Pyrite Morphology and δ 34 S as Indicators of Deposition Environment in Organic-Rich Shales

: This study is focused on the mineralogical, chemical, and isotopic characterization of pyrites from the rocks of the Bazhenov Formation (Upper Jurassic–Lower Cretaceous organic-rich shales, Western Siberia, Russia). Scanning electron microscopy (SEM) revealed pyrites of different morphologies: small and large framboids, small crystals, and large euhedral crystals; all morphotypes were usually combined into aggregates. Isotope ratio mass spectrometry (IRMS) and secondary ion mass spectrometry (SIMS) showed that small framboids and microcrystalline pyrite are isotopically light, with δ 34 SCDT varying from − 55 to − 20‰. Large framboids and euhedral crystals of pyrite are isotopically heavy with δ 34 SCDT up to +26‰. Both morphology and δ 34 S were suggested to be controlled by the redox conditions and sedimentation regime. The abundance of small framboids suggests that pyrite sedimentation occurred under anoxic conditions; the presence of the large framboids and euhedral crystals of pyrite suggest the accumulation of sediments occurred at suboxic conditions, possibly in the presence of oxygen.


Introduction
The Bazhenov Formation of West Siberia is one of the largest source rock formations in the world. It has been actively studied for more than two decades, and interest in understanding this formation remains high. The potential of unconventional hydrocarbon reservoirs is defined by the amount and quality of organic matter, which controls the genesis of hydrocarbons [1]. Organic matter is actively accumulated and well preserved in stagnant basins lacking oxygen, i.e., processes that occur at low oxygen (suboxic), oxygenfree (anoxic), and hydrogen sulfide (euxinic) conditions [2][3][4].
Pyrite (FeS 2 ) is an authigenic mineral which is abundant in multicomponent organicrich rocks [5]. It was shown that the shape and size of pyrite crystals/aggregates are sensitive to redox conditions, and therefore, the distribution of pyrite morphotypes helps reconstruct the evolution of paleoredox conditions during sedimentation [6][7][8][9]. Pyrite is present in the rocks as framboids (densely packed spherical aggregates of microcrystals), as well as crystals of different shapes and sizes, which often form layers and replace organic remnants.
One of the main directions of our research on unconventional hydrocarbons is studying the size distribution of framboids and individual pyrite crystals, as well as their sulfur isotope compositions [10][11][12][13][14][15]. Wang et al. [16] showed that the bulk sulfur isotopic composition could be misleading for determining the sedimentation environment, as pyrite undergoes significant changes during dia-and cata-genesis. Therefore, it is very important to look at the isotopic composition of individual crystals and distinguish pyrite morphology in the rocks. The main source of sulfur in such a sedimentary system is seawater, where the predominant marine aqueous species of sulfur is sulfate-ion (SO 4 2− ). Under oxygen-depleted conditions and with the presence of sulfate-reducing bacteria, sulfur is being reduced by changing its oxidation state from +6 to −2. As a result, hydrogen sulfide (H 2 S) is formed. There are four stable isotopes of sulfur ( 32 S, 33 S, 34 S, and 36 S), and 32 S and 34 S are the most abundant and represent more than 99% of the total sulfur isotopes [17,18]. Delta-notation δ 34 S is used to express the deviation of 34 S/ 32 S measured in a system from a standard material: where R sample is a ratio of 34 S/ 32 S in a sample and R CDT is 34 S/ 32 S (=0.04416) in the Canyon Diablo Troilite standard [19,20]. The goal of our work is to estimate the potential of morphology and δ 34 S of pyrite in reconstructing redox conditions in the well section of the late Jurassic-early Cretaceous sedimentary rocks of the Bazhenov Formation. We conducted a lithological and geochemical study on selected rocks and minerals by first distinguishing pyrite morphotypes and their distribution in the well section, and then comparing δ 34 S in the bulk rock determined by Isotope Ratio Mass Spectrometry (IRMS) with δ 34 S in the individual pyrite crystals and framboids evaluated by Secondary Ion Mass Spectrometry (SIMS). This combined instrumental study of the sediments in the Bazhenov Formation is novel.

Geological Setting
The study area is situated in the North of the West Siberian Basin in the area of the Gydan Peninsula and encompasses the interval of the Late Jurassic to Early Cretaceous ( Figure 1). Deposition of the sediments under investigation lasted for more than 8 Ma at the Jurassic-to-Cretaceous transition and corresponded to environments of the transgression of a marine basin area of more than 2 million km 2 . Deposits of the Bazhenov Formation are underlain by marine terrigenous deposits of the Abalak or Georgiev Formation, Upper Jurassic, and are overlain by Lower Cretaceous argillites and sand-clay clinoforms. In the opinion of Afanasiev I.S. et al. the Bazhenov basin was rather shallow-water, that being witnessed by a 200-300 m difference between occurrence depths of the shelf (undafrom) and deep-water (fondoform) parts of the Lower Cretaceous clinoforms [21]. Deposits of the central parts of the basin show biogenic sedimentation and are represented by predominantly siliceous, argillaceous-siliceous, and carbonate-argillaceous-siliceous deposits enriched in an organic matter [22][23][24][25].
Various amounts of terrigenous material as an admixture to the above-said lithology are associated with different stages of the sea-level stand. It was believed for a long time that near-bottom layers of the Bazhenov basin were hydrogen sulfide contaminated, but findings of deposit eaters and geochemical studies confirm the presence of hydrogen sulfide-free areas [26].
The Bazhenov Formation deposits are spread all over the territory of West Siberia, with the thickness of 10 to 60 m, being 30 m thick on average; occurrence depths of the formation top increase northward and are 600 m minimum and 3800 m maximum [25,27,28].
In terms of mineral composition, the Bazhenov Formation is represented by multicomponent rocks consisting of silica, clay minerals, calcite, pyrite, and feldspars in various percentage proportions. Rock composition is different in the central and peripheral parts of the formation's distribution area. The mean content of organic carbon in Bazhenov Formation rocks is 5-10% wt.; in exceptional cases, it can reach 25%. Pyrite is a ubiquitous component of the rock. Several stages of rock pyritization are observed: finely dispersed early-generation pyrite forms organo-mineral complexes with organic matter, while later-generation pyrite is developed non-uniformly and may form thin laminae and lenses [21].
Close to the study area, large and giant gas, gas-condensate, and oil fields are situated, though quite a small number of wells with coring from the Bazhenov deposits have been drilled in the North of West Siberia, that make the material for the performed studies quite unique.
Geosciences 2021, 11, x FOR PEER REVIEW 3 of 16 component of the rock. Several stages of rock pyritization are observed: finely dispersed early-generation pyrite forms organo-mineral complexes with organic matter, while latergeneration pyrite is developed non-uniformly and may form thin laminae and lenses [21]. Close to the study area, large and giant gas, gas-condensate, and oil fields are situated, though quite a small number of wells with coring from the Bazhenov deposits have been drilled in the North of West Siberia, that make the material for the performed studies quite unique.

Materials and Methods
Samples for our study were obtained from the well core, which is in the north part of Western Siberia. These bulk samples are high-carbon thin clay-siliceous rocks that belong to the Bazhenov Formation, which corresponds to late Jurassic-early Cretaceous geologic time. Preliminarily, gamma-ray logging was recorded on a full-size core to calibrate the core recovery depths to measured depths during logging and to characterize the natural radioactivity of the rocks. Then, we applied lithological, mineralogical, and geochemical methods to evaluate: (i) rock texture; (ii) distribution of organic matter and pyrite in geological section; (iii) morphology of pyrite at micro and nanoscales; and iv) δ 34 S of pyrite.

Materials and Methods
Samples for our study were obtained from the well core, which is in the north part of Western Siberia. These bulk samples are high-carbon thin clay-siliceous rocks that belong to the Bazhenov Formation, which corresponds to late Jurassic-early Cretaceous geologic time. Preliminarily, gamma-ray logging was recorded on a full-size core to calibrate the core recovery depths to measured depths during logging and to characterize the natural radioactivity of the rocks. Then, we applied lithological, mineralogical, and geochemical methods to evaluate: (i) rock texture; (ii) distribution of organic matter and pyrite in geological section; (iii) morphology of pyrite at micro and nanoscales; and (iv) δ 34 S of pyrite.

Characteristics of Texture and Organic Matter
Texture and organic matter of the selected samples were evaluated at the Skolkovo Institute of Science and Technology (Russia). A total of thirty-nine thin sections (2 cm × 3 cm in size) were prepared by slicing the core rock in the direction perpendicular to sedimentary layers for textural characterization. A total of 150 samples were collected to determine the abundance of organic matter or total organic carbon (TOC). Rock texture and distribution of organic matter were examined with a Carl Zeiss Axio Imager transmitted light polarization microscope. Micro-textures of the rocks, pyrite morphology, and major element content were examined on the freshly chipped rock fragments with two scanning electron microscopes (SEM): JSM 6610 LV (JEOL) equipped with an energy dispersive X-ray spectrometer (EDS) (IE350, OXFORD INSTRUMENTS) and Quattro S (Thermo Fisher Scientific) equipped with EDS (XFlash 6|60, Bruker). Mineral compositions of the rocks were determined with X-ray diffraction (XRD) on a Rigaku SmartLab instrument. The organic matter content and type of kerogen was evaluated with the pyrolysis Rock-Eval method on a Hawk Resource Workstation (Wildcattechnology).

Mass Spectrometry Analyses
Sulfur isotope composition and sulfur content in the bulk rocks were determined by isotope ratio mass spectrometry using an EA-IRMS system, consisting of an isotope ratio mass spectrometer IRMS Delta V plus (Thermo Fisher Scientific, Bremen, Germany) coupled to an Elemental Analyzer Thermo Scientific Flash 2000 (Thermo Fisher Scientific, Milan, Italy). The system provides combustion of total sulfur from the sample in SO 2 -gas at 1020 • C, with subsequent measurement of its sulfur isotope composition. The reference materials used for calibration were: Sulfate NBS 127 δ 34 S: +21.3‰, Methionine δ 34 S: +9.3‰, and 4,4-Diaminodiphenyl sulfone δ 34 S: +6.0‰. δ 34 S values are reported relative to Vienna Canyon Diablo troilite (V-CDT). Two or more technical replicates of each sample were measured with an uncertainty lower than 0.5‰.
The sulfur isotope compositions of individual crystals and pyrite aggregates were analyzed using the CAMECA ims-1290 ion microprobe in the Department of Earth, Planetary, & Space Sciences, University of California Los Angeles (USA). The samples, together with the standards (Ruttan pyrite and Chisel sphalerite) were mounted in epoxy (Buehler) and polished using 600, 800, and 1200 grit silicon carbide paper (Buehler) followed by 1 µm Al 2 O 3 powder (Mark V Laboratories). Prior to ion microprobe, the mounts were examined with optical microscopy, coated with a 20 nm layer of gold, and examined by SEM at MSU. Sulphides were clearly visible with both microscopy techniques. EDS spectra allowed distinction between Fe and Zn sulphides embedded in the aluminosilicate matrix.
Single-spot sulphur isotope analyses (δ 34 S) were conducted using a~0.25 nA Cs+ primary beam. Data were acquired in multicollection mode, with two Faraday Cups (FC for 32 S and FC2 for 34 S). Mass resolution was set at 3000 to isolate the peaks of interest from interferences (e.g., 32 S from O 2 ). Prior to signal collection, each spot was pre-sputtered for 30 seconds to achieve sputtering equilibrium. Each spot analysis consisted of 10 cycles for both standards and samples, with a five-second counting time in each cycle. For the standards, the average 32 S count rates on the FC detector were~4 × 10 8 and 5 × 10 8 cps (counts per second) for Ruttan pyrite and Chisel sphalerite respectively. The stability of the instrument was constantly monitored, and the instrumental mass fractionation was determined by analyzing the standard several times throughout each analytical session. The reproducibility of the standard (mounted together with samples) was better than 0.3‰ (standard deviation [SD]).
Instrumental mass fractionation (IMF) was calculated as: where R = 34 S/ 32 S of the standard was measured by SIMS (R Measured ) and determined independently (R True ). IMF varied from 0.9998 to 0.9996 (i.e., from −0.4 to −1.2‰) for Ruttan pyrite, and was more constant (IMF = 0.995) for Chisel sphalerite (i.e., −5‰). According to Kozdon et al. [30], instrumental mass fractionation for chalcopyrite (−1.7‰) is lower than for sphalerite (−3.2‰). Therefore, data reported for chalcopyrite are subject to a few ‰ systematic uncertainty and should be used with extra caution. The same is applicable to isotope data in FeS.

Lithology and Organic Matter
Late Jurassic-Early Cretaceous sedimentary rocks were collected from the Georgiev and the Bazhenov formations. The section consists of argillaceous-siliceous-rocks with rare lamination of carbonate rocks ( Figure 2).
where R = 34 S/ 32 S of the standard was measured by SIMS (R Measured ) and determined independently (R True ). IMF varied from 0.9998 to 0.9996 (i.e., from −0.4 to −1.2‰) for Ruttan pyrite, and was more constant (IMF = 0.995) for Chisel sphalerite (i.e., −5‰). According to Kozdon et al. [30], instrumental mass fractionation for chalcopyrite (−1.7‰) is lower than for sphalerite (−3.2‰). Therefore, data reported for chalcopyrite are subject to a few ‰ systematic uncertainty and should be used with extra caution. The same is applicable to isotope data in FeS.

Lithology and Organic Matter
Late Jurassic-Early Cretaceous sedimentary rocks were collected from the Georgiev and the Bazhenov formations. The section consists of argillaceous-siliceous-rocks with rare lamination of carbonate rocks (Figure 2).  The rock classification scheme is based on structural characteristics and the amount of organic matter:
Argillaceous-siliceous indistinctly laminated mudstone The rocks have similar mineral compositions and contain on average 42% clay minerals (predominantly illite), 38% quartz, up to 15% feldspars, up to 17% pyrite, and in some cases, up to 3% dolomite according to XRD analysis. Furthermore, ZnS (sphalerite or wurzite) and CuFeS 2 (chalcopyrite) were found using EDS SEM in the rocks. Based on the geochemical and lithological characteristics, eight members can be distinguished ( Figure 2).
The amount of total organic carbon (TOC) is one of the main criteria in classification of lithotypes and characteristics. The TOC content varied from 0.5 to 12% within the section, having an average of 4.8%. Kerogen type II was the predominant species, but kerogen of the mixed type (II+III) was present in some members. The base of the studied well corresponded to the Georgiev horizon and contained a low abundance of organic matter, on average 2.3%. The higher TOC content was present in the Bazhenov Formation, with an average value of 4.8%. In the Bazhenov Formation, members 1, 2, 3, 5, and 7 were enriched in TOC, whereas members 4, 6, and 8 were relatively depleted in TOC. In the first member, TOC varied from 2.8 to 9.9% (with an average of 5.8%) and the kerogen type was II. In the second member, the TOC variation was smaller (3.4-6.7%) and the average TOC values was 4.7%, which was slightly lower than TOC in the first member. The higher TOC values (6.4-11.6%, 8.8% on average) were found in the third member with kerogen type-II. In the higher member (№4), the kerogen changed to mixed type (II+III) with TOC of 1.4-7.3% (4.3% in average). In members 5 and 7, the predominant kerogen type was II and TOC contents were higher: 4.0-8.7% with an average of 7.1% in member 5 and 6.3-9.5% with an average of 7.6% in member 7. The mixed type of kerogen (II+III) occurred in members 6 and 8 with lower TOC of 0.5-4.3% (2% in average).

Pyrite Morphology
According to XRD analysis, the abundance of pyrite in the rocks varied from 3 to 17% (7% on average). Four morphological types of pyrite were identified using SEM ( Figure 3):

1.
Small framboids (diameter 5-10 µm) concentrations of sub-micron pyrite crystals 0.5-1 µm in size (Figure 3a). These framboids were porous, abundant in the sediment layers with high carbon content, and were often associated with organic matter.

2.
Large framboids were denser and less porous (without space between micro-crystals) than small framboids. Diameter of large framboids could reach 30 µm and consisted of microcrystals with the size range of 2-5 µm (Figure 3b).

3.
Small crystals (1-2 µm in size) were non-uniformly distributed in lenses and layers (Figure 3c,d). Small crystals were often abundant within bioturbation traces and organic remnants.
Geosciences 2021, 11, x FOR PEER REVIEW    Additional SEM images showed the textural difference between the aggregate interior and the rim (Figure 4a-h). The inner part of the aggregate has numerous pits filled with nano-sized crystals (Figure 4e,g), whereas the outer parts have coarse crystals (Figure 4f,i,h). These coarse crystals (>10 µm) have pronounced grain boundaries (Figure 4a-c). The boundary between inner and outer portions of the aggregate is shown on Figure 4f  Imaging with SEM EDS yielded homogeneous distribution of S and Fe within large pyrite aggregate confirming that it consist entirely of FeS2. In contrast to S and Fe, analysis shows that Mn was depleted in the center and enriched near the rim of the aggregate ( Figure 5).

Sulfur Isotope Composition (Bulk IRMS)
Seventeen samples were analyzed with IRMS for evaluating bulk δ 34 S in the core (Figure 2). The accuracy of bulk isotope analysis is supported by the predominance of pyrite, Imaging with SEM EDS yielded homogeneous distribution of S and Fe within large pyrite aggregate confirming that it consist entirely of FeS 2 . In contrast to S and Fe, analysis shows that Mn was depleted in the center and enriched near the rim of the aggregate ( Figure 5).

Sulfur Isotope Composition (Bulk IRMS)
Seventeen samples were analyzed with IRMS for evaluating bulk δ 34 S in the core (Figure 2). The accuracy of bulk isotope analysis is supported by the predominance of

Sulfur Isotope Composition of Individual Pyrites (SIMS)
To determine δ 34 S of the identified morphotypes of pyrite, four samples were selected from the different members of the section (Figure 2). Thirty-six spot analyses were conducted using SIMS (Table A1) and the ranges of isotope data for each morphology in each sample are presented in Table 1. Overall, δ 34 S CDT varied from −55 to +26‰. The lightest sulfur (δ 34 S CDT = −55.38‰) was observed in small pyrite framboids (Figure 6a). The heaviest sulfur (δ 34 S CDT = +26.49‰) was trapped in the outer portion (rim) of the aggregate of large euhedral pyrite crystals (Figure 5c). The inner portion (center) of the same aggregate is isotopically lighter (−32.79 ≤ δ 34 S CDT ≤ −21.72‰) and the border between "heavy" rim and "light" center is visible on reflected light and SEM images (Figure 6c).

Discussion
Sulfur isotopes are useful for studying the origin of pyrite (i.e., involvement of microorganisms in crystallization process). The range of sulfur isotope composition, observed in our study, suggests pyrite crystallization was caused by microbial sulfate reduc- 34 34

Discussion
Sulfur isotopes are useful for studying the origin of pyrite (i.e., involvement of microorganisms in crystallization process). The range of sulfur isotope composition, observed in our study, suggests pyrite crystallization was caused by microbial sulfate reduction [31], which produced isotopically light hydrogen sulfide, i.e., δ 34 Ssulfide-δ 34 Ssulfate was as large as −77‰ [32,33]. The observed variation in sulfur isotope compositions and morphologies of pyrite was likely controlled by changes in redox conditions during sedimentation and burial processes as well as by Ryleigh fractionation. At reduced or oxygen-depleted conditions, isotopically light pyrite could crystallize above the sediment-water boundary or in the upper layer of sediments having an excess of marine sulfate (i.e., open system). During diagenesis, crystallization of pyrite could occur in porewaters, where access to sulfate is restricted (i.e., system is closed) and progressively reduced sulfate became enriched in 34 S. As a result, sulfide and pyrite became isotopically heavy compared to those in the open system.
In an open system, crystallization of small framboids of isotopically light syngenetic pyrite (δ 34 S = −62 to −50‰) is expected [32,33], which was likely the case for pyrites from our study where δ 34 S values of small pyrite framboids were as low as −55.38‰ (Figure 6a), suggesting pyrite formed in anoxic conditions or/and in the presence of hydrogen sulfide. The change in depositional environment from open to closed system can explain the observed wide isotopic range (Table A1). In the large pyrite aggregate, δ 34 S was negative in the center and positive at the rim (δ 34 S > + 15‰) (Figure 6c). The porous nature of the aggregate center suggests that it consists of small framboids and probably was deposited at a high rate, as was previously observed [8,34]. An increase of δ 34 S (from −33 to +26‰) with this aggregate can be explained by Rayleigh fractionation (discussed further in the text). In the late Jurassic, δ 34 Ssulfate varied from +15 to +20‰ [35] or from +17 to +18‰ [36], with the average value of 17.5‰, and therefore, the expected δ 34 S of accumulated pyrite should not exceed this δ 34 Ssulfate value. This restriction is not necessary for instantaneously precipitated pyrite. A gap between negative and positive values of δ 34 S in pyrite aggregate can be explained by suppression of the pyrite crystallization due to temporary oxygenation of the depositional environment. The later diagenesis allowed the development of sulfatereducing and oxygen-depleting anaerobic microorganisms, which caused an increase in hydrogen sulfide in porewaters and crystallization of isotopically heavy, large pyrite crystals (in the rim of the large aggregate) at a probably slow growth rate. Magnall et al. [37] made a similar conclusion on their work in Canada. The lack of sulfate exchange between porewater and seawater caused progressive 34 S enrichment in sulfate, hydrogen sulfide, and pyrite.
The effort was taken to explain this isotope fractionation with a Rayleigh model. A Rayleigh equation was applied to explain the isotopic fractionation by calculating the evolution of δ 34 S of the dissolved sulfate, instantaneous sulfide, and accumulated pyrite as a function of the fraction of sulfate (f) controlled by microbial sulfate reduction: δ 34 S sul f ide = δ 34 S sul f ate − 72.88 (4) here R sul f ate 0 is an initial ratio of 34 S/ 32 S in sulfate (when f = 1); isotope difference between sulfide and sulfate (i.e., δ 34 S sul f ide − δ 34 S sul f ate ) is −72.88‰ [33]; α is an apparent fractionation factor between dissolved sulfide and sulfate: α = R sul f ide R sul f ate = 1 + δ 34 S sul f ide −δ 34 S sul f ate 1000 = 1 − 0.07288 = 0.927; α is not a thermodynamic fractionation factor as it depends on the reduction rate of sulfate (e.g., Fry et al. [38]). The chosen fractionation factor (0.927) is within the range of previously reported values [33,39,40]. We selected fractionation magnitude (−72.88) as a difference between δ 34 S of the minimum measured value in pyrite and sulfate of Jurassic seawater. The selected fractionation is bigger than those (up to −66‰) obtained in the cultured experiment of Sim et al. [41]. However, it is at the highest end of fractionation observed in natural systems (up to −72‰) by Canfield et al. [42], within the theoretic limit (−75.7‰) Otake et al. [43], and within the ranges (up to −77‰) reported in Rudnicki [32] and Brunner [33].
The selected magnitude fractionation (−72.88‰) allowed fitting bulk and SIMS data with the Rayleigh model, via adjusting the f value for accumulated and instantaneous sulfide, respectively ( Figure 7). If fractionation of −66‰ from Sim et al. [41] is assumed, then the Rayleigh model cannot fit the most negative SIMS datum (i.e., −55.38‰). Evolution curves of δ 34 S of the remaining sulfate and instantaneous sulfide with f are always different by −72.88‰ (Figure 7a, Equations (3) and (4)). No isotope fractionation between aqueous sulfide and pyrite was assumed while plotting measured δ 34 Spyrite versus f determined by Equation (5). The reason why offsets between pyrite and aqueous sulfur species vary with f is that as soon as pyrite crystallized, it did not interact with sulfate or sulfide. When sulfate reduction was completed, the accumulated pyrite has δ 34 S values equal to the initial sulfate δ 34 S: δ 34 (3) and (4)). No isotope fractionation between aqueous sulfide and p was assumed while plotting measured δ 34 Spyrite versus determined by Equation (5) reason why offsets between pyrite and aqueous sulfur species vary with is that as soo pyrite crystallized, it did not interact with sulfate or sulfide. When sulfate reduction was c pleted, the accumulated pyrite has δ 34 S values equal to the initial sulfate δ 34 S: 34 The Rayleigh fractionation curves for accumulated pyrite and instantaneous su (pyrite) fit δ 34 Spyrite data measured with bulk IRMS and in situ SIMS techniques, res tively (Figure 7b). The large gap between negative and positive SIMS δ 34 Spyrite data ure 7b) suggests that there was an interruption in pyrite growth (or at least a decrea the crystallization rate), which is also shown by the distribution of δ 34 S within the p  Figure 7b) suggests that there was an interruption in pyrite growth (or at least a decrease of the crystallization rate), which is also shown by the distribution of δ 34 S within the pyrite aggregate (Figure 6c). This fact confirms the change in pyrite growth regime and suggests that the progressive development of anoxic conditions along with the exhausting of sulfate led to a reduction in Mn-oxides from sediment porewater under low Eh conditions. The higher concentration of Mn in pore water at the late stage of diagenesis contributed to the coprecipitation of Mn during formation of outer parts of pyrite aggregates [44,45]. The observed isotopic, elemental, and textural features suggest that the growth of the outer layer occurred at conditions different from those that enhanced growth of the pyrite aggregate interior. The large variability in sulfur isotopes and trace elements within a single grain of pyrite was reported by Nabhan et al. [46]. They suggested that the growth of the pyrite rim, enriched in Ni and Co and depleted in δ 34 S, occurred from pore fluids enriched in isotopically light sulfide and in Ni and Co derived from weathering of ultramafic rocks.

Conclusions
Four types of pyrite were identified in the sediments of the Bazhenov Formation of Siberia: small framboids, aggregation of small crystals, large framboids, and large euhedral crystals. Sulfur isotopes analyses show small framboids and microcrystalline pyrite are isotopically light, with δ 34 SCDT varying from −55 to −20‰. Large framboids and euhedral crystals of pyrite are isotopically heavy with δ 34 SCDT up to +26‰.
The dominating abundance of small framboids suggests that pyrite sedimentation in the section of the Bazhenov Formation occurred under anoxic conditions with the presence of hydrogen sulfide above sediment-water boundary or in the upper layer of sediments with an excess of marine sulfate. These conditions are favorable for the accumulation of organic matter. The presence of the large framboids and euhedral crystals of pyrite in several sediment layers suggests another pattern of sulfide formation. The accumulation of sediments occurred at suboxic conditions; the depositional environment was unfavorable for pyrite formation. Anoxic conditions, necessary for microbial sulfate reduction and sulfide formation, have been developed in porewater of these layers during the burial of sediments, which decreased the connectivity between the porewater and overlying seawater and formed in diagenesis a restriction in respect to the oxygen and sulphate system. Therefore, the presence of the large framboids and euhedral pyrite crystals suggest conditions unfavorable for the accumulation and preservation of organic matter.   Isotope data for CuFeS 2 and FeS are subject to a few permille errors because of the absence of matrix-matched SIMS standards. Therefore, these data should be used with caution.