Large sulfur isotope fractionations associated with Neoarchean microbial sulfate reduction

See allHide authors and affiliations

Science  07 Nov 2014:
Vol. 346, Issue 6210, pp. 742-744
DOI: 10.1126/science.1256211

Dissecting ancient microbial sulfur cycling

Before the rise of oxygen, life on Earth depended on the marine sulfur cycle. The fractionation of different sulfur isotopes provides clues to which biogeochemical cycles were active long ago (see the Perspective by Ueno). Zhelezinskaia et al. found negative isotope anomalies in Archean rocks from Brazil and posit that metabolic fluxes from sulfate-reducing microorganisms influenced the global sulfur cycle, including sulfur in the atmosphere. In contrast, Paris et al. found positive isotope anomalies in Archean sediments from South Africa, implying that the marine sulfate pool was more disconnected from atmospheric sulfur. As an analog for the Archean ocean, Crowe et al. measured sulfur isotope signatures in modern Lake Matano, Indonesia, and suggest that low seawater sulfate concentrations restricted early microbial activity.

Science, this issue p. 703, p. 742, p. 739; see also p. 735


The minor extent of sulfur isotope fractionation preserved in many Neoarchean sedimentary successions suggests that sulfate-reducing microorganisms played an insignificant role in ancient marine environments, despite evidence that these organisms evolved much earlier. We present bulk, microdrilled, and ion probe sulfur isotope data from carbonate-associated pyrite in the ~2.5-billion-year-old Batatal Formation of Brazil, revealing large mass-dependent fractionations (approaching 50 per mil) associated with microbial sulfate reduction, as well as consistently negative Δ33S values (~ –2 per mil) indicative of atmospheric photochemical reactions. Persistent 33S depletion through ~60 meters of shallow marine carbonate implies long-term stability of seawater sulfate abundance and isotope composition. In contrast, a negative Δ33S excursion in lower Batatal strata indicates a response time of ~40,000 to 150,000 years, suggesting Neoarchean sulfate concentrations between ~1 and 10 μM.

The Archean Eon [4.0 to 2.5 billion years ago (Ga)] is generally characterized as a time in Earth history when oxygen was a trace constituent in the atmosphere (1) and oceanic sulfate abundance was lower than its present-day concentration by as much as a factor of 150 (2, 3). Although sulfur isotope signatures in marine sedimentary rocks suggest that microbial sulfate reduction (MSR) was used as a metabolic strategy as far back as 3.5 Ga (4), overwhelming evidence for mass-independent sulfur isotope fractionations derived from atmospheric reactions (5) and preserved in Archean records argues against the prominence of contemporaneous MSR in the oceans (69). Iron speciation and trace metal abundances indicate a Neoarchean (2.5 to 2.8 Ga) role for MSR and sulfide production, implying euxinic (lacking oxygen but containing free hydrogen sulfide) conditions in some basins (10). Moreover, recent high-precision secondary ion mass spectrometry (SIMS) analyses of pyrite (1113) and organic matter (14) in Archean successions reveal evidence of micrometer-scale mass-dependent sulfur isotope fractionation (MDF-S).

To further investigate the extent of MSR and the magnitude of MDF-S, we used multiple isotopic techniques to constrain temporal and spatial changes in sulfur and carbon cycling in a Neoarchean sedimentary succession from Brazil (15). We focused on samples from the ~2.5 Ga Batatal Formation (fig. S1) collected from an exploration drill core (GDR-117 provided by AngloGold Ashanti) intersecting ~180 m of strata (Fig. 1). The shallow marine Batatal platform, which is composed of two shale-rich intervals above and below a stromatolitic carbonate, is considered to be a time equivalent of once-contiguous deep-water slope successions in the Hamersley basin of Western Australia and the Transvaal basin of South Africa (16) (fig. S2). This geographically and spatially discrete sedimentary unit thus provides a unique opportunity to examine similarities and differences in sediment geochemistry across the terminal Neoarchean world.

Fig. 1 High-resolution time-series sulfur and carbon isotopic data from ~2.5 Ga Batatal Formation in Brazil.

The most negative δ34S and Δ33S values are recorded exclusively in the carbonate interval. Although the carbonates record shows large variations in δ13C between shallow- and deep-water facies, the magnitude of carbon isotope fractionation between coexisting organic and inorganic phases (Δcarb–org) is relatively constant and unusually small. Shaded area represents extent of carbon isotope fractionation in Neoarchean formations from Western Australia [Mt. McRae Shale Formation, from (21)] and South Africa [Nauga Formation, from (22)].

Distal shale and proximal carbonate facies of the Batatal Formation reveal large sulfur isotope variations (Fig. 1). Pyrite in carbonate (and to a lesser degree in carbonaceous shale) is depleted in 34S and 33S relative to black shale facies. In microdrilled and bulk carbonate-associated pyrite (CAP), δ34S values (17) were as low as –30 per mil (‰) and –14‰, respectively, and both phases were found to have consistently negative Δ33S values (18). In contrast, samples from shale-rich facies are relatively more enriched in both 34S and 33S, depending on carbonate content (fig. S4). SIMS pyrite analyses in a subset of five carbonate samples were conducted to examine the extent of micrometer-scale sulfur isotopic variation within grains and across textural boundaries (fig. S6). Although the Δ33S values determined by SIMS mirror those of bulk and microdrilled samples, the ion probe δ34S measurements reveal even greater degrees of 34S depletion, with pyrite δ34S values as low as –38‰. Notably, SIMS analyses yielding the most negative δ34S values also have constant Δ33S (Fig. 2A). Although some macroscopic pyrite samples drilled in the upper Batatal carbonaceous shale have negative Δ33S values and δ34S signatures as low as –13.6‰, we did not study these samples by the SIMS technique and thus do not know whether these pyrite grains preserve the fine-scale variability and extreme 33S and 34S depletions seen in the carbonate samples. The SIMS pyrite data from carbonate samples in the interval between 1261 and 1272 m form arrays indicating similar levels of 33S depletion (~ –2‰), whereas the array of measurements from the sample 36 m lower in the core (1308 m) defines a Δ33S floor at –3‰.

Fig. 2 Sulfur isotope data for the Batatal Formation in the context of published Neoarchean measurements.

(A) Cross-plot of Δ33S versus δ34S values of pyrites from Neoarchean successions from South Africa (20, 22) and Western Australia (20, 23, 31) as well as bulk and SIMS measurements from Batatal shale and carbonate samples. The bulk Batatal data show large sulfur isotope fractionation in CAP attributed to microbial sulfate reduction; the SIMS analyses reveal two different floors of Δ33S values at negative δ34S. (B) Cross-plot of Δ36S versus Δ33S values of various Neoarchean successions (20, 22); note that Batatal carbonates have consistently negative Δ33S and consistently positive Δ36S values, likely defining end-member compositions on the Archean array.

The SIMS determinations of Δ36S/Δ33S for the Batatal CAP grains are consistent with those of bulk and microdrilled analyses (Fig. 2B), and most fall in the range of previously published Neoarchean data (5, 12) characterized by a Δ36S/Δ33S slope of ~ –1. In some cases (e.g., samples 1308.08, 1266.72, and 1261.90), however, coupled values lie below the reference array, indicating greater depletion in 36S abundances (fig. S8). We attribute the additional 36S depletion in CAP to a metabolic response of MSR at the time the Batatal carbonates were deposited (see fig. S8). Experiments on modern sulfate reducers (19) show that variations within and relationships between multiple sulfur isotopes are produced by bacteria grown under different conditions, likely reflecting the differential transport of sulfate into and out of the cells. On the other hand, Rayleigh distillation (20) is also known to produce variations in Δ36S, but this process also generates large variability in δ34S that is not observed in the data from this study (15) (fig. S9).

Carbon isotope measurements of shale and carbonate in the Batatal Formation reveal a pronounced difference between deep- and shallow-water facies. The magnitude of carbon isotope fractionation between carbonate carbon and organic carbon (Δ13Ccarb–org) in shale and carbonate facies is remarkably constant, with values generally ranging between 15 and 20‰ (Fig. 1). This range of fractionation is much smaller than that measured in the broadly equivalent units from Western Australia and South Africa [~30‰ (21, 22)] (Fig. 1 and fig. S10). Reduced fractionation is unlikely the result of metamorphism but may be biologically mediated in the deep ocean environment. Sedimentological and petrographic observations coupled with a comparison of modern shallow marine environments suggest that the reduced fractionation may be the result of carbon limitation in an evaporitic setting (15) where oceanic sulfate was potentially concentrated.

Sulfur isotope compositions of CAP also differ between this study and prior research on temporally equivalent Neoarchean successions in Western Australia (21) and South Africa (23). In those reports, δ34S values in mixed shale and carbonate facies that accumulated in continental slope settings fell to a nadir of –10‰ in bulk rock measurements and down to –21‰ for SIMS analysis (12). New SIMS results from the Batatal Formation reveal even more negative δ34S pyrite values as low as –38‰ (Fig. 2). Coupled with the estimated range of terminal Neoarchean seawater sulfate compositions (+6 to +15‰) [compare with (3, 24)], this observation suggests that MDF-S values were as much as 40 to 50‰. The magnitude of this fractionation is considerable, even in comparison with those recorded in modern environments (6).

Habicht et al. (2) argued that large fractionations would not be preserved in typical shale facies unless pore water sulfate concentrations were higher than 200 μM (for comparison, modern seawater SO42– is 28 mM). This threshold now appears to be an upper limit, as recent observations of modern meromictic lakes (25, 26) with SO42– as low as 25 to 100 μM indicate that isotopic fractionation of more than 20‰ is possible. The discrepancy between microscale sulfur isotope measurements that reveal evidence for strongly negative δ34S values [e.g., (13, 14)] and bulk measurements that fail to yield a clear indication of strongly negative δ34S signatures may reflect SO42– transport limitation (closed or partially closed) in Archean sediments. The strongly negative and variable δ34S with constant Δ33S and variable Δ36S at the micrometer scale supports this interpretation applied to the Batatal depositional environment; this further provides a link between Batatal sediment pore fluids and overlying seawater sulfate. The greater 34S depletion in the microscale analyses relative to bulk measurements also implies mass-dependent redistribution of sulfur isotopes by MSR within the sediments. However, the average negative δ34S pyrite values (~ –10‰) preserved in bulk samples relative to the inferred positive δ34S composition of coeval seawater SO42– imply that the system was not completely closed [the proportional fraction of sulfate reduced would be ~0.5 to 0.7, assuming a starting sulfate composition of 15‰ and a fractionation of 50‰ (15)].

The large sulfur isotope fractionations observed here imply that MSR did not deplete and fractionate residual pore water sulfate to a large enough extent that transport limitation controlled the expression of isotopic compositions. Following recent suggestions (27) that the availability of electron donors, cell-specific growth rates, and sulfur isotope fractionation are interdependent, we interpret the production and preservation of sulfide with strongly negative δ34S values to reflect a lower proportion of electron donors (i.e., organic matter or H2) relative to sulfate ([e donor]/[SO42–]) in the Batatal shallow marine carbonate-rich environment relative to those found in typical Archean shales. Indeed, organic carbon contents in the Brazilian sediments are much lower than contemporary deep-water facies from Western Australia and South Africa. We further suggest that Batatal CAPs derive their sulfur predominantly from sulfate reduction, rather than from an atmospherically derived source carrying positive δ34S and Δ33S compositions. Our observations of the shallow and potentially evaporitic Batatal carbonate are in agreement with model predictions (28) suggesting greater contributions of atmospherically derived elemental sulfur (with positive Δ33S) to distal environments, whereas sulfate sourced from the atmosphere (with negative Δ33S) and from continental weathering was concentrated in proximal settings.

A notable feature of the SIMS analyses of samples at 1266.72, 1261.90, and 1272.20 m depth is the almost constant Δ33S [average –1.7 ± 1.02‰ (2σ)] for the most highly fractionated (most negative δ34S values) pyrite grains (Fig. 2). This array suggests variations in MDF-S associated with MSR from a starting sulfate pool with a constant negative Δ33S. The relatively constant Δ33S over nearly 60 m of the Batatal Formation carbonate is interpreted to reflect a long-term stability [~0.4 to 1.5 million years, based on a carbonate accumulation rate of 40 to 150 m per million years (29)] in the Δ33S composition of sulfate supplied to the shallow marine environment (i.e., the proportion of atmospheric and nonatmospheric sulfate was stable over the depositional interval).

The negative shift in Δ33S values of samples between core depths of 1302 to 1308 m, however, suggests a secular change in the isotopic composition of oceanic sulfate. The amount of time necessary for shallow marine carbonate accumulation of ~6 m is relatively short (40,000 to 150,000 years) but is still sufficiently long to suggest that Δ33S was well mixed in the oceans. According to our calculations, the residence time of sulfate was lower than the response time by a factor of 2. Thus, order-of-magnitude constraints can be placed on oceanic sulfate concentrations of 10–5 to 10–6 mol/liter assuming only volcanic fluxes (~1011 mol/year), or higher if weathering fluxes are also included (15). Our Neoarchean sulfate concentration estimate is consistent with previous assessments (2, 3).

The connection between pyrite with negative Δ33S and oceanic sulfate implies that at the time the Batatal Formation was deposited, the source of oceanic sulfate had a predominantly negative Δ33S composition. If correct, the negative oceanic Δ33S signature indicates that the atmospheric source of sulfate dominated over that of elemental sulfur, and that sources of sulfate from the oxidation of sedimentary sulfides with positive Δ33S (10) were negligible. This conclusion implies either that late Neoarchean oxidative weathering occurred at low levels, or that weathering fluxes of sulfate were episodic and somehow were not captured by the Batatal CAPs.

Evaluated in the context of the long-term Archean bias toward positive pyrite Δ33S compositions (30) (fig. S11), our analyses suggest that MSR was an important sink for sulfate with negative Δ33S in the Neoarchean ocean. Support for this view comes from analyses of CAP with negative Δ33S in slightly older carbonate-rich Neoarchean strata of Western Australia (Fig. 2A) (24, 31), including the Carawine Dolomite, which contains textural evidence for shallow marine evaporitic conditions (32), as inferred from the sedimentology of the Batatal carbonates. Given the low concentration of CAP in Batatal bulk samples and relative abundances of carbonate and shale in Neoarchean successions (33), however, we estimate that carbonates could represent only about 1% of the negative Δ33S sink necessary to balance the sulfur cycle (15). Other chemical inventories in the oceans, including basinal banded iron formations (30) and abyssal hydrothermal volcanogenic massive sulfide deposits (3), must therefore have been the dominant sinks for atmospherically derived sulfate with negative Δ33S compositions.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S12

Tables S1 and S2

References (3472)

References and Notes

  1. See supplementary materials on Science Online.
  2. Sulfur has four stable isotopes: two more abundant (32S and 34S) and two rare (33S and 36S) isotopes. Sulfur isotope composition of materials is reported as δxS = 1000 × [(xS/32S)sample/(xS/32S)standard – 1], where x is 33, 34, or 36.
  3. The deviation from the mass-dependent relationships is calculated by the following equations: Δ33S = δ33S – 1000 × [(δ34S/1000 + 1)0.515 – 1] and Δ36S = δ36S – 1000 × [(δ34S/1000 + 1)1.90 – 1].
  4. Acknowledgments: We thank C. Noce, N. Geboy, and A. Shrestha, as well as the facilities and the scientific and technical assistance of the Australian Microscopy and Microanalysis Research Facility at the Centre for Microscopy, Characterization and Analysis at the University of Western Australia, a facility funded by the University, State, and Commonwealth governments. Supported by the Fulbright Program (grantee ID 15110620, I.Z.), NASA Astrobiology Institute grant NNA09DA81A (J.F.), and NSF Frontiers of Earth Surface Dynamics program grant 432129 (A.J.K.). All data are provided in the supplementary materials.
View Abstract

Stay Connected to Science

Navigate This Article