates.

The elements measured were ²⁴Mg, ²⁷Al, ²⁹Si, ³⁴S, ³⁹K, ⁴²Ca, ⁴⁵Sc, ⁴⁷Ti, ⁵¹V, ⁵³Cr, ⁵⁵Mn, ⁵⁷Fe, ⁵⁹Co, ⁶⁰Ni, ⁶³Cu, ⁶⁶Zn, ⁷¹Ga, ⁷³Ge, ⁷⁵As, ⁷⁷Se, ⁸⁵Rb, ⁸⁸Sr, ⁸⁹Y, ⁹⁰Zr, ⁹⁵Mo, ¹⁰⁷Ag, ¹¹¹Cd, ¹¹⁵In, ¹¹⁸Sn, ¹²¹Sb, ¹²⁵Te, ¹³³Cs, ¹³⁵Ba, ¹³⁷Ba, ¹³⁹La, ¹⁴⁰Ce, ¹⁴¹Pr, ¹⁴⁶Nd, ¹⁴⁷Sm, ¹⁵³Eu, ¹⁵⁷Gd, ¹⁵⁹Tb, ¹⁶³Dy, ¹⁶⁵Ho, ¹⁶⁶Er, ¹⁶⁹Tm, ¹⁷²Yb, ¹⁷⁵Lu, ¹⁹⁷Au, ²⁰²Hg, ²⁰⁵Tl, ²⁰⁸Pb, ²⁰⁹Bi, ²³²Th, and ²³⁸U. Dwell times were set to 0.005 s for most elements and 0.01 s for the rare-earth elements (REEs) and Y. Accuracy for most elements was <10% error; however, for the heavy rare-earth elements (HREEs) the error was ∼20%. Precision error for trace elements is estimated to be <10%. Data reduction was performed using the “trace elements next” data reduction scheme in Iolite 4 [70,71,72].

4.4. Secondary Ion Mass Spectrometry (SIMS)

A secondary ion mass spectrometer (SIMS) at the Manitoba Isotope Research Facility (MIRF), University of Manitoba (Winnipeg, MB, Canada), was used to analyse in situ S (baryte) isotopes. Multiple sonic baths were used to clean the samples prior to gold coating. The operating conditions included a 3 nA caesium (Cs⁺) primary beam accelerated at 10 kV with a 20 μm sputtering diameter, a 300 V sample offset, a −9 keV secondary accelerating voltage, a 247 μm slit, an 18 ns dead time, and a mass resolving power of 347. A Balzers SEV 1217 electron multiplier coupled to an ion-counting system was used to detect ions.

Stable isotopes of sulphur (δ³⁴S) are reported in per mil (‰) relative to Vienna Canyon Diablo Troilite (V-CDT). An in-house baryte standard (δ³⁴S = 31.5‰) was used and repeatedly analysed during analytical sessions for precision assessment. There was an error of 0.3‰ for S isotopes. The average spot-to-spot reproducibility (1σ) was 0.4‰ for the baryte standard.

5. Results

5.1. Petrography

Three general stages in the evolution of the Walton deposit (pre-, syn-, and post-main-stage sulphides) were established from detailed petrographic studies and are summarised paragenetically and systematically in Figure 4a,b, respectively. In this paper, the focus is on the non-sulphide phases; sulphides will be described in detail in a complementary future publication. Representative photomicrographs of the mineralogy of the deposit are depicted in Figure 5, Figure 6 and Figure 7.

5.1.1. Pre-Main Stage

Unmineralised and unaltered laminated Macumber Formation limestone from Cheverie (~30 km from the Walton deposit) provides insight into the nature of the host rocks prior to alteration and mineralisation. The limestone is a peloidal wackestone/packstone with peloids up to 500 µm, carbonate mud, and minor primary porosity (Figure 5a,b) that corresponds to the laminated carbonate rock of the upper Macumber Formation [53]. The limestone is referred to as C0. The edges of the pore spaces in C0 are lined with a clear calcite cement (C1) and locally occluded by botryoidal hematite of unclear timing (Figure 5c). Limestone was not present in any samples from the deposit area used in this study.

At the deposit, local dolostone (D0) consists of finely crystalline (<50 µm) dolomite (Figure 5d). Locally, porosity in the dolostone is occluded by carbonate cement (A1) that is mostly ankerite with some areas of dolomite (Figure 5d,e) that typically goes extinct as a single crystal and is interpreted as one phase.

The siderostone (Sd0) at the deposit exhibits depositional characteristics that resemble those of the Macumber Formation limestone elsewhere (Figure 5f,g) and contains no calcite or dolomite. Abundant hydrocarbon inclusions are present in siderostone based on examination under UV light (Figure 5h,i). The siderostone is subdivided into sideritic Macumber Formation (Sd0a; Figure 5f,g) and sideritic Pembroke breccia clasts (Sd0b; Figure 5j), both consisting of finely crystalline (<10 µm) siderite. Dolomite and siderite (regardless of type) do not co-occur, but locally, siderite replaces ankerite in the host rock.

It has generally been assumed that the massive baryte ore body, which is localised to the top of the Macumber Formation, formed after the siderostone [32]. In this study, samples of the baryte ore body in the upper Macumber Formation were extremely limited and mostly come from dump material, which does not allow for a detailed first-hand account of baryte characteristics and distribution. Therefore, our observations are integrated with historical accounts. Samples representing the main baryte ore body consist of primarily finely crystalline (<25 µm) baryte (B0; Figure 5k–o) with local coarsely crystalline baryte that is described later (Figure 5k–q). The finely crystalline baryte has various colours (i.e., shades of beige and brown to orange) and has irregular dark material as patches or thin (mm-scale) layers of possible bitumen [37].

Pores in the siderostone are occluded by large (0.5 to 2 mm), locally curved, siderite cement crystals (Sd1; Figure 5r and Figure 6a) followed by tabular baryte (B1) (Figure 5g and Figure 6b). The cores of Sd1 are cloudy (Sd1a), related to abundant small fluid inclusions, and overgrown by a clear siderite rim (Sd1b). Clay locally occludes porosity (Figure 6c) but rarely co-occurs with B1, such that the timing of clay and B1 development remains ambiguous.

Predating the main stage mineral phases are a variety of pyrite types found in siderostone: framboids, cubes, overgrowths, and replacive (Figure 6d–f). Stylolite swarms of coalesced pressure-solution seams (up to 1 mm; Figure 5f and Figure 6g–l) are common in laminated siderostone and contain pyrite framboids, siderite, rutile, quartz, K-feldspar, chalcopyrite, tennantite, and baryte.

5.1.2. Main Stage

This stage commenced with massive and banded, colloform pyrite that recrystallised to coarser grains lacking the precursor textures (Figure 6m,n). These forms of pyrite are post-dated by various phases including tennantite, sphalerite, galena, chalcopyrite, and bornite (Figure 6m–s).

Locally, a coarsely crystalline siderostone (20–200 µm crystals; Sd2) is associated with elongated, displacive euhedral baryte laths (B2; Figure 7a–c). Pseudomorphs of siderite-baryte-galena after an earlier mineral (identity uncertain; possibly a sulphate such as anhydrite) are abundant (Figure 6r and Figure 7d,e). Each pseudomorph consists of one or more of the three minerals. The siderite in the pseudomorphs (Sd3) is seen to be overgrown or replaced by baryte (B3) followed by galena. The timing of the sulphate dissolution and precipitation of siderite-baryte-galena in relation to Sd1-Sd2 and B1-B2 and their associated mineral phases is unclear, but the appearance of galena suggests an association with sulphide mineralisation.

Fractures in Sd0 that crosscut main-stage sulphides contain siderite cement and vein-type baryte (Figure 7f–h). The siderite cement (Sd4) is equant, clear, or yellow, whereas the baryte (B4) has numerous forms including radial and coarse. Locally, B4 is associated with pressure-solution seams or stylolites (Figure 6g–l). Abundant hydrocarbon inclusions (i.e., petroleum), as seen as light blue under UV light, decorate fracture planes cutting B4 (Figure 7i,j) and are part of complex fluid inclusion assemblages that include saline aqueous inclusions and mixed CO₂-CH₄ types [38].

Siderite cement is also present in the Horton Group (Sd5) in pores among quartz grains. Veins of baryte (B5) are noted in the Horton Group (Figure 7k) and could be the same as any of the earlier baryte cement types (i.e., B1–B4).

5.1.3. Post-Main Stage

Ankerite 2 (A2) occupies veins that crosscut P6 (Figure 7l). The dolomite veins (D1) crosscut D0 (Figure 7m). Calcite cement (C2) is present in the matrix of sideritised and mineralised Pembroke breccia (Figure 5j). The matrix also contains the same minerals as the stylolite swarms. Unaltered Macumber Formation limestone is crosscut by veins of blocky calcite (C3; Figure 5a and Figure 7n).

In some samples of the massive baryte, the finely crystalline baryte locally coarsens (i.e., ripens) to radial baryte having a spherulitic texture (B6) (Figure 5m,n). The coarse baryte contains abundant fluid inclusions that occlude cores (primary origin) or decorate fractures (secondary) (Figure 5p,q and Figure 7i,j). The amount of B6, and its distribution in the massive baryte ore body, is unclear; it was described in previous studies [30,31].

5.2. Baryte Cathodoluminescence

Cathodoluminescence (CL) imaging of baryte revealed bright and dull luminescence. Images of B1 show a brightly luminescent core containing multiple euhedral growth zones with a uniformly dull overgrowth on the corroded interior (Figure 6b). These zoned growth features in B1 are not visible in transmitted light. Bright luminescence is also documented in B2-B4 and B6 (Figure 5l and Figure 7c,e,h) with most baryte showing bright luminescence lacking zoning. Finely crystalline B0 has slightly dull luminescence relative to B6 (Figure 5l). No CL imaging was performed on B5.

5.3. Carbonate Mineral Chemistry

5.3.1. Major and Trace Element Chemistry

Calcite, dolomite, ankerite, and siderite compositions obtained via LA-ICP-MS are depicted in Figure 8 and Figure 9. The complete dataset is available in ESM S1. Binary plots show a slight negative correlation for Mg versus Mn and a slight positive correlation for Mn versus Fe in calcite (Figure 8).

Dolostone is non-stoichiometric and contains, on average, 9.91 wt. % Mg, 0.78 wt. % Mn, and 0.55 wt. % Fe (n = 6). Ankerite cement contains, on average, 4.54 wt. % Mg, 2.36 wt. % Mn, and 8.16 wt. % Fe (n = 6). Vein dolomite is non-stoichiometric and contains, on average, 9.64 wt. % Mg, 0.63 wt. % Mn, and 0.28 wt. % Fe (n = 5). Strontium is higher in the dolostone (avg. = 98.50 ppm; n = 6) and dolomite vein (avg. = 103.05 ppm; n = 5) than A1 (avg. = 27.26 ppm; n = 6). Manganese and Sr have a negative correlation.

All siderite types are non-stoichiometric (Figure 9). Sideritic Macumber Formation contains, on average, 0.75 wt. % Mg, 0.36 wt. % Ca, and 5.11 wt. % Mn (n = 20). In contrast, sideritic Pembroke breccia contains, on average, 8.30 wt. % Mn, 1.86 wt. % Mg, and 1.11 wt. % Ca (n = 7), with the Ca and Mn higher than other siderite types. The core of early siderite cement contains, on average, 0.91 wt. % Mg, 0.18 wt. % Ca, and 5.02 wt. % Mn (n = 9), whereas the rim of early siderite cement contains, on average, 2.50 wt. % Mg, 0.13 wt. % Ca, and 4.91 wt. % Mn (n = 3). Horton Group siderite is the only siderite type with Sc consistently above the limit of detection (avg. = 17.51 ppm; n = 5).

5.3.2. REEY Chemistry

The ΣREEY (sum of rare-earth elements and yttrium) in carbonates ranges from 1 to 103 ppm and is generally lowest in dolomite and ankerite and highest in siderite (Figure 8d and Figure 9e). The Y/Ho ratio (Figure 8e and Figure 9f) generally decreases from limestone (avg. = 37.87; n = 4) to C1 (avg. = 32.48; n = 8) to dolostone (avg. = 31.64; n = 6) to ankerite (avg. = 20.93; n = 6) to siderite (avg. = 16.14; n = 93) (excluding that of the Horton Group, which averages 19.93 (n = 5)).

Spider diagrams for carbonates are PAAS-normalised (Figure 10) and show three or four different REEY patterns. The patterns are mostly LREE-depleted (LREE<MREE) with lesser convex-like, flat, and upward sloping (i.e., HREE-enriched) patterns. Lanthanum and Ce anomalies were quantified using equations in Bau and Dulski [75] (Figure 11). The summary of REEY data is in

Table 1. Summary of REEY data for the different carbonate types.

	Statistic	C0	C1	D0	A1	Sd0a	Sd0b	Sd1a	Sd1b	Sd2	Sd3	Sd4	Sd5	D1	C2	C3
ΣREEY	n	4	4	6	6	20	7	9	3	15	14	25	5	5	4	4
(ppm)	max	47.3	35.1	5.5	34.2	91.4	103	36.1	6.9	40.3	36.7	42.9	63.7	6.2	45.2	42.2
	min	9.7	31	2.4	7.9	14.8	36.4	13.2	4.2	4.5	9.2	13.9	26	4.4	20.3	0.7
	mean	22.6	32.5	4.2	17.7	42.6	62.6	22.8	5.6	25.3	23.3	24.9	39	5.5	33.5	18.9
	st. dev.	16.9	1.8	1.2	10.3	20	24	7.6	1.2	10.7	10	8.7	17	0.7	12.2	17.6
Y/Ho	n	4	4	6	6	20	7	9	3	15	14	25	5	5	4	4
	max	40.1	35.7	38.8	23.5	24.5	24.7	17.5	13.7	16	25.7	23.2	22	36.9	30.5	40.1
	min	34.5	31.7	27.7	19.2	12	18.4	12.3	10.6	10.1	11.4	12.7	17.9	28.7	25.9	29.7
	mean	37.9	33.5	31.6	20.9	16.6	20.6	15.7	12.2	12.9	16	17.2	19.9	33.4	27.9	37.1
	st. dev.	2.4	1.7	4.5	1.6	3.4	2.2	2.1	1.5	1.7	4.6	3.4	1.9	3.9	2	5
Ce/Ce*	n	4	4	6	6	19	7	8	<LOD	15	9	20	5	5	4	3
	max	1.17	0.93	1	0.89	1.96	0.85	0.75		1.16	0.86	0.98	1.24	0.95	1.06	0.19
	min	0.84	0.8	0.82	0.54	0.48	0.57	0.42		0.48	0.48	0.25	0.77	0.86	0.86	0.05
	mean	1.03	0.89	0.9	0.69	0.77	0.68	0.58		0.77	0.63	0.59	1.03	0.9	0.97	0.1
	st. dev.	0.14	0.06	0.07	0.14	0.34	0.11	0.11		0.18	0.15	0.18	0.31	0.04	0.09	0.08
Eu/Eu*	n	4	4	6	6	20	7	9	1	15	13	24	5	5	4	3
	max	1.44	1.61	1.35	1.21	1.08	0.96	0.98	0.55	1.46	1.47	1.93	1.66	1.37	1.26	1.31
	min	1.01	1.19	1.12	0.96	0.68	0.83	0.51		0.5	0.87	0.47	0.71	1.25	1.15	1.29
	mean	1.18	1.32	1.2	1.1	0.93	0.88	0.78		0.82	1.04	0.95	1.09	1.28	1.2	1.3
	st. dev.	0.21	0.19	0.09	0.09	0.1	0.06	0.17		0.25	0.23	0.28	0.39	0.05	0.06	0.01
Y/Yo*	n	4	4	6	6	20	7	7	3	15	14	25	5	5	4	4
	max	1.53	1.35	1.52	0.87	0.88	0.93	0.69	0.6	0.63	0.89	0.9	0.86	1.41	1.03	1.51
	min	1.35	1.16	0.96	0.69	0.48	0.73	0.54	0.54	0.42	0.47	0.53	0.7	1.01	0.87	1.04
	mean	1.41	1.25	1.17	0.74	0.64	0.81	0.63	0.57	0.52	0.62	0.67	0.78	1.22	0.93	1.29
	st. dev.	0.08	0.08	0.23	0.06	0.11	0.07	0.06	0.03	0.06	0.15	0.12	0.06	0.17	0.07	0.25

St. dev. = standard deviation. <LOD = below the limit of detection. Ce* = Ce/(0.5*La_SN + 0.5*Pr_SN). Eu* = Eu/(0.67*Sm_SN + 0.33*Tb_SN). Y* = Y/(0.5*Dy_SN + 0.5*Ho_SN). SN = normalised to PAAS. C0—limestone, C1—calcite 1, D0—dolostone, A1—ankerite 1, Sd0a—siderostone (Macumber Formation), Sd0b—siderostone (Pembroke breccia), Sd1a—siderite 1 (core), Sd1b—siderite 1 (rim), Sd2—siderite 2, Sd3—siderite 3, Sd4—siderite 4, Sd5—(siderite 5), D1—dolomite 1, C2—calcite 2, and C3—calcite 3.

.

The carbonates with convex-like patterns are C0, C1, D0, D1, and C2, whereas patterns for A1, Sd1–Sd5, and C3 are mostly LREE-depleted. The LREE-depleted patterns are further subdivided based on the presence of positive (C3) or negative (A1 and Sd1-Sd5) Y anomalies. Two out of five spot analyses of Sd5 are relatively flat. Most spot analyses have a positive Ce anomaly (Figure 11).

5.4. Baryte Mineral Chemistry

The REEY in baryte are mostly below the limit of detection (<LOD) and therefore are not further discussed. Other trace element data are summarised according to the different baryte types above (B0-B6) in Figure 12. For some of the elements there is a separation of the data in terms of type. Strontium shows a large variation from <LOD to 2.7 wt. % with the highest values in B0 and B6 (avg. = 2.12 wt. %), intermediate values for B3 and B5 (avg. = 0.74 wt. %), and the least enrichment for B1 and B4 (avg. = 0.37 wt. %). This range in Sr is larger than that reported by Burtt [29] (0.06 to 1.45 wt. %). The same occurs for Zn, but with a limited spread from 550 to 650 ppm with the highest values (avg. = 616.54 ppm; n = 16) in tabular B4. Additionally, Zn shows a negative correlation with Sr (Figure 12a). Both Ca (avg. = 897.72 ppm; n = 13) and Pb (avg. = 23.03 ppm; n = 13) are higher in B3 than in other baryte types but lack any correlation, whereas Zn and Ca show a positive trend (Figure 12b,c). Iron is low in baryte (<800 ppm) and generally <LOD except for B3 (Figure 12d). The complete dataset is available in ESM S2.

5.5. S Isotopes

Sulphur isotopes were obtained for B0, B1, B4, and B6. Data are provided in

Table 2. Sulphur isotope (SIMS) data.

Stage	δ34SVCDT (‰)	Stage	δ34SVCDT (‰)	Stage	δ34SVCDT (‰)
B0	15.8	B1 (rim)	19.8	B4	16.6
B0	15.3	B1 (rim)	20.7	B4	15.0
B0	14.1	B1 (rim)	20.3	B4	16.7
B0	15.0	B1 (rim)	20.2	B4	15.2
B0	15.2	B1 (rim)	20.3	B4	14.5
B0	14.4	B1 (rim)	20.3	B4	15.8
B0	14.4	B1 (rim)	20.3	B6	11.1
B0	14.4	B1 (rim)	20.1	B6	9.4
B0	14.1	B1 (rim)	19.9	B6	9.7
B0	13.3	B4	15.0	B6	9.5
B1 (core)	17.0	B4	13.9	B6	8.8
B1 (core)	16.2	B4	13.5	B6	10.1
B1 (core)	16.7	B4	14.1	B6	9.1
B1 (core)	16.6	B4	15.7	B6	9.9
B1 (core)	16.4	B4	14.4	B6	9.2
B1 (core)	14.9	B4	15.1	B6	9.3
B1 (rim)	19.7	B4	14.7

All analyses are of baryte. The error (1σ) is 0.3‰ for all analyses.

, with spot locations in Appendix A (Figure A1, Figure A2 and Figure A3). The δ³⁴S_VCDT values for baryte types are 13.3 to 15.8‰ for B0 (avg. = 14.6‰; n = 10; Figure A1), 14.9 to 17.0‰ for B1 core (avg. = 16.3‰; n = 6; Figure A2), 19.7 to 20.7‰ for B1 rim (avg. = 20.2‰; n = 10; Figure A2), 13.5 to 16.7‰ for B4 (avg. = 15.0‰; n = 14; Figure A3), and 8.8 to 11.1‰ for B6 (avg. = 9.6‰; n = 10; Figure A3). Noted again are the different CL responses for B1, with data from the bright luminescent core (avg. 16.3‰) contrasting with the dull luminescent overgrowth (avg. 20.2‰; Figure A2b). These values are mostly lower than baryte S isotope data from Boyle et al. [65], which ranged from 14.8 to 33.6‰.

6. Discussion

6.1. Origin of the Carbonates

6.1.1. Limestone (C0) and Early Calcite Cement (C1)

The unaltered Macumber Formation limestone sample from the Cheverie area has a shale-normalised REEY pattern at 0.1 to 1 times PAAS (Figure 10a) that is slightly convex-like and differs from seawater because it has elevated LREEs and lacks a negative Ce anomaly (e.g., [80]). The reason for this is unclear, but it could be because of burial diagenesis, with the slightly positive Ce anomaly indicating reduced conditions (e.g., [83]), which is consistent with the elevated Fe and Mn contents (Figure 8c). Calcite 1 has the same pattern and likely a similar interpretation as C0.

6.1.2. Dolostone (D0) and Ankerite Cement (A1)

Dolostone from the Walton deposit has REEY patterns (Figure 10c) similar to C0 and C1, but with Y/Ho ratios that approximate chondritic values (27.7 ± 2.7; [74]). The dolostone replaced C0 and likely inherited its pattern. The ankerite cement (Figure 10d) in the dolostone contrasts with D0, C0, and C1 in being LREE-depleted, which indicates it did not equilibrate with a reservoir enriched in LREEs. Possible reservoirs include the underlying Horton Group, which has previously been the suggested fluid aquifer and/or proposed source of the metals in the fluids for Walton [38] and the much deeper Meguma Supergroup rocks (inferred at depth from seismic lines in [43]). The sub-chondritic (<27.7 ± 2.7) Y/Ho ratios and lack of negative Ce anomalies for A1 negate a seawater origin (see Figure 10o). The ankerite cement, as with D0, is also enriched in Mn and Fe, which can only substitute under reduced conditions. The geochemical attributes of D0 and A1 therefore support precipitation under reducing conditions and exclude shallow subsurface conditions (<1 km) where oxidised fluids dominate. Their geochemical features suggest instead that the carbonates formed in an intermediate (1 to 3 km) or deeper (>3 km) burial environment [84].

The LREE-depleted signal of the ankerite is significant (Figure 10d). As was noted above for C0, C1, and D0, the LREE depletion indicates the related fluids did not equilibrate with an LREE-enriched reservoir, such as the rocks of the Horton Group or the underlying Meguma Supergroup. Additionally, this fluid did not equilibrate with the host rock (C0 or D1). These LREE-depleted patterns have also been reported for carbonates from other ore settings: ankerite and siderite at the MacMillan Pass SEDEX district (NWT, Canada; [3]), calcite and dolomite from George Fisher (Mount Isa, Australia; [16]), and siderite and Mg-Fe carbonates from iron ore deposits in Spain [85,86]. The LREE depletion for the MacMillian Pass and George Fisher settings was related to chloride complexation of the LREEs by a saline hydrothermal fluid that caused LREE retention in the fluid (i.e., produces HREE enrichment). This interpretation is also possible for ankerite at Walton, but the exact temperature of formation is unclear. However, it is worth noting that Perry and Gysi [87,88] conducted experiments at similar conditions to Walton mineralisation (100–250 °C and 20 wt.% eq. NaCl) that favoured the incorporation of HREE into calcite, which might be applied to ankerite at Walton.

6.1.3. Siderostone (Sd0) and Siderite Cements (Sd1–Sd5)

The siderostone is replacive and fabric retentive, as evidenced from the inheritance of the porosity and textures of the prior limestone (Figure 5f,g); these observations and conclusions are consistent with previous studies (e.g., [28]). Siderite likely precipitated under reducing conditions because of the need for the Fe-carrying fluid to carry significant amounts of Fe²⁺ (e.g., [85]), which is also indicated by the Mn enrichment in siderite. Europium is redox sensitive and is reduced to Eu²⁺ under either reduced conditions or at >250 °C [89,90]. Because of its larger size, Eu²⁺ is typically excluded from siderite (e.g., [91]). Therefore, positive Eu anomalies in siderite can only form when Eu³⁺ is dominant, which would be under oxidised conditions or at temperatures <200 °C [7]. The high Fe²⁺ and Mn²⁺ content of Walton siderites preclude oxidised conditions, so the slight positive Eu anomalies are more likely due to the fluids being <200 °C, thus consistent with δ¹⁸O_siderite data [28].

The REEY patterns for most siderite types (Figure 10e–l) are consistently LREE-depleted with negative Y anomalies, which contrasts markedly with the other carbonates (except A1) that are either not LREE-depleted (C0, C1, D0, D1, and C2) or lack negative Y anomalies (all other carbonates excluding A1), and the whole-rock data from Sangster et al. [33] (Figure 10p).

The ΣREEY in siderite is generally higher than in the other carbonate types, which could be due to the breakdown of organic matter and/or reduction of Fe-oxide complexes [83], but it is opposite of what would be expected because REEY are typically not strongly incorporated into non-Ca-bearing carbonates (e.g., [7]). Specifically, Bau and Möller [7] determined that the HREEs (charge of 3+) are incorporated into the siderite crystal lattice because they are smaller than the LREEs (i.e., lanthanide contraction) and are higher charged than Fe²⁺. This could also partly explain why most of the Walton siderites have LREE-depleted patterns. The inferred low temperatures suggested for siderite formation exclude the likelihood that chloride complexation of LREEs by a saline hydrothermal fluid led to the LREE-depleted patterns because this process only occurs at >150 °C (e.g., [87]).

Emsbo [92] related the breakdown of Fe-oxyhydroxides during hydrothermal fluid influx to the release of Fe, Mn, REEs, and PO₄. In the case of Walton, it is possible that the formation of siderite, which is enriched in Fe, Mn, and REEs relative to precursor limestone (or even dolostone), was tied directly to this process. Organic matter and Fe-Mn-oxyhydroxides preferentially scavenge REEs (but not Y) [83,93,94] and the release of REEs from those particles would increase the ΣREEY in siderite. This is possible at Walton, because there is significant organic matter in the Macumber Formation limestone, present both as oil inclusions and bitumen [95]. The C isotopic signature of Walton siderite is posited to be depleted relative to the limestone because of the incorporation of light C from organic matter [28], and hence, it is also consistent with the negative Y anomalies in siderite that are attributed to organic matter and/or Fe-Mn-oxyhydroxides [79,91].

However, it should be noted that although Fe-oxyhydroxides and organic matter can have LREE-enrichments [83,93,94], the reason for LREE depletion in siderite is explained by the aforementioned exclusion of LREEs into siderite’s crystal lattice. The LREE-depleted signal for most of the Walton siderite could also indicate that the fluids did not equilibrate with a reservoir enriched in LREEs or that there was LREE retention in the fluid (similar to that explained for A1 above). Ultimately, this suggests a very different fluid than the one(s) that formed the earlier calcite (C0-C1) and D0.

Siderite in the Horton Group at Walton is different from the others with two of five analyses having a relatively flat REEY pattern (Figure 10l). This means that some siderite cement precipitated from fluids that partially equilibrated with a shale-like rock, but not the Horton Group, which has a strongly positive Y anomaly (Figure 10q; [81]) as opposed to the negative Y anomalies in these siderites. These flat siderite patterns could be the result of the interaction of the sideritising fluid with both the Horton Group and with organic matter and/or Fe-Mn-oxyhydroxides. The flat siderite patterns are similar to Mn carbonates from other settings (e.g., [79]) and to Fe-Mn crusts from the Central Pacific Ocean [78], as highlighted in Figure 10l. The higher Sc and V in these siderite cements could be due to incorporation from the host rock. Unfractionated PAAS-normalised spider diagrams of the Horton Bluff Formation from Goodarzi et al. [82] (Figure 10r) are characterised by negative Eu anomalies and mostly chondritic to sub-chondritic Y/Ho values (avg. = 24.66), which are unlike any of the patterns in this study.

The reason for the sub-chondritic Y/Ho values and negative Y anomalies in the siderites could also be because of the breakdown of organic matter and/or reduction of Fe-oxide complexes (e.g., [83]). These particles typically have negative Y anomalies [79,91], which would lower the Y/Ho values by adding more Ho than Y.

In banded iron formations (BIFs), dissimilatory iron reduction (DIR) [96] during early diagenesis has been invoked to precipitate siderite (e.g., [97,98]). Under relatively reducing conditions, DIR occurs when Fe-reducing bacteria couple organic matter oxidation to the reduction of Fe-oxyhydroxides to produce siderite where carbonate or bicarbonate is present, as in reaction (1) [96,97]:

4Fe(OH)₃ + CH₂O + 3HCO₃⁻ → 4FeCO₃ + 3OH⁻ + 7H₂O

(1)

A similar reaction also occurs with Mn-oxyhydroxides and in both cases, sorbed metals such as Co, Ni, V, and REEs are released [99]. Notably, these elements are enriched in Walton siderites (Figure 9). Fe-Mn-oxyhydroxide formation occurs in the oxic layer of a stratified water column and particulates can become buried with marine carbonates or dissolve as they settle to the seafloor through the anoxic layer of the water column (e.g., [98,100]). Similar to BSR, temperatures of <100 °C are typically necessary for Fe-reducing bacteria to operate [99] and mediate the above reactions. Sideritisation at Walton was therefore the result of DIR with contemporaneous host rock carbonate dissolution related to an Fe-rich fluid (Figure 13). This fluid could have been of hydrothermal origin, as posited for siderite formation in other settings (e.g., [98]). Reaction 1 also illustrates that two C sources are involved in siderite precipitation. One C source has the signature of the dissolved precursor carbonate, whereas the other has that of the organic matter, which could account for the slightly depleted δ¹³C signature of Sd0 compared to the precursor carbonate.

Negative δ¹³C signatures, such as those previously reported for Walton siderite [28], can be generated during or immediately after bacterial sulphate reduction (BSR) or thermochemical sulphate reduction (TSR) (e.g., [101,102]) due to the oxidation of organic matter and/or petroleum. Evidence supporting sulphate reduction would be the presence of pre-ore pyrite types that formed from reduced evaporitic sulphur, which suggests that sulphate reduction and siderostone formation could be coupled.

Figure 13. Model of sideritisation and baryte formation modified after Burtt [29], Kelley et al. [103], Tang et al. [98], Reynolds et al. [104], and Grema et al. [105]. (a) Setting at the Walton deposit prior to alteration and mineralisation. (b) Fe-Mn-rich fluids from the basement utilised structures and partially replaced the Macumber Formation limestone with siderostone. (c) Ba- and CH₄-rich fluids that also utilised basement structures mixed with sulphate-rich porewaters to form the overlying baryte ore body. (d) On-site sulphate reduction (with corresponding oxidation of methane) led to sulphide precipitation and enriched residual sulphate in ³⁴S. The probable heat source was from underlying intrusions.

Figure 13. Model of sideritisation and baryte formation modified after Burtt [29], Kelley et al. [103], Tang et al. [98], Reynolds et al. [104], and Grema et al. [105]. (a) Setting at the Walton deposit prior to alteration and mineralisation. (b) Fe-Mn-rich fluids from the basement utilised structures and partially replaced the Macumber Formation limestone with siderostone. (c) Ba- and CH₄-rich fluids that also utilised basement structures mixed with sulphate-rich porewaters to form the overlying baryte ore body. (d) On-site sulphate reduction (with corresponding oxidation of methane) led to sulphide precipitation and enriched residual sulphate in ³⁴S. The probable heat source was from underlying intrusions.

However, the significant production of reduced S would have led to the widespread formation of pre-ore pyrite instead of siderite, because of the presence of abundant reduced Fe, but this was not the case. Therefore, the amount of reduced S present at this time must have been minor. Berner [106] termed this environment post-oxic and interpreted it to form when sufficient organic matter is deposited to consume all dissolved oxygen during organic matter decomposition. Conversely, as there is not sufficient organic matter for sulphidic conditions to prevail, this leads to further organic matter decomposition during nitrate, Mn, and Fe reduction but not sulphate reduction. The subsequent influx of a Ba- and methane-rich fluid to this sulphate-dominant environment would form baryte.

6.1.4. Late Calcite Cements (C2-C3) and Dolomite Veins (D1)

Calcite cement in the matrix of the Pembroke breccia (C2) and D1 have similar REEY patterns to the dolostone, which likely indicate the return of the system to pre-ore conditions. Calcite could have precipitated after the fluid lost Mg, and the formation of both C2 and D1 is likely related to diagenesis, similar to C0 and C1.

The final calcite (C3) has an REEY pattern similar to seawater (i.e., negative Ce and positive Y anomalies and similar REEY profile; Figure 10o; e.g., [80]). The super-chondritic Y/Ho ratios (i.e., >27.7 ± 2.7; [74]) of C0 and C3 are also typical of seawater (e.g., 67–102; [80]) with Y/Ho ratios above normal chondritic values that indicate the complexation of REEs over Y by organic matter and Fe-oxides (e.g., [107]). The strong negative Ce anomalies are common in oxidised fluids in general (e.g., [108]).

6.1.5. Integration of Carbonate Data

A zonation in carbonate chemistry has been noted at Walton [29] as reflected by the fact siderite loses Fe and gains Ca-Mn up stratigraphy and that ankerite is restricted to the uppermost part of the upper Macumber Formation. Ankerite was also noted to be associated with dolostone (Figure 5d,e), which is even more Fe-depleted than ankerite. Away from the deposit, the Macumber Formation limestone contains less Fe on average than the dolostone (Figure 8c). This zonation is similar to that present in SEDEX deposits, such as at the Lady Loretta, Queensland, Australia [109], where the following zonation is noted away from ore: Mn siderite to Mn ankerite to ferroan dolomite to dolomite.

The Y/Ho ratios in the carbonates decrease through the paragenesis from C0 (avg. = 37.87) to D0 (avg. = 31.64 to A1 (avg. = 20.93) to siderite (avg. = 16.14, excluding Horton Group siderite). This progressive change from super-chondritic to sub-chondritic is attributed to being coincident with the breakdown of organic matter and/or Fe-Mn-oxyhydroxides. Negative Y anomalies have been documented in carbonates associated with sulphides in several studies (e.g., [16,110]), with Rieger et al. [16] postulating that sub-chondritic Y/Ho values may indicate the fractionation of Y relative to Ho and that sub-chondritic Y/Ho ratios may be useful in identifying hydrothermal carbonates. Therefore, the decreasing Y/Ho ratios in carbonates through the paragenesis could record the transition from diagenetic to a more hydrothermal-dominant fluid regime.

The ΣREEY in carbonates at Walton range from 1 to 103 ppm, which is much lower than in carbonates at the Scotia Mine deposit (14.74 to 1337.03 ppm; avg. = 202.26 ppm; [9]). Kontak and Jackson [9] proposed that the high concentration of REEY in carbonates at Scotia Mine was likely because of leaching from apatite during fluid flow through the Horton Group, which is dominated by Meguma Supergroup detritus, or by flow through the Meguma Supergroup proper. However, the significantly lower ΣREEY at Walton implies that the Horton Group there may differ in its REEY-rich minerals compared to Scotia Mine. Alternatively, some undocumented differences in the chemistry of the mineralising fluid(s) at Walton compared to Scotia Mine could be responsible for the differences, i.e., such fluid(s) at Walton were less effective at mobilising REEY.

Whole-rock REEY data for the least altered and sideritised Macumber Formation limestone, the former from outside the mineralised area, have flat PAAS-normalised REEY patterns [33] (Figure 10p) that are shale-like but relatively depleted in ΣREEY compared to spot analyses for C0 and Sd0. These data likely include the contribution of organic matter and clays, which are analysed as a part of bulk analyses versus the in situ analyses.

6.2. Origin of the Baryte

6.2.1. Early Baryte Types (B0 and B1)

Sulphide ore bodies at numerous SEDEX deposits are now interpreted as forming due to the replacement of overlying precursor baryte [34,103,104,111]. This model is also adopted for Walton, thereby accounting for the observation that sulphides are localised to the base of the dominantly finely crystalline massive baryte ore body (B0, [29]). All previous studies consider B0 as a post-sideritisation of the Macumber Formation and also a post formation of the overlying evaporites based on crosscutting relationships and because B0 has a δ³⁴S_VCDT signature that overlaps with, and is likely inherited from, the evaporites [32,37,65]. We also suggest that the minor porosity in the Macumber Formation likely contained sulphate-rich porewaters that precipitated brightly luminescent B1, which has the same δ³⁴S_VCDT values as B0 and is therefore likely synchronous with B0.

The global Carboniferous δ³⁴S_sulphate curve shows a decrease from ~20‰ in the early Carboniferous to ~15‰ at 334 Ma ([112] and references therein), the latter equating to the ages of Windsor Group evaporites [51]. Three baryte types (B0, cores of B1, and B4) have δ³⁴S_VCDT values similar to the lower end of this range (avg. = 14.6‰, 16.3‰, and 15.0‰, respectively; Figure A1, Figure A2 and Figure A3) and comparable to some, but not all, previous bulk analyses for B0 (14.8‰ to 33.6‰, avg. = 23.8‰, n = 4; [65]). Assuming precipitation from a fluid with the signature of Carboniferous seawater sulphate, these low values indicate little, if any, fractionation; thus, the baryte δ³⁴S_VCDT values are suitable proxies for Viséan seawater [112,113].

The formation of B0 could have been induced by a bacteria-driven process called sulphate-driven anaerobic oxidation of methane (SDAOM) (e.g., [114,115]):

CH₄ + SO₄^2– + 2H⁺ → H₂S + CO₂ + 2H₂O

(2)

or

CH₄ + SO₄^2– → HCO₃⁻ + HS⁻ + H₂O

(3)

The decomposition of organic matter during burial forms reduced, methane-rich pore waters, which can lead to the depletion of sulphate and are thereby capable of dissolving baryte (e.g., [116]) and anhydrite (e.g., [117]). Barium, sourced from the dissolved baryte and/or from underlying rocks such as the Horton Group, is then transported with upward-diffusing methane-rich fluids to shallower porewaters or the seafloor, which are sulphate-rich and methane-poor thus reprecipitating baryte [116,118]. The sulphate–methane transition (SMT) occurs where the upward-diffusing methane meets with downward-diffusing seawater sulphate and as a result is an area of significant SDAOM [118,119]. The presence of organic matter in the system is supported by the siderite trace elements presented in this study, in addition to C isotope [28] and fluid inclusion [38] data from previous studies.

Carbonate and pyrite are also documented as products of the SDAOM [115]. At Walton, Ba- and methane-rich fluids emanating from the underlying Horton Bluff Formation could have mixed with sulphate-rich pore waters to form B0 and B1. The mixing of two fluids, one Ba-rich and the other sulphate-rich, was theorised by Tenny [30] as leading to the formation of the massive baryte at Walton. Baryte precipitation was followed by sulphide precipitation (cubic, overgrowth, and replacive pyrites of the pre-main stage) as the SDAOM made reduced sulphate available. The formation of B0 and B1 under these conditions would indicate that they are diagenetic barytes [116].

The presence of CL growth zones in B1 (Figure 6b) suggests fluctuation in chemical composition and/or temperature, as is commonly reported for hydrothermal carbonate and quartz [120,121,122], whereas the uniform luminescence in B1 suggests relatively uniform physico-chemical conditions affected the overgrowths. The interface between these zones is corroded, likely owing to coupled dissolution–precipitation (CDP; [123]), which has been described for minerals such as quartz (e.g., [120,122]) and calcite (e.g., [121]). The CDP reaction could have been facilitated by hydrocarbon-rich fluids, as is documented with the methane-enhanced dissolution of anhydrite (e.g., [117]). Increasing temperature and the presence of the chlorides increases the solubility of baryte (e.g., [124]); thus the mineralising fluid (100 to 300 °C and 20–25 wt. % NaCl eq.; [38]) should be capable of dissolving baryte.

The rims of B1 have δ³⁴S_VCDT values similar to the upper end of the range for Carboniferous seawater sulphate (avg. = 20.2‰, n = 10; Figure A2b). This suggests that during the CDP of B1, as is evident in Figure A2b, the fractionation of the dissolved sulphur from the bright CL B1 (δ³⁴S_VCDT values of ~15‰) occurred, such that newly precipitated dull CL B1 was enriched in δ³⁴S_VCDT (~20‰). This fractionation indicates that some sulphate reduction occurred (e.g., [125]), with H₂S production (see reactions 2, 3) facilitating post-B1 sulphide precipitation.

6.2.2. Later Baryte Types (B2–B6)

Baryte inferred to pseudomorph sulphate (B3) has more Ca than other baryte types, which is likely an effect of inheritance from the precursor (i.e., anhydrite or gypsum). For comparison, Ca in Walton baryte is lower than in baryte from hydrothermal vent deposits, such as the Endeavour Segment, Juan de Fuca Ridge (up to 1.2 wt. %, [20]). Also, the overall low trace element contents (excluding Sr) typify baryte in other studies, such as for volcanogenic massive sulphide (VMS) deposits [21,22], and this has been attributed to the large size of Ba, which precludes substitution.

The baryte trace element data fall into two groups, (1) B0 and B6 and (2) B1-B5, which correspond to differing paragenetic stages (Figure 13). The former group corresponds to the early finely crystalline baryte and its later recrystallised complement, whereas the latter group corresponds to the coarse varieties of tabular baryte. Of note is the high Sr in early baryte versus the higher Zn and Pb and lower Sr contents of later stage baryte. Therefore, distinct fluid signatures are recorded by differing baryte types, which supports their separation paragenetically.

6.2.3. Integration of the Baryte Data

The highest δ³⁴S_VCDT value for baryte at Walton is 33.6‰ [65], which is ~14‰ heavier than Carboniferous seawater, but less than the fractionation measured in other studies (e.g., >40‰; [126]). Sivan et al. [127] determined that the addition of Fe oxides to experiments involving the SDAOM results in DIR and a decrease in the magnitude of S and O isotopic fractionation of sulphate. In those experiments, S isotopic fractionation was up to ~20‰ with hematite reduction but up to ~40‰ without it. This supports an environment of siderite precipitation via DIR followed by baryte precipitation.

Boyle [32] noted that most of the 4.5 Mt baryte ore body is composed of finely crystalline baryte (B0 of this study). Notable are the textural similarities of B0 and B6 (Figure 5k–o) with that of the recrystallisation and coarsening (i.e., textural ripening) of silica from opal to mosaic textures and equant grains from the sampling of geothermal wells (e.g., [128]) and epithermal quartz veins (e.g., [129]). Therefore, we suggest that these observations are consistent with a model involving the initial formation of an amorphous-like baryte material, perhaps due to the mixing of sulphate-rich and Ba-rich fluid reservoirs, as posited above as part of the SDAOM. A similar fluid-mixing model has been posited for the formation of early finely crystalline baryte at the Red Dog SEDEX deposits, Alaska, that also later underwent recrystallisation [103].

Post-main stage barytes (B4) are present as coarser grains confined to discordant vein features and reflect a paragenetically later fluid event than B0, which is supported by its Zn-Pb enrichment and Sr depletion compared to B0 (Figure 12). This same fluid event could be responsible for facilitating the recrystallisation of B0 to B6 (Figure 5k–o), which is consistent with the observations of B0 by Dawson [130], Boyle [32], and Kontak and Sangster [38]. Fluid inclusion studies on B4 indicate that these fluids, which are most likely related to main stage sulphide mineralisation, were heated (i.e., 100 to 300 °C), highly saline (20–25 wt. % eq. NaCl), and Ca-rich [38].

The B6 δ³⁴S_VCDT values (8.8 to 11.1‰, avg. = 9.6‰, n = 10) are lower than the evaporitic sulphate source, and could be the result of the oxidation of H₂S [131,132], which is reported for some barytes of the Red Dog SEDEX district [133]. This indicates that there was either excess H₂S that did not form sulphides and/or the oxidation of pre-existing sulphides. Baryte with lower δ³⁴S_VCDT values than its precursor has also been described for stratiform baryte-sulphide deposits near Aberfeldy, Scotland, with these values attributed to pre-metamorphic diagenetic processes involving fluid fluxes between the mineralised beds and the host sediments [134]. Alternatively, the low values of B6 can also be attributed to mixing of light and heavy δ³⁴S_VCDT values to average 10‰.

Alternatively, the depletion of methane would also cause dissolved baryte to reprecipitate because baryte has retrograde solubility at the temperature range indicated by fluid inclusions (100–300 °C; [38,67]), if there is no methane present (e.g., [135]). The depletion of methane would also cease the SDAOM and production of reduced S. Baryte precipitation by the SDAOM in SEDEX deposits has been invoked in other studies including Magnall et al. [111], Reynolds et al. [104], and Grema et al. [105].

Methane in baryte-hosted fluid inclusions at Walton [38] could have been sourced from the organic-matter-rich Horton Bluff Formation of the Lower Horton Group [136] and generated from its heating (e.g., [137]) (Figure 13). Methane could also originate from the Macumber Formation itself, which is also organic matter rich [53]. Rogers and Savard [95] analysed oil inclusions from the Jubilee deposit that is hosted by the Macumber Formation in the River Denys sub-basin (Cape Breton, Nova Scotia) and determined that the oil inclusions were sourced from the underlying Horton Group. It is possible that the same can be said for the organic matter in the Macumber Formation at Walton.

In a study of the Guaymas Basin in the Gulf of California, Berndt et al. [138] determined that rift magmatism can trigger methane venting in sedimentary basins due to the heat provided by the intrusions. Rifting during the early Viséan is evident in Nova Scotia and the Maritimes Basin [50,61] and has previously been suggested as the source of heat during Walton mineralisation [38], which Conliffe et al. [27] also recently advocated. This deeper underlying magmatism (possibly mafic intrusions are present elsewhere in the Kennetcook sub-basin) are the likely heat source for the hot (up to 300 °C) metalliferous fluids [38] and may have caused the recrystallisation of Sd0 and B0 to Sd2 and B6, respectively, and the formation of the coarser, late baryte (B2-B5). Precipitation from these heated fluids indicates that B2–B6 are hydrothermal barytes and formed from a distinctly different environment than B0 and B1.

7. Conclusions

The Walton baryte-sulphide ore deposit is analogous to SEDEX deposits wherein there occurs the juxtaposition of a pre-ore siderostone unit and a massive baryte replacement body. The widespread sideritisation is herein tied to pre-main stage DIR-mediated breakdown of Fe-Mn-oxyhydroxides and organic matter that were a sink for elements such as Fe, Mn, and REEs. The siderostone partially incorporated organic-matter-sourced C and Fe, Mn, REEs, etc., released from Fe-Mn-oxyhydroxides and organic matter. The finely crystalline baryte formation was the result of evaporite dissolution related to the SDAOM followed by an ingress of a basement-derived fluid enriched in Ba and methane, with methane likely sourced from the Horton Bluff Formation. Reduced S formation owing to the SDAOM led to minor pre-ore pyrite precipitation. The subsequent influx of a heated metalliferous fluid caused the formation of widespread sulphide mineralisation and initiated the local recrystallisation of earlier siderostone and massive microcrystalline baryte units, along with the formation of veinlets of both phases. The heat that drove this mineralised setting may have been related to rifting and associated magma underplating during the early Viséan, which also led to methane venting. The post-ore events consisted of renewed siderite and baryte precipitation. The complex δ³⁴S_VCDT values for barytes reflect sulphur originally sourced from the Viséan evaporites (~15‰) that was subsequently modified to heavier (to ~20‰) and depleted (~10‰) values due to fractionation related to synchronous sulphate reduction and the oxidation of H₂S, respectively, suggesting an excess of reduced S.

Future work to further resolve the origin of the siderite should include in situ C isotopes on the siderite and the organic matter for insights into the incorporation of organic C into siderite. The identification of distinct S isotopic populations in baryte would not have been possible without integrating petrography, CL imaging, and SIMS analysis and thus highlights the application of the microanalytical protocol employed. Cathodoluminescence should be more widely used in studies of baryte because it highlights the textural complexity of baryte and can aid in targeting areas for subsequent in situ analyses.