We report sulfur isotopic compositions of sulfides of various paragenetic stages in the giant Sukhoi Log sediment-hosted orogenic Au deposit in Russia. The overall mean value and the significant variability in early pyrite indicate that the sulfur was from the reduction of seawater sulfate. The later generations of sulfide have δ34S values in successively smaller ranges, coincident with the mode that is around the median value of the whole data set. Together with textural evidence, sulfide trace element data, and gold occurrence, the data demonstrate that metamorphism has gradually homogenized the early sulfur, accompanied by the segregation of quartz and the release of Au from the lattice of early pyrite and its reprecipitation as inclusions in later pyrite. The S isotopic compositions of sulfides in Sukhoi Log, and many other major orogenic Au deposits hosted in sedimentary rocks of various ages, show a pattern generally parallel to the seawater sulfate curve through geologic time, indicating that the sulfur in most sediment-hosted orogenic Au deposits was probably also originally from the reduction of seawater sulfate. We conclude that sulfidation and gold mineralization in many sediment-hosted orogenic gold deposits was early during basin evolution when seawater was the principal active fluid, rather than later, during or after basin inversion, as proposed in current models.
Because gold is believed to be mostly transported as bisulfide complexes in orogenic deposits (e.g., Groves et al., 2003; Goldfarb et al., 2005), the source of the sulfur is important to the understanding of the source of Au, the source of the hydrothermal fluids, and the genesis of the deposits. However, the S source of such deposits has been equivocal for many decades. Various hypotheses have been proposed, for example, S introduced by metamorphic fluids from wall rock (e.g., Sangster, 1992; Goldfarb et al., 1997) or from deep crust at amphibolite or granulite facies (e.g., Kerrich and Fryer, 1979; Phillips et al., 1987), S from the mantle depths (e.g., Perring et al., 1987; Barley and Groves, 1990), and S from magmas (e.g., Burrows and Spooner, 1985).
In this paper we present S isotope data from the giant Sukhoi Log orogenic gold deposit, Russia (1100 t Au averaging 2.45 g/t; Distler et al., 2004), and discuss the genesis. Then, by comparing the S isotope pattern of Sukhoi Log with other sediment-hosted orogenic gold deposits in relation to the secular seawater sulfate S isotope curve, we argue that the S from other deposits was probably also from the reduction of seawater sulfate. This conclusion has major implications for the genesis of sediment-hosted orogenic deposits.
SUKHOI LOG GOLD DEPOSIT
Sukhoi Log is within the Patom Highlands orogen (Fig. 1), a Neoproterozoic passive margin sequence that accumulated on the eastern margin of the Siberian craton and was subsequently deformed and metamorphosed at 516 ± 22 Ma (Laverov et al., 2000). The sequence also contains sedimentary exhalative (SEDEX) style base metal deposits, including the Kholodninskoe Pb-Zn deposit (Fig. 1; Distanov et al., 1982). The host rock is an organic-bearing pyritic black shale and siltstone sequence with a maximum sedimentation age of 600 Ma (Meffre et al., 2008). The mineralization occurs in the core of a tight overturned anticline. It consists of discontinuous bedding-parallel quartz-sulfide (dominantly pyrite) segregations and veins, disseminated bedding-parallel pyrite, and minor pyrite nodules, coarse euhedral pyrite, and cross-cutting quartz veins. The mineralized zone forms a shallow-dipping tabular body parallel to the axial plane cleavage, with a high-grade gold zone (>4 g/t) concentrated within the general hinge zone of the fold, where folded quartz-pyrite veinlets are more intensely developed (Distler et al., 2004). The best gold grades are associated with the narrow bedding-parallel quartz-pyrite segregations and veins that have been folded by the main deformation event. No sulfate minerals have been found or reported.
Detailed paragenetic studies of the mineralization and host rocks by Large et al. (2007) defined six stages of pyrite (py) development, including syngenetic framboidal or fine-grained py1, early diagenetic py2, late diagenetic to early metamorphic py3, synmetamorphic to late metamorphic py4, late metamorphic py5, and late-stage py6 (Fig. 2). Typically py4 grains have narrow quartz pressure shadows and there is little py1 in their immediate vicinity. In fold hinge zones there are quartz-pyrite segregations and irregular, discontinuous veins that have been interpreted as coalesced segregations. Pyrite in such segregations and veins (vein pyrite) is typically zoned from core to rim as arsenopyrite → py3 → py4 → py5 (Fig. 2F). Pyrrhotite, chalcopyrite, sphalerite, and arsenopyrite first appear as inclusions in py3, and minor pyrrhotite and arsenopyrite are also associated with py4 and py5.
Large et al. (2007) reported laser ablation–inductively coupled plasma–mass spectrometry analyses of the various pyrite types. The synsedimentary py1 contains the highest levels of lattice-bound Au and As and a suite of trace elements (Mo, Sb, Ni, Co, Se, Te, Ag, Cu, Pb, Zn, Mn, Ba, Cr, U, V) that are similar to those concentrated by organic processes in euxinic sedimentary environments. Later generations of pyrite, from py2 to py5, including vein pyrite, contain progressively lower contents of lattice-bound Au and most other trace elements, but significant inclusions of free gold and gold telluride (Large et al., 2007). Monazite geochronology and Pb-Pb dating of pyrite indicate that the formation of pyrite and introduction and/or remobilization of gold was a multistage process, from sedimentation and/or diagenesis through deformation and metamorphism (Meffre et al., 2008).
SUKHOI LOG S ISOTOPES
Sulfur isotopes were measured using both conventional (±0.3‰, 2σ; Robinson and Kasakabe, 1975) and laser ablation methods (±1.0‰, 2σ; Huston et al., 1995); the results are presented in Figure 3. The 35 analyses of py1 and py2 range from −6.5‰ to +19.3‰, with a median of 10.6‰, and 50% of the data are in a range from 3.4‰ to 12.1‰ (1st and 3rd quartile). Spatially the S isotope compositions of py1 vary greatly over a short distance, e.g., 8.7‰ over 7 cm (Fig. 2A). The 26 analyses of py3 are from −2.7‰ to +16.7‰, with a median of 11.5‰, and 50% of the data are in a range from 7.7‰ to 13.0‰. The range for the 55 py4 analyses is 4.3‰–24.0‰, the median 10.9‰, and 50% of the data range from 9.7‰ to 13.5‰. Although the total range of the py4 δ34S values is only slightly smaller than the earlier pyrites, the spatial variation of S isotopes is much smaller. For example, py4 crystals 9 cm apart have almost identical S isotope compositions (Fig. 2E), and large variations occur only over large distances (>~15 m). We analyzed py5 by the laser ablation method; the range of the 25 measurements is much smaller, 7.1‰–12.6‰. The conventional analyses of drilled powders of vein pyrite, mostly py5, but likely containing py3 and/or py4, yielded a similarly small range but with slightly higher δ34S values, 11.1‰–16.9‰. The S isotope compositions of pyrrhotite (12 analyses) and arsenopyrite (2 analyses) are also concentrated in the 7‰–13‰ range. No correlation between S isotope compositions and stratigraphy or lithology was found. Overall, the 183 S isotope analyses have a large variation, from −6.5‰ to +24.0‰, with the majority in the range of 7‰–17‰, and a median of 10.4‰ (Fig. 3).
COMPARISON WITH OTHER SEDIMENT-HOSTED OROGENIC GOLD DEPOSITS
Sulfur isotope data from Sukhoi Log and 15 other sediment-hosted gold deposits were plotted in Figure 4. The other deposits were selected to cover a large age span. Deposits hosted in volcanic or intrusive rocks were not included in this compilation. The S isotope data were plotted against the sedimentation age of the host rocks, together with the δ34S seawater sulfate curve (Claypool et al., 1980; Strauss, 2004). The statistics of the data are summarized in GSA Data Repository Table DR11.
Figure 4 shows that the variation in S isotope compositions of the deposits through geologic time has a pattern similar to the seawater sulfate curve. The δ34S values of the sulfides are typically ~15‰–20‰ lighter than the coeval seawater sulfate, though the difference can be smaller (~10‰) or larger (~30‰).
DISCUSSION AND CONCLUSIONS
Source of Sulfur in the Sukhoi Log Deposit
The S isotope compositions of early sulfides (syngenetic and diagenetic: py1 and py2) in the Sukhoi Log deposit have large variations, in terms of both total range (~26‰) and stratigraphic variation (e.g., ~9‰ change over 7 cm). Such large variations preclude the possibility that the S is from either a magmatic source (Ohmoto, 1986), or a homogenized deep metamorphic source (e.g., Groves et al., 2003). Instead, the S isotope data indicate that the S is sourced from seawater sulfate during sedimentation and diagenesis, consistent with the textural evidence of a syngenetic origin (Large et al., 2007). This interpretation is further supported by the similarity in δ34S ranges and medians of the data sets at Sukhoi Log (−6.5‰ to +24.0‰, median 10.4‰) and at the nearby, and age equivalent, Kholodniskoe SEDEX Zn-Pb-Ag deposit (range −9.5‰ to +25.4‰, median 9.5‰; n = 269: Distanov et al., 1988). Typically, S in SEDEX deposits is interpreted to be from seawater (e.g., Goodfellow et al., 1993).
The seawater sulfate could have been reduced by several processes, or a combination of them. Bacterial sulfate reduction can explain the large variations (e.g., Ohmoto, 1986); experiments have shown that biogenic sulfides can have δ34S values 2‰–46‰ lighter than the seawater sulfate (Detmers et al., 2001, and references therein). Abiotic thermochemical reduction of sulfate can cause a fractionation of 10‰–25‰ (Harrison and Thode, 1957; Seal, 2006, and references therein). At Sukhoi Log, the framboidal texture of some pyrite may indicate a bacterial origin, whereas the presence of an abnormal amount of Au-bearing syngenetic pyrite may be the result of synsedimentary hydrothermal activity, as proposed by Large et al. (2007). Therefore it is possible that both bacterial reduction and abiotic reduction related to synsedimentary hydrothermal activity have played a role in fixing the S, Au, and As into the sediments as gold-bearing arsenian pyrite.
Homogenization of Sulfur Related to Metamorphism and Au Concentration
From early synsedimentary to late metamorphic, the different generations of pyrite at Sukhoi Log have successively narrower ranges of δ34S values, whereas the median values stay similar (Fig. 3). Although the overall ranges for py3 and py4 are only slightly narrower than those of py1 and py2, the spatial variation is much smaller and the 1st to 3rd quartile ranges narrower. The latest metamorphic pyrite (py5) and pyrrhotite have δ34S values in a small range (7.1‰–12.6‰, 8.2‰–12.2‰), concentrating around the median value of the whole data set (10.4‰). Their total ranges are similar to the mode of all sulfide data at Sukhoi Log (7‰–13‰; Fig. 3). We interpret this trend as the result of gradual homogenization of S during metamorphism.
The increasing effect of metamorphism on py1 to py5 was demonstrated by Large et al. (2007) to also coincide with the evolution of Au occurrence from purely lattice-bound Au in synsedimentary pyrite (py1) to totally free Au as inclusions in py5 and quartz-pyrite segregations. This indicates that the homogenization of S isotopes is accompanied by the remobilization and concentration of Au. In fact, py5, the only pyrite type that exhibits significant S isotope homogenization at the deposit scale, is mostly hosted in quartz-sulfide segregations and veinlets that bear the highest Au grades (as free gold) in the deposit.
In summary, sulfur, arsenic and gold were introduced into the sedimentary rock as syngenetic and diagenetic gold-bearing arsenian pyrite, with the sulfur produced by a mixture of both hydrothermal reduction and bacterial reduction of seawater sulfate. The original S in pyrite was gradually homogenized during metamorphism, together with the release of invisible gold and other elements from the early syngenetic and diagenetic pyrite, resulting in the formation of free gold in the later metamorphic pyrite. Significant S isotope homogenization only occurred late in the history of the deposit during maximum deformation and folding, when py5 overgrew the earlier pyrite stages and quartz segregations developed in pressure shadows.
Sulfur Source of Sediment-Hosted Orogenic Au Deposits and its Implications
The fact that the S isotope compositions of many orogenic sediment-hosted deposits vary in a parallel fashion to the seawater sulfate curve (Fig. 4) indicates that the sulfur in the deposits was derived by reduction of seawater sulfate. A similar pattern has previously been reported for volcanic-hosted massive sulfide (VHMS) and SEDEX Zn-Pb-Cu deposits (Sangster, 1968; Large, 1992; Goodfellow et al., 1993; Huston, 1999), and remains one of the strongest arguments that the S in these deposits is principally reduced seawater sulfate rather than magmatic or metamorphic sulfur. The sulfide δ34S values in most sediment-hosted gold deposits have a mode consistently ~15‰–20‰ lighter than the coeval sulfates (Fig. 4), again similar to base metal SEDEX and VHMS deposits, indicating that at least part of the reduction was abiotic and hydrothermal (Δ34Ssulfate-sulfide = 10‰–25‰; Harrison and Thode, 1957), especially in deposits hosted in Phanerozoic sediments. If there were no hydrothermal reduction, the sulfate-sulfide S isotope fractionation would be >40‰, commonly ~60‰, in Phanerozoic sedimentary rocks (Strauss, 1997). Such a large fractionation is believed to be caused by initial bacterial sulfate reduction (Δ34Ssulfate-sulfide = 2‰–46‰; e.g., Habicht and Canfield, 1997) and subsequent disproportionation reactions (Δ34 Ssulfate-sulfide = 7‰–20‰; e.g., Bottcher et al., 2001). On the other hand, the large δ34S variations present in some sediment-hosted gold deposits were probably caused by bacterial reduction in addition to the hydrothermal reduction. Goldfarb et al. (1997) also noticed that δ34S of sulfides in orogenic Au deposits in Alaska has a temporal varia tion pattern similar to that of seawater sulfate, on the basis of which they proposed that the S source reservoir of the Alaska deposits was the ore-hosting sediments. Our study argues further that the S in similar sediment-hosted deposits worldwide is ultimately from seawater, with S and Au both deposited during sea-floor hydrothermal activity. The later homogenization of S and concentration of Au during metamorphism are supported by sulfide textures and trace element characteristics.
There is general consensus that Au deposition in orogenic gold deposits is by a process of sulfidation of preexisting Fe minerals such as Fe carbonates, magnetite, or Fe silicates (e.g., Phillips and Groves, 1984; Groves et al., 2003; Goldfarb et al., 2005). This implies that the Au and S are transported together in the ore fluid, and added to the rock during the ore-forming process. The relationship between S isotopes in gold-ore pyrite and the seawater sulfate curve through time revealed here is compelling evidence that the ore fluid for Sukhoi Log and most other sediment-hosted orogenic deposits contained a major proportion of seawater S. The sulfidation process responsible for gold mineralization was thus a very early event during the basin history, when contemporaneous seawater was an active crustal fluid. This result is contrary to current theories on orogenic gold deposits that require a very late timing for the sulfidation and gold mineralization process, usually during deformation and metamorphism or sometimes after metamorphism (e.g., Groves et al., 2000; Hagemann and Cassidy, 2000). We still consider the role of metamorphism to be important in producing economic deposits, due to the associated recrystallization of early syngenetic and dia genetic gold-bearing pyrite, which release invisible gold as free gold, to become upgraded and coarsened in grain size in structural sites such as anticlines and shear zones.
This research was supported by the Australia Research Council through the Centre of Excellence program. We appreciate the critical and constructive reviews of Dave Craw, Daniel Kontak, Robert Seal, Mike Solomon, and Dave Cook.