Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Sulphur cycling in a Neoarchaean microbial mat

Sulphur cycling in a Neoarchaean microbial mat University of St Andrews, St Andrews, UK Multiple sulphur (S) isotope ratios are powerful proxies to understand the complexity Department of Earth and Planetary Sciences, Washington University, St. Louis, of S biogeochemical cycling through Deep Time. The disappearance of a sulphur mass- MO, USA independent fractionation (S- MIF) signal in rocks <~2.4 Ga has been used to date a Correspondence dramatic rise in atmospheric oxygen levels. However, intricacies of the S- cycle before N. R. Meyer, Department of Earth System the Great Oxidation Event remain poorly understood. For example, the isotope com- Science, Stanford University, Stanford, CA, USA. position of coeval atmospherically derived sulphur species is still debated. Furthermore, Email: nrmeyer@stanford.edu variation in Archaean pyrite δ S values has been widely attributed to microbial sul- and A. L. Zerkle, School of Earth and phate reduction (MSR). While petrographic evidence for Archaean early- diagenetic Environmental Sciences, University of pyrite formation is common, textural evidence for the presence and distribution of St Andrews, St Andrews, UK. Email: az29@st-andrews.ac.uk MSR remains enigmatic. We combined detailed petrographic and in situ, high- 34 33 resolution multiple S- isotope studies (δ S and Δ S) using secondary ion mass spec- Current address N. R. Meyer, Department of Earth System trometry (SIMS) to document the S- isotope signatures of exceptionally well- preserved, Science, Stanford University, Stanford, CA pyritised microbialites in shales from the ~2.65- Ga Lokammona Formation, Ghaap 94305, USA Group, South Africa. The presence of MSR in this Neoarchaean microbial mat is sup- Funding information ported by typical biogenic textures including wavy crinkled laminae, and early- Geological Society of London’s Alan & Charloett Welch Fund; A Palaeontological diagenetic pyrite containing <26‰ μm- scale variations in δ S and Association Undergraduate Research, Grant/ 33 34 Δ S = −0.21 ± 0.65‰ (±1σ). These large variations in δ S values suggest Rayleigh Award Number: PA-UB201504; University of St Andrew’s Zannah Stephen Memorial Travel distillation of a limited sulphate pool during high rates of MSR. Furthermore, we iden- Scholarship; Natural Environment Research tified a second, morphologically distinct pyrite phase that precipitated after lithifica- Council Fellowship, Grant/Award Number: 34 33 NE/H016805/1; Natural Environment tion, with δ S = 8.36 ± 1.16‰ and Δ S = 5.54 ± 1.53‰ (±1σ). We propose that the Research Council Standard, Grant/Award S- MIF signature of this secondary pyrite does not reflect contemporaneous atmos- Number: NE/J023485/2; NSF/EAR, Grant/ Award Number: 1229370; Packard Fellowship pheric processes at the time of deposition; instead, it formed by the influx of later- stage sulphur- bearing fluids containing an inherited atmospheric S- MIF signal and/or from magnetic isotope effects during thermochemical sulphate reduction. These in- sights highlight the complementary nature of petrography and SIMS studies to resolve multigenerational pyrite formation pathways in the geological record. 33 34 1  |  INTR ODUCTION follow a thermodynamically determined, linear δ S/δ S relationship 33 34 (δ S = 0.515 × δ S). However, the geological S- isotope record also The sulphur isotope record has played an integral role in shaping captures evidence of mass- independent fractionation (MIF), where 33 34 our understanding of key events in Earth’s geological and biolog- δ S and δ S deviate from the predicted terrestrial mass fractionation ical history. Surficial S- cycling principally involves biological and line, quantified by the capital- delta (Δ) notation. Prior to the Great abiotic mass- dependent fractionation (MDF) processes, which Oxidation Event (GOE) at ~2.4 Ga, S- bearing minerals show large This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2017 The Authors Geobiology Published by John Wiley & Sons Ltd. Geobiology. 2017;15:353–365. wileyonlinelibrary.com/journal/gbi     353 MEYER Et al . 354       33 33 34 33 positive and negative Δ S values. Aerft ~2.4 Ga, Δ S ratios dimin- S- isotope ratios that largely follow a linear trend in δ S versus Δ S ish towards values that tightly cluster around zero (Δ S = 0 ± 0.2‰). along an Archaean reference array. Combining these data with photo- Archaean sulphur MIF has been widely attributed to atmospheric chemical models, Ono et al. (2003) suggested that the Δ S sign was photochemical reactions involving SO ; these reactions would have positive for elemental sulphur and negative for sulphate. Furthermore, been blocked in the Palaeoproterozoic due to the UV- shielding eects ff Palaeoarchaean barites (BaSO ) also show a 0 to −1.5‰ Δ S signal caused by increased atmospheric O and O concentrations (e.g., (e.g., Roerdink, Mason, Farquhar, & Reimer, 2012; Ueno, Ono, Rumble, 2 3 Farquhar, Bao, & Thiemens, 2000). In addition, the delivery of S- MIF to & Maruyama, 2008), but these could have formed in the marine water the Earth’s surface requires sulphur to leave the atmosphere via mul- column, in the sediments or in hydrothermal environments not neces- tiple exit channels at different redox states, which are homogenised sarily reflective of the seawater sulphate pool (Paytan, Mearon, Cobb, & −5 when atmospheric oxygen exceeds 10 of present atmospheric lev- Kastner, 2002; Van Kranendonk, 2006). Additional photochemical exper- els (Pavlov & Kasting, 2002). Therefore, measuring multiple sulphur iments produce S- isotope slopes that contrast to Ono et al.’s (2003) re- 34 33 isotopes (δ S, Δ S) in Archaean sediments can provide information sults (e.g., see review of Paris, Adkins, Sessions, Webb, & Fischer, 2014); on both atmospheric chemistry and biogeochemical sulphur cycling in moreover, recent photochemical models demonstrate that the sign in Archaean palaeoenvironments. Δ S of experimentally produced sulphur will be highly dependent on the The biogeochemical cycling of sulphur in the Archaean was funda- wavelength of UV light used to simulate photolysis (Claire et al., 2014). mentally different to the present- day cycle. The largest flux of sulphur Further poorly constrained components of the Archaean S- cycle into the modern oceans is riverine sulphate, derived from the oxidative are the concentration and δ S of seawater sulphate. Archaean seawa- weathering of pyrite. However, this flux was less significant before the ter sulphate concentrations were expected to be lower than modern GOE due to low atmospheric pO ; therefore, the two most significant values because of the low atmospheric pO before the GOE. Habicht, 2 2 S fluxes into the Archaean oceans were likely hydrothermally sourced Gade, Thamdrup, Berg, and Canfield (2002) used culturing experi- and atmospherically derived sulphur species (e.g., Fike, Bradley, & ments of modern sulphate reducers to observe decreasing fraction- Rose, 2015). The Archaean S- MIF signal is thought to be dominated ation factors with decreasing sulphate concentrations. Assuming this by photochemical reactions involving atmospheric SO (e.g., Farquhar relationship can be extrapolated into the Archaean, they estimated et al., 2000; Pavlov & Kasting, 2002). In contrast, the shielding eect ff seawater sulphate concentrations at <200 μM. However, MSR frac- of atmospheric oxygen and ozone prevents the creation of an S- MIF tionation factors are dependent on microbial species (Bradley et al., signal in the contemporary S- cycle, with the exception of photochem- 2016) and other factors including reduction rate, type of substrate ical reactions generating stratospheric sulphate aerosols from volcanic and temperature (e.g., Fike et al., 2009; and references therein). More eruptions (Baroni, Thiemens, Delmas, & Savarino, 2007). Furthermore, recent results have suggested Archaean seawater sulphate concen- with low atmospheric and marine pO , homogenisation of discrete trations could have been even lower, <10 μM (Crowe et al., 2014; 2− atmospheric sulphur species (e.g., SO , S , SO ) was limited, further Zhelezinskaia, Kaufman, Farquhar, & Cliff, 2014). 4 n 2 favouring preservation of the Archaean S- MIF signal. The δ S ratio of seawater sulphate depends on the isotope compo- Aerft formation, the photochemically derived sulphur species, sition of the sulphate input into the oceans, the steady state burial flux 2− generally assumed to be sulphate (SO ) and elemental sulphur (S ), of sulphides and sulphate and the fractionation factor between coeval 4 n were deposited in the oceans. Sulphate is soluble in water and would sulphate- and sulphide- bearing species (Fike et al., 2015). Particularly, homogenise rapidly. It could form sulphate- bearing phases, such as if the sulphate concentrations in the Archaean oceans were low and carbonate- associated sulphate (CAS), gypsum (CaSO .2H O) or barite basins were restricted, MSR could have changed the residual sulphate 4 2 (BaSO ); or it could be reduced to hydrogen sulphide (H S) via microbi- composition of seawater through Rayleigh distillative fractionation 4 2 ally mediated MDF redox reactions or thermochemical sulphate reduc- (Roerdink et al., 2012). Therefore, seawater sulphate S- isotope compo- tion (TSR) before capture in the rock record as pyrite. These S- cycling sitions may have been heterogeneous, both spatially and temporally, 34 34 processes can significantly change the isotope composition of δ S with as proposed from the scaertt of δ S values preserved in Neoarchaean small (<0.2‰) differential eects ff on Δ S (Johnston et al., 2005). The carbonates (e.g., Paris et al., 2014; Zhelezinskaia et al., 2014). fate of S is less certain, but once deposited on the seafloor, it could In addition, the distribution and occurrence of microbial sulphate react with H S produced by microbial sulphate reduction (MSR) to form reduction in the Archaean is debated. Bulk δ S data from Australian 2− a reactive, mobile, soluble form of polysulphide S . Reactive polysul- Palaeoarchaean barites (Shen & Buick, 2004; Shen, Buick, & Canfield, phide and hydrogen sulphide can then react to form sulphide- bearing 2001) along with phylogenetic studies (Wagner, Roger, Flax, Brusseau, minerals, such as pyrite (Farquhar et al., 2013). Therefore, the S- isotope & Stahl, 1998) suggest that microbial sulphate reduction is likely an composition of pyrite generally records the biogeochemical MDF pro- ancient metabolism. However, the link between MSR S- isotope finger - 2− cesses that generated S and H S from atmospheric precursors. prints and pyrite textures remains enigmatic. Ono, Beukes, and Rumble n 2 Despite the narrative outlined above, there remain several poorly (2009) sampled multiple pyrite phases in ~2.5- Ga upper Prieska facies constrained aspects of the Archaean sulphur cycle. Firstly, the Δ S from the GKP01 Agouron drill core. They correlated different py- 2− signatures of the atmospheric products (e.g., S and SO ) produced rite phases with the Archaean reference array (Ono et al., 2003) and n 4 through photochemical reactions remain a subject of debate. Archaean suggested nodular and layered pyrite had S- isotope signatures most rocks from the Hamersley Basin, Western Australia, contain pyrite with consistent with MSR. Similarly matching pyrite crystal morphology to MEYER Et al .       355 the array, Kamber and Whitehouse (2007) proposed spheroidal py- rite concretions capture an MSR S- isotope fingerprint in the 2.52- Ga Upper Campbellrand Subgroup, Transvaal, South Africa. Moreover, Fischer et al. (2014) showed that δ S in nodular pyrite were charac- terised by systematic enrichment towards the rims (with the lowest δ S signature present in the centre of the nodules). Additional studies in the Griqualand West Basin are required to test whether these tex- tural and geochemical interpretations can be applied to pyrite from other palaeoenvironments and depositional ages. To address these gaps in our current understanding of Archaean sulphur cycling, we measured in situ, μm- scale, multiple sulphur iso- 34 33 tope ratios (δ S and Δ S) in exceptionally well- preserved pyritised microbialites in shales (sample- 3184) from the ~2.65- Ga Lokammona Formation, Ghaap Group, South Africa. By combining petrography with S- isotope data from secondary ion mass spectrometry (SIMS), we aim to increase our understanding of sulphur cycling in a Neoarchaean micro- bial mat, and the secondary (late- diagenetic/post- lithification) processes that may impact the primary (depositional/early- diagenetic) S- isotope signal. By petrographically correlating pyrite phases with sulphur isotope fingerprints, we can determine the MDF and MIF processes that contrib- uted to the formation of multiple pyrite generations in these sediments. 2  |  MATERIALS AND METHODS 2.1 | Geological setting and core material Our sample (3184) was collected from the BH1- SACHA core, drilled through the Neoarchaean Transvaal Supergroup in the Griqualand West Basin (Figure 1). The core material was obtained from the National Core Library at Donkerhoek (Pretoria, South Africa). The core contains carbon- FIGURE 1 Simplified geological map of Transvaal and Griqualand ates, siliciclastics and iron formations as well as several igneous intrusive West sediments, the approximate location of BH1- SACHA (star) bodies (Figure 2). Sample- 3184 was obtained from a depth of 3184 m and the inferred fault trace of the Kheis sole thrust fault. Insert: the from the Lokammona Formation, Schmidtsdrif Subgroup, Ghaap group, red rectangle shows the location of the main map relative to other Transvaal sediments and the Kaapvaal craton (beige). Adapted from Transvaal Supergroup. The Lokammona (Clearwater) Formation is domi- Altermann and Wotherspoon (1995) [Colour figure can be viewed at nated by shale, tuff layers intercalated with black shale, and minor do- wileyonlinelibrary.com] lomites (Figure 2). The high proportion of fine- grained sediments is consistent with deposition in a low- energy environment; Altermann and been identified (Altermann & Siegfried, 1997). We will consider these Siegfried (1997) suggested sedimentation occurred in a deep, shelf en- igneous and metasomatic processes when interpreting geochemical vironment with shale- carbonate/shale cycles representing shallowing- data from BH1- SACHA. upward cycles. They also noted the presence of microbial laminites, which suggested deposition in the photic zone. The interpretation is 2.2 | Imaging and EPMA supported by studies suggesting that modern microbial mats that pro- duce microbially induced sedimentary structures (MISS) are dominated We used the VHX- 2000 super- resolution digital microscope housed by benthic photoautotrophs (Noffke, 2009). The Lokammona Formation in the School of Earth and Environmental Sciences at the University has a model age of 2.650 ± 0.008 Ga from SHRIMP U- Pb analyses of of St Andrews (Scotland, UK) to image thick sections of sample- zircons from tuff laminae (Knoll & Beukes, 2009). 3184. This facilitated the detailed mapping of structures and fabrics The area sampled by the BH1- SACHA core is characterised by little within the sample and selection of target areas for SIMS. Backscatter subsequent tectonic deformation (Beukes, 1987), and it has been sub- electron (BSE) imaging was carried out using the School of Earth jected to sub- greenschist facies metamorphism (Button, 1973; Miyano and Environmental Sciences’ Jeol JCXA- 733 Superprobe electron & Beukes, 1984). The core has been penetrated by dykes and sills; microprobe analyser (EPMA). Prior to analysis, sample- 3184 thick the largest continuous igneous intrusion occurs at a depth of 922.3– sections were coated with a ~40- nm- thick graphite layer. The analy- 1201.27 m. In the Lokammona Formation, there is some evidence of sis was performed using a primary ion current of ~17–21 nA and an secondary mineralisation, particularly at ~3,165 m where galena has acceleration voltage of 15 kV. MEYER Et al . 356       FIGURE 2 BH1- SACHA bulk sulphur isotope data plotted against core depth (m) with the corresponding sedimentary log. Brecciated zones largely reflect post- depositional tectonic deformation. The pyritised microbialites analysed in this study were sampled at a depth of 3,184 m (grey band). Lokam., represents Lokammona. Data from Izon et al. (2015) and stratigraphic log adapted from Altermann & Siegfried, 1997 [Colour figure can be viewed at wileyonlinelibrary.com] a high spatial resolution, precision and mass resolution while minimis- 2.3 | Multiple sulphur isotope analyses via SIMS ing destruction to the sample. Epoxy thick sections of sample- 3184 The sulphur isotope data are reported relative to the Vienna Canyon with a 2.54 cm diameter were made, polished to <1 μm surface rough- Diablo Troilite (V- CDT) international reference standard. The delta (δ) ness and covered with a ~50- nm gold coating. notation denotes the deviation of the sample from V- CDT in permil −8 Samples were outgassed in an ancillary chamber at <1 × 10 (‰; Equations 1–2). The capital- delta (Δ) reflects mass- independent mB prior to analysis, and the pressure in the sample chamber was fraction, quantifying the deviation of a sample from the expected ter- allowed to equilibrate for >1 hr aerft sample introduction, prior to restrial mass fractionation line (Equation 3). analysis. Thick sections were analysed in a vacuum with a pressure of −9 + 3 × 10 mB. A focused ~2 μm Cs primary ion beam of 0.9–1.8 nA was 33 32 ( S∕ S) sample δ S= −1 ×1000 (1) rastered over a 10 μm by 10 μm area of interest. Each measurement 33 32 ( S∕ S) V−CDT was divided into 15 cycles and required 7–8 min to complete. The 34 32 secondary ions were collected using faraday cups (FC). The 7f- GEO is ( S∕ S) sample δ S= −1 ×1000 (2) 34 32 specifically designed for precise isotope ratio measurement using two ( S∕ S) V−CDT FCs optimised for major (FC1) and minor (FC2) isotope data acquisition � � 0.515 ⎛ ⎞ in “charge mode” (Peres, de Chambost, & Schuhmacher, 2008). Each δ S V−CDT 33 33 ⎜ ⎟ Δ S=δ S −1000× 1− V−CDT ⎜ ⎟ (3) isotope is selected by magnet switching, and the corresponding sec- ⎝ ⎠ ondary ion counts were collected in sequence of ascending mass. The 34 33 Multiple sulphur isotope analyses (δ S and δ S) were conducted desired FC is selected using a deflector situated in the detector assem- using the CAMECA IMS 7f- GEO secondary ion mass spectrome- bly. This configuration allows for the initiation of data acquisition from ter housed in the Department of Earth and Planetary Sciences at FC2 prior to complete dissipation of the signal from FC1. This alternat- Washington University (St Louis, MO, USA). Secondary ion mass spec- ing collection arrangement results in improved signal- to- noise ratios trometry (SIMS) enables the in situ measurement of S- isotope ratios to and thereby greater precision per unit time. In this study, two minor MEYER Et al .       357 isotopes were to be analysed. As back- to- back data collection using R, respectively. Although IMF can vary according to instrumental FC2 would defeat the signal- to- noise advantage described above; thus, conditions, the S- isotope offset between bracketing standards and un- a pseudomeasurement, at mass 33.5, was made on FC1 so that each knowns should not vary. Furthermore, the measured S- isotope ratios 32 33 34 cycle became S (FC1), S(FC2), 33.5(FC1), S(FC2). The background of a session’s bracketing standards were observed to determine drift. 4 3 level for FC1 was ~3 × 10 counts/s and for FC2 ~3 × 10 counts/s. The drift value was usually <1‰ over 24 hr. To correct for this, a linear 32 8 33 6 Typically, secondary ion yields were: S~2 × 10 counts/s, S~10 and constant drift was assumed within a session. The magnitude of 34 6 counts/s and S~10 counts/s. this drift was much smaller than the signals observed and was there- Background and ion yields were recalibrated on a daily basis. Mass fore unlikely to have impacted the results. calibration, as well as automatic peak centring of the field and contrast apertures, was performed at the start of each measurement; this corrects 3  |  RESULT S for any drift in the secondary ion beam or the magnetic field (Fike et al., 2009). To remove the 50- nm gold coating, each measurement was pre- 3.1 | Textural analysis sputtered for 2 min prior to the start of the 15 cycles. A minimum mass 33 32 resolving power (MRP) of 3900 is required to separate the S and SH Sample- 3184 is a black shale composed of ~1- mm- thick laminae with peaks, and therefore, a MRP of ~4300 was used. Analysis was preferen- varying proportions of pyrite and matrix (Figure 3). The matrix is com- tially carried out in regular spaced intervals in a grid- like paern tt (Fike, posed of silt- sized particles (~40 μm) of aluminosilicates (clay min- Gammon, Ziebis, & Orphan, 2008). Generally, the analysis spots were erals; 80% modal abundance) and quartz (20%). Uncommon <5- μm placed at ~50- μm intervals. However, deviations from the ideal grid were detrital grains of rutile (TiO ) and apatite (Ca (PO ) (F,Cl,OH) ) make 2 10 4 6 2 necessary to consistently sample pyrite instead of siliciclastic matrix. up <1% of the matrix volume. Laminae contain variable proportions The internal error (standard error of n = 15 cycles) varied inversely of pyrite- to- matrix ratios, ranging from 10%–70% pyrite to 30%–90% with the S count rate (Figure S1), suggesting that precision was lim- matrix. There are three distinguishable pyrite phases: A) disseminated ited by counting statistics (Fike et al., 2009). Anomalous data were pyrite crystals, <5 μm in size; B) type 1 pyrite: irregular aggregates of omitted from further analysis if any of the following applied to the coalesced pyrite crystals. Clusters can range from 10 μm to 500 μm measurement: across, and are composed of anhedral to subhedral, 1- to 20- μm pyrite crystals. (Figure 4); C) type 2 pyrite: euhedral cubic pyrite with a crystal 1. The sputter area clearly targeted only matrix as subsequently size ranging from 10 to 100 μm (Figure 4). Some euhedral pyrite crys- determined from BSE and super-resolution digital microscope tals are isolated within the matrix, while other crystals overgrow and images. encrust type 1 aggregates. Overgrowths occur on the edges of type 1 2. The uncorrected relative standard error for S counts was >1%. pyrite aggregates and within clusters where matrix lenses occur. Type 34 32 3. The uncorrected relative standard error for S/ S was >0.1%. 1 and type 2 pyrites are chemically distinct phases, as determined from BSE images (Figures 4 and S6). Aerft excluding the anomalous measurements, the internal error Laminae containing type 1 pyrite show sedimentary structures 34 33 was typically <1‰ for δ S and <0.5‰ for δ S (1SE). An in- house that suggest it precipitated during early diagenesis, prior to soft sed- Washington University pyrite standard was mounted in a polished and iment deformation. These include wavy crinkled laminae with typical gold-c oated thin section and placed in a separate holder. Its composition wavelengths of ~500 μm and wave heights of ~200 μm, isoclinal folds, was determined by measuring the S- isotope composition of the internal 1- mm rip- up structures and over- folded laminae. Particularly, <2- mm standard 12 times and the Balmat standard 20 times. The Balmat stan- irregular, ellipsoidal concretions are composed of polycrystalline type 34 33 dard (δ S = 15.1‰ and δ S = 7.7‰) analyses bracketed the in- house 1 pyrite. They cross- cut lamination but also cause draping of laminae standard. The in-house standard has a known isotope composition of on either side of the concretion (Figures 4 and S4); this suggests type 34 33 δ S = 0.13 ± 0.30‰ and δ S = 0.13 ± 0.20‰ (1SE). The external errors 1 pyrite precipitated aerft deposition but before lithification, during (1 standard deviation of multiple adjacent points on the in- house stan- early diagenesis. 34 33 dard) were typically 0.38‰ and 0.32‰ for δ S and δ S, respectively A normal microfault cross- cuts the entire core section of sam- (n = 35). Within sample- 3184, there were 26‰ variability in measured ple- 3184 (Figure 3). The ~50- μm- wide microfault has been infilled by 34 33 δ S values and 14‰ variability in δ S. This variation far exceeds the aluminosilicates and silica. The microfault cross- cuts type 1 pyrite; type calculated internal and external errors. 2 pyrite crystals cross- cut the microfault itself (Figure 4). Furthermore, On average, a session consisted of ~50 measurements of un- type 2 pyrite overgrows the sedimentary structures composed of type knowns and was bracketed by ≥4 measurements of the in- house stan- 1 pyrite. According to these cross- cutting relationships, type 2 pyrite dard on both sides of the unknowns (Figure S2). Instrumental mass precipitated post- lithification. fractionation (IMF) was corrected for by standard-sample bracketing. The mean, uncorrected isotope composition of repeated analyses of 3.2 | Multiple sulphur isotope data the in- house pyrite standard on the SIMS was δ S = −0.05 ± 0.38‰ and δ S = 1.00 ± 0.32‰ (1σ; n = 35; relative to V- CDT). Therefore, All the SIMS data measured in sample- 3184 are presented in the IMF (IMF = R /R ) was typically 1.000 and 1.001 for R and Figure 5. There is no obvious stratigraphic trend; however, there is raw known MEYER Et al . 358       (a) (b) FIGURE 3 (a) Reflected light scanner image of sample- 3184 from the Lokammona Formation showing the areas of SIMS analyses (white dots). Blue dot = Figure 6 analysis area; green dot = Figure 7 analysis area. Divisions on the scale bar represent 1 mm. (b) A trace of the image in (a) showing laminae compositions, type examples of sedimentary structures discussed in the text and the location of the normal microfault. Modal percentages of laminae compositions are shown in brackets (pyrite/matrix) [Colour figure can be viewed at wileyonlinelibrary.com] a relationship between pyrite crystal shape and S- isotope composi- and type 2 pyrite end members. When these measurements were sub- tions. The first data set shows a mean of δ S = 12.54 ± 4.98‰ and sequently examined in their petrographic contexts, photomicrographs Δ S = −0.21 ± 0.65‰ (±1σ, n = 177) and corresponds to type 1 py- suggest the analysis spots targeted a mixture of both pyrite phases. rite (Figures 6 and 7). In sample- 3184, type 1 pyrite δ S values vary by 26‰; within a ~500 × 500 μm area, the range of δ S is 15‰ 4  |  DISCUSSION 34 33 (Figure 6). Wavy crinkled laminae show δ S and Δ S values consist- ent with typical type 1 pyrite (Figure S3). The second data set has a 4.1 | Early- diagenetic (type 1) pyrite 34 33 mean of δ S = 8.36 ± 1.16‰ and Δ S = 5.54 ± 1.53‰ (±1σ, n = 18) and was measured in type 2 pyrite (Figures 7 and S5). To a 95% confi- Laminae composed of type 1 pyrite are wavy and crinkled on a dence level, the S- isotope values for type 1 and type 2 pyrite are signif- sub- mm scale (Figures 3 and 4) and are consistent with sedimen- icantly different in their medians and variances (p = .000 for δ S and tary structures that have been microbially induced (Noffke, 2009). Δ S in two- sample Wilcoxon and Levene’s tests). Some data points Prokaryotes and eukaryotes in microbial mats produce a matrix correspond to an intermediary S- isotope signature between the type 1 of extracellular polymeric substances (EPS), which are composed MEYER Et al .       359 FIGURE 4 BSE images of different sedimentary structures in sample- 3184. (a) Wavy crinkled internal lamination composed of type 1 pyrite and disseminated pyrite. (b) Wavy crinkled internal lamination composed of type 1 pyrite with type 2 pyrite overgrowths. (c) Pyrite concretion. Note that the concretion both cross- cuts and causes deformation of the lamination. (d) The normal microfault that is infilled by quartz and clay minerals. Note the microfault and vein cross- cut type 1 pyrite; type 2 pyrite cross- cuts the microfault and vein. Therefore, the relative order of formation is as follows: pyrite 1 precipitated first, brittle deformation caused microfault formation, the microfault was infilled, and finally type 2 pyrite precipitated. py, py 1, py 2 and qtz represent pyrite, type 1 pyrite, type 2 pyrite and quartz, respectively [Colour figure can be viewed at wileyonlinelibrary. com] FIGURE 5 Plot of multiple S- isotope 34 33 data (δ S and Δ S) measured via SIMS. SIMS error bars are 1SE for each measurement (n = 15 cycles). Green triangles represent a mixed signal, where the area of analysis sampled both type 1 and type 2 pyrite. The orange line is the Archaean reference array as described by Ono et al. (2003) [Colour figure can be viewed at wileyonlinelibrary.com] of polysaccharides, proteins, humic substances and nucleic acids is eroded and turned over at its edges (Schieber, 1999). However, (Nielsen, Jahn, & Palmgren, 1997). The matrix has several physi- wavy crinkled laminae can also form through the differential com- cal functions, including adhesion to surfaces, aggregation of cells, paction of phyllosilicates around, for example, microconcretions, silt stabilisation of microbial mat structure, sorption of exogenous or- lenses or silica spherules (Schieber, 2007). Examples of compaction ganic molecules and retention of water (Laspidou & Rittmann, 2002). around such structures were absent in BSE images of sample- 3184 Importantly, EPS is responsible for the formation of wavy crinkled (Figure 4). Thus, the sedimentary structures indicate high cohesion lamination due to its ability to trap detrital grains and its high cohe- of the depositional layers and are consistent with typical microbialite sive properties. This cohesion is also responsible for the formation textures, supporting a biogenic control on the formation of type 1 of over- folded layers (Figure 3), which occur when the mat surface pyrite. MEYER Et al . 360       FIGURE 6 Combined petrography with SIMS S- isotope data shows the typical 34 33 textural, δ S and Δ S characteristics of type 1 pyrite. (a) Reflected light photomicrograph of the analytical grid location. (b) BSE image of the analysis area overlain by a SIMS δ S image constructed by spline interpolation of the analytical grid (n = 31). Note that the analysis area is composed of a ~500-μ m aggregate of anhedral to subhedral, 1- to 20- μm pyrite crystals. (c) Plot to show SIMS δ S against 33 34 Δ S data (‰). Note the variation in δ S data is significant; the variation in Δ S data is not significant. Error bars represent 1SE for each measurement (n = 15 cycles) (d) BSE image of the analysis area overlain by a SIMS Δ S image constructed by spline interpolation of the analytical grid (n = 31) [Colour figure can be viewed at wileyonlinelibrary.com] 34 34 32 34 The S- isotope composition of type 1 pyrite is distinguishable by R = S/ S) and variable seawater δ S values. Figure 5 shows sulphate 33 34 a Δ S value of −0.21 ± 0.65‰ (±1σ) and a 26‰ range in δ S within that few data occur above 18‰, and thus, the curve showing the 33 34 sample- 3184. The Δ S error bars for type 1 pyrite within an area of δ S of the instantaneous product must be in the distillative tail sulphide 33 34 analysis overlap (Figure 6); therefore, the Δ S variation within a lamina when δ S > 18‰. Values of 6‰–12‰ for α , 4‰–16‰ for sulphide MSR 34 34 is statistically insignificant. The considerable variation in δ S values is δ S and f values from ~0.5 to 0.9 are all consistent with the sulphate consistent with MSR—modern microbial mats with sulphate reducers data (Figure 8). show ~15–53‰ variations in δ S on a 1- mm scale (Fike et al., 2009; This 6‰–12‰ estimate of α is at the lower end of the MSR Wilbanks et al., 2014), similar to what we observe here. Thus, both spectrum for the modern MSR cell- specific fractionation factors of textural evidence and the S- isotope composition are consistent with α = 2‰–66‰ (Fike et al., 2015). Small MSR isotope fractionation MSR the expected biosignatures of an ancient microbial mat containing can occur when sulphate concentrations are low, when sulphate re- sulphate reducers. The 26‰ range in δ S within sample- 3184 could duction rates or sulphide oxidation rates are high, when H is used be explained by non- uniformity in MSR fractionation factors, which as an electron donor instead of organic carbon, and/or when sulphur depend on the microbe strain, type of electron donors, reduction rate, disproportionation levels are low (e.g., Fike et al., 2009 and references temperature and the presence of additional S metabolisms such as ox- therein). Although the relationship between sulphate concentration idation and disproportionation (e.g., Fike et al., 2009). and MSR fractionation factor is complex (see Bradley et al., 2016), Alternatively, the scaertt and enrichment in δ S sulphide relative our estimates of α are consistent with other studies that suggest MSR to coeval sulphate could be produced by fractionation during MSR in low seawater sulphate concentrations and high rates of MSR in the a (partially) closed system, analogous to a mat environment periodi- Archaean (e.g., Crowe et al., 2014; Habicht et al., 2002; Zhelezinskaia cally flushed with seawater sulphate. We used a Rayleigh distillation et al., 2014). Furthermore, the seawater δ S compositions we esti- model to explore this scenario, using a suite of realistic fractionation mate from closed- system modelling of the SIMS data are consistent 34 34 34 factors for MSR (α = 1,000 × [ R / R − 1], where with estimates of seawater sulphate δ S from Neoarchaean CAS MSR sulphate sulphide MEYER Et al .       361 FIGURE 7 The contrasting textural and S- isotope signatures of type 1 and type 2 pyrite. (a) Reflected light photomicrograph of the area of analysis. (b) BSE image of the analysis area overlain by a SIMS δ S image constructed by spline interpolation of the analytical grid (n = 15). Note the ~300- μm cubic type 2 pyrite grain, overgrowing type 1 pyrite aggregates. (c) Plot to show 34 33 SIMS δ S against Δ S data (‰). Note the significant difference in Δ S for type 1 and 2 pyrite. Error bars represent 1SE for each measurement (n = 15 cycles). Blue circles = type 1 pyrite; green circles = type 2 pyrite. (d) BSE image of the analysis area overlain by a SIMS Δ S image constructed by spline interpolation of the analytical grid (n = 15) [Colour figure can be viewed at wileyonlinelibrary.com] (e.g., Domagal- Goldman, Kasting, Johnston, & Farquhar, 2008; Guo crystals have all been formed from the same, uniform S- isotope 34 33 et al., 2009; Paris et al., 2014). However, the small MSR fractionation source. The δ S and Δ S signature of type 2 pyrite falls within Ono factors and variation in Δ S values in the literature imply Archaean et al.’s (2003) estimates of the composition of atmospheric elemen- seawater sulphate had a short residence time due to a small reser- tal sulphur as inferred from Archaean pyrite data and photochemical voir size relative to the fluxes into and out of the reservoir; thus, the experiments (Figure 5), and Paris et al.’s (2014) Neoarchaean CAS val- 34 33 S- isotope composition of seawater sulphate was likely spatially and ues. Similar δ S and Δ S ratios in euhedral grains sampled from GKF- temporally heterogeneous (Fischer et al., 2014; Zhelezinskaia et al., 01 (a core sampling deeper water equivalents of the BH1- SACHA; 2014). Schröder, Lacassie, & Beukes, 2006) were interpreted as having an We therefore conclude that type 1 pyrite reflects the morphologi- atmospheric elemental sulphur origin (Farquhar et al., 2013). Farquhar cal and geochemical signatures of sulphate reducers in a Neoarchaean et al. (2013) hypothesised that solid atmospherically derived S parti- microbial mat, as inferred from (i) the biogenicity of sedimentary cles could remain in an unreactive form as they fell through the water structures like wavy crinkled lamination, (ii) textural evidence of early- column. After deposition, they could then react with H S in the sedi- diagenetic precipitation of type 1 pyrite, and (iii) models of type 1 ment, producing reactive polysulphide. Finally, the polysulphide could pyrite δ S values supporting plausible MSR fractionation factors ex- react with FeS to form pyrite. Farquhar et al. (2013) predicted that the pressed within a restricted seawater sulphate pool. pyrite would closely reflect the S- isotope signature of atmospheric elemental sulphur if the mass contribution of H S was small relative to polysulphide in the pyrite product. 4.2 | Secondary (type 2) pyrite A similar interpretation of type 2 pyrite formation in sample- 3184 As discussed above, simple cross- cutting relations suggest type 2 py- is less likely, due to its inferred later timing of formation based on rite precipitated after type 1 pyrite formation and post- lithification, the petrographic relationships we describe above. As type 1 pyrite because it overgrows deformed aggregates of type 1 pyrite, and it likely formed from sulphide generated via microbial sulphate reduc- cross- cuts a normal microfault with vein infill (Figure 4). The low tion, H S was abundant during early diagenesis. Therefore, any atmo- standard deviation of type 2 pyrite S- isotope ratios suggests these spheric, unreactive elemental sulphur particles that deposited into the MEYER Et al . 362       FIGURE 8 Graphs showing the calculated δ S values of the hydrogen sulphide instantaneous product relative to the proportion of sulphate consumed (f), the initial δ S composition and the MSR fractionation factor (α ). The grey rectangles correspond to the range of sulphate source-product δ S compositions measured in type 1 pyrite in sample- 3184. (a) α = 6‰, (b) α = 12‰ [Colour figure can be viewed at wileyonlinelibrary.com] sulphide Neoarchaean microbial mat would have simultaneously reacted with noted evidence of thrust faults in BH1- SACHA at depths >2,800 m in H S to form a soluble, reactive form of polysulphide. During early dia- the Monteville Formation. They suggested this is the intersection of genesis, the mobile polysulphide, a precursor to pyrite synthesis, would BH1- SACHA with the Kheis sole thrust fault (Figure 1). Allochthonous have reacted with an iron monosulphide to form FeS . However, petro- sections of Archaean Transvaal Supergroup rocks (including rocks graphic relationships suggest type 2 pyrite formed post- lithification, that were sampled by the BH1- SACHA core) as well as Proterozoic not coevally with type 1 pyrite during diagenesis (Figure 4). Therefore, Waterberg and Olifantshoek Groups were horizontally displaced and type 2 pyrite in sample- 3184 unlikely formed from an elemental sul- now lie unconformably on the autochthonous Archaean package of phur source by the mechanism proposed by Farquhar et al. (2013). rocks (Altermann & Wotherspoon, 1995; Martini, Eriksson, & Snyman, Hence, we suggest the Δ S = 5.5‰ signature of type 2 pyrite was not 1995). The age assigned to thrusting ranges from 2.20 to 1.75 Ga produced from contemporaneous atmospheric S products at the time (Grobbelaar, Burger, Pretorius, Marais, & Van Niekerk, 1995). If the sample- 3184 was formed. Instead, we propose later- stage alteration thrust is older than ~2.0 Ga, it suggests that the hanging wall rocks processes could explain the Δ S values measured in type 2 pyrite. were exhumed before the key igneous and metasomatic events asso- The BH1-S ACHA core has been altered by several post- ciated with the Bushveld Complex. Hence, the metamorphic grade of depositional hydrothermal events that could have caused multi - BH1- SACHA in the hanging wall may be lower than the footwall MVT generational pyrite genesis. Stratigraphic logs indicate evidence of deposits examined by Huizenga et al. (2006). In addition, De Kock secondary mineralisation of galena, ~20 m stratigraphically higher et al. (2009) showed that similar strata preserved within the GKP- 01 than sample- 3184. Furthermore, intrusion of the igneous body at drill core through the Ghaap Group show pervasive remagnetisation 922.3–1201.27 m formed a contact metamorphic aureole (Altermann by 2.5- to 1.8- Ga nanoscale pyrrhotite. Furthermore, high- resolution & Siegfried, 1997). To the best of our knowledge, this dyke has not palaeomagnetic and geochemical evidence from GKF- 01 Neoarchaean been dated; however, it may be co- genetic to other mafic intrusions nodular pyrite suggests that some of the Transvaal Supergroup pyrites related to the emplacement of the Bushveld complex, 2.06–2.05 Ga have been post- depositionally altered, ~0.5 Ga aerft their deposition (Hartzer, 1995). Associated igneous and metamorphic fluids can ca- (Fischer et al., 2014). Thus, even undeformed Ghaap Group strata can talyse metasomatic reactions, causing the loss and gain of elements show a complex history of iron sulphide mineral precipitation. in circulating fluids. The occurrence of metasomatism in BH1- SACHA The geological evidence therefore suggests sulphur could have is supported by regional geological evidence; a plaorm- tf wide fluid circulated in fluids as soluble sulphur species and participated in hy- flow event at ~2.0 Ga could have caused the Pb- Zn- Cu- mineralisation drothermal reactions in the Griqualand West Basin. There are multiple in the Campbellrand Subgroup, Ghaap Group. Huizenga, Gutzmer, sources of subsurface S- rich fluids (Figure 9): In modern settings, sul- Greyling, and Schaefer (2006) sampled Mississippi Valley- type (MVT) phate can be acquired from (buried) seawater (Ohmoto, 1972). If the hy- deposits in the Griqualand West area, ~100–120 km straight- line drothermal alterations occurred aerft the GOE, seawater was a potential distance from Kathu. They estimated a regional fluid temperature of source (A, Figure 9). Furthermore, pore waters expelled from sedimen- 200–240°C and pressure of 0.8–1.5 kbar during mineralisation of the tary rocks during compaction, fluids formed by magmatic/metamor - MVT deposits. It is possible that BH1- SACHA experienced similar phic processes or meteoric- or seawater- derived fluids can cause the metamorphic conditions. However, Altermann and Siegfried (1997) dissolution of S- bearing minerals. Dissolution of sulphate, such as CAS MEYER Et al .       363 FIGURE 9 Hypothesised formation of post- lithification pyrite with an anomalous Δ S signal. Sulphur can be sourced via three pathways: (A) seawater, (B) dissolution of sulphate- or sulphide- bearing minerals or (C) the disproportionation of SO in magmatic- hydrothermal or magmatic steam environments [Colour figure can be viewed at wileyonlinelibrary.com] 33 34 or sulphate- bearing minerals (e.g., gypsum, anhydrite or barite), occurs of S relative to S in the polysulphide product, and therefore, the readily, while pyrite dissolution can also occur in the presence of an oxi- TSR product carries a positive Δ S signal (Oduro et al., 2011). 3+ dant, for example, O or Fe (Descostes, Vitorge, & Beaucaire, 2004; B, Thermochemical sulphate reduction occurs at 100–300°C Figure 9). Moreover, both sulphate and sulphide can be produced from (Johnston, 2011); these temperatures are equivalent to sub- the disproportionation of magmatic SO . For example, during the con- greenschist to greenschist facies, which correspond to the maximum densation of a magmatic plume, H SO and H S form from the dispro- metamorphic grade that the lower Transvaal Supergroup, including 2 4 2 portionation of magmatic SO at temperatures 200–400°C and pH < 3. BH1- SACHA, experienced during intrusion of the Bushveld Complex Sulphate can be leached by fluids, causing the production of an acidic, in the Palaeoproterozoic (Sumner & Beukes, 2006). Therefore, the sulphate- bearing hydrothermal fluid that can infiltrate surrounding wall temperature regime during contact metamorphism of BH1- SACHA rock (C, Figure 9; Rye, 2005). The intrusion of the Bushveld Complex and the surrounding rocks was sufficiently high to support TSR. The at ~2.06 Ga could have produced magmatic sulphur species (Hartzer, petrographic evidence and isotope fingerprint of type 2 pyrite, cou- 1995). Finally, these subsurface S- rich fluids could have infiltrated sur- pled with the history of hydrothermal alteration described above, rounding rocks via fracture flow or porous flow, and precipitated pyrite. could be consistent with MIE during TSR causing an anomalous en- 33 34 We suggest two scenarios whereby the circulation of S- rich fluids richment in S relative to S. However, there is currently no evidence from the above sources could have contributed to the formation of of anomalous Δ S signals associated with TSR in bulk rock analyses type 2 pyrite and its associated S- MIF signal: one via thermochemi- from the geological record. Furthermore, TSR causes a MIF eect ff in 33 36 cal sulphate reduction of S- bearing fluids from a more recent sulphur S, but not S (Oduro et al., 2011); therefore, this hypothesis cannot source, and one via migration of S- rich fluids from stratigraphically be conclusively demonstrated without the inclusion of Δ S data. older sediments carrying an inherited Archaean S- MIF signal. More likely, the MIF signal measured in type 2 pyrite could have If the subsurface sulphur- rich fluid were sourced from modern or formed from the dissolution and reprecipitation of Archaean S- bearing post- GOE seawater, the dissolution of post- GOE sulphur minerals or minerals originally deposited with a positive Δ S signature (Figure 9). magmatic sulphate, it would not carry a primary MIF signal. Therefore, Notably, bulk pyrite shows a Δ S composition within error of the type to precipitate type 2 pyrite with a Δ S fingerprint of +5.5‰ from one 2 pyrite S- isotope fingerprint, 20–30 m stratigraphically lower than of these sources, a post- depositional process would be necessary to sample- 3184 (Figure 2). Therefore, dissolution of these pyrite phases cause S- MIF. Experimental investigations have shown that thermo- and migration of Δ S = +5.5‰ sulphur- rich fluids upwards are a likely chemical sulphate reduction (TSR) can produce a < 13‰ enrichment in source of the type 2 pyrite MIF signal. Therefore, grain- scale Archaean Δ S relative to the sulphate source (Oduro et al., 2011). An anomalous MIF signatures may reflect secondary pyrite formation mechanisms as 33 34 enrichment of S relative to S during TSR occurs due to magnetic well as atmospheric processes at the time of deposition. These mul- isotope eects ff (MIE). In the sulphur isotope system, the S- isotope tiple pyrite phases can reflect a protracted history of pyrite precipita- contains magnetic nuclei; during TSR, S undergoes the spin- selective tion, which can be recognised using high- resolution, in situ S- isotope 34 32 reaction faster than S and S. This results in an anomalous enrichment geochemistry. MEYER Et al . 364       NE/H016805/1 and Natural Environment Research Council Standard 5  |  C ONCL USIONS Grant NE/J023485/2 (to A.Z.) and an NSF/EAR Grant (#1229370) Sulphate reduction in Neoarchaean microbial mats: Our textural and sul- and a Packard Fellowship (to D.F.) phur isotope data echo other studies that suggest MSR influenced S- cycling in the Neoarchaean (e.g., Habicht et al., 2002; Ueno et al., 2008; REFERENCES Zerkle, Claire, Domagal- Goldman, Farquhar, & Poulton, 2012). As typi- cal microbialite textures are generally associated with photoautotrophs Altermann, W., & Siegfried, H. P. (1997). Sedimentology and facies devel- opment of an Archaean shelf: Carbonate plaorm tf transition in the (Noffke, 2009), this study proposes that microbial sulphate reducers may Kaapvaal Craton, as deduced from a deep borehole at Kathu, South have commonly occurred in the same community as photosynthesisers, Africa. Journal of African Earth Sciences, 24, IN1–IN4. forming a microbial mat ecosystem similar to those of the modern. Altermann, W., & Wotherspoon, J. M. (1995). The carbonates of the Neoarchaean seawater sulphate concentrations and S- isotope compo- Transvaal and Griqualand West Sequences of the Kaapvaal craton, with special reference to the Lime Acres limestone deposit. Mineralium sition: SIMS S- isotope data and closed- system modelling suggest the Deposita, 30, 124–134. local sulphur reservoir had a Δ S = −0.21 ± 0.65‰ (±1σ, n = 177) and Baroni, M., Thiemens, M. H., Delmas, R. J., & Savarino, J. (2007). Mass- δ S = 4‰–12‰ during formation of type 1 pyrites in sample- 3184 independent sulfur isotopic compositions in stratospheric volcanic (Figure 8). These values are consistent with previous estimates of eruptions. Science, 315, 84–87. Beukes, N. J. (1987). Facies relations, depositional environments and dia- the δ S- isotope composition of Archaean sulphate (e.g., Domagal- genesis in a major early Proterozoic stromatolitic carbonate plaorm tf Goldman et al., 2008; Guo et al., 2009; Paris et al., 2014). However, to basinal sequence, Campbellrand Subgroup, Transvaal Supergroup, the S- isotope composition of seawater sulphate may have been spa- Southern Africa. Sedimentary Geology, 54, 1–46. tially and temporally variable, particularly at the low sulphate levels Bradley, A. S., Leavitt, W. D., Schmidt, M., Knoll, A. H., Girguis, P. R., & Johnston, D. T. (2016). Paerns tt of sulfur isotope fractionation during estimated for the Neoarchaean (Zhelezinskaia et al., 2014). Closed- Microbial Sulfate Reduction. Geobiology, 14, 91–101. system modelling of sample- 3184 indicates small MSR fractionation Button, A. (1973). Algal stromatolites of the early Proterozoic Wolkberg factors (α = 6‰–12‰), consistent with high reduction rates and MSR Group, Transvaal sequence. Journal of Sedimentary Research, 43, 160–167. low sulphate concentrations. The small MSR fractionation factors and Claire, M. W., Kasting, J. F., Domagal-Goldman, S. D., Stüeken, E. E., Buick, scatter in Δ S values in the literature imply sulphate had a short resi- R., & Meadows, V. S. (2014). Modeling the signature of sulfur mass- independent fractionation produced in the Archean atmosphere. dence time in Archaean oceans due to a small reservoir size and com- Geochimica et Cosmochimica Acta, 141, 365–380. paratively large fluxes into and out of the reservoir. Additional studies Crowe, S. A., Paris, G., Katsev, S., Jones, C., Kim, S. T., Zerkle, A. L., … assessing the S- isotope composition of CAS will help constrain the Farquhar, J. (2014). Sulfate was a trace constituent of Archean seawa- 34 33 evolution of seawater sulphate δ S and Δ S during the Archaean. ter. Science, 346, 735–739. De Kock, M. O., Evans, D. A. D., Kirschvink, J. L., Beukes, N. J., Rose, E., & The utility of combining petrography and SIMS: This study demon- Hilburn, I. (2009). Paleomagnetism of a Neoarchean- Paleoproterozoic strates that grain- scale MIF signals could record secondary processes carbonate ramp and carbonate plaorm tf succession (Transvaal in addition to contemporaneous atmospheric processes at the time of Supergroup) from surface outcrop and drill core, Griqualand West re- deposition. MIF signals can be generated by MIE during later TSR or, gion, South Africa. Precambrian Research, 169, 80–99. Descostes, M., Vitorge, P., & Beaucaire, C. (2004). Pyrite dissolution in more probably, leached from surrounding Archaean sedimentary rocks acidic media. Geochimica et Cosmochimica Acta, 68, 4559–4569. and reprecipitated following lithification. Furthermore, it demonstrates Domagal-Goldman, S. D., Kasting, J. F., Johnston, D. T., & Farquhar, J. that Archaean rocks can preserve a complex, protracted history of pyrite (2008). Organic haze, glaciations and multiple sulfur isotopes in the formation. Pyrite genesis can occur over several generations; precipita- Mid- Archean Era. Earth and Planetary Science Leer tt s, 269, 29–40. Farquhar, J., Bao, H., & Thiemens, M. (2000). Atmospheric influence of tion can range from pre- so ft sediment deformation to post- lithification. Earth’s earliest sulfur cycle. Science, 289, 756–758. This study highlights the eectiv ff eness of pairing petrography with high- Farquhar, J., Cliff, J., Zerkle, A. L., Kamyshny, A., Poulton, S. W., Claire, resolution SIMS sulphur isotope studies, to unravel the diagenetic pro- M., … Harms, B. (2013). Pathways for Neoarchean pyrite formation cesses that have contributed to multigenerational pyrite genesis. constrained by mass- independent sulfur isotopes. Proceedings of the National Academy of Sciences, 110, 17638–17643. Fike, D. A., Bradley, A. S., & Rose, C. V. (2015). Rethinking the ancient sulfur A CKNO WLEDGMENT S cycle. Annual Review of Earth and Planetary Sciences, 43, 593–622. Fike, D. A., Finke, N., Zha, J., Blake, G., Hoehler, T. M., & Orphan, V. J. (2009). We thank Dr. Clive Jones for helping with SIMS analysis, Mr. Donald High resolution SIMS- based sulfide δ S: A new tool for characterizing Herd for assistance with imaging and the staff at the South African microbial activity in a variety of depositional environments. Geochimica et Cosmochimica Acta, 73, 6187–6204. National Core Library in Donkerhoek for facilitating access to the Fike, D. A., Gammon, C. L., Ziebis, W., & Orphan, V. J. (2008). Micron- scale core materials. The authors thank Tony Prave, Tim Raub, Gareth Izon, mapping of sulfur cycling across the oxycline of a cyanobacterial mat: A Mark Claire and James Farquhar for their input and discussion. This paired nanoSIMS and CARD- FISH approach. ISME Journal, 2, 749–759. study was financially supported by the Geological Society of London’s Fischer, W. W., Fike, D. A., Johnson, J. E., Raub, T. D., Guan, Y., Kirschvink, J. L., & Eiler, J. M. (2014). SQUID–SIMS is a useful approach to uncover Alan & Charlotte Welch Fund, A Palaeontological Association primary signals in the Archean sulfur cycle. Proceedings of the National Undergraduate Research Grant (PA- UB201504) and the University Academy of Sciences, 111, 5468–5473. of St Andrew’s Zannah Stephen Memorial Travel Scholarship Grobbelaar, W. S., Burger, M. A., Pretorius, A. I., Marais, W., & Van Niekerk, (all to N.M.), Natural Environment Research Council Fellowship I. J. M. (1995). Stratigraphic and structural setting of the Griqualand MEYER Et al .       365 West and the Olifantshoek Sequences at Black Rock, Beeshoek and Paytan, A., Mearon, S., Cobb, K., & Kastner, M. (2002). Origin of marine bar- Rooinekke Mines, Griqualand West, South Africa. Mineralium Deposita, ite deposits: Sr and S isotope characterization. Geology, 30, 747–750. 30, 152–161. Peres, P., de Chambost, E., & Schuhmacher, M. (2008). CAMECA IMS 7f- Guo, Q., Strauss, H., Kaufman, A. J., Schröder, S., Gutzmer, J., Wing, B., … Geo: Specialized SIMS tool for geosciences. Applied Surface Science, Farquhar, J. (2009). Reconstructing Earth’s surface oxidation across the 255, 1472–1475. Archean- Proterozoic transition. Geology, 37, 399–402. Roerdink, D. L., Mason, P. R., Farquhar, J., & Reimer, T. (2012). Multiple Habicht, K. S., Gade, M., Thamdrup, B., Berg, P., & Canfield, D. E. (2002). sulfur isotopes in Paleoarchean barites identify an important role for Calibration of sulfate levels in the Archean ocean. Science, 298, 2372–2374. microbial sulfate reduction in the early marine environment. Earth and Hartzer, F. J. (1995). Transvaal Supergroup inliers: Geology, tectonic de- Planetary Science Leer tt s, 331, 177–186. velopment and relationship with the Bushveld Complex, South Africa. Rye, R. O. (2005). A review of the stable- isotope geochemistry of sulfate Journal of African Earth Sciences, 21, 521–547. minerals in selected igneous environments and related hydrothermal Huizenga, J. M., Gutzmer, J., Greyling, L. N., & Schaefer, M. (2006). Carbonic systems. Chemical Geology, 215, 5–36. fluid inclusions in Paleoproterozoic carbonate- hosted Zn- Pb deposits Schieber, J. (1999). Microbial mats in terrigenous clastics; the challenge of in Griqualand West, South Africa. South African Journal of Geology, 109, identification in the rock record. Palaios, 14, 3–12. 55–62. Schieber, J. (2007). Microbial mats on muddy substrates—examples of pos- Izon, G., Zerkle, A. L., Zhelezinskaia, I., Farquhar, J., Newton, R. J., Poulton, S. sible sedimentary features and underlying processes. Atlas of Microbial W., … Claire, M. W. (2015). Multiple oscillations in Neoarchaean atmo- Mat Features Preserved within the Siliciclastic Rock Record Atlases in spheric chemistry. Earth and Planetary Science Leer tt s, 431, 264–273. Geoscience, 2, 117–133. Johnston, D. T. (2011). Multiple sulfur isotopes and the evolution of Earth’s Schröder, S., Lacassie, J. P., & Beukes, N. J. (2006). Stratigraphic and geo- surface sulfur cycle. Earth- Science Reviews, 106, 161–183. chemical framework of the Agouron drill cores, Transvaal Supergroup Johnston, D. T., Farquhar, J., Wing, B. A., Kaufman, A. J., Canfield, D. E., & (Neoarchean–Paleoproterozoic, South Africa). South African Journal of Habicht, K. S. (2005). Multiple sulfur isotope fractionations in biological Geology, 109, 23–54. systems: A case study with sulfate reducers and sulfur disproportion- Shen, Y., & Buick, R. (2004). The antiquity of microbial sulfate reduction. ators. American Journal of Science, 305, 645–660. Earth- Science Reviews, 64, 243–272. Kamber, B. S., & Whitehouse, M. J. (2007). Micro- scale sulphur isotope evi- Shen, Y., Buick, R., & Canfield, D. E. (2001). Isotopic evidence for microbial dence for sulphur cycling in the late Archean shallow ocean. Geobiology, sulphate reduction in the early Archaean era. Nature, 410, 77–81. 5, 5–17. Sumner, D. Y., & Beukes, N. J. (2006). Sequence stratigraphic development Knoll, A. H., & Beukes, N. J. (2009). Introduction: Initial investigations of a of the Neoarchean Transvaal carbonate plaorm, tf Kaapvaal Craton, Neoarchean shelf margin- basin transition (Transvaal Supergroup, South South Africa. South African Journal of Geology, 109, 11–22. Africa). Precambrian Research, 169, 1–14. Ueno, Y., Ono, S., Rumble, D., & Maruyama, S. (2008). Quadruple sulfur iso- Laspidou, C. S., & Rittmann, B. E. (2002). A unified theory for extracellular tope analysis of ca. 3.5 Ga Dresser Formation: New evidence for micro- polymeric substances, soluble microbial products, and active and inert bial sulfate reduction in the early Archean. Geochimica et Cosmochimica biomass. Water Research, 36, 2711–2720. Acta, 72, 5675–5691. Martini, J. E. J., Eriksson, P. G., & Snyman, C. P. (1995). The early protero- Van Kranendonk, M. J. (2006). Volcanic degassing, hydrothermal circula- zoic Mississippi Valley- type Pb- Zn- F deposits of the Campbellrand and tion and the flourishing of early life on Earth: A review of the evidence Malmani subgroups, South Africa. Mineralium Deposita, 30, 135–145. from c. 3490–3240 Ma rocks of the Pilbara Supergroup, Pilbara Craton, Miyano, T., & Beukes, N. J. (1984). Phase relations of stilpnomelane, Western Australia. Earth- Science Reviews, 74, 197–240. ferri- annite, and riebeckite in very low- grade metamorphosed iron- Wagner, M., Roger, A. J., Flax, J. L., Brusseau, G. A., & Stahl, D. A. (1998). formations. South African Journal of Geology, 87, 111–124. Phylogeny of dissimilatory sulfite reductases supports an early origin of Nielsen, P. H., Jahn, A., & Palmgren, R. (1997). Conceptual model for pro- sulfate respiration. Journal of Bacteriology, 180, 2975–2982. duction and composition of exopolymers in biofilms. Water Science and Wilbanks, E. G., Jaekel, U., Salman, V., Humphrey, P. T., Eisen, J. A., Facciotti, Technology, 36, 11–19. M. T., … Orphan, V. J. (2014). Microscale sulfur cycling in the phototro- Noe,ffk N. (2009). The criteria for the biogeneicity of microbially induced phic pink berry consortia of the Sippewissett Salt Marsh. Environmental sedimentary structures (MISS) in Archean and younger, sandy deposits. Microbiology, 16, 3398–3415. Earth- Science Reviews, 96, 173–180. Zerkle, A. L., Claire, M. W., Domagal-Goldman, S. D., Farquhar, J., & Poulton, Oduro, H., Harms, B., Sintim, H. O., Kaufman, A. J., Cody, G., & Farquhar, S. W. (2012). A bistable organic- rich atmosphere on the Neoarchaean J. (2011). Evidence of magnetic isotope eects ff during thermochemical Earth. Nature Geoscience, 5, 359–363. sulfate reduction. Proceedings of the National Academy of Sciences, 108, Zhelezinskaia, I., Kaufman, A. J., Farquhar, J., & Cliff, J. (2014). Large sulfur 17635–17638. isotope fractionations associated with Neoarchean microbial sulfate Ohmoto, H. (1972). Systematics of sulfur and carbon isotopes in hydrother- reduction. Science, 346, 742–744. mal ore deposits. Economic Geology, 67, 551–578. Ono, S., Beukes, N. J., & Rumble, D. (2009). Origin of two distinct multiple- sulfur isotope compositions of pyrite in the 2.5 Ga Klein Naute Formation, SUPPOR TING INF ORMATION Griqualand West Basin, South Africa. Precambrian Research, 169, 48–57. Additional Supporting Information may be found online in the support- Ono, S., Eigenbrode, J. L., Pavlov, A. A., Kharecha, P., Rumble, D., Kasting, J. F., & Freeman, K. H. (2003). New insights into Archean sulfur cycle from ing information tab for this article. mass- independent sulfur isotope records from the Hamersley Basin, Australia. Earth and Planetary Science Leer tt s, 213, 15–30. How to cite this article: Meyer NR, Zerkle AL, Fike DA. Paris, G., Adkins, J. F., Sessions, A. L., Webb, S. M., & Fischer, W. W. (2014). Sulphur cycling in a Neoarchaean microbial mat. Geobiology. Neoarchean carbonate–associated sulfate records positive Δ33S anomalies. Science, 346, 739–741. 2017;15:353–365. https://doi.org/10.1111/gbi.12227 Pavlov, A. A., & Kasting, J. F. (2002). Mass- independent fractionation of sulfur isotopes in Archean sediments: Strong evidence for an anoxic Archean atmosphere. Astrobiology, 2, 27–41. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Geobiology Wiley

Sulphur cycling in a Neoarchaean microbial mat

Meyer, N. R.; Zerkle, A. L.; Fike, D. A.
Geobiology , Volume 15 (3) – May 1, 2017
Download PDF
13 pages

Loading next page...
 
/lp/wiley/sulphur-cycling-in-a-neoarchaean-microbial-mat-3JdcQl6cu0

References (56)

Publisher
Wiley
Copyright
Copyright © 2017 John Wiley & Sons Ltd
ISSN
1472-4677
eISSN
1472-4669
DOI
10.1111/gbi.12227
pmid
28128527
Publisher site
See Article on Publisher Site

Abstract

University of St Andrews, St Andrews, UK Multiple sulphur (S) isotope ratios are powerful proxies to understand the complexity Department of Earth and Planetary Sciences, Washington University, St. Louis, of S biogeochemical cycling through Deep Time. The disappearance of a sulphur mass- MO, USA independent fractionation (S- MIF) signal in rocks <~2.4 Ga has been used to date a Correspondence dramatic rise in atmospheric oxygen levels. However, intricacies of the S- cycle before N. R. Meyer, Department of Earth System the Great Oxidation Event remain poorly understood. For example, the isotope com- Science, Stanford University, Stanford, CA, USA. position of coeval atmospherically derived sulphur species is still debated. Furthermore, Email: nrmeyer@stanford.edu variation in Archaean pyrite δ S values has been widely attributed to microbial sul- and A. L. Zerkle, School of Earth and phate reduction (MSR). While petrographic evidence for Archaean early- diagenetic Environmental Sciences, University of pyrite formation is common, textural evidence for the presence and distribution of St Andrews, St Andrews, UK. Email: az29@st-andrews.ac.uk MSR remains enigmatic. We combined detailed petrographic and in situ, high- 34 33 resolution multiple S- isotope studies (δ S and Δ S) using secondary ion mass spec- Current address N. R. Meyer, Department of Earth System trometry (SIMS) to document the S- isotope signatures of exceptionally well- preserved, Science, Stanford University, Stanford, CA pyritised microbialites in shales from the ~2.65- Ga Lokammona Formation, Ghaap 94305, USA Group, South Africa. The presence of MSR in this Neoarchaean microbial mat is sup- Funding information ported by typical biogenic textures including wavy crinkled laminae, and early- Geological Society of London’s Alan & Charloett Welch Fund; A Palaeontological diagenetic pyrite containing <26‰ μm- scale variations in δ S and Association Undergraduate Research, Grant/ 33 34 Δ S = −0.21 ± 0.65‰ (±1σ). These large variations in δ S values suggest Rayleigh Award Number: PA-UB201504; University of St Andrew’s Zannah Stephen Memorial Travel distillation of a limited sulphate pool during high rates of MSR. Furthermore, we iden- Scholarship; Natural Environment Research tified a second, morphologically distinct pyrite phase that precipitated after lithifica- Council Fellowship, Grant/Award Number: 34 33 NE/H016805/1; Natural Environment tion, with δ S = 8.36 ± 1.16‰ and Δ S = 5.54 ± 1.53‰ (±1σ). We propose that the Research Council Standard, Grant/Award S- MIF signature of this secondary pyrite does not reflect contemporaneous atmos- Number: NE/J023485/2; NSF/EAR, Grant/ Award Number: 1229370; Packard Fellowship pheric processes at the time of deposition; instead, it formed by the influx of later- stage sulphur- bearing fluids containing an inherited atmospheric S- MIF signal and/or from magnetic isotope effects during thermochemical sulphate reduction. These in- sights highlight the complementary nature of petrography and SIMS studies to resolve multigenerational pyrite formation pathways in the geological record. 33 34 1  |  INTR ODUCTION follow a thermodynamically determined, linear δ S/δ S relationship 33 34 (δ S = 0.515 × δ S). However, the geological S- isotope record also The sulphur isotope record has played an integral role in shaping captures evidence of mass- independent fractionation (MIF), where 33 34 our understanding of key events in Earth’s geological and biolog- δ S and δ S deviate from the predicted terrestrial mass fractionation ical history. Surficial S- cycling principally involves biological and line, quantified by the capital- delta (Δ) notation. Prior to the Great abiotic mass- dependent fractionation (MDF) processes, which Oxidation Event (GOE) at ~2.4 Ga, S- bearing minerals show large This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2017 The Authors Geobiology Published by John Wiley & Sons Ltd. Geobiology. 2017;15:353–365. wileyonlinelibrary.com/journal/gbi     353 MEYER Et al . 354       33 33 34 33 positive and negative Δ S values. Aerft ~2.4 Ga, Δ S ratios dimin- S- isotope ratios that largely follow a linear trend in δ S versus Δ S ish towards values that tightly cluster around zero (Δ S = 0 ± 0.2‰). along an Archaean reference array. Combining these data with photo- Archaean sulphur MIF has been widely attributed to atmospheric chemical models, Ono et al. (2003) suggested that the Δ S sign was photochemical reactions involving SO ; these reactions would have positive for elemental sulphur and negative for sulphate. Furthermore, been blocked in the Palaeoproterozoic due to the UV- shielding eects ff Palaeoarchaean barites (BaSO ) also show a 0 to −1.5‰ Δ S signal caused by increased atmospheric O and O concentrations (e.g., (e.g., Roerdink, Mason, Farquhar, & Reimer, 2012; Ueno, Ono, Rumble, 2 3 Farquhar, Bao, & Thiemens, 2000). In addition, the delivery of S- MIF to & Maruyama, 2008), but these could have formed in the marine water the Earth’s surface requires sulphur to leave the atmosphere via mul- column, in the sediments or in hydrothermal environments not neces- tiple exit channels at different redox states, which are homogenised sarily reflective of the seawater sulphate pool (Paytan, Mearon, Cobb, & −5 when atmospheric oxygen exceeds 10 of present atmospheric lev- Kastner, 2002; Van Kranendonk, 2006). Additional photochemical exper- els (Pavlov & Kasting, 2002). Therefore, measuring multiple sulphur iments produce S- isotope slopes that contrast to Ono et al.’s (2003) re- 34 33 isotopes (δ S, Δ S) in Archaean sediments can provide information sults (e.g., see review of Paris, Adkins, Sessions, Webb, & Fischer, 2014); on both atmospheric chemistry and biogeochemical sulphur cycling in moreover, recent photochemical models demonstrate that the sign in Archaean palaeoenvironments. Δ S of experimentally produced sulphur will be highly dependent on the The biogeochemical cycling of sulphur in the Archaean was funda- wavelength of UV light used to simulate photolysis (Claire et al., 2014). mentally different to the present- day cycle. The largest flux of sulphur Further poorly constrained components of the Archaean S- cycle into the modern oceans is riverine sulphate, derived from the oxidative are the concentration and δ S of seawater sulphate. Archaean seawa- weathering of pyrite. However, this flux was less significant before the ter sulphate concentrations were expected to be lower than modern GOE due to low atmospheric pO ; therefore, the two most significant values because of the low atmospheric pO before the GOE. Habicht, 2 2 S fluxes into the Archaean oceans were likely hydrothermally sourced Gade, Thamdrup, Berg, and Canfield (2002) used culturing experi- and atmospherically derived sulphur species (e.g., Fike, Bradley, & ments of modern sulphate reducers to observe decreasing fraction- Rose, 2015). The Archaean S- MIF signal is thought to be dominated ation factors with decreasing sulphate concentrations. Assuming this by photochemical reactions involving atmospheric SO (e.g., Farquhar relationship can be extrapolated into the Archaean, they estimated et al., 2000; Pavlov & Kasting, 2002). In contrast, the shielding eect ff seawater sulphate concentrations at <200 μM. However, MSR frac- of atmospheric oxygen and ozone prevents the creation of an S- MIF tionation factors are dependent on microbial species (Bradley et al., signal in the contemporary S- cycle, with the exception of photochem- 2016) and other factors including reduction rate, type of substrate ical reactions generating stratospheric sulphate aerosols from volcanic and temperature (e.g., Fike et al., 2009; and references therein). More eruptions (Baroni, Thiemens, Delmas, & Savarino, 2007). Furthermore, recent results have suggested Archaean seawater sulphate concen- with low atmospheric and marine pO , homogenisation of discrete trations could have been even lower, <10 μM (Crowe et al., 2014; 2− atmospheric sulphur species (e.g., SO , S , SO ) was limited, further Zhelezinskaia, Kaufman, Farquhar, & Cliff, 2014). 4 n 2 favouring preservation of the Archaean S- MIF signal. The δ S ratio of seawater sulphate depends on the isotope compo- Aerft formation, the photochemically derived sulphur species, sition of the sulphate input into the oceans, the steady state burial flux 2− generally assumed to be sulphate (SO ) and elemental sulphur (S ), of sulphides and sulphate and the fractionation factor between coeval 4 n were deposited in the oceans. Sulphate is soluble in water and would sulphate- and sulphide- bearing species (Fike et al., 2015). Particularly, homogenise rapidly. It could form sulphate- bearing phases, such as if the sulphate concentrations in the Archaean oceans were low and carbonate- associated sulphate (CAS), gypsum (CaSO .2H O) or barite basins were restricted, MSR could have changed the residual sulphate 4 2 (BaSO ); or it could be reduced to hydrogen sulphide (H S) via microbi- composition of seawater through Rayleigh distillative fractionation 4 2 ally mediated MDF redox reactions or thermochemical sulphate reduc- (Roerdink et al., 2012). Therefore, seawater sulphate S- isotope compo- tion (TSR) before capture in the rock record as pyrite. These S- cycling sitions may have been heterogeneous, both spatially and temporally, 34 34 processes can significantly change the isotope composition of δ S with as proposed from the scaertt of δ S values preserved in Neoarchaean small (<0.2‰) differential eects ff on Δ S (Johnston et al., 2005). The carbonates (e.g., Paris et al., 2014; Zhelezinskaia et al., 2014). fate of S is less certain, but once deposited on the seafloor, it could In addition, the distribution and occurrence of microbial sulphate react with H S produced by microbial sulphate reduction (MSR) to form reduction in the Archaean is debated. Bulk δ S data from Australian 2− a reactive, mobile, soluble form of polysulphide S . Reactive polysul- Palaeoarchaean barites (Shen & Buick, 2004; Shen, Buick, & Canfield, phide and hydrogen sulphide can then react to form sulphide- bearing 2001) along with phylogenetic studies (Wagner, Roger, Flax, Brusseau, minerals, such as pyrite (Farquhar et al., 2013). Therefore, the S- isotope & Stahl, 1998) suggest that microbial sulphate reduction is likely an composition of pyrite generally records the biogeochemical MDF pro- ancient metabolism. However, the link between MSR S- isotope finger - 2− cesses that generated S and H S from atmospheric precursors. prints and pyrite textures remains enigmatic. Ono, Beukes, and Rumble n 2 Despite the narrative outlined above, there remain several poorly (2009) sampled multiple pyrite phases in ~2.5- Ga upper Prieska facies constrained aspects of the Archaean sulphur cycle. Firstly, the Δ S from the GKP01 Agouron drill core. They correlated different py- 2− signatures of the atmospheric products (e.g., S and SO ) produced rite phases with the Archaean reference array (Ono et al., 2003) and n 4 through photochemical reactions remain a subject of debate. Archaean suggested nodular and layered pyrite had S- isotope signatures most rocks from the Hamersley Basin, Western Australia, contain pyrite with consistent with MSR. Similarly matching pyrite crystal morphology to MEYER Et al .       355 the array, Kamber and Whitehouse (2007) proposed spheroidal py- rite concretions capture an MSR S- isotope fingerprint in the 2.52- Ga Upper Campbellrand Subgroup, Transvaal, South Africa. Moreover, Fischer et al. (2014) showed that δ S in nodular pyrite were charac- terised by systematic enrichment towards the rims (with the lowest δ S signature present in the centre of the nodules). Additional studies in the Griqualand West Basin are required to test whether these tex- tural and geochemical interpretations can be applied to pyrite from other palaeoenvironments and depositional ages. To address these gaps in our current understanding of Archaean sulphur cycling, we measured in situ, μm- scale, multiple sulphur iso- 34 33 tope ratios (δ S and Δ S) in exceptionally well- preserved pyritised microbialites in shales (sample- 3184) from the ~2.65- Ga Lokammona Formation, Ghaap Group, South Africa. By combining petrography with S- isotope data from secondary ion mass spectrometry (SIMS), we aim to increase our understanding of sulphur cycling in a Neoarchaean micro- bial mat, and the secondary (late- diagenetic/post- lithification) processes that may impact the primary (depositional/early- diagenetic) S- isotope signal. By petrographically correlating pyrite phases with sulphur isotope fingerprints, we can determine the MDF and MIF processes that contrib- uted to the formation of multiple pyrite generations in these sediments. 2  |  MATERIALS AND METHODS 2.1 | Geological setting and core material Our sample (3184) was collected from the BH1- SACHA core, drilled through the Neoarchaean Transvaal Supergroup in the Griqualand West Basin (Figure 1). The core material was obtained from the National Core Library at Donkerhoek (Pretoria, South Africa). The core contains carbon- FIGURE 1 Simplified geological map of Transvaal and Griqualand ates, siliciclastics and iron formations as well as several igneous intrusive West sediments, the approximate location of BH1- SACHA (star) bodies (Figure 2). Sample- 3184 was obtained from a depth of 3184 m and the inferred fault trace of the Kheis sole thrust fault. Insert: the from the Lokammona Formation, Schmidtsdrif Subgroup, Ghaap group, red rectangle shows the location of the main map relative to other Transvaal sediments and the Kaapvaal craton (beige). Adapted from Transvaal Supergroup. The Lokammona (Clearwater) Formation is domi- Altermann and Wotherspoon (1995) [Colour figure can be viewed at nated by shale, tuff layers intercalated with black shale, and minor do- wileyonlinelibrary.com] lomites (Figure 2). The high proportion of fine- grained sediments is consistent with deposition in a low- energy environment; Altermann and been identified (Altermann & Siegfried, 1997). We will consider these Siegfried (1997) suggested sedimentation occurred in a deep, shelf en- igneous and metasomatic processes when interpreting geochemical vironment with shale- carbonate/shale cycles representing shallowing- data from BH1- SACHA. upward cycles. They also noted the presence of microbial laminites, which suggested deposition in the photic zone. The interpretation is 2.2 | Imaging and EPMA supported by studies suggesting that modern microbial mats that pro- duce microbially induced sedimentary structures (MISS) are dominated We used the VHX- 2000 super- resolution digital microscope housed by benthic photoautotrophs (Noffke, 2009). The Lokammona Formation in the School of Earth and Environmental Sciences at the University has a model age of 2.650 ± 0.008 Ga from SHRIMP U- Pb analyses of of St Andrews (Scotland, UK) to image thick sections of sample- zircons from tuff laminae (Knoll & Beukes, 2009). 3184. This facilitated the detailed mapping of structures and fabrics The area sampled by the BH1- SACHA core is characterised by little within the sample and selection of target areas for SIMS. Backscatter subsequent tectonic deformation (Beukes, 1987), and it has been sub- electron (BSE) imaging was carried out using the School of Earth jected to sub- greenschist facies metamorphism (Button, 1973; Miyano and Environmental Sciences’ Jeol JCXA- 733 Superprobe electron & Beukes, 1984). The core has been penetrated by dykes and sills; microprobe analyser (EPMA). Prior to analysis, sample- 3184 thick the largest continuous igneous intrusion occurs at a depth of 922.3– sections were coated with a ~40- nm- thick graphite layer. The analy- 1201.27 m. In the Lokammona Formation, there is some evidence of sis was performed using a primary ion current of ~17–21 nA and an secondary mineralisation, particularly at ~3,165 m where galena has acceleration voltage of 15 kV. MEYER Et al . 356       FIGURE 2 BH1- SACHA bulk sulphur isotope data plotted against core depth (m) with the corresponding sedimentary log. Brecciated zones largely reflect post- depositional tectonic deformation. The pyritised microbialites analysed in this study were sampled at a depth of 3,184 m (grey band). Lokam., represents Lokammona. Data from Izon et al. (2015) and stratigraphic log adapted from Altermann & Siegfried, 1997 [Colour figure can be viewed at wileyonlinelibrary.com] a high spatial resolution, precision and mass resolution while minimis- 2.3 | Multiple sulphur isotope analyses via SIMS ing destruction to the sample. Epoxy thick sections of sample- 3184 The sulphur isotope data are reported relative to the Vienna Canyon with a 2.54 cm diameter were made, polished to <1 μm surface rough- Diablo Troilite (V- CDT) international reference standard. The delta (δ) ness and covered with a ~50- nm gold coating. notation denotes the deviation of the sample from V- CDT in permil −8 Samples were outgassed in an ancillary chamber at <1 × 10 (‰; Equations 1–2). The capital- delta (Δ) reflects mass- independent mB prior to analysis, and the pressure in the sample chamber was fraction, quantifying the deviation of a sample from the expected ter- allowed to equilibrate for >1 hr aerft sample introduction, prior to restrial mass fractionation line (Equation 3). analysis. Thick sections were analysed in a vacuum with a pressure of −9 + 3 × 10 mB. A focused ~2 μm Cs primary ion beam of 0.9–1.8 nA was 33 32 ( S∕ S) sample δ S= −1 ×1000 (1) rastered over a 10 μm by 10 μm area of interest. Each measurement 33 32 ( S∕ S) V−CDT was divided into 15 cycles and required 7–8 min to complete. The 34 32 secondary ions were collected using faraday cups (FC). The 7f- GEO is ( S∕ S) sample δ S= −1 ×1000 (2) 34 32 specifically designed for precise isotope ratio measurement using two ( S∕ S) V−CDT FCs optimised for major (FC1) and minor (FC2) isotope data acquisition � � 0.515 ⎛ ⎞ in “charge mode” (Peres, de Chambost, & Schuhmacher, 2008). Each δ S V−CDT 33 33 ⎜ ⎟ Δ S=δ S −1000× 1− V−CDT ⎜ ⎟ (3) isotope is selected by magnet switching, and the corresponding sec- ⎝ ⎠ ondary ion counts were collected in sequence of ascending mass. The 34 33 Multiple sulphur isotope analyses (δ S and δ S) were conducted desired FC is selected using a deflector situated in the detector assem- using the CAMECA IMS 7f- GEO secondary ion mass spectrome- bly. This configuration allows for the initiation of data acquisition from ter housed in the Department of Earth and Planetary Sciences at FC2 prior to complete dissipation of the signal from FC1. This alternat- Washington University (St Louis, MO, USA). Secondary ion mass spec- ing collection arrangement results in improved signal- to- noise ratios trometry (SIMS) enables the in situ measurement of S- isotope ratios to and thereby greater precision per unit time. In this study, two minor MEYER Et al .       357 isotopes were to be analysed. As back- to- back data collection using R, respectively. Although IMF can vary according to instrumental FC2 would defeat the signal- to- noise advantage described above; thus, conditions, the S- isotope offset between bracketing standards and un- a pseudomeasurement, at mass 33.5, was made on FC1 so that each knowns should not vary. Furthermore, the measured S- isotope ratios 32 33 34 cycle became S (FC1), S(FC2), 33.5(FC1), S(FC2). The background of a session’s bracketing standards were observed to determine drift. 4 3 level for FC1 was ~3 × 10 counts/s and for FC2 ~3 × 10 counts/s. The drift value was usually <1‰ over 24 hr. To correct for this, a linear 32 8 33 6 Typically, secondary ion yields were: S~2 × 10 counts/s, S~10 and constant drift was assumed within a session. The magnitude of 34 6 counts/s and S~10 counts/s. this drift was much smaller than the signals observed and was there- Background and ion yields were recalibrated on a daily basis. Mass fore unlikely to have impacted the results. calibration, as well as automatic peak centring of the field and contrast apertures, was performed at the start of each measurement; this corrects 3  |  RESULT S for any drift in the secondary ion beam or the magnetic field (Fike et al., 2009). To remove the 50- nm gold coating, each measurement was pre- 3.1 | Textural analysis sputtered for 2 min prior to the start of the 15 cycles. A minimum mass 33 32 resolving power (MRP) of 3900 is required to separate the S and SH Sample- 3184 is a black shale composed of ~1- mm- thick laminae with peaks, and therefore, a MRP of ~4300 was used. Analysis was preferen- varying proportions of pyrite and matrix (Figure 3). The matrix is com- tially carried out in regular spaced intervals in a grid- like paern tt (Fike, posed of silt- sized particles (~40 μm) of aluminosilicates (clay min- Gammon, Ziebis, & Orphan, 2008). Generally, the analysis spots were erals; 80% modal abundance) and quartz (20%). Uncommon <5- μm placed at ~50- μm intervals. However, deviations from the ideal grid were detrital grains of rutile (TiO ) and apatite (Ca (PO ) (F,Cl,OH) ) make 2 10 4 6 2 necessary to consistently sample pyrite instead of siliciclastic matrix. up <1% of the matrix volume. Laminae contain variable proportions The internal error (standard error of n = 15 cycles) varied inversely of pyrite- to- matrix ratios, ranging from 10%–70% pyrite to 30%–90% with the S count rate (Figure S1), suggesting that precision was lim- matrix. There are three distinguishable pyrite phases: A) disseminated ited by counting statistics (Fike et al., 2009). Anomalous data were pyrite crystals, <5 μm in size; B) type 1 pyrite: irregular aggregates of omitted from further analysis if any of the following applied to the coalesced pyrite crystals. Clusters can range from 10 μm to 500 μm measurement: across, and are composed of anhedral to subhedral, 1- to 20- μm pyrite crystals. (Figure 4); C) type 2 pyrite: euhedral cubic pyrite with a crystal 1. The sputter area clearly targeted only matrix as subsequently size ranging from 10 to 100 μm (Figure 4). Some euhedral pyrite crys- determined from BSE and super-resolution digital microscope tals are isolated within the matrix, while other crystals overgrow and images. encrust type 1 aggregates. Overgrowths occur on the edges of type 1 2. The uncorrected relative standard error for S counts was >1%. pyrite aggregates and within clusters where matrix lenses occur. Type 34 32 3. The uncorrected relative standard error for S/ S was >0.1%. 1 and type 2 pyrites are chemically distinct phases, as determined from BSE images (Figures 4 and S6). Aerft excluding the anomalous measurements, the internal error Laminae containing type 1 pyrite show sedimentary structures 34 33 was typically <1‰ for δ S and <0.5‰ for δ S (1SE). An in- house that suggest it precipitated during early diagenesis, prior to soft sed- Washington University pyrite standard was mounted in a polished and iment deformation. These include wavy crinkled laminae with typical gold-c oated thin section and placed in a separate holder. Its composition wavelengths of ~500 μm and wave heights of ~200 μm, isoclinal folds, was determined by measuring the S- isotope composition of the internal 1- mm rip- up structures and over- folded laminae. Particularly, <2- mm standard 12 times and the Balmat standard 20 times. The Balmat stan- irregular, ellipsoidal concretions are composed of polycrystalline type 34 33 dard (δ S = 15.1‰ and δ S = 7.7‰) analyses bracketed the in- house 1 pyrite. They cross- cut lamination but also cause draping of laminae standard. The in-house standard has a known isotope composition of on either side of the concretion (Figures 4 and S4); this suggests type 34 33 δ S = 0.13 ± 0.30‰ and δ S = 0.13 ± 0.20‰ (1SE). The external errors 1 pyrite precipitated aerft deposition but before lithification, during (1 standard deviation of multiple adjacent points on the in- house stan- early diagenesis. 34 33 dard) were typically 0.38‰ and 0.32‰ for δ S and δ S, respectively A normal microfault cross- cuts the entire core section of sam- (n = 35). Within sample- 3184, there were 26‰ variability in measured ple- 3184 (Figure 3). The ~50- μm- wide microfault has been infilled by 34 33 δ S values and 14‰ variability in δ S. This variation far exceeds the aluminosilicates and silica. The microfault cross- cuts type 1 pyrite; type calculated internal and external errors. 2 pyrite crystals cross- cut the microfault itself (Figure 4). Furthermore, On average, a session consisted of ~50 measurements of un- type 2 pyrite overgrows the sedimentary structures composed of type knowns and was bracketed by ≥4 measurements of the in- house stan- 1 pyrite. According to these cross- cutting relationships, type 2 pyrite dard on both sides of the unknowns (Figure S2). Instrumental mass precipitated post- lithification. fractionation (IMF) was corrected for by standard-sample bracketing. The mean, uncorrected isotope composition of repeated analyses of 3.2 | Multiple sulphur isotope data the in- house pyrite standard on the SIMS was δ S = −0.05 ± 0.38‰ and δ S = 1.00 ± 0.32‰ (1σ; n = 35; relative to V- CDT). Therefore, All the SIMS data measured in sample- 3184 are presented in the IMF (IMF = R /R ) was typically 1.000 and 1.001 for R and Figure 5. There is no obvious stratigraphic trend; however, there is raw known MEYER Et al . 358       (a) (b) FIGURE 3 (a) Reflected light scanner image of sample- 3184 from the Lokammona Formation showing the areas of SIMS analyses (white dots). Blue dot = Figure 6 analysis area; green dot = Figure 7 analysis area. Divisions on the scale bar represent 1 mm. (b) A trace of the image in (a) showing laminae compositions, type examples of sedimentary structures discussed in the text and the location of the normal microfault. Modal percentages of laminae compositions are shown in brackets (pyrite/matrix) [Colour figure can be viewed at wileyonlinelibrary.com] a relationship between pyrite crystal shape and S- isotope composi- and type 2 pyrite end members. When these measurements were sub- tions. The first data set shows a mean of δ S = 12.54 ± 4.98‰ and sequently examined in their petrographic contexts, photomicrographs Δ S = −0.21 ± 0.65‰ (±1σ, n = 177) and corresponds to type 1 py- suggest the analysis spots targeted a mixture of both pyrite phases. rite (Figures 6 and 7). In sample- 3184, type 1 pyrite δ S values vary by 26‰; within a ~500 × 500 μm area, the range of δ S is 15‰ 4  |  DISCUSSION 34 33 (Figure 6). Wavy crinkled laminae show δ S and Δ S values consist- ent with typical type 1 pyrite (Figure S3). The second data set has a 4.1 | Early- diagenetic (type 1) pyrite 34 33 mean of δ S = 8.36 ± 1.16‰ and Δ S = 5.54 ± 1.53‰ (±1σ, n = 18) and was measured in type 2 pyrite (Figures 7 and S5). To a 95% confi- Laminae composed of type 1 pyrite are wavy and crinkled on a dence level, the S- isotope values for type 1 and type 2 pyrite are signif- sub- mm scale (Figures 3 and 4) and are consistent with sedimen- icantly different in their medians and variances (p = .000 for δ S and tary structures that have been microbially induced (Noffke, 2009). Δ S in two- sample Wilcoxon and Levene’s tests). Some data points Prokaryotes and eukaryotes in microbial mats produce a matrix correspond to an intermediary S- isotope signature between the type 1 of extracellular polymeric substances (EPS), which are composed MEYER Et al .       359 FIGURE 4 BSE images of different sedimentary structures in sample- 3184. (a) Wavy crinkled internal lamination composed of type 1 pyrite and disseminated pyrite. (b) Wavy crinkled internal lamination composed of type 1 pyrite with type 2 pyrite overgrowths. (c) Pyrite concretion. Note that the concretion both cross- cuts and causes deformation of the lamination. (d) The normal microfault that is infilled by quartz and clay minerals. Note the microfault and vein cross- cut type 1 pyrite; type 2 pyrite cross- cuts the microfault and vein. Therefore, the relative order of formation is as follows: pyrite 1 precipitated first, brittle deformation caused microfault formation, the microfault was infilled, and finally type 2 pyrite precipitated. py, py 1, py 2 and qtz represent pyrite, type 1 pyrite, type 2 pyrite and quartz, respectively [Colour figure can be viewed at wileyonlinelibrary. com] FIGURE 5 Plot of multiple S- isotope 34 33 data (δ S and Δ S) measured via SIMS. SIMS error bars are 1SE for each measurement (n = 15 cycles). Green triangles represent a mixed signal, where the area of analysis sampled both type 1 and type 2 pyrite. The orange line is the Archaean reference array as described by Ono et al. (2003) [Colour figure can be viewed at wileyonlinelibrary.com] of polysaccharides, proteins, humic substances and nucleic acids is eroded and turned over at its edges (Schieber, 1999). However, (Nielsen, Jahn, & Palmgren, 1997). The matrix has several physi- wavy crinkled laminae can also form through the differential com- cal functions, including adhesion to surfaces, aggregation of cells, paction of phyllosilicates around, for example, microconcretions, silt stabilisation of microbial mat structure, sorption of exogenous or- lenses or silica spherules (Schieber, 2007). Examples of compaction ganic molecules and retention of water (Laspidou & Rittmann, 2002). around such structures were absent in BSE images of sample- 3184 Importantly, EPS is responsible for the formation of wavy crinkled (Figure 4). Thus, the sedimentary structures indicate high cohesion lamination due to its ability to trap detrital grains and its high cohe- of the depositional layers and are consistent with typical microbialite sive properties. This cohesion is also responsible for the formation textures, supporting a biogenic control on the formation of type 1 of over- folded layers (Figure 3), which occur when the mat surface pyrite. MEYER Et al . 360       FIGURE 6 Combined petrography with SIMS S- isotope data shows the typical 34 33 textural, δ S and Δ S characteristics of type 1 pyrite. (a) Reflected light photomicrograph of the analytical grid location. (b) BSE image of the analysis area overlain by a SIMS δ S image constructed by spline interpolation of the analytical grid (n = 31). Note that the analysis area is composed of a ~500-μ m aggregate of anhedral to subhedral, 1- to 20- μm pyrite crystals. (c) Plot to show SIMS δ S against 33 34 Δ S data (‰). Note the variation in δ S data is significant; the variation in Δ S data is not significant. Error bars represent 1SE for each measurement (n = 15 cycles) (d) BSE image of the analysis area overlain by a SIMS Δ S image constructed by spline interpolation of the analytical grid (n = 31) [Colour figure can be viewed at wileyonlinelibrary.com] 34 34 32 34 The S- isotope composition of type 1 pyrite is distinguishable by R = S/ S) and variable seawater δ S values. Figure 5 shows sulphate 33 34 a Δ S value of −0.21 ± 0.65‰ (±1σ) and a 26‰ range in δ S within that few data occur above 18‰, and thus, the curve showing the 33 34 sample- 3184. The Δ S error bars for type 1 pyrite within an area of δ S of the instantaneous product must be in the distillative tail sulphide 33 34 analysis overlap (Figure 6); therefore, the Δ S variation within a lamina when δ S > 18‰. Values of 6‰–12‰ for α , 4‰–16‰ for sulphide MSR 34 34 is statistically insignificant. The considerable variation in δ S values is δ S and f values from ~0.5 to 0.9 are all consistent with the sulphate consistent with MSR—modern microbial mats with sulphate reducers data (Figure 8). show ~15–53‰ variations in δ S on a 1- mm scale (Fike et al., 2009; This 6‰–12‰ estimate of α is at the lower end of the MSR Wilbanks et al., 2014), similar to what we observe here. Thus, both spectrum for the modern MSR cell- specific fractionation factors of textural evidence and the S- isotope composition are consistent with α = 2‰–66‰ (Fike et al., 2015). Small MSR isotope fractionation MSR the expected biosignatures of an ancient microbial mat containing can occur when sulphate concentrations are low, when sulphate re- sulphate reducers. The 26‰ range in δ S within sample- 3184 could duction rates or sulphide oxidation rates are high, when H is used be explained by non- uniformity in MSR fractionation factors, which as an electron donor instead of organic carbon, and/or when sulphur depend on the microbe strain, type of electron donors, reduction rate, disproportionation levels are low (e.g., Fike et al., 2009 and references temperature and the presence of additional S metabolisms such as ox- therein). Although the relationship between sulphate concentration idation and disproportionation (e.g., Fike et al., 2009). and MSR fractionation factor is complex (see Bradley et al., 2016), Alternatively, the scaertt and enrichment in δ S sulphide relative our estimates of α are consistent with other studies that suggest MSR to coeval sulphate could be produced by fractionation during MSR in low seawater sulphate concentrations and high rates of MSR in the a (partially) closed system, analogous to a mat environment periodi- Archaean (e.g., Crowe et al., 2014; Habicht et al., 2002; Zhelezinskaia cally flushed with seawater sulphate. We used a Rayleigh distillation et al., 2014). Furthermore, the seawater δ S compositions we esti- model to explore this scenario, using a suite of realistic fractionation mate from closed- system modelling of the SIMS data are consistent 34 34 34 factors for MSR (α = 1,000 × [ R / R − 1], where with estimates of seawater sulphate δ S from Neoarchaean CAS MSR sulphate sulphide MEYER Et al .       361 FIGURE 7 The contrasting textural and S- isotope signatures of type 1 and type 2 pyrite. (a) Reflected light photomicrograph of the area of analysis. (b) BSE image of the analysis area overlain by a SIMS δ S image constructed by spline interpolation of the analytical grid (n = 15). Note the ~300- μm cubic type 2 pyrite grain, overgrowing type 1 pyrite aggregates. (c) Plot to show 34 33 SIMS δ S against Δ S data (‰). Note the significant difference in Δ S for type 1 and 2 pyrite. Error bars represent 1SE for each measurement (n = 15 cycles). Blue circles = type 1 pyrite; green circles = type 2 pyrite. (d) BSE image of the analysis area overlain by a SIMS Δ S image constructed by spline interpolation of the analytical grid (n = 15) [Colour figure can be viewed at wileyonlinelibrary.com] (e.g., Domagal- Goldman, Kasting, Johnston, & Farquhar, 2008; Guo crystals have all been formed from the same, uniform S- isotope 34 33 et al., 2009; Paris et al., 2014). However, the small MSR fractionation source. The δ S and Δ S signature of type 2 pyrite falls within Ono factors and variation in Δ S values in the literature imply Archaean et al.’s (2003) estimates of the composition of atmospheric elemen- seawater sulphate had a short residence time due to a small reser- tal sulphur as inferred from Archaean pyrite data and photochemical voir size relative to the fluxes into and out of the reservoir; thus, the experiments (Figure 5), and Paris et al.’s (2014) Neoarchaean CAS val- 34 33 S- isotope composition of seawater sulphate was likely spatially and ues. Similar δ S and Δ S ratios in euhedral grains sampled from GKF- temporally heterogeneous (Fischer et al., 2014; Zhelezinskaia et al., 01 (a core sampling deeper water equivalents of the BH1- SACHA; 2014). Schröder, Lacassie, & Beukes, 2006) were interpreted as having an We therefore conclude that type 1 pyrite reflects the morphologi- atmospheric elemental sulphur origin (Farquhar et al., 2013). Farquhar cal and geochemical signatures of sulphate reducers in a Neoarchaean et al. (2013) hypothesised that solid atmospherically derived S parti- microbial mat, as inferred from (i) the biogenicity of sedimentary cles could remain in an unreactive form as they fell through the water structures like wavy crinkled lamination, (ii) textural evidence of early- column. After deposition, they could then react with H S in the sedi- diagenetic precipitation of type 1 pyrite, and (iii) models of type 1 ment, producing reactive polysulphide. Finally, the polysulphide could pyrite δ S values supporting plausible MSR fractionation factors ex- react with FeS to form pyrite. Farquhar et al. (2013) predicted that the pressed within a restricted seawater sulphate pool. pyrite would closely reflect the S- isotope signature of atmospheric elemental sulphur if the mass contribution of H S was small relative to polysulphide in the pyrite product. 4.2 | Secondary (type 2) pyrite A similar interpretation of type 2 pyrite formation in sample- 3184 As discussed above, simple cross- cutting relations suggest type 2 py- is less likely, due to its inferred later timing of formation based on rite precipitated after type 1 pyrite formation and post- lithification, the petrographic relationships we describe above. As type 1 pyrite because it overgrows deformed aggregates of type 1 pyrite, and it likely formed from sulphide generated via microbial sulphate reduc- cross- cuts a normal microfault with vein infill (Figure 4). The low tion, H S was abundant during early diagenesis. Therefore, any atmo- standard deviation of type 2 pyrite S- isotope ratios suggests these spheric, unreactive elemental sulphur particles that deposited into the MEYER Et al . 362       FIGURE 8 Graphs showing the calculated δ S values of the hydrogen sulphide instantaneous product relative to the proportion of sulphate consumed (f), the initial δ S composition and the MSR fractionation factor (α ). The grey rectangles correspond to the range of sulphate source-product δ S compositions measured in type 1 pyrite in sample- 3184. (a) α = 6‰, (b) α = 12‰ [Colour figure can be viewed at wileyonlinelibrary.com] sulphide Neoarchaean microbial mat would have simultaneously reacted with noted evidence of thrust faults in BH1- SACHA at depths >2,800 m in H S to form a soluble, reactive form of polysulphide. During early dia- the Monteville Formation. They suggested this is the intersection of genesis, the mobile polysulphide, a precursor to pyrite synthesis, would BH1- SACHA with the Kheis sole thrust fault (Figure 1). Allochthonous have reacted with an iron monosulphide to form FeS . However, petro- sections of Archaean Transvaal Supergroup rocks (including rocks graphic relationships suggest type 2 pyrite formed post- lithification, that were sampled by the BH1- SACHA core) as well as Proterozoic not coevally with type 1 pyrite during diagenesis (Figure 4). Therefore, Waterberg and Olifantshoek Groups were horizontally displaced and type 2 pyrite in sample- 3184 unlikely formed from an elemental sul- now lie unconformably on the autochthonous Archaean package of phur source by the mechanism proposed by Farquhar et al. (2013). rocks (Altermann & Wotherspoon, 1995; Martini, Eriksson, & Snyman, Hence, we suggest the Δ S = 5.5‰ signature of type 2 pyrite was not 1995). The age assigned to thrusting ranges from 2.20 to 1.75 Ga produced from contemporaneous atmospheric S products at the time (Grobbelaar, Burger, Pretorius, Marais, & Van Niekerk, 1995). If the sample- 3184 was formed. Instead, we propose later- stage alteration thrust is older than ~2.0 Ga, it suggests that the hanging wall rocks processes could explain the Δ S values measured in type 2 pyrite. were exhumed before the key igneous and metasomatic events asso- The BH1-S ACHA core has been altered by several post- ciated with the Bushveld Complex. Hence, the metamorphic grade of depositional hydrothermal events that could have caused multi - BH1- SACHA in the hanging wall may be lower than the footwall MVT generational pyrite genesis. Stratigraphic logs indicate evidence of deposits examined by Huizenga et al. (2006). In addition, De Kock secondary mineralisation of galena, ~20 m stratigraphically higher et al. (2009) showed that similar strata preserved within the GKP- 01 than sample- 3184. Furthermore, intrusion of the igneous body at drill core through the Ghaap Group show pervasive remagnetisation 922.3–1201.27 m formed a contact metamorphic aureole (Altermann by 2.5- to 1.8- Ga nanoscale pyrrhotite. Furthermore, high- resolution & Siegfried, 1997). To the best of our knowledge, this dyke has not palaeomagnetic and geochemical evidence from GKF- 01 Neoarchaean been dated; however, it may be co- genetic to other mafic intrusions nodular pyrite suggests that some of the Transvaal Supergroup pyrites related to the emplacement of the Bushveld complex, 2.06–2.05 Ga have been post- depositionally altered, ~0.5 Ga aerft their deposition (Hartzer, 1995). Associated igneous and metamorphic fluids can ca- (Fischer et al., 2014). Thus, even undeformed Ghaap Group strata can talyse metasomatic reactions, causing the loss and gain of elements show a complex history of iron sulphide mineral precipitation. in circulating fluids. The occurrence of metasomatism in BH1- SACHA The geological evidence therefore suggests sulphur could have is supported by regional geological evidence; a plaorm- tf wide fluid circulated in fluids as soluble sulphur species and participated in hy- flow event at ~2.0 Ga could have caused the Pb- Zn- Cu- mineralisation drothermal reactions in the Griqualand West Basin. There are multiple in the Campbellrand Subgroup, Ghaap Group. Huizenga, Gutzmer, sources of subsurface S- rich fluids (Figure 9): In modern settings, sul- Greyling, and Schaefer (2006) sampled Mississippi Valley- type (MVT) phate can be acquired from (buried) seawater (Ohmoto, 1972). If the hy- deposits in the Griqualand West area, ~100–120 km straight- line drothermal alterations occurred aerft the GOE, seawater was a potential distance from Kathu. They estimated a regional fluid temperature of source (A, Figure 9). Furthermore, pore waters expelled from sedimen- 200–240°C and pressure of 0.8–1.5 kbar during mineralisation of the tary rocks during compaction, fluids formed by magmatic/metamor - MVT deposits. It is possible that BH1- SACHA experienced similar phic processes or meteoric- or seawater- derived fluids can cause the metamorphic conditions. However, Altermann and Siegfried (1997) dissolution of S- bearing minerals. Dissolution of sulphate, such as CAS MEYER Et al .       363 FIGURE 9 Hypothesised formation of post- lithification pyrite with an anomalous Δ S signal. Sulphur can be sourced via three pathways: (A) seawater, (B) dissolution of sulphate- or sulphide- bearing minerals or (C) the disproportionation of SO in magmatic- hydrothermal or magmatic steam environments [Colour figure can be viewed at wileyonlinelibrary.com] 33 34 or sulphate- bearing minerals (e.g., gypsum, anhydrite or barite), occurs of S relative to S in the polysulphide product, and therefore, the readily, while pyrite dissolution can also occur in the presence of an oxi- TSR product carries a positive Δ S signal (Oduro et al., 2011). 3+ dant, for example, O or Fe (Descostes, Vitorge, & Beaucaire, 2004; B, Thermochemical sulphate reduction occurs at 100–300°C Figure 9). Moreover, both sulphate and sulphide can be produced from (Johnston, 2011); these temperatures are equivalent to sub- the disproportionation of magmatic SO . For example, during the con- greenschist to greenschist facies, which correspond to the maximum densation of a magmatic plume, H SO and H S form from the dispro- metamorphic grade that the lower Transvaal Supergroup, including 2 4 2 portionation of magmatic SO at temperatures 200–400°C and pH < 3. BH1- SACHA, experienced during intrusion of the Bushveld Complex Sulphate can be leached by fluids, causing the production of an acidic, in the Palaeoproterozoic (Sumner & Beukes, 2006). Therefore, the sulphate- bearing hydrothermal fluid that can infiltrate surrounding wall temperature regime during contact metamorphism of BH1- SACHA rock (C, Figure 9; Rye, 2005). The intrusion of the Bushveld Complex and the surrounding rocks was sufficiently high to support TSR. The at ~2.06 Ga could have produced magmatic sulphur species (Hartzer, petrographic evidence and isotope fingerprint of type 2 pyrite, cou- 1995). Finally, these subsurface S- rich fluids could have infiltrated sur- pled with the history of hydrothermal alteration described above, rounding rocks via fracture flow or porous flow, and precipitated pyrite. could be consistent with MIE during TSR causing an anomalous en- 33 34 We suggest two scenarios whereby the circulation of S- rich fluids richment in S relative to S. However, there is currently no evidence from the above sources could have contributed to the formation of of anomalous Δ S signals associated with TSR in bulk rock analyses type 2 pyrite and its associated S- MIF signal: one via thermochemi- from the geological record. Furthermore, TSR causes a MIF eect ff in 33 36 cal sulphate reduction of S- bearing fluids from a more recent sulphur S, but not S (Oduro et al., 2011); therefore, this hypothesis cannot source, and one via migration of S- rich fluids from stratigraphically be conclusively demonstrated without the inclusion of Δ S data. older sediments carrying an inherited Archaean S- MIF signal. More likely, the MIF signal measured in type 2 pyrite could have If the subsurface sulphur- rich fluid were sourced from modern or formed from the dissolution and reprecipitation of Archaean S- bearing post- GOE seawater, the dissolution of post- GOE sulphur minerals or minerals originally deposited with a positive Δ S signature (Figure 9). magmatic sulphate, it would not carry a primary MIF signal. Therefore, Notably, bulk pyrite shows a Δ S composition within error of the type to precipitate type 2 pyrite with a Δ S fingerprint of +5.5‰ from one 2 pyrite S- isotope fingerprint, 20–30 m stratigraphically lower than of these sources, a post- depositional process would be necessary to sample- 3184 (Figure 2). Therefore, dissolution of these pyrite phases cause S- MIF. Experimental investigations have shown that thermo- and migration of Δ S = +5.5‰ sulphur- rich fluids upwards are a likely chemical sulphate reduction (TSR) can produce a < 13‰ enrichment in source of the type 2 pyrite MIF signal. Therefore, grain- scale Archaean Δ S relative to the sulphate source (Oduro et al., 2011). An anomalous MIF signatures may reflect secondary pyrite formation mechanisms as 33 34 enrichment of S relative to S during TSR occurs due to magnetic well as atmospheric processes at the time of deposition. These mul- isotope eects ff (MIE). In the sulphur isotope system, the S- isotope tiple pyrite phases can reflect a protracted history of pyrite precipita- contains magnetic nuclei; during TSR, S undergoes the spin- selective tion, which can be recognised using high- resolution, in situ S- isotope 34 32 reaction faster than S and S. This results in an anomalous enrichment geochemistry. MEYER Et al . 364       NE/H016805/1 and Natural Environment Research Council Standard 5  |  C ONCL USIONS Grant NE/J023485/2 (to A.Z.) and an NSF/EAR Grant (#1229370) Sulphate reduction in Neoarchaean microbial mats: Our textural and sul- and a Packard Fellowship (to D.F.) phur isotope data echo other studies that suggest MSR influenced S- cycling in the Neoarchaean (e.g., Habicht et al., 2002; Ueno et al., 2008; REFERENCES Zerkle, Claire, Domagal- Goldman, Farquhar, & Poulton, 2012). As typi- cal microbialite textures are generally associated with photoautotrophs Altermann, W., & Siegfried, H. P. (1997). Sedimentology and facies devel- opment of an Archaean shelf: Carbonate plaorm tf transition in the (Noffke, 2009), this study proposes that microbial sulphate reducers may Kaapvaal Craton, as deduced from a deep borehole at Kathu, South have commonly occurred in the same community as photosynthesisers, Africa. Journal of African Earth Sciences, 24, IN1–IN4. forming a microbial mat ecosystem similar to those of the modern. Altermann, W., & Wotherspoon, J. M. (1995). The carbonates of the Neoarchaean seawater sulphate concentrations and S- isotope compo- Transvaal and Griqualand West Sequences of the Kaapvaal craton, with special reference to the Lime Acres limestone deposit. Mineralium sition: SIMS S- isotope data and closed- system modelling suggest the Deposita, 30, 124–134. local sulphur reservoir had a Δ S = −0.21 ± 0.65‰ (±1σ, n = 177) and Baroni, M., Thiemens, M. H., Delmas, R. J., & Savarino, J. (2007). Mass- δ S = 4‰–12‰ during formation of type 1 pyrites in sample- 3184 independent sulfur isotopic compositions in stratospheric volcanic (Figure 8). These values are consistent with previous estimates of eruptions. Science, 315, 84–87. Beukes, N. J. (1987). Facies relations, depositional environments and dia- the δ S- isotope composition of Archaean sulphate (e.g., Domagal- genesis in a major early Proterozoic stromatolitic carbonate plaorm tf Goldman et al., 2008; Guo et al., 2009; Paris et al., 2014). However, to basinal sequence, Campbellrand Subgroup, Transvaal Supergroup, the S- isotope composition of seawater sulphate may have been spa- Southern Africa. Sedimentary Geology, 54, 1–46. tially and temporally variable, particularly at the low sulphate levels Bradley, A. S., Leavitt, W. D., Schmidt, M., Knoll, A. H., Girguis, P. R., & Johnston, D. T. (2016). Paerns tt of sulfur isotope fractionation during estimated for the Neoarchaean (Zhelezinskaia et al., 2014). Closed- Microbial Sulfate Reduction. Geobiology, 14, 91–101. system modelling of sample- 3184 indicates small MSR fractionation Button, A. (1973). Algal stromatolites of the early Proterozoic Wolkberg factors (α = 6‰–12‰), consistent with high reduction rates and MSR Group, Transvaal sequence. Journal of Sedimentary Research, 43, 160–167. low sulphate concentrations. The small MSR fractionation factors and Claire, M. W., Kasting, J. F., Domagal-Goldman, S. D., Stüeken, E. E., Buick, scatter in Δ S values in the literature imply sulphate had a short resi- R., & Meadows, V. S. (2014). Modeling the signature of sulfur mass- independent fractionation produced in the Archean atmosphere. dence time in Archaean oceans due to a small reservoir size and com- Geochimica et Cosmochimica Acta, 141, 365–380. paratively large fluxes into and out of the reservoir. Additional studies Crowe, S. A., Paris, G., Katsev, S., Jones, C., Kim, S. T., Zerkle, A. L., … assessing the S- isotope composition of CAS will help constrain the Farquhar, J. (2014). Sulfate was a trace constituent of Archean seawa- 34 33 evolution of seawater sulphate δ S and Δ S during the Archaean. ter. Science, 346, 735–739. De Kock, M. O., Evans, D. A. D., Kirschvink, J. L., Beukes, N. J., Rose, E., & The utility of combining petrography and SIMS: This study demon- Hilburn, I. (2009). Paleomagnetism of a Neoarchean- Paleoproterozoic strates that grain- scale MIF signals could record secondary processes carbonate ramp and carbonate plaorm tf succession (Transvaal in addition to contemporaneous atmospheric processes at the time of Supergroup) from surface outcrop and drill core, Griqualand West re- deposition. MIF signals can be generated by MIE during later TSR or, gion, South Africa. Precambrian Research, 169, 80–99. Descostes, M., Vitorge, P., & Beaucaire, C. (2004). Pyrite dissolution in more probably, leached from surrounding Archaean sedimentary rocks acidic media. Geochimica et Cosmochimica Acta, 68, 4559–4569. and reprecipitated following lithification. Furthermore, it demonstrates Domagal-Goldman, S. D., Kasting, J. F., Johnston, D. T., & Farquhar, J. that Archaean rocks can preserve a complex, protracted history of pyrite (2008). Organic haze, glaciations and multiple sulfur isotopes in the formation. Pyrite genesis can occur over several generations; precipita- Mid- Archean Era. Earth and Planetary Science Leer tt s, 269, 29–40. Farquhar, J., Bao, H., & Thiemens, M. (2000). Atmospheric influence of tion can range from pre- so ft sediment deformation to post- lithification. Earth’s earliest sulfur cycle. Science, 289, 756–758. This study highlights the eectiv ff eness of pairing petrography with high- Farquhar, J., Cliff, J., Zerkle, A. L., Kamyshny, A., Poulton, S. W., Claire, resolution SIMS sulphur isotope studies, to unravel the diagenetic pro- M., … Harms, B. (2013). Pathways for Neoarchean pyrite formation cesses that have contributed to multigenerational pyrite genesis. constrained by mass- independent sulfur isotopes. Proceedings of the National Academy of Sciences, 110, 17638–17643. Fike, D. A., Bradley, A. S., & Rose, C. V. (2015). Rethinking the ancient sulfur A CKNO WLEDGMENT S cycle. Annual Review of Earth and Planetary Sciences, 43, 593–622. Fike, D. A., Finke, N., Zha, J., Blake, G., Hoehler, T. M., & Orphan, V. J. (2009). We thank Dr. Clive Jones for helping with SIMS analysis, Mr. Donald High resolution SIMS- based sulfide δ S: A new tool for characterizing Herd for assistance with imaging and the staff at the South African microbial activity in a variety of depositional environments. Geochimica et Cosmochimica Acta, 73, 6187–6204. National Core Library in Donkerhoek for facilitating access to the Fike, D. A., Gammon, C. L., Ziebis, W., & Orphan, V. J. (2008). Micron- scale core materials. The authors thank Tony Prave, Tim Raub, Gareth Izon, mapping of sulfur cycling across the oxycline of a cyanobacterial mat: A Mark Claire and James Farquhar for their input and discussion. This paired nanoSIMS and CARD- FISH approach. ISME Journal, 2, 749–759. study was financially supported by the Geological Society of London’s Fischer, W. W., Fike, D. A., Johnson, J. E., Raub, T. D., Guan, Y., Kirschvink, J. L., & Eiler, J. M. (2014). SQUID–SIMS is a useful approach to uncover Alan & Charlotte Welch Fund, A Palaeontological Association primary signals in the Archean sulfur cycle. Proceedings of the National Undergraduate Research Grant (PA- UB201504) and the University Academy of Sciences, 111, 5468–5473. of St Andrew’s Zannah Stephen Memorial Travel Scholarship Grobbelaar, W. S., Burger, M. A., Pretorius, A. I., Marais, W., & Van Niekerk, (all to N.M.), Natural Environment Research Council Fellowship I. J. M. (1995). Stratigraphic and structural setting of the Griqualand MEYER Et al .       365 West and the Olifantshoek Sequences at Black Rock, Beeshoek and Paytan, A., Mearon, S., Cobb, K., & Kastner, M. (2002). Origin of marine bar- Rooinekke Mines, Griqualand West, South Africa. Mineralium Deposita, ite deposits: Sr and S isotope characterization. Geology, 30, 747–750. 30, 152–161. Peres, P., de Chambost, E., & Schuhmacher, M. (2008). CAMECA IMS 7f- Guo, Q., Strauss, H., Kaufman, A. J., Schröder, S., Gutzmer, J., Wing, B., … Geo: Specialized SIMS tool for geosciences. Applied Surface Science, Farquhar, J. (2009). Reconstructing Earth’s surface oxidation across the 255, 1472–1475. Archean- Proterozoic transition. Geology, 37, 399–402. Roerdink, D. L., Mason, P. R., Farquhar, J., & Reimer, T. (2012). Multiple Habicht, K. S., Gade, M., Thamdrup, B., Berg, P., & Canfield, D. E. (2002). sulfur isotopes in Paleoarchean barites identify an important role for Calibration of sulfate levels in the Archean ocean. Science, 298, 2372–2374. microbial sulfate reduction in the early marine environment. Earth and Hartzer, F. J. (1995). Transvaal Supergroup inliers: Geology, tectonic de- Planetary Science Leer tt s, 331, 177–186. velopment and relationship with the Bushveld Complex, South Africa. Rye, R. O. (2005). A review of the stable- isotope geochemistry of sulfate Journal of African Earth Sciences, 21, 521–547. minerals in selected igneous environments and related hydrothermal Huizenga, J. M., Gutzmer, J., Greyling, L. N., & Schaefer, M. (2006). Carbonic systems. Chemical Geology, 215, 5–36. fluid inclusions in Paleoproterozoic carbonate- hosted Zn- Pb deposits Schieber, J. (1999). Microbial mats in terrigenous clastics; the challenge of in Griqualand West, South Africa. South African Journal of Geology, 109, identification in the rock record. Palaios, 14, 3–12. 55–62. Schieber, J. (2007). Microbial mats on muddy substrates—examples of pos- Izon, G., Zerkle, A. L., Zhelezinskaia, I., Farquhar, J., Newton, R. J., Poulton, S. sible sedimentary features and underlying processes. Atlas of Microbial W., … Claire, M. W. (2015). Multiple oscillations in Neoarchaean atmo- Mat Features Preserved within the Siliciclastic Rock Record Atlases in spheric chemistry. Earth and Planetary Science Leer tt s, 431, 264–273. Geoscience, 2, 117–133. Johnston, D. T. (2011). Multiple sulfur isotopes and the evolution of Earth’s Schröder, S., Lacassie, J. P., & Beukes, N. J. (2006). Stratigraphic and geo- surface sulfur cycle. Earth- Science Reviews, 106, 161–183. chemical framework of the Agouron drill cores, Transvaal Supergroup Johnston, D. T., Farquhar, J., Wing, B. A., Kaufman, A. J., Canfield, D. E., & (Neoarchean–Paleoproterozoic, South Africa). South African Journal of Habicht, K. S. (2005). Multiple sulfur isotope fractionations in biological Geology, 109, 23–54. systems: A case study with sulfate reducers and sulfur disproportion- Shen, Y., & Buick, R. (2004). The antiquity of microbial sulfate reduction. ators. American Journal of Science, 305, 645–660. Earth- Science Reviews, 64, 243–272. Kamber, B. S., & Whitehouse, M. J. (2007). Micro- scale sulphur isotope evi- Shen, Y., Buick, R., & Canfield, D. E. (2001). Isotopic evidence for microbial dence for sulphur cycling in the late Archean shallow ocean. Geobiology, sulphate reduction in the early Archaean era. Nature, 410, 77–81. 5, 5–17. Sumner, D. Y., & Beukes, N. J. (2006). Sequence stratigraphic development Knoll, A. H., & Beukes, N. J. (2009). Introduction: Initial investigations of a of the Neoarchean Transvaal carbonate plaorm, tf Kaapvaal Craton, Neoarchean shelf margin- basin transition (Transvaal Supergroup, South South Africa. South African Journal of Geology, 109, 11–22. Africa). Precambrian Research, 169, 1–14. Ueno, Y., Ono, S., Rumble, D., & Maruyama, S. (2008). Quadruple sulfur iso- Laspidou, C. S., & Rittmann, B. E. (2002). A unified theory for extracellular tope analysis of ca. 3.5 Ga Dresser Formation: New evidence for micro- polymeric substances, soluble microbial products, and active and inert bial sulfate reduction in the early Archean. Geochimica et Cosmochimica biomass. Water Research, 36, 2711–2720. Acta, 72, 5675–5691. Martini, J. E. J., Eriksson, P. G., & Snyman, C. P. (1995). The early protero- Van Kranendonk, M. J. (2006). Volcanic degassing, hydrothermal circula- zoic Mississippi Valley- type Pb- Zn- F deposits of the Campbellrand and tion and the flourishing of early life on Earth: A review of the evidence Malmani subgroups, South Africa. Mineralium Deposita, 30, 135–145. from c. 3490–3240 Ma rocks of the Pilbara Supergroup, Pilbara Craton, Miyano, T., & Beukes, N. J. (1984). Phase relations of stilpnomelane, Western Australia. Earth- Science Reviews, 74, 197–240. ferri- annite, and riebeckite in very low- grade metamorphosed iron- Wagner, M., Roger, A. J., Flax, J. L., Brusseau, G. A., & Stahl, D. A. (1998). formations. South African Journal of Geology, 87, 111–124. Phylogeny of dissimilatory sulfite reductases supports an early origin of Nielsen, P. H., Jahn, A., & Palmgren, R. (1997). Conceptual model for pro- sulfate respiration. Journal of Bacteriology, 180, 2975–2982. duction and composition of exopolymers in biofilms. Water Science and Wilbanks, E. G., Jaekel, U., Salman, V., Humphrey, P. T., Eisen, J. A., Facciotti, Technology, 36, 11–19. M. T., … Orphan, V. J. (2014). Microscale sulfur cycling in the phototro- Noe,ffk N. (2009). The criteria for the biogeneicity of microbially induced phic pink berry consortia of the Sippewissett Salt Marsh. Environmental sedimentary structures (MISS) in Archean and younger, sandy deposits. Microbiology, 16, 3398–3415. Earth- Science Reviews, 96, 173–180. Zerkle, A. L., Claire, M. W., Domagal-Goldman, S. D., Farquhar, J., & Poulton, Oduro, H., Harms, B., Sintim, H. O., Kaufman, A. J., Cody, G., & Farquhar, S. W. (2012). A bistable organic- rich atmosphere on the Neoarchaean J. (2011). Evidence of magnetic isotope eects ff during thermochemical Earth. Nature Geoscience, 5, 359–363. sulfate reduction. Proceedings of the National Academy of Sciences, 108, Zhelezinskaia, I., Kaufman, A. J., Farquhar, J., & Cliff, J. (2014). Large sulfur 17635–17638. isotope fractionations associated with Neoarchean microbial sulfate Ohmoto, H. (1972). Systematics of sulfur and carbon isotopes in hydrother- reduction. Science, 346, 742–744. mal ore deposits. Economic Geology, 67, 551–578. Ono, S., Beukes, N. J., & Rumble, D. (2009). Origin of two distinct multiple- sulfur isotope compositions of pyrite in the 2.5 Ga Klein Naute Formation, SUPPOR TING INF ORMATION Griqualand West Basin, South Africa. Precambrian Research, 169, 48–57. Additional Supporting Information may be found online in the support- Ono, S., Eigenbrode, J. L., Pavlov, A. A., Kharecha, P., Rumble, D., Kasting, J. F., & Freeman, K. H. (2003). New insights into Archean sulfur cycle from ing information tab for this article. mass- independent sulfur isotope records from the Hamersley Basin, Australia. Earth and Planetary Science Leer tt s, 213, 15–30. How to cite this article: Meyer NR, Zerkle AL, Fike DA. Paris, G., Adkins, J. F., Sessions, A. L., Webb, S. M., & Fischer, W. W. (2014). Sulphur cycling in a Neoarchaean microbial mat. Geobiology. Neoarchean carbonate–associated sulfate records positive Δ33S anomalies. Science, 346, 739–741. 2017;15:353–365. https://doi.org/10.1111/gbi.12227 Pavlov, A. A., & Kasting, J. F. (2002). Mass- independent fractionation of sulfur isotopes in Archean sediments: Strong evidence for an anoxic Archean atmosphere. Astrobiology, 2, 27–41.

Journal

Geobiology – Wiley

Published: May 1, 2017

There are no references for this article.