Abstract The spark to put together this volume on banded iron formation (BIF)-related high-grade iron ore was born in 2005 during a steamy night in Carajás where the iron research group from the Universidade Federal Minas Gerais, Vale geologists, Carlos Rosière and Steffen Hagemann, were hotly debating the hypogene alteration genesis for the high-grade, jaspilite-hosted Serra Norte iron ore deposits. A couple of caipirinhas later we decided that the time was opportune to put together a volume that captured the new and innovative research that was being conducted on BIF-related high-grade iron ores throughout the world. We had little problem convincing our South African colleagues Jens Gutzmer and Nic Beukes to join the effort and decided that the 2008 biannual Society of Economic Geologists' (SEG) meeting in South Africa would be the perfect place to present this project through a combined field trip and workshop near Sishen. The enthusiastic support that we received from the research community, SEG, and industry to put this volume together was generated by the significant increase in exploration activity, and with it the need for more detailed information on what exactly controls the location of high-grade iron orebodies, and renewed research interest around the world in models for the genesis of BIF-related high-grade iron ore, and particularly the relative importance of hypogene and supergene processes in formation of high-grade ore.

Abstract This paper provides a critical review of advances made in understanding of sedimentary environments, geochemical processes, and biological systems that contributed to the deposition and diagenetic evolution of the exceptionally well-preserved and large iron formations of the late Neoarchean to very early Paleoproterozoic Ghaap-Chuniespoort Group of the Transvaal Supergroup on the Kaapvaal craton (South Africa) and the time equivalent Hamersley Group on the Pilbara craton (Western Australia). These iron formations are commonly assumed to have formed coevally but in separate basins, and they are often used as proxies for global ocean chemistry and paleoenvironmental conditions at ~2.5 Ga. However, lithostratigraphic and paleogeographic reconstructions show that the iron formations formed in a single large partly enclosed oceanic basin along the margins of the ancient continent of Vaalbara. Furthermore, although large relative to other preserved iron formations, the combined Transvaal-Hamersley basin is miniscule compared to marginal basins of the modern ocean system so that the succession probably documents secular changes in depositional environments of that basin rather than of the global ocean at the time. The iron formations comprise a large variety of textural and mineralogical rock types that display complex lateral and vertical facies variations on basinal scale. Based on detailed analyses of these variations it is concluded that the iron formations were deposited in environments that ranged from very deep-water basinal settings far below storm-wave base and the photic zone to very shallow-platform settings above normal wave base. Precipitation of both iron and silica took place from hydrothermal plumes in a dynamically circulating ocean system that was not permanently stratified. Ferric oxyhydroxide was the primary iron precipitate in virtually all of the iron formation facies. This primary precipitate is now represented by early diagenetic hematite in some of the iron formations. However, in both deep- and shallow-water iron formations most of the original ferric oxyhydroxides have been transformed by dissimilatory iron reduction to early diagenetic siderite and/or magnetite in the presence of organic carbon. Precipitation of ferric oxyhydroxides in very deep water below the photic zone required a downward flux of photosynthetically-derived free oxygen from the shallow photic zone. In these deep-water environments, under microaerobic conditions, chemolithoautotrophic iron-oxidizing bacteria may have played an important role in precipitation of ferric oxyhydroxides and acted as a source of primary organic matter. With basin fill even shallow-shelf embayments were invaded by circulating hydrothermal plume water from which ferric oxyhydroxides could be precipitated in oxygenated environments with high primary organic carbon productivity and thus iron reduction to form hematite-poor siderite- and magnetite-rich clastic-textured iron formations. Depositional models derived from the study of the iron formations along the Neoarchean-Proterozoic boundary can be applied to iron formations of all ages in both the Archean and later Paleoproterozoic. The facies architecture of the iron formations determines to a large degree the textural attributes, composition, and stratigraphic setting of high-grade iron ores hosted by them. Detailed facies information thus would assist in improving genetic models for high-grade iron ore deposits. Future research should be guided in this direction, especially in some of the very large iron ore districts of Brazil and India where very little is known about the composition and facies variations of the primary iron formation hosts and possible controls on localization of high-grade ores.

Abstract The processes responsible for the transformation of banded iron formations to hard high-grade hematite ore, and their timing, remain poorly understood despite many recent advances. The paleomagnetic method allows for the estimation of ore genesis timing as a complement to other techniques. The effectiveness of the paleomagnetic method at dating, and testing proposed models for, the genesis of hard high-grade hematite ore deposits is illustrated by two South African examples. A new dataset is reported for the Thabazimbi deposit that independently constrains the age of ore formation between 2054 and 1930 Ma, while previously published data from the Sishen-Beeshoek deposits highlight the association of those deposits with weathering preceding the development of a marked Paleoproterozoic-aged unconformity (older than 2060 Ma). Paleomagnetic results are in both cases consistent with proposed models of ore genesis (i.e., extensive carbonate metasomatism and meteoric fluid interaction at Thabazimbi and ancient supergene processes at Sishen-Beeshoek). The antiquity of these South African examples appears to reflect a common theme among other hard high-grade hematite deposits from around the world, as revealed by a review and reevaluation of existing paleomagnetic literature. This review represents a first attempt at providing a synopsis of hard high-grade hematite deposits within a temporal framework. The apparent Paleoproterozoic to Mesozoic age distribution of deposits as discussed in this review, which must be tested and verified by both the expansion of the database and improvement of current available data, has important implications for proposed models of ore genesis, as well as for exploration.

Abstract Iron enrichment in banded iron formation (BIF)-hosted high-grade iron deposits is the final result of sequential removal or replacement of gangue minerals from the host by hydrothermal and supergene processes. Apart from the presence of the host BIF, structure is the most important control on the location of these deposits. Also, the distinct structural setup of the mineralizing environment results in iron ore of distinct textural features and consequently variable physical properties. In the Hamersley province of Western Australia pre-Upper Wyloo Group extensional faults are most often associated with high-grade hematite deposits in the Paleoproterozoic Brockman Iron Formation. The most important faults provide a fluid pathway between underlying dolomites of the Wittenoom Formation, through a sequence of shales and cherts, and into the overlying BIF. Iron ore in the Kaapvaal province of South Africa is hosted within BIFs of similar age to the Pilbara craton. The BIFs in the Kaapvaal province rest directly on dolomite, and Paleoproterozoic karst structures form the main spatial control on the high-grade iron ore. In contrast, low-angle thrust faults are the principal structural control on large deposits in the Marra Mamba BIF in the Hamersley province. These structures provided a more effective fluid pathway between the BIF and the overlying dolomites. A very similar structural scenario controls the very large Paleoproterozoic iron deposits in the Quadrilátero Ferrífero province in Brazil, although individual deposits are often highly complex due to postmineralization deformation during the Brasiliano orogeny. Structural reconstruction suggests that early structures, particularly thrust faults and tight folds that link a potential fluid source such as the dolomites of the Gandarela Formation with the underlying BIFs, form the most important control on ore formation in this province. Iron deposits hosted by Archean BIFs are less well understood. In the Carajás province of Brazil, fluids derived from granitoid intrusions are interpreted to have caused the initial hypogene alteration of the BIF which later focused the supergene ore fluids that led to high-grade hematite formation. Major structures that linked these granitoids with the BIF were crucial in the formation of the protores. In all these districts, mineralizing structures are those that provided the most effective link between a source of hydrothermal, silica-undersaturated fluids and iron formation, or allowed the influx of surface-derived meteoric waters to control the sites of ore formation in the BIF. Another important effect of structures is that they locally caused a differential pressure gradient during deformation and concentrated fluids into low-strain or dilational sites of iron ore formation. Most high-grade iron deposits formed close to (paleo)-unconformity surfaces and are, therefore, prone to rapid erosion. The structural setting can play a major role in preservation of these deposits. Ore deposits near normal faults in extensional grabens and karst structures are particularly favorable to ore preservation because the faults usually caused downthrow of the mineralized zones and burial by younger sediments. Compressional structures such as thrusts were far less favorable, because they usually caused uplift and erosion of the orebodies within them. Orebodies controlled by these structures require postmineralization preservation events, such as a major postore orogeny, or formed relatively recently, and therefore erosion did not progress far enough to erode them.

Abstract Hydrothermal alteration in structurally controlled, high-grade banded iron formation (BIF)-related iron deposits at Carajás (Brazil), Hamersley (Australia), and Thabazimbi and the Zeekoebaart prospect (South Africa) exhibit significant similarities and differences in geologic setting and hypogene alteration. In Carajás, Paleoproterozoic hematite deposits are hosted in low-metamorphic grade Archean jaspilites that are encased in metabasalts. The Paleoproterozoic BIF-hosted deposits of the Hamersley district, the Thabazimbi deposit, and the Zeekoebaart prospect are surrounded by shales. At Carajás, the hydrothermal alteration of jaspilites is characterized by a distal alteration zone with magnetite-calcite-quartz-pyrite where the primary microcrystalline hematite → magnetite (±kenomagnetite). The intermediate alteration zone consists of martite-microplaty hematite-quartz with magnetite → martite, whereas the proximal alteration zone contains hematite ± carbonate ± quartz with martite → microlamellar hematite → anhedral hematite → euhedral-tabular hematite. The proximal alteration zone represents the high-grade ore (i.e., porous hard to soft and hard ores). Hydrothermal alteration also affected mafic wall rocks with chlorite-quartz-carbonate ± hematite in distal alteration zones, and chlorite-hematite-quartz-albite-mica-carbonate ± titanite ± magnetite ± sulfides and hematite-chlorite-quartz-albite-mica-carbonate ± titanite ± magnetite ± sulfides in intermediate and proximal alteration zones, respectively. At the Mount Tom Price deposit in the Hamersley district, three spatially and compositionally distinct hydrothermal alteration zones are distinguishable: (1) distal magnetite-siderite-iron silicate, where the shape of the magnetite is suggestive of it being pseudomorphous after preexisting minerals, likely siderite; (2) intermediate hematite-ankerite-magnetite, with euhedral and bladed magnetite showing minor replacement by martite along crystal boundaries and replacement of iron-silicates by anhedral and microplaty hematite; and (3) proximal martite-microplaty hematite zones, where carbonate is removed. Martite and anhedral hematite replace magnetite and iron silicates of the intermediate alteration assemblage, respectively. The Thabazimbi deposit and the Zeekoebaart prospect lack unequivocal evidence for the formation of paragenetically early hydrothermal magnetite. Chert in ore zones has been replaced by microplaty hematite or has been leached, giving rise to porosity. Veins contain coarse tabular hematite and coarse crystalline quartz. High-grade hematite-martite orebodies are the result of SiO 2 leaching and associated volume loss that created widespread brecciation of the high-grade hematite ore. In addition to high-grade hematite-martite ores, four mineralogically distinct types of iron ore have been recognized: (1) goethite-rich, (2) low-grade dolomite-hematite, (3) low-grade calcite-hematite, and (4) talc-hematite. The comparison of hydrothermal alteration characteristics in the three case study areas revealed: (1) a similar paragenetic sequence of iron oxides, marked by an abundance of open-space filling and replacement textures; (2) distinct lack of a penetrative fabric in alteration lithologic units and high-grade ores; and (3) the importance of porosity and brecciation to accommodate volume loss. Differences include: (1) the formation of carbonate in different hydrothermal alteration zones of each deposit; (2) the presence of stilpnomelane in BIF that is surrounded by shales and hosted in sedimentary basins but absence in BIF that is bounded by mafic rocks; (3) the presence of significant amount of siderite in distal alteration zone in the Hamersley deposits but absence in the Carajás and Thabazimbi deposits; (4) the presence of significant amount of sulfides in the Carajás deposits but absence in the Hamersley and Thabazimbi deposits; and (5) significant amounts of chlorite, talc, white mica, and albite in basalt-hosted iron ore deposits (e.g., Carajás) or mafic dikes that are spatially and temporally associated with iron mineralization (e.g., in the Hamersley province). The systematic documentation of hydrothermal-alteration minerals and assemblages has significant implications for the exploration of concealed high-grade iron orebodies, because key hydrothermal alteration minerals such as chlorite, talc, carbonates or iron silicates are an expression of the hydrothermal footprint of the BIF iron-ore mineral system and, therefore, can be used as mineral vectors.

Abstract A range of techniques has been applied to the mineralogical and microchemical characterization of high-grade iron ore hosted by banded iron formation (BIF), including reflectance spectroscopy, X-ray diffraction, Raman spectroscopy, scanning electron microscope, electron microprobe, and proton induced X-ray emission analysis (PIXE). These tools provide key physicochemical properties of the main ore minerals, such as magnetite, kenomagnetite, maghemite, hematite, and goethite, which in turn determine the grade of the deposit and its economic viability. For instance, current automated HyLogging™ systems, based on reflectance spectroscopy, provide quick and objective measurements of hematite, goethite, and gangue mineralogy on large volume of cores and drill chips. X-ray diffraction used on powders offers a full account of the bulk mineralogy of the sample as well as aluminum substitution in the structure of hematite, goethite, and maghemite. On the other hand, Raman spectroscopy provides in situ iron oxide mineralogy and cation substitution at the thin section scale. In situ microchemical analyses, using scanning electron microscopy, electron microprobe, and PIXE, emphasize the mineralogical relationship and distribution of deleterious elements such as P, Al, and Si that underpins the development of downstream processing methods for assessing upgradability and exploitation of iron ore deposits.

Abstract The whole-rock geochemistry of banded iron formation-hosted high-grade iron ores has long been ignored as a possible source of constraints on the physicochemical conditions of ore formation. In this contribution, available geochemical data, including major, trace, and rare earth element concentrations, from a selected number of high-grade hematite-martite deposits that represent supergene and hypogene ore-forming environments are collated. Geochemical data for high-grade iron ores are evaluated against the average composition of the BIF protolith, to gauge important trends of enrichment and depletion. Results reveal a generally very similar distribution of major and minor elements, irrespective of deposit type. The marked enrichment of iron is in all cases attributable to the effective removal of SiO 2 , MgO, CaO, as well as CO 2 . The often invoked immobility or even introduction of iron during high-grade iron ore formation is called into question by the observation that the increase in concentration of Al 2 O 3 exceeds that of iron in almost all deposits. Furthermore, the distribution of redox-sensitive elements, such as Mn and V, suggests that during the transformation from BIF to high-grade hematite-martite ore f o2 remained effectively buffered by the oxidation of magnetite to hematite. Distinct enrichment of certain trace elements holds the promise to establish geochemical fingerprints to distinguish high-grade iron ore deposit types of different origin. This applies in particular to supergene high-grade hematite-martite ores, which are characterized by distinctly elevated concentrations of Sr and Ba and the efficient fractionation of LREE from HREE. Hydrothermal, magmatic-hydrothermal and supergene-modified hydrothermal deposits, on the other hand, appear not to have unique geochemical fingerprints. Enrichment of trace metals is usually restricted to single deposits but nevertheless provides an indication that more thorough studies may yield meaningful geochemical signatures to also distinguish different types of hypogene hematite-martite deposits.

Abstract Finite difference modeling of fluid flow in response to topography, extensional collapse, and thermal structure has been employed to simulate processes leading to the genesis of Proterozoic iron ores, using input data appropriate to the Hamersley district of Western Australia and other iron ore districts. The geologic history and questions that provide the motivation for the modeling include the presence of a mountain range formed by pre-ore genesis convergent deformation, extensional collapse of that mountain range, and evidence at the deposits for two or more different fluid types, including a deep-seated (reduced) and a surface-derived (oxidized and 18 O-depleted) fluid. In terms of fluid-flow rates, topographically driven downward fluid flow is seen to be comparable to both deformation-driven flow and also to heating and/or basal overpressures for comparable permeability structures and mountains with elevations in excess of 1 km. During extensional deformation at geologically realistic strain rates, downward flow is created by the combination of dilation produced by deformation with the inability of the fluid always to flow quickly enough to account for the dilatant volume change, producing areas of fluid under pressure, particularly across permeability interfaces. This effect is most pronounced where extensional faults cut through low-permeability basement. Upward fluid flow of heated fluids, as has been proposed to initiate genesis of these giant iron ore deposits, can be achieved at the start of extensional deformation if the deep fluid is initially overpressured, for example, due to input of fluids derived from magmas or to heating and/or devolatilization deep in the system. This initial upward flow can produce substantial temperature anomalies at relatively shallow depths, particularly in the hanging wall of dipping faults. However, with time, the extension and topography drives cooler meteoric fluids downward, which competes with and then eventually swamps the initial upflow. This scenario matches the envisaged sequence of events at the major deposits of the Hamersley district and also explains how different deposits record different degrees of preservation of the early-formed high-temperature assemblages, depending on the extent to which later surface-derived fluids have utilized the same structures as the initial upflowing fluid. Questions remaining from this modeling, and in consideration of the geochemical and stable isotope data, relate to which of the fluids (or both) was largely responsible for silica dissolution and whether both deep-seated and shallow fluids are prerequisite ingredients for genesis of this ore type.

Abstract The Hamersley province of northwest Western Australia is one of the world's premier iron ore regions. The high-grade iron ore deposits are mostly hosted within banded iron formation (BIF) sequences of the Brockman and Marra Mamba Iron Formations of the Hamersley Group and consist of two types: martite-microplaty hematite containing between 60 and 68 wt percent Fe, and martite-goethite containing between 56 and 63 wt percent Fe. Examples of martite-microplaty hematite include Mount Whaleback, Mount Tom Price, and Paraburdoo and examples of martite-goethite ore deposits include Mining Area C (Area C), Hope Downs, and the Chichester Range. The high-grade martite-microplaty hematite ores, which formed in the Paleoproterozoic, have a three-stage origin. Stage 1 involved the release, from the underlying sedimentary successions, of low (110°C) to high (280°C) temperature, highly saline (20–25.5 wt % NaCl-CaCl 2 equiv; Ca > Na > K) basinal brines that interacted with the underlying Wittenoom Formation and moved upward in normal faults, such as the Southern Batter fault at Mount Tom Price, the 4E fault at Paraburdoo, and the Central and Eastern Footwall faults at Mount Whaleback, into the host BIF. The hypogene fluids migrated laterally within large-scale folds with permeability controlled by shale layers and northwest-trending dolerite dike sets. The BIF was laterally and vertically altered into magnetite-siderite-stilpnomelane and hematite-ankerite ± magnetite assemblages at Mount Tom Price, a hematite-dolomite-chlorite-pyrite assemblage at Paraburdoo, and formed a dolomite-chlorite assemblage in the Mount McRae Shale at Mount Whaleback. Stage 2 involved deeply circulating, low-temperature (<110°C), Na-rich meteoric waters that interacted with evaporites prior to their interaction with the BIF. The descending meteoric waters interacted with the carbonate-altered BIF to produce a martite-microplaty hematite-apatite assemblage prior to supergene alteration. Stage 3, the supergene stage during the Mesozoic to Tertiary, is the final stage in the transformation of BIF to high-grade ore. Shallow supergene fluids interacted with the martite-microplaty hematite-apatite assemblage to form a highly porous high-grade (>63 wt % Fe) martite-microplaty hematite ore. Supergene alteration is likely to have occurred for at least 80 m.y. and close to the present topographic surface. High-pressure (>0.10 wt %) martite-microplaty hematite assemblages can therefore form and may remain concealed beneath BIF, below Proterozoic erosion surfaces. The martite-goethite bedded orebodies resulted from late Mesozoic supergene alteration of BIF. During this process magnetite was oxidized to martite, whereas silicates and carbonates were oxidized and hydrated to goethite or leached without replacement. The controls on the localization of supergene martite-goethite deposits, for example, the Hope Downs, Cloud Break, and Area C deposits include preexisting structures, such as faults, thrusts, and folds. These structures acted as fluid conduits that directed descending supergene fluids into the host BIF. Dolerite dikes and shale layers further focused and controlled fluid flow. High iron grades at the Area C and Hope Downs deposits are associated with synclinal structures where increased supergene fluid flow caused multiple phases of goethite leaching, precipitation, and cementation. Microplaty hematite encompasses a variety of sizes, ranging from 20 to 300 μm, and textures, ranging from platy to tabular. Microplaty hematite is commonly associated with supergene-modified hydrothermal deposits but can also form in the hydration zone of supergene deposits. The phosphorus (P) in supergene and supergene-modified hydrothermal deposits was repeatedly remobilized by both hypogene and/or supergene fluids. The P distribution was controlled by several factors, such as fluid flux in fault zones, permeability of shale layers, and synclinal folds, which resulted in locally high concentrations (>0.10 wt %) of P in the deposits. It is unlikely that a single model for the formation of the martite-microplaty hematite ore deposits can explain all the structural, stratigraphic, hypogene alteration, and ore characteristics at the Mount Whaleback, Mount Tom Price, and Paraburdoo deposits. Continued collaborative research directed at elucidation of a single tectonic history of the Pilbara, based on collection of similar structural and geochemical data sets from these deposits, will advance genetic ore models and aid in exploration for concealed orebodies.

Abstract The Quadrilátero Ferrífero district, located on the southern portion of the San Francisco craton in Minas Gerais, Brazil, comprises Archean greenstone terranes of the Nova Lima Supergroup and the Paleoproterozoic cratonic cover sequences of the Minas Supergroup that consist of quartzites, metaconglomerates, phyllites, dolomites, and banded iron formations. The Minas Supergroup was affected by two orogenic events—the Paleoproterozoic Transamazonian-Mineiro (2.1–2.0 Ga) orogeny and the Neoproterozoic to Early Paleozoic Brasiliano-Araçuaí (0.65–0.50 Ga) orogeny, resulting in complex deformation and metamorphic grades that increase from greenschist facies in the West to amphibolite facies in the East. Metamorphosed iron formations, referred to as itabirites, are found in three compositionally distinct lithofacies, namely quartz itabirite, dolomitic itabirite, and amphibolitic itabirite; these lithofacies are host to a large number of economically important high-grade iron ore deposits that give rise to the name Quadrilátero Ferrífero, or "Iron Quadrangle." High-grade iron ores replace itabirites in tectonically favorable, low-strain sites. faults acted as conduits while large fold hinges were sinks for mineralizing fluids. Hard and fine-grained hematite and/or magnetite orebodies are in the western low-strain domain of the Quadrilátero Ferrífero. Subsequent deformation led to recrystallization and development of distinctly schistose high-grade hematite ores characteristic of the eastern high-strain domain. A combination of hypogene and geologically recent supergene processes is thus invoked to explain the formation of the high-grade iron ores of the Quadrilátero Ferrífero. Three stages of hypogene ore formation are distinguished. The first two of these stages took place early during the Transamazonian orogeny (2.1–2.0 Ga) and are well preserved in the western low-strain domain. During the first stage metamorphic fluids leached SiO 2 and carbonates and mobilized iron, which resulted in the formation of massive magnetite bodies, Fe oxide veins, and Fe-rich itabirite bodies; during the second stage, low-temperature, low-salinity fluids caused oxidation of magnetite and Fe-rich dolomite to hematite. The resulting ore is porous to massive and has a granoblastic fabric. The third and final hypogene stage of ore formation is related to thrusts of uncertain age (Transamazonian or Brasiliano orogeny), which dominate the tectonic structure of the eastern high-strain domain of the Quadrilátero Ferrífero. Crystallization of tabular hematite and large platy specularite crystals that overprint the preexisting granular fabric in the presence of high-salinity hydrothermal fluids are characteristic of this stage. During the Neogene, supergene residual enrichment processes gave rise to the formation of soft to friable hematite orebodies. The larger soft orebodies that surround some smaller hard high-grade orebodies are typically associated with dolomitic itabirite. Together, both ore types comprise the giant high-grade iron deposits typical for the Quadrilátero Ferrífero, resulting from the superposition of both hypogene and supergene processes. Pure supergene deposits are considerably smaller and do not extend to deeper levels below the erosion surface.