Chemical Geology 218 (2005) 135–169 www.elsevier.com/locate/chemgeo Biogeochemical cycling of iron in the Archean–Paleoproterozoic Earth: Constraints from iron isotope variations in sedimentary rocks from the Kaapvaal and Pilbara Cratons Kosei E. Yamaguchia,b,c, Clark M. Johnsonb,c,T, Brian L. Beardb,c, Hiroshi Ohmotoc,d aInstitute for Research on Earth Evolution (IFREE), Japan Agency for Marine-Earth Science and Technology (JAMSTEC), 2-15 Natsushima, Yokosuka, 237-0061, Japan bDepartment of Geology and Geophysics, University of Wisconsin - Madison, 1215 W. Dayton St., Madison, WI 53706, USA cNASA Astrobiology Institute, United States dAstrobiology Research Center and Department of Geosciences, The Pennsylvania State University, 435 Deike Building, University Park, PA 16802, USA Received 4 April 2004; received in revised form 21 October 2004; accepted 26 January 2005 Abstract Iron isotope compositions of low-metamorphic grade samples of Archean–Paleoproterozoic sedimentary rocks obtained from fresh drill core from the Kaapvaal Craton in South Africa and from the Pilbara Craton in Australia vary by ~3x in 56Fe/54Fe ratios, reflecting a variety of weathering and diagenetic processes. Depositional ages for the 120 samples studied range from 3.3 to 2.2 Ga, and Fe, C, and S contents define several compositional groups, including samples rich in Fe, organic carbon, carbonate, and sulfide. 56 The d Fe values for low-Corg, low-Ccarb, and low-S sedimentary rocks are close to 0x, the average of igneous rocks. This range is essentially the same as that of Corg-poor late Cenozoic loess, aerosol, river loads, and marine sediments and those of 56 Corg-poor Phanerozoic–Proterozoic shales. That these d Fe values are the same as those of igneous rocks suggests that Fe has behaved conservatively in bulk sediments during sedimentary transport, diagenesis, and lithification since the Archean. These observations indicate that, if atmospheric O2 contents rose dramatically between 2.4 and 2.2 Ga, as proposed by many workers, such a rise did not produce a significant change in the bulk Fe budget of the terrestrial sedimentary system. If the Archean atmosphere was anoxic and Fe was lost from bedrock during soil formation, any isotopic fractionation between aqueous ferrous 2+ Fe (Feaq ) and Fe-bearing minerals must have been negligible. In contrast, if the Archean atmosphere was oxic, Fe would have been retained as Fe3+ hydroxides during weathering as it is today, which would produce minimal net isotopic fractionation in bulk detrital sediments. 56 2+ Siderite-rich samples have d Fe values of À0.5F0.5x, and experimentally determined Feaq -siderite fractionation factors 2+ 56 56 2+ suggest that these rocks formed from Feaq that had similar or slightly higher d Fe values. The d Fe values calculated for Feaq T Corresponding author. Department of Geology and Geophysics, University of Wisconsin - Madison, 1215 W. Dayton St., Madison, WI 53706, USA. Tel.: +1 608 262 1710, +1 608 265 6798; fax: +1 608 262 0693. E-mail address: [email protected] (C.M. Johnson). 0009-2541/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.chemgeo.2005.01.020 136 K.E. Yamaguchi et al. / Chemical Geology 218 (2005) 135–169 2+ 56 overlaps those of modern submarine hydrothermal fluids, but it is also possible that Feaq had d Fe values higher than those of 2+ modern hydrothermal fluids, depending upon the Feaq –Fe carbonate fractionation factor that is used. In contrast, Corg-rich samples and magnetite-rich samples have strongly negative d56Fe values, generally between À2.3x and À1.0x, and available fluid–mineral fractionation factors suggest that the Fe-bearing minerals siderite and magnetite in these rocks formed in the 2+ 56 3+ presence of Feaq that had very low d Fe values, between À3x and À1x. Reduction of Fe hydroxide by sulfide, precipitation of sulfide minerals, or incongruent dissolution of silicate minerals are considered unlikely means to produce 56 2+ 3+ significant quantities of low-d Fe Feaq . We interpret microbial dissimilatory Fe reduction (DIR) as the best explanation for 56 2+ producing such low d Fe values for Feaq , and our results suggest that DIR was a significant form of respiration since at least 2.9 Ga. D 2005 Elsevier B.V. All rights reserved. Keywords: Iron; Isotope; Archean; Proterozoic; Biology 1. Introduction by a variety of inorganic and organic redox reactions, many of which produce significant isotopic fractiona- Evolution of the Archean biosphere and its tions. In the case of Fe, biological and nonbiological influence on redox conditions remains a first-order isotopic fractionations are well demonstrated to occur problem in studies on the evolution of the Earth. in nature and in experiments (e.g., Beard and Johnson, Evidence from molecular phylogeny suggests that 1999, 2004a,b; Beard et al., 1999, 2003a,b; Anbar, photosynthesis evolved quite early in Earth’s history 2004; Anbar et al., 2000; Zhu et al., 2000, 2001, 2002; (Xiong et al., 2000). Biomarker evidence also suggests Brantley et al., 2001, 2004; Bullen et al., 2001; that organisms related to cyanobacteria (oxygenic Matthews et al., 2001, 2004; Sharma et al., 2001; photosynthesizers) evolved by at least 2.7 Ga (Brocks Johnson et al., 2002, 2003, 2004, 2005; Severmann et et al., 1999, 2003). It has also been proposed, on the al., 2004a; Levasseur et al., 2004; Skulan et al., 2002; basis of molecular phylogeny, that dissimilatory Fe3+ Welch et al., 2003; Kehm et al., 2003; Roe et al., 2003; reduction (DIR) may have been one of the earliest Rouxel et al., 2003, 2004; Icopini et al., 2004). forms of respiration on Earth (Vargas et al., 1998), Because some of the largest isotopic fractionations preceding other important respiratory processes, such occur during redox changes (Polyakov and Mineev, as sulfate reduction, nitrate reduction, and oxygen 2000; Schauble et al., 2001; Johnson et al., 2002; reduction (Lovley, 1993). Although such a view is Welch et al., 2003; Anbar et al., 2005), Fe isotopes consistent with known phylogenies of extant organ- hold particular promise in tracing biologically Fe2+ isms, it is impossible to validate these hypotheses oxidation and Fe3+ reduction (Johnson et al., 2004). without corroborating geochemical evidence that Despite great interests in Fe isotope geochemistry, little provides a timeline for the process (Benner et al., work has been done on Archean sedimentary rocks. 2002). The possibility that DIR occurred early in In this contribution, we examine the evidence for Earth’s history is particularly important because of its redox cycling of Fe in the Archean. The dataset implications for the chemical and redox evolution of include low organic-C (Corg) clastic rocks that likely the early Earth. If little Fe3+ existed on the early Earth, reflect terrestrial weathering processes, as well as an Fe3+-reductase would not have functioned because rocks that probably formed in relatively anoxic marine there would be no advantage in assimilatory and environments. A major theme we develop centers on dissimilatory metabolism of Fe (or other metals) to the evidence that dissimilatory Fe3+ reduction (DIR) sustain life (Nealson and Saffarini, 1994). may have been an active metabolic pathway by at Specific metabolisms may produce unique isotopic least 2.9 Ga. One hundred and twenty samples of fractionations of essential elements such as C, H, N, carbonaceous shales, non-carbonaceous shales, mag- and S, which may be detected in the rock record (e.g., netite-rich shales, siderite-rich shales, and greywackes Hayes, 2001; Canfield, 2001). The biogeochemical were selected for this study using drill cores from 10 cycles of Fe, like those of C, H, N, and S, are mediated drill holes representing 14 geologic formations recov- K.E. Yamaguchi et al. / Chemical Geology 218 (2005) 135–169 137 ered in South Africa and Australia. They are 3.25 to significant metamorphism, generally less than greens- 2.20 Ga in age, have undergone minimal metamor- chist facies. South African samples are from the ~3.3 phism, and are free from the effects of modern Ga Fig Tree Group of the Swaziland Supergroup, the 3+ 2+ weathering. We report data for Fe ,Fe ,Corg, and ~3.0 Ga West Rand Group of the Witwatersrand Ccarb (carbonate carbon), and S contents and Fe Supergroup, the ~2.7 Ga Platberg Group of the isotope compositions. Ventersdorp Supergroup, the ~2.6 Ga Wolkberg Group, the ~2.6 Ga Chuniespoort Group, and the ~2.2 Ga Pretoria Group of the Transvaal Supergroup 2. Geological settings and samples (Fig. 3). Australian samples are from the ~2.7 Ga Fortescue Group and the ~2.6 Ga Hamersley Group Archean and Paleoproterozoic strata are exception- of the Mt. Bruce Supergroup (Fig. 4). ally well preserved and exposed on the Kaapvaal Craton of southern Africa (Fig. 1) and on the Pilbara 2.1. Swaziland Supergroup Craton in the Pilbara–Hamersley regions of NW Australia (Fig. 2). Sedimentary successions uncom- The Swaziland Supergroup is exposed in the 3.4 to formably overlie the cratonic basements. We have 3.1 Ga Barberton Greenstone Belt located in the analyzed 103 laminated and nonlaminated black and eastern part of the Kaapvaal Craton (Fig. 1). The greenish shale samples, 9 greywacke samples, and 8 Swaziland Supergroup consists of three groups: from laminated red shale samples (in 14 formations) from lower to upper, the 3.48–3.45 Ga Onverwacht Group, drill cores, which are free of weathering products (120 3.33–3.23 Ga Fig Tree Group, and 3.22–3.10 Ga samples in total). The rocks have not been subject to Moodies Group (Armstrong et al., 1990; Kro¨ner et al., Fig. 1. Simplified geological map of South Africa showing the distribution of the Transvaal, Ventersdorp, and the Witwatersrand Supergroups, the Sabie-Pilgrim’s Rest region, the Griqualand West region, the Witwatersrand Basin, and the Barberton Greenstone Belt, which includes the Swaziland Supergroup.