Search

Iron mobilization during continental weathering was pervasive before the Great Oxidation Event (GOE) that started at around 2.43 billion years (Ga) ago, due to the soluble nature of reduced iron. However, various geochemical proxies indicate transient oxygenation during deposition of the Mesoarchean (∼2.95 Ga) Pongola Supergroup, South Africa, which suggests that continental weathering could have also occurred under transiently oxic conditions before the GOE. We analyzed trace elemental and Fe, Ti, and Cr isotopic compositions of the ca. 2.95 Ga Denny Dalton paleosol in the Pongola Supergroup to better understand continental weathering and redox conditions in the ancient critical zone, and the nature of geochemical fluxes from the continents to the oceans and marine sediments. Iron isotope systematics are consistent with a model where Fe was released during intense leaching from the paleosol to concentrate in Fe-rich groundwaters in the deeper part of the soil horizon. We show for the first time that Fe isotopic fractionation during Mesoarchean continental weathering was limited, and Fe enrichments and depletions are coupled with those of divalent transition metals, Co, Ni, and particularly Zn. This suggests that Fe redox cycling was not involved in paleosol formation, and Fe2+ was mobilized under anoxic conditions. Chromium isotopes are also unfractionated relative to the parent igneous rock in this paleosol, which precludes removal of isotopically heavy Cr6+ and thus supports anoxic continental weathering. We show that previously reported Cr isotopic fractionation in another Denny Dalton paleosol profile does not follow a Cr6+ leaching trend, but instead scales with Cr enrichment and may reveal Cr enrichment from post-burial fluids. Thus, there is no clear evidence for an oxidative continental weathering during deposition of the Pongola Supergroup. Titanium isotopes are not significantly fractionated in the paleosol, suggesting that continental weathering and erosion in the Archean did not fractionate Ti isotopes. Similarly, Ni/Co and Th/Sc ratios are reasonably conserved, which validates their use as a robust proxy for upper continental crust composition in shales, whereas La/Sc, Cr/Zn and Cr/U ratios are highly variable relative to the provenance composition, which suggests that caution should be used when applying these ratios in shale studies of the ancient upper continental crust composition.

The Archaean continental crust comprises two major groups of silicon-rich granitoids: the tonalite–trondhjemite–granodiorite and granite–monzonite–syenite suites, which differ in their sodium-to-potassium ratios. How these felsic granitoids evolved from their mafic precursors remains elusive and the subject of great debate. Here, we present silicon isotopic constraints on the formation of representative trondhjemitic and granitic plutons from the Kaapvaal craton that range in age from 3.51–2.69 billion years ago. We identified very consistent silicon isotopic signatures, all uniformly 0.1–0.2‰ heavier than rocks of the modern continental crust. This unusual composition is explained by the melting of a mafic source that included significant proportions (15–35 wt%) of silicified basalts, which were common supracrustal rocks before 3 billion years ago. Before the melting event that formed the granitoid magmas at depth, portions of the mafic source rocks were enriched in silica by interaction with silica-saturated seawater. The addition of silica depresses the stability of amphibole at similar water activity, allowing trondhjemitic and granitic melt production at lower temperatures from protoliths with contrasting silica contents: 52–57 and ≥60 wt%, respectively. This explains why granitoids were able to form very early in Earth history but did not emerge in significant amounts on other rocky planets. Granitic continental crust in the Archaean formed from a basaltic source that was enriched in silica due to interaction with the early oceans before melting, according to silicon isotope analyses on rocks from the Kaapvaal craton.

The question of whether high-grade metamorphism and crustal melting in the early Archaean were associated with modern-style plate tectonics is a major issue in unravelling early Earth crustal evolution, and the eastern Kaapvaal craton has featured prominently in this debate. We discuss a major ca. 3.2 Ga tectono-magmatic-metamorphic event in the Ancient Gneiss Complex (AGC) of Swaziland, a multiply deformed medium- to high-grade terrane in the eastern Kaapvaal craton consisting of 3.66–3.20 Ga granitoid gneisses and infolded greenstone remnants, metasedimentary assemblages and mafic dykes. We report on a 3.2 Ga granulite-facies assemblage in a metagabbro of the AGC of central Swaziland and relate this to a major thermo-magmatic event that not only affected the AGC but also the neighbouring Barberton granitoid-greenstone terrane. Some previous models have related the 3.2 Ga event in the eastern Kaapvaal craton to subduction processes, but we see no evidence for long, narrow belts and metamorphic facies changes reflecting lithospheric suture zones, and there is no unidirectional asymmetry in the thermal structure across the entire region from Swaziland to the southern Barberton granite-greenstone terrane as is typical of Phanerozoic and Proterozoic belts. Instead, we consider an underplating event at ca. 3.2 Ga, giving rise to melting in the lower crust and mixing with mantle-derived under- and intraplated mafic magma to generate the voluminous granitoid assemblages now observed in the AGC and the southern Barberton terrane. This is compatible with large-scale crustal reworking during a major thermo-magmatic event and the apparent lack of a mafic lower crust in the Kaapvaal craton as shown by seismic data.

In the Kaapvaal craton of southern Africa, as well as other Archaean cratons worldwide, the progression from dominant tonalite–trondhjemite–granodiorite (TTG) to granite-monzogranite-syenogranite (GMS) rock types is interpreted to reflect progressive reworking and differentiation of the continental crust. Here we re-evaluate the early Archaean evolution of the Kaapvaal craton and propose a unified view of the plutonic and volcanic records based on elemental and isotopic (Nd, Hf) data and zircon U–Pb ages. We also report new whole-rock major and trace element analyses, zircon U–Pb ages and Hf-in-zircon analyses of igneous clasts from a conglomerate of the 3.2 Ga Moodies Group of the Barberton Greenstone Belt. Many of these clasts are derived from shallow intrusive rocks of granitic composition, which are scarcely represented in outcrop. Despite alteration, the volcanic rocks can be classified based on their trace element contents into two main groups by comparison with plutonic rocks. One group has characteristics resembling TTGs: relatively low and fractionated rare earth element concentrations with no Eu anomaly and relatively low concentrations of high field strength elements (Nb mostly ≤12 ppm). The second group has GMS-like characteristics: less fractionated REE, marked negative Eu anomalies and HFSE-increasing trends with progressing fractionation (Nb ≤ 50 ppm or more, Th up to 30–40 ppm). In addition, igneous clasts of Moodies Group conglomerate have chemical, mineralogical and isotopic characteristics that link them to GMS. New analyses of some of these clasts indicate elevated high field strength elements (Nb up to 20 ppm) and εHf(t) of zircon down to −3.5. These rocks imply the presence of an already differentiated felsic crust at >3.5 Ga, which has Nd and Hf model ages indicating mantle extraction ages extending back to the Eoarchaean. The combined record of plutonic and volcanic rocks of the Kaapvaal craton provides a more complex scenario than previously suggested and indicates that TTG and GMS-like felsic magmas were emplaced broadly coevally in multiple pulses between ∼3.5 Ga and 3.2 Ga. Keywords: Archean, Kaapvaal craton, TTG, Greenstone, Barberton

The Archean spherule layers (SLs) of the Barberton Greenstone Belt (BGB, South Africa) and Pilbara Craton (Australia) are the only known evidence of early, large impact events on Earth. Spherules in these layers have been, alternatively, interpreted as molten impact ejecta, condensation products from an impact vapor cloud, or ejecta from impact craters melted during atmospheric re-entry. Recently, a new exploration drill core (CT3) from the northern BGB revealed 17 SL intersections. Spherules are densely packed, sand-sized, and variably rounded or deformed. The CT3 SLs are intercalated with black and brown shale, and laminated chert. The determination of the original number of impact events that are represented by these multiple SLs is central to the present paper. A comprehensive study of the sedimentary and petrographic characteristics of these SLs involved the determination of the size, shape and types of individual spherules, as well as their mineralogy. CT3 SLs consist of K-feldspar, phyllosilicate, siderite, dolomite, quartz, Ti- and Fe-oxides, as well as apatite. In addition, small amounts of carbonaceous, presumably organic material are observed in several spherules at 145 and 149 m depth. Only Ni-rich Cr-spinel (up to 11 wt% NiO) crystals, rare zircon grains, and alloys of platinum group elements ± Fe or Ni represent primary phases in these thoroughly altered strata. The 0.3 to 2.6-mm-sized spherules can be classified into four types: 1. Spherules crystallized completely with secondary K-feldspar (subtype 1A) or phyllosilicate (subtype 1B); spherules completely filled with Ti- and Fe-oxides (subtype 1C); spherules containing disordered or radially oriented, fibrous and lath-shaped K-feldspar textures (subtype 1D); or subtype 1B spherules that contain significant Cr-spinel (subtype 1E); 2. zoned compositions with these types 1A and/or 1B minerals (subtype 2A); spherules that contain central or marginal vesicles (subtype 2B); subtype 1B spherules whose rims consist of Ti and Fe-oxides (subtype 2C); 3. deformed spherules (subtype 3A) - of all types; (B) subtype 1B spherules are assimilated into groundmass; (C) open spherules or spherules with collapsed rims; and 4. interconnected spherules of type 1A. A few spherules show botryoidal devitrification textures interpreted to result from rapid cooling/devitrification of former melt droplets. SL 15 at a depth of 145 m is unique in being the only grain-size sorted SL; this bed may have been deposited by fallout through a water column. The SL and their host rocks can be easily distinguished by their significant differences in micro-XRF elemental distribution maps. Depending on which aspects of the SLs are primarily considered (such as similar geochemistry, similar layering, SL occurrences abundant at three different depth intervals), the 17 CT3 SLs can be assigned to three or up to 13 individual impact events. Uncertainty about the actual number of impact events represented remains, however, due to the complex folding deformation observed throughout the drill core.

Mass-independent fractionation of sulfur isotopes (MIF-S) in Archaean sediments results from photochemical processing of atmospheric sulfur species in an oxygen-depleted atmosphere. Geological preservation of MIF-S provides evidence for microbial sulfate reduction (MSR) in low-sulfate Paleoarchaean (3.8–3.2 billion years ago (Ga)) and Neoarchaean (2.8–2.5 Ga) oceans, but the significance of MSR in Mesoarchaean (3.2–2.8 Ga) oceans is less clear. Here we present multiple sulfur and iron isotope data of early diagenetic pyrites from 2.97-Gyr-old stromatolitic dolomites deposited in a tidal flat environment of the Nsuze Group, Pongola Supergroup, South Africa. We identified consistently negative Δ33S values in pyrite, which indicates photochemical reactions under anoxic atmospheric conditions, but large mass-dependent sulfur isotope fractionations of ~30‰ in δ34S, identifying active MSR. Negative pyrite δ56Fe values (−1.31 to −0.88‰) record Fe oxidation in oxygen-bearing shallow oceans coupled with biogenic Fe reduction during diagenesis, consistent with the onset of local Fe cycling in oxygen oases ~3.0 Ga. We therefore suggest the presence of oxygenated near-shore shallow-marine environments with ≥5 μM sulfate at this time, in spite of the clear presence of an overall reduced Mesoarchaean atmosphere.

The only known case of natural fission reactors is hosted by high-grade uranium (U) deposits at Oklo-Okelobondo and Bangombe in sandstones of the ~2.1 Ga Francevillian Group, Gabon. However, the geochemical influence of the depositional environment on this unique natural nuclear phenomenon has not been clearly established. Localized, unusually high vanadium (V) enrichments are thought to have prevented such natural nuclear fission reactors from occurring in other Francevillian U deposits (e.g., Mounana, Boyindzi, and Mikouloungou). However, while U-bearing detrital monazite derived from Archean rocks surrounding the Francevillian basin is viewed as the main source of U, the source of V remains poorly constrained. Here, we combine petrographic and whole-rock geochemical data for the Francevillian Group sedimentary rocks, coupled with previously documented geochemical data for the Archean basement. These data suggest that, although ultramafic to mafic igneous rocks of the Mesoarchean Belinga Group, and to some extent Archean granitoids, were likely important sources of V to the Francevillian U deposits, they were not the only source of the abnormally high V concentrations in the U deposits that did not produce natural nuclear reactors. Instead, hydrocarbon migration from V-rich black shales of the Upper Francevillian Group, deposited during widespread and protracted deoxygenation of the Paleoproterozoic ocean at the end of the Lomagundi Carbon Isotope Excursion (LE) at ~2.11–2.06 Ga, resulted in a redox front that precipitated the U deposits. These migrated V-rich hydrocarbons likely account for the high V concentrations, which ultimately prevented natural fission reactions from occurring in these U deposits. Similarities with other pyrobitumen-bearing Paleoproterozoic U deposits worldwide suggest that organic-rich source rocks, which deposited in open-marine settings under widespread hyper-euxinic conditions in the aftermath of the LE, played a key role in preventing the Oklo natural nuclear reactor phenomenon from reaching a larger extent.

Hafnium and oxygen isotopic compositions measured in-situ on U-Pb dated zircon from Archaean sedimentary successions belonging to the 2.9-2.8 Ga Belingwean/Bulawayan groups and previously undated Sebakwian Group are used to characterize the crustal evolution of the Zimbabwe Craton prior to 3.0 Ga. Microstructural and compositional criteria were used to minimize effects arising from Pb loss due to metamorphic overprinting and interaction with low-temperature fluids. 207 Pb/206 Pb age spectra (concordance >90%) reveal prominent peaks at 3.8, 3.6, 3.5, and 3.35 Ga, corresponding to documented geological events, both globally and within the Zimbabwe Craton. Zircon delta O-18 values from +4 to +10% point to both derivation from magmas in equilibrium with mantle oxygen and the incorporation of material that had previously interacted with water in near-surface environments. In epsilon(Hf)-time space, 3.8-3.6 Ga grains define an array consistent with reworking of a mafic reservoir ((176) Lu/(177) Hf similar to 0.015) that separated from chondritic mantle at similar to 3.9 Ga. Crustal domains formed after 3.6 Ga depict a more complex evolution, involving contribution from chondritic mantle sources and, to a lesser extent, reworking of pre-existing crust. Protracted remelting was not accompanied by significant mantle depletion prior to 3.35 Ga. This implies that early crust production in the Zimbabwe Craton did not cause complementary enriched and depleted reservoirs that were tapped by later magmas, possibly because the volume of crust extracted and stabilised was too small to influence (asthenospheric) mantle isotopic evolution. Growth of continental crust through pulsed emplacement of juvenile (chondritic mantle-derived) melts, into and onto the existing cratonic nucleus, however, involved formation of complementary depleted subcontinental lithospheric mantle since the early Archaean, indicative of strongly coupled evolutionary histories of both reservoirs, with limited evidence for recycling and lateral accretion of arc-related crustal blocks until 3.35 Ga. (C) 2017 Elsevier Ltd. All rights reserved.

Seafloor alteration provides large amounts of solutes to the hydrosphere. In order to investigate gold mobility during water-rock interaction prior to 3-billion-years ago, low detection limit analysis of Au concentrations was carried out on rocks from marine alteration zones. Stratiform zones recording low-temperature (≤150 °C) seafloor alteration are a characteristic feature of greenstone belts older than 3.0 Ga. Hydrothermal processes were operating on, and immediately below, the seafloor, giving rise to extensive silicification of sub-seafloor volcanic rocks and silicification of seafloor sediments. In order to investigate gold mobility during silicification, unaltered and variably silicified volcanic rocks and associated cherts from Palaeoarchaean greenstone successions (c. 3.4 Ga) of South Africa were analyzed. Results show mobility of gold during silicification of mafic/ultramafic rocks and transfer to the Archaean ocean. Some gold was incorporated into carbonaceous marine sediments overlying the alteration zones. A combination of pervasive silicification, rarity of black shales, and low gold content in komatiites can explain the low mineralization potential of Palaeoarchaean greenstone belts for orogenic gold deposits.

The Barberton Greenstone Belt hosts abundant structurally controlled gold mineralisation of Mesoarchaean age. More than 300 gold occurrences have been reported, although most of the gold production so far (>350 tonnes Au) has come from a handful of deposits located along the northern margin of the greenstone belt. Most deposits are hosted by greenschist-facies metasedimentary and metamafic rocks, with the notable exception of the amphibolite-facies rocks at New Consort mine. Mineralisation is associated with quartz–carbonate veins that truncate major compressional structures at the greenstone belt scale. The age of mineralisation is loosely constrained at circa 3080–3030 Ma, based on U–Pb dating of hydrothermal rutile and titanite. In greenschist-facies deposits, the ore assemblage is dominated by pyrite and arsenopyrite, which contain up to thousands of ppm of ‘invisible’ gold, Ni–As–Sb sulphides and native gold. At New Consort mine, mineralisation includes massive replacement-style ore and vein-hosted or disseminated types. Both structural studies in the field and microstructural observation point to a multistage ore deposition process, which is reflected in the re-activation of brittle to ductile structures and the overprinting of sulphide assemblages. The presence of mass-independently fractionated S isotopes (Δ33S = –0.6 to +1.0‰) in pyrite from Sheba and Fairview mines suggests that hydrothermal fluids mobilised S from volcanic and sedimentary rocks of the greenstone belt and places constraints on the origin of the Au itself.

The southeastern Kaapvaal Craton is a Palaeoarchaean granitoid–greenstone terrain. Supracrustal rocks are dominated by metamorphosed mafic–ultramafic volcanic rocks intercalated with minor felsic volcanic and chemical sedimentary rocks, including carbonaceous chert and minor iron formation. Siliciclastic sedimentary rocks are rare. The greenstones occur in the Schapenburg and Dwalile fragments close to the Barberton greenstone belt, the Assegaai, De Kraalen, Witrivier and Commondale fragments in the vicinity of Piet Retief, and the Nondweni and Ilangwe greenstone belts together with several smaller fragments in the southern part of the craton. The greenstones are locally in tectonic contact with compositionally layered grey gneisses as old as ~3.5 Ga and, in the southern part, are intruded by the ~3.25 Ga Anhalt granitoid suite, providing a minimum age for the greenstones. The oldest felsic volcanic rocks so far identified are 3.53 Ga in age. The belts show evidence of polyphase deformation, with metamorphism at greenschist- to amphibolite-facies grade, the peak of which was at ~3.2 Ga. Ore deposits are restricted to sub-economic gold and rare massive sulphide Cu–Zn deposits.

This study was funded by the University of Johannesburg; the National Research Foundation of South Africa [Department of Science and Technology Innovation Research Fellowship (F.O.O.), Grant 75892 (to A.H.), and the Centre of Excellence for Integrated Mineral and Energy Resource Analysis hosted by the University of Johannesburg]; the German Research Foundation Grant SCHO1071/7-1 under the DFG Priority Programme SPP 1833 “Building a Habitable Earth” (to R.S.); and the University of Lausanne. S.W.P. acknowledges support from a Royal Society Wolfson Research Merit Award and a Leverhulme Research Fellowship. A.B. acknowledges support from the Natural Sciences and Engineering Research Council of Canada Discovery and Accelerator grants. The Archean Eon was a time of predominantly anoxic Earth surface conditions, where anaerobic processes controlled bioessential element cycles. In contrast to “oxygen oases” well documented for the Neoarchean [2.8 to 2.5 billion years ago (Ga)], the magnitude, spatial extent, and underlying causes of possible Mesoarchean (3.2 to 2.8 Ga) surface-ocean oxygenation remain controversial. Here, we report δ15N and δ13C values coupled with local seawater redox data for Mesoarchean shales of the Mozaan Group (Pongola Supergroup, South Africa) that were deposited during an episode of enhanced Mn (oxyhydr)oxide precipitation between ∼2.95 and 2.85 Ga. Iron and Mn redox systematics are consistent with an oxygen oasis in the Mesoarchean anoxic ocean, but δ15N data indicate a Mo-based diazotrophic biosphere with no compelling evidence for a significant aerobic nitrogen cycle. We propose that in contrast to the Neoarchean, dissolved O2 levels were either too low or too limited in extent to develop a large and stable nitrate reservoir in the Mesoarchean ocean. Since biological N2 fixation was evidently active in this environment, the growth and proliferation of O2-producing organisms were likely suppressed by nutrients other than nitrogen (e.g., phosphorus), which would have limited the expansion of oxygenated conditions during the Mesoarchean. Postprint

The Pongola Supergroup, deposited between 2.99 and 2.87 Ga on the southeast margin of the Kaapvaal Craton, is one of the least tectonically disturbed Mesoarchaean supracrustal sequences on Earth. Two groups comprise the Pongola Supergroup: the lower Nsuze Group, dominated by volcanic rocks ranging from basalt to basalt-andesite through intermediate to rhyolite compositions, and the upper Mozaan Group, dominated by sedimentary units, which may correlate with portions of the Witwatersrand Supergroup. Pongola Supergroup rocks are largely pristine, frequently preserving original sedimentary and volcanic structures and primary compositions. The depositional environment was an epicontinental sea subjected to sea-level oscillations, resulting in predominantly marginal marine siliciclastic deposits. BIFs, stromatolitic carbonates, glacial diamictites, Mn-rich shales and palaeosols elucidate Mesoarchaean surface processes and suggest shallow-marine oxygen oases under an anoxic atmosphere. The tectonic setting involved a marginal basin undergoing progressive thermal subsidence after magmatism. Palaeomagnetic results are inconsistent but suggest that the Kaapvaal Craton underwent significant wandering. The Pongola Supergroup exhibits two dominant structural domains, including north–northwest-trending rifts and thrust-related structures in the Northern and Central domains, and east–west-trending folds and thrust faults in the south. Economically, the Pongola Supergroup is insignificant apart from small gold deposits, and subeconomic deposits of iron, manganese and uranium.

Multiple tools have been used to determine the sensitivity of phosphates from the early Archaean Barberton greenstone belt to transformation. The assessment of the degree of transformation is crucial for verifying data about the parameters of the paleo-environment. From the obtained results, three generations of phosphates can be distinguished. Group A is observed in cherts and banded iron formation BIF early-generation fluor-hydroxyapatite that precipitated from seawater. It is characterized by flat rare earth element (REE) patterns with a positive Eu anomaly and high Y/Ho ratio in the range of 54–70. Apatites in this group lack any visible indicators of secondary alterations at the micro- and nanoscales. Fourier transform infrared spectra indicate that these apatites are relatively rich in water, and, due to cationic substitution, their OH-stretching regions exhibit complex ordering and numerous component bands. The characteristics observed in the cherts and silicified felsic volcaniclastics of group B imply advanced metasomatic alteration. They exhibit light and heavy REE depletion and an absence of water in the halogen site. Nanoscale investigations reveal cracks, pores, nanofluid inclusions and nanochannellike structures, as well as inclusions. Group C is represented by igneous-derived apatites that partially reflect their igneous origin. The phosphates are predominantly fluorapatite with typical magmatic apatite REE distribution patterns. Imaging at the micro- and nanoscales indicates that they partially preserve the signature of igneous origin. It seems that some of the analyzed apatite partially preserved their primordial features; therefore, they might be used for the reconstruction of Archaean abiotic systems. peerReviewed

The antiquity of life in marine environments has been demonstrated, with examples of microfossils and stromatolites extending back to at least 3.5 billion years ago (Ga). In contrast, emerged land was likely a more challenging environment during the Archaean, and only sparse evidence of life in non-marine environments has so far been identified. Here we document the abundance of isotopically light carbon (with δ 13 C values from −46.6 to −31.3‰), diagnostic of a biogeochemical methane cycle or acetogenesis, in shale and sandstone deposited in ∼3 billion-years-old acidic volcanic lakes on the Kaapvaal Craton of southern Africa. A distinctive Al-rich mineral assemblage with abundant pyrophyllite in lacustrine sedimentary rocks bears similarity to modern volcanic rocks affected by circulation of hot acidic fluids. This is compounded with an enrichment of Ni, Mo, W, As and Cu in whole-rock analyses of sedimentary rocks, which is also observed in geothermal areas of modern volcanic environments. Analysis of early diagenetic pyrite in these sedimentary rocks indicates high nutrient level in the lake, which might reflect hydrothermal input with leaching of volcanic material. Despite the restricted and ephemeral nature of volcanic lakes, a highly productive and complex ecosystem established itself in this environment. Volcanic lakes during the Mesoarchaean thus served as an ecological niche for the development and diversification of microbial life on emerged continental landmasses.

Oxidative weathering of continental rocks is the major source of redox-sensitive elements to the ocean. The timing for the initiation of oxidative weathering remains strongly debated. Here we report new δ 238 U data for the authigenic component of well-preserved ca. 3.07–2.90 Ga marine and lacustrine shales from the Mozaan and Dominion groups of the Kaapvaal Craton, South Africa. The δ 238 Uauth values in marine shales (−0.71‰ to −0.21‰ relative to CRM 112a) are lower than the bulk silicate Earth (BSE, −0.40‰ to −0.17‰), while those of lacustrine shales (−0.68‰ to 0.05‰) scatter above and below the BSE range. In addition, the δ 238 Uauth values of marine shales deposited in shallow seawater are significantly lower than those of the shales deposited in deep seawater. We propose that U(VI) was adsorbed onto Fe-Mn oxides in oxic shallow water and then shuttled to the seafloor. The U isotope fractionations in lacustrine shales are best explained by reduction of U(VI) to U(IV) below the sediment-water interface. Together, these observations bolster the case that a mildly oxidative weathering regime had been established at the Earth's surface by ∼3.0 Ga, paving the way to the onset of aerobic metabolism in both continental and shallow marine environments.

The Paleoproterozoic sandstones of the ca. 2.1 Ga Francevillian Group, southeast Gabon, host the oldest known redox-controlled high-grade uranium (U) deposits. Uranium-bearing detrital minerals (e.g., monazite) in the Francevillian Group sediments were derived by erosion of Archean basement rocks from the East Gabonian block of the northwestern Congo craton surrounding the Francevillian Basin. They are viewed as primary sources for U, but the nature of these sources remains poorly constrained. Here, U-Pb ages and Hf isotope data for detrital zircon grains are combined with literature geochemical data for igneous rocks from the Archean basement, to characterize the nature, age, and origin of the provenance for the Francevillian Basin and the ultimate source of U for the high-grade uranium deposits. Detrital zircon ages range from ca. 3.1 to 2.6 Ga with a dominant population between ca. 2.90 and 2.80 Ga, suggesting a major contribution from ca. 2.88 to 2.81 Ga high-K granitoids with a limited contribution from ca. 3.1 to 2.84 Ga TTGs and 2.75 to 2.7 Ga granitoids of the East Gabonian block. Most detrital zircon grains share similar depleted-mantle Hf model ages between ca. 3.08 and 3.30 Ga, indicating significant recycling of a Paleoarchean to Mesoarchean crust. We suggest that ca. 2.9 to 2.8 Ga high K, U, and Th granitoids, which reflect hybrid melts of both mantle and crustal parentage were the major detrital sources of U to the Francevillian Basin and, ultimately, for the U deposits. We propose that ca. 2.9 to 2.8 Ga detrital zircons with Hf TDM ages between ca. 3.1 and 3.3 Ga may thus allow for vectoring U deposits in sedimentary basins across the Congo craton.

Pyrite is the most abundant ore mineral in gold-bearing quartz pebble conglomerates (QPCs) of the Witwatersrand Basin and similar Archaean sedimentary units. Much of the pyrite in Archaean conglomerates is detrital in origin, implying that it survived weathering, transport, and reworking under anoxic conditions. Detrital pyrite is generally found together with authigenic pyrite formed in situ as a result of syn- to post-depositional processes. As pyrite is frequently associated with gold mineralization, many authors have developed different pyrite classification schemes in order to better understand the nature and origin of the mineralization, although the different classifications cannot always be easily compared as they are largely based on inferred pyrite genesis. We propose a revised, entirely descriptive classification scheme for detrital and authigenic pyrite in Archaean QPCs. The scheme was created in order to provide a consistent framework for the petrographic description of pyrite in clastic sedimentary rocks. Detrital pyrite is subdivided into massive, inclusion-bearing, and coarsely crystalline types. Authigenic pyrite includes euhedral, overgrowth, infill, aggregate, and pseudomorphic types. We evaluate the methods used in pyrite classification, propose formation pathways for detrital and authigenic pyrites, and apply the classification scheme to some well-known occurrences of detrital pyrite-bearing conglomerates.

Oxygenic photosynthesis fundamentally transformed all major biogeochemical cycles and increased the size and complexity of Earth's biosphere. However, there is still debate about when this metabolism evolved. As oxygenic photosynthesis is the only significant source of O-2 at Earth's surface, O-2 -sensitive trace element enrichments and isotopic signatures in Archean sedimentary rocks can potentially be used to determine the onset of oxygenic photosynthesis by tracking shifts in the oxidative capacity of Earth's surface environment. Here, we present an extensive new Archean U isotope record from iron formations, organic-rich shales, and paleosols. Variability in delta U-238 values gradually increased from Archean to Phanerozoic, consistent with current view of gradual oxidation of Earth's surface. In addition, statistical analysis on available delta U-238 data indicates a turning point of delta U-238 variability at roughly 3.0 billon years ago. We suggest that such a turning point in delta U-238 variability indicates the initiation of relatively large-scale oxidative weathering of U(IV)-bearing minerals, implying that oxygenic photosynthesis may have evolved before 3.0 billion years ago.

The history of the growth of continental crust is uncertain, and several different models that involve a gradual, decelerating, or stepwise process have been proposed1–4. Even more uncertain is the timing and the secular trend of the emergence of most landmasses above the sea (subaerial landmasses), with estimates ranging from about one billion to three billion years ago5–7. The area of emerged crust influences global climate feedbacks and the supply of nutrients to the oceans 8 , and therefore connects Earth’s crustal evolution to surface environmental conditions9–11. Here we use the triple-oxygen-isotope composition of shales from all continents, spanning 3.7 billion years, to provide constraints on the emergence of continents over time. Our measurements show a stepwise total decrease of 0.08 per mille in the average triple-oxygen-isotope value of shales across the Archaean–Proterozoic boundary. We suggest that our data are best explained by a shift in the nature of water–rock interactions, from near-coastal in the Archaean era to predominantly continental in the Proterozoic, accompanied by a decrease in average surface temperatures. We propose that this shift may have coincided with the onset of a modern hydrological cycle owing to the rapid emergence of continental crust with near-modern average elevation and aerial extent roughly 2.5 billion years ago. The use of triple-oxygen-isotope data from continental shales spanning the past 3.7 billion years suggests that continental crust with near-modern average elevation and extent emerged about 2.5 billion years ago.

Here, we report the first occurrence of two (Ca-Y)-phosphate phases in apatite crystals from ancient rocks from both the Barberton greenstone belt and the Pilbara Craton. First, a cubic Ca3Y(PO4)3 phase was observed in a sample of silicified tuff from the Mendon Formation from the Barberton greenstone belt. A second phase, corresponding to a synthetic compound with the formula CaYP7O20, was observed in a sample of black banded chert from the Hooggenoeg Formation of the Onverwacht Group and in a sample of chert from the Strelley Pool Chert Formation (East Pilbara Terrane). Based on the presence of these phosphates and specific textures revealed by transmission electron microscopy, we argue for the importance of dissolution-reprecipitation processes in the formation of these phosphate phases. Temperature was likely not the primary parameter controlling the crystallization of the Ca3Y(PO4)3 and CaYP7O20 phases. Instead, the REE-F complexes in an H2O solution and the specific budget of REEs and Y in apatite were likely responsible for the nucleation and formation of the (Ca-Y)-phosphate phases in the Archean rocks of the Barberton greenstone belt and Pilbara Craton.

Chemical zoning of pyrites can record the evolution of mineralising fluids at widely varying P–T conditions ranging from diagenesis to medium-grade metamorphism. If preserved, zoning can reveal growth textures, brecciation and veining, resorption and recrystallisation events, thus shedding light on the processes that contributed to ore formation. Chemical zoning of sulfides is invisible in optical microscopy, but can be studied by chemical etching, high-contrast back-scattering electron images, and elemental imaging. In this study we compared micro-PIXE and WDS-EPMA elemental maps on the chemically zoned pyrites in mineralised vein-bearing samples from the Sheba and Fairview gold mines in the Barberton Greenstone Belt, South Africa. Elemental images show complex distribution of trace elements, suggesting multiple events of pyrite crystallisation and gold deposition. EPMA maps show fine-scale variations reflecting growth and recrystallisation textures marked, in particular, by variations of As, Ni, and Co. In PIXE maps, gold occurs both as finely-distributed and discrete inclusions, suggesting incorporation in the pyrite structure as solid solution, and deposition as electrum inclusions, respectively. Micro-PIXE and EPMA provide complementary information, forming together a powerful tool to obtain information on chemical zoning of pyrites in ore deposits.

In the Barberton greenstone belt (BGB) of South Africa, four distinct spherule horizons (S1–S4) with ages between ca. 3.5 and 3.2 Ga are among the oldest records of large asteroid impacts on Earth. Spherules in these layers are interpreted as molten impact ejecta, condensation products from impact plumes, or impact ejecta melted during atmospheric reentry. Past research has shown that some spherule layer samples from the BGB carry ultrahigh abundances of siderophile elements, including platinum group elements (PGEs), in some cases as much as four times the chondritic abundances. Inferences for very large projectile sizes responsible for these impact layers have been made on the strength of these data. Drilling by the International Continental Scientific Drilling Program has yielded 4 new spherule layer intersections in a 22-cm-long core segment. Scanning electron microscopy–energy dispersive X-ray spectrometry at high spatial resolution identified local areas of PGE enrichment with Ni-rich chromium spinel (Ni-Cr spinel) clusters. They are associated with PGE-rich metal alloy and sulfarsenide phases 600–1400 nm in size. The metal alloys are interpreted as primary objects that formed during the impact process. PGE sulfarsenides are the result of secondary alteration by S- and As-rich fluids. Thus, a micronugget effect caused by PGE phases and Ni-Cr spinel is responsible for the anomalously high extraterrestrial component in some spherule layer samples. This heterogeneous incorporation of meteoritic components due to primary heterogeneous fallout from the vapor plume must be taken into consideration in any attempt to estimate the global fallout of extraterrestrial components and, thus, to constrain projectile sizes.

The Older Metamorphic Group and the Older Metamorphic Tonalite Gneiss are classic examples of Palaeaoarchaean high-grade metamorphic rocks of the Singhbhum Craton of India. The Older Metamorphic Group is a supracrustal assemblage that was probably deposited as a greenstone belt-type succession before c. 3.32 Ga, the low-grade equivalents of which are found in parts of the Iron Ore Group. The Older Metamorphic Tonalite Gneiss represents a suite of granitoids that are tectonically interleaved with the supracrustal gneisses and that formed over an extended period of time, 3.53–3.45 Ga and possibly later, by processes of Archaean crustal growth.

We report zircon ages, Hf-in-zircon isotopes as well as whole-rock geochemistry and Hf–Nd isotopic systematics for Palaeoarchaean grey gneisses of the Ancient Gneiss Complex of Swaziland, the oldest components of the Kaapvaal craton, southern Africa. The Hf-in-zircon isotopic compositions in these compositionally heterogeneous, multicomponent, migmatitic gneisses are highly variable, even in the oldest zircons dating back to 3.66 Ga, suggesting growth of zircon from melts representing a mix of juvenile and anatectic material derived from differentiated continental crust of Eoarchean to late Hadean age. In contrast, the initial Nd and Hf whole-rock isotopic compositions are frequently not in agreement with the Hf-in-zircon data that mostly show approximately chondritic initial values for Nd and strongly radiogenic initial values for Hf. We consider it likely that both the Lu–Hf and Sm–Nd whole-rock isotopic systems were disturbed and partly reset during later episodes of partial melting and crustal reworking, most likely during a pervasive 3.2 Ga tectono-metamorphic event. Primitive mantle-normalized trace element patterns show the variable influence of residual plagioclase and garnet in the sources as well as high contents of strongly incompatible elements. In conjunction with the Hf-in-zircon isotopic data the trace element contents are best explained by the incorporation of older continental crustal material into the sources of the grey gneisses. Our data support evidence from other Palaeoarchean terranes that crustal recycling, as seen in even the oldest crustal components, played an important role in early continental evolution. Rocks previously classified as a subduction-related tonalite–trondhjemite–granodiorite suite are complex, and their chemistry alone cannot be used to reconstruct tectonic settings.

Detrital pyrite is an accessory mineral in clastic sedimentary rocks deposited prior to c. 2.4 Ga. It is a redox-sensitive mineral and an indicator of low oxygen levels prior to the rise of atmospheric oxygen in the early Paleoproterozoic. Resistance to mechanical abrasion, availability in Archean source rocks, and high density make it a common component of Archean quartz pebble conglomerates, some of which may be gold-bearing. The provenance of pyrite may include sediments and sedimentary rock successions, volcanogenic or sedimentary exhalative massive sulfide deposits, magmatic-hydrothermal deposits, and igneous rocks. The morphology and texture of pyrite grains, the nature of mineral inclusions, trace element geochemistry, and isotopic composition as well as age are the main variables that can be used to evaluate their source. In the Mesoarchean Witwatersrand Basin of South Africa, the world's largest gold deposit, detrital pyrite is mainly derived from sedimentary sources and synsedimentary precipitates.

An unusual sediment-hosted manganese deposit is described from the Mesoarchean Mozaan Group, Pongola Supergroup, South Africa. MnO contents up to 15 wt.% were observed in marine clastic and chemical sedimentary rocks. Mn enrichment is interpreted to have resulted from the hydrothermal alteration of manganiferous shale and BIF parent rocks, the primary MnO contents of which are as high as 8.5 wt.%. A detailed mineralogical and petrographic study shows that these parent rocks are characterized by manganoan siderite, ferroan rhodochrosite and other Mn–Fe-rich mineral phases, such as kutnohorite and Fe–Mn-chlorite. Their hypogene alteration gave rise to a diversification of mineral assemblages where ferroan tephroite, calcian rhodochrosite, rhodochrosite, pyrochroite, pyrophanite, cronstedtite, manganoan Fe-rich chlorite and manganoan phlogopite partially or totally replaced the previous mineral assemblage. Thermodynamic modeling performed on chlorite phases associated with the described mineral assemblages illustrates a decrease of average crystallization temperatures from ca. 310 °C during early metamorphic stages to ca. 250 °C during a hydrothermal stage. Mineral transformation processes were thus related to retrograde metamorphism and/or hydrothermal alteration post-dating metamorphism and gave rise to progressive Mn enrichment from unaltered parent to altered rocks. The timing of hypogene alteration was constrained by 40 Ar/ 39 Ar dating to between about 1500 and 1100 Ma ago, reflecting tectonic processes associated with the Namaqua-Natal orogeny along the southern Kaapvaal Craton margin. Manganiferous shale and BIF of the Mozaan Group may represent the oldest known examples of primary sedimentary Mn deposition, related to oxidation of dissolved Mn(II) by free oxygen in a shallow marine environment. Oxygenic photosynthesis would have acted as a first-order control during Mn precipitation. This hypothesis opens a new perspective for better constraining secular evolution of sediment-hosted mineral deposits linked to oxygen levels in the atmosphere-hydrosphere system during the Archean Eon.

The influence of continental weathering on oceanic seawater compositions in the Precambrian is poorly understood. Here we combine Si isotopes and Ge/Si ratios of a Mesoarchean paleosol (~ 2.95 Ga) and overlying shales as proxies for weathering processes and Si mass transfer at the early Earth's surface. Results show that, as with modern soils, neoformation of secondary clay minerals in the paleosol was associated with the fractionation of Si isotopes and Ge/Si ratios in response to chemical weathering and soil desilication. Furthermore, the loss of Fe-bearing minerals, most likely Fe-bearing smectites, produced additional controls on Si and Ge mobility, through the coupled dissolution and re-precipitation of clay minerals under reductive conditions. Opposite fractionation behaviors are observed: neoformed secondary clay minerals acted as a 28 Si and Ge sink, whereas the leaching of Fe-bearing minerals released 28 Si and Ge into soil solutions. An attempt to estimate δ 30 Si and Ge/Si signatures of released dissolved elements into soil solution provided δ 30 Si and Ge/Si signatures of about + 0.1‰ and 1.4 μmol/mol, respectively. Furthermore, shales deposited shortly after paleosol formation display δ 30 Si and Ge/Si compositions that may be explained as mixtures of the recognized paleosol components. The shale record suggests that weathering-induced desilication and leaching of Fe-bearing minerals, to a lesser extent, might have been widely effective during the Mesoarchean for the transfer of Si from the continents to the hydrosphere. Comparison of Si isotopes recorded in detrital sediments and chemical precipitates over time might thus provide a useful supplementary tool to decipher the impact of putative increases in continental weathering and desilication, relative to changes in hydrothermal and continent-derived solute inputs to ocean and ocean cooling.

International audience; The Barberton Greenstone Belt of southern Africa hosts several Mesoarchaean gold deposits. The ores were mostly formed in greenschist facies conditions, and occur as hydrothermal alteration zones around extensional faults that truncate and post-date the main compressional structures of the greenstone belt. Ore deposition was accompanied by the intrusion of porphyries, which has led to the hypothesis that gold may have been sourced from magmas. Because the transport of Au in the hydrothermal fluids is widely believed to have involved S complexes, tracing the origin of S may place strong constraints on the origin of Au. We measured multiple S isotopes in sulfide ore from Sheba and Fairview mines of the Barberton Greenstone Belt to distinguish ‘‘deep” S sources (e.g. magmas) from ‘‘surface” S sources (i.e. rocks of the volcano-sedimentary succession that contain S processed in the atmosphere preserved as sulfide and sulfate minerals). Ion probe (SIMS) analyses of pyrite from ore zones indicate mass-independent fractionation of S isotopes (D33S = 0.6‰ to +1.0‰) and the distribution of the analyses in the D33S–d34S space matches the distribution peak of previously published analyses of pyrite from the entire volcano-sedimentary succession. Notwithstanding that the H2O–CO2 components of the fluids may have been introduced from a deep source external to the greenstone belt rocks, the fact that S bears an atmosphericsignature suggests the hypothesis that the source of Au should also be identified in the supracrustal succession of the greenstone belt. Our findings differ from conclusions of previous studies of other Archaean shear-hosted Au deposits based on mineralogical and isotopic evidence, which suggested a magmatic or mantle source for Au, and imply that there is no single model that can be applied to this type of mineralisation in the Archaean.

We report SHRIMP zircon ages for felsic volcanic rocks of the early Archaean Nondweni greenstone belt (NGB) located in northern KwaZulu-Natal Province, South Africa. The NGB is part of a chain of greenstone belt remnants that occurs south of the well-known Barberton greenstone belt (BGB). Two samples of felsic volcanic rocks of the basal Toggekry Formation yielded zircon ages of ~ 3.53 Ga, whereas significantly younger ages of ~ 3.41 Ga were obtained for two samples of felsic rocks from near the top of the Witkop Formation. The latter date now constrains the age of stromatolite-like structures in cherts associated with the felsic rocks, which probably represent some of the oldest preserved evidence of life on Earth. The zircon ages also indicate that the NGB correlates well chronologically with the BGB and is one of the oldest volcano-sedimentary successions in the southeastern Kaapvaal Craton. Geochemical data for the felsic NGB rocks suggest formation by partial melting or crystal fractionation of a mafic source that was contaminated by older crust, possibly in a back-arc environment. Initial e Nd values for NGB rocks range from − 1.2 to + 0.3 and suggest involvement of pre-greenstone continental crust that was possibly extensive in the southeastern part of the Kaapvaal Craton during Palaeoarchaean times.

Fe isotopes can provide new insight into redox-dependent biogeochemical processes. Precambrian iron formations (IF) are deserving targets for Fe isotope studies because they are composed predominantly of authigenic Fe phases and record a period of unprecedented iron deposition in Earth’s history. We present Fe isotope data for bulk samples from 24 Archean and Proterozoic IF and eight Phanerozoic Fe oxide-rich deposits. These data reveal that many Archean and early Paleoproterozoic iron formations were a sink for isotopically heavy Fe, in contrast to later Proterozoic and Phanerozoic Fe oxide-rich rocks. The positive δ 56 Fe values in IF are best explained by delivery of particulate ferric oxides formed in the water column to the sediment–water interface. Because IF are a net sink for isotopically heavy Fe, there must be a corresponding pool of isotopically light Fe in the sedimentary record. Earlier work suggested that Archean pyritic black shales were an important part of this light sink before 2.35 billion years ago (Ga). It is therefore likely that the persistently and anomalously low δ 56 Fe values in shales are linked with the deposition of isotopically heavy Fe in IF in the deeper parts of basins. IF deposition produced a residual isotopically light dissolved Fe pool that was captured by pyritic Fe in shales. Local dissimilatory Fe reduction in porewater and associated diagenetic reactions resulting in pyrite and carbonate precipitation may have further enhanced Fe isotope heterogeneity in marine sediments, and an ‘iron shuttle’ may have transported isotopically light Fe from shelf sediments to the basin. Nevertheless, water-column processing of hydrothermally delivered Fe likely had the strongest influence on the bulk iron isotope composition of Archean and Paleoproterozoic iron formations and other marine sediments.

The Assegaai, De Kraalen and Witrivier greenstone belts in the southeast part of the Kaapvaal craton contain strongly deformed volcano-sedimentary successions affected by amphibolite- to granulite-facies metamorphism at ∼3.22 to 3.20 Ga. These greenstone belts consist predominantly of mafic-ultramafic volcanic successions intercalated with chert and BIF. Single zircon SHRIMP ages of 3222 ±8 and 3193±5 Ma for two granitoids intrusive into the Assegaai greenstone belt provide a minimum age for the succession. Calc-silicates and amphibolites of the De Kraalen and Witrivier greenstone belts record a high-pressure metamorphic event (M 1 ) at ∼12 to 15 kbar and ∼600 °C and 800 °C, respectively. This high-pressure event was overprinted by a medium-pressure (7-8 kbar) amphibolite-facies event (M 2A ) that was recorded in all three greenstone belts, suggesting that the De Kraalen and Witrivier supracrustal rocks were buried to a deeper crustal level than the Assegaai rocks during crustal thickening. Decompression of the rocks along clockwise P-T paths possibly occurred during subsequent crustal extension at ∼3.22 to 3.20 Ga. The low geothermal gradient of this area (15-20 °C/km) contrasts with higher geothermal gradients reported for other Archean terrains worldwide and may support the view that crustal thickening processes related to convergence in a setting similar to subduction zones were operating in Paleoarchean times.

We present baddeleyite U–Pb ages of Neoarchaean to Palaeoproterozoic dyke swarms and the Mashonaland sill province in Zimbabwe. The 2575.0 ± 1.5 Ma age of the Umvimeela dyke is indistinguishable from the 2575.4 ± 0.7 Ma result (Oberthur et al., 2002) for a pyroxenite layer of the Great Dyke and testifies to synchronous emplacement of the Great Dyke and its satellites. Three samples of WNW- to NNW-trending dykes of the Sebanga swarm yielded ages of 2512.3 ± 1.8 Ma, 2470.0 ± 1.2 Ma and 2408.3 ± 2.0 Ma, the latter of which dates the Sebanga Poort Dyke of this swarm. These results suggest that emplacement took place over a protracted period which involved at least three generations of dykes within the swarm and, more importantly, invalidate previous inferences of a genetic link between the Sebanga swarm and the Mashonaland sills. Crystallisation ages of 1877 ± 2.2 Ma, 1885.9 ± 2.4 Ma and 1875.6 ± 1.6 Ma for three dolerite samples of the extensive Mashonaland sills from different parts of the Zimbabwe craton were also obtained. This is the oldest common igneous event that is recorded in the Zimbabwe and Kaapvaal cratons. Collectively with previous published geochronological and petrological evidence in favour of a major 2.0 Ga event within the Limpopo Belt, these results suggest that the Zimbabwe and Kaapvaal cratons did not form a coherent unit (Kalahari) until ca. 2.0 Ga. In the global barcode record for Archaean cratons, all the three generations of dykes within the Sebanga swarm match exactly events of dyke intrusion and LIPs (Large Igneous Provinces) in the Superior, Hearne and Kola-Karelia cratons, and corroborates the recognition of Superia as a supercontinent in the Neoarchaean. The 2408.3 ± 2.0 Ma result of the Sebanga Poort Dyke also provides a link to Western Australia, i.e. to the giant Widgiemooltha swarm of the Yilgarn craton. We conclude that in the context of the global barcode record, both the Zimbabwe and Yilgarn cratons fit into a Superia reconstruction.

The ocean and atmosphere were largely anoxic in the early Precambrian, resulting in an Fe cycle that was dramatically different than today's. Extremely Fe-rich sedimentary deposits i.e., Fe formations are the most conspicuous manifestation of this distinct Fe cycle. Rare Earth Element (REE) systematics have long been used as a tool to understand the origin of Fe formations and the corresponding chemistry of the ancient ocean. However, many earlier REE studies of Fe formations have drawn ambiguous conclusions, partially due to analytical limitations and sampling from severely altered units. Here, we present new chemical analyses of Fe formation samples from 18 units, ranging in age from ca. 3.0 to 1.8 billion years old (Ga), which allow a reevaluation of the depositional mechanisms and significance of Precambrian Fe formations. There are several temporal trends in our REE and Y dataset that reflect shifts in marine redox conditions. In general, Archean Fe formations do not display significant shale-normalized negative Ce anomalies, and only Fe formations younger than 1.9 Ga display prominent positive Ce anomalies. Low Y/Ho ratios and high shale-normalized light to heavy REE (LREE/HREE) ratios are also present in ca. 1.9 Ga and younger Fe formations but are essentially absent in their Archean counterparts. These marked differences in Paleoproterozoic versus Archean REE Y patterns can be explained in terms of varying REE cycling in the water column. Similar to modern redox-stratified basins, the REE Y patterns in late Paleoproterozoic Fe formations record evidence of a shuttle of metal and Ce oxides across the redoxcline from oxic shallow seawater to deeper anoxic waters. Oxide dissolution mainly of Mn oxides in an anoxic water column lowers the dissolved Y/Ho ratio, raises the light to heavy REE ratio, and increases the concentration of Ce relative to the neighboring REE (La and Pr). Fe oxides precipitating at or near the chemocline will capture these REE anomalies and thus evidence for this oxide shuttle. In contrast, Archean Fe formations do not display REE Y patterns indicative of an oxide shuttle, which implies an absence of a distinct Mn redoxcline prior to the rise of atmospheric oxygen in the early Paleoproterozoic. As further evidence for reducing conditions in shallow-water environments of the Archean ocean, REE data for carbonates deposited on shallow-water Archean carbonate platforms that stratigraphically underlie Fe formations also lack negative Ce anomalies. These results question classical models for deposition of Archean Fe formations that invoke oxidation by free oxygen at or above a redoxcline. In contrast, we add to growing evidence that metabolic Fe oxidation is a more likely oxidative mechanism for these Fe formations, implying that the Fe distribution in Archean oceans could have been controlled by microbial Fe uptake rather than the oxidative potential of shallow-marine environments. (C) 2010 Elsevier Ltd. All rights reserved.

Iron formations are economically important sedimentary rocks that are most common in Precambrian sedimentary successions. Although many aspects of their origin remain unresolved, it is widely accepted that secular changes in the style of their deposition are linked to environmental and geochemical evolution of Earth. Two types of Precambrian iron formations have been recognized with respect to their depositional setting. Algoma-type iron formations are interlayered with or stratigraphically linked to submarine-emplaced volcanic rocks in greenstone belts and, in some cases, with volcanogenic massive sulfide (VMS) deposits. In contrast, larger Superior-type iron formations are developed in passive-margin sedimentary rock successions and generally lack direct relationships with volcanic rocks. The early distinction made between these two iron-formation types, although mimimized by later studies, remains a valid first approximation. Texturally, iron formations were also divided into two groups. Banded iron formation (BIF) is dominant in Archean to earliest Paleoproterozoic successions, whereas granular iron formation (GIF) is much more common in Paleoproterozoic successions. Secular changes in the style of iron-formation deposition, identified more than 20 years ago, have been linked to diverse environmental changes. Geochronologic studies emphasize the episodic nature of the deposition of giant iron formations, as they are coeval with, and genetically linked to, time periods when large igneous provinces (LIPs) were emplaced. Superior-type iron formation first appeared at ca. 2.6 Ga, when construction of large continents changed the heat flux at the core-mantle boundary. From ca. 2.6 to ca. 2.4 Ga, global mafic magmatism culminated in the deposition of giant Superior-type BIF in South Africa, Australia, Brazil, Russia, and Ukraine. The younger BIFs in this age range were deposited during the early stage of a shift from reducing to oxidizing conditions in the ocean-atmosphere system. Counterintuitively, enhanced magmatism at 2.50 to 2.45 Ga may have triggered atmospheric oxidation. After the rise of atmospheric oxygen during the GOE at ca. 2.4 Ga, GIF became abundant in the rock record, compared to the predominance of BIF prior to the Great Oxidation Event (GOE). Iron formations generally disappeared at ca. 1.85 Ga, reappearing at the end of the Neoproterozoic, again tied to periods of intense magmatic activity and also, in this case, to global glaciations, the so-called Snowball Earth events. By the Phanerozoic, marine iron deposition was restricted to local areas of closed to semiclosed basins, where volcanic and hydrothermal activity was extensive (e.g., back-arc basins), with ironstones additionally being linked to periods of intense magmatic activity and ocean anoxia. Late Paleoproterozoic iron formations and Paleozoic ironstones were deposited at the redoxcline where biological and nonbiological oxidation occurred. In contrast, older iron formations were deposited in anoxic oceans, where ferrous iron oxidation by anoxygenic photosynthetic bacteria was likely an important process. Endogenic and exogenic factors contributed to produce the conditions necessary for deposition of iron formation. Mantle plume events that led to the formation of LIPs also enhanced spreading rates of midocean ridges and produced higher growth rates of oceanic plateaus, both processes thus having contributed to a higher hydrothermal flux to the ocean. Oceanic and atmospheric redox states determined the fate of this flux. When the hydrothermal flux overwhelmed the oceanic oxidation state, iron was transported and deposited distally from hydrothermal vents. Where the hydrothermal flux was insufficient to overwhelm the oceanic redox state, iron was deposited only proximally, generally as oxides or sulfides. Manganese, in contrast, was more mobile. We conclude that occurrences of BIF, GIF, Phanerozoic ironstones, and exhalites surrounding VMS systems record a complex interplay involving mantle heat, tectonics, and surface redox conditions throughout Earth history, in which mantle heat unidirectionally declined and the surface oxidation state mainly unidirectionally increased, accompanied by superimposed shorter term fluctuations.

Multiple S ( δ 34 S and δ 33 S) and Fe ( δ 56 Fe) isotope analyses of rounded pyrite grains from 3.1 to 2.6 Ga conglomerates of southern Africa indicate their detrital origin, which supports anoxic surface conditions in the Archaean. Rounded pyrites from Meso- to Neoarchaean gold and uranium-bearing strata of South Africa are derived from both crustal and sedimentary sources, the latter being characterised by non-mass dependent fractionation of S isotopes ( Δ 33 S as negative as − 1.35‰) and large range of Fe isotope values ( δ 56 Fe between − 1.1 and 1.2‰). Most sediment-sourced pyrite grains are likely derived from sulphide nodules in marine organic matter-rich shales, sedimentary exhalites and volcanogenic massive sulphide deposits. Some sedimentary pyrite grains may have been derived from in situ sulphidised Fe-oxides, prior to their incorporation into the conglomerates, as indicated by unusually high positive δ 56 Fe values. Sedimentary sulphides without significant non-mass dependent fractionation of S isotopes were also present in the source of some conglomerates. The abundance in these rocks of detrital pyrite unstable in the oxygenated atmosphere may suggest factors other than high p O 2 as the cause for the absence of significant non-mass dependent fractionation processes in the 3.2–2.7 Ga atmosphere. Rounded pyrites from the c. 2.6 Ga conglomerates of the Belingwe greenstone belt in Zimbabwe have strongly fractionated δ 34 S, Δ 33 S and δ 56 Fe values, the source of which can be traced back to black shale-hosted massive sulphides in the underlying strata. The study demonstrates the utility of combined multiple S and Fe isotope analyses for provenance reconstruction of Archaean sedimentary successions.

In order to evaluate the potential of felsic sediments as a CO2-sink in the Archean, we studied felsic volcaniclastic/epiclastic sedimentary rocks of the 3.45 Ga Hooggenoeg Formation, Barberton greenstone belt, which were affected by metasomatic processes during seafloor alteration and diagenesis. Water–rock interactions leading to K-, Si- and CO2-metasomatism were quantified. The precursor rock, a K-metasomatized dacite, was leached of Ca, Mg, Fe, Na, Sr, and Ba and enriched in K, Rb and Si prior to erosion and deposition. The formation of K-mica and quartz (and minor K-feldspar) in the dacites suggests low pH metasomatic conditions, likely produced by hydrothermally induced circulation of hot, acidic and reduced Archean seawater in equilibrium with a high P C O 2 atmosphere. Erosion and transport of K-metasomatized dacitic detritus away from the felsic volcanic centers resulted in the deposition of conglomerate, sandstone and shale by mass flow processes. Early diagenetic silicification affected mainly the fine-grained sediments with higher silica sorption capacity, forming impermeable layers, while sand-rich sediments were partly silicified and remained permeable. Trapped fluids precipitated two generations of Fe-rich dolomites and finally calcite. Up to 30 vol.% of siliciclastic coarse-grained sediment was replaced by carbonates in a shallow-burial, high heat-flow diagenetic regime (depth: ∼750 m, temperature: 80–160 °C), and likely throughout deposition of overlying volcano–sedimentary units. The carbon isotopic composition of Fe-rich dolomites (δ13CPDB = +1.9 to +2.4‰) and the strong Fe–Ca–Mg leaching of the Paleoarchean volcanic formations support the influence of seawater-derived fluids throughout CO2-metasomatism. For each gram of eroded dacite, the overall chemical exchange involved by K–Si–CO2-metasomatism was characterized by a mass transfer of Fe (1.2 mmol/g), Na (2.1 mmol/g) and O2− (2.0 mmol/g) to seawater. In contrast, seawater was depleted in Si (10 mmol/g), Ca (0.51 mmol/g), Mg (0.43 mmol/g), K (1.5 mmol/g) and H (0.93 mmol/g) during incorporation of these elements in the volcanic and sedimentary rocks. The average CO2 uptake by the sedimentary rocks studied here is estimated to be 1.8 mmol/g, in the same order of magnitude as previous estimates for the Paleoarchean basaltic crust. Although mafic-ultramafic rocks are the most abundant rocks in Paleoarchean greenstones belts, and represent the most important atmospheric CO2-sink upon seafloor alteration in the Paleoarchean, coarse felsic sedimentary rocks provide a non-negligible contribution to the build-up of the continental CO2 reservoir.

Silicification of volcanic rocks, ranging from komatiitic to dacitic in composition, is a common phenomenon of the 3.5 to 3.2 Ga old volcano-sedimentary succession of the Barberton greenstone belt. Silica enrichment occurs in zones several tens of metres thick at the top of volcanic sequences that are capped by sedimentary chert horizons. SiO2 contents increase upsection, from the original igneous value up to 92%. Silicification is associated with a depletion of most elements commonly mobile during water–rock interaction, while K2O, Rb and Ba are enriched in these zones. Some transition metals (Sc, V, and Cr) were immobile, whereas Ni and Co and some other metals (Cu, Zn) were mobile and are depleted. The δ18O values of volcanic rocks show a positive linear relationship with silica content and range from 9.0 to 17.3‰. The element depletion-enrichment patterns and oxygen isotope data indicate low-temperature (c. 100–150 °C) hydrothermal processes for the origin of the alteration zones. The loss of most elements coupled with the addition of SiO2 as well as REE systematics indicate high water–rock ratios and a fluid REE composition typical of Archaean seawater that had a δ18O value of ~ 0‰. Hydrothermal activity is attributed to heat derived from cooling of volcanic rocks and a high regional heat flow in an oceanic plateau-like setting and resulted in the establishment of shallow subseafloor convection cells. Diffuse venting of hydrothermal fluids over broad areas of the ancient ocean floor provided an ideal environment for both seafloor and subsurface biota 3.5 Ga ago.

The 3.5-3.2 Ga old volcano-sedimentary succession of the Barberton greenstone belt (South Africa) is characterized by lithological units that are repeated in a regular manner. Komatiitic, basaltic, and dacitic volcanic and volcaniclastic sequences are capped by zones of silica enrichment, followed by bedded carbonaceous cherts. Stratiform and crosscutting carbonaceous chert veins are common in silica alteration zones and bedded cherts. A detailed field study of several chert horizons and chert veins that range in age from 3.47 to 3.30 Ga revealed the importance of syndepositional hydrothermal activity for their origin. Bedded cherts consist of silicified detrital and tuffaceous sediments that were deposited on the seafloor. Silicification took place at the sediment-water interface as a result of diffuse upflow of low-temperature hydrothermal fluids, which gave rise to the formation of impermeable chert caps. Fluid overpressure resulted in the breaching of the cap rocks at times. Chert veins contain angular host rock fragments, replace wall rocks, and show evidence of multiple vein fillings and in situ brecciation of earlier generations of vein fillings. They represent hydraulic fractures that were initiated by overpressuring of the hydrothermal system. The vein networks were infilled, partly by hydrothermal chert precipitates, and partly by still unconsolidated (not yet silicified) sedimentary material derived from overlying sedimentary horizons. Field, petrographic, isotopic, and trace element evidence indicate that most carbonaceous matter represents sedimentary material that originated by biogenic processes in the Archean oceans and not by hydrothermal processes in the subsurface.

The 3.45 Ga old felsic volcanic rocks from the Barberton Greenstone Belt of southern Africa include a submarine succession of felsic lavas and volcaniclastic rocks and shallow subvolcanic intrusions. Sea floor alteration strongly affected rock compositions shortly after emplacement, making comparisons with coeval intrusions challenging. We analysed well-preserved quartz-hosted melt inclusions from the felsic rocks, which offers a unique opportunity to gain insight into pre-alteration melt compositions. Melt inclusion compositions are then compared with whole-rock analyses to evaluate the significance of the volcanic complex in the context of Archaean felsic magmatism in the Barberton Greenstone Belt. Whole-rock immobile trace element concentrations, including high-field strength and rare earth elements, indicate strong similarities with coeval tonalite–trondhjemite–granodiorite intrusions. These same geochemical characteristics distinguish these rocks from other Archaean felsic rocks in the Barberton Greenstone Belt (e.g. the 3.54 Ga Theespruit Formation), which have characteristically higher Th, Nb and REE. Low CaO (≤ 1 wt.%) and relatively high Nb and Th (mostly ≤ 9 and ≤ 13 ppm, respectively) in comparison with least altered whole-rock samples, and the presence of negative Eu anomalies are consistent with the melt inclusions being relatively evolved, rather than representing primitive compositions. Melt inclusion analyses suggest the presence of two melts with varying K2O/Na2O. Melt inclusions from extrusive samples, which contain remnants of K-feldspar phenocrysts, have K2O/Na2O > 1 by weight, whereas subvolcanic intrusions, which contain Na-rich plagioclase, have K2O/Na2O

The ca. 3.25 Ga old Fig Tree Group in the southern part of the Barberton greenstone belt consists of deep- to shallow-water shale, greywacke, jaspilitic banded iron formation (BIF), and carbonaceous chert. The sequence is more than 1200 m thick and crops out as tectonically duplicated, but stratigraphically distinct tectono-stratigraphic units. Chemical weathering of the source terrain of Fig Tree strata was minor. Instead, hydrothermal-metasomatic events affected the sedimentary rocks in the study area, resulting in the depletion of alkaline earth elements and K metasomatism. Provenance modelling using REEs and trace element ratios indicate varying contributions from ultramafic to mafic greenstones, TTGs and HREE-undepleted granites. A clear stratigraphic control on the composition, in the form of increasing trace metal contents and decreasing La N /Yb N ratios, reflects progressive increase of ultramafic and mafic volcanic rocks in the source. Banded iron formation shows REE-Y patterns consistent with its precipitation in a marine environment from Eu-enriched seawater. Carbonaceous cherts at the base of the Fig Tree Group do not represent chemical precipitates out of a hot Archaean ocean, but represent carbonaceous sediments that were silicified during low-temperature hydrothermal fluid emanations on the seafloor.

Late Archaean (∼2.68 Ga) clastic sedimentary rocks of the Shamvaian Group form a shear zone-bounded lithotectonic unit in the Midlands greenstone belt (MGB) of Zimbabwe. The sedimentary rocks are wedged in between mid-Archaean granitoid gneisses and late Archaean mafic and bimodal volcanic rocks. The ≥2000-m thick sedimentary sequence consists of marine turbidite deposits overlain by high-energy shelf sandstone, fluvial braid-plain pebbly sandstone and alluvial fan conglomerate. Deposition took place in a northward-deepening basin with fluvial systems draining a source dominated by granitoid gneiss and banded chert. Basin formation was associated with a west-directed thrusting event, and the basin fill was incorporated into the thrust stack soon after deposition. The Shamvaian Group is similar to modern foreland and forearc successions and possibly accumulated during collision of a continental fragment with an island arc-like complex.

Shallowing-upward carbonate cycles in the 2650 Ma Cheshire Formation, Belingwe greenstone belt (Zimbabwe), closely resemble their Proterozoic and Phanerozoic counterparts. The cycles form part of a karstified carbonate-ramp sequence that is overlain by, and grades basinward into, siliciclastic turbidites. A single section of 74 cycles (1.5 m average thickness) was studied in detail. Two basic cycle types are recognized, both with an asymmetric facies stacking pattern. One cycle type contains open marine, subtidal shale at the base. Shale is intercalated with storm-generated sandstone and grainstone beds that become more common and thicken upward, indicating progressive shallowing. Wave-rippled ooid-intraclast grainstone beds and bedsets overlie shale or form the base of the second cycle type. Grainstone formed at or above fair-weather wave base as shoreface sand sheets in an agitated, shallow subtidal setting. Microbial laminites constitute the top of both cycle types and are interpreted as peritidal deposits. In the upper part of the studied section, microbial boundstones with aragonite pseudomorphs are intercalated with, or overlie, laminites and formed in a supratidal environment. The vertical facies distribution within a cycle is indicative of rapid submergence followed by gradual shallowing of relative sea level. High-frequency eustatic sea-level changes are favored over an autocyclic mechanism and tectonically induced allocyclicity as the controlling mechanism for the cyclicity. Hierarchies of stratigraphic cyclicity occur on different scales and may be a result of the combined effects of several orders of sea-level oscillations. Cycle recurrence ratios correspond well to the Milankovitch frequencies calculated for the Late Archean, suggesting that orbital climatic forcing may have been in operation in Archean times.

Metre-thick horizons of ironstone lithologically similar to ‘sulphide-facies’ banded iron formation, but interpreted as silicified and sulphide-impregnated shear zones, are a common component of the greenstone stratigraphy of the Archaean Zimbabwe craton. Such tectonic ironstones separate different lithostratigraphic units commonly regarded as autochthonous rock sequences. On outcrop scale, shearing along ironstone horizons is indicated by anastomosing foliation domains, folding, boudinage and mylonitic fabrics, and, on a regional scale, by truncation of bedding and/or foliation, an anastomosing geometry of the horizons and duplication or juxtaposition of lithostratigraphic units. Tectonic ironstone formation is attributed to in situ silicification and iron (sulphide)-impregnation of rocks, commonly sediments, by mineralized fluids that penetrated the shear zones and their immediate wall rocks. The shear zones formed as a result of thin-skinned thrust tectonics that gave rise to horizontal accretion, imbrication and juxtaposition of volcanic and sedimentary rock units before deformation associated with granitoid diapirism. As a result, ‘layer-cake’ stratigraphic models of greenstone sequences containing tectonic ironstone ‘layers’ have to be treated with care, and the use of ironstone horizons as stratigraphic markers should be discouraged.

The Chirwa dome in northeast Zimbabwe is situated at the boundary between the Zimbabwe craton and Archaean gneisses of the Zambezi mobile belt. This circular granite intrusion has long been regarded as Archaean granite which was remobilized during Pan-African times and emplaced as a mantled gneiss dome into a Proterozoic metasedimentary sequence. Field mapping, structural work and new zircon dates indicate that the Chirwa dome and surrounding rocks underwent a very different history. The supracrustal sequence was deposited and moderately deformed between ~2613 and ~2601 Ma ago, as bracketed by a provenance date and an intrusive date, respectively. Shortly after deposition, the sequence was thrust westward and juxtaposed against a greenstone terrain of the Zimbabwe craton. This active thrust stack was intruded by successive syntectonic granitoids at ~2601 and ~2593 Ma. The Chirwa dome granite, ~2570 Ma old, represents the last intrusive event before cessation of deformation. Our data indicate that while the central Zimbabwe craton already behaved as a stable crustal block at ~2600 Ma, tectonism was active along its northeastern margin.

Detailed structural data from the Midlands greenstone belt show that, in contrast to pre-existing models, the evolution of the greenstone belt involved an early episode of thin-skinned thrusting affecting separate lithological domains. These display different sedimentary and structural-metamorphic histories prior to their juxtaposition across steep, west-directed thrust zones. The domains include 3.5 Ga old gneiss, 2.67–2.88 Ga, mafic–felsic volcanic units and early-syn-tectonic, clastic sedimentary sequences. Concomitant with thrusting along domain boundaries, upright folding and ‘minor’ shear accommodated strain within domains. The early thin-skinned, and later, steeper, thrusting events can be interpreted as progressive stages in an accretionary and crustal thickening sequence, possibly associated with under-plating or low-angle subduction. The clastic sedimentary sequences contain intraformational clasts and show coarsening upward cycles below major early thrust horizons, and may have formed on alluvial fans developing in front of west-moving nappes. Strike-slip motion on the main shear zones in the Midlands greenstone belt only occurred late in the tectonic history and there is no evidence that this resulted in large displacements. A shift of σ 1 , from E–W to N–S during the late stages of Archaean deformation of the greenstone belt can be attributed to extensional collapse following E–W shortening and crustal thickening of the Zimbabwe craton.

Pb–Pb isochrons have been obtained for stromatolitic limestones from the late Archaean Belingwe Greenstone Belt of Zimbabwe, providing direct age constraints on the deposition of these shallow water marine sediments. Samples from the Manjeri Formation and stratigraphically higher Cheshire Formation yield age estimates of 2706±49 Ma (MSWD=11.5) and 2601±49 Ma (MSWD=0.93), respectively. These data are in agreement with published U–Pb zircon and Pb–Pb whole rock ages of associated volcanics, and we, therefore, interpret our Pb–Pb ages as representing the timing of early diagenesis, thus providing a minimum age for carbonate precipitation. A 2543±70 Ma age (MSWD=5.1) for one sample from the Cheshire Formation is considered to reflect either late-stage diagenesis, a craton wide thermal/chemical disturbance or tectonic activity along the crustal-scale Mtshingwe fault. Calculated model μ1-values for Manjeri and Cheshire limestones are 8.40±0.02 and 9.02±0.01, similar to values for approximately 3.5 and 2.9 Ga Archaean basement units adjacent to the Belingwe Belt. Negative eNd (Tdeposition) and fSm/Nd suggest derivation of the REE from old, LREE enriched continental crust. Two-stage Nd model ages for the carbonates indicate that precursor rocks were extracted from the mantle at around 3.5 Ga (Cheshire Formation) and 3.3 Ga (Manjeri Formation), in good agreement with mantle-extraction ages for local basement units (model TDM: 3.5–3.3 Ga).

Iron and sulfur isotope compositions recorded in ancient rocks and minerals such as pyrite (FeS2) have been widely used as a proxy for early microbial metabolisms and redox evolution of the oceans. However, most previous studies focused on only one of these isotopic systems. Herein, we illustrate the importance of in-situ and coupled study of Fe and S isotopes on two pyrite nodules in a c. 2.7 Ga shale from the Bubi Greenstone Belt (Zimbabwe). Fe and S isotope compositions were measured both by bulk-sample mass spectrometry techniques and by ion microprobe in-situ methods (Secondary Ion Mass Spectrometry, SIMS). Spatially-resolved analysis across the nodules shows a large range of variations at micrometer-scale for both Fe and S isotope compositions, with δ56Feδ56Fe and δ34Sδ34S values from −2.1 to +0.7‰+0.7‰ and from −0.5 to +8.2‰+8.2‰, respectively, and Δ33SΔ33S values from −1.6 to +2.9‰+2.9‰. The Fe and S isotope variations in these nodules cannot be explained by tandem operation of Dissimilatory Iron Reduction (DIR) and Bacterial Sulfate Reduction (BSR) as was previously proposed, but rather they reflect the contributions of different Fe and S sources during a complex diagenetic history. Pyrite formed from two different mineral precursors: (1) mackinawite precipitated in the water column, and (2) greigite formed in the sediment during early diagenesis. The in-situ analytical approach reveals a complex history of the pyrite nodule growth and allows us to better constrain environmental conditions during the Archean.

Iron formations are economically significant, iron- and silica-rich sedimentary rocks that are restricted to Precambrian successions. There are no known modern or Phanerozoic analogues for these deposits that are comparable in terms of areal extent and thickness. Although many aspects of iron formation origin remain debatable, it is generally accepted that secular changes in the style of deposition are genetically linked to plate tectonic processes, mantle plume events, and evolution of Earth's surface environments. Two types of Precambrian iron formations have been recognized based on depositional and tectonic settings. Iron formations formed proximal to volcanic centers are interlayered with or laterally linked to submarine volcanic rocks and, in some cases, with volcanogenic massive sulfide (VMS) deposits. In contrast, larger sedimentary rock-hosted iron formations are developed in passive-margin settings and typically lack a direct association with volcanic rocks. A full gradation between these two end-members exists in the rock record. Texturally, iron formations are divided into two groups. Banded iron formation (BIF) is predominant in Archean to earliest Paleoproterozoic successions, whereas granular iron formation (GIF) is more common in middle to late Paleoproterozoic successions, having been deposited in shallow-marine settings after the rise of atmospheric oxygen at ~2.4 Ga. Secular changes in the style of iron formation deposition have been linked to a diverse array of environmental changes. Geochronologic studies emphasize the periodicity in deposition of giant iron formations, which are coeval with large igneous provinces (LIPs). Giant sedimentary rock-hosted iron formations first appeared ~ 2.6 Ga, possibly when the construction of large continents changed the heat flux across the core–mantle boundary. From ~ 2.6 to ~ 2.4 Ga, global mafic-to-ultramafic magmatism culminated in the deposition of giant sedimentary rock-hosted iron formations in South Africa, Australia, Brazil, Russia, and Ukraine. The younger BIFs in this age range were deposited immediately before a shift from reducing to oxidizing conditions in the ocean–atmosphere system. Counterintuitively, enhanced magmatism at 2.50–2.45 Ga, which likely delivered large amounts of reductants to shallow-marine environments, may have triggered atmospheric oxidation. After the rise of atmospheric oxygen ~ 2.4 Ga, GIF became more abundant in the rock record than BIF. Iron formations largely disappeared ~ 1.85 Ga, reappearing at the end of the Neoproterozoic, again tied to periods of intense magmatic activity and also, in this case, to global-scale glaciations, during the so-called snowball Earth events. In the Phanerozoic, deeper-water iron formation deposition became restricted to local areas of closed to semi-closed basins, where volcanic and hydrothermal activity was extensive, such as in back-arc basins. In contrast, episodically deposited, basin-scale Phanerozoic oolitic and pisolitic ironstones are linked to periods of intense magmatic activity and ocean anoxia. Late Paleoproterozoic iron formations and at least some Paleozoic ironstones were deposited at the redoxcline, where biological and nonbiological oxidation occurred. In contrast, older iron formations were deposited in anoxic oceans, where ferrous iron oxidation by anoxygenic photosynthetic bacteria was likely an important process. Endogenic and exogenic factors contributed to the production of the conditions necessary for deposition of iron formation. Mantle plume events that led to the emplacement of LIPs also enhanced spreading rates of mid-ocean ridges and resulted in higher growth rates of oceanic plateaus; both processes thus contributed to a higher hydrothermal iron flux to the oceans. Oceanic and atmospheric redox states determined the fate of this flux. When the hydrothermal flux overwhelmed the ocean oxidation state, iron was transported and deposited distally from hydrothermal vents. Where the hydrothermal flux was insufficient to overwhelm the ocean redox state, iron was deposited mainly proximally, generally as oxides or sulfides; manganese was more mobile. It is concluded that occurrences of BIF, GIF, Phanerozoic ironstones, and hydrothermal sedimentary rocks of exhalative origin (exhalites) surrounding VMS systems are a record of a complex interplay among mantle plume events, plate tectonics, and ocean redox conditions throughout Earth's history, in which mantle heat unidirectionally decreased and the surface oxidation state mainly unidirectionally increased, accompanied by superimposed shorter-term fluctuations.