1. Introduction
The Proterozoic eon (2.5–0.54 Ga) marks the transition from an anoxic Archaean to oxic Phanerozoic ocean–atmosphere system (Holland, Reference Holland1984; Des Marais et al. Reference Des Marais, Strauss, Summons and Hayes1992). Although oxygenation of deep ocean began at c. 1.8 Ga (Holland, Reference Holland1984; Holland & Beukes, Reference Holland and Beukes1990), recent studies on biomarkers of sulphur bacteria (Brocks et al. Reference Brocks, Love, Summons, Knoll, Logan and Bowden2005), molybdenum and sulphur isotope (δ34S) compositions in black shales (Arnold et al. Reference Arnold, Anbar, Barling and Lyons2004), and pyrites (Canfield, Reference Canfield1998, Reference Canfield2004; Poulton, Fralick & Canfield, Reference Poulton, Fralick and Canfield2004) strongly advocate continuation of the sulphidic–anoxic oceanic state throughout the Mesoproterozoic and parts of the Neoproterozoic period until the pO2 approached modern levels after c. 1 Ga. Bio-limiting trace metals might have been scarce in such a stressed environment, causing a restricted nitrogen cycle and marine primary productivity, or in turn hindering evolution of early life like eukaryotes or triploblastic animals (Hoffman et al. Reference Hoffman, Kaufman, Halverson and Schrag1998; Anbar & Knoll, Reference Anbar and Knoll2002; Kah, Lyons & Frank, Reference Kah, Lyons and Frank2004).
In particular, the exact timing of appearance and explosion of triploblastic animals or eukaryotes is an issue of debate. The oldest known records of both triploblastic trace fossils (Seilacher, Bose & Pfluger, Reference Seilacher, Bose and Pflüger1998; Rasmussen et al. Reference Rasmussen, Bose, Sarkar, Banerjee, Fletcher and McNaughton2002) and multicellular eukaryotes (Bengtson et al. Reference Bengtson, Belivanova, Rasmussen and Whitehouse2009) have recently been recorded from c. 1.6 Ga old rocks of the Vindhyan basin, India. It is peculiar that these metazoans/algae did not evolve and proliferate in the subsequent phases of Vindhyan sedimentation. Also, these purported traces are not found in any other correlative basinal successions of the world, making them controversial (Jensen, Droser & Gehling, Reference Jensen, Droser and Gehling2005). Understanding the geochemical conditions (e.g. sulphur cycle) of Indian Proterozoic basins, therefore, becomes important, as this can give important insight into the Proterozoic oxygenation and evolution of early life.
The fractionation of sulphur between sulphate (SO=4) and sulphide (H2S) species in oceans is driven by bacterial sulphate reduction (BSR). Sulphide produced by BSR is either recycled back as sulphate or combines with available iron to form pyrite (FeS2). Burial of sedimentary pyrite (and organic carbon) eventually controls oxygen in an ocean–atmosphere system (Berner, Reference Berner1984; Bottrell & Newton, Reference Bottrell and Newton2006). BSR preferentially fractionates 32S in pyrite, making the remaining sulphate enriched in 34S. Since the amount of reduction depends both on the redox state at the sediment–water interface and the supply of sulphate in the water column, the sulphur isotope compositions (δ34S) of sedimentary sulphides can be used as potential tracers of Proterozoic ocean chemistry.
Here we report preliminary δ34S data of sedimentary pyrites from three large Mesoproterozoic basins of India: Vindhyan, Cuddapah and Chattisgarh, which add to the already existing database and suggest that the Mesoproterozoic anoxic sulphidic ocean was indeed global in nature. Our data, although limited, provide additional input to the proposed model of the Proterozoic ‘Canfield ocean’ (Buick, Reference Buick2007) and suggest that both oxygenation of biosphere and evolution of metazoa or eukaryotes were possibly delayed due to this extreme adverse environmental condition.
2. Geology
Epeiric seaways were over-represented in the Proterozoic, low-gradient, and much wider and shallower than modern shelves (maximum ~ 100 m deep with no discernible shelf–slope break: Shaw, Reference Shaw1964; Irwin, Reference Irwin1965; Friedman, Sanders & Kopaska-merkel, Reference Friedman, Sanders and Kopaska-Merkel1992; Eriksson et al. Reference Eriksson, Condie, Tirsgaard, Mueller, Altermann, Miall, Aspler, Catuneanu and Chiarenzelli1998), but exposed to open-ocean tides and waves. However, widespread carbonate deposits and/or phosphate nodules in black shales indicate that depth varied from as low as 30 m to > 100 m, respectively (Friedman, Sanders & Kopaska-Merkel, Reference Friedman, Sanders and Kopaska-Merkel1992; Eriksson et al. Reference Eriksson, Condie, Tirsgaard, Mueller, Altermann, Miall, Aspler, Catuneanu and Chiarenzelli1998).
Locations, composite stratigraphies and important chronological levels of Vindhyan, Chattisgarh and Cuddapah basins are shown in Figure 1. Palaeocurrent analysis (Bose et al. Reference Bose, Sarkar, Chakrabarty and Banerjee2001) and basin-scale presence of tidal currents or waves (Banerjee, Reference Banerjee, Valdiya, Bhatia and Gaur1982; Chakraborty & Bose, Reference Chakraborty and Bose1990; Paul & Chakraborty, Reference Paul and Chakraborty2003) suggest the Vindhyan to be an epicratonic sea with an open ocean connection in the northwest (Chanda & Bhattacharya, Reference Chanda, Bhattacharya, Valdiya, Bhatia and Gaur1982). Of particular interest in the present study is the Bijaygarh Shale (BS; Fig. 1b). Gradational contact with the underlying shallow marine Lower Kaimur Sandstone, fining-upward sand-free character, absence of current and wave features, occasional phosphatic and strata-bound pyrites indicate the Bijaygarh Shale as the product of maximum flooding (Bose et al. Reference Bose, Sarkar, Chakrabarty and Banerjee2001) following a transgressive systems tract (TST). The pyritiferous part is interpreted as a condensed zone (Banerjee et al. Reference Banerjee, Dutta, Paikaray and Mann2006). Within this overall transgressive framework, however, the Bijaygarh Shale registers intermittent storm incursions (e.g. profuse gutter casts: Chakraborty, Reference Chakraborty2006). Tidal bundles or double mud-layers at its lower and upper parts also suggest occasional oxygenation of the Vindhyan shelf (Bose & Chaudhuri, Reference Bose and Chaudhuri1990).
(a) Location map of the three large Proterozoic basins of India: Vindhyan, Chattisgarh and Cuddapah. Composite stratigraphy of (b) Vindhyan, (c) Chattisgarh and (d) Cuddapah basins. Also indicated are chronology (including the dating methods adopted by different workers) of various horizons and positions of the pyritiferous units analysed in the present study (open star); solid star indicates the horizons analysed by earlier workers. Solid circle indicates the barite layer in the Cuddapah basin.
In the Chattisgarh basin, the depositional environments varied from fluvial (Patranabis Deb et al. Reference Patranabis Deb, Bickford, Hill, Chaudhuri and Basu2007) through estuarine to distal marine shelf below storm wave base (Murti, Reference Murti and Radhakrishna1987; Chakraborty & Paul, Reference Chakraborty and Paul2008). A TST across the Rehatikhol–Saraipali transition in a fan delta-shelf environment is surmised (Chakraborty et al. Reference Chakraborty, Sarkar, Das and Das2009; Fig. 1c). The overlying Charmuria Limestone also shows a retrogradational stacking pattern representing a transgressive event. The limestone is dominantly micritic and pyritiferous. Though rare in the Phanerozoic, such micritic bedded carbonates were abundantly formed in anoxic intertidal conditions in both the Palaeo- and Mesoproterozoic (Sumner & Grotzinger, Reference Sumner and Grotzinger1996; Bartley et al. Reference Bartley, Knoll, Grotzinger, Sergeev, Grotzinger and James2000). Cyanobacteria is reported at some locations (De, Reference De2007). From its evenly bedded character (0.25 to 1.25 cm thick), the presence of a ‘birds eye’ structure, complete absence of siliciclastic grains and waves or tide, a sub-tidal, low-energy euxinic depositional environment (an isolated platform in a cratonic seaway) has been conceived (Moitra, Reference Moitra1995; Patranabis Deb, Reference Patranabis Deb2004).
A coastal to shallow marine depositional environment is proposed for Cuddapah sediments (Nagaraja Rao et al. Reference Nagaraja Rao, Rajurkar, Ramalingaswamy, Ravindra Babu and Radhakrishna1987; Basu et al. Reference Basu, Gangadharan, Kumar, Sharma, Rai and Chaki2007; Fig. 1d). Without any ‘sequence analysis’, based solely on the repeated alternation between quartzite, shale or limestone at the formation scale, recurrent transgressive–regressive cycles and unconformities were suggested (Ramam & Murthy, Reference Ramam and Murthy1997). In particular, the barite–pyrite-bearing Cumbum shale has been interpreted as a product of transgression in a deep-water cratonic basin (Chaudhuri et al. Reference Chaudhuri, Saha, Deb, Patranabis Deb, Mukherjee and Ghosh2002; Mukhopadhyay, Ghosh & Nandi, Reference Mukhopadhyay, Ghosh and Nandi2006). All sedimentological evidence indicates absence of any open ocean connectivity for both the Chattisgarh and Cuddapah basins (Patranabis Deb, Reference Patranabis Deb2004).
3. Sampling
Sample locations and stratigraphic levels are shown in Figure 1. Strata-bound pyrite from the Amjhore mine, Shahbad district, Bihar, hosted within the Bijaygarh Shale (Vindhyan), was sampled. In the Chattisgarh and Cuddapah basins, pyrites from the Charmuria limestone, Cumbum black shale and Narji limestone were sampled. For assessing the sulphur isotope fractionation between dissolved sulphate and sulphide (Ohmoto & Rye, Reference Ohmoto, Rye and Barnes1979), barite, hosted in the Cumbum shale (locally known as the Pullampet Formation), was sampled from the Mangampeta mine. The barite is grey, fine-grained and thick-bedded. In addition, selected carbonates were sampled for oxygen (δ18O) and carbon (δ13C) isotope analysis from these basins for assessing the extent of burial diagenesis.
4. Chronology of pyrite-bearing strata
Pb–Pb dates of the limestones from the Kajrahat and Rohtas formations are 1721 ± 90 Ma (Fig. 1b; Sarangi, Gopalan & Kumar, Reference Sarangi, Gopalan and Kumar2004; Ray, Reference Ray2006) and 1601 ± 130 Ma (Ray, Veizer & Davis, Reference Ray, Veizer and Davis2003) and 1599 ± 48 Ma (Sarangi, Gopalan & Kumar, Reference Sarangi, Gopalan and Kumar2004), respectively. Although the Pb–Pb ages are inherently not very reliable and give a tentative chronology for the lower Vindhyans, the zircon U–Pb dates (both conventional and SHRIMP techniques) of 1630.7 ± 0.4 Ma (Ray et al. Reference Ray, Martin, Veizer and Bowring2002) and 1628 ± 0.8 Ma (Rasmussen et al. Reference Rasmussen, Bose, Sarkar, Banerjee, Fletcher and McNaughton2002) of the Porcellanite Formation occurring between the Kajrahat and Rohtas formations are well consistent with the obtained Pb–Pb dates. Recent palaeomagnetic studies and a LA-ICPMS U–Pb detrital zircon age of 1020 Ma from the Upper Bhander sandstone (Malone et al. Reference Malone, Meert, Banerjee, Pandit, Tamrat, Kamenov, Pradhan and Sohl2008) nearly bracket the upper limit of the Upper Vindhyans. The syn-sedimentary pyrites from the Bijaygarh Shale, therefore, suggest an age between 1.6 Ga and 1.0 Ga, that is, Mesoproterozoic (Fig. 1b).
The lower part of the Cuddapah basin provides ages between 1817 ± 24 Ma (Rb–Sr date of the Pulivendla mafic sill emplaced into the Tadpatri shales: Bhaskar Rao et al. Reference Bhaskar Rao, Pantulu, Damodara Reddy and Gopalan1995; Fig. 1d) and 1756 ± 29 Ma (Pb–Pb age of uranium mineralization and stromatolitic dolomite from the Vempalle and Tadpatri formations: Zachariah et al. Reference Zachariah, Bhaskar Rao, Srinivasan and Gopalan1999). The 40Ar–39Ar age of 1418 ± 8 Ma of lamproite dykes emplaced within the Cumbum shale (Chalapathi Rao et al. Reference Chalapathi Rao, Miller, Gibson, Pyle and Madhavan1999) marks the termination of the Cuddapah sedimentation. Although both Rb–Sr and Ar–Ar ages can be easily reset, the ranges suggest that both the pyrite and barite samples of the Cumbum shale, analysed in the present work, have a Mesoproterozoic age between 1.8 Ga and 1.4 Ga. The pyrites of the Narji limestone of the Kurnool Group are younger than 1.4 Ga, but the exact age is difficult to estimate.
Das et al. (Reference Das, Yokoyama, Chakraborty and Sarkar2009) calculated a c. 1455 ± 47 Ma age (EPMA U–Th–Pb dates of monazite grains from bedded tuff) for the basal part of the Chattisgarh succession. The U–Pb SHRIMP date on zircons from the tuff layers in the uppermost Tarenga shale provides an age range of 990–1020 Ma (Fig. 1c; Patranabis Deb et al. Reference Patranabis Deb, Bickford, Hill, Chaudhuri and Basu2007). Bounded between the two well-dated horizons, the pyritiferous Charmuria Limestone undoubtedly indicates a Mesoproterozoic age. Summarizing, the pyrites studied here were formed during Mesoproterozoic times when regional black shale/limestone deposition was concurrently taking place in all three sedimentary basins of India.
5. Analytical methods
The pyrite samples were polished and examined under a reflected light microscope. For isotopic analysis, powdered (micro-drilled) pyrite and barite samples were packed in tin capsules with a mixture of V2O5 to promote full combustion. Samples were combusted in a quartz tube reactor, pre-filled with WO3 and pure Cu, and kept at a temperature of 1050 °C within a Flash HT 1112 elemental analyser. Purified SO2 gas was measured in a Delta Plus XP continuous-flow mass spectrometer. Routine precision (monitored by running both NBS-123 and an internal BaSO4 standard) of ~ ± 0.2 ‰ was obtained for δ34S. The δ34S (‰) values are expressed relative to the Canyon Diablo Troilite standard (CDT). For δ18O and δ13C analyses, powdered carbonates were reacted with orthophosphoric acid and purified CO2 measured in the same mass spectrometer. The system was calibrated by the NBS-19 standard. Analytical precision was ~ ± 0.1 ‰ for both δ18O and δ13C.
6. Results and interpretation
6.a. Mode of occurrence and petrography of pyrites
Mesoscopically the pyrites occur as massive, bedded (~ 0.8 m thick) cryptocrystalline variety (e.g. at Amjhore, Vindhyan), microclots (maximum size being ~ 0.3 cm; e.g. in Charmuria Limestone), disseminated and large euhedral crystals (e.g. Narji limestone or Cumbum shale). Under the microscope, Amjhore pyrite displays both framboidal and euhedral habits. The framboids show a concentration of elliptical to oval (Fig. 2a) or elongated grains that radiate from possible growth nuclei (Fig. 2c). The wavy carbonaceous and pyritiferous laminae in the Bijaygarh Shale have recently been inferred to be microbial mats of cyanobacterial origin (Banerjee et al. Reference Banerjee, Dutta, Paikaray and Mann2006). From the growth of quartz cement between the pyrite grains in the form of ‘teeth and socket’ structures, both pyrite and quartz have been interpreted as early diagenetic products (Sur, Schieber & Banerjee, Reference Sur, Schieber and Banerjee2006). While this may not be strictly the case, the framboids do indicate a definite early diagenetic origin. The euhedral grains (Fig. 2b, d) could have formed from these original framboids (Sawlowicz, Reference Sawlowicz1993), where variation in grain size possibly suggests different extents of diagenetic recrystallization. Mangampeta pyrites exhibit microscopic-scale banding of small euhedral grains (Fig. 2e). Pyrites from Charmuria and Kurnool are characteristically euhedral in nature (Fig. 2f–h).
(a–c) Reflected light photomicrographs of pyrite showing framboidal texture from Amjhore; (d) euhedral grains with more diagenetic recrystallization from Amjhore; (e, f, h) laminated bands of small euhedral grains from Mangampeta, Charmuria and Kurnool, respectively; (g) euhedral pyrite crystals in a hand specimen of Kurnool limestone.
6.b. Sulphur isotopes in pyrites
Tables 1 and 2 summarize δ34S and δ18O/δ13C values, respectively, along with the sample details. Our data show moderate to very heavy pyrite δ34S values ranging from minimum 8.1 to maximum 38.8 ‰ in these basins. When compared and put together with the published data, the δ34S values of the Chattisgarh and Vindhyan show means of 26.3 ± 0.9 ‰ (n = 12) and 25.5 ± 8.7 ‰ (n = 42), respectively (Sinha et al. Reference Sinha, Raju, Bhaskar and Asha2001; Guha, Reference Guha1971; Table 1). It is important to note that the earlier work on the Vindhyan pyrites reported a mean value of ~ 9.3 ‰, much lower than the maximum enriched value, up to 31.8 ‰, obtained in the present work. This indicates distinct geochemical differences (namely, the amount of sulphate reduction) that operated during the precipitation of these two sets of pyrite samples, as evident from the sedimentological signature of alternating depositional environments in the Bijaygarh Shale (see Section 2). No published record of syn-sedimentary Cuddapah pyrite is available, and the limited data show a mean δ34S of 28.3 ± 11.8 ‰ (n = 6). When all available data are considered, the δ34S of Indian pyrites range from 4.5 ‰ to 40.7 ‰.
Sulphur isotope composition of Indian pyrites
The δ34S (‰) values are expressed relative to the Canyon Diablo Troilite standard (CDT).
δ18O, δ13C values of carbonates
Frequency distributions of all δ34S data from the three basins are shown in Figure 3a–c. No facies-dependent variation in the δ34S values (see Table 1 for facies types), as reported by Shen, Canfield & Knoll (Reference Shen, Canfield and Knoll2002), is observed. All Indian pyrites show high 34S enrichment, irrespective of the nature of the pyrite (framboidal, euhedral or laminated), and none shows moderate to extreme negative δ34S values as observed in most sedimentary pyrites of Phanerozoic age (Strauss, Reference Strauss1999; Canfield & Raiswell, Reference Canfield and Raiswell1999; Canfield, Reference Canfield2004 and references therein). The δ34S values as high as ~ 40 ‰ are closer to or even higher than the known δ34S value of Proterozoic marine sulphates (Strauss, Reference Strauss1997; Kah, Lyons & Chesley, Reference Kah, Lyons and Chesley2001; Gellatly & Lyons, Reference Gellatly and Lyons2005; Canfield, Reference Canfield2004) and require discussion.
Frequency distribution of δ34S values of pyrite from (a) Vindhyan; (b) Chattisgarh; (c) Cuddapah basins. Open bars: data from present work; solid bar: published data. (d) Calculated distribution of sulphide δ34S formed from Proterozoic seawater sulphate (+30 ‰) by biogenic sulphate reduction under relatively open (assumed SO=4 – S= fractionation factor ~ 1.040, Ohmoto & Rye, Reference Ohmoto, Rye and Barnes1979) and closed (assumed fractionation factor ~ 1.025) systems. Also shown is the calculated enrichment in sulphate in a closed system capable of producing ~ 42 ‰ value like the Cuddapah barite (c).
7. Discussion
7.a. Diagenetic effect on sulphur istotope composition
In modern environments, authigenic pyrite is formed by BSR; the magnitude of reduction depends on specific geochemical conditions (Goldhaber & Kaplan, Reference Goldhaber, Kaplan and Goldberg1974; Berner, Reference Berner1984; Canfield, Reference Canfield2004). Normally BSR is completed within a few metres of the sediment–water interface (Riciputi, Cole & Machel, Reference Riciputi, Cole and Machel1996). Large amounts of H2S are produced by this mechanism, provided the diagenetic environment is anaerobic (Eh < −100 mV), has a sufficient supply of dissolved SO=4 and nutrients for bacterial growth (e.g. high amount of organic matter) and has a lower temperature range of 0–60 °C (Riciputi, Cole & Machel, Reference Riciputi, Cole and Machel1996; Machel, Krouse & Sassen, Reference Machel, Krouse and Sassen1995). At higher temperatures (80–200 °C), abiogenic thermochemical sulphate reduction (TSR; Machel, Krouse & Sassen, Reference Machel, Krouse and Sassen1995) can occur if there is sufficient dissolved sulphate in pore water.
Initial reduction (and consequent rupturing of S–O bonds) in BSR involves a kinetic fractionation of up to 40 ‰, however, formation of intermediate compounds such as thiosulphate can cause an extended range of fractionation, producing H2S as depleted as −70 ‰ (Rees, Reference Rees1973; Ohmoto & Rye, Reference Ohmoto, Rye and Barnes1979; Jørgensen, Reference Jørgensen1990). The resulting pyrites, thus formed, can be depleted by 15–65 ‰, compared to the parent marine sulphate. The kinetic fractionation during the SO=4 → S= reduction in TSR progressively decreases from ~ 20 ‰ at 100 °C to only ~ 10 ‰ at 200 °C (Kiyosu & Krouse, Reference Kiyosu and Krouse1990). In many deep gas reservoirs, insignificant fractionation has been observed between the sulphate and metal sulphide, thereby producing enriched δ34S values of pyrites almost similar to that of sedimentary sulphate (Krouse, Reference Krouse1977). The only available marine sulphate, the Cuddapah barite, has a δ34S value of +42.3 ‰ and is similar to the earlier reported data (41.8 to 45.5 ‰; Clark, Pooleb & Wang, Reference Clark, Pooleb and Wang2004). The δ34S values of the Cuddapah barite are not only enriched by about 10–15 ‰ more than the contemporary (Mesoproterozoic) marine sulphate (Strauss, Reference Strauss1999; Kah, Lyons & Chesley, Reference Kah, Lyons and Chesley2001; Kah, Lyons & Frank, Reference Kah, Lyons and Frank2004; Gellatly & Lyons, Reference Gellatly and Lyons2005), but also have exceptionally small variation compared to other Proterozoic barite deposits of the world (Clark, Pooleb & Wang, Reference Clark, Pooleb and Wang2004; Strauss, Reference Strauss1999; Kah, Lyons & Chesley, Reference Kah, Lyons and Chesley2001; Kah, Lyons & Frank, Reference Kah, Lyons and Frank2004; Gellatly & Lyons, Reference Gellatly and Lyons2005). The maximum pyrite value of ~ +40 ‰ is not very different from this sulphate value. Hence the distinct possibility exists that TSR created at least some of the enriched values. A deep burial diagenesis involving TSR will also cause substantial change in both texture and δ18O–δ13C values of basinal carbonates through water–rock interaction. In general, the well-preserved laminations, dominant micritic character and preservation of primary textures, such as stromatolite laminae, pelloidal and intraclastic texture in lime mudstone (Burns & Matter, Reference Burns and Matter1993), suggest early diagenetic character. Although textural preservation is possible even in a post-depositional fluidized diagenetic regime, we consider it to be negligible, based on the ranges of stable isotopes of associated carbonates as discussed below.
Figure 4a shows the cross-plot of δ18O and δ13C values of plane-laminated deep-water pyritiferous limestones of the Chattisgarh and Cuddapah basins. In the Vindhyan basin, although pyrite is hosted in black shale, plain laminated Bhander limestone was also plotted for comparison (Table 2; Fig. 4a; data taken from Chakraborty et al. Reference Chakraborty, Sarkar, Bhattacharya and Sanyal2002). All data plot within the ellipse representing the δ18O/δ13C range of global marine carbonates deposited between 1.5 and 0.8 Ga (Kasting et al. Reference Kasting, Tazewell Howard, Wallmann, Veizer, Shields and Jaffrés2006). Although there is a considerable spread, the mean δ18O value of these carbonates (−6.6 ± 2.9 ‰) is considerably depleted compared to modern marine carbonates. This is quite consistent with the observed δ18O values of Proterozoic carbonates which are often depleted by as much as 10 ‰ compared to the Late Phanerozoic period (Kasting et al. Reference Kasting, Tazewell Howard, Wallmann, Veizer, Shields and Jaffrés2006). Such depleted values, earlier thought to be a diagenetic effect, are increasingly considered as products of depleted ocean water caused by low-temperature water–rock interaction in a hot ocean (Kasting et al. Reference Kasting, Tazewell Howard, Wallmann, Veizer, Shields and Jaffrés2006). The mean δ13C value of these carbonates is +2.9 ± 1 ‰ with all data plotting towards the most enriched end of the global range (Fig. 4a, b). Global compilation of the marine δ13C values between 1.5 Ga and 1 Ga also shows enriched values between 0 and 5 ‰ (Brasier & Lindsay, Reference Brasier and Lindsay1998; Kah et al. Reference Kah, Sherman, Narbonne, Kaufman, Knoll and James1999; Bartley & Kah, Reference Bartley and Kah2004), consistent with the observed δ13C range (Fig. 4b). One possible reason for such enrichment is a reduction in the global dissolved inorganic carbon reservoir and consequent increase in onshore carbonate productivity (Bartley & Kah, Reference Bartley and Kah2004; Chakraborty et al. Reference Chakraborty, Sarkar, Bhattacharya and Sanyal2002). The δ13C values of carbonates formed during TSR, on the other hand, are generally depleted, ranging from −20 to −70 ‰, depending on the species of hydrocarbon (e.g. kerogen, crude or bitumen) aiding the sulphate reduction (Machel, Krouse & Sassen, Reference Machel, Krouse and Sassen1995). Further, the Bhander Limestone (Vindhyan) reportedly preserves the original marine Sr isotope ratio (87Sr/86Sr ~ 0.705 to 0.706; Ray, Veizer & Davis, Reference Ray, Veizer and Davis2003). A pervasive diagenesis, however, introduces radiogenic Sr, increasing the Sr ratio to as high as 0.709 (Shields & Veizer, Reference Shields and Veizer2002; Ray, Veizer & Davis, Reference Ray, Veizer and Davis2003). Based on these lines of evidence, we believe that deep burial diagenesis was not a common feature for most carbonates (and associated sediments) in these basins, and hence TSR played an insignificant role in the formation of these pyrites. Instead we consider these pyrites to be the product of BSR in widely varying geochemical conditions, such as redox state and extent of sulphate supply, as discussed below.
(a) δ18O–δ13C cross-plot of Indian Proterozoic carbonates. Note enriched δ13C values and cluster of data away from probable TSR field (for details see text). The ellipse is the global range of marine carbonates between 1.5 and 0.8 Ga (data source: http://www.science.uottawa.ca/geology/isotope_data/; Kasting et al. Reference Kasting, Tazewell Howard, Wallmann, Veizer, Shields and Jaffrés2006). (b) Secular marine δ13C variation through Mesoproterozoic (modified from Brasier & Lindsay, Reference Brasier and Lindsay1998); the range of Indian carbonates shown by dotted line.
7.b. Geochemical modelling
The δ34S values of H2S or metal sulphides, formed during BSR, depend on the kinetic isotope effect (k1/k2) during the SO=4 ↔ S= reduction and SO=4/H2S ratio of different geochemical systems that are either open or closed with respect to SO=4 or H2S (Ohmoto & Rye, Reference Ohmoto, Rye and Barnes1979). The value of k1/k2 varies from 1.025 in a closed system to 1.065 in a relatively open system with high SO=4 supply, the absolute δ34S value of sulphide being dependent on the initial value of marine sulphate and its concentration in the water column. Figure 3d shows the calculated distribution of δ34S values of SO=4 and sulphide in closed systems with no SO=4 supply, but with fast and slow pyrite removal (Rees, Reference Rees1973; Schwarcz & Burnie, Reference Schwarcz and Burnie1973; Ohmoto & Rye, Reference Ohmoto, Rye and Barnes1979). Also shown in Figure 3d are the δ34S values of sulphides in a relatively open system, where SO=4 is supplied by slow diffusion from the upper layer of the ocean, a situation often described for the so-called ‘euxinic’ basin of the Black Sea. Reported δ34S values of barite from the Proterozoic Nagthat Formation of lesser Himalaya, deposited in a continental shelf setting, range from +26.5 to +29.5 ‰ (Sharma, Verma & Law, Reference Sharma, Verma and Law2006) and could have been the sea-water sulphate value in this region during the Proterozoic; a range of δ34S sulphate values between +25 and +35 ‰ have been obtained from Proterozoic evaporites from different parts of the world (Strauss, Reference Strauss1997, Reference Strauss1999). We assumed an initial sulphate value of +30 ‰ for calculating the change in δ34S of either SO=4 or H2S. Figure 3d suggests that the highly enriched δ34S value (~ 42 ‰) of the Cuddapah barite possibly resulted due to progressive reduction of sulphate in a closed system that enriched the δ34S of the remaining dissolved SO=4. Our estimation suggests that a ~ 40% decrease in the initial SO=4 concentrations can produce the value observed in the Cuddapah barite.
Comparison of calculated values of sulphides with the observed frequency distribution of δ34S of the Chattisgarh and Cuddapah basins (Fig. 3b, c) is consistent with sulphur derived by BSR in closed systems with no SO=4 supply. Very high δ34S values (> 35 ‰) in both the basins suggest that pyrite removal or burial was much faster compared to the rate of sulphate reduction (Ohmoto & Rye, Reference Ohmoto, Rye and Barnes1979; Jørgensen, Reference Jørgensen1979; Strauss & Schieber, Reference Strauss and Schieber1990). The higher frequency of lower δ34S values in the Vindhyan pyrite (Fig. 3a), on the contrary, suggests a semi-confined marine depositional environment where SO=4 flux was still higher (though limited) compared to the other two basins. Chemical tracers like U/Th and V/Cr also suggest a fluctuating oxic/anoxic condition during the depositional span of the shale unit (Bijaygarh Shale) that contained these pyrites (Sur, Schieber & Banerjee, Reference Sur, Schieber and Banerjee2004). Another possible reason for the variability in δ34S of Vindhyan pyrite could be the greater sensitivity of isotopic composition to very low sulphate concentration (Gellatly & Lyons, Reference Gellatly and Lyons2005). It must, however, be mentioned that our dataset is limited, and more extensive sulphur isotope studies in each of these basins are required to assess the Proterozoic ocean chemistry in general.
8. Implications of heavy δ34S values
The high enriched mean δ34S values of > 25 ‰ in different types of sedimentary pyrites from various stratigraphic levels of the Indian Mesoproterozoic basins call for a likely common parameter of low sulphate availability. A large body of data on δ34S of pyrite and barite shows a much lower SO=4–H2S fractionation during the Proterozoic and was explained by high rate of pyrite burial with respect to very low concentrations of available sulphate (Canfield & Raiswell, Reference Canfield and Raiswell1999; Canfield, Reference Canfield2004). A rate-dependent model for sulphur isotope change indicates that the marine sulphate concentration was between 1.5 and 4.5 mM (~ 5–15% of modern marine value: Kah, Lyons & Frank, Reference Kah, Lyons and Frank2004), suggesting an anoxic bottom water (Shen, Canfield & Knoll, Reference Shen, Canfield and Knoll2002) throughout the Proterozoic. This prompted the hypothesis of a widespread sulphidic ocean during this period (Logan et al. Reference Logan, Hayes, Hieshima and Summons1995; Canfield, Reference Canfield2004; Hurtgen et al. Reference Hurtgen, Arthur, Suits and Kaufmann2002). Based on the Fe speciation data, a classic case for a Mesoproterozoic sulphidic ocean has indeed been demonstrated across the 1.8 Ga old Animikie Group of Canada (Poulton, Fralick & Canfield, Reference Poulton, Fralick and Canfield2004). The sedimentary pyrites from Tapley Hill, Australia, showed a range of δ34S values (+10 to +55 ‰) that were enriched over the coexisting marine sulphate and strongly suggested a sulphate-poor anoxic ocean during early Neoproterozoic time as well (Canfield, Reference Canfield2004). Total organic carbon (TOC) content in some of the Indian pyrite-bearing shales, such as the Bijaygarh Shale, has been found to be ~> 4% (Banerjee et al. Reference Banerjee, Dutta, Paikaray and Mann2006). This, along with the definite presence of cyanobacterial pyritiferous mat in this shale (Banerjee et al. Reference Banerjee, Dutta, Paikaray and Mann2006), suggests anoxic stratified bottom water conditions, where the SO=4 was only being supplied by diffusion from above. The high positive δ34S data of sedimentary pyrite, therefore, suggest that even in the intra-cratonic basins a low sulphate–high sulphidic geochemical condition prevailed during Mesoproterozoic time. Our data provide additional evidence that the continued bottom ocean anoxia through much of Proterozoic time was indeed global in nature. The present level of knowledge, however, is insufficient to infer whether these Proterozoic basins were part of an integrated global ocean or not, even though temporal similarity in their geochemical evolution suggests a common driving force, such as low pO2–low sulphate of the ocean.
The marine sulphate concentration probably remained low at the ~ 4.5 mM level until about 1.2 Ga and rose to ~ 7–10 mM only after 0.8 Ga (Canfield, Reference Canfield1998; Azmy et al. Reference Azmy, Veizer, Misi, Olivia and Dardenne2001; Hurtgen et al. Reference Hurtgen, Arthur, Suits and Kaufmann2002; Kah, Lyons & Frank, Reference Kah, Lyons and Frank2004; Gellatly & Lyons, Reference Gellatly and Lyons2005). Since accumulation of sulphate in the ocean is also controlled by oxidative weathering of the continents, the low sulphate levels possibly implied delayed oxygenation of the earth's ocean–atmosphere system until Neoproterozoic time. Such an inference also has implications for the evolution of early life. The sulphidic ocean might have induced an acute shortage of redox-sensitive bio-limiting trace metals like Cu, Mo, Cd and Zn in the water column by removing them onto sediments (Lewis & Landing, Reference Lewis and Landing1992; Helz et al. Reference Helz, Miller, Charnock, Mosselmans, Pattrick, Garner and Vaughan1996; Anbar & Knoll, Reference Anbar and Knoll2002). Because these metals regulate nitrogen fixation, acute nitrogen stress might have delayed the evolution of eukaryotes, since they need to assimilate fixed nitrogen from ambient water (Anbar & Knoll, Reference Anbar and Knoll2002). As mentioned before, the oldest evidence of metazoan burrows of ‘wormlike undermat miners’ or possible eukaryotes are recorded from the c. 1.6 Ga Vindhyan sediments occurring below the pyritiferous Bijaygarh Shale (Seilacher, Bose & Pfluger, Reference Seilacher, Bose and Pflüger1998; Bengtson et al. Reference Bengtson, Belivanova, Rasmussen and Whitehouse2009; Fig. 1b). Diversification of animal life, however, was very slow until the Early Cambrian, when hard skeletons appeared in an evolutionary explosion (Knoll, Reference Knoll, Lipps and Signor1992). This was coincident with the complete oxygenation of the Ediacaran ocean and a large increase in sulphur isotope fractionation (Canfield & Teske, Reference Canfield and Teske1996; Canfield, Reference Canfield2004; Fike et al. Reference Fike, Grotzinger, Pratt and Summons2006). It is tempting to speculate that the early metazoans or eukaryotes developed only in the Vindhyan basin due to its limited open ocean connection which locally provided intermittent oxygen supply in an otherwise sulphidic ocean. The pace of the evolution was retarded, however, due to the overall prolonged anoxic sulphidic conditions in the oceans that deposited sedimentary pyrites with heavy δ34S, as found in the present study and studies of many other contemporary basins of the world.
9. Conclusions
Analysis of sulphur isotope compositions of sedimentary pyrites from the three large Proterozoic basins of India, Vindhyan, Chattisgarh and Cuddapah, gave mean heavy δ34S values of +25.5 ± 8.7 ‰, 26.3 ± 0.9 ‰ and 28.3 ± 11.8 ‰, respectively. The δ34S values do not show any relationship either with the mesoscopic/microscopic nature of pyrites or sedimentary facies. The data provide supplementary evidence in support of a hypothesized global Proterozoic sulphidic anoxic ocean, where a very low concentration of marine sulphate, bacterially reduced in closed systems, produced δ34S values in pyrites similar to or even heavier than open ocean marine sulphate. The anoxic ocean could have been responsible for the delayed oxygenation of the biosphere and may have retarded evolution of multicellular life.
Acknowledgements
All isotopic data have been generated in the National Stable Isotope Facility of IIT, Kharagpur. We sincerely thank the Department of Science and Technology, New Delhi, for funding the mass spectrometer laboratory at this facility.
