Earth and Planetary Science Letters 303 (2011) 121–132

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

journal homepage: www.elsevier.com/locate/epsl

Iron and carbon isotope evidence for microbial iron respiration throughout the Archean

Paul R. Craddock ⁎, Nicolas Dauphas

Origins Laboratory, Department of the Geophysical Sciences and Enrico Fermi Institute, The University of Chicago, 5734 South Ellis Avenue, Chicago, IL 60637, United States article info abstract

Article history: Banded Iron-Formations (BIFs) are voluminous chemical sediments that are rich in iron-oxide, carbonate and Received 17 August 2010 silica and whose occurrence is unique to the Precambrian. Their preservation in the geological record offers Received in revised form 20 December 2010 insights to the surface chemical and biological cycling of iron and carbon on early Earth. However, many details Accepted 22 December 2010 regarding the role of microbial activity in BIF deposition and diagenesis are unresolved. Laboratory studies have Available online 22 January 2011 + shown that reaction between carbon and iron through microbial iron respiration [2Fe2O3∙nH2O+CH2O+7H → 2+ − Editor: R.W. Carlson 4Fe +HCO3 +(2n+4)H2O+chemical energy] can impart fractionation to the isotopic compositions of these elements. Here, we report iron (δ56Fe, vs. IRMM-014) and carbon isotopic (δ13C, vs. V-PDB) compositions of Keywords: magnetite and of iron-rich and iron-poor carbonates in BIFs from the late Archean (~2.5 Ga) Hamersley Basin, iron-formation Australia and the early Archean (~3.8 Ga) Isua Supracrustal Belt (ISB), Greenland. The range of δ56Fe values Hamersley measured in the Hamersley Basin, including light values in magnetite and heavy values in iron-rich carbonates Isua (up to +1.2‰), are incompatible with their precipitation in equilibrium with seawater. Rather, the data together iron carbonates with previously reported light δ13C values in iron-rich carbonates record evidence for diagenetic reduction of iron respiration ferric oxide precursors to magnetite and carbonate through microbial iron respiration (i.e., dissimilatory iron reduction, DIR). Iron and carbon isotope data of iron-rich metacarbonates from the ISB are similar to those of late Archean BIFs. The isotopic signatures of these metacarbonates are supportive of an early diagenetic origin despite metasomatic overprint, and preserve evidence of microbial iron respiration within the oldest recognized sedimentary rocks on Earth. © 2010 Elsevier B.V. All rights reserved.

1. Introduction et al., 2000; Farquhar et al., 2000; Kasting, 1987; Ono et al., 2003; Pavlov and Kasting, 2002), is uncertain. Photochemical oxidation of Fe(II) Banded Iron-Formations (BIFs) are conspicuously laminated marine in surface ocean waters owing to interaction with incident UV radiation chemical sediments that are characterized by high concentrations of iron- has been proposed as an entirely abiological means of accounting for bearing minerals (20–40 wt.% bulk Fe) commonly interbedded with ferric iron in BIFs (Braterman et al., 1983; Cairns–Smith, 1978). Oxidation layers of silica, and whose occurrence is unique to the Precambrian of Fe(II) by O2 produced via photosynthesis has also been suggested (James, 1954, 1983). The mineralogy of the best-preserved BIFs consists of (Cloud, 1965, 1973), implying an indirect biological influence on BIF combinations of four dominant facies: oxide (magnetite, hematite), formation and hinting at the presence of free O2 oases in the Archean carbonate (siderite, ankerite, Fe–dolomite and, less commonly, calcite), surface ocean. Alternatively, direct biological activity has been implicated, chert and silicate (stilpnomelane, riebeckite, greenalite, minnesotaite), via anoxygenic photosynthesis that coupled oxidation of Fe(II) to and locally sulfide (pyrite) and phosphate (apatite). Most known BIFs reduction of inorganic carbon to yield organic compounds (Garrels have ages in the range ~3.8 to 1.8 Ga, but these formations also occur to a et al., 1973; Kappler et al., 2005; Konhauser et al., 2007; Widdel et al., lesser extent in the Neoproterozoic at ~700 Ma (Klein, 2005). The study of 1993). It is also uncertain the extent to which the mineral assemblages these formations preserved in the rock record offers critical insights to preserved in BIFs reflect either primary precipitates from seawater, surface geochemical cycles and chemical evolution of the ocean and possibly in near-chemical equilibrium with the ocean and atmosphere, or atmosphere in the Precambrian, and in particular the Archean (≥2.5 Ga). are authigenic minerals formed during early sedimentary diagenesis and Despite significant scientific interest and research, however, there is burial metamorphism. For example, the mineralogical, chemical (e.g., rare no consensus on the origin of BIFs. The primary mechanism for oxidation earth element) and isotopic (e.g., δ13C, δ18O) characteristics of iron-rich of Fe(II)aq in an Archean ocean that was purportedly anoxic (Canfield carbonates such as siderite [FeCO3]andankerite[Ca0.5(Fe,Mg)0.5CO3]in BIFs have been used to argue either for primary precipitation from an anoxic and stratified water column (Beukes et al., 1990; Kaufman et al., ⁎ Corresponding author. Tel.: +1 773 834 3997. 1990; Klein and Beukes, 1989) or for an authigenic origin (Becker and E-mail address: [email protected] (P.R. Craddock). Clayton, 1972; Heimann et al., 2010; Walker, 1984).

0012-821X/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2010.12.045 122 ..Cadc,N apa at n lntr cec etr 0 21)121 (2011) 303 Letters Science Planetary and Earth / Dauphas N. Craddock, P.R.

Table 1 Iron and carbon isotopic compositions of oxides and carbonate from drill core samples of the Brockman Iron Formation, Hamersley Basin.

Samplea Hole a Macroband Mesoband Type Depthb δ13 C δ56 Fe δ57 Fe (m) (‰)c (‰) d (‰)d

Ankerite Siderite Ankerite/Siderite Magnetite Hematite Ankerite/Siderite Magnetite Hematite

Dales Gorge Member, Brockman Iron Formation Drill core from Wittenoom Gorge Area (118° 28′ E; 22° 25′ S) 13324 M1 27 BIF 12 Magnetite 85.2 0.424±0.031 0.630±0.043 13327 WC2 27 BIF 11 Chert 98.4 −11.41 −11.08 −0.289±0.031 0.205±0.031 −0.442±0.049 0.316±0.049 13327 WC1 27 BIF 11 Chert 98.4 −15.05 0.013±0.031 0.465±0.031 0.008±0.046 0.699±0.043 13321 HC 27 BIF 10 Chert+hematite 99.5 −6.98 −7.45 −0.705±0.031 0.908±0.038 −1.023±0.083 1.369±0.044 13318 M3 27 BIF 7 Magnetite 122.2 −9.24±0.03 0.006±0.036 0.005±0.057 13318 Pl 27 BIF 7 Chert 122.2 −9.83±0.04 −0.584±0.029 −0.865±0.051 13318 FM1 27 BIF 7 Chert 122.2 −9.72±0.06 0.118±0.031 −0.210±0.034 0.373 ±0.029 0.198±0.046 −0.298±0.054 0.561±0.051 13318 M1 27 BIF 7 Magnetite 122.3 0.159±0.029 0.230±0.051 13318 M2 27 BIF 7 Magnetite 122.3 0.112±0.031 0.179±0.046 13316 M1 27 BIF 6 Magnetite 128.1 0.650±0.038 0.949±0.058 13316 CA1 27 BIF 6 Fine−band combination 128.2 −9.31 0.641±0.031 0.636±0.033 0.933±0.083 0.914±0.060 13316 CB1 27 BIF 6 Coarse-band combination 128.2 −9.01 −9.41 0.178±0.033 0.196±0.031 0.283±0.060 0.244±0.049 13309 FM1 27 BIF 1 Chert 162.2 −9.87 0.057±0.031 0.588±0.031 0.070±0.046 0.852±0.043 13309 PC1 27 BIF 1 Chert 162.3 −10.06 −0.161±0.031 0.490±0.030 −0.236±0.043 0.710±0.046 13309 QIO1 27 BIF 1 Chert-matrix 162.4 0.675±0.031 0.790 ±0.034 0.990±0.046 1.144±0.054 13309 M2 27 BIF 1 Magnetite 162.4 0.488±0.030 0.694±0.044

13326 M1 28 BIF 12 Magnetite 110.9 0.418±0.031 0.610±0.046 – 13328 WC2 28 BIF 11 Chert 114.9 −11.07 −12.48 −0.555±0.031 −0.757±0.049 132 13328 WC1 28 BIF 11 Chert 115.0 −12.86 0.426±0.038 0.630±0.044 13322 WC1 28 BIF 10 Chert 125.2 −9.73 0.453±0.035 0.662±0.047 13322 WCIA 28 BIF 10 Chert 125.2 −9.61 0.294±0.027 0.419±0.040 13322 M1 28 BIF 10 Magnetite 125.2 0.730±0.033 1.107±0.060 13322 HC 28 BIF 10 Chert+hematite 125.2 −6.76 −8.26 −0.792±0.038 0.398±0.030 −1.142±0.044 0.584±0.057 13322 WHC 28 BIF 10 Chert+hematite 125.2 −6.50±0.03 −7.80 −1.081±0.033 −1.601±0.060 13322 FM1 28 BIF 10 Chert 125.3 −8.96 −9.79 0.787±0.033 1.194±0.033 1.158±0.044 1.778±0.044 13322 QIO1 28 BIF 10 Chert-matrix 125.3 0.566±0.038 0.998±0.038 0.847±0.058 1.477±0.058 13319 FM3 28 BIF 7 Chert 146.5 −9.74±0.03 −0.318±0.029 −0.287±0.030 −0.470±0.043 −0.422±0.044 13319 FM2 28 BIF 7 Chert 146.6 −10.18±0.01 0.086±0.030 0.064±0.030 0.156±0.044 0.077±0.044 13319 M1 28 BIF 7 Magnetite 146.6 −0.022±0.030 0.014±0.044 13319 FM1 28 BIF 7 Chert 146.6 −9.90 0.075±0.030 −0.149±0.029 0.093±0.044 −0.212±0.043 13313 M1 28 BIF 2 Magnetite 174.5 1.085±0.029 1.582±0.043 13310 PC1 28 BIF 1 Chert 186.7 −0.181±0.028 0.159±0.035 −0.303±0.047 0.241±0.047 M 2 40 BIF 2 Magnetite 196.8 0.551±0.035 0.810±0.047 P 2 40 BIF 2 Chert 196.8 −9.70±0.02 P 2A 40 BIF 2 Chert 196.8 −9.80±0.06 Table 1 (continued) Samplea Hole a Macroband Mesoband Type Depthb δ13 C δ56 Fe δ57 Fe (m) (‰)c (‰) d (‰)d

Ankerite Siderite Ankerite/Siderite Magnetite Hematite Ankerite/Siderite Magnetite Hematite

Dales Gorge Member, Brockman Iron Formation Drill core from Wittenoom Gorge Area (118° 28′ E; 22° 25′ S) FM 2 40 BIF 2 Chert 196.8 −10.28±0.02 M 3 51 BIF 12 Magnetite 104.5 0.314±0.028 0.478±0.047 QIO 51 51 BIF 5 Chert-matrix 155.4 −0.189±0.028 −0.282±0.047 M 1 51 BIF 5 Magnetite 155.6 −0.092±0.035 −0.111±0.047 P l 51 BIF 5 Chert 155.6 −8.65 Drill core from Yampire Gorge Area (118° 37′ E; 22° 27′ S) 13325 WC2 Y3 BIF 12 Chert 66.2 −10.95±0.02 −0.305±0.041 0.202±0.029 −0.448±0.072 0.298±0.041 13325 M2 Y3 BIF 12 Magnetite 66.2 0.351±0.029 0.515±0.041 13325 M1 Y3 BIF 12 Magnetite 66.3 0.344±0.051 0.519±0.133 13325 FM1 Y3 BIF 12 Chert 66.3 −10.96 0.301±0.033 0.297±0.029 0.410±0.083 0.433±0.041 13325 WC1 Y3 BIF 12 Chert 66.4 −10.84±0.01 −0.404±0.029 0.194±0.029 −0.588±0.043 0.282±0.133 13329 M1 Y3 BIF 11 Magnetite 70.7 −12.59±0.50 −0.005±0.031 −0.017±0.083 13329 FM1 Y3 BIF 11 Chert 70.8 −11.11 0.244±0.033 0.408±0.033 0.674 ±0.051 0.395±0.083 0.632±0.083 1.057 ±0.133 ..Cadc,N apa at n lntr cec etr 0 21)121 (2011) 303 Letters Science Planetary and Earth / Dauphas N. Craddock, P.R. 13323 M1 Y3 BIF 10 Magnetite 78.2 0.727±0.034 1.029±0.054 13320 FM1 Y3 BIF 7 Chert 98.4 −9.53 0.013±0.033 −0.156±0.029 −0.029±0.083 −0.253±0.041 13317 CA1 Y3 BIF 6 Fine-band combination 104.4 −9.36 −9.34 0.605±0.038 0.870±0.058 13317 CBI Y3 BIF 6 Coarse-band combination 104.4 −9.06 −9.31 0.456±0.038 0.710±0.044 13317 M1 Y3 BIF 6 Magnetite 104.5 0.665±0.085 0.887±0.058 13314 M3 Y3 BIF 3 Magnetite 125.3 0.735±0.029 1.073±0.041 13314 M2 Y3 BIF 3 Magnetite 125.3 0.877±0.028 1.283±0.047 13314 H1 Y3 BIF 3 Hematite 125.3 1.055 ±0.033 1.568 ±0.083 13314 M1 Y3 BIF 3 Magnetite 125.3 0.886±0.041 1.301±0.072 13314 FM1 Y3 BIF 3 Chert 125.3 0.963±0.028 1.419±0.047 13314 Carb 1 Y3 BIF 3 Chert+carbonate 125.5 −9.48 1.212±0.041 1.041±0.041 1.850±0.072 1.549±0.072 13311 PC1 Y3 BIF 1 Chert 137.6 −9.75 −0.033±0.029 0.600±0.027 −0.053±0.041 0.868±0.040 13311 PC1 Y3 BIF 1 Chert 137.6 0.096±0.035 0.135±0.047

a Sample numbers follow the nomenclature adopted by Becker (1971). Core samples were recovered by drilling operations in two areas: Wittenoom Gorge and Yampire Gorge. Hole designates core number. b Depth in meters below modern surface. c Carbon isotopic (δ13C) ratios are per mil difference relative to PDB standard and are those published by Becker and Clayton (1972). Quoted uncertainties are the standard deviation of replicate analyses as reported by Becker and Clayton (1972). d Iron isotopic (δ56 Fe, δ57 Fe) ratios are reported in per mil difference relative to iron metal standard IRMM-014 (δ57 Fe~1.5xδ56 Fe). Uncertainties are 95% confidence intervals. – 132 123 124 P.R. Craddock, N. Dauphas / Earth and Planetary Science Letters 303 (2011) 121–132

In order for BIFs to offer a faithful and critical perspective of the which late Archean and Paleoproterozoic (~2.6 to 2.45 Ga) volcanics Earth's early ocean and atmosphere biogeochemical cycles, an under- and sediments of the Hamersley Group were deposited (Trendall and standing of processes leading to their formation [i.e., primary Fe(II) Blockley, 1970). The Hamersley Group is of particular significance oxidation and precipitation versus diagenetic transformations] is owing to the presence of four major and several minor conspicuously 56 56 54 essential. Here, we report coupled iron, δ Fe=[( Fe/ Fe)sample/ laminated, iron-rich stratigraphic units (Banded Iron Formation, BIF), 56 54 3 13 13 12 13 ( Fe/ Fe)IRMM-014 −1]×10 , and carbon, δ C=[( C/ C)sample/( C/ including the Brockman Iron Formation (MacLeod, 1966; Trendall, 12 3 C)V-PDB −1]×10 , isotope compositions of iron oxides and carbonates 2002; Trendall and Blockley, 1970). The Hamersley Group also from one of the most well-preserved and least metamorphosed BIFs, the contains the Wittenoom Dolomite, which comprises between 200 ~2.5 Ga Brockman Iron Formation of the Hamersley Basin, western and 300 m of interbedded carbonate, chert and shale with the Australia, and in one of the oldest known Fe-rich chemical sedimentary carbonate occurring typically as mostly massive calcite and dolomite. formations, that of the ~3.8 Ga Isua Supracrustal Belt (ISB), southern West Samples from the Hamersley Basin were selected both from an iron- Greenland. Our data are used to constrain the probable chemical and formation (Dales Gorge Member of the Brockman Iron Formation) and biological pathways for the precipitation of iron-bearing minerals in from a carbonate formation (Wittenoom Dolomite). Samples of iron- Archean BIFs. A primary motivation for this research is the study of Becker formation included iron oxide (magnetite) and iron-rich carbonates and Clayton (1972) that reported δ13C of iron-rich carbonates (siderite, (siderite, ankerite) and were from fresh drill core material that was ankerite) in the Brockman Iron Formation, Hamersley Basin, mostly recovered by the Australian Blue Asbestos Co. as part of a prospecting between −8and−11‰, significantly more depleted than that of typical program under subsidy from the Western Australian Government iron-poor marine carbonates (calcite, dolomite) with δ13C between −2 (Trendall and Blockley, 1970). Samples of carbonate formation were and +2‰. The source of light carbon in the iron-formation was suggested hand specimens of massive, iron-poor carbonate (calcite, dolomite) to derive from organic carbon, possibly resulting from oxidation– obtained from chipping out fresh surfaces of the Wittenoom Dolomite reduction reactions involving microbially metabolized Fe(III). We from weathered exposures. All samples were originally collected and examine specifically whether the range of iron and carbon isotope ratios prepared for carbon and oxygen isotope studies by Becker and Clayton recorded in iron-rich carbonates and magnetite from the Brockman Iron (1972,1976). Formationisconsistentwithsuchaprocessandthusrecordsevidencefor The Isua Supracrustal Belt (ISB) is part of the ~3.8 Ga high-grade microbial iron respiration (i.e., dissimilatory iron reduction; DIR) in the Itsaq Gneiss Complex and is one of the oldest metasedimentary- late Archean, a process that was originally suggested by Walker (1984) bearing formations on Earth. The ISB assemblage comprises a variety and Baur et al. (1985). Johnson et al. (2008) have recently reported the of metavolcanic, clastic and metasedimentary rock types, the iron isotope compositions of magnetite and siderite from the Brockman stratigraphy of which has been discussed in detail by previous Iron Formation in order to distinguish between primary precipitation in contributions (Boak and Dymek, 1982; Dymek and Klein, 1988; seawater and diagenetic iron transformations. These authors suggested Myers, 2001; Nutman and Friend, 2009; Nutman et al., 1984, 1996; that the range of measured iron isotopic compositions could be consistent Rosing et al., 1996; van Zuilen et al., 2003). Protolith identification of with an authigenic origin, but they did not have access to carbon isotopic rocks in the ISB is complicated owing to the diversity of lithological data for the same carbonates to validate a DIR pathway. Complementary units observed in the ISB and metamorphic and/or metasomatic iron and carbon isotopic studies of magnetite and carbonates from the 2.5 overprint. For this study, metacarbonate sequences were targeted in to 2.4 Ga Kuruman Iron Formation, Transvaal Craton, South Africa, have order to better identify their probable origin (chemical sedimentary also explored evidence for microbial iron respiration during this period versus metasomatic) and possible association with genuine iron- of Earth's history (Heimann et al., 2010; Johnson et al., 2003). Heimann formations. Iron-poor metacarbonates occur typically at the contacts et al. (2010) concluded that the range of δ13Candδ56Fe values with ultramafic intrusions in the southern ISB and have been documented in iron-rich carbonates was consistent with authigenic interpreted as originating by leaching of ultramafic protoliths by pathways for carbonate formation through microbial DIR. metasomatic fluids (Rose et al., 1996; Rosing et al., 1996). Iron-rich Further, we assess whether a record of microbial iron respiration metacarbonates occur within a range of host rocks in the northern ISB, extends to the early Archean. Dauphas et al. (2007) have reported iron including in association with banded magnetite–quartz rocks of a isotopic compositions of iron-rich and iron-poor metacarbonates from chemical sedimentary origin. We report and contrast here the carbon the ~3.8 Ga ISB. Metacarbonates from the ISB typically show a isotopic ratios of iron-rich and iron-poor metacarbonates and metasomatic relationship with their host rocks (Rose et al., 1996; Rosing combine these new data with existing iron isotopic data for the et al., 1996). Rose et al. (1996) and Rosing et al. (1996) have argued that same samples (Dauphas et al., 2007) to better constrain their primary metacarbonates were formed by metasomatic alteration of igneous formation, in particular identifying whether the isotopic signatures protoliths. However, the iron isotopic and trace element (e.g., REE, Fe/Ti) record a chemical sedimentary origin despite metasomatic overprint. characteristics of iron-rich metacarbonates are more similar to that of co- existing BIFs in the ISB and are more supportive of a chemical sedimentary 2.2. Sample preparation and analytical methods origin (Bolhar et al., 2004; Dauphas et al., 2007). Here, we integrate new carbon and existing iron isotopic data of iron-rich and iron-poor Samples of magnetite and iron-rich carbonate from the Brockman metacarbonates from the ISB to re-examine the origin of these rocks, in Iron Formation were originally taken from drill core material and particular evaluating whether these isotopic signatures are supportive of separated and crushed by Richard Becker to obtain milligram-sized a sedimentary origin prior to metasomatic overprint and record evidence monomineralic powders for the analysis of carbon and oxygen of microbial iron respiration within the oldest recognized sedimentary isotopes (Becker and Clayton, 1972, 1976). In addition, samples of rocks on Earth. iron-poor carbonate from the Wittenoom Dolomite were cut from unweathered faces of hand specimens and approximately one gram of 2. Materials and methods this material was crushed to a powder using an agate mortar and pestle. Magnetite and iron-rich and iron-poor carbonates from the ISB 2.1. Geology and sample descriptions were taken from powders of hand specimens originally prepared for iron isotopic analysis (see Dauphas et al., 2007). The geological setting, description and analytical procedures for Iron isotopic analyses of samples from the Brockman Iron Formation isotopic measurement of iron-formations in this study are described and Wittenoom Dolomite were carried out following the procedures in detail in the Supporting Online Material, and summarized below. developed in our laboratory (Craddock and Dauphas, 2010; Dauphas et The Hamersley Basin, western Australia is a roughly elliptical basin in al., 2004a, 2009b). Iron isotopic measurements were performed on a P.R. Craddock, N. Dauphas / Earth and Planetary Science Letters 303 (2011) 121–132 125

seawater seawater Carbon isotopic analyses were carried out on iron-rich and iron- 2+ - Fe aq HCO3 , aq 15 Hamersley poor metacarbonates from the ISB at RSMAS, University of Miami. The ~2.5 Ga procedures followed those developed by Swart et al. (1991). Carbon magnetite Fe-rich isotope compositions are reported relative to the V-PDB scale. carbonate 10 Fe-rich Fe-poor carbonate 3. Results carbonate frequency Fe-poor δ56 5 carbonate Iron isotopic ( Fe) compositions of magnetite samples from the Brockman Iron Formation, Hamersley Basin, analyzed in this study range from −0.3 to +1.2‰ (Table 1; Fig. 1). These data are similar to those 0 measured in magnetite samples from the same formation by Johnson et al. (2008). In the two studies combined, magnetite from the Brockman 15 Kuruman Iron Formation extends down to δ56Fe=−1.0‰ and has a mean iron ~2.5 Ga isotopic composition of +0.17‰ (n=94). Two types of carbonates from

Fe-poor the Hamersley Basin were studied: iron-rich carbonates (siderite, 10 Fe-rich carbonate ankerite) from the Brockman Iron Formation and iron-poor carbonates carbonate (calcite, dolomite) from the underlying Wittenoom Dolomite [using

frequency Fe-rich notations from van Zuilen et al. (2003) and Dauphas et al. (2007),iron- 5 magnetite carbonate rich carbonates have an Fe atomic ratio, Fe/(Fe+Mn+Ca+Mg), greater Fe-poor carbonate than 0.40, whereas iron-poor carbonates have an Fe atomic ratio less than 0.15]. The δ56Fe values of iron-rich carbonates analyzed in this study range 0 from −1.1 to +1.2‰ (Table 1; Fig. 1). Considering the iron isotopic data -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 -14 -12 -10 -8 -6 -4 -2 0 2 for siderite published by Johnson et al. (2008), the range extends down to δ56Fe, IRMM-014 (‰) δ13C, V-PDB (‰) −2.0‰. The iron isotopic compositions of magnetite and iron-rich carbonates sampled from adjacent bands in the Brockman Iron Formation Fig. 1. Histograms of iron (δ56Fe) and carbon (δ13C) isotope ratios in iron-rich carbonates δ56 (siderite, ankerite), iron-poor carbonates (calcite, dolomite) and magnetite from the ~2.5 Ga have a broadly positive correlation. The Fe values of iron-poor Brockman Iron Formation, Hamersley Basin (top panel) and Kuruman Iron Formation, carbonates from the Wittenoom Dolomite analyzed in this study are all Transvaal Craton (bottom panel). Iron isotope data for the Brockman Iron Formation are from very negative, between −0.5 and −1.0‰ (Table 2; Fig. 1). Carbon isotope Table 1. Carbon isotope data for the Brockman Iron Formation are from Becker and Clayton compositions (δ13C) of the carbonate samples analyzed in this study for (1972). Carbon and iron isotope data for the Kuruman Iron Formation are from Johnson et al. (2003) and Heimann et al. (2010). Bin widths for iron and carbon isotope values are 0.025‰ their iron isotopic compositions were previously reported by Becker and 13 and 1.0‰, respectively. The vertical arrows at δ56Fe=−0.2‰ and δ13C=0‰ are the Clayton (1972). All iron-rich carbonates have light δ C compositions 13 estimated compositions of Fe(II)aq and dissolved inorganic carbon (DIC) in Archean seawater, (Fig. 1). Siderite δ Crangefrom−12.6 to −7.5‰ and cluster around the respectively. The range of iron and carbon isotopic ratios in these samples are incompatible average of −9.6‰ (n=11). Ankerite δ13C are slightly more variable, with formation of magnetite and iron-rich carbonates as direct precipitates from seawater, ranging from −15.0 to −6.5‰, but cluster around a similar average of but can be reconciled by a model invoking diagenetic mobilization of primary ferric oxides − ‰ and organic carbon through dissimilatory iron reduction (DIR). 9.9 (n=32). Iron-poor carbonates from the Wittenoom Dolomite have a limited range of heavier δ13Cwithanaverage−2‰ (n=15; Fig. 1). For comparison, the iron and carbon isotope compositions of iron- Thermo Scientific Neptune MC-ICPMS at the University of Chicago. All rich and iron-poor carbonates and of magnetite from the penecontem- iron isotope data are calibrated relative to the IRMM scale (Craddock poraneous Kuruman Iron Formation, Transvaal Supergroup, South Africa and Dauphas, 2010; Taylor et al., 1992). (Heimann et al., 2010; Johnson et al., 2003) are illustrated (Fig. 1). The

Table 2 Iron and carbon isotopic compositions of iron-poor carbonate hand specimens from the Wittenoom Dolomite, Hamersley Basin.

Samplea Lithology δ13C δ56Fe δ57Fe (‰)b (‰)c (‰)c

Calcite Dolomite Calcite Dolomite Calcite Dolomite

Wittenoom Dolomite Formation 12523 Dolomite 0.05±0.01 −0.995±0.068 −1.427 ±0.092 12523 vein Dolomite vein 0.38±0.02 −0.774±0.068 −1.135 ±0.092 13354 Calcite+quartz −1.09±0.02 −0.728±0.055 −1.059±0.038 13355 top Calcite±dolomite −1.80±0.02 −1.56 −0.472±0.055 −0.674±0.038 13355 bottom −1.83 −1.84 13356 Calcite −0.13±0.05 −0.921±0.055 −1.313±0.038 13357 chert Calcite+quartz −1.22 −0.860±0.055 −1.344±0.038 13358 #1 Calcite −0.30±0.05 −0.882±0.055 −1.338±0.038 13358 #2 −0.450±0.055 −0.692±0.038 13359 chert Black chert+calcite −1.06±0.05 −0.887±0.057 −1.366±0.085 13360 Calcite −0.82±0.06 −0.685±0.057 −1.024±0.085 RB-22a* Calcite+silicate −4.91±0.10 RB-22b* −4.68±0.02 RB-23a* Calcite+dolomite+ −6.36±0.04 RB-23b* silicate −4.75±0.03 −5.67±0.03

a Sample numbers follow the nomenclature adopted by Becker (1971). Detailed descriptions given by Becker (1971). b Carbon isotopic (δ13C) ratios are per mil difference relative to PDB standard and are those published by Becker and Clayton (1972). Quoted uncertainties are the standard deviation of replicate analyses as reported by Becker and Clayton (1972). c Iron isotopic (δ56Fe, δ57Fe) ratios are reported in per mil difference relative to iron metal standard IRMM-014 (δ57Fe~1.5 xδ56Fe). Uncertainties are 95% confidence intervals. * Indicates iron-poor carbonates sampled from the base of the Wittenoom Dolomite that occur as fine-layers in apparently microbanded rocks containing silica and silicate and that are distinct in appearance to more massive calcite and dolomite sampled elsewhere in the Wittenoom Dolomite (see Becker, 1971). These samples were not available for the analysis of iron isotopes. 126 P.R. Craddock, N. Dauphas / Earth and Planetary Science Letters 303 (2011) 121–132

Table 3 Iron and carbon isotopic compositions of bulk metacarbonates from Isua Supracrsutal Belt, southern West Greeland.

Sample Description Fe (wt.%) δ13C(‰) a δ56Fe (‰) b δ57Fe (‰) b

Fe-rich metacarbonates IS-04-01 Siderite, northwest of belt 23.6 −5.50±0.06 0.236±0.059 0.296 ±0.093 IS-04-02 Siderite, northwest of belt 23.7 −4.52±0.05 0.318±0.125 0.500 ±0.169 IS-04-03 Siderite, northwest of belt 24.7 −4.73±0.02 0.448±0.074 0.763 ±0.151 IS-04-07 Metacarbonate mix, east of belt 59.4 −4.10 0.759±0.038 1.129 ±0.074 IS-04-07* Replicate, this study 58.1 0.743±0.039 1.098 ±0.045 IS-04-08 Siderite, east of belt 31.1 −4.52±0.06 0.400±0.034 0.587 ±0.048 IS-04-09 Graphite-rich metacarbonate 47.3 −4.75±0.05 0.294±0.069 0.496 ±0.165 IS-04-10 Siderite, east of belt 49.6 −5.94±0.05 0.281±0.084 0.400 ±0.110 IS-04-11 Graphite-rich metacarbonate 45.7 −4.58±0.04 0.204±0.056 0.383 ±0.114

Fe-poor metacarbonates IS-04-05 Calcite/dolomite, west of belt 2.6 −1.98±0.07 −0.742±0.046 −1.096 ±0.093 IS-04-05* Replicate, this study 2.9 −0.754±0.037 −1.138 ±0.043 AL-04-G18 Carbonated Amitsoq gneiss 3.2 −1.52±0.04 −0.901±0.177 −1.456 ±0.219 Al-04-G23 Carbonated Amitsoq gneiss 2.5 −0.34±0.06 −0.725±0.177 −1.051 ±0.219

a Carbon isotopic (δ13C) ratios are reported in per mil difference relative to the V-PDB standard. Uncertainties are two standard deviation for replicate analyses of the same sample. b Iron isotopic (δ56Fe, δ57Fe) ratios are those previously published by Dauphas et al. (2007b), except for two samples IS-04-07 and IS-04-05 indicated by *, which were independently measured for their iron isotopic composition as part of the present study for confirmation of accuracy. Iron isotopic ratios are reported in per mil difference relative to iron metal standard IRMM-014. Quoted uncertainties are 95% confidence intervals. data show a distribution of iron and carbon isotopic ratios that is similar to 4. Discussion that documented in the Hamersley Basin. To compare the isotopic signatures of metacarbonates from the 4.1. Evidence for an authigenic origin of iron-rich carbonates and ~3.8 Ga Isua Supracrustal Belt (ISB) with those from genuine late sedimentary microbial iron respiration at 2.5 Ga Archean BIFs, we combine new carbon isotope data of iron-rich and iron-poor metacarbonates measured in this study with previously Iron-rich carbonates in BIFs from the Hamersley Basin exhibit a reported iron isotope data for the same metacarbonates from Dauphas wide range of iron and carbon isotopic compositions. Can this isotopic et al. (2007). Iron-poor metacarbonates have light δ56Fe (−0.90 to heterogeneity be explained entirely by precipitation of dissolved Fe −0.73‰), whereas iron-rich metacarbonates have heavy δ56Fe (II) and inorganic carbon in the water column in near-isotopic (+0.24 to +0.76‰; Table 3, Fig. 2). Iron-poor metacarbonates have equilibrium? Or, are the isotopic signatures consistent with an δ13C values between −2.0 and 0‰, whereas iron-rich metacarbonates authigenic origin in marine sediments, possibly involving microbial have distinctly lighter δ13C values between −6.0 to −4.1‰ (Table 3; iron and carbon respiration? Fig. 2). Several studies have previously reported δ13C values of The δ13C of iron-poor carbonates of different ages ranging from metacarbonates from the ISB (Oehler and Smith, 1977; Perry and ~3.8 to 2.5 Ga are remarkably uniform and similar to that of platform Ahmad, 1977; Schidlowski et al., 1979; Ueno et al., 2002; van Zuilen carbonates deposited throughout the Phanerozoic (~0‰)(Becker and et al., 2003); only van Zuilen et al. (2003) have distinguished Clayton, 1972; Schidlowski et al., 1975; Shields and Veizer, 2002; geochemically between the carbon isotopic signatures of iron-poor Veizer et al., 1989, 1992). This argues for no significant changes (i.e., and iron-rich variants and none have iron isotopic data for the same b10‰) in the bulk carbon isotopic reservoir of dissolved inorganic samples that enable a direct comparison with our study. carbon (DIC) in seawater (c.f. Beukes et al., 1990; Kaufman et al., 1990). The uniform δ13C of iron-poor carbonates deposited synchro- nously in the late Archean across a range of water column depths from continental shelves (platform) to abyssal plains (basinal) argues also against vertical stratification with respect to carbon isotopes of DIC at seawater seawater 10 2+ - any given time (Fischer et al., 2009). Based on the experimentally Fe aq Isua HCO3 , aq determined carbon isotopic fractionation between siderite and DIC, ~3.8 Ga Δ13 − ‰ CHCO3-Sid = 0.5±0.2 (Jimenez-Lopez and Romanek, 2004), siderite precipitated from the water column in isotopic equilibrium Fe-rich Fe-poor 13 metacarbonate metacarbonate with DIC should have near-uniform δ C~0±2‰ (Fig. 3). The range 5 of δ13C between −6 and −15‰ documented in iron-rich carbonates

frequency magnetite from the Brockman Iron Formation in the Hamersley Basin is clearly Fe-poor Fe-rich metacarbonate metacarbonate different from that expected for, and argues against, formation of these carbonates in isotopic equilibrium with Archean seawater (Baur et al., 1985; Becker and Clayton, 1972). 0 Evidence from iron isotope data of the same iron-rich carbonates -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 -14 -12 -10 -8 -6 -4 -2 0 2 further argues against primary precipitation in Archean seawater. The δ56 δ13 Fe, IRMM-014 (‰) C, V-PDB (‰) prevailing consensus is that iron was delivered to the Archean ocean as ferrous Fe(II) , primarily via hydrothermal activity (e.g., Bau and Möller, Fig. 2. Histograms of iron (δ56Fe) and carbon (δ13C) isotope ratios in iron-rich aq metacarbonates (siderite, ankerite), iron-poor metacarbonates (calcite, dolomite) and 1993; Dymek and Klein, 1988; Jacobsen and Pimental-Klose, 1988). The 56 magnetite from the ~3.7 to 3.8 Ga Isua Supracrustal Belt, SW Greenland. Iron isotopic data δ FeFe(II) composition of modern hydrothermal fluids ranges from −0.1 are from Dauphas et al. (2007). Bin widths for iron and carbon isotope values are 0.025‰ and to −0.6‰, averaging −0.20‰ (n=19) (Beard et al., 2003; Severmann 1.0‰, respectively. The vertical arrows are the same as those depicted in Fig. 1.Despitea et al., 2004; Sharma et al., 2001). Given that the range of iron isotopic complex metamorphic history and metasomatic overprint, the iron and carbon isotopic data composition of Archean igneous rocks is limited and similar to modern aresimilartothoseofgenuinelate-ArcheanBIFsandarecompatiblewithderivationofiron- rich metacarbonates in the ISB from reduction of ferric oxides (ferrihydrite), probably (e.g., Dauphas et al., 2009a), hydrothermal alteration of oceanic crust coupled to oxidation of organic carbon. in the Archean should contribute Fe(II) with an isotopic composition also P.R. Craddock, N. Dauphas / Earth and Planetary Science Letters 303 (2011) 121–132 127

1.5 Iron Formation are heavier than −0.5‰ andextendupto+1.2‰ (Fig. 3). ~2.5 Ga These isotopic signatures cannot be explained by direct precipitation in 1.0 near isotopic equilibrium with bulk Archean seawater.

Kuruman If iron-rich carbonates were not directly precipitated from seawater, then what were the chemical and/or biological transformations of iron 0.5 Fe-rich and carbon that prevailed to their formation? The range of light δ13C Fe (‰)

56 ratios measured in iron-rich carbonates from the Brockman Iron δ 0.0 Formation requires a source of carbon in the iron-formation with a light isotopic composition. Becker and Clayton (1972) have previously -0.5 Brockman argued that the most reasonable source of this light carbon was organic Fe-rich

carbonate matter. The carbon isotopic compositions of fossil organic matter in the Hamersley Basin and other Precambrian iron-formations fall in a -1.0 restricted range of very light δ13Cbetween−30 and −33‰ (Barghoorn et al., 1977; Brocks et al., 1999; Perry et al., 1973). Fossil biomarkers Fe-poor carbonates 1.5 preserved in the Marra Mamba Iron and Jeerinah Formations of the ~3.8 Ga Hamersley Basin (Brocks et al., 1999) and in iron- and carbonate- formations of the penecontemporaneous Transvaal Supergroup (Wald- 1.0 bauer et al., 2009) provide support for oxygenic/anoxygenic photosyn- Isua, Fe-rich metacarbonates thesis and organic carbon production in surface seawater in the late 0.5 Archean. Diagenetic oxidation of organic carbon would deliver a pool

Fe (‰) of isotopically-light carbon (as CO2 or dissolved carbonate) that 56 seawater δ 0.0 could dilute the existing pore water reservoir of DIC and accumulate in marine sediments until saturation with respect to authigenic car- bonates was obtained. Whereas carbonates with δ13Cratios~−30‰ -0.5 would have formed entirely from an organic-derived pool of carbonate carbonate carbon, iron-rich carbonates in the Brockman Iron Formation (average Isua, Fe-poor metacarbonates -1.0 δ13C=−9.8±0.2‰) precipitated from a pore water reservoir of carbonate carbon contributed by both organic and DIC sources, the latter possibly supported by partial exchange of DIC across the -1.5 -16 -14 -12 -10 -8 -6 -4 -2 0 sediment–seawater interface. carbonate δ13C (‰) Oxidation of organic carbon must be coupled to reduction of an appropriate electron acceptor. In modern marine sediments the order of δ56 δ13 Fig. 3. Coupled Fe and C ratios iron-rich and iron-poor carbonates from the oxidant consumption is O2 NMn-oxidesNnitrateNFe-oxidesNsulfate Brockman Iron Formation (Becker and Clayton, 1972, this study) and Kuruman Iron (Froelich et al., 1979). In anoxic marine sediments of the Archean, Formation (Heimann et al., 2010; Johnson et al., 2003) (top panel) and from the Isua oxygen would have been severely limited (if not absent) as the primary Supracrustal Belt (Dauphas et al., 2007, this study) (bottom panel). The white box delineates the estimated isotopic composition of Archean seawater. The gray box is the electron acceptor, as would Mn-oxide and nitrate (e.g., Anbar and predicted range of isotopic compositions for iron-rich carbonates precipitated directly Holland, 1992; Anbar et al., 2007; Chapman and Schopf, 1983; Walker, from and in isotopic equilibrium with Archean seawater based on experimental (Jimenez- 1984). Banded iron-formations are anomalously rich in iron and Lopez and Romanek, 2004; Wiesli et al., 2004) and theoretical estimates (Anbar et al., reduction of ferric oxides may have permitted oxidation of organic 2005; Blanchard et al., 2009; Schauble et al., 2001) of equilibrium isotope fractionation carbon (Dimroth and Chauvel, 1973; Walker, 1984). Studies have between Fe(II)aq-siderite and HCO3-siderite. Iron-poor carbonates from both the ~2.5 Ga Brockman and Kuruman Iron Formations and the ~3.8 Ga Isua Supracrustal Belt have iron demonstrated that oxidation–reduction between ferric oxide and and carbon isotopic compositions very similar to that expected for carbonate precipitation organic carbon is effectively mediated by microbes via the process of in Archean seawater. In contrast, iron-rich carbonates from these iron-formations have a DIR (Lovley, 1993), which can be illustrated by the stoichiometric wide range of iron and carbon isotopic ratios that are inconsistent with their precipitation reaction for ferrihydrite, in equilibrium with common seawater.

d þ→ 2+ – ðÞ: 2Fe2O3 nH2O+CH2O+7H 4Fe +HCO3 +2n+4H2O ð Þ ~−0.20‰. It is likely that the Archean ocean, at least below the surface 1 mixed layer (~100 m), was well-mixed and homogeneous with respect to − iron isotopes because the reservoir of Fe(II)aq required to form BIFs must We write HCO3 as the carbonate product, reflecting the dominance of − be considerable (20 to 100 mg Fe L 1) and continually replenished this species in seawater at equilibrium (e.g., Walker, 1983). Previous (Ewers and Morris, 1981; Morris, 1993). In effect, the residence time of studies have applied simple mass balance models to estimate the iron in the Archean ocean was significantly longer than the oceanic overall potential for Fe(III) reduction in the formation of iron oxides mixing time (Johnson et al., 2003). Iron-rich carbonates precipitated in and carbonates in BIFs and implicate a significant role for microbial equilibrium with Archean seawater are predicted to have light iron iron respiration in the late Archean (e.g., Konhauser et al., 2005; isotopic ratios between ~−0.5 and −2‰ (Fig. 3), based on the net isotope Walker, 1984). fractionation between carbonate (siderite) and Fe(II)aq determined by Our isotopic data for magnetite and iron-rich carbonates from the 56 experiments, Δ FeSid–Fe(II) =−0.5±0.2‰ (Wiesli et al., 2004), and Brockman Iron Formation provide an independent assessment of 56 theoretical calculations, Δ FeSid–Fe(II) =−1.6 to −2.1‰ (Anbar et al., these results and demonstrate that the isotopic signatures imprinted 2005; Blanchard et al., 2009; Schauble et al., 2001). Such low δ56Fe values in BIFs are consistent with extensive microbial iron respiration at are found in iron-poor carbonates from the Wittenoom Dolomite (Fig. 3), ~2.5 Ga. Ferric iron in magnetite requires the oxidation and a platform carbonate formation in the Hamersley Group underlying the precipitation of dissolved Fe(II)aq. The process by which large δ56 Brockman Iron Formation. Similar light Fe values are observed in amounts of Fe(II)aq were oxidized in the Archean ocean that was platform carbonates associated with other late Archean and Paleoproter- anoxic is debated, but the primary Fe(III) precipitates in all cases were ozoic BIFs (Heimann et al., 2010). In contrast, the iron isotopic likely amorphous ferric oxides (Ewers and Morris, 1981; Kappler and compositions measured in most iron-rich carbonates from the Brockman Newman, 2004; Morris, 1993; Trendall and Blockley, 1970), such as 128 P.R. Craddock, N. Dauphas / Earth and Planetary Science Letters 303 (2011) 121–132 ferrihydrite. Experimental studies have demonstrated that amor- 2.5 phous ferric oxides from partial Fe(II)aq oxidation are enriched in the 1:1 heavy isotopes of iron relative to starting Fe(II)aq, with an equilibrium 2.0 56 fractionation factor ranging from Δ Fe ~ +0.9‰ for O2-mediated oxidation (Bullen et al., 2001) to +2.0‰ for direct microbial oxidation 1.5 (Balci et al., 2006; Beard et al., 2010; Croal et al., 2004). Thus, partial oxidation of Fe(II) in Archean seawater should produce ferrihydrite aq Fe (‰) 56 1.0 with positive δ Fe values up to +1.5‰. Near-complete Fe(II) 56 δ oxidation in surface seawater would produce ferrihydrite with a 56 0.5 δ Fe value of ~0‰, similar to that of the starting reservoir of Fe(II)aq, 56 Δ which is a lower limit on the range of δ Fe in the primary ferric Sid-Mgt oxides. Magnetite in the Brockman Iron Formation was likely formed carbonate 0.0 by the reaction of primary ferrihydrite with ferrous iron (Ewers and Morris, 1981; Morris, 1993): -0.5 -1.8 ‰ -2.3 ‰ equilibrium – Fe O d nH O+Fe2+ + 2OH →Fe O +n+1ðÞH O: ð2Þ 2 3 2 3 4 2 -1.0 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 Magnetite formed by reaction between ferrihydrite and Fe(II)aq in the magnetite δ56Fe (‰) water column can have positive δ56Fe values between ~0 and +1.0‰, 56 reflecting the contributions of iron from ferrihydrite (δ Fe ~0 to +1.5‰) Fig. 4. Iron isotopic compositions of paired magnetite and iron-rich carbonates (siderite, 56 56 and Fe(II)aq (δ Fe~ −0.2‰) in the ratio 2:1. Measured δ Fe ratios of ankerite) from the same laminations (chert-matrix mesobands) from the Brockman Iron magnetite from the Brockman Iron Formation are as light as −1.0‰, Formation. The isotopic fractionation between siderite and magnetite (light gray lines) is estimated to be between −1.8‰, based on laboratory experiments (Johnson et al., 2005; which is difficult to explain unless the pool of Fe(II) from which aq Wiesli et al., 2004), and −2.3‰, based on spectroscopy and theoretical calculations of − − ‰ magnetite formed was also very light, approaching 1to 2 .Themost vibrational states of isotopic bonding (Blanchard et al., 2009; Polyakov et al., 2007). The 56 likely source of Fe(II)aq with very light δ Fe values is from partial gray parallelogram delineates the range of iron isotopic ratios expected for iron-rich reduction of ferrihydrite in marine sediments via microbial DIR (Eq. (1); carbonates and magnetite precipitated in isotopic equilibrium with Archean seawater. The also see Johnson et al. 2008; Heimann et al. 2010). Indeed, experimental measured iron isotopic compositions of most magnetite and iron-rich carbonates are fi incompatible with this model. Note the overall positive correlation between the iron studies have documented a signi cant isotopic effect associated with isotopic compositions of magnetite and siderite, which may reflect formation of most iron- partial Fe(III) reduction via DIR that yields reduced Fe(II)aq depleted by rich carbonates and magnetite in association during diagenetic iron and carbon cycling. −1.0 to −2.5‰ in δ56Fe relative to precursor ferric oxides (Beard et al., The blue band defines the 95% confidence interval of the regression (slope=0.69±0.33; 56 1999, 2010; Crosby et al., 2007; Icopini et al., 2004; Johnson et al., 2005; the two data with light carbonate δ Fe are omitted from this regression because these data may reflect near-isotopic equilibrium with seawater). Tangalos et al., 2010). Iron-rich carbonates in the Brockman Iron Formation exhibit a wide range of δ56Fe values (Fig. 1). Iron-rich carbonates with heavy δ56Fe compositions up to +1‰ could only have precipitated from a idea that the relative proportions of organic carbon and ferric iron 56 pool of Fe(II)aq with positive δ Fe. This reservoir of iron was not controlled the diagenetic fate of the sediment; preserving magnetite common seawater, but was an authigenic reservoir contributed from only when ferric iron was in excess relative to organic matter. near-complete reduction in marine sediments of primary ferrihydrite Most, if not all, Archean and Paleoproterozoic BIFs have experi- that had heavy δ56Fe values following partial Fe(II) oxidation in surface enced varying degrees of metamorphism, which can affect the stable seawater. Microbial DIR (Eq. (1)) produces both dissolved Fe(II) and mineral assemblage observed in iron-formations (e.g., French, 1973; carbonate that would accumulate in sediment pore waters until Klein, 1983, 2005). A possible burial metamorphic origin for carbonate saturation with respect to iron-rich carbonates was attained: in the Brockman Iron Formation via inorganic reaction between primary ferric iron, such as ferrihydrite, and organic carbon at 2+ –→ þ: ð Þ Fe + HCO3 FeCO3 +H 3 elevated temperature and pressure can, however, be discounted on the basis of mass balance for iron and carbon (DIC) in BIFs. During Light δ13C values of the same iron-rich carbonates (Fig. 3) have metamorphism, the sediment is isolated from exchange with the been shown to be consistent with derivation of carbonate ion from ocean. The maximum concentration of inorganic carbonate in pore oxidation of organic matter during microbial DIR (Baur et al., 1985; water at the onset of metamorphism can be estimated by assuming

Becker and Clayton, 1972; Walker, 1984). equilibrium with atmospheric CO2. The partial pressure, pCO2, of the Additional support for microbial iron respiration is provided by the Archean atmosphere is debated (e.g., Hessler et al., 2004; Kasting, correlated iron isotopic ratios of magnetite and iron-rich carbonates 1993; Lowe and Tice, 2004; Rosing et al., 2010; Rye et al., 1995; sampled from the same mesobands and microbands within the Brock- Walker, 1985), but pCO2 between 0.1 and 1 bar (≫100 times present) man Iron Formation (Fig. 4). Following reduction of ferrihydrite, is a reasonable estimate. A kilogram of BIF in the Dales Gorge Member authigenic magnetite and siderite can precipitate from the same of the Brockman Iron Formation contains on average 6 moles of iron 56 reservoir of Fe(II)aq. Thus, the δ Fe signature of Fe(II)aq produced by (46.37 wt.% Fe2O3; Ewers and Morris, 1981). This quantity of microbial iron reduction can be inherited by both magnetite and iron- sediment corresponds to a pore volume of ~3.4×103 cm3 assuming rich carbonate. Our interpretations are consistent with recent micro- a density of the iron-formation of ~3.4 g/cm3 (Ewers and Morris, analytical studies of the Brockman Iron Formation that have documented 1981) and accounting for post-depositional compaction by up to iron isotopic variations in magnetite and iron-rich carbonates on spatial 90% (Trendall and Blockley, 1970). This pore volume could contain scales less than one millimeter, which are suggested to reflect post- up to ~0.12 moles of total dissolved carbonate. According to reaction 4 depositional redistribution of iron within the sediment (Steinhoefel (net reaction between ferrihydrite and organic matter to yield Fe- et al., 2010). Geochemical studies have shown that for the penecontem- rich carbonate; note that 3 moles of DIC is needed for each mole of paraneous Kuruman Iron Formation, carbonate facies contain higher organic C): residual organic carbon contents up to an order of magnitude higher than in oxide (magnetite) facies in the iron-formation (Beukes et al., 2Fe O :nH O þ CH O þ 3HCO− þ 3Hþ→4FeCO þð2n þ 4ÞH O; ð4Þ 1990; Fischer et al., 2009). These observations are consistent with the 2 3 2 2 3 3 2 P.R. Craddock, N. Dauphas / Earth and Planetary Science Letters 303 (2011) 121–132 129 this reservoir of dissolved carbonate would be sufficient to convert up origin. In support, alteration of oceanic crust at modern seafloor to 0.15 moles of iron in ferrihydrite to siderite, which is only ~2% of hydrothermal environments yields secondary mineral assemblages the total inventory of Fe in the iron-formation. Magnetite and iron with a heavy iron isotopic composition and an implied fluid with a carbonate constitute on average 30 wt.% and 15 wt.%, respectively, of complementary light iron isotopic ratio (Rouxel et al., 2003). Heavy the Dales Gorge Member in the Brockman Iron Formation (Trendall δ56Fe ratios up to +0.80‰ in iron-rich metacarbonates, however, are and Blockley, 1970). Assuming carbonate is siderite, the average mole inconsistent with mobilization by metasomatic fluids of isotopically fraction of iron present as carbonate is 25% (assuming ankerite gives a light iron from an igneous protolith. Dauphas et al. (2007) showed that minimum of 10%), which is considerably greater than the 2% of Fe that iron-rich metacarbonates had similar heavy iron isotopic compositions could be converted to carbonate through metamorphism. The to magnetite from BIFs but the study was inconclusive as to the nature of calculations imply that an external source of carbonate is required this relationship. New carbon isotopic data coupled to existing iron to yield the large quantities of iron-rich carbonates observed in late isotopic data for the iron-rich and iron-poor metacarbonates from the Archean iron-formations. We suggest that this carbonate was made ISB provide further constraints on a possible chemical sedimentary available both from oxidation of organic carbon and through partial versus metasomatic origin for the metacarbonates. An authigenic origin, exchange of DIC between sediment pore water and the overlying similar to that indicated for iron-rich carbonates from the younger water column, and could only proceed during early diagenetic Hamersley Basin, would implicate the evolution of microbial metabolic transformation. This idea is consistent with several petrographic pathways (oxygenic or anoxygenic photosynthesis, DIR) by ~3.8 Ga. studies that suggested an early diagenetic origin for iron-rich The ISB metacarbonates fall into two isotopically distinct groups carbonates in BIFs (e.g., Ewers and Morris, 1981; Morris, 1993; Pecoits (Fig. 3). Iron-poor metacarbonates that have light δ56Fe have near-zero et al., 2009; Trendall and Blockley, 1970). δ13Cratios(−3to0‰), whereas Fe-rich metacarbonates that have Finally, we note that an interpretation to explain the light δ56Fe heavy δ56Fe have distinctly lighter δ13Cratios(meanof−4.8±0.6‰). values of minerals from late Archean BIF sequences (e.g., pyrite in An important consideration in interpreting these data is whether these shale units) has been proposed by Rouxel et al. (2005) whereby isotopic signatures are primary and record the conditions of precipita- partial oxidation and precipitation of ferric oxides in BIFs in the water tion, or if these have been disturbed by metamorphism. All rocks in the column leaves the residual Fe(II)aq reservoir enriched in the light ISB have been subject to amphibolite facies metamorphism, with peak isotopes of iron. This reservoir is subsequently precipitated as ferrous- temperatures ~500 to 550 °C and pressures ~5 kbar (e.g., Boak and bearing minerals. We have shown that the light δ56Fe values of iron- Dymek, 1982). P–T phase relations of the reaction (Lamb, 2005), bearing minerals in BIFs can be produced by iron cycling during → ð Þ diagenesis and concur with Johnson et al. (2008) that the highly- FeCO3 +SiO2 FeSiO3 +CO2;g 5 fractionated and stratigraphically variable iron isotopic compositions of iron oxides and carbonates in oxide facies BIFs primarily reflect suggest that siderite formed during diagenesis would have survived diagenetic pathways for their formation. peak metamorphic conditions (Fig. 5). We interpret the different iron and carbon isotopic compositions of 4.2. Iron and carbon isotope evidence to support microbial iron iron-poor and iron-rich metacarbonates in the ISB as reflecting their respiration in the early Archean original formation through distinct pathways. The field relationship between iron-poor metacarbonates and ultramafic host rocks are

We now turn our attention to interpreting the iron and carbon supportive of metasomatic overprint by leaching of iron by a CO2-bearing isotope signatures of metacarbonates from the ~3.8 Ga Isua Supracrustal fluid from an ultramafic protolith (Rose et al., 1996; Rosing et al., 1996). Belt (ISB). The origin of ISB metacarbonates is contentious, with The light iron isotopic compositions of iron-poor metacarbonates are protolith identification ranging from sedimentary (Bolhar et al., 2004; possibly consistent with derivation from iron mobilized from ultramafic Dymek and Klein, 1988; Mojzsis et al., 1996; Nutman et al., 1984)to rocks (Dauphas et al., 2007). The carbon and iron isotopic data combined, entirely metasomatic (Myers, 2001; Rose et al., 1996; Rosing et al., however, reveal that the isotopic characteristics of iron-poor metacarbo- 1996). Iron-rich metacarbonates were initially interpreted as carbonate nates from the ISB are very similar to those of iron-poor carbonates facies iron-formations (Dymek and Klein, 1988; Nutman et al., 1984). In this interpretation framework, metacherts and calc-silicates associated with iron-rich metacarbonates were formed by reaction of sedimentary 9 a carbonates and quartz during burial and high-grade metamorphism, Siderite + Qtz = Ferrosilite + CO2 and the banding documented in metacherts was indicated to preserve 8 b that of the sedimentary protolith (Dymek and Klein, 1988; Nutman e c et al., 1984). Trace element characteristics (e.g., REE+Y) of these rocks 7 d were interpreted as being consistent with their deposition as chemical XCO2 = 1 sediments in seawater (Bolhar et al., 2004; Frei and Polat, 2007). f 6 Alternatively, iron-rich and iron-poor metacarbonates and calc-silicates have been interpreted as metasomatic in origin, formed by carbonation Pressure (kbar) Isua P-T and desilication of igneous protoliths. A metasomatic origin for some 5 metacarbonates, in particular iron-poor variants, was indicated by the discordant nature of calc-silicate and metacarbonate units, veining, 4 500550 600 650700 750 replacive textures and of the igneous lithologies within which these units occur (Rose et al., 1996; Rosing et al., 1996). Temperature (˚C) In a recent publication, Dauphas et al. (2007) have reported the iron Fig. 5. Pressure–temperature phase relationship for the reaction siderite+quartz→ isotopic compositions and trace element geochemistry of iron-rich and ferrosilite+CO2,g (shown by black line, a). Shown in the solid gray lines are equivalent iron-poor metacarbonates from the ISB in order to better distinguish a phase boundaries for magnesite+quartz→enstatite+CO2,g (b), dolomite+quartz→ → metasomatic versus sedimentary origin. Iron-poor and iron-rich diopside+CO2,g (c), ankerite+quartz hedenbergite+CO2,g (d), calcite+enstatite+ → → metacarbonates have light and heavy δ56Fe ratios, respectively quartz diopside+CO2,g (e) and calcite+ferrosilite+quartz hedenbergite+CO2,g (f). Phase boundaries are reproduced from Lamb (2005). The peak metamorphic P–T conditions (Fig. 2). The light δ56Fe of iron-poor metacarbonates was interpreted (~5 kbar, 550 °C) to which rocks in the ISB were subjected (e.g., Boak and Dymek, 1982)are fl as re ecting mobilization of isotopically light iron from pre-existing shownbythegraybox.Thephaserelationsindicatethatsiderite(andothercarbonates) maficorultramafic protoliths by fluid, consistent with a metasomatic would have been stable during peak metamorphism. 130 P.R. Craddock, N. Dauphas / Earth and Planetary Science Letters 303 (2011) 121–132 preserved in association with late Archean and Paleoproterozoic BIFs 3. Iron-rich carbonates with heavy δ56Fe (up to +1.2‰) must have 56 (Fig. 3). By analogy, these isotopic signatures can instead be interpreted as precipitated from a reservoir of Fe(II)aq with very positive δ Fe. 56 reflecting formation of iron-poor metacarbonates as primary precipitates The source of heavy δ Fe–Fe(II)aq was likely from near-complete in Archean seawater. The REE patterns of these carbonates also appear to microbial reduction of ferric oxides (ferrihydrite) in marine be supportive of their formation as chemical precipitates from seawater sediments. The light δ13C of the same iron-rich carbonates was (Dauphas et al., 2007). Thus, while iron-poor metacarbonates have derived from carbonate produced by the oxidation of organic previously been suggested to be entirely metasomatic in origin, we carbon that was probably coupled to reduction of iron. caution that metasomatic mobilization of a primary (i.e., chemical 4. The combined iron and carbon isotopic data support an authigenic sedimentary) carbonate cannot be excluded. Continued study of the origin for iron-rich carbonates in the Hamersley Basin via coupled isotopic and chemical signatures of iron-poor metacarbonates is required organic carbon oxidation and ferrihydrite reduction. This process is to unambiguously determine the primary derivation of these rocks. effectively mediated by microbes in marine sediments though Iron-rich metacarbonates from the ISB have iron and carbon isotopic dissimilatory iron reduction (DIR), which implicates extensive compositions similar to those of other iron-rich carbonates of known BIF microbial iron respiration in the formation of late Archean BIFs. association (Fig. 3). To the extent that iron-rich metacarbonates of the 5. Iron-rich metacarbonates from the ~3.8 Ga ISB have δ56Fe and δ13C ISB share the same isotopic characteristics with those of late Archean signatures similar to those of carbonates in late Archean BIFs. The iron-formations, they may have formed by the same process. The heavy isotopic data are interpreted as reflecting formation of these iron- δ56Fe values of iron-rich metacarbonates cannot be explained by rich metacarbonates as marine authigenic precipitates through metasomatic mobilization of isotopically-light iron from igneous microbial iron respiration. Despite metasomatic overprint, iron- protoliths. Instead, it is now recognized from studies of genuine BIFs rich metacarbonates in the ISB preserve primary isotopic char- that iron-rich carbonates of a known authigenic origin can carry heavy acteristics supporting evolution of microbial iron catabolism by δ56Fe resulting from microbial iron respiration of primary ferric oxide ~3.8 Ga during the formation of some of the oldest recognized precipitates also with heavy δ56Fe (Heimann et al., 2010, this study). sedimentary-bearing rocks on Earth. Magnetite in BIFs from the ISB has heavy δ56Fe (Dauphas et al., 2004b, 2007; see also Whitehouse and Fedo, 2007). Microbial Fe(III) reduction Acknowledgments of this magnetite or other precursor ferric oxide would transfer the heavy isotopic composition to the ferrous iron product and subse- Drill core material of the Brockman Iron Formation from Hamersley quently to iron-rich carbonates, which is exactly that observed. Further was made available by the Geological Survey of Western Australia. R. N. evidence that microbial iron respiration (DIR) was involved in the Clayton provided hand specimens of the Wittenoom Dolomite from authigenic formation of iron-rich metacarbonates comes from the light Hamersley. We gratefully acknowledge contributions to this research by δ13C values of these samples. Direct evidence for reduced carbon of R. H. Becker who previously carried out the mineral separation of the biogenic origin in the ISB is disputed (e.g., Perry and Ahmad, 1977; Hamersley samples for carbon and oxygen isotope analysis. Metacarbo- Mojzsis et al., 1996; Rosing, 1999; Schidlowski et al., 1979; van Zuilen nate samples from Isua were provided by M. van Zuilen and A. Lepland. et al., 2002; 2003). Still, the light carbon isotopic compositions of iron- The manuscript benefited from discussions with R. N. Clayton, and from rich metacarbonates from the ISB mirror those measured of late Archean constructive reviews by K. Konhauser and an anonymous reviewer. This BIFs (Baur et al., 1985; Becker and Clayton, 1972; Heimann et al., 2010), research was supported by National Science Foundation through grant which are consistent with derivation of carbonate from oxidation of EAR-0820807 (Geobiology), National Aeronautics and Space Adminis- organic carbon during iron respiration. We conclude that iron-rich tration through grant NNX09AG59G (Cosmochemistry) and a Packard metacarbonates associated with iron-formation in the ISB formed by Fellowship to N.D. similar microbially-mediated iron and carbon transformations as documented in late Archean BIFs. The antiquity of microbial iron Appendix A. Supplementary data respiration, as suggested by our iron and carbon isotope data from one of the oldest known sedimentary sequences, is consistent with results of Supplementary data to this article can be found online at phylogenetic studies. Ferric iron reduction has been documented as a doi:10.1016/j.epsl.2010.12.045. metabolic pathway within a large diversity of extent Bacteria and Archaea including those most closely related to the last common References ancestor of modern life, which points toward iron respiration as one of the earliest forms of microbial metabolism (Liu et al., 1997; Lovley, Anbar, A.D., Duan, Y., Lyons, T.W., Arnold, G.L., Kendall, B., Creaser, R.A., Kaufman, A.J., 1991; Vargas et al., 1998; Weber et al., 2006). Gordon, G.W., Scott, C., Garvin, J., Buick, R., 2007. A whiff of oxygen before the Great Oxidation Event? Science 317, 1903–1906. Anbar, A.D., Holland, H.D., 1992. The photochemistry of manganese and the origin of 5. Summary and conclusions banded iron formations. Geochim. Cosmochim. Acta 56, 2595–2603. Anbar, A.D., Jarzecki, A.A., Spiro, T.G., 2005. Theoretical investigation of iron isotope 3+ 2+ fractionation between Fe(H2O)6 and Fe(H2O)6 : implications for iron stable Iron and carbon isotopic analyses of iron oxides and carbonates in isotope geochemistry. Geochim. Cosmochim. Acta 69, 825–837. BIFs can be used to constrain the pathways of their formation. Here, we Balci, N., Bullen, T.D., Witte-Lien, K., Shanks, W.C., Motelica, M., Mandernack, K.W., report the iron (δ56Fe, vs. IRMM-014) and carbon (δ13C, vs. V-PDB) 2006. Iron isotope fractionation during microbially stimulated Fe(II) oxidation and Fe(III) precipitation. Geochim. Cosmochim. Acta 70, 622–639. isotopic compositions of magnetite and of iron-rich and iron-poor Barghoorn, E.S., Knoll, A.H., Dembicki Jr., H., Meinschein, W.G., 1977. Variation in stable carbonates from the 2.5 Ga Brockman Iron Formation in the Hamersley carbon isotopes in organic matter from the Gunflint Iron Formation. Geochim. Basin, Australia, and the ~3.8 Ga Isua Supracrustal Belt (ISB), West Cosmochim. Acta 41, 425–430. Greenland. The key results and implications from this study are: Bau, M., Möller, P., 1993. Rare earth element systematics of the chemically precipitated component in early precambrian iron formations and the evolution of the terrestrial atmosphere–hydrosphere–lithosphere system. Geochim. Cosmochim. 1. Magnetite and iron-rich carbonates (siderite, ankerite) from the Acta 57, 2239–2249. Hamersley Basin preserve a wide range of δ56Fe values that are Baur, M.E., Hayes, J.M., Studley, S.A., Walter, M.R., 1985. Millimeter-scale variations of stable isotope abundances in carbonates from banded iron-formations in the incompatible with direct precipitation in isotopic equilibrium with Hamersley Group of Western Australia. Econ. Geol. 80, 270–282. Archean seawater. Beard, B.L., Handler, R.M., Scherer, M.M., Wu, L., Czaja, A.D., Heimann, A., Johnson, C.M., 2. Magnetite with light δ56Fe (≪0‰) must have precipitated from a 2010. Iron isotope fractionation between aqueous ferrous iron and goethite. Earth – δ56 − − ‰ Planet. Sci. Lett. 295, 241 250. pool of Fe(II)aq with light Fe between 1 and 2 , which was Beard, B.L., Johnson, C.M., Cox, L., Sun, H., Nealson, K.H., Aguilar, C., 1999. Iron isotope likely produced through microbial partial reduction of ferrihydrite. biosignatures. Science 285, 1889–1892. P.R. Craddock, N. Dauphas / Earth and Planetary Science Letters 303 (2011) 121–132 131

Beard, B.L., Johnson, C.M., Von Damm, K.L., Poulson, R.L., 2003. Iron isotope constraints Garrels, R.M., Perry, E.A., Mackenzie, F.T., 1973. Genesis of Precambrian iron-formations on Fe cycling and mass balance in oxygenated Earth oceans. Geology 31, 629–632. and the development of atmospheric oxygen. Econ. Geol. 68, 1173–1179. Becker, R.H., 1971. Carbon and oxygen isotope ratios in iron-formation and associated Heimann, A., Johnson, C.M., Beard, B.L., Valley, J.W., Roden, E.E., Spicuzza, M.J., Beukes, rocks from the Hamersley Range of Western Australia and their implications. Ph.D. N.J., 2010. Fe, C, and O isotope compositions of banded iron formation carbonates Thesis, The University of Chicago. demonstrate a major role for dissimilatory iron reduction in 2.5 Ga marine Becker, R.H., Clayton, R.N., 1972. Carbon isotopic evidence for the origin of a banded environments. Earth Planet. Sci. Lett. 294, 8–18. iron-formation in Western Australia. Geochim. Cosmochim. Acta 36, 577–595. Hessler, A.M., Lowe, D.R., Jones, R.L., Bird, D.K., 2004. A lower limit for atmospheric Becker, R.H., Clayton, R.N., 1976. Oxygen isotope study of a Precambrian banded iron- carbon dioxide levels 3.2 billion years ago. Nature 428, 736–738. formation. Hamersley Range. Western Australia. Geochim. Cosmochim. Acta 40, Icopini, G.A., Anbar, A.D., Ruebush, S.S., Tien, M., Brantley, S.L., 2004. Iron isotope fractionation 1153–1165. during microbial reduction of iron: the importance of adsorption. Geology 32, 205–208. Beukes, N.J., Klein, C., Kaufman, A.J., Hayes, J.M., 1990. Carbonate petrography, kerogen Jacobsen, S.B., Pimental-Klose, M.R., 1988. A Nd isotopic study of the Hamersley and distribution and carbon and oxygen isotope variations in an Early Proterozoic Michipicoten banded iron formations: the source of REE and Fe in Archean oceans. transition from limestone to iron-formation deposition, Transvaal Supergroup, Earth Planet. Sci. Lett. 87, 29–44. South Africa. Econ. Geol. 85, 663–690. James, H.L., 1954. Sedimentary facies of iron-formation. Econ. Geol. 49, 235–293. Blanchard, M., Poitrasson, F., Méheut, M., Lazzeri, M., Mauri, F., Balan, E., 2009. Iron James, H.L., 1983. Distribution of Banded Iron-Formation in space and time. In: Trendall,

isotope fractionation between pyrite (FeS2), hematite (Fe2O3) and siderite (FeCO3): A.F., Morris, R.C. (Eds.), Iron-Formation: Facts and Problems. Elsevier, Amsterdam, a first-principles density functional theory study. Geochim. Cosmochim. Acta 73, pp. 471–490. 6565–6578. Jimenez-Lopez, C., Romanek, C.S., 2004. Precipitation kinetics and carbon isotope partitioning Boak, J.L., Dymek, R.F., 1982. Metamorphism of the ca. 3800 Ma supracrustal rocks at of inorganic siderite at 25 °C and 1 atm. Geochim. Cosmochim. Acta 68, 557–571. Isua, West Greenland: implications for early Archaean crustal evolution. Earth Johnson, C.M., Beard, B.L., Beukes, N.J., Klein, C., O'Leary, J.M., 2003. Ancient geochemical Planet. Sci. Lett. 59, 155–176. cycling in the Earth as inferred from Fe isotope studies of banded iron formations Bolhar, R., Kamber, B.S., Moorbath, S., Fedo, C.M., Whitehouse, M.J., 2004. Characterisa- from the Transvaal Craton. Contrib. Mineralog. Petrol. 144, 523–547. tion of early Archaean chemical sediments by trace element signatures. Earth Johnson, C.M., Beard, B.L., Klein, C., Beukes, N.J., Roden, E.E., 2008. Iron isotopes Planet. Sci. Lett. 222, 43–60. constrain biologic and abiologic processes in banded iron formation genesis. Braterman, P.S., Cairns-Smith, A.G., Sloper, R.W., 1983. Photo-oxidation of hydrated Fe2+ — Geochim. Cosmochim. Acta 72, 151–169. significance for banded iron formations. Nature 303, 163–164. Johnson, C.M., Roden, E.E., Welch, S.A., Beard, B.L., 2005. Experimental constraints on Fe Brocks, J.J., Logan, G.A., Buick, R., Summons, R.E., 1999. Archean molecular fossils and isotope fractionation during magnetite and Fe carbonate formation coupled to the early rise of eukaryotes. Science 285, 1033–1036. dissimilatory hydrous ferric oxide reduction. Geochim. Cosmochim. Acta 69, 963–993. Bullen, T.D., White, A.F., Childs, C.W., Vivit, D.V., Schulz, M.S., 2001. Demonstration of Kappler, A., Newman, D.K., 2004. Formation of Fe(III)-minerals by Fe(II)-oxidizing significant abiotic iron isotope fractionation in nature. Geology 29, 699–702. photoautotrophic bacteria. Geochim. Cosmochim. Acta 68, 1217–1226. Cairns-Smith, A.G., 1978. Precambrian solution photochemistry, inverse segregation, Kappler, A., Pasquero, C., Konhauser, K.O., Newman, D.K., 2005. Deposition of banded and banded iron formations. Nature 276, 807–808. iron formations by anoxygenic phototrophic Fe(II)-oxidizing bacteria. Geology 33, Canfield, D.E., Habicht, K.S., Thamdrup, B., 2000. The Archean sulfur cycle and the early 865–868. history of atmospheric oxygen. Science 288, 658–661. Kasting, J.F., 1987. Theoretical constraints on oxygen and carbon dioxide concentrations Chapman, D.J., Schopf, J.W., 1983. Biological and biochemical effects of the development in the Precambrian atmosphere. Precambrian Res. 34, 205–229. of an aerobic environment. In: Schopf, J.W. (Ed.), Earth's Earliest Biosphere: Its Kasting, J.F., 1993. Earth's early atmosphere. Science 259, 920–926. origin and Evolution. Princeton University Press, Princeton, NJ, pp. 302–320. Kaufman,A.J.,Hayes,J.M.,Klein,C.,1990.Primaryanddiageneticcontrolsofisotopic Cloud, P., 1965. Significance of the Gunflint (Precambrian) microflora. Science 148, compositions of iron-formation carbonates. Geochim. Cosmochim. Acta 54, 3461–3473. 27–35. Klein, C., 1983. Diagenesis and metamorphism of Precambrian banded iron-formations. Cloud, P., 1973. Paleoecological significance of the banded iron-formation. Econ. Geol. In: Trendall, A.F., Morris, R.C. (Eds.), Iron-Formation: Facts and Problems. Elsevier, 68, 1135–1143. Amsterdam, pp. 417–469. Craddock, P.R., Dauphas, N., 2010. Iron isotopic compositions of geological reference Klein, C., 2005. Some Precambrian banded iron-formations (BIFs) from around the materials and chondrites. Geostandards and Geoanalytical Research. doi:10.1111/ world: their age, geologic setting, mineralogy, metamorphism, geochemistry, and j.1751-908X.2010.00085.x. origins. Am. Mineralog. 90, 1473–1499. Croal, L.R., Johnson, C.M., Beard, B.L., Newman, D.K., 2004. Iron isotope fractionation by Fe(II)- Klein, C., Beukes, N.J., 1989. Geochemistry and sedimentology of a facies transition from oxidizing photoautotrophic bacteria. Geochim. Cosmochim. Acta 68, 1227–1242. limestone to iron-formation deposition in the early Proterozoic Transvaal Crosby, H.A., Roden, E.E., Johnson, C.M., Beard, B.L., 2007. The mechanisms of iron Supergroup, South Africa. Econ. Geol. 84, 1733–1774. isotope fractionation produced during dissimilatory Fe(III) reduction by Shewanella Konhauser, K., Newman, D.K., Kappler, A., 2005. The potential significance of microbial putrefaciens and Geobacter sulfurreducens. Geobiology 5, 169–189. Fe(III) reduction during deposition of Precambrian banded iron formations. Dauphas, N., Craddock, P.R., Asimow, P.D., Bennett, V.C., Nutman, A.P., Ohnenstetter, D., Geobiology 3, 167–177. 2009a. Iron isotopes may reveal the redox conditions of mantle melting from Konhauser, K.O., Amskold, L., Lalonde, S.V., Posth, N.R., Kappler, A., Anbar, A.D., 2007. Archean to Present. Earth Planet. Sci. Lett. 288, 255–267. Decoupling photochemical Fe(II) oxidation from shallow-water BIF deposition. Dauphas, N., Janney, P.E., Mendybaev, R.A., Wadhwa, M., Richter, F.M., Davis, A.M., van Earth Planet. Sci. Lett. 258, 87–100. Zuilen, M., Hines, R., Foley, C.N., 2004a. Chromatographic separation and multi- Lamb, W., 2005. Carbonates in feldspathic gneisses from the granulite facies:

collection-ICPMS analysis of iron. Investigating mass-dependent and independent implications for the formation of CO2-rich fluid inclusions. In: Thomas, H. (Ed.), isotope effects. Anal. Chem. 76, 5855–5863. Metamorphism and Crustal Evolution. Atlantic Publishers and Distributors, New Dauphas, N., Pourmand, A., Teng, F.-Z., 2009b. Routine isotopic analysis of iron by HR- Delhi, pp. 163–181. MC-ICPMS: how precise and how accurate? Chem. Geol. 267, 175–184. Liu, S.V., Zhou, J., Zhang, C., Cole, D.R., Gajdarziska-Josifovska, M., Phelps, T.J., 1997. Dauphas, N., van Zuilen, M., Busigny, V., Lepland, A., Wadhwa, M., Janney, P.E., 2007. Thermophilic Fe(III)-reducing Bacteria from the deep subsurface: the evolutionary Iron isotope, major and trace element characterization of early Archean implications. Science 277, 1106–1109. supracrustal rocks from SW Greenland: protolith identification and metamorphic Lovley, D.R., 1991. Dissimilatory Fe(III) and Mn(IV) reduction. Microbiol. Rev. 55, 259–287. overprint. Geochim. Cosmochim. Acta 71, 4745–4770. Lovley, D.R., 1993. Dissimilatory metal reduction. Annu. Rev. Microbiol. 47, 263–290. Dauphas, N., van Zuilen, M., Wadhwa, M., Davis, A.M., Marty, B., Janney, P.E., 2004b. Lowe, D.R., Tice, M.M., 2004. Geologic evidence for Archean atmospheric and climatic

Clues from Fe isotope variations on the origin of early Archean BIFs from Greenland. evolution: fluctuating levels of CO2,CH4, and O2 with an overriding tectonic control. Science 306, 2077–2080. Geology 32, 493–496. Dimroth, E., Chauvel, J.J., 1973. Petrography of the Sokoman iron formation in part of MacLeod, W.N., 1966. The geology and iron deposits of the Hamersley Range area, the central Labrador trough, Quebec, Canada. Geol. Soc. Am. Bull. 84, 111–134. Western Autstralia: Geological Survey of Western Australia Bulletin, 117, p. 170. Dymek, R.F., Klein, C., 1988. Chemistry, petrology and origin of banded iron-formation Mojzsis, S.J., Arrhenius, G., McKeegan, K.D., Harrison, T.M., Nutman, A.P., Friend, C.R. lithologies from the 3800 Ma Isua Supracrustal Belt, West Greenland. Precambrian L., 1996. Evidence for life on Earth before 3, 800 million years ago. Nature 384, Res. 39, 247–302. 55–59. Ewers, W.E., Morris, R.C., 1981. Studies of the Dales Gorge Member of the Brockman Morris, R.C., 1993. Genetic modelling for banded iron-formation of the Hamersley Iron Formation, Western Australia. Econ. Geol. 76, 1929–1953. Group, Pilbara Craton, Western Australia. Precambrian Res. 60, 243–286. Farquhar, J., Bao, H., Thiemens, M., 2000. Atmospheric influence of Earth's earliest sulfur Myers, J.S., 2001. Protoliths of the 3.8–3.7 Ga Isua greenstone belt, West Greenland. cycle. Science 289, 756–758. Precambrian Res. 105, 129–141. Fischer, W.W., Schroeder, S., Lacassie, J.P., Beukes, N.J., Goldberg, T., Strauss, H., Nutman, A.P., Allaart, J.H., Bridgwater, D., Dimroth, E., Rosing, M., 1984. Stratigraphic Horstmann, U.E., Schrag, D.P., Knoll, A.H., 2009. Isotopic constraints on the Late and geochemical evidence for the depositional environment of the early Archaean Archean carbon cycle from the Transvaal Supergroup along the western margin of Isua Supracrustal Belt, southern West Greenland. Precambrian Res. 25, 365–396. the Kaapvaal Craton, South Africa. Precambrian Res. 169, 15–27. Nutman, A.P., Friend, C.R.L., 2009. New 1:20,000 scale geological maps, synthesis and Frei, R., Polat, A., 2007. Source heterogeneity for the major components of 3.7 Ga Banded history of investigation of the Isua supracrustal belt and adjacent orthogneisses, Iron Formations (Isua Greenstone Belt, Western Greenland): tracing the nature of southern West Greenland: a glimpse of Eoarchaean crust formation and orogeny. interacting water masses in BIF formation. Earth Planet. Sci. Lett. 253, 266–281. Precambrian Res. 172, 189–211. French, B.M., 1973. Mineral assemblages in diagenetic and low-grade metamorphic Nutman, A.P., McGregor, V.R., Friend, C.R.L., Bennett, V.C., Kinny, P.D., 1996. The Itsaq iron-formation. Econ. Geol. 68, 1063–1074. Gneiss Complex of southern West Greenland; the world's most extensive record of Froelich, P.N., Klinkhammer, G.P., Bender, M.L., Luedtke, N.A., Heath, G.R., Cullen, D., early crustal evolution (3900–3600 Ma). Precambrian Res. 78, 1–39. Dauphin, P., 1979. Early oxidation of organic matter in pelagic sediments of the eastern Oehler, D.Z., Smith, J.W., 1977. Isotopic composition of reduced and oxidized carbon in equatorial Atlantic: suboxic diagenesis. Geochim. Cosmochim. Acta 43, 1075–1090. early Archaean rocks from Isua, Greenland. Precambrian Res. 5, 221–228. 132 P.R. Craddock, N. Dauphas / Earth and Planetary Science Letters 303 (2011) 121–132

Ono, S., Eigenbrode, J.L., Pavlov, A.A., Kharecha, P., Rumble III, D., Kasting, J.F., Steinhoefel, G., von Blanckenburg, F., Horn, I., Konhauser, K.O., Beukes, N.J., Gutzmer, J., Freeman, K.H., 2003. New insights into Archean sulfur cycle from mass- 2010. Deciphering formation processes of banded iron formations from the Transvaal independent sulfur isotope records from the Hamersley Basin, Australia. Earth and the Hamersley successions by combined Si and Fe isotope analysis using UV Planet.Sci.Lett.213,15–30. femtosecond laser ablation. Geochim. Cosmochim. Acta 74, 2677–2696. Pavlov, A.A., Kasting, J.F., 2002. Mass-independent fractionation of sulfur isotopes in Swart, P.K., Burns, S.J., Leder, J.J., 1991. Fractionation of the stable isotopes of oxygen and Archean sediments: strong evidence for an anoxic Archean atmosphere. Astrobiology carbon in carbon dioxide during the reaction of calcite with phosphoric acid as a 2, 27–41. function of temperature and technique. Chem. Geol. 86, 89–96. Pecoits, E., Gingras, M.K., Barley, M.E., Kappler, A., Posth, N.R., Konhauser, K., 2009. Tangalos, G.E., Beard, B.L., Johnson, C.M., Alpers, C.N., Shelobolina, E.S., Xu, H., Konishi, Petrography and geochemistry of the Dales Gorge banded iron formation: paragenetic H., Roden, E.E., 2010. Microbial production of isotopically light iron(II) in a modern sequence, source and implications for palaeo-ocean chemistry. Precambrian Res. 172, chemically precipitated sediment and implications for isotopic variations in ancient 163–187. rocks. Geobiology 8, 197–208. Perry Jr., E.C., Ahmad, S.N., 1977. Carbon isotope composition of graphite and carbonate Taylor, P.D.P., Maeck, R., De Bievre, P., 1992. Determination of the absolute isotopic minerals from 3.8-AE metamorphosed sediments, Isukasia, Greenland. Earth Planet. composition and atomic weight of a reference sample of natural iron. Int. J. Mass Sci. Lett. 36, 280–284. Spectrom. Ion Processes 121, 111–125. Perry Jr., E.C., Tan, F.C., Morey, G.B., 1973. Geology and stable isotope geochemistry of Trendall, A.F., 2002. The significance of iron-formation in the Precambian stratigraphic the Biwabik Iron Formation, Northern Minnesota. Econ. Geol. 68, 1110–1125. record. Spec. Publ. Int. Assoc. Sedimentol. 33, 33–66. Polyakov, V.B., Clayton, R.N., Horita, J., Mineev, S.D., 2007. Equilibrium iron isotope Trendall, A.F., Blockley, J.G., 1970. The iron formations of the Precambrian Hamersley fractionation factors of minerals: reevaluation from the data of nuclear inelastic Group, Western Australia: Geological Survey of Western Australia Bulletin, 119, p. 370. resonant X-ray scattering and Mössbauer spectroscopy. Geochim. Cosmochim. Acta Ueno, Y., Yurimoto, H., Yoshioka, H., Komiya, T., Maruyama, S., 2002. Ion microprobe 71, 3833–3846. analysis of graphite from ca. 3.8 Ga metasediments, Isua supracrustal belt, West Rose, N.M., Rosing, M.T., Bridgwater, D., 1996. The origin of metacarbonate rocks in the Greenland: relationship between metamorphism and carbon isotopic composition. Archaean Isua supracrustal belt, West Greenland. Am. J. Sci. 296, 1004–1044. Geochim. Cosmochim. Acta 66, 1257–1268. Rosing, M.T., 1999. 13C-depleted carbon microparticles in N3700 Ma sea-floor van Zuilen, M.A., Lepland, A., Arrhenius, G., 2002. Reassessing the evidence for the sedimentary rocks from West Greenland. Science 283, 674–676. earliest traces of life. Nature 418, 627–630. Rosing, M.T., Bird, D.K., Sleep, N.H., Bjerrum, C.J., 2010. No climate paradox under the van Zuilen, M.A., Lepland, A., Teranes, J., Finarelli, J., Wahlen, M., Arrhenius, G., 2003. faint early Sun. Nature 464, 744–747. Graphite and carbonates in the 3.8 Ga old Isua supracrustal belt, southern West Rosing, M.T., Rose, N.M., Bridgwater, D., Thomsen, H.S., 1996. Earliest part of Earth's Greenland. Precambrian Res. 126, 331–348. stratigraphic record: a reappraisal of the N3.7 Ga Isua (Greenland) supracrustal Vargas, M., Kashefi, K., Blunt-Harris, E.L., Lovley, D.R., 1998. Microbiological evidence for sequence. Geology 24, 43–46. Fe(III) reduction on early Earth. Nature 395, 65–67. Rouxel, O.J., Bekker, A., Edwards, K.J., 2005. Iron Isotope Constraints on the Archean and Veizer, J., Clayton, R.N., Hinton, R.W., 1992. Geochemistry of Precambrian carbonates. Paleoproterozoic Ocean Redox State. Science 307, 1088–1091. IV. Early Paleoproterozoic (2.25±0.25 Ga) seawater. Geochim. Cosmochim. Acta Rouxel, O.J., Dobbek, N., Ludden, J., Fouquet, Y., 2003. Iron isotope fractionation during 56, 875–885. oceanic crust alteration. Chem. Geol. 202, 155–182. Veizer, J., Hoefs, J., Lowe, D.R., Thurston, P.C., 1989. Geochemistry of Precambrian carbonates: Rye, R., Kuo, P.H., Holland, H.D., 1995. Atmospheric carbon dioxide concentrations II. Archean greenstone belts and Archean sea water. Geochim. Cosmochim. Acta 53, before 2.2 billion years ago. Nature 378, 603–605. 859–871. Schauble, E.A., Rossman, G.R., Taylor, H.P., 2001. Theoretical estimates of equilibrium Fe- Waldbauer, J.R., Sherman, L.S., Sumner, D.Y., Summons, R.E., 2009. Late Archean isotope fractionations from vibrational spectroscopy. Geochim. Cosmochim. Acta 65, molecular fossils from the Transvaal Supergroup record the antiquity of microbial 2487–2497. diversity and aerobiosis. Precambrian Res. 169, 28–47. Schidlowski, M., Appel, P.W.U., Eichmann, R., Junge, C.E., 1979. Carbon isotope Walker, J.C.G., 1983. Possible limits on the composition of the Archaean ocean. Nature geochemistry of the 3.7 x 109 yr-old Isua sediments, West Greenland: implications 302, 518–520. for the Archaean carbon and oxygen cycles. Geochim. Cosmochim. Acta 43, 189–199. Walker, J.C.G., 1984. Suboxic diagenesis in banded iron formations. Nature 309, 340–342. Schidlowski, M., Eichmann, R., Junge, C.E., 1975. Precambrian sedimentary carbonates: Walker, J.C.G., 1985. Carbon dioxide in the early Earth. Orig. Life 16, 117–127. carbon and oxygen isotope geochemistry and implications for the terrestrial Weber, K.A., Achenbach, L.A., Coates, J.D., 2006. Microorganisms pumping iron: oxygen budget. Precambrian Res. 2, 1–69. anaerobic microbial iron oxidation and reduction. Nat. Rev. Microbiol. 4, 752–764. Severmann, S., Johnson, C.M., Beard, B.L., German, C.R., Edmonds, H.N., Chiba, H., Green, Whitehouse, M.J., Fedo, C.M., 2007. Microscale heterogeneity of Fe isotopes in N3.71 Ga D.R.H., 2004. The effect of plume processes on the Fe isotope composition of banded iron formation from the Isua Greenstone Belt, southwest Greenland. Geology hydrothermally derived Fe in the deep ocean as inferred from the Rainbow vent 35, 719–722. site, Mid-Atlantic Ridge, 36°14′N. Earth Planet. Sci. Lett. 225, 63–76. Widdel, F., Schnell, S., Heising, S., Ehrenreich, A., Assmus, B., Schink, B., 1993. Ferrous Sharma, M., Polizzotto, M., Anbar, A.D., 2001. Iron isotopes in hot springs along the Juan iron oxidation by anoxygenic phototrophic bacteria. Nature 362, 834–836. de Fuca Ridge. Earth Planet. Sci. Lett. 194, 39–51. Wiesli, R.A., Beard, B.L., Johnson, C.M., 2004. Experimental determination of Fe isotope Shields, G., Veizer, J., 2002. Precambrian marine carbonate isotope database: Version fractionation between aqueous Fe (II), siderite and “green rust” in abiotic systems. 1.1. Geochem. Geophys. Geosyst. 3. doi:10.1029/2001GC000266. Chem. Geol. 211, 343–362.