Interpreting the Complexity of Sulfur, Carbon, and Oxygen Isotopes from Sulfides and Carbonates in a Precious Metal Epithermal Field: Insights from the Permian Drake Epithermal Au-Ag Field of Northern New South Wales, Australia

Abstract

The Drake Goldfield, also known as Mount Carrington, is located in north-eastern New South Wales, Australia. It contains a number of low–intermediate-sulfidation epithermal precious metal deposits with a current total resource of 724.51 metric tons of Ag and 10.95 metric tons of Au. These deposits occur exclusively within the Drake Volcanics, a 60 × 20 km NW-SE trending sequence of Late Permian volcanics and related epiclastics. Drilling of the Copper Deeps geochemical anomaly suggests that the volcanics are over 600 m thick. The Drake Volcanics are centered upon a geophysical anomaly called “the Drake Quiet Zone” (DQZ), interpreted to be a collapsed volcanic caldera structure. A total of 105 fresh carbonate samples were micro-drilled from diamond drillcores from across the field and at various depths. A pXRD analysis of these carbonates identified five types as follows: ankerite, calcite, dolomite, magnesite, and siderite. Except for three outlier values (i.e., −21.32, −19.48, and 1.42‰), the δ¹³C_VPDB generally ranges from−15.06 to −5.00‰, which is less variable compared to the δ¹⁸O_VSMOW, which varies from −0.92 to 17.94‰. μ-XRF was used to analyze the elemental distribution, which indicated both syngenetic/epigenetic relationships between calcite and magnesite. In addition, a total of 53 sulfide samples (primarily sphalerite and pyrite) from diamond drillcores from across the Drake Goldfield were micro-drilled for S isotope analysis. Overall, these have a wide range in δ³⁴S_CDT values from −16.54 to 2.10‰. The carbon and oxygen isotope results indicate that the fluids responsible for the precipitation of carbonates from across the Drake Goldfield had complex origins, involving extensive mixing of hydrothermal fluids from several sources including those of magmatic origin, meteoric fluids and fluids associated with low-temperature alteration processes. Sulfur isotope ratios of sulfide minerals indicate that although the sulfur was most likely derived from at least two different sources; magmatic sulfur was the dominant source while sedimentary-derived sulfur was more significant for the deposits distal from the DQZ, with the relative importance of each varying from one deposit to another. Our findings contribute to a greater understanding of Au-Ag formation in epithermal environments, particularly in collapsed calderas, enhancing exploration strategies and models for ore deposition.

1. Introduction

Epithermal deposits, which form at shallow depths and low temperatures, are significant sources of precious metals, such as Au and Ag [1]. Collapsed calderas are considered ideal locations for these types of deposits, as they provide the necessary conditions for the formation of mineralization from ascending thermal waters. These deposits typically occur in the shallow parts of magmatic systems, at depths ranging from ~50 to 1500 m, with ore minerals precipitating at temperatures between <150 °C and ~300 °C, often within associated volcano-sedimentary basins [2,3,4,5].

The Drake Goldfield, also known as Mount Carrington, is located in north-eastern New South Wales (NSW) Australia, ~5 km north-east of Drake Village, 44 km east of Tenterfield, and ~800 km north of Sydney (Figure 1). It contains a number of low–intermediate-sulfidation epithermal precious metal deposits (Figure 2), including Kylo, Strauss, Red Rock, Lady Hampden, Silver King, White Rock, and White Rock North [6], along with numerous other small deposits and prospects. Lady Hampden is unusual for the Drake Goldfield in being a Au-Ag deposit; All Nations, Gladstone Hill, and West Copper Deeps are all Cu prospects. The Strauss, Kylo, Silver King, and Lady Hampden deposits are collectively known as the Mount Carrington Group, as they are in close proximity to the historical Mount Carrington Au-Ag deposit. These deposits occur exclusively within the Drake Volcanics, which comprise a 60 × 20 km NW-SE trending sequence of Late Permian shallow volcanics and related epiclastics.

The first significant precious metal deposit discovered in the Drake Volcanics was the White Rock silver deposit which was discovered in 1886, and its geology was first described by Andrews [9]. Subsequently, many deposits of differing styles were discovered and mined. Although being known and mined for over 100 years, there has been no detailed study on the deposits associated with the Drake Volcanics as a whole, apart from that of Perkins [10] on the deposits of the Red Rock Field, which is one of the smallest of the fields within the larger Drake Goldfield. Recent work by White Rocks Minerals suggests that most of the economic precious metal deposits within the Drake Volcanics are centered upon a geophysical anomaly called “the Drake Quiet Zone” (DQZ), interpreted to be a collapsed volcanic caldera structure (Figure 3). The Drake Goldfield has been mined since the late 1800s for precious metals, particularly Au and Ag [6,9] with the most recent mining activity by Mt. Carrington Mines (1988 to 1990) producing 0.714 metric tons of Au and 434,870 oz of Ag [11]. Recent estimates suggest that the Drake Goldfield contains an Indicated Resource of 724.51 metric tons of Ag and 10.95 metric tons of Au [6]. The present study builds upon more recent unpublished studies on the relationships between mineralization, grade, and alteration assemblages/intensity on the White Rock and the Lady Hampden deposits [12,13]; the relationship between primary volcanic facies and mineralization and correlation between the White Rock and White Rock North deposits [14]; the structural controls and relationship between mineralization and alteration for the Strauss deposit [15]; and a detailed mineralogical and geochemical study on selected Ag-rich deposits within the Drake Goldfield [16].

As there were no comprehensive, published studies across the Drake Goldfield itself and on mineralization within the Drake Volcanics, the main objective of this study was to document the isotopic compositional variations of sulfides and carbonates that precipitated during different stages across the Drake Goldfield. A number of vein-like carbonates and sulfides from mineralized horizons from diamond drillcores from across the Drake Goldfield and at various depths were systematically studied for their carbon, oxygen, and sulfur isotopes in order to better understand the source of the mineralizing and alteration fluids and to provide information about the genesis of the Drake Goldfield.

2. Geological Setting

2.1. The Southern New England Orogen

The Drake Goldfield is near the north-eastern boundary of the Southern New England Orogen (SNEO) or Southern New England Fold Belt [6,18] (Figure 1). The Southern New England Orogen is a sub-province of the New England Orogen (NEO), the youngest sub-province of the Tasmanides, whose formation is widely accepted to comprise two main cycles of compression and extension since the early Cambrian [7,19,20]. The NEO extends ~1600 km along the east coast of Australia from Townsville in northern Queensland to Newcastle in New South Wales (Figure 1). It is divided by the Clarence–Moreton Basin into northern and southern provinces, which are the Yarrol–Gympie Province and SNEO, respectively [19,21]. The SNEO is further divided into two regions, the Tamworth Belt and the Tablelands Complex, divided by the Peel-Manning Fault System [7,22]. The Tamworth Belt represents a relatively weakly deformed fore arc environment, while the Tablelands Complex comprises a series of accretionary wedge/subduction complexes [19,21]. During the uppermost Carboniferous to Lower/Middle Permian, the SNEO had been through an extensional cycle followed by widespread orogenic deformation, granitoid intrusion and volcanism [18,20]. The host rocks of the Drake Goldfield are mostly volcanics which formed during this period [6,23].

2.2. The Drake Volcanics

The Drake Volcanics were first briefly described by Andrews [9] and much later as a sequence of volcanic rocks some 20 km wide and 60 km long trending NNW, close to the north-eastern margin of the Tablelands Complex within the SNEO and bounded to the west by the Demon Fault (Figure 2 and Figure 3) [7]. It has been shown through U-Pb zircon dating that the Drake Volcanics comprise middle Permian (~265 Ma [23]) acid to intermediate volcanics dominated by volcaniclastic andesitic units intruded by sub-volcanic andesite and rhyolite porphyries [18,24,25]. The Drake Volcanics have an estimated thickness of ~600 m in the Red Rock–White Rock–Lady Hampden area [18], while Thomson [8] estimated a maximum thickness of 900 m elsewhere.

The Drake Volcanics are calc-alkaline, conformably overlie the Razorback Creek Mudstone, and are conformably overlain by the Gilgurry Mudstone (Figure 2 and Figure 3) [6,8,25,26,27]. The Carboniferous to Early Permian sedimentary Emu Creek Formations both underlie the Drake Volcanics and are fault-bounded on its eastern boundary, hosting the gold mineralization of the Tooloom and Lunatic Goldfields, while to the west, the Drake Volcanics are intruded by the Early Triassic Stanthorpe Monzonite pluton and structurally bound by the Demon Fault (Figure 2 and Figure 3) [6]. The volcanics are relatively undeformed, with only slightly inclined bedding (~20°). Some researchers have suggested that the Drake Volcanics formed during rifting or graben formation [28,29], while others have suggested that they formed due to transform/strike slip fault movement [30].

Airborne magnetic imagery revealed a 250 km² roughly circular area of low magnetic response (the Drake Quiet Zone or DQZ) (Figure 3). This region with a diameter of approximately 20 km has been interpreted to represent a collapsed volcanic caldera structure with the circular shape and the quiet magnetic signature explained by the presence of a caldera filled with less magnetic volcanics than the surroundings [11,31]. This type of structure is a relatively common feature in a range of other epithermal Au-Ag deposits, including Creede (2488.28 metric tons Ag), Colorado, USA [32,33,34] and Round Mountain (513.21 metric tons Au), Nevada, USA [35,36,37]. As a collapsed caldera, the structure has been proposed to control the location of post-collapse intrusives, mineralization, and the extensive alteration within the region. Although Craighead and Gordon [11] believed that the flat to moderately dipping bedding planes acted as major fluid feeder structures, both field relationships and drill holes conclusively show that the mineralized veins are mostly subvertical in orientation, as expected, following epigenetic extensional faulting [37]. This is supported by the conclusions of Herbert [26], Bottomer [18], Perkins [27], and Houston [31] who all believed that structure played an important role in ore localization. Thick lobes and layers of dacitic–lithic and rhyolitic quartz–feldspar fiamme breccias typically observed throughout the area as a thick sequence have been interpreted as representing the caldera-forming eruption products [38].

3. Samples and Methodology

Prior to sample preparation for further analysis, samples were described, imaged, then made into thin sections or polished blocks, and analyzed using transmitted and reflected-light microscopy.

3.1. Sample Description

Carbonate deposition is mostly very late stage and is unrelated to the main mineralization event. The deposition was a multi-stage process involving a Ca-Mg carbonate stage, calcite forming cross-cutting veins, euhedral calcite lining cavities, late-stage calcite filling fractures, and siderite of meteoric origin on fracture surfaces of the Strauss open cut, comprising drusy euhedral crystals (to a few mm) of dark brown siderite lining fractures on altered andesite.

The carbonate samples were selected from carbonate veins from throughout the Drake Goldfield, covering a diverse range in terms of shape, thickness, and lateral continuity. The thickness of the veins ranged widely from <5 mm to >10 cm. The wide range in vein morphologies included one massive vein with straight and sharp boundaries (Figure 4b,e), stockwork veins (Figure 4a), breccia with carbonate-fill (Figure 4d) and sub-parallel veins (Figure 4b), vuggy infill with euhedral carbonate crystals (mostly calcite, Figure 4f), carbonate spots (Figure 4c), and zonation within individual veins occurred in some places (Figure 4e).

The principal occurrence of carbonates within the Drake Goldfield are as follows: (1) replacement/overprinting of earlier rock-forming minerals (e.g., plagioclase) in volcanic tuffs (Figure 5a,d); (2) vein infill as a main component of the carbonate ± quartz stage (Figure 5b,c,f). The vein carbonates are dominantly composed of fine to coarsely crystalline calcite or dolomite (Figure 5b,c,f). Three calcite generations were found within the veins: (1) fine-grained granular-textured calcite veinlets; (2) calcite replacing earlier coarse-grained dolomite; and (3) massive veins comprising coarse euhedral calcite crystals. Ankerite is rare, occurring as fine-grained veinlets, mostly replacing/overprinting calcite (Figure 5e). Dolomite occurs as massive veins and is locally banded with magnesite (Figure 4e) or calcite. Magnesite is more massive and coarsely crystalline, and interbanded with calcite/dolomite (Figure 4a,b). Siderite was observed as euhedral crystals on fracture surfaces within the Strauss open-cut or rimming/cross-cutting earlier magnesite in drill cores.

Sulfides were chosen from the vein and veinlet ores of the Drake Goldfield. Pyrite, sphalerite, and chalcopyrite are by far the most dominant sulfide species and were thus chosen for the sulfur isotope analyses, mainly occurring as sulfide ± quartz veins (Figure 6). Figure 6a–i show the main sulfide assemblages from the Drake Goldfield, including Ccp + Py + Qtz veins (Figure 6a), Py + Sph + Qtz veins (Figure 6b), Py veins, low-Fe Sph + Ccp + Qtz veins (Figure 6d,f), and low-Fe Sph + Ccp + Gn + Py + Qtz veins (Figure 6e).

The overall mineralogy for the deposits within the Drake Goldfield is relatively simple with the main ore comprising pyrite (some with micro to submicron inclusions of gold), with lesser sphalerite, chalcopyrite, and galena; minor electrum/gold and Ag minerals; along with supergene covellite and chalcocite–digenite. A variety of Ag minerals were identified, primarily of the tetrahedrite–tennantite group, along with minor occurrences of polybasite–pearceite species, and trace amounts of acanthite, native Ag, stephanite, pyrargyrite, and jalpaite. There are at least three generations of pyrite in the Drake Goldfield (Figure 6a–i): as small subhedral–euhedral disseminated pyrite in the host rocks (Figure 6c,h) (pyrite I); as euhedral coarse-grained pyrite (pyrite II, Figure 6c,g,i), and as framboidal-like or spheroidal textured pyrite (pyrite III, Figure 6h). Pyrite I formed before the main mineralization stage, while pyrites II and III formed during the main mineralization stage. Chalcopyrite occurs in two different generations, as exsolutions within sphalerite (Figure 6i) or as coarse grains rimming sphalerite or pyrite (Figure 6d,g). Sphalerite is often rimming pyrite (Figure 6g,i). Galena commonly exhibits mutually embayed grain boundaries with sphalerite suggesting co-crystallization (Figure 6i). For a detailed paragenetic sequence, please refer to the electronic Supplementary Material Figures S5–S7.

3.2. Sample Preparation

Fresh carbonate vein minerals were sampled from diamond drillcores covering a wide geographical and vertical distribution from across the Drake Volcanics and selected according to drill logs [6], with all either associated or spatially close to mineralization. A total of 105 samples were selected from 33 diamond drill holes from the Kylo, Lady Hampden, Silver King, All Nations, Mozart, Gladstone Hill, West Copper Deeps, Red Rock, Strauss, White Rock, and White Rock North deposits and one from the Strauss open cut. These deposits are distributed across the entire Drake Goldfield (Figure 2 and Figures S1–S4 in the Supplementary Material).

A further 53 sulfide samples from the main mineralization stage (mainly coarse-grained pyrite and sphalerite along with minor chalcopyrite and galena) were selected from across the Drake Goldfield (Figures S1–S4 in Supplementary Material) to investigate any spatial variation, and samples were also selected from various depths to investigate any vertical variation. Where possible, only coarse-grained and pure separates of sphalerite and pyrite were chosen.

The samples were then collected as ground powders using a tungsten-tipped micro-drill. They were then placed in small sample holders before being ground to <50 µm using an agate mortar and pestle to ensure compositional and grain-size homogeneity. After each sample was collected, the micro drill head was thoroughly cleaned with 85% acetone solution.

3.3. Portable X-Ray Diffraction

Carbonate species identification was made using an Olympus Terra 6400 portable X-ray diffraction (pXRD) (Olympus, Melburne, Australia) analyzer at the University of New South Wales, Sydney, Australia. In total, 15 mg of the sample with a uniform grain size <150 µm were loaded into the vibrating sample holder (VSH), which vibrated the sample without macroscopic movement of the holder [39]. This exposed crystallites in each sample to the X-ray beam at random orientations, thus helping to reduce orientation effects and allowing for superior analytical statistics [39]. The VSH was then introduced into the Olympus Terra pXRD, and the sample was subjected to XRD analysis using CoKα radiation, with a data collection range of 5 to 55° 2θ and an increment of 0.25°. Collected data were then analyzed and interpreted using HighScore Plus and Siroquant^TM (v.4) commercial interpretation software. Phase identification was conducted using HighScore Plus, while Siroquant^TM was used to estimate the quantitative abundance of mineral phases within the sample [40]. For full details on the use of pXRD for quantitative analysis of epithermal assemblages, see Burkett et al. [41].

3.4. Micro X-Ray Fluorescence

Micro X-ray fluorescence (μ-XRF) mapping of two carbonate veins, one from each of the All Nations and Kylo deposits, containing magnesite and calcite, was conducted using a Bruker M4 Tornado μ-XRF at the X-ray Fluorescence Laboratory, University of New South Wales, Sydney, Australia. The instrument is equipped with a Rh tube with a maximum power of 30 W, a tube voltage of 45 kV, a current of 600 μA, and a 25 μm spot size in order to achieve a high spatial resolution. One silicon drift detector was used with an active area of 30 mm². For element-mapping analysis, spectra were acquired every 50 μm with an acquisition time set at 120 ms per pixel. All analyses were performed under vacuum (20 mbar) to facilitate the detection of light elements. Data analysis was performed in standardless mode using the Bruker M4 TORNADO built-in software FP MQuant.

3.5. Stable Isotope Measurements

Carbon and oxygen isotope analyses were carried out in the Bioanalytical Mass Spectrometry Facility, Mark Wainwright Analytical Centre, University of New South Wales, while sulfur isotopes were determined at the Isotope Ratio Mass Spectrometry facility, Central Science Laboratory, University of Tasmania.

For carbon and oxygen isotope analysis, 40 to 60 µg of carbonates were weighed out for each sample and then analyzed using a MAT 253 isotope ratio mass spectrometer equipped with a Kiel IV carbonate device (Thermo Fisher Scientific, Bremen, Germany). The carbonates were reacted with 2 drops of phosphoric acid at 70 °C to release CO₂. The reacted gasses were then subjected to cryogenic removal of water vapor and other non-condensable gasses, and the pure CO₂ was measured against a reference gas. For sulfur isotope determination, between 0.4 and 1.0 mg of finely powdered mineral sample was weighed in tin capsules and analyzed using flash combustion isotope ratio mass spectrometry on a varioPYRO cube coupled to an Isoprime100 mass spectrometer (Elementar Analysensysteme, Langenselbold, Germany). The SO₂ produced during combustion is separated from other reaction products in a trap, then fed into the mass spectrometer and measured against a reference gas.

Stable isotope abundances are reported in delta (δ) values as the deviations from conventional standards in parts per mil (‰) using the following equations:

δ^HI (‰) = [(R_sample/R_standard − 1) × 1000],

with ^HI = ¹³C, ¹⁸O, or ³⁴S and R = ¹³C/¹²C, ¹⁸O/¹⁶O, or ³⁴S/³²S

δ¹³C values are expressed as relative to VPDB (Vienna Pee Dee Belemnite). δ¹⁸O can be standardized against either the VPDB or VSMOW (Vienna Standard Mean Ocean Water) scale. For the present work, the VPDB scale was adopted. The following equation was used for conversion [42] between VPDB and VSMOW scales:

δ¹⁸O_VSMOW = 1.3086 δ¹⁸O_VPDB + 30.86

δ ³⁴S values are reported relative to CDT (Canyon Diablo Troilite, a meteorite which created the Barringer crater, Arizona, USA).

International reference standards with known isotopic compositions, sulfur: IAEA-S-1 (−0.3‰), IAEA-S-2 (22.66‰), IAEA-S-3 (−32.3‰), IAEA-SO5 (0.49‰), NBS 123 (17.44‰), and NBS 127 (20.32‰); carbon and oxygen: NBS18 (−5.01 and −23.2‰ for carbon and oxygen, respectively) and NBS 19 (1.95 and 2.2‰ for carbon and oxygen, respectively), were measured for instrument calibration and quality assurance. The analytical performance of the instrumentation, drift correction, and linearity performance were calculated from the repetitive analysis of these standards. The analytical precision was 0.1‰ for both δ¹³C and δ¹⁸O and 0.2‰ for δ³⁴S.

4. Results

4.1. Carbonate Species

The results of the pXRD analysis for the carbonates from the Drake Goldfield are presented in Supplementary Material Table S1. Five species of carbonates were identified, including ankerite, calcite, dolomite, magnesite, and siderite. Although commonly found in many epithermal deposits (e.g., Eroy et al. [43]), no manganese carbonates, such as kutnohorite or rhodochrosite, were found in the Drake Goldfield. The quantitative Siroquant results are also shown in Table S1.

4.2. μ-XRF

μ-XRF geochemical maps for two carbonate veins from the Drake Goldfield (Figure 7) show the relative variations in chemistry of the veins. There are relatively pure granular magnesite veins in ANDD012 123.55 (Figure 7a,b), which contain a thin prominent pure calcite vein enclosed within the magnesite (Figure 7b), indicating a clear epigenetic origin, and the calcite appears to be a late-stage cavity infill. In contrast, the carbonate veins in KYDD003 188 have a relatively complex texture (Figure 7c,d). The Mg and Ca distributions (Figure 7d) collectively show that there was a regularly changing Mg/Ca ratio during precipitation with coexisting magnesite, calcite, and magnesium-bearing calcite, indicating a syngenetic relationship. The pXRD confirmed the presence of magnesite, calcite, and magnesium-bearing calcite in these samples, consistent with the μ-XRF analysis.

4.3. Stable Isotope Results

4.3.1. Carbon and Oxygen Isotopes

The detailed carbon and oxygen isotope analyses from across the Drake Goldfield are presented in Supplementary Material Table S1. The range of C and O isotopic compositions for these carbonates are presented in

Table 1. The range in carbon and oxygen isotopic compositions of carbonates from Drake Goldfield.

Mineral	Number of Samples	δ13CVPDB (‰)	Average Value for δ13CVSMOW (‰)	δ18OVSMOW (‰)	Average Value for δ18OVSMOW (‰)
Ankerite	4	−9.33–−5.92	−7.39	+5.81–+7.99	+6.63
Calcite	68	−21.32–+1.42	−9.13	−0.92–+17.94	+7.77
Dolomite	17	−13.02–−5.15	−8.47	+3.40–+10.59	+5.47
Magnesite	15	−10.44–−5.37	−7.99	+8.15–+15.84	+11.89
Siderite	1	−13.71	-	+17.11	-

.

4.3.2. Sulfur Isotopes

The results for sulfur isotopes are presented in Supplementary Material Table S2. A total of 53 sulfide samples (primarily sphalerite and pyrite) were analyzed, and these have a wide range of δ³⁴S_CDT values from −16.54 to 2.10‰. Specifically, δ³⁴S_CDT in pyrite ranges between −8.77 and +2.4‰ (mean −2.86‰—12 analyses) with one outlier of −16.33‰ from White Rock; in sphalerite between −11.50 and −0.53‰ (mean −3.62‰—23 analyses); and in a mixture of sulfides ranging from 8.38‰ to −0.64‰ (mean −4.03‰—16 analyses). To this, a single analysis each for chalcopyrite (−2.64‰) and galena (−11.57‰) are added.

5. Discussion

The study of stable isotopes can provide information regarding the diverse origins of fluids, metal sources, chemical conditions, as well as temperatures of ore formation (e.g., [Ohmoto et al. [44]; Huston [45]; Hoefs [46]). In this discussion, we compare the differences within the Drake Goldfield and attempt to explain the spatial distribution of the gold (copper) and silver-dominant deposits.

5.1. Sources of Fluid for the Carbonates

Late-stage carbonate samples from the Drake Goldfield have δ¹³C_VPDB values ranging from −21.32 to 1.42‰, with 102 samples having a relatively limited variation between −15.06 to −5.00‰. This range for most of the analyses is consistent with carbonates from other epithermal deposits, such as the Kushikino Au-Ag deposit, Japan, and the Qiucun Au deposit, China [47,48,49], while the mean value of −8.65‰ is close to the mantle range (Figure 8) [50]. The δ¹³C_VPDB value range for ankerite (−9.33 to −5.92‰), dolomite (13.02 to −5.15‰) and magnesite (−10.44 to −5.37‰) are relatively limited to the mantle range, while calcite (−21.32 to 1.42‰) has a much wider range, and siderite has a more negative value (−13.71‰). This suggests that carbon in ankerite, dolomite, and magnesite has a magmatic dominant source, while the carbon in calcite more likely results from significant mixing of fluids from a predominant magmatic source with those from a sedimentary organic carbon source. As the siderite occurs, lining fractures near the surface in the open cut, the negative value most likely indicates a groundwater source, which could be interpreted as meteoric water that has undergone interaction with the host andesites and organic matter as the waters percolated downwards via fracture networks from the surface. In the δ¹⁸O–δ¹³C diagram (Figure 9 and Supplementary Material Figures S8–S10), the analyses collectively form an overall linear trend centered on an initial magmatic source trending to a more meteoric influence on one side and a low-temperature alteration influence on the other. Moreover, no data plot along any clear trend, indicating mixing of carbon derived from contaminated sedimentary rocks or high-temperature effects. The Razorback Creek Mudstone is an underlying sedimentary host rock; however, the effects of mixing with this potential carbon source are limited.

The carbon isotopes do not show any significant variation between the Au-Cu and Ag deposits, while the oxygen isotopes are slightly different (Figure 9). The δ¹⁸O for carbonates from the Ag deposits trend to heavier values (concentrated in the range of 10 to 15‰), compared to those for the Au (Cu) deposits (concentrated in the range of 3 to 7‰). This difference also indicates that the carbon for the carbonates in the Au (Cu) deposits was dominantly from a magmatic source compared to that for the Ag deposits.

For these samples, the carbon and oxygen isotope trends are generally near horizontal in the δ¹⁸O–δ¹³C diagram (Figure 9). This near-horizontal distribution of carbon and oxygen isotope values may be due to two processes: (1) degassing of CO₂; (2) reaction between the fluid and surrounding host rocks [52,53,54]. If the distribution of carbon and oxygen isotopes is caused by the degassing of CO₂, then the hydrothermal fluid is generally dominated by H₂O, and the degassing of CO₂ has no significant effects on the oxygen isotope composition but has a significant effect on the carbon isotopes within the fluid [55]. Though the number of samples is limited, the silver-rich deposits have a relatively wider range of δ¹³C, e.g., Lady Hampden (−21.32 to −6.16‰), Silver King (−13.72 to 1.4‰), and White Rock (North) (−19.48 to −6.21‰). This likely indicates that the carbonates in the silver-rich deposits have precipitated due to hydrothermal degassing. The observation of coexisting vapor-rich and liquid-rich fluid inclusions from Red Rock [27], and the presence of direct deposition of quartz (throughout the Drake Goldfield), adularia (mostly in the silver-rich deposits), and calcite (usually platy, present in White Rock and West Copper Deeps) into open spaces or colloform banded quartz (Red Rock and Kylo) could be the result of boiling in the Drake Goldfield [54,56]. However, the gold and copper-rich deposits have a narrower range of δ¹³C, e.g., Kylo (−7.83 to −6.79‰), Red Rock (−14.82 to −6.05‰), Strauss (−13.72 to −5.15‰), and West Copper Deeps (−15.06 to −8.84‰). Thus, the degassing of CO₂ might not be the only factor influencing the precipitation of carbonates across the Drake Goldfield. In hydrothermal fluids, the solubility of carbonates increases with temperature, decreases with pressure, and increases with decreasing solution pH [57,58]. Cooling in a closed system is unlikely to induce precipitation of carbonate from the hydrothermal fluid [51]. Although water–rock reactions involving magmatic-derived fluids and country rocks are common in epithermal deposits [59], they may have no significant effects on the carbon isotopes within the fluids [55], due to minor quantities of carbonates in the host Drake Volcanics. Therefore, the precipitation mechanism of carbonate in the gold- and copper-dominant deposits is most likely dominated by the boiling process with minor effects due to water–rock reactions.

On the δ¹⁸O versus δ¹³C diagram (Figure 9), data from the Gladstone Hill and Mozart deposits plot in the magma–mantle field. For the Silver King, White Rock, and White Rock North deposits, most of the points plot in the magma–mantle field with a slight shift to the negative δ¹⁸O direction, suggesting that the fluids from which these carbonates precipitated from were partly meteoric in origin. For the All Nations, Kylo, and West Copper Deeps deposits, the data suggest that the carbon was derived from a magma–mantle source with a significant influence of low-temperature alteration fluids and also exhibits some characteristics indicative of recrystallization (Figure 5a). Only a few of the analyses for the Lady Hampden and Strauss deposits plot in the magma–mantle zone, with most trending towards higher δ¹⁸O values and a few to lower values, presumably due to fluid reactions with the surrounding host rocks and minor localized mixing with meteoric water. For the Red Rock deposit, the carbon appears to have been mainly derived from a magma–mantle source but with a strong meteoric water influence.

The carbon in the carbonate veins throughout the Drake Goldfield likely originated from a magma–mantle source, subsequently mixed to varying extents with fluids linked to low-temperature hydrothermal alteration and meteoric water. Additionally, a few carbonates exhibit characteristics indicative of recrystallization. The spatial distribution of the Drake Goldfield deposits in the DQZ is shown in Figure 2, from which it can be seen that, except for the Red Rock field, all the gold- and copper-rich deposits (including Kylo, Strauss, Lady Hampden, All Nations, Gladstone Hill, and West Copper Deeps) are closer to the center of the DQZ than the silver-rich deposits (White Rock, White Rock North, Silver King, and Mozart). All the deposits in the Drake Goldfield are hosted within rhyolitic to andesitic volcanic tuffs. However, for the silver-rich deposits and the Red Rock deposit, mineralization is associated with intensive brecciation and adularia ± illite ± quartz ± pyrite alteration, characteristic of a typical rock-buffered low-sulfidation epithermal system [5,16]. In addition, the White Rock (North) deposits are hosted within a magmatic breccia, whereas the other deposits are hosted within various tuff and tuff–breccia units. On the other hand, the gold (copper)-rich deposits show characteristics of an intermediate-sulfidation epithermal system, having a common alteration assemblage comprising quartz ± kaolinite ± albite ± illite ± muscovite ± chlorite ± calcite ± siderite ± pyrite, associated with pyrite + chalcopyrite + low-Fe sphalerite + galena ± minor tennantite group minerals, tetrahedrite, and polybasite mineralization [15]. As we have discussed above, the fluids responsible for at least the carbonates in the gold and copper deposits (except for the Red Rock field) and Lady Hampden (δ¹⁸O shift towards positive values in Figure 9) show an initial magmatic origin but were then more greatly affected by fluids responsible for low-temperature alteration. The carbonate-forming fluids for the silver-rich deposits show an initial magmatic fluid origin followed by later mixing with minor meteoric water during their evolution. This is supported by a δ¹⁸O shift towards negative values, as illustrated in Figure 9 and ankerite overprinting calcite (Figure 5e). The metal zonation of the deposits within the central part of DQZ may relate to the localized distance from the fluid and/or heat source. For the Red Rock area, the carbonate-forming fluids also show a magmatic origin but have a much greater meteoric water influence. This may indicate a separate fluid and/or heat source for the Red Rock area. Although the carbon and oxygen isotope data for the carbonate veins from the White Rock and White Rock North deposits overlap with those of the Mt Carrington group silver-rich deposits, field relationships and other data (e.g., petrographic analysis, whole rock geochemistry) strongly suggest that these silver-rich deposits are from separate shallow level volcanic centers.

5.2. Sources of Sulfur

Sulfur in ore-forming fluids can originate from three main sources: (I) mantle-derived sulfur (0 ± 3‰ [60]), (II) seawater sulfur (>20‰), and (III) strongly reduced (sedimentary) sulfur (<0‰ [50]). The δ³⁴S values for the Drake Goldfield epithermal deposits are summarized in Figure 10a and Supplementary Material Table S2, from which it can be seen that the δ³⁴S values for the Drake Goldfield all cluster around −6 to 0‰. Pyrite is the most abundant sulfide and has a wider range in δ³⁴S values (−16.33 to 2.4‰), while sphalerite has a narrower range in composition (−6 to 0‰). Potential fractionation of δ³⁴S within the sulfides in the main stage is unlikely, as supported by the mutually embayed grain boundaries between galena and sphalerite (Figure 6i); similar δ³⁴S values were found for the galena and sphalerite in the same drill hole (WRDD019 in Supplementary Materials Table S2) and in the similar values for coexisting pyrite and sphalerite (SRDD020 10 and WRDD026 78.5 in Supplementary Materials Table S2).

The similar δ³⁴S values observed for both galena and sphalerite within the same drill hole (WRDD019, Supplementary Material Table S2), as well as for coexisting pyrite and sphalerite (SRDD020 10 and WRDD026 78.5 in Supplementary Material Table S2), suggest that isotopic reversals may have occurred. Sulfide minerals generally show isotopic fractionation in the following sequence: δ³⁴S pyrite > δ³⁴S sphalerite > δ³⁴S chalcopyrite > δ³⁴S galena [44]. The slight isotopic disequilibrium observed in our study is also supported by petrographic evidence, which shows early sulfides being replaced by later generations (Figure 6g,i).

Some deposits have a narrow compositional range and therefore suggest a single source of sulfur, while other deposits (e.g., White Rock) have a wide range that extends to more negative values. The broadening of the ranges may suggest mixing with other isotopic reservoirs or the operation of redox processes during epithermal mineralization [64]. However, there is no evidence for a significant change in oxidation state observed between the sulfides. The wide range of sulfur isotopes may in part be due to thermochemical sulfate reduction, as the reduced Razorback Creek Mudstone underlies the Drake Volcanics in the study area and could act as a reductant.

Herbert and Smith [63] reported δ³⁴S values for sulfides from 17 deposits (host rock included) from the Drake Volcanics ranging from −29.5 to 3‰ (Golden Age, Addisons, Red Rock ‘K’, Emu Creek, Lady Hampden, Mt Carrington, Kylo, Guy Bell, Pioneer, White Rock, Adeline, Mascotte, Lady Jersey, Golden Drake, Just In Time, Just In Time North, and the Drake Volcanics), most clustering between −4 and 2‰. They believed that this sulfur was derived from the local country rocks. Perkins [27] concluded that the sulfur for the Red Rock deposits was dominantly magmatic sulfur, while sulfur for the other deposits in the Drake Goldfield and the Drake Volcanics may be the result of seawater sulfur mixing with magmatic sulfur.

In this study, sulfides from the All Nations, Gladstone Hill, Kylo, Lady Hampden, Red Rock, Strauss, and West Copper Deeps have δ³⁴S values ranging from −6.96 to +2.4‰, with 89% of these close to 0 ± 4‰ (Figure 10a,b), similar to the range reported from other low-sulfidation epithermal deposits (−6 to 5‰ [61]). These sulfides have a relatively narrow range in δ³⁴S compositions, suggesting that they formed under stable physical and chemical conditions and were derived from a relatively homogeneous source [44]. The δ³⁴S values for these deposits are clustered near to 0‰, suggesting that the sulfur has a magmatic origin or has been leached from the host volcanic sequence (~1‰ [63]).

In contrast, the δ³⁴S data for Mozart, Silver King, White Rock, and White Rock North deposits show a wider and more negative range (Figure 10b) from −16.54 to −3.98‰, from which White Rock and White Rock North have a range from −16.54 to −4.4‰, similar to previously reported results (~ −15 to ~ −4‰ [63]). Variation in δ³⁴S values is common in epithermal deposits [45,46,47], and Richards [59] summarized that the δ³⁴S values in epithermal systems range widely from −15 to +8‰ (though mostly <0‰). Sulfur isotope values as high as −3.98‰ (close to 0 ± 3‰) suggest a magmatic sulfur source. The presence of a −16.54‰ value implies that a second source contributed minor amounts of sulfur to the Drake Goldfield and was similar to that of reduced sulfur in sedimentary rocks [50], suggesting that remobilized sulfur from a sedimentary source may have been incorporated into the hydrothermal fluids. Furthermore, remobilization of sulfur from a sedimentary source would be expected to produce a wider spread of δ³⁴S values [65]. Given that the Razorback Creek Mudstone underlies the Drake Volcanics in the study area, magmatic and sedimentary-derived sulfur mixing is certainly possible.

Furthermore, most of the deposits which have a magmatic dominant δ³⁴S source are gold- and copper-dominant, while the deposits that have a mixed δ³⁴S source are silver-dominant. Importantly, there is no noticeable δ³⁴S variation with depth either for individual deposits or across the Drake Goldfield as a whole.

6. Conclusions

Mineralization within the Drake Goldfield mainly occurs in the felsic intrusive units and andesitic/dacitic tuffs. There is a wide range of carbonate species within the Drake Goldfield, including ankerite, calcite, dolomite, magnesite, and siderite. Most of the carbonates having a vein or stockwork structure are of very late-stage origin and appear to have been the last phases to have precipitated from hydrothermal fluids across the Drake Goldfield. These carbonates are likely not related to the main mineralization event. The geochemistry and morphology show both syngenetic/epigenetic relationships between calcite, magnesium-bearing calcite, and magnesite, providing strong evidence for fluctuating Mg/Ca ratios in the fluids during carbonate precipitation and disequilibrium assemblages.

Apart from three outlier values, the carbonates are characterized by a wide range in δ¹⁸O_VSMOW but a relatively narrow range of δ¹³C_VPDB. Carbon and oxygen isotopes of vein carbonates from the Drake Goldfield deposits indicate that the carbon had a mantle-derived source and that this then underwent limited mixing with meteoric waters and fluids associated with low-temperature alteration. Deposits dominated by gold and copper (except for the Red Rock field) show an initial magmatic origin for the late-stage fluid, which then mixed with fluids associated with low-temperature alteration processes during their evolution. In contrast, for the silver dominant deposits and the Red Rock deposit, although also having an initial magmatic origin, the fluids responsible for carbonate formation were more greatly affected by mixing with meteoric water.

Sulfur isotope ratios of sulfide minerals indicate that the sulfur was most likely derived from at least two different sources: a predominantly magmatic source (especially for the gold and copper dominant deposits) and a minor reduced sedimentary source (only for the silver-dominant deposits), the relative importance of each varying between deposits.

The combined carbon, oxygen, and sulfur isotopes clearly show that although the carbonate veins were derived from a wide range of fluids, these were different to those responsible for sulfide deposition. The different carbon, oxygen, and sulfur isotope characteristics from different deposits located across the Drake Goldfield collectively suggest a major heat and metal source below the center of the DQZ, which best explains the spatial distribution of the gold (copper) and silver-dominant deposits. The data also suggest that the Red Rock area may have had a second localized heat source, though this requires further work.

This study also has implications for further exploration for Ag-Au (Cu) and low–intermediate-sulfidation epithermal deposits within the Drake Goldfield. Au (Cu) intermediate-sulfidation deposits mainly occur within the center of the DQZ, while Ag and Au low-sulfidation epithermal deposits mainly occur within the periphery. In general, the Au (Cu) deposits within the Drake Goldfield show a limited variation in values for sulfur isotopes of the main mineralization stage sulfides (mostly close to 0‰ of δ³⁴S) and oxygen isotopes of the carbonates, indicating magmatic dominant fluid sources. In contrast, the sulfur and oxygen isotopes for the Ag deposits have larger variations, suggesting substantial mixing processes during the fluid’s evolution. The carbon and oxygen isotopes for the low-sulfidation epithermal deposits within the Drake Goldfield show an initial magmatic source which then mixed with meteoric fluids, while the intermediate-sulfidation epithermal deposits show an initial magmatic source, which then mixed with fluids associated with low-temperature alteration. In addition, the Ag low-sulfidation epithermal deposits show a greater variation in carbonate isotope values compared to the Au low-sulfidation epithermal deposits, reflecting increased fluid-mixing processes within the Ag low-sulfidation epithermal deposits.