Plate tectonics caused the demise of banded iron formations.: History Edit
Subjects: Geology

Early models for the origin of banded iron formations (BIFs) assumed that under a reducing atmosphere the supergene anoxic hydrolysis of mafic silicate minerals would provide an abundance of ferrous iron in solution to be transported to the oceans, there to be oxidised and precipitated by locally produced oxygen in shallow seas. The sparse distribution of BIFs in the Archaean era, its greater abundance during the Palaeoproterozoic era and perceived absence thereafter was considered to be essentially linked to the concentration of O2 in the atmosphere and this was a key factor in hypotheses of atmospheric evolution from anoxic to oxidising. Several assumptions regarding the deposition of BIFs were used to infer atmospheric evolution; however, chemical considerations indicate that the deposition of BIF is independent of atmospheric oxygen levels. Ferrous iron is only soluble in acid solution and in anoxic conditions in natural waters is precipitated as ferrous hydroxide or by carbon dioxide as ferrous carbonate. Silica in solution
typically exists in equilibrium between ionic solution and colloidal suspension. Flocculation of colloidal silica is catalysed by the presence of cations and polymerises to a hydrophobic gel, thus removing silica from solution. It is highly unlikely that the oceans could ever have been the reservoir of iron and silica for the deposition of BIFs.
Modern interpretations consider BIFs as deep sea sediments with the source of the iron and silica derived from reactions between circulating sea water and hot mafic to ultramafic rocks producing hydrothermal systems venting onto the sea floor. The solubility of ferrous and ferric iron and silica is greatly increased at elevated temperatures and hydrothermal solutions would immediately precipitate iron hydroxide and iron silicates on quenching by cold seawater, even in the absence of ambient oxygen, to form hydrothermal plumes and mound deposits, subsequently resedimented by turbidity and density currents across the ocean floor. No transportation of ferrous iron in solution at ambient temperatures and no external source of O2 or Fe2+ is necessary and the deposition of BIFs was independent of atmospheric oxygen, biogenic processes and continental sources of dissolved ferrous iron and silica, although any or all of these may have been present during deposition.
A similar process occurs today with black smoker hydrothermal vents depositing large quantities of iron hydroxide and iron silicates on the ocean floor that will eventually be subducted beneath the continents by plate tectonic processes. During the Archaean era, shallow oceans and immature continents, preserved sea floor deposits as greenstone belts and in marginal sedimentary basins until the formation of large buoyant continents caused the subduction of deep sea ocean crust including BIF deposits. The temporal distribution of BIFs is thus related to the preservation of deep ocean sedimentary rocks that since the onset of modern style plate tectonics in the late Archaean to Proterozoic eras have largely been destroyed by subduction of the ocean floor beneath the continents.

Introduction
Banded iron formations (BIFs) are very important to civilisation as the major source of iron ore in support of the steel industry, but their conditions of formation and the origin of their component elements are poorly understood and still widely debated. Banded iron formation  is defined as a sedimentary rock containing >15 wt%Fe and consisting of alternating cm-scale iron-rich and iron-poor bands (James, 1954) typically consisting of magnetite and micro-quartz with accessory iron silicate and carbonate. Typically found only in Precambrian rocks (with rare exceptions) and with no apparent modern analogues, numerous hypotheses based on various assumptions have been presented regarding their mode of deposition and particularly the origin of the vast quantities of iron that they contain. Because they are typically finely laminated with no evidence of detrital
particles it was assumed that they were chemical sediments and the iron and silica was precipitated directly from seawater. It was assumed that in the absence of free oxygen the oceans formed a reservoir of ferrous iron in solution from anoxic sub aerial weathering, and that oxygen from photolysis (Braterman and Cairns-Smith,1987) and/or photosynthesis was required to precipitate the ferrous iron as the insoluble ferric hydroxide in BIFs; essentially linking the deposition of BIFs to the removal of traces of oxygen from the atmosphere. Biogenic origin through the activity of iron metabolising bacteria has also been proposed as the agent of precipitation of ferrous iron as ferric hydroxide (La Berge, 1973; Konhauser et al., 2002; Kappler et al., 2005). 
Although common constituents of Archaean supracrustal sequences (Algoma type) and forming massive deposits in Palaeoproterozoic continental sedimentary basins (Superior type) they apparently ceased to form after 1.8 Gy ago apart from a resurgence in the Neoproterozoic era and other rare occurrences (Fig. 1A). Banded iron
formation typically consists almost entirely of iron oxides (predominantly magnetite) and silica in the form of quartz or iron silicates with less than 1 wt% Al2O3, and less than
2 wt% total CaO and MgO. With no apparent detrital textures and no modern analogues to show how they formed, models were proposed based on very rare sedimentary structures and a number of assumptions based on a poor understanding of the geochemistry of iron and silica. The absence of detrital particles and recognised sedimentation textures suggested chemical precipitation and it was assumed that ferrous iron in solution was precipitated by traces of O2 and silica was precipitated by evaporative  concentration. The need for oxygen produced by photolysis of water or by organic photosynthesis requiring sunlight together with evaporative precipitation of silica inferred shallow water deposition. This model was then used to form further models of the early history of the Earth, especially the development of the oxygenated atmosphere and the Great Oxidation Event.
Oxygen (O2) is the major volumetric constituent of planetary mantles and crusts in combination with silica and other elements but free oxygen is typically absent from  planetary atmospheres. The unique abundance of free O2 in the Earth’s atmosphere is due to the photosynthetic production of carbohydrates and O2 from
water and carbon dioxide by living organisms and implies that until photosynthetic organisms evolved the early Earth’s atmosphere was virtually devoid of O2 (Cloud, 1973; Holland, 1973, 1984, 2002, 2006; Kasting, 1993; Shaw, 2008). The implication of an anoxic early atmosphere is supported by mass independent fractionation of sulphur (MIS, Fig. 1B)) isotopes (Farquhar et al., 2001, 2007; Pavlov and Kasting, 2002; Mojzsis et al., 2003), iron,molybdenum and carbon isotopic ratios in sediments (Grassineau et al., 2001;  Siebert et al., 2005) and the absence of red beds, evaporites (sulphate) and phosphorites prior to ∼2.5 Ga (Cook and Shergold, 1986; Beukes et al., 2002a; Melezhik et al., 2005) but a key factor in the origin of the Great Oxidation Event (GOE) hypothesis was the distribution of BIF over time (Garrels et al., 1973; James, 1983; Towe, 1983; Derry and Jacobsen, 1990; Kump and Holland, 1992; Law et al., 2002). Some other evidence proposed to support the anoxic early atmosphere and hydrosphere has been disputed and a number of local instances of apparently oxidising conditions have been published; i.e. conflicting evidence of oxidised minerals in Archaean sedimentary rocks (Ohmoto, 1993, 2003; Huston and Logan, 2004; Hoashi et al., 2009;Kato et al., 2009; Czaja et al., 2012) in contrast with unoxidised magnetite, pyrite and uraninite detrital grains in sedimentary
rocks of Archaean to Palaeoproterozoic age (Roscoe, 1968; Grandstaff, 1980; Rasmussen and Buick, 1999; England et al., 2002). The presence of high field strength elements (Zr, Y, REE, Hf, Ta, Th and U; Derry and Jacobsen, 1990; Bau and Alexander, 2009) have been used as indicators of anoxic weathering, erosion and deposition environments but the significance of these factors to the redox value of seawater has been disputed (Graf, 1978). The leaching of iron as Fe2z from palaeosols was also considered as evidence of anoxic weathering (Rye and Holland, 1998) but the leaching of iron below the ferruginous layer is an integral feature of lateritic soil profiles and is not confined to anoxic weathering (Beukes et al., 2002b). With the rise of atmospheric oxygen, the formation of continents rising above sea level and the onset of subaerial weathering, the oxidation of pyrite would then contribute sulphate ions to the oceans and the absence of BIFs after the GOE has been attributed to the replacement of BIFs by sedex sulphide deposition in
euxinic deep oceans (Canfield, 1998). Mesoproterozoic sedex sulphide deposits are commonly associated with BIFs (Stanton, 1972, 1976) or extensive iron oxides and silicates (Lambert, 1976) but typically the deposits occur in sedimentary basins preserved by orogenic uplift, whereas the extensive contemporaneous sea floor BIFs would be destroyed by subduction.
In the absence of silica extracting organisms until the Phanerozoic era it was assumed that silica would accumulate in the oceans until precipitated by evaporative concentration (Trendall and Blockley, 1970; Morris, 1985, 1993) also pointing towards shallow water deposition (Morris and Horwitz, 1983; Garrels, 1987). The
reactions of silica with water are highly complex (Iler, 1979). The oft-quoted solubility of quartz at ∼8 ppm and the solubility of chalcedony at ∼125 ppm (Siever, 1957)
are measurements of solution in pure water, and do not consider the effects of cations in seawater, equilibrium between solutes and colloids, dispersion of amorphous silica, colloidal miscibility with water or the aging of colloidal silica in water. Deposition of silica is typically the result of polymerisation of SiO2.nH2O to form a hydrophobic gel and subsequently to solidify by syneresis and is a product of aging or catalysis independent of concentration.
Later models propose hydrothermal fluids leaching iron and silica from heated mafic rocks (Seyfried and Janecky, 1985) and then venting into the ocean as the origin for much of the iron and silica in BIFs (Isley, 1995; Isley and Abbott, 1999; Abbott and Isley, 2001), but still supported the oceans as a reservoir of dissolved ferrous iron and still required an oxygenated surface layer in the ocean to precipitate dissolved ferrous iron as ferric hydroxide. The hydrothermal model was extended by Lascelles (2007) proposing immediate precipitation of colloidal (ferrous) iron hydroxide and hydrous iron silicate as the hydrothermal fluids were quenched by cold seawater (cf. black smokers). The primary unsorted precipitate, consisting of water-saturated mounds of colloidal iron hydroxide, hydrous iron silicate (Severmann et al., 2004) and accessory iron carbonate and sulphide, was then re-sedimented by turbidity currents (Barrett and Fralick, 1985, 1989;Krapez et al., 2003), forming granular iron formations (GIFs) at intermediate depths, and
density currents to produce widespread finely laminated BIFs on the sea floor (Lascelles, 2007). The episodic nature and rapid deposition of turbidity and density currents,  lasting a few hours to days with a hiatus between flows, is in direct contrast to the slow deposition of annual micro-laminations over millions of years attributed to earlier models (Trendall and Blockley, 1970; Morris, 1985, 1993; Garrels, 1987).

Many authors have proposed a biogenic process for the precipitation of iron oxides in BIFs either by iron metabolising bacteria or by photosynthetically produced oxygen (La Berge, 1973; Hartman, 1984; Konhauser et al., 2002; Kappler et al., 2005; Posth et al., 2008; Czaja et al., 2013) but do not show any evidence for the process from the actual rocks. Traces of bacteria and possibly cyanobacteria have been reported (Brocks et al., 2003) but their extreme rarity does not support an overall biogenic origin for BIFs and suggests only minor modification by anaerobic iron and sulphur metabolising extremophiles (Wacey et al., 2011). Present day subsea hydrothermal vents have abundant organisms including exremophiles encrusting the vicinity of black smokers and hot springs but the precipitation of the iron oxides and silicates comes directly from hot acidic (pH ∼5, Butterfield et al., 1997) hydrothermal fluids contacting the cold neutral water of the ocean and not from biological processes. 

BIFs and the evolution of the atmosphere
Current models for the origin of BIFs still assume that iron and silica were alternately precipitated directly from sea water and to be wholly dependent on the presence of free oxygen to precipitate ferrous iron (Fe2+) from  solution as the insoluble ferric form (Fe3+) in an anoxic ocean (Kappler et al., 2005). Degens and Stoffers (1976) describe  pycnoclines in the Black Sea and some African lakes separating oxic surface water from euxinic deeper water and the sediments relating to the different environments, and compared these sediments to deep sea cores and sequences through geologic time. Later authors (Morris and Horwitz, 1983; Abbott and Isley, 2001; Tsikos et al., 2010)  proposed a model whereby the surface of the oceans became oxidised while the deep oceans remained anoxic. Evidence from REE, oxygen and iron isotopes showed a volcanogenic component to BIFs (Bolhar et al., 2005; Van den Boom et al., 2007) and an additional hydrothermal origin for the deep ocean reservoir of iron was proposed (Isley, 1995) with precipitation as ferric hydroxide during periods of upwelling or overturn of the stratification. With the earliest known BIFs dated at ca 3?75 Ga and numerous deposits during the rest of the Archaean era, this was used to imply the early episodic formation of an oxic surface layer in the ocean (by photolysis, Braterman and Cairns-Smith, 1987 or photosynthesis, Czaja et al., 2012).
Although a large proportion of the early oxygen was sequestered by weathering of both subaerial and submarine mafic rocks a slow but exponential build-up of oxygen  occurred in the atmosphere. The action of ultraviolet radiation (UV) and electrical discharges splitting traces of diatomic oxygen molecules (Atkins et al., 2010) in the  atmosphere leads to the formation of ozone (O3) that blocks further penetration of biological damaging UV to the surface. This reduction of UV penetration to the oceans enlarged the habitable region, originally limited to benthic organisms in a narrow zone beyond the reach of UV radiation, and simultaneously increased the available light for photosynthesis allowing accelerated growth of photosynthetic organisms and more production of oxygen. The earliest appearance of photosynthetic organisms is not well constrained with estimates ranging from 3.4 to 2.8 Ga based on the interpretation of early stromatolites as being formed by photosynthesising cyanobacteria or anaerobic extremophiles. The first clear evidence of photosynthetic organisms appears to be ca 2.7 Ga (Shaw, 2008) but most of the O2 produced as a by-product would be sequestered by
seafloor weathering as well as subaerial weathering (Towe, 1983) with minor diffusion into the atmosphere, leading to an initially slow but exponential rise of free O2 in the atmosphere and ocean and consequently ozone in  the atmosphere, reducing the penetration of harmful UV radiation thus allowing the development of photosynthetic
organisms. With the exponential growth of photosynthesis in the late Archaean era, the abundance of BIFs increased culminating in the very large deposits in the Palaeoproterozoic era (Fig. 1A) after which the presence of an oxygenated atmosphere inhibited the supply of ferrous iron to the ocean thus terminating the formation of BIFs.

Assumptions underlying current models of BIF deposition
Rationalising existing models by interpreting evidence according to the model is a common tendency in the literature (Wunsch, 2010) but scientific method requires deliberate testing of existing models, reassessment as new evidence is discovered and in particular reexamination of the assumptions underlying the model. 

1st assumption – dissolved ferrous iron accumulated in anoxic seawater until precipitated by free oxygen as insoluble ferric hydroxide
With the exception of some organic complexes ferric iron compounds are virtually insoluble in natural waters at ambient temperatures and pressures but so too are ferrous sulphide, ferrous carbonate (compounds that are found in abundance in natural sediments) and ferrous hydroxide (Garrels and Christ, 1965). The assumed very low level of O2 in the early ocean and atmosphere was believed to lead to a build-up of ferrous iron in solution from the anoxic weathering of surface rocks with the corollary that oxidation is required to precipitate the ferrous iron as the insoluble ferric hydroxide. The paucity of terrigenous sediments apart from immature volcanogenic tuffs throughout the Palaeo- to Meso-Archaean era suggests very little subaerial outcrop to produce large quantities of ferrous iron by erosion and weathering but hydrolysis of silicate minerals comprising
the ocean floor was suggested as a possible source. 
However, ferrous iron is only stable in acid solution and on neutralisation precipitates as ferrous hydroxide. Ferrous iron also reacts with silica and alumina in solution
or suspension to form authigenic iron-rich clay minerals leading to the rapid removal of iron from the oceans even in the absence of oxidising conditions. Ferrous bicarbonate is soluble in the presence of H2CO3 in slightly acid solution (Barnett and Wilson, 1957) provided Eh values are low due to absorption of oxygen by decaying organic matter. High concentrations of atmospheric CO2 could facilitate the transport of Fe2+ in freshwater as Fe(HCO3)2 and humic Fe complexes derived from the decay of soil bacteria may also have been a significant transporter of iron to the oceans; however, ferrous bicarbonate is not stable in seawater (pH 8.1 for modern seawater; Millero, 2004) and ferrous carbonate is precipitated. Even assuming high concentrations of CO2 in the Archaean atmosphere, the presence of Ca and Mg ions in solution would make the water alkaline (Garrels and Christ, 1965) and there is no evidence that Archaean oceans could have been more than very slightly acidic (pH>6.5) due to dissolved CO2 and it seems highly
unlikely therefore that the oceans could have contained large quantities of ferrous iron in solution.

2nd assumption – in the absence of silica metabolising organisms, silica would accumulate in seawater until precipitated by evaporation
The solubility of silica (SiO2) in water is highly complex forming an equilibrium state between colloidal suspension (SiO2.nH2O) of hydrophilic silica that is completely
miscible with pure water and ionic solution (H3SiO4-;  Iler, 1979).  Measurements of the solubility of quartz in pure water (Siever, 1957) were frequently quoted as representative of the solubility of silica, particularly with regard to supergene enrichment models (Morris, 1980, 1983, 1985; Trendall and Blockley, 1970) but whereas quartz has a very stable crystalline hydrophobic structure that resists dispersion, amorphous silica is readily dispersed as a hydrophilic colloid in pure water. However, colloidal silica is rapidly  polymerised by the presence of ions in solution to form hydrophobic gels and also reacts with cations in solution to form hydrous silicates (clay minerals) and compounds such as magardiite. Thus the solubility of amorphous silica (SiO2) in water varies with the presence of cations and the aggregation state and degree of hydration of the silica molecule. Saltwater induces polymerisation and precipitation of silica in the form of chert preventing the accumulation of dissolved silica in the oceans.

3rd assumption – BIF was only deposited in shallow water where sunlight enabled photolytic or photosynthetic production of O2
Much of the evidence given for supporting shallow sedimentary environments and subaerial exposure is inconcl usive because ripple marks, convoluted bedding and  laminations are also typical of turbidites, and the presence of turbidite deposits associated with iron formations is well documented (Krapez et al., 2003; Lascelles,
2007). Desiccation features attributed to subaerial exposure of the newly deposited sediment are typically syneresis cracks in chert bands and are not seen in silicate or oxide
(mud) layers and the interpreted raindrop pits on bedding planes are probably weathered out small spherical calcareous concretions (cf. bed of holes in the Brockman Formation, e.g. Trendall and Blockley, 1970). The main reason for the assumption of shallow water was the requirement for photolytically derived O2 to precipitate the
dissolved ferrous iron as ferric hydroxide (Morris and Horwitz, 1983).
The solubility of gases in water in contact with the atmosphere is proportional to the partial pressure of the gas in air; i.e. at zero partial pressure O2 is insoluble and biogenic O2 diffused into the atmosphere (Shaw, 2008) prohibiting the co-existence of an oxic layer at the ocean surface and an anoxic atmosphere. The inconsistencies of an oxic surface layer co-existing with an anoxic atmosphere for ∼1.5 Ga and the improbability of solutions of ferrous iron derived by subaerial weathering passing through this oxidising surface water to reach the deep ocean were never explained (Morris and Horwitz, 1983; Garrels, 1987; Morris, 1993; Tsikos et al., 2010). Absence of O2 also means the absence of O3 and the uninterrupted passage of life-inhibiting UV through the atmosphere and into the upper surface of the oceans (Farquhar et al., 2007). The earliest forms of life on Earth  appear to have been prokaryotes (Wacey et al., 2011) similar to modern extremophiles depending on volcanic heat and oxidation/reduction processes (e.g. Fe2+ to Fe3+) for their energy and early prokaryotes were able to survive in the depths of the oceans away from light. Traces of oxygen may have been produced by photolysis of water during the early Archaean era but the evolution of organisms capable of photosynthesis and producing O2 as a by-product was the major development leading to an oxic atmosphere.

4th assumption – the volume and temporal distribution of BIF observed today reflects BIF deposition
Numerous charts have been published (Gole and Klein, 1981; James, 1983; Abbott and Isley, 2001; Huston and Logan, 2004; Holland, 2006), showing a steady increase in volume of BIFs throughout the Archaean Era and culminating in the large BIF deposits of the Palaeoproterozoic era and their subsequent absence except for a burst of deposition during the Neoproterozoic era (Fig. 1A). This temporal distribution of iron formations was a key factor in the formulation of models of atmospheric (and hydrosphere) evolution and  the GOE (Cloud, 1973).
The oldest supracrustal rocks at ca 3.8 Ga in the Isua greenstone belt contain BIFs but are from a single region, in West Greenland and Canada (Appel, 1987) and there is a long gap until the next oldest supracrustal rocks at ca 3.45 Ga in the Pilbara region of Western Australia and the Barberton Region of South Africa (Fig. 1A). From 3.3 to 3.0 Ga the number of occurrences of greenstone belts and BIF increases with examples on several of the larger cratons and from 3.0 to 2.5 Ga almost every craton contains greenstone belts and BIFs (Huston and Logan, 2004). The apparent increasing abundance of BIFs during the Archaean is a function of the preservation of greenstone belts; the oldest rocks have very limited surface exposure and the area of younger rocks increases exponentially (Fig. 1C; Wilkinson et al., 2009). Greenstone belts are typically deformed and steeply dipping but stratigraphic correlation (Watkins and Hickman, 1990) and reconstruction of the greenstone belts suggests that they may have formed as ocean-wide sea floor deposits rather than in narrow troughs or rifts as previously assumed and their current topology and limited distribution is due to anatexis, tectonic dismemberment and erosion. Furthermore, while greenstone belts in some areas (e.g. North America, South Africa, Brazil and Australia) have been studied in detail and their stratigraphy and chronology is fairly well established, the data on others is sketchy to non-existent. After ca 2?5 Ga the classic greenstone belts are virtually absent from the geological record only occurring as thin slivers, commonly with interbedded BIFs, in orogenic regions (O’Rourke, 1961).
BIF associated with greenstone belts is classified as Algoma type (Gross, 1983) but the greatest thickness of BIFs is found in Superior type relatively undeformed continental margin sedimentary basins around the time of the Archaean-Proterozoic boundary that have unconformable contacts on granite-greenstone terrains. Surface outcrop of Algoma BIF is greatly limited due to tectonic dismemberment and typically steep dip whereas Superior BIF commonly covers wide areas with shallow dipping to sub-horizontal strata giving a disproportional abundance of outcropping BIF. Superior BIF also tends to show a much greater thickness of BIF units but nevertheless the apparent sharp peak in production of BIF ca 2.6–2.4 Ga, and in GIF from 2.5 to 1.8 Ga is largely due to preservation and area of outcrop and it is proposed that the apparent sharp reduction in BIF after ∼1.8 Ga, (Cloud, 1973; James, 1983; Towe, 1983; Holland, 1984; Kump and Holland, 1992;Morris, 1993) is also a function of preservation instead of production, as numerous minor occurrences of BIFs are found after 1.8 Ga (O’Rourke, 1961; Stanton, 1972, 1976) including a major series of Rapitan type deposits in the Neoproterozoic era (Ilyin, 2009).
Although less common, particularly if compared with the major Palaeoproterozoic basins, BIF is by no means absent from younger formations. Oxidative weathering after the GOE of sulphide minerals supplying sulphate to the oceans is reported as replacing the precipitation of ferric oxide with sulphide as an explanation for the purported absence of BIFs after 1.8 Ga; however, the association of BIFs with Mesoproterozoic VHMS and sedex deposits such as Broken Hill is well documented, although typically overshadowed by the interest in the sulphide ore (Stanton, 1972, 1976; Slack et al., 2007): nor can the volume of VHMS deposits remotely approach the volume of BIFs in the Palaeoproterozoic
era. The Neoproterozoic Rapitan type deposits appear to range in age from 750 to 550 Ma (Young, 1976; Klein and Beukes, 1993; Klein and Ladeira, 2004; Ilyin, 2009) and though commonly described as containing hematite instead of magnetite this is not always the case (Whitten, 1970). BIF is also recorded from the Palaeozoic era (O’Rourke, 1961;  Quade, 1976) and large deposits of iron oxides and silicates are forming at the present day in the deep ocean basins (James, 1969; Müller and Förstner, 1973; Butuzova et al., 1990; Bennett et al., 2008; Sun et al., 2012) as well as minor occurrences of the Lahn-Dill type (Butuzova, 1969; Krautner, 1977). It seems highly unlikely that the hydrothermal systems associated with oceanic spreading ridges and mantle plumes ceased in the Palaeoproterozoic era and suddenly resumed at the present day, and far more probable
that they have existed throughout geological history and that BIF was also deposited throughout geological history on the deep ocean floor.
Granular iron formations first appear in quantity in continental basins ca 2.5Ga with maximum deposition ca 1.9–1.8 Ga and commonly contain associated BIF facies.
The paucity of GIF (Superior type) in the Archaean could also be due to the lack of preservation as minor
occurrences are reported from Canada and Australia (Barrett and Fralick, 1985, 1989; Lascelles, 2014).

BIF sedimentology
Banded iron formation is characterised by alternating iron-rich and iron-poor laminations and fine bands without discernible detrital grains deposited as submicroscopic to colloidal particles showing few sedimentary features apart from the banding. Current and ripple bedding is rarely observed and slump structures and intraformational folding are visible in some formations (Lascelles, 2006, 2007, 2014) consistent with deposition from density currents and hydrothermal plumes. There is increasing evidence that BIF was deposited as deep sea sediments (Barrett and Fralick, 1985; Krapez et al., 2003; Pickard et al., 2004; Lascelles, 2007) in oceanic basins at depths >200 m with the source of the iron and silica derived from reactions between circulating sea water and hot mafic to ultramafic rocks producing hydrothermal systems venting onto the sea floor similar to present day deep sea ‘smokers’ (Isley, 1995; Dekov et al., 2007; Lascelles, 2007; Sun et al., 2012). It was proposed that the chert bands in BIFs derived from the diagenetic dissociation of Al-poor nontronite into iron hydroxide and amorphous silica (Lascelles, 2007) and that the ferruginous sediments associated with sea floor hydrothermal vents on the modern sea floor are the progenitors of future BIFs. The association of BIFs with pillow lavas that show no evidence of formation at depths similar to modern ocean basins (McCall, 2010) is compatible with comparatively shallow ocean basins in the Archaean era and does not conflict with greenstone belts being remnants of ocean crust or an ocean floor deposition of BIFs.
The solubility of silica and both forms of iron is markedly increased by increases of temperature and pressure (Garrels and Christ, 1965) and high temperature fluids are thus capable of extracting and transporting large quantities of silica and ferrous iron (together with minor ferric iron; McGuire et al., 1989) from mafic rocks (Seyfried and Janecky, 1985). Venting of high temperature fluids into cold neutral seawater causes rapid quenching and supersaturation leading to immediate precipitation of colloidal particles of ferrous hydroxide and hydrous ferrous silicate (equation (1); Dekov et al., 2007, Decarreau et al., 2008; Sun et al., 2012) forming unstable water saturated mounds of colloidal iron hydroxide and hydrous iron silicate surrounding the vents and hydrothermal plumes of fine colloidal particles depositing BIFs across the ocean basin.
          6Fe2++8SiO2+2H2O→Fe6Si8O18(OH)4
                                                      nontronite                                 (1)
Ferrous hydroxSlumping of the mound deposits causes the precipitates to be resedimented by turbidity currents with coarse particles forming GIFs at the foot of the slope, and density currents containing colloidal particles widely distributing finely laminated BIFs over the ocean floor (Barrett and Fralick, 1985; Krapez et al., 2003; Pickard et al., 2004; Lascelles, 2007). Both density currents and hydrothermal plumes are intimately mixed with sea water and the colloidal particles thus carry the isotopic signatures of both hydrothermal fluids and ambient sea water. Thus no external source of O2, continental sources of dissolved ferrous iron and silica or biological agency is necessary for the deposition of BIFs and although any of these may have been present they were not the cause of deposition and no free O2 is required for the transformation of ferrous hydroxide into magnetite during diagenesis as ferrous hydroxide ionises water to form magnetite with the evolution of H2 (equation (2), Shipko and Douglas, 1956) and ferrosoferric hydroxide simply dehydrates to magnetite and goethite (equation (3)) during diagenesis                                                                                                                                                                  3Fe(OH)2       →       Fe3O4  +  H+  2H2
 ferrous hydroxide    magnetite                                          (2)
       2(Fe3+ , Fe2+)(OH)5      →     Fe3O4    +   FeOOH + 2H2O                                                                                                                                                                                                                                                  ferrosoferric hydroxide      magnetite   goethite             (3)

The episodic nature and rapid deposition of turbidity and density currents, lasting a few hours to days with a hiatus between flows, during which normal pelagic sediments accumulate, is in direct contrast to the deposition of annual microbands over millions of years attributed to earlier models (Trendall and Blockley, 1970; Garrels, 1987; Posth et al., 2008). Modern ferruginous sediments are precipitated in an O2-rich environment and consist largely of colloidal goethite and Al-poor Fe3+ nontronite with relatively minor sulphides concentrated around the vent (cf. VHMS deposits) and little or no chert (Dekov et al., 2007; Sun et al., 2012). Unfortunately, no BIF has been identified to date since
sampling has either not gone deep enough into the sediment to show the effects of diagenesis or the act of deep core sampling would probably disperse amorphous or  gelatinous silica bands before recovery. The origin of GIF is less clear; some deposits are clearly turbidites (Lascelles, 2007, 2014) but others, particularly oolitic GIF may be outer shelf deposits similar to Phanerozoic oolitic ironstones (Young, 1989). It is proposed that marine iron formations should be grouped in four classes according to mode and place of deposition:                                                                                                                                                                                                                                                                                                                                                   (i) banded iron formations – colloidal density current deposits covering the deep sea floor
     (ii) granular iron formations – turbidity current deposits on or near mid ocean ridges or continental slopes with limited extension on deep sea floor
    (iii) oolitic iron formations – chemical precipitates on outer continental shelves
    (iv) Lahn-Dill iron formations – associated with volcanoes in orogenic belts; both granular and banded iron formations may be present.

Evolution of the Earth’s crust
The early history of the Earth is still largely speculative but it is assumed that agglomeration of rocky planetismals ca 4.5 Ga produced a molten globe with a dense metallic core and a mafic silicate mantle (Dehant et al., 2012). The initially high surface temperature >1500oC prevented retention of volatiles until the surface cooled by radiating heat into space to form a solid crust. It has been inferred that the Late Heavy Bombardment (LHB) that affected the moon at ca 4.2–3.8 Ga was even more intense on the Earth due to its greater gravitational potential and caused remelting of the Earth’s surface since no trace of the LHB remains on Earth (Albarede et al., 2013) and it has also been suggested that the LHB contained a high proportion of volatile cometary material that was retained to form an early atmosphere and ocean. Van Kranendonk et al. (2013) suggest the oldest granite gneiss indicates the local existence of a thin sialic crust that may have been produced by partial melting of the mantle by the LHB.                                                                                As the Earth continued to cool, lnterior pressure instabilities with hot spots in the mantle due to random distribution of heat-producing radioactive elements caused the development of plumes of hot upwelling mantle material (Moore and Webb, 2013) forming domes at the surface. As the upwelling hot mantle material approached the surface the reduction in pressure caused partial melting of the less refractive elements forming magmas that erupted mainly as pillow lavas onto the surface of sialic crust or mantle forming a ∼10 km thick crust (cf. Hawaii) with volatile elements degassing to add to the thin early atmosphere and hydrosphere (Figs. 1 and 2). The result was a shallow ocean covering the whole of the Earth with minor subaerial land above upwelling mantle plumes and possibly small areas of sialic crust providing rare clastic sediments (Dunlop and Buick, 1981; Erikson, 1999; McCall, 2010). Seawater percolated into the hot lava through gaps between pillows and scoria hydrolysing silicate minerals and releasing cations and silica into hydrothermal systems (Seyfried and Janecky, 1985) that were rapidly quenched on venting into the ocean, precipitating iron oxides and silicates to form BIFs covering both the new crust and the exposed mantle sea floor.
The denser residue from partial melting was pushed replacing the mantle plume at depth (Moore and Webb, 2013). However, the lighter surface crust of solidified lava flows, accumulated sediments and BIFs remained at the surface forming the oceanic crust. Small and localised at first, motion of the mantle was predominantly vertical with  essentially immobile sialic cratons but, as the mantle differentiated due to partial melting, convection cells formed and the number and size of convection cells increased with corresponding increase of sialic cratons and increasing lateral motion of the upper mantle was transferred to the crust. Interference between adjacent new ocean crust caused the newly formed crust to deform as a unit producing kilometre scale folds (Angerer and Hagemann 2010; Lascelles and Tsiokos, 2014) thickening the crust to ∼50 km (Fig. 2.2). The increased temperature and pressure at depth renewed partial melting of the crustal material generating andesitic magmas and volcanism forming ephemeral islands rising above the shallow ocean basins that were floored with predominantly ferromagnesian crustal rocks (Fig. 2.2). Explosive volcanism and subaerial erosion of the islands  deposited intermediate to felsic tuffs and immature sediments on the ocean floor overlying the older ferromagnesian lavas and BIFs.                                                                                    The further addition of folded crust and overlying volcaniclastic and terrigenous sediments at the intersections of convection cells (Fig. 2.3) produced greater thickening of the crust with further partial melting at the base and gave rise to anatectic and diapiric granitisation and the formation of the primordial cratons (Fig. 2.3) consisting of a granitoid basement overlain by folded oceanic crust and underlain by the mafic residue from the granitoid basement. Uplift and erosion of the buoyant continents resulted in stable continents consisting of granitoid terrains with remnants of oceanic crust in the form of greenstone belts with included Algoma type BIFs (Fig. 2.4). The formation of new crust and the amalgamation of the primordial cratons continued throughout the Archaean resulting in the formation of supercontinents (Rogers and Santosh, 2003) and gradual
deepening of the oceans while the increasing size and number of mantle convection cells caused interference and the transformation of plumes into linear combinations
as primitive spreading ridges.
As continents grew larger and thicker (Veizer and Jansen, 1979; Barley et al., 2005), during the Meso-Archaean era (Fig. 1C), the centres rose above the sea and subaerial weathering and erosion produced mature sediments interbedded with thick sequences of mafic to felsic lava flows that accumulated as deep sedimentary basins (e.g. Witwatersrand, Hamersley) on thinner craton margins (Fig. 2.5). Due to the small continental freeboard (Erikson, 1999) and comparatively shallow oceans (∼2 km), these basins were depressed below the level of the ocean floor and the input of turbidity and density currents from the ocean deposited thick sequences of Superior type BIFs. Note that BIF is typically absent from shallow parts of the basins due to its derivation from the oceans and not the continents. It is suggested that the Rapitan type deposits are similar in origin to the Superior type and formed when the weight of ice on the continents during extreme glaciation (Hoffman et al., 1998) temporarily depressed continental margin sedimentary basins to the level of the ocean floor.
Continental derived sediments encroached on the sea floor burying the Superior type BIF deposits and spilling onto the adjacent ocean crust forming new basins on the
ocean floor (Fig. 2.6). Depression of the ocean floor by the sedimentary basins was compensated by uplift of the continents increasing erosion and sedimentation into the
basins until the floor of the basins and underlying ocean crust sank to the zone of partial melting with the formation of new granitoids. Regional and thermal metamorphism and folding of the basin contents was followed by uplift to form an orogenic domain accreted onto the craton margin (Fig. 2.7). Only minor folding occurs at the margins of the continental basins which now form part of the buoyant continents.
The increasing growth of continents, with deep roots and marginal sedimentary basins in the Neo-Archaean era, their amalgamation into supercontinents, the reduced area of the oceans and the greater contrast between continental crust and ocean basins, causes the oceans to deepen and gave rise to a major change in the relation of crust and mantle convective cells. The increased density contrast between ocean crust and the thick sialic continents was accompanied by isostatic uplift of the continents and relative depression of the ocean crust and ultimately subduction of the cold ocean crust beneath the continents. Ocean crust no longer overlies continental areas and is subducted beneath the accreted sialic orogens and continents (Fig. 2.7).

Discussion
The realisation that BIF is a deep-water marine sediment dictates the present-day occurrence of BIFs to the preservation and exposure of deep ocean floor sedimentary rocks on the continents. A new model relating BIF deposition and preservation to crustal evolution is proposed. Banded iron formation deposited on the primitive ocean crust during the early to mid Archaean era was preserved in the remnants of the multi-kilometre scale folded regions in the greenstone belts (Algoma type BIF). Terrigenous sedimentary rocks are relatively rare in Archaean supracrustal rocks and clastic sediments in greenstone belts commonly consist of immature arkose, greywacke and minor pelagic shale
interbedded with volcanogenic rocks (Dunlop and Buick, 1981) indicative of derivation from active volcanic islands rather than stable continental regions exposed to long periods of subaerial weathering (Polat and Frei, 2005; Wilkinson et al., 2009). The land area exposed to subaerial weathering was minimal and with erosion rates generally exceeding weathering rates, especially in the absence of land vegetation, contributed little soluble weathering products to th e oceans. The oldest known widespread siliciclastic sedimentary deposits first formed ca 3 Ga on the thinner (continental shelf) margins of the cratons with the deposition of the Witwatersrand Group in South Africa and the later ca 2.7 Ga deposits such as the Fortescue Group in Western  Australia and the Kuruman Formation in South Africa.
Small continents and lesser ocean volume in the Archaean era meant shallower oceans of probably 1–2 km depth. These basins commonly contain relatively minor craton-derived mature psammitic and pelitic sedimentary rocks in the early stages but with the input of abundant volcanogenic rocks, they developed into deep basins below the level of the ocean floor and into which density currents and hydrothermal plumes from oceanic hydrothermal vents deposited thick accumulations of BIF (Superior type BIF)  interbedded with fine-grained turbidites, pelagic shale and volcaniclastic rocks; generally followed by more siliciclastic sediments as the basins filled. It is noteworthy that BIF is typically absent from contemporary shallow sequences such as the eastern part of the Hamersley Group in Australia and that the reason for the inability to locate the  hydrothermal sources of the iron and silica in BIF is because they were derived from the oceanic regions and not from the continent. They typically show comparatively minor
deformation and low-grade burial metamorphism, compared to the greenstone belts or adjacent younger orogenic basins, as a result of their more or less rigid cratonic base and subsequently added to the thickness and stability of the craton.
Assuming that subaerial exposure was more or less limited to active volcanic islands before the formation of cratonic continents, active erosion would preclude prolonged
weathering, particularly in the absence of vegetation to hold soils together, and the sediments derived from them would be immature arkose and greywacke and with low fO2 magnetite, pyrite and uraninite would be unoxidised (Roscoe, 1968; Grandstaff, 1980; Kimberley et al., 1980; Rasmussen and Buick, 1999; England et al., 2002) – typical of most Archaean sedimentary rocks. However, pyrite and magnetite are relatively resistant to oxidation and commonly occur in modern immature sediments from rapidly uplifted regions where erosion rates exceed weathering and although carbonate ions in solution can react with pyrite to form sulphate the reaction at ambient temperatures is slow and the supply of sulphate to the early ocean would be minimal (Huston and Logan, 2004; Caldeira et al., 2010).
Deep weathering requires the formation of stable continents where weathering rates exceed erosion rates. Terrigenous sedimentary rocks show increasing dominance through the Proterozoic and the Phanerozoic eras whereas pelagic sedimentary deposits are much less commonly preserved in post-Archaean rocks. Detrital sediments from the  continents, including the uplifted marginal basins, formed deep subsiding sedimentary basins on the ocean floor around the margins of the c ratons and in rifts as the  supercontinents were broken apart by plate tectonic movements. Collisions between continental plates turned the deep sedimentary basins into orogenic fold and thrust belts that were accreted onto the margins of the cratons, causing relatively minor  deformation of the continental sedimentary basins, i.e. Ophthalmia and Capricorn orogenies in Western Australia.
Conclusion
Reassessment of the rarely questioned assumptions underlying the BIF genesis models indicates that BIF deposition is independent of atmospheric oxygen levels and most
probably formed on the ocean floor whenever mantle plumes, active sea floor spreading or underwater volcanism occurred. Basic chemistry precludes even anoxic oceans
from becoming reservoirs of iron and silica in solution but it seems highly probable that the hydrothermal systems and iron formation deposits associated with undersea lava
flows have existed throughout geological history. As a deep-water marine sediment sourced from hydrothermal vents and redistributed by density currents across the ocean floor, it is therefore concluded that the deposition of BIFs on the sea floor occurred throughout geological history whenever undersea mafic lavas were erupted. The temporal distribution and abundance of BIF is controlled by the preservation or destruction of deep sea sediments but since the onset of modern style plate tectonics the ocean crust and overlying deep sea sediments including BIFs have been subducted beneath the continents. The apparent sharp reduction in BIFs after ~1.8 Ga is due to subduction of the ocean floor and BIF deposits; not to discontinuity of deposition and is unrelated to the atmospheric oxygen levels and the GOE.
Detrital sediments from the continents, including the uplifted marginal basins, accumulated on the ocean floor around the margins of the cratons and in rifts as the continents were broken apart by plate tectonic movements.  The light sediment piles depressed the underlying ocean crust forming deep accumulations but the  tops of the piles were  never below the ocean floor and thus did not contain BIFs. Converging continental plates turned the deep sedimentary basins into orogenic fold and thrust belts that were accreted onto the margins of the cratons, causing relatively minor deformation and uplift of the continental sedimentary basins (cf. Hamersley Province; Tyler and Thorne,  1990), while the oceanic crust underneath the basins was forced down into the mantle, and the ocean floor, on which BIFs continued to be deposited, was also subducted beneath the accreting orogenic pile and then descended into the mantle similar to modern plate tectonic processes (Fig. 2.5). It is apparent that present day ocean floor iron formations will also be subducted and destroyed: a fate that must have applied to other Mesoproterozoic to Phanerozoic BIF.  Lahn-Dill type BIF is commonly preserved due to
its origin with late stage orogenic volcanoes (Butuzova, 1969; Quade, 1976) and minor occurrences of BIFs are preserved in thrust slices of ocean crust embedded in orogenic belts (O’Rourke 1961).
In conclusion, the deposition of banded iron formations was independent of atmospheric oxidation and generally had no effect on the evolution of the atmosphere.