—In the present paper we demonstrate that most sulfides of the studied deposits of the Archean Sumozero–Kenozero greenstone belt within the Karelian Craton on the Fennoscandian Shield have nonzero Δ33S values. This indicates that proportions of seawater sulfate and elemental sulfur in Mesoarchean, included into the ores and resulting from UV photolysis, are different. Our results show that systematics of sulfur isotopes of sulfides generally reflects the mixing of mass-independently fractionated sulfur reservoirs with positive and negative Δ33S values. Pyrite is depleted in 34S isotope, which was interpreted as evidence for microbial sulfate reduction. Variations in the positive Δ33S anomalies of the Leksa deposit and the general tendency for Δ33S sulfide content to increase with stratigraphic levels in certain boreholes most likely reflect the change in temperature and the fluid mixing throughout the life of the hydrothermal system. The presence of sulfides with strongly negative Δ33S anomalies suggests that atmospheric sulfur and seawater sulfate, rather than volcanic sulfur, were the prevailing source for mineral systems of the studied deposits. The presented data require the Mesoarchean seawater to contain sulfates at least locally.

Volcano-sedimentary massive sulfide (VHMS or VMS) deposits are strata bound accumulations of precipitated sulfide minerals and represent major sources of Cu, Pb, Zn, Ag, and Au (Barrie and Hannington, 1999; Allen and Weihed, 2002; Ross and Mercier-Langevin, 2014). Volcano-sedimentary massive sulfide deposits formed starting from the Archean eon and up to the present (Smirnov, 1976; Groves et al., 2005). They are considered to be related genetically to the undersea volcanic activity being precipitates from a mixture of cold seawater and ascending hot metal-laden hydrothermal solutions at or near the seawater-rock boundary (Ohmoto and Rye, 1974; Smirnov, 1976; Solomon and Walshe, 1979; Lister, 1980; Ohmoto et al., 1983; Lisitsyn et al., 1992; Grichuk, 2000).

Metal-rich fluids, circulating in ore-magmatic systems and forming sulfide deposits, have very high S concentrations. In these fluids, S is in different oxidation states (from S2– to S6+) and phases (minerals, liquids and gases), and participates in many chemical reactions, which leads to isotopic fractionation between coexisting S-bearing minerals and is good for geochemical research (Grinenko and Grinenko, 1974; Grichuk and Lein, 1991; Butler et al., 1998; Seal, 2006; Schauble, 2008). S-isotopes bear valuable information about the source of the material that participated in the ore formation process, and about its physical and chemical evolution (for example, about changes in pH, oxidation-reduction potential and temperature).

During deposit formation, sulfur can come from different sources, such as igneous and sedimentary rocks, atmosphere, seawater. The MDF-S can provide data on geochemical reservoirs as sources of sulfur, and certain geochemical processes (based on δ34S variations) (Herzig et al., 1998; Bogdanov et al., 2002; Bortnikov et al., 2003; Lein et al., 2003; Vikent’ev, 2004). However, δ34S alone is not enough to make a comprehensive description of dynamic mineral systems, in which chemical processes, such as dissolution, sedimentation, liquid phase decomposition, and oxidation-reduction reactions are varying parameters regulating mobilization, transport and metal deposition (Hutchison et al., 2020). It is often difficult to identify which of these processes is responsible for isotopic variability.

The discovery and theoretical understanding of mass-independent fractionation (MIF-S) of minor sulfur isotopes, 33S and 36S, (Farquhar et al., 2000; Farquhar and Wing, 2003) provided a tool for studying sulfur sources in Archean ores (Bekker et al., 2009; Johnston, 2011; Selvaraja et al., 2017; Vysotskiy et al., 2019). This is very important for understanding the sulfur cycle in the Archean eon, the origin of different microbial metabolisms, and the chemistry of hydrothermal and surface waters (Bekker et al., 2009; Golding et al., 2011; Selvaraja et al., 2017).

MIF-S (based on Δ33S variations) is a chemically conservative indicator showing deviations from mass-independent fractionation processes. MIF-S is a mark, transmitted to sulfur molecules that photochemically interacted with UV rays in the atmosphere. Before the increase in oxygen concentration during the Great Oxygenation Event (GOE) around 2.4 Ga, sulfur with such a mark had been accumulated in sedimentary rocks (Farquhar et al., 2000). Consequently, MIF-S is a process that occurs in the atmosphere but it also is an isotopic mark preserved in the only primary sulfur reservoir – Archean supracrustal rocks, where photolytic sulfur was concentrated and accumulated in sulfides and sulfates.

The Δ33S mark is not influenced by dynamic geochemical processes, it can only be diluted (Vysotskiy et al., 2019). Thus, a combination of δ34S and Δ33S in a mineral system enables us to identify both sulfur sources and varying chemical parameters of the environment (e.g., pH, P, T, fO2, fS2). This is very important as the tracking of changes in chemical parameters of the fluid that contains metals is the key to understanding the processes controlling their sedimentation (Ohmoto, 1986).

In this paper, we discuss the processes that influenced the formation of Mesoarchean (~2.9 Ga) sulfide deposits and were identified by multi-isotope characteristics of sulfur. The studied deposits are located in the Archean Sumozero–Kenozero greenstone belt of the Karelian Craton of the Fennoscandian shield (Kulikov et al., 2017). The deposits subsequently underwent a slight tectonic deformation with maximal metamorphic grade not exceeding greenschist facies (Kuleshevich, 1992). Some of them have well preserved primary textures and potential primary biota (Vysotskii et al., 2019).

Geologic position of the deposit

VHMS ore deposits used for this study are located in the Kamennoozero structure of the Archean Sumozero–Kenozero greenstone belt which occupies the southeastern part of the Karelian Craton (Fig. 1a). The oldest greenstone complexes (3.1–2.9 Ga) are found within the Vodlozero terrain of the Craton (Slabunov et al., 2006; Kulikov et al., 2017). The Sumozero–Kenozero greenstone belt being ~400 km in length and up to 50 km in width is among the largest structure of such class (Puchtel et al., 1999; Glebovitskii, 2005; Kuleshevich et al., 2005; Slabunov et al., 2006). It represents a system of tectonic sheets (Fig. 1b) composed of weakly altered volcanogenic, volcano-sedimentary, and intrusive formations overlaying tonalite-trondhjemite-granodiorite (TTG) gneisses-granites and partially overlapped by Paleoproterozoic rocks (Kuleshevich et al., 2005). The Kamennoozero structure (Fig. 1b) consists of two tectonically superposed units (Puchtel et al., 1999). The lowermost (basalt-komatiite) unit is a combination of metamorphosed, amygdaloidal and pillow, predominantly tholeiitic basalts and rare komatiites flows retaining spinifex textures which are host rocks for Cu–Ni sulfide ores, and thin horizons of carbon-containing slates, chemogenic quartzites, and carbonate–quartz–sericite schists. The upper basalt–andesite–rhyolite–dacite (BARD) unit is made up of metamorphosed tuffogenic-sedimentary rocks and lavas of rhyolites, dacites, andesites and andesibasalts. Metabasalts with a pillow texture are also common, interlayers of carbon-containing slates are found, as well as subvolcanic bodies of the adakite series rhyolites. A characteristic feature of this unit is that it hosts horizons of massive sulfide ores, which have become objects of the present research (Fig. 1b).

A geochemical study of volcanic rocks of the structure has shown that the units formed in different tectonic settings (Puchtel et al., 1999). The basalt–komatiite lower sequence originated in the seafloor conditions and is regarded as a fragment of an oceanic plateau aged ~2.9 Ga: the Sm–Nb isochron showed an age of 2916 ± 117 Ma, and Pb–Pb dating of metabasalt showed an age of 2892 ± 130 Ma (Puchtel et al., 1999).

The formation of felsic volcanic BARD- and adakite-series rocks of the upper unit is associated with subduction processes in an island-arc setting. ID-TIMS dating of zircon from the rhyolites yielded U–Pb ages 2875 ± 5 and 2876 ± 5 Ma of their emplacement, respectively (Puchtel et al., 1999). VMS ores as well as carboniferous sediments (schists) are closely related to intermediate-felsic island-arc volcanic rocks of the Kamennoozero structure (Fig. 1b), suggesting that they were deposited underwater in back-arc basins of the island-arc system.

Therefore, massive sulfide mineralization is characterized by the concordance with bedding and constant association with island-arc volcano-sedimentary, mainly BARD-series facies. All the massive sulfide deposits typically form conformable sheet-like lodes several centimeters to tens of meters thick at different stratigraphic levels. The mineral composition of the ores is dominated by pyrite and pyrrhotite with subordinate nonferrous metal sulfides (Rybakov, 1987; Kuleshevich, 1992; Rubakov and Golubev, 1999).

Brief description of deposits

At the end of the last century, over 10 sulfide deposits were explored in the Kamennoozero structure by exploratory drilling (Fig. 1b). Based on the lithological classification of host rocks (Barrie and Hannington, 1999; Franklin et al., 2005), we classify these deposits as the bimodal-mafic type VHMS. The volcanic rocks of this deposit type are dominated by basalt, andesite, and felsic flows and pyroclastic rocks. These deposits mark the large hydrothermal vents fields commonly associated with back-arc volcanoes and spreading centers (Hannington et al., 2005).

Research samples were collected from drill-cores within the Leksa, Zolotye Porogi, Central Vojma, and Northern Vojma deposits (Fig. 1b). The studied deposits formed underwater at different distances from zones of active hydrothermal activity (vents).

The Leksa deposit represents two lenses of pyritic ores (30–70% of sulfides) up to 50 m thick bedded within a weakly dislocated series of quartz–albite–sericite and carbon-containing shales that underwent greenschist-facies metamorphism (Kuleshevich, 1992). Mafic lavas were not found during drilling. The ores themselves repeat the weakly dislocated pattern of the host rocks but retain their primary structure.

The ores are characterized by a major-mineral assemblage dominated by pyrite (much less frequently pyrrhotite or marcasite), other sulfide minerals are present in trace amounts, polymetallic concentrations at the level of background values (Kuleshevich and Belashev, 1998). The ore of the Leksa deposit occurs as banded to disseminated-banded, massive and disseminated-streaky pyritic seams. They are weakly metamorphosed, as evidenced by the high porosity of the ores, low strength of the samples, and texture heterogeneity on the thin section scale. Massive ores contain primary textures such as idiomorphic crystals pyrite, framboidal and botryoidal pyrite, laminated nodules, etc. In some samples, exhalative (chemical) sedimentary rocks (silicagel) are found here. Considering these data and the metal zoning established in many VHMS deposits (Eldridge et al., 1983; Hannington et al., 1999; Goodfellow and McCutcheon, 2003), a relatively low-temperature sulfide ore (bedded pyrite), distal from the venting zone, was drilled in the Leksa deposit.

The structure of the Zolotye Porogi deposit comprises three strata (Kuleshevich et al., 2005). The lower stratum is composed of basalts, komatiites, tuffs and overlying chlorite-talc-carbonate schists. The average is represented by volcano-sedimentary rocks of acidic-intermediate composition, carbon-containing shales, quartzites, VHMS ore. The upper stratum consists of andesite-basalts, basalts, tuffs and overlying schists.

The ore of the Zolotye Porogi deposit occurs as banded to disseminated-banded, massive and disseminated-streaky pyritic seams (sulfide content 30–80%) of variable (up to 50 m) thickness intercalated within carbon-containing shales and quartzites (Kuleshevich, 1992, 2005; Kuleshevich and Belashev, 1998). The ore mineral composition is dominated by pyrite but also contains pyrrhotite (5–10%), chalcopyrite and sphalerite (2–3%), arsenopyrite (up to 3–5% in the zones of As–Sb superimposed mineralization), tetrahedrite, and other Sb-bearing minerals. The content of polymetals in ores (Ni, Co, As, Sb) is insignificant, although it is higher than in the Leksa deposit. Probably, the ores of the Zolotye Porogi deposit formed closer to the venting zone and at a higher temperature.

The Central Vojma ore occurrence is confined to a lens of carbon-bearing and albite-quartz-sericite-chlorite schists squeezed between serpentinized ultramafic rocks, and the western and eastern sections of the Vojma massif (Kuleshevich et al., 2005). Mineralization in the slates is represented by layers of pyrite-rich finely disseminated, disseminated, massive and banded ores with well-preserved stratified structure and a thickness varying from several to hundred meters. The rocks are mostly composed of pyrite or pyrrhotite–pyrite with a small amount of sphalerite, chalcopyrite, and scarce galena (sporadic grains). The average content of polymetals is low (Co ~ 0.003, Ni ~ 0.02, Cu ~ 0.038, Zn ~ 0.13%).

Based on the analysis of the structural position of the deposit, mineral paragenesis, and zonality of metals (Kuleshevich, 1992; Kuleshevich and Belashov, 1998), we believe that the Central Vojma deposit was formed in a rift valley near the hydrothermal venting zone.

The Northern Vojma ore occurrence is located north of the Vojma ultramafic massif and is confined to a NE-trending fault cutting chlorite-sericite host schists (Kuleshevich et al., 2005). The ore body can be traced for over 800 km along strike and 400 m down dip at the angle of 65° and is 5 to 30 m thick. Its uppermost part has massive brecciated structure, whereas lower and remote horizons exhibit streaky disseminated and disseminated banded structure semi-conformable to the banded structure of the host rocks. The massive sulfides are compact, fine-grained aggregates of intergrown sulfide minerals with irregular grain boundaries. Pyrite forms cubic crystals split and cemented by chalcopyrite and sphalerite. The ore body core is composed of sphalerite–chalcopyrite–pyrrhotite, with galena and subordinate arsenopyrite, cobaltite, Bi-containing minerals (bismuth, bismuthite), naumannite, Ag–Au association, copper gold, bogdanovichite [AgBiSe2–AgBi(Se,S)2], weibullite [Pb6Bi8(Se,S)18], tetrahedrite and accessory minerals including cassiterite, monazite, and xenotime, the presence of which emphasizes the relation to acid magmatism. The content of zinc in ores is up to 12.1%, copper – up to 14.0%, gold (0.1 to 1.2 ppm) and silver (5.0 to 167.0 ppm). Predicted resources of zinc are estimated at 133.7 thousand tons, copper – 15.1 thousand tons (http://nedrark.karelia.ru/mnia/pb_zn_karelia.htm). Analyses of the structural position of the deposit, mineral paragenesis, and metal zonality indicate that the Northern Vojma deposit formed in the active hydrothermal vent zone of a rift valley.

Minerals have been analyzed at the Laboratory of X-Ray Methods of the Analytical Center of the Far East Geological Institute, Far Eastern Branch of the Russian Academy of Sciences (FEGI FEB RAS) in Vladivostok, using a JEOL JXA 8100 (Japan) electron microprobe equipped with three wave and one energy dispersive spectrometers (Oxford Instruments Inca, UK), under a resolution of 137eV MnKα, an accelerating voltage of 20 kV and a measure current of 1·10–8 A. The volume of analyzed material at point measurements ranged between 1 and 3 μm. Microphotographs have been obtained via scanning the backscattered electrons. Pure metals, glasses and minerals analyzed by other methods as well as MAC Calibration Standards have been used for reference. In calculations, total Fe was equivalent to Fe+2.

Minerals and microfossils images have been obtained at the Laboratory of Micro and Nano research of the FEGI FEB RAS Analytical Center using a TESCAN LYRA 3 XMH (Schottky barrier diode) double-beam scanning electron microscope with the Oxford Instrument’s EDS software AztecEnergy.

Sulfur isotope analyses have been carried out at the Laboratory of Stable Isotopes of the FEGI FEB RAS Analytical Center following the local laser method described in (Ignatiev et al., 2018). Sulfur isotope ratios have been measured at mass values of 127, 128, and 128 (S32F5+, S33F5+, and S34F5+) in a triple-inlet mode using a MAT 253 isotope ratio mass spectrometer. The δ33S and δ34S data for sulfide inclusions with spatial resolution of 100 μm are reported relative to the V-CDT standard, with 1σ errors of ± 0.20 and ±0.15‰ for δ33S and δ34S, respectively, and no more than ±0.050‰ for Δ33S. Isotopic relations in the samples have been measured relative to SF6 gas standard, calibrated according to the IAEA-S-1, IAEA-S-2, IAEA-S-3, and NBS-123 international standard. The reproducibility of results (1σ) of the repeated analyses of the IAEA-S-1 international standard is 0.15, and 0.02‰ for δ34S and Δ33S, respectively, the diameter of the ablation crater is 100 μm and its depth is 40 μm.

Sulfide mineralization features

The dominant sulfide ore mineral in the studied VHMS deposits is pyrite or pyrrhotite. The following most common ore minerals, chalcopyrite, sphalerite, and galena, are found in various quantities, but they are not found anywhere in concentrations that exceed the content of iron sulfide. Their maximum concentrations were found in the North Vojma, and Central Vojma deposits.

The change in mineral paragenesis in space and time enables to determine the location of the VHMS relative to the venting zone. The paragenesis of the ore minerals of the described deposits (Kuleshevich, 1992, 2005) is consistent with the generally accepted pattern of zoning of the VHMS ores (Galley et al., 2007), which relates to a change in fluid temperature. Weakly dislocated ores are characterized by globular sulfide aggregates and oolite concretions from thinbladed crystals of marcasite and pyrite. We have found two main associations of sedimentary iron disulfide: syngenetic pyrite-marcasite concretions, which are deemed to have formed at the sediment-water interface, and diagenetic pyrite, the formation of which is related to the sediments beneath this boundary.

The syngenetic pyrite association is presented by colloform aggregates (crusts), globular aggregates and laminated concretions, oval, and ellipsoidal morphology and clear concentric zoning (Fig. 2). This zoning is the result of alternation of massive pyrite and thin-platy marcasite that often makes up stellar bunches. Marcasite crystals are embedded in a compound matrix of chlorite (chamosite), sericite (muscovite) and quartz, sometimes with a little carbon. Small oolites lack zoning and are entirely composed of lamellar and needle-shaped crystals of marcasite. X-ray structural analysis results showed the presence of pyrite and marcasite in concretions and coatings.

The diagenetic pyrite association contains framboidal pyrite, small euhedral (idiomorphic) crystals and microspherical pyrite.

Euhedral (idiomorphic) pyrite includes scattered isometric microcrystals, small intergrowths and veinlets in the host rocks (Fig. 3a, b, f), as well as rims surrounding the oolites.

Framboidal pyrite represents relatively compact (50–200 μm and larger), tightly-packed aggregates composed of small (2–3 μm) isometric microcrystals approximately equal in size and morphology (Fig. 3). They represent pyrite shells 1–2 to 10 μm in diameter, rarely larger, round and subspherical in shape (Fig. 3). The shells are thick-walled with denticulate sculpture (Fig. 3b, d, e), sometimes tubular (Fig. 3j). Both single forms and accumulations are found. Each individual differs in shape, size, and position. According to microprobe analysis, the core inside the shells is composed of silicate minerals – quartz, chlorite (chamosite), muscovite, or their mixture.

Historically these structures have been called “mineralized bacteria”. These structures were interpreted as pyritized bacterial colonies or remains of primitive life forms (Ramdohr, 1958; Schidlowski, 1965; Saager, 1970; Russell and Hall, 2006; Schopf et al., 2018). In our case certain microobjects can morphologically be compared with silicate tubular structures described in Mesoarchean silicites (Medvedev et al., 2014).

Sulfur isotope geochemistry

A total of 153 δ33S and δ34S isotope analyses have been conducted for sulfide minerals from rocks within the Leksa, Zolotye Porogi, Central Vojma, and Northern Vojma deposits.

These data are presented in Fig. 4. The sulfide texture and associated variation of the S isotopes within individual samples are summarized in Table 1. The difference between the measured values of δ33S and the expected values calculated from δ34S is presented as Δ33S, which has been calculated by using the following equation:

Δ33S=δ33S(1000·((1+δ34S/1000)0.5151)).

Among the studied deposits, there are two types (Table 1). One of them (Leksa deposit) is characterized by variable δ34S values (−10.2 to +27.5‰) and positive Δ33S values (up to +2.64‰). Positive Δ33S values imply that sulfur was largely derived from a photochemical elemental sulfur reservoir (Fig. 4).

The widest range of δ34S values has been measured for concretions and crusts (Fig. 4a) of the Leksa deposit. Local laser method allowed establishing that δ34S values within one concretion range from –9.8 to +27.5‰ (Table 1), and isotopic composition of sulfur from pyritic interlayers is lighter than that from marcasitic ones. Additionally, a weak signal of δ33S anomaly (Δ33S ≈ 0.4‰) is detected in the middle part of the zonal concretion, while in other parts it disappeared. Such a range of δ34S values may be considered as a result of sulfate-reducing bacteria live activity.

In contrast to concretions, idiomorphic pyrite crystals show a minor variation of δ34S, from 5.2 to 7.0‰. The study has revealed abnormally high δ33S content, its magnitude being Δ33S = +2.64‰, which points to the fact that these sulfides contain sulfur that was involved in photochemical transformations in an anoxic Archean atmosphere. The δ34S and δ33S values falling on a MIF line with a slope of Δ33S/δ34S ≈ 1 indicate the involvement of sulfur that underwent a sulfur cycle in that atmosphere (the so-called Archean reference array, ARA: Ono et al., 2003; Kaufman et al., 2007; Philippot et al., 2012).

The second type has a narrow range of δ34S values (~0 ± 4‰) and small negative Δ33S values (~–0.5 ± 0.25‰). Negative Δ33S values imply that sulfur was derived from a photochemical sulfate sulfur reservoir. These are the Zolotye Porogi, Northern Vojma, and Central Vojma deposits.

The isotopic composition of these ores is characterized by a sustainably heavier composition of δ34S and δ33S for pyrite relative to coexistent pyrrhotite, arsenopyrite and sphalerite, with magnitude of Δ33S remaining nearly the same. Figurative points of the δ34S/δ33S ratio measured for minerals in one sample show linear dependence (Fig. 4b), and form several trends subparallel to the y = 0.515x line, pointing to mass-dependent fractionation conditions during crystallization of sulfide minerals.

Thus, although during mineral formation sulfur isotope fractionation follows the mass-dependent fractionation law, the sulfur involved in this process already underwent transformation in the atmosphere, and bears an associated mark. The amount of this “marked” sulfur varies in different samples, and at different levels of ore bodies, reflecting the dynamics of hydrothermal mineralization.

The δ34S values for pyrite are ~2.9–3.5‰ higher than for other sulfides. In equilibrium fractionation it corresponds to temperature interval of 70–100 °C (Kajiwra and Krouse, 1971) and muscovite–chlorite facies of epigenesist (Drits and Kossovskaya, 1991), but not to greenschist-facies metamorphism. However, mineralogical thermometers show that sulfide ores crystallized at 200–400 °C (Kuleshevich, 1992; Kuleshevich and Belashev, 1998). It may probably be just a primary signature of coexisting non-equilibrium sulfide minerals that precipitate at different temperatures due to different solubility (Fouquet et al., 1997; Keith et al., 2014, 2016).

The S isotopic composition of sulfides in the described deposits varies greatly both in core samples from wells and in thin sections (Figs. 45; Table 1). This is due to a change in the formation conditions of ore minerals. We have identified several likely scenarios for the behavior of sulfur under different conditions. On the one hand, these scenarios have a local character typical for concrete parts of the studied deposits. However, they can be of a greater importance, if at different deposits sulfides form under similar conditions.

The first scenario (Fig. 5a) is characterized by variable values of δ34S (from –10.2 to +27.5‰) and Δ33S (from –0.27 to +0.44‰). Pyrite and pyrite–marcasite concretions and coatings of the Leksa deposit are associated with this type. Their crystallization took place simultaneously with the formation of sediments from the colloidal solutions at seawater-sediment boundary. They formed in a depositional environment dominated by clay- to silt-sized clastic sediments that, during growth, captured some detritus (quartz, clay minerals – now metamorphosed to chamosite and muscovite), and carbonaceous matter. The concretions and coatings have a layered texture of alternating massive pyrite and reticulate marcasite in the form of needles and thin laminae with inclusions of chamosite, sericite and quartz. A periodic change in crystallization conditions leads to the formation of either pyrite or marcasite in a growing concretion or coating, thus increasing it or creating the observed lamination. This lamination reflects the change in physical-chemical conditions of the environment, specifically in the pH of the seawater. It has become known from experiments that pyrite and marcasite required different pH for their precipitation. Murowchick and Barnes (1986), Schoonen and Barnes (1991a, b), and Benning et al. (2000) have calculated the optimal pH for marcasite, between 4 and 5, which is lower than the typical pH of pyrite. At the same time, within the black shales, during the interaction of sediments with interstitial water, the growth of euhedral, framboidal and microspherical pyrite occurred – the pH of solutions inside the sediments changed towards less acidic, rather neutral values.

The study of the matrix of the black shales using SEM and XEDS analytical techniques showed mineral association of quartz, chamosite and muscovite akin to the minerals composing the silicate matrix of marcasite domains in the layered concretions. Chamosite and muscovite are mineral phases formed under reduction conditions when neutral pH dominated in the environment at the time of their formation (Merino et al., 1989).

Nonpermanent isotopic composition of sulfur in pyritemarcasite concretions and coatings indicates that microorganisms were actively involved in the process of their formation (Shanks, 2001). This is supported by findings of the presumed pyritized microfossils in black shales (Fig. 3). It is quite possible that fluctuation of the pH created favorable conditions for the growth of certain minerals, and controlled the growth of sulfate-reducing bacteria. This assumption is confirmed by the ratio of S isotopes in the concretions, where pyrite layers have a lighter isotopic composition compared to the marcasite ones (Fig. 5a). Sulfur isotope fractionation is probably associated with the process of biogenic sulfate reduction. In this case, the magnitude of isotope fractionation can reach values of 60–70‰ relative to the seawater sulfate (Canfield et al., 2010). It can be assumed that sulfur-reducing bacteria were quite active as indicated not only by the magnitude of δ34S (~37‰), but also by the almost coming-to-naught values of Δ33S anomaly (from Δ33S = 0.44 to Δ33S ≈ 0). Preservation of primary MIF-S signal in sedimentary sulfides is possible only in case of non-intensive biological sulfur cycle; the opposite leads to isotopic homogeneity of sulfides due to microbial reduction of sulfur compounds in the seawater (Halevy, 2013). We consider thus that biological sulfate reduction played an essential role in the formation of Mesoarchean massive sulfide deposits.

The second scenario (Fig. 5a) is characterized by a narrow range of δ34S values (4.6~7.0‰) and positive Δ33S (from +1.55 to +2.64‰). This scenario is only observed at the Leksa deposit for the late pyrite forming rims round the concretions and separate idiomorphic crystals. Positive Δ33S values mean that sulfur came from a photochemical reservoir of elemental sulfur. Here, pyrite crystallized together with quartz. The presence of quartz points to a weak acidic environment that caused a massive crystallization of pyrite, in the composition of which atmospheric elemental sulfur played a significant role. This is possibly related to an inflow of large portions of hot hydrothermal solutions into a colder silt, at the periphery of the hydrothermal field. The ore body is overlapped by schist strata (presumably felsic tuffs at the beginning), which, possibly, points to an active explosive volcanic activity during this period.

The third scenario of sulfur isotopes behavior has been revealed in the ores of the Zolotye Porogi deposit (Fig. 5b). The change in sulfur isotope ratios during crystallization of pyrite concretions at the Zolotye Porogi deposit clearly shows the pulse-like character of atmospheric sulfur inflow into the basin and its connection to acidic volcanism. When volcanic activity was weak, sulfides and a notable amount of light sulfur isotopes began to grow (Philippot et al., 2012). In this case points of the compositions of pyrite sulfur are drawn to the area of mixing of sulfate aerosol and volcanic sulfate along the APA line. During acidic explosions a large amount of solid matter erupted, which obstructed UV radiation leakage (Philippot et al., 2012). A larger quantity of sulfur and heavy isotopes got into the ocean. In this case points of the compositions of pyrite sulfur form a field elongated along the FVA line, which indicates that sulfur came from three sources (pools) – seawater sulfates, aerosol, and magmatic sulfur. The beginning of the field is on the line of mixing of sulfate aerosol with seawater close to the Archean baryte fields. Further on, figurative points of pyrite compositions trend towards the volcanic sulfur field. This, probably, reflects an inflow of a large amount of sulfur resulting from an acidic explosion.

The fourth scenario relates to the infinite reservoir of seawater sulfate and, probably, reflects very deep-level conditions. For these processes a significant amount of sulfur needed for sulfide formation was brought from seawater (open or pore) in the form of sulfate (Fig. 5c).

It is realized at the Central Vojma and Northern Vojma deposits, the ores of which are characterized by a narrow range of δ34S values (~0 ± 4‰) and negative Δ33S values (~–0.5 ± 0.25‰). If data on δ34S were studied separately, a direct interpretation would allow assuming a purely magmatic source (Huston et al., 2010). However, a multi-isotope analysis of sulfur in sulfide ores clearly shows the presence of anomalous 33S/32S ratios (Fig. 5c), indicating the involvement of surface sulfur in the process of mineralization. Negative Δ33S values mean that during sulfide formation some amount of sulfur was obtained due to reduction of the seawater sulfate that was brought, among other sources, from atmospheric photochemical sulfate reservoir (Farquhar et al., 2001; Ono et al., 2003). At the same time, a relevant change in sulfur isotope ratios related to the mineral crystallization temperature is observed in the ores. Sulfides with different composition (pyrite, pyrrhotite, galena, etc.) have different ratios of δ34S, δ33S, although Δ33S value remains constant. This indicates that, despite the crystallization temperature for these minerals being different, the input ratios of different sulfur sources were the same.

Numerous multi-isotope studies of sulfur ores, host rocks and sediments provided basis for studying the inputs of sulfur of the Archean atmosphere and hydrosphere at the VHMS deposits (Bekker et al., 2009; Johnston, 2011; Guy et al., 2012; Sharman et al., 2015). Moreover, the obtained knowledge can help identify a deposit’s position in the oreforming system (Fig. 6).

The Leksa and Zolotye Porogi deposits probably formed on the slope of a central-type volcano in the zone of volcanism similar to that of modern island arcs (Fig. 6a). Host rocks are represented by metamorphosed volcanic and volcano-sedimentary rocks. The ores are associated with quartz-albite-sericite, and carbon-containing slates, which points out the proximity of the strata to the terrigenous material source. Pyrite and insignificant inclusions of chalcopyrite and galena dominate in the ores. No selenium was found in these sulfides, which indicates that they formed at temperatures lower than those of ores at the Central Vojma, and Northern Vojma deposits. The original sedimentary Fe sulfides in various morphologic forms such as laminated and massive coatings and concretions, framboids, and small idiomorphic crystals were found in slightly metamorphosed banded pyrite. The character of changes in isotopic composition of sulfur in pyrite-marcasite concretions and coatings points to the active involvement of microorganisms in the process of their formation.

At the same time, pyrites of the Leksa deposit have an increased concentration of sulfur with positive Δ33S values. It is assumed that atmospheric sulfate with negative Δ33S values and elemental sulfur (S8) with a corresponding positive Δ33S value precipitated in equal amounts both on land and in the oceans. As a result of bacterial reduction some amount of photolytic sulfate on land turned into sulfide, while elemental sulfur was brought into the ocean (Maynard et al., 2013). Consequently, excessive elemental sulfur, which produced sulfides with a purely positive Δ33S mark, got into the seawater, especially in the coastal area (Fig. 6a).

Host rocks of the Central Vojma and Northern Vojma deposits are also represented by metamorphosed volcanic, and volcano-sedimentary rocks that form fragments inside ultrabasite massifs. Ores of these deposits are massive, brecciated and disseminated. Mineralization consists of two mineral associations: chalcopyrite-sphalerite-pyrite, and chalcopyrite-pyrite. Ore minerals formed at 300–430 °C (Kuleshevich and Belashev, 1998), which is also confirmed by the presence of selenium in the ores of the Northern Vojma deposit. As is known, concentration of selenium in sulfides depends on physico-chemical parameters of ore solutions during sedimentation. High Se values are found in high-temperature mineral associations (Auclair et al., 1987; Layton-Matthews et al., 2005).

The abovementioned allows assuming that ore deposition at the Central Vojma and Northern Vojma deposits occurred under hydrothermal conditions in the discharge zone, accompanied by an intensive circulation of high-temperature fluids. We suppose that these deposits formed within structures analogous to modern hydrothermal systems of rift zones of oceanic or back-arc basins (Fig. 6c). In underwater hydrothermal ore-generating systems, seawater is the base for hydrothermal fluid (Bogdanov et al., 2002; Galley et al., 2007).

Hydrothermal fluids related to the Archean submarine volcanism contained sulfur with a negative Δ33S value. This sulfur was obtained from dissolved seawater sulfates, and from surrounding leached sedimentary strata containing sulfides that formed earlier and showed a negative anomaly due to photolytic sulfate S involvement into their formation. During an intensive circulation with an unlimited volume of seawater, the seawater sulfate S prevailed over sulfur from other sources when sulfide ores were formed. Consequently, sulfides of these deposits acquired a negative for the most part Δ33S value. This anomalous sulfur was sufficient for the mark to remain undiluted and not be removed by volcanic sulfur coming from a magma chamber and leached from surrounding igneous rocks.

Our study of the varying sulfur isotope composition VHMS deposits in the Archean Sumozero–Kenozero greenstone belt has revealed the following features that may be more generally applicable to the Mesoarchean deposits in other terranes:

  1. Most sulfides of the studied deposits of the Sumozero– Kenozero greenstone belt of the Karelian Craton have zero Δ33S values, which point to different proportions of sulfate and elemental sulfur resulting from UV photolysis and later included into ores.

  2. Compositions of sulfur isotopes in volcanic-sedimentary deposits (VHMS) within the Sumozero–Kenozero greenstone belt are direct evidence that seawater sulfate was involved in the formation of these deposits. In addition, atmospheric elemental sulfur also influenced sulfide formation at certain sections. Thus, the Archean atmosphere and hydrosphere significantly contributed to the VHMS mineralization in the region.

  3. Laminated pyrite-marcasite concretions and coatings formed in low-energy environments such as submarine calderas, foot of submarine volcanoes or ore hills, where pyritic mud and other fine-grained detritus accumulated. Concretions grew layer to layer during the early diagenesis, when grains were at the surface of the muddy substrate or beneath a shallow sedimentary cover. They formed under the conditions of fluctuation of the pH, pulsation of oxidation-reduction conditions, and microbial activity. The difference in δ34S values between layers of a single grain points to the change in sulfate concentrations probably caused by the biological sulfate reduction.

  4. Data on multi-isotope sulfur composition confirm the conceptual model of evolution of the (VHMS) Sumozero–Kenozero greenstone belt, which can be generalized in the following manner. The deposits possibly formed within structures analogous to modern back-arc basins. Structures and composition of sulfide ores, and formation temperature of sulfide minerals show that the Central Vojma and Northern Vojma deposits, among all studied, have the highest temperature. Mineralogical thermometers determined the highest values of temperature and high Se values, typical for high-temperature mineral associations. This group of deposits formed directly above or near the intensive hydrothermal venting zone, probably, in the rift zone. It is likely that the Central Vojma deposit marks a feeder vent (ore breccia in the upper part, and stockwork ores deep down), and the Northern Vojma deposit marks the subsurface ores at the periphery of the hydrothermal venting zone. The Leksa and Zolotye Porogi deposits have a lower temperature and, probably, formed on the slope of a central-type volcano or in a submarine caldera.

This work was financially supported by the Russian Science Foundation, grant No. 21-17-00076.