Genesa Iron - World

02005 Society of Eoonomk Geologists. Jnc. Eronomic CAOloo lOOt" Annra..r..rry VollI ll'Wl pp. 643-679

Precambrian Iron Formations and Iron Formation-Hosted Iron Ore Deposits

J. M. F. CLOUT'

Fortescue Metals Group Ltd., 50 /Gngs Park Road, West Perth, Western Australia 6872

AND B. M. SIMONSON

Geology Department, Oberlin CoUege, Oberti" , Ohio 44074-1052

Abstract

Iron formations are the most important precursors to et.'Onomic iron ore deposits. Iron formations originated as chemical sediments rich in iron and silica that accumulated almost exclusively on Archean and early Paleo· proteroZOiC sea floors. Most are banded and known as banded iron formations (BIF); they originated as thinly layered chemical muds. The less common granular iron formations (GIF) are rich in sand·sizecl de tritus known as granules and generally crossbedded, indicating deposition in shallower, higher energy environments. Banded iron formations display a heterogeneous suite of iron·rich minerals including oxides, silicates, carbonates, and sulfides; iron oxide and silicate minerals dominate GIF, although a few are rich in iron carbonates . The acme of iron sedimentation was reached between -2.65 and 1.85 Ga when large iron formations were deposited globally due to a unique confluence of (1) a large supply of aqueous iron from oceanic hydrothennal systems, (2) the appearance of large continental she lves to serve as depositional repoSitOries, and (3) a stratified ocean capable of connecting the two. Depositional mechanism(s) are still being debated, but evidence for the involvement of microbes is increasing.

Iron formation·hosted iron ore deposits account for the majority of current world iron ore production and consist of three classes: (1) iron-rich primary iron formation with typically 30 to 45 wt percent Fe, (2) martltegoethite ore with abundant hydrous iron oxides containing 56 to 63 wi percent Fe, and (3) high-grade hematite ores with 60 to 68 wt percent Fe. The high·grade hematite ores, which account for the majority of world re selVes of high· grade iron ore (>31,000 Mt), can be further subdivided into hematite and microplaty hematite ore types. Individual iron ore deposits range from a few millions of tons to over 2 billion tons at >64 wt per· cent Fe, although most are within the range of 200 to 500 Mt. Many depoSitional features of parent BIF and GIF, especially microbanding. have been preserved during ore formation for the martite-goethite and highgrade hematite ores.

There is a general oonsensus that martite-goethite ores from Australia fonned as a result of relatively recent supergene enrichment of iron formation through replacement of gangue by goethite beneath Cretaceous to Tertiary weathering profiles. Likewise, a supergene origin is well supported for the soft, high·grade hematite ores from the Quadrilatero Fenifero in Brazil, but this involved leaching of iron formation gangue and resid· ual concentration of hematite. The origins of both microplaty hematite and Brazilian hard high-grade hematite depoSits are still controversial; alternatives proposed vary from supergene to initial supergene with subsequent burial metamorphism, hypogene, and supergene-modified hypogene-hydrothermal involving warm basinal brines plus ascending or descending heated meteoric fluids . Although the supergene-modified hypogene-hydrothermal has received widespread support, it is unlikely that a Single hypogene model can explain the wide diversity of deposits around the world. •

Introduction

IRON FORMATION-HOSTED iron ore deposits account for the majority of current world production and >31,000 Mt of highgrade hematite resources of iron are (Dalstra and Guedes, 2(04). Archean and Proterowic iron formations with 30 to 45 wi percent Fe are still mined in the United States, Canada, and especially in China, the latter where -100 Mt was mined as ore in 2004 and beneficiated to a -64 to 67 wi percent Fe product (]. Clout, unpub. data). ApprOximately 700 Mt of run-of-mine ore and 200 Mt of magnetite concentrates were produced worldwide from iron formations in 2002 (Astier, 2003). Iron formations are also the precursor for martitegoethite and high-grade hematite (56--68 wi % Fe) orebodies.

Given the primacy of iron formations as hosts for large iron ore deposits, we first describe the sedimentary characteristics of iron formations and genetic models proposed to explain them. Variability in the characteristics of unenriched iron

t Corresponding author: [email protected]

fonnations set important constraints on ore formation ; some iron formations are even mineable without subsequent enrichment. Another important constraint on the formation of large iron ore deposits is simply the amount of iron Originally deposited on the sea floor. In general terms, the largest iron ore deposits are found where paleoenvironmental conditions permitted the accumulation of the largest iron formations. For example, the Carajas and Quadrilatero Ferrife ro (Iron Quadrangle) provinces in Brazil and the Hamersley province in Western Australia are among the largest iron formations on Earth and account for the bulk of the world's iron resources, in the form of iron formation-hosted high-grade iron ore. Higb-grade economic bodies do not form, however, without subsequent events that upgrade the iron formation, typically increasing from 30 wi percent Fe in an unenriched iron formation to 56 to 68 wi percent Fe in martite-goethite and high-grade hematite ores. To give a sense of the diversity of processes by which this can be accomplisbed, we dedicate the second and greater part of the paper to describing the

643

644 CLOUT AND SIMONSON

different types of high-grade iron ore depoSits and ore genesis models formulated to explain their origins. The ultimate goal is to better understand the full spectrum of processes that create economic iron depoSits from iron fonnations . especially high-grade iron ore, from which an iron ore product can be recovered, either through crushing and screening or beneficiation. This in tum will lead to more efficient exploration for as yet undiscovered deposits of iron ore to meet growing demand.

Our knowledge of the nature and genesis of iron formations and high-grade iron ore hosted by iron formations has evolved through time from the early petrographic and genetic work on iron formations (Van Hise and Leith, 1911; James, 1954; Cole and Klein, 1981; James and Trendall, 1982) as well as high-grade iron ores from Australia (MacLeod, 1966; Kneeshaw, 1975; Morris, 1980, 1985; Ewers and Morris, 1981; Ewers, 1983), Brazil (Guild, 1957; Dorr, 1964, 1965), and Canada (Gross, 1965). Despite the economic importance of high-grade iron ores, their origin is still subject to intense research and debate. The early work by Morris (1980) established the importance of supergene leaching in the upgrading of iron formation to high-grade iron ores, especially for hematite-goethite ores, and also led to the proposal of a supergene-metamorphic model for generating high-grade microplaty hematite ores. Renewed interest in iron ore genesis has led various workers to propose a supergene-modified hypogene hydrothermal origin for the upgrading of iron formation to form high-grade hematite ores in the Hamersley province of Australia (Barley et al. , 1999; Hagemann et al., 1999; Taylor et al ., 2001; Dalstra and Guedes, 2004; Thorne et al. , 2004) and the Caraj:\s province in Brazil (Dalstra and Guedes, 2004; Rosiere et al ., 2004;), as well as the hard, highgrade ores in the Quadrilatero Ferrifero district of Minas Gerais in Brazil (Spier et al., 2003; Rosiere and Rios, 2004), Krivoy Rog in the Ukraine (Dalstra and Guedes, 2004), and the Thabazimbi deposit in South Africa (Beukes et al., 2002; Netshiozwi, 2(02). This work has greatly expanded our understanding of iron ore formation , most notably on the nature of the ore fluids via detailed fluid inclusion and stable isotope studies (e.g., Hagemann et al., 1-999; Webb et al., 2003; Rosiere and Rios, 2004; Thorne et al., 2004). Given their diverse characteristics, it is unclear at present whether or not researchers will eventually converge on a Single unifYing model that can explain all large iron ore depOSits.

Review of Iron Formation Sedimentology

Where depositional features are not obliterated by metamorphism, iron formations can be subdivided into banded and granular varieties based on their original grain size. Even though their original grains are masked by diageneSiS, it is clear that banded iron formations, or BIF as they are commonly known (Figs. lA-C, 2A-C, 3A-C), were deposited as chemical muds. In contrast, the detrital textures of granular iron formations, or GIF, are generally retained and consist of well-sorted chemical sands (Figs. ID-F, 2D-F, 3D-F), analogous to those of calcarenites (Dimroth and Chauvel, 1973). Most of the clasts in GIF appear to have formed via erosion and intrabasinal redeposition of chemical muds like those that became BIF (e.g., Beukes and Klein, 1990). The acronym BIF is widely used as a blanket term for all iron formations because

most iron formations have thin layers or "bands," but it should be noted that GIF usually have thicker, more discontinuous bedding instead of thin, well-developed, and laterally continuous banding. The fundamental dichotomy between BIF and GIF has been recognized for years. For example, the "slaty" vs. "cherty" iron formations of the Lake Superior region (Morey, 1983) are essentially BIF and GIF, respectively.

In addition to the banded andlor granular distinction, James (1954) subdivided iron formations into four facies (Table 1). Although he called them "sedimentary facies," his subdivisions were more along the lines in which metamorphic petrologists use the te rm facies, as they were based entirely on what type(s) of iron-bearing minerals were present. Chert was not incorporated into his scheme because it is a near-ubiquitous component of iron formations . James did emphasize that different mineralogical facies of iron formations are likely to show different suites of sedimentary features, and this observation has stood the test of time. Specifically, the vast majority of GIF belong to his oxide andlor silicate mineral facies, whereas BIF span a much broader spectrum that includes Significant thicknesses of oxide, silicate, and carbonate facies (James, 1954; Simonson, 1985). Sulfide-facies BIF are rarer and less extensive but nonetheless occur in some successions (e.g., Goodwin et al., 1985). Debate continues as to which of the mineral constituents in iron formation (if any) represent original preCipitates and which formed during diagenesis. Excellent overviews of the chemistry and mineralogy of iron formation can be found in James (1954, 1966), Klein (1983), and Lepp (1987). As with all sediments, depoSitional processes and paleoenvironments are best inferred from textures and structures rather than from mineralOgical compositions alone. We summarize the depositional features first because they set important constraints on porosity, permeability, and rheology, which in turn help set the stage for any subsequent enrichment to form ore. We start with GIF because, unlike BIF, their primary detrital constituents are coarse enough to be readily visible in hand sample or thin section.

TABLE 1. Names and Idealized Compositions of the Iron-Bearing Minerals Characteristic of Each of James' (1954. 1966) Four Minera10gicaI Facies of

Unenriched, Unmetamorphosed Iron Fonnations

Mineral facies

Oxide

Silk-ate

Carbonate

Sulfide

Principal iron-rich mineral (s)

Hematite Magnetite

Greenalite l

Minnesotaite l

Stilpnomelane

"Chlorite"2 Riebeckite

Siderite Ankerite

Pyrite

Chemical fannula

Fe2D3 F~04

Fe,S;,o, (OH), (Mg.Fe),S40",(OH )' (K,N a, Calo .• ( Mg.F"" . F "").SW

(O,OH),,-2-4H,O (Fe,A1;Mg)' (S;,AI) ,o,(OH), Na:aFe3+Fe3· 2SiIlO~(OH )2

FeC03 Ca(Fe,Mg.Mn)(CO,) ,

FeS2

Fonnulas are from Deer et aI. (1992) unless indicated otherwise I from Miyano (1987) 2 from KJein and Bricker (1977)

!HON FOIlMAT10NS AND ASSOCIATED /HON ONE DEPOSITS

Flc. I. Iron formations in outcrop. A. The entire ISO-m thic:kness of Dales Gorge Member of the Brockman Iron Formation exposed in \Vi ttcnoom Gorge, Western Australia. Prominent macrobanding consists of banded iron fl:mnation (HI F) laye rs (oxidized resistant ledges) alternating with shaly layers of volcanic origin (slope-fanning intervals with extensive spin ifex gnL~s (.·over); Trendal! (1983. fig. 3--.5) shows part of same cliff with macrobands labeled. Kate geologist in white shi l1 on road in lower right comer. B. Part of single 131 F macroband ncar base of Dales Corge Member in Dales COr<Je with alternating mesoband ... of reddish white hematitic chert and dark iron oxides; note chert pods (whitish ovals ) in thicker iron oxide mcsobands and seam of crocidolite at base (sh iny, bluish strip just above the coin). Strata in image are 30 cm thick: coin in lower left is 2 elll in diam. C. E:qx)sure of silicate-carbonate BIF ncar base of the I ranwood Iron Formation, Cogehic Bange. northern Wisconsin; layers rich in iron silicates and carbonates appear greenish and tannish, respecti\'e1y. Coin is 2 em in diam. D. Cross-bedded granular iron formation (G IF) in the Sokoman Iron Formation near SchefTerville, Quehec; imhricated white diagenetic mottles dipping to len help define one 25-cm-thick crossbcd coset; layer above coset is rk:her in iron oxides and has small er scale trough crossbedding, reflected in more wavy bedding. Coin ill upper right is 2 cm in diam. E. Flat pebble conglomerate layer 20 em thick in same Sokoman exposure as (D); pebbles are intraclastic rounded disks of hematitic chert, space between them is filled with cherty C I F. Gray layers at top and bottom are C I F rich in coarsely crystalline magnetite. Coin ill upper right is 2 cm in diam. F. Interbedded Sokoman S ir and GlF deposited in deeper water east of SchefTerville, Quebec. Light-colored lenses are up to 10 em thick, ('"Ollsist of coarse GI F (see Fig. 2D), ha\'e internal crossbed.~, and rorm trains of stalvet! dune-size bedforms, some of which were deformed during compaction. Thinly IUlIlinatcd B I F rich in iron silicates (dark) encloses lenses. Coin above and to left of center is 2 em in (liam.

645

Granular ironjo111wtiol1s (GIF)

Three pri maly textural components are readily recognizable in GIF, as in most arenites: (1) a framework of clasts, (2) matrix (fine r grained interstit ial material), and (3) cement (authigenic minerals fillin g interstitial voids). Fine-grained detrital matrix is present locally (e.g. , Simonson, 1987, fi g.

8A) but rare in GIF' overall. The framework clasts mostly range in size Ii'om fin e to coarse sand (Mengel, 1973) and generally consist or a finely crystalli ne mixture of iron oxides, iron silicates, and/or chert internally. They have long been relim'ed to as granules (Figs. 2D, 3D, F ). They are analogous in Illany ways to the peloids and intraclasts of carbonate grainstones (Oimroth and Chauvel, 1973; Dimroth, 1976) and


FIG. 2. Iron formation in hand sample (all shown in correct stratigraphic orientation ). A. Microbanclcd he matiti c che rt layer capped by dark layer rich in magnetite with more indistinct lamination from Dales Gorge Me mbe r (same locali ty as Fig. lA)o B. Thinly laminated 131 F <:onsisting largely of side rite and chelt from deeper watc r part of the Sokoman Formation near Sche fTerville, Quebec. Gray maten,ll is unweathe red; dark brown rind along edge is from oxidation of iron ca rbonates. C , BIF from the \Vittenoom Fonnation , Western Allstralia, wi th parts of two hematitic chert pods: de flections o f laminations near ends of che rt pods arc due to diffe ren tial (;ompaction (see also Fig. 3C ). D. e l F' from le ns in Figu re I F, consisting of coarse sand to flat fine pebble.size intraclasts o r he matHic che rt; intergranular ce ment consists or t ransparent quartz and che rt. Staple in lowe r right ror scale. E. Cross section o r trough crossbed wit h tangential base rrom Sokoman e lF in the Howells Hiver area (Klein and Fink, 1976); originally homogeneous sand now varies in com position rm m red hematitic che rt to greenish silicate-rich chert to dark iron oxides. F GI l" rrom Sokoman ncar Schd TcJVi lle, Quebec, cu t by si nuous, near~ vertical cf"J.ck fi lled with dmsy megaquart i' .. e lF surrounding crack is cement rich and uncompacted. G. T)1)ical irreb'lIlar he(kling in G IF rrom the Sokoman Formation (sec Fig. I D ); early chalcedony and dms), quartz cements are abundant in hematitic cherts (reddish areas) but scarce in magnetite.li ch areas (dark). H. Small m \u mnar (fingc rlikc) che rt stromatolites (rom the Biwabik [ron Formation , Mesabi Bange, northern Minnesota; areas betwee n columns are fill ed with oolites or che lt and iron o.\y hydroxides.

m ON FOHMA'fIONS AND ASSOCIATED IRON OIlE DEPOSITS

F IG. 3. TextufCs of unenri ched iron format ion. A. Dales Go rge Member BIF (between c rossed polarizers) with microbands alternately rich in iron oxides and chert; chert-rich laminations have abundant coarse replacivc an kerite crystals (l ight) and iron oxide-ri ch layers (dark) have internal lami nations that arc fi nely wavy, probably from dirrerential cemen tation an d compaction. H. Thinly lam in ated BlF from the Kuru man Iron Formation , Transvaal SlLpergroup, South Africa (between c rossed polarizcrs); most of the rock (.'Ollsists of finely clystalline greenalite and siderite, but coarser ankerite (light) and magnetite crystals (black) selectively replace certain laminae. Sample from core CN-109 (I3eukes and Klein , 1990), cou rtesy of C. Klein. C. Contact between chert pod (clear) and adjaeent iron oxide-rich, chert-poor sediment (most ly opaque) in Dales Gorge Member BIF (plane-pohuized light); laminations inside chert pod are much thicker and very similar to those in c'Ontinuous chert layers (e.g., Fig. 3A). D. Typical oxide faci es GIF from the Gunfl int Iron Formation of CUll /lint Range, wcstem Ontario (between crossed polarize rs witli gypsum plate inserted). Granules are fairly homogeneous intemally and range from nearly pure chert {magenta} to al most opaque with iron oxides (black); most of original porosity was filled wi th chalcedonic cemcnt (o riented quartz fibe rs evidenced hy strips of blu e and yellow extinct ion). E. I lematiti<.: chert oolite from the Sokoman Formation near Scheffe rvilie, Quehec, with interstitial cement of drusy quartz On plane-polarized light). :\Iote delicate concentric laminations in ooid cortices and some compound nuclei. F. Sokoman Formation G IF' from same crop as Figure 2E (plane-polarized light); granules consist of chert with minor silicates (probably greennlite); intergranu lar cement is dru s), quartz. Diagenetic crystals of euhedral magnetite (black) and a fibrous iron silicate (probably minnesotui te) cut across both gran ules and c"Cments.

647

range in shape from well rounded to angular (Mengel, 1973). Concentrically laminated ooids are locally abundant in some GIF (Fig. 3E) but much rare r than granules overall. Many researchers believe that much of the material now found in the granules is de rived from the Original sedimentary material with relatively little change in composition (e .g. , Klein and

Bricker, 1977; Beukes, 1984). Many granules and ooids contain small septarian-style cracks formed by postdepositional shrinkage, suggesting they originally consisted of amorphous, gelatinous materials (see below).

Cements nIl the former pores between the granules in many of the undeformed GlF, most commonly iron-poor


drusy quartz anellor chalcedony. These siliceous cements generally show textures formed during VOid-filling precipitation rather than via later recrystallization (Fig. 3D). Much of this silica cement was emplaced very rapidly and close to the sediment-water interface (Simonson, 1987; Maliva et al. , 2005). Many GIF also contain large cavities, cracks, andlor vugs filled with siliceous cements (Fig. 2F) that may contain quartz crystals up to a centimeter long. Such cement-filled cracks cut across both granules and cements and form a continuum with septarian-style cracks confined to individual granules (Maliva et al., 2(05). The intergranular cracks have morpholOgies that have not been reported from any type of sediment other than iron formations and have been attributed to true syneresis, Le., shrinkage due to the dewatering of a gelatinous silica precursor (Gross, 1972; Dimroth and Chauvel 1973; Beukes, 1984). However, early silica cementation is not universal; many GIF were heavily compacted as evidenced by tight frameworks and distorted clasts. In addition to primary textural constituents, even undeformed GIF contain varying amounts of secondary diagenetic phases that are generally more coarsely crystalline and cut across clearly detrital textures. Because they consist of reactive precipitates, recrystallization of iron-bearing minerals is widespread, even in iron formations where surrounding units such as sandstones show little evidence of diagenetic reorganization (e.g., Klein and Fink, 1976). Han (1978, 1988) demonstrated via a series of careful textural studies that most if not all of the coarsely crystalline magnetite in GIF and BIF is diagenetic in origin rather than a direct precipitate from basin waters. Thus diagenesis alone can give rise to low-grade, syngenetic ore deposits, called taconite, that contain enough magnetite for it to be concentrated economically via grinding and magnetic separation. As a result of heterogeneous cementation, wide variations in the size and abundance of quartz crystals pose a challenge for beneficiation of taconite-type ore deposits (see beloW). However, quartz crystals in GIF grow larger and more uniform \vith progressive metamorphism (Gross, 1961). Both GIF and BIF can be rich enough in magnetite to be considered taconite ore, but the quartz crystals in a typical BIF tend to be much finer and more uniform than those in a GIF, unless it was metamorphosed.

Dune-scale cross stratification is the dominant depositional structure and evident in most GlF that have not been highly altered by diagenetic processes or tectonic deformation (Figs. ID, 2E; Simonson, 1985). The few paleocurrents that have been measured on these cross beds show complex polymodal patterns with hints of herringbone structures, typical of shallow marine sands (Ojakangas, 1983). Flat-pebble conglomerates (Fig. IE) are a minor but widespread component of GlF. Pebbles in intraclastic layers in GIF are generally siliceous, indicating that they were preferentially derived from silicarich layers; this is consistent with the early silica cementation discussed above. Other depositional structures are scarce, but siliceous stromatolites (Fig. 2H) are locally abundant in some GIF and have characteristics like those of siliceous sinters deposited by hot springs (e.g., Walter, 1972; Hall and Goode, 1978; Fralick, 1988; Maliva et al., 2(05). Due in part to pervasive early silica cementation, these stromatolites and associated strata in iron fonnations contain some of the best-preserved early Precambrian biotas in the world (Walter and

Hofmann, 1983), including the oldest macroscopic carbonaceous body fossils (Han and Runnegar, 1992).

Layers of pure GIF thicker than a few meters are rare, whereas BIF can continue uninterrupted by GIF or other iron-poor sediments for thicknesses up to a 100 m stratigraphically (Simonson and Hassler, 1996; Trendall, 2(02). Iron formations with a mixture of BIF and GIF are probably more abundant than pure GIF, and they show a style of bedding that is intermediate in character, Le., thicker and more lenticular on average than pure BIF but thinner and with more lamination than pure GIF (Fig. IF). The GIF in mixed iron formations usually occurs as discontinuous lenses enclosed in BIF. Some of these lenses represent starved bedforms generated by storm waves and currents (Simonson, 1985), but others appear to be products of differential compaction, Le., concretionlike volumes of sediment that were preferentially cemented with silica prior to Significant burial. The latter are analogous to the chert pods of BIF (described below). GIF lenses in mixed iron formations are commonly zoned with an oxidized, jaspery core and a more reduced outer rind; the latter is probably a reaction rim formed by incomplete equilibration between oxidized sands and the reduced muds encasing them during diagenesis.

Banded iron formations (BlF)

BIF are too intrinsically fine grained to reveal much about their sedimentary nature via petrographic analysiS (Fig. 3A, B). Even though diageneSiS has overprinted all primary detrital textures, unaltered BIF are typically remarkably fine grained and uniform and show more diversity in iron mineralogy than GIF. Most of these mineral assemblages are thought to have compositions close to the phases originally precipitated from basin waters, e.g., siderite, ferric hydroxides, and poorly ordered precursors of silicate minerals such as greenalite (Klein and Bricker, 1977). One clear exception is stilpnomelane, whose presence usually reflects contamination with volcaniclastic detritus (LaBerge, 1966a, b; Pickard, 2002, 2(03). Chert is generally interpreted as a primary precipitate (Maliva et al., 2(05), but its content can vary tremendously along a given stratigraphic level. The lateral variations are mainly via the presence of structures known as chert pods, which are localized pockets rich in silica, much of which appears to have been added early in diageneSiS (see below).

As the name implies, most BIF have well-developed thin lamination to thin bedding with alternating iron-rich and -poor layers (Figs. lA-C, 2A, B). A hierarchical nomenclature for layers at different scales has been developed for the BIF that have been studied in most detail , those of the Hamersley basin. Within stratigraphic units of BIF, bedding (generally referred to as banding) is commonly cyclic at three different scales, to which Trendall and Blockley (1970) gave the names macrobands, mesobands, and microbands. Thicknesses are typically on the order of meters for macro bands (Fig. lA), centimeters for mesobands (Fig. IB), and millimeters for microbands (Fig. 2A). Microbands were originally defined as couplets of layers, one rich in iron minerals and the other in chert. Subsequent study revealed that this simple nomenclature was not adequate to cover the complexity of all the variations in the Hamersley BIF (Ewers and Morris, 1981; Trendall, 1983). Additional terms such as aftbands (Trendall, I

IRON FORMATIONS AND ASSOCIATED IRON ORE DEPOSITS 649

1983) and microlaminae (Morris, 1993) have since been introduced, as well as entirely diffe rent nomenclatures (e.g., the ferhythmite of Beukes, 1980). The terms microband, mesoband, and macroband have gained the widest usage, but they may have somewhat different meanings when used by different researchers. Moreover, these terms have rarely been rigorously applied to BIF in other basins, one exception being the Kuruman BIF of South Africa, a contemporary of the Hamersley BIF (Pickard, 2003). Consequently, the terms micro-, meso- and macroband are used here primarily to convey a sense of scale consistent witll their original definition by Trendall and Blockley (1970).

Thin lamination is the norm in fme-grained Precambrian strata, given the lack of burrowing biota, but the layers in BIF (particularly those rich in iron oxides) are among the most striking seen in any sediment. Macrobands can generally be correlated throughout the ca. 60,000 km' within which the Hamersley BIF are preserved (Trendall, 1983). In some cases, mesobands and even microbands can be correlated for over 100 km (Trendall and Blockley, 1970; Ewers and Morris, 1981; McConchie, 1987), but it is not a given that such uniformity is typical of all BIF, and perhaps it has been overemphasized (Morris, 1993). Correlations at this level of detail have not been attempted in very many units outside of the Hamersley basin (Trendall, 2002). The correlations in the Hamersley basin are aided by the fact that the layers in these BIF are highly cyclic at various scales (described above). Given their scale and rhythmiCity, the microbands are generally interpreted as varves (Trendall and Blockley, 1970; Ewers and Morris, 1981; Morris, 1993). Trendall (1972) attempted to relate various cycles in the Hamersley BIF to orbital parameters, but they have yet to be adequately tested for the periodicities typical of Milankovitch forcing.

The behavior of the bands or layers in and around chert pods (Fig. 2C) reveals much about the original nature of the sediments that became BIF. The shapes of the chert pods and their relationships to enclOSing sediments are highly analogous to concretions in other types of sediment. For example, chert pods typically have ovoi·o cross sections, they are circular to amoeboid in plan, and microbands commonly continue through, but thicken sharply inside of, chert pods (Figs. 3C). By analogy to sediments of other compositions, the chertpoor BIF adjacent to the pods have been compacted relative to their original thickness, and material added near the sediment-water interface protected the sediment inside the pods from similar compaction (Dimroth, 1976; Beukes, 1984; Simonson, 1987). The cherty nature of the pods indicates that this material was siliceous cement. The observed reductions in microband thickness indicate that the depoSitional porosities of the precursor sediments to BIF were comparable to fine-grained sediments of other compositions (70-90% in argillite: Singer and Miiller, 1983; 80-95% in carbonate: Cook and Egbert, 1983), which in tum implies that most of the chert in the pod was added as early cement. Early cementation also helps account for how the Hamersley BIF responded to the rare high-energy events that happened during their depoSition. For example, high-energy waves andlor currents associated with an asteroid impact preferentially entrained those layers rich in silica in the Dales Gorge BIF (Hassler and Simonson, 2001; Pickard et al. , 2004). Likewise,

the fact that the rapid emplacement of several large carbonate debris flows did not cause any Significant soft-sediment deformation in the JolTre BIF immediately underneath (Kepert, 2001 ) is consistent with strengthening by early cementation. Early concretions typically shield minerals from chemical alteration as well as physical C'Ompaction. A range of finely crystalline iron-rich minerals are preserved inside chert pods, suggesting tl,at the original sediment had a range of compositions similar to the four facies shown by present-day BIF, rather than any Single precursor mineral (Simonson, 2(03). Pods in the Hamersley BIF may also have a preferred orientation andlor developed as stacked pods (Trendall and Blockley, 1970) due to dilTerential extension during compaction.

Depositionol environments of iron fonnation

Interpreting the environmental signillcance of iron formation is not as straightforward as it is for most types of sediment because of a lack of similar present-day analogs. The approach that has arguably shed the most light on the depositional setting of iron formations has been the study of sedimentary units witll which tlley are associated, particularly those with which they are in conformable contact. A wide array of different sedimentary and volcanic rock types is associated with iron formations, implying an equally wide array of different possibilities for subsequent development of iron ore deposits. The diverse sedimentary characteristics of iron formations themselves likewise require depoSition in a range of different environments. Moreover. in classifying iron-formations Simplistically into either banded or granular varieties, BIF or GIF, there are nevertheless large variations within each category. In short, iron formations constitute a diverse class of sedimentary rocks that show a range of depoSitional lithofacies. While some have textures that are analogous to, for example, certain limestones or phosphOrites, their distinct chemical composition indicates that processes or conditions rarely if ever active in the Phaneroroic were a prerequisite for the depoSition of iron formations. However. iroI) formations were depoSited in environments ranging from deep basinal shelf and slope areas well below wave base to shallow, higbenergy platform settings. Key sedimentary data bearing on the original environments of iron formation depoSition are summarized below, fOCUSing on the stratigraphic settings of the larger iron formations.

Despite the variety of different rock types associated witll large iron formations , generalized patterns have emerged at the broadest level. Most notably, the great majority of large iron formations are intimately associated with demonstrably marine units, and there is a high degree of correlation between the nature of an iron formation (GIF vs. BIF) and that of the associated units. Large GIF are typically underlain by shallow marine deposits such as tidally influenced quartz arenites (Goode et al. , 1983; Ojakangas, 1983; Simonson, 1984) or platformal carbonates (Beukes, 1983, 1986), whereas BIF are typically associated with deeper water shale-rich successions with turbidites whose composition varies from siliciclastic to volcaniclastic to carbonate (Larue, 1981; Klein and Beukes, 1989; Hassler, 1993; Simonson et al., 1993; Pickard et al., 2004). In addition, the successions in which large iron formations occur have proven amenable to sequence-stratigraphic analyses, and some of the largest BIF and GIF are in


successions that closely resemble those deposited in younger marginal marine settings (Blake and Barley, 1992; Morey and Southwick, 1995; Krapez and Martin, 1999). Although it is certainly' possible that some smaller iron formations are lacustrine in origin (Eriksson, 1983; Beukes, 1984), the nature of their stratigraphic context indicates most, if not all, large iron formations were deposited in open marine settings. An additional argument against a nonmarine origin is the lack of chemical and mineralOgical variability one would expect if they had precipitated from waters that would of necessity be highly variable in solute composition (Gole and Klein, 1981; Lepp, 1987). The sheer size and lack of internal variability of the largest iron formations is another argument in favor of the marine origin of iron formations (Kimberley, 1989; Simonson and Hassler, 1996).

Another common element in the deposition of large iron formations is that they were preceded and/or accompanied by transgression, i.e., deepening of the connection with the ocean. This was clearly the case with the large GIF (e.g., Ojakangas, 1983; Simonson, 1984; Klein and Beukes, 1989), where the onset of iron formation sedimentation may Signal a chemocline migrating into shallow areas from which it had previously been excluded (Simonson and Hassler, 1996). Deposition on a continental margin also enhanced the preservation potential of iron formations, since they were less likely to be subducted than if they had been deposited on oceanic crust. It is more difficult to determine if large BIF were likewise associated with transgression, i.e., deepening of the water column, because there is less variability in the associated sediments. However, this appears to have been the case in the two basins containing the largest and most extensive BIF, the Hamersley and Transvaal successions (Beukes, 1983, 1984; Klein and Beukes, 1989; Simonson and Goode, 1989; Simonson et al., 1993; Simonson and Hassler, 1996; Thome and Trendall, 2001).

Hoffman (1987) suggested the transgressions that commonly accompanied the deposition of large iron formations were due to subsidence induced by the approach of thrust sheets. Loading of the continental lithosphere induces a flexural response that creates a sediment-staIVed repository known as a foredeep (or foreland basin) that migrates laterally in front of the advancing thrust sheets. The sediment-starved phase in foredeeps is typically succeeded by a thick succession of shallow-water clastic material. The GIF-rich Danielskuil Member (Griquatown Iron Formation, South Africa) is one possible example; it shallows upward conformably into the lacustrine Pietersberg Member (Beukes, 1983). Dating of units from the Lake Superior region also appears to be consistent with the migrating foredeep model (Schneider et al. , 2002). However, large GIF are typically overlain by deeper water successions rich in shales and turbidites, many in conformable contact (Simonson, 1985), radler than dIe shallowwater successions predicted by the foredeep model. Morever, many large iron formations accumulated for extended periods on stable-shelf to upper-slope environments with little or no evidence of synsedimentary tectonism. In the Hamersley basin, for example, folding and thrusting began no earlier than 2.2 Ga (Tyler and Thome 1990), whereas the first major iron formation was deposited before 2.6 Ga. Finally, iron formations were depoSited in a number of different tectonic

environments (Gross, 1983). Some tectonic configurations may have even been unique to early Earth history, e.g., the centers of convective descent model of Trendall (2002). The scarcity of examples of post-GIF and -BIF shallOwing also raises questions about the simple model of chemocline migration. Many iron formations contain thin but widespread volcaniclastic interbeds (e.g. , Ewers and Morris, 1981; Hassler and Simonson, 1989; Barley et al., 1997; Pickard, 2002), suggesting that their depoSition was more a result of heightened volcanic activity than conditions endemic to anyone sedimentary or tectonic environment (discussed below).

Changes in iron formations through tim£

Iron formations range in age from early Archean to Neoproterowic, but they were not formed in equal measure throughout this long time span (Kerrich, 2005, fig. 12). BIF are found among the oldest well-preserved sedimentary successions on Earth (Nutman et al., 1984), although the sedimentary origins of some of dIe oldest BIF have been called into question (Fedo and Whitehouse, 2002). At the other extreme, iron-rich rocks widely referred to as iron formations were deposited on various continents in the Neoproterozoic. The Neoproterozoic units differ from early Precambrian iron formations in having a simple iron mineralogy dominated by hematite and being less cherty on average (James and Trendall, 1982; Beukes and Klein, 1992). However, thinly laminated cherty beds similar to microbanded BIF and pelOidal layers that resemble GIF occur locally in Neoproterozoic iron formations (Klein and Ladeira, 2004). Unlike early Precambrian iron fonnations, the Neoproterozoic examples are intimately associated with glaciogenic sediments (Young, 1976) and are much smalle r on average. The largest iron formations were all depOSited during an interval of -800 m.y. in the Neoarchean to Paleoproterozoic which ended rather abruptly at or before 1.8 Ga (Gole and Klein, 1981; Trendall, 2002). Researchers are starting to realize that this may consist of two or more peaks of iron accumulation rather than a single plateau of iron formation deposition (Isley and Abbott, 1999). Clearly, there were secular changes in both the size and depositional environments of iron formation , as follows.

Statistically, iron formations that are Paleo- to Mesoarchean in age tend to be smaller than those of the Neoarchean to Paleoproterozoic. This could simply reflect greater degrees of tectonic dismemberment with age. were it not for the fact that older iron formations show a different mix of depoSitional features and stratigraphic associations. Gross (1965, 1983) therefore subdivided iron formations into two major varieties, Superior type and Algoma type. In general, Algoma-type iron fonnations are smaller, consist exclUSively of BlF, and are intimately associated with volcanic rocks, whereas Superiortype iron formations are associated primarily with sedimentary strata (which commonly have a volcanic component), and may contain GIF as well as BIF. Algoma-type iron formations are typical of Archean greenstone belts, whereas Superiortype iron formations occur in continental margin successions, are Neoarchean to Paleoproterozoic in age. and include the largest iron formations. James and Trendall (1982) assessed the size variation in iron formations as a function of age by placing major iron formations from five continents into four categories: small (1010 or fewer tons of iron), moderate (on


the order of 101LI0" tons of iron), large (on the order of 1013

tons of iron), and very large (1014 or more tons of iron). Their data set confirms that the largest iron formations are all Neoarchean through Paleoproterowic in age. In contrast, smaller iron formations range in age from Paleoarchean through Paleoproterowic, including the time span during which the large iron formations were deposited. Although the smaller size of Algoma-type iron formations is generally taken to mean deposition in smaller hasins, Cole and Klein (1981, p . 170) correctly noted that some Algoma-type iron formations "may have been quite extensive prior to deformation and disruption."

The largest individual iron formations known from any point in geolOgiC time are the Neoarchean to Paleoproterozoic iron formations of the Hamersley basin of Western Australia and the Transvaal basin of South Africa. Examples of James and Trendall's (1982) · very large" iron formations are found on all five continents, but those of the Hamersley and Transvaal hasins contain the highest estimated tonnages of original iron. Although there are five major iron formations within tl,e Hamersley succession (Trendall, 1983, fig. 8) and two in the Transvaal succession (Beukes, 1984), the Transvaal BIF contain a larger total mass of iron because the area over which they are preserved is roughly twice that of the Hamersley BIF. The exceptional size of the iron formations in these two hasins becomes even more remarkable since they may actually be two parts of a single basin. Button (1976) summarized a number of striking similarities in their deposits (both sedimentary and economic) and geologic evolution. Cheney (1996) formalized this hypothesis by suggesting the name ·Vaalbara" for the combined landmass. Not everyone is persuaded, but detailed studies have revealed some striking geolOgiC parallels between these two successions even at very fine scales (e.g., Simonson and Camey, 1999; Pickard, 2003). IsotopiC dates compiled by Nelson et al. (1999) point to certain inconsistencies in the ages of stratigraphically comparable units on the two continents, but dates from the BIF themselves indicate they are essentially contemporaneous (Pickard, 2003). Either individually or jOintly, the Hamersley and Transvaal basins constitute the largest repositories of sedimentary iron on Earth.

Although iron formations grow larger on average around the time of the Archean-Proterowic boundary, the average energy of the environments in which they were deposited did not increase dramatically at first. The typically older Algomatype iron formations are generally associated with volcanic rocks and deep-water turbidites and consist almost excluSively of BIF (Dunbar and McCall, 1971; Barrett and Fralick, 1965, 1989; Shegelski, 1987). Occurrences of GIF in Algomatype iron formations (e.g., Manikyamba. 1999) are extremely rare. The older of the Superior-type iron formations likewise accumulated in deep shelf to pOSSibly upper slope environments (Trendall, 1983; Simonson, 2003; Pickard et al. , 2004) and consist largely of BIF, but more GIF are present in these older Superior-type iron formations (Simonson and Goode, 1989; Beukes and Klein, 1990). Unlike any of tl,e younger examples, some of the older GIF are siderite dominated. Recent work on the Fe isotope compositions of the siderite suggests they are indeed primary precipitates from the water column rather than diagenetic products involving reduction

of ferne iron-bearing phases after burial (Johnson et al., 2003, 2004).

Between their first appearance around 2.65 Ga and their disappearance around 1.8 Ga, Superior-type iron formations changed Significantly in character. Most of the large iron formations in the Lake Superior area and Labrador trough of North America were depoSited from around 2.0 to 1.85 Ga, and most are rich in GlF that have a high proportion of iron oxides and silicates but little siderite (G ross and Zajac, 1983; Morey, 1983; Dimroth, 1986; Fralick and Barrett , 1995; Fralick et al. , 2002; Schneider et aI. , 2002). Similarly, the iron formations of the Nabberu basin of West em Australia are young in age and rich in oxide-facies GIF (Hall and Goode, 1978; Goode et al. , 1983). In addition, many of these younger GIFrich Superior-type iron formations are in conformable contact with tidally crossbedded quartz arenites and stromatolitic dolomites (Hall and Goode, 1978; Morey, 1983; Ojakangas, 1983; Simonson, 1985). The increase in the abundance of GIF (Kernch et al., 2005, fig. 12) and the nature of the associated units both indicate that the average depositional energy of Superior-type iron formations increased through time.

It is unclear at present whether iron formations changed gradually and progressively through time or if the changes were abrupt, discontinuous, or possibly even oscillatory in nature. The transition from Algoma- to Superior-type iron formations was gradual in the sense that Algoma-type iron formations were still accumulating on other continents at the time the oldest Superior-type iron formations were being deposited in the Hamersley basin of Western Australia and the Transvaal basin of South Africa (ca. 2.6 Ga). Moreover, some iron fonnations deposited on the margins of the Kaapvaal and Zimbabwe cratons at -3.0 Ga appear to be intermediate in character between Algoma- and Superior-type iron formations (Watchorn 1980; Fedo and Eriksson 1996). The abruptness of the shift from virtually all BIF to a mixture of BIF and GIF within the Superior-type category is more difficult to assess because most iron formations with well-constrained ages are concentrated in relatively short time windows, the most prominent of which occur near 2.7, 2.45, and 1.9 Ga (Isley and Abbott, 1999). This clustering may itself be a sign that the evolutionary changes in iron formation were not evenly distributed in time. Moreover, a contrast in the isotopic variability of iron suggests that the younger Superior-type iron formations may have been deposited by different mechanisms than the older ones (Rouxel et al. , 2005).

Models for the Deposition of Large Iron Formations

There are no close present-day analogs of iron formation, and this may be the reason that an unusually broad range of theories has been proposed for their origin. As Trendall (2002, p. 60) so eloquently put it, iron fonnations •... have often been described as bizarre or unusual rocks, and correspondingly exceptional conditions have been advanced to explain their presence in the stratigraphie record; .. . it should not be asked what strange circumstances led to the depoSition of BIF, but instead in what respects were the ordinary environments of the Precambrian Earth radically different from those now existing." One of the first and most creative researchers to take this approach was Cloud, whose suggestions (e.g., Cloud, 1968) stimulated much new thinking about iron


formations and early Earth mnditions in general (see Trendall, 2002, for a nice summary of Cloud's contributions). Although consensus has yet to be reached on the specific mechanisms whereby iron and silica were precipitated and their physical states, most researchers now favor models involving derivation of the dissolved iron from hydrothermal vent sources in the open ocean, deposition of the larger iron formations on sediment-starved continental-shelf to upper-slope environments via the precipitation of iron-rich phases along a chemocline in a stratified water column, as well as coprecipitation of silica with the iron, and involvement of microbes in the precipitation and diagenetic reorganization of these phases. We elaborate briefly on these important points of broad agreement below (see Simonson, 2003, for a fuller discussion).

Banded iron formation deposition has been linked to hydrothermal activity via stratigraphic context and facies relationships for some Algoma-type iron formations (e.g., Goodwin et al., 1985), and hydrothermal geochemical signatures, e.g., in Tare earth element ratios and isotopic systems such as sulfur and neodymium (cited in Simonson, 2003) have been detected in all types of iron formations (Klein and Beukes, 1992). As Fryer et al . (1979) pointed out, sea-floor hydrothermal systems would have injected large masses of reduced species into the Archean ocean from the bottom, most notably ferrous iron. The need for a stratified water column stems mainly from the fact that, even though normal marine surface waters were clearly not well oxygenated in the Neoarchean to Paleoproterowic, they were still too oxic to carry much dissolved ferrous iron (Trendall, 2002). The mineralogy of iron formations themselves support this model in that fully oxidized hematite is the dominant iron mineral among the least altered GIF, which were deposited in the shallowest waters, whereas the BIF deposited in deeper water show a much broader range of iron minerals, including large amounts of reduced phases such as siderite.

Contrasts in the trace element and isotopic compositions of iron formations and coeval iron-poor strata (Klein and Beukes, 1989; Carrigan and Cameron, 1991; Winter and Knauth, 1992; Rouxel et al ., 2005) support a stratified ocean model. Consensus has yet to be reached on the character and causes of that stratification. Some workers envision a surface layer depleted in iron and a large reservoir of bottom water with relatively uniform concentrations of dissolved ferrous iron (e.g., Jacobsen and Pimentel-Klose, 1988; Huston and Logan, 2004). Others beUeve dissolved iron concentrations reached a maximum at some intermediate water depth Q\ving to higher concentrations of hydrogen sulfide in deeper waters (Cameron, 1983). Sulfide concentrations were probably low in early Precambrian oceans overall because of low sulfate production during weathering in an atmosphere with Uttle oxygen (Farquhar et al. , 2000), resulting in low inputs of reducible sulfate into mid-ocean ridge hydrothermal systems (Kump and Seyfired, 2005). This is consistent with stratigraphiC patterns shown by many Superior-type iron formations (Simonson and Hassler, 1996). In either case, Isley (1995) demonstrated the feasibility of connecting open-ocean hydrothermal sources with shelf sinks via lateral dispersal at shallow to intermediate water depths, even for the large Superior-type iron formations.

Given this situation, models for the deposition of large iron formations should focus on processes active along chemoclines between deeper iron-rich and shallower iron-poor water masses (e.g., Beukes and Klein, 1992). For example, iron could be precipitated via oxidation along the chemocUne in a manner somewhat analogous to the formation of particulate MnO. in the present-day Black Sea (Force and Maynard, 1991). Microbes are apt to take advantage of any steep chemical gradients, as they appear to have done for billlons of years (Johnson et al., 2004), so microbes were probably active along these chemoclines. Iron isotopes show more variability in iron formations than in any other natural material (Beard et al ., 1999; Johnson et al., 2003). This was initially taken as a sign of microbial mediation of redox reactions, but separating biotic from abiotic fractionation effects is challenging (Beard and Johnson, 2004; Johnson et al., 2004). Moreover, microbially mediated reactions could have affected the isotopic composition of iron either wben it was first fixed from the water column or during reorganization in the pore waters after depOSition (or both). Recently, Rouxel et al. (2005) attributed the observed variations to rapid changes througb time in the isotopic composition of the reservoir of dissolved iron in Archean seawater, Whether they were responsible or not, calculations by Konhauser et al . (2002) indicate that the number of microbes needed to Hx the mass of iron present is not unreasonable, even in large iron formations. Trendall (2002) marshals additional arguments faVOring the involvement of the biosphere in the depoSition of iron formations. Given the variety of iron minerals found in iron formations, a variety of different precipitation mechanisms were probably involved at different times and places (see review by Morris, 1993, p. 254-256). Pinning down the specific mechanisms and explaining cycUc pattenlS in BIF are the greatest challenges remaining in understanding the precipitation of iron formations,

There is near-universal agreement that tile high chert content of iron formations reflects higher ambient concentrations of silica in Precambrian seawater due to the absence of silicafIxing organisms (Mallva et al. , 1989, 2005). Agreement is more elusive concerning the source of the silica or the cause of its coaccumulation with iron-nch phases, Mechanisms proposed for silica precipitation include direct or indirect fixation by microbes in the water column (LaBerge et al. , 1987), sUght evaporative concentration, coprecipitation with iron (Ewers, 1983), and polymerization due to electrolyte changes (Morris, 1993). The high silica content of iron formations is not excluSively a depositional feature; there is saUd evidence (outUned above) that a Significant fraction of the silica in iron formations, both BIF and GIF, actually preCipitated in the shallow subsurface as vOid-Hlling cement shortly after depoSition, presumably abiogenically (Simonson, 1987; Mallva et al., 2005). Recent geochronologiC work suggests that the sedimentation rates of iron formations were faster than those of other sediment types with which they are interbedded, e.g., the S macrobands in the Dales Gorge BIF, which are shales rich in fine volcaniclastic material (Fig. lA; Pickard, 2002, 2003; Trendall, 2002; Trendall et al. , 2004). This in turn suggests a situation where there were relatively short-Uved pulses of iron and silica input, again consistent with the behavior one might expect from hydrothermal sources. Higher geothennal

IRON FORMATIONS AND ASSOCIATED IRON ORE DEPOSITS (WI

gradients could have also increased the Ilux of silica from below by acc..,lerating its dissolution at depth and reprecipitation in the shallow subsurface (Simonson, 1987). However, recent work on GeiSi ratios has indicated the silica may be at least partly a product of continental weathering (Hamade et al.,2oo3).

In addition to points of agreement summarized above, there is also a broad consensus as to why iron formations grew Significantly larger on average during the Neoarchean and Paleoproterowic. The transition from small Algoma- to large Superior-type iron formations -2.6 billion years ago apparently rellects the Hrst appearance of extensive continental shelf environments. Continental margins offer larger, more unifo rm repositories than volcanic terraness, and stable-she lf deposits of various types increased dramatically in size during the Neoarchean. For example, the first platformal carbonates comparable in size to Phanerowic bUild-ups appeared in the Neoarchean in tl,e same basins as the Hrst large iron formations (Beukes, 1983; Klein and Beukes, 1989; Simonson et al. , 1993; Grotzinger, 1994). The expansion of shelf area presumably rellects a Neoarchean surge in the growtll of continental crust and associated rise in sea levels (Goodwin, 1991; Lowe, 1992; Groves et al., 2005). The highly diachronous nature of cratonization (Eriksson and Donaldson, 1986) may help explain why the largest iron formations differ in age on different continents (Trendall, 2002). However, the increase in the average size of iron formations may not be entirely a product of a tectonic shift. The dramatic increases in the abundance of GIF, Hrst from Algoma- to Superior-type iron formations and then from olde r to younger Superior-type iron formations, indicate tllat they were deposited in progreSSively shallower waters. This implies progressive shallOWing of chemoclines that could rellect changes in the chemistI)' of the atmosphere anellor seawater. The scarcity of Superior-type iron formations with well-constrained ages between ca. 2.45 and 1.9 Ga (Isley and Abbott, 1999) makes it difficult to determine if this shift took place gradually and monotonically or was rapid anellor episodic.

Finally, there is a consensus that tl,e seemingly abrupt termination of iron formation deposition in the Paleoproterozoic rellects evolutional)' shifts in atmospheriC and hydrospheriC chemistI)' (Knoll , 2003; Canfield, 2005). Prior to ca. 1.9 Ga, dissolved iron could neither accumulate in high concentrations nor be dispe rsed over long distances in the ocean's surface waters but must have done so in the deeper parts of the ocean. The mobility of dissolved iron in deeper waters was clearly radically reduced at about 1.9 Ga. Until recently, this has generally been attributed to ventilation, i.e., oxygenation, of the deep ocean. However, the Hrst dramatic rise in atmospheriC oxygen appears to have taken place around 2.4 Ga (Bekker et al., 2004; Kerr, 2005), which predates the end of iron formation depoSition by a wide margin. An alternative model tl,at invokes increased levels of dissolved sulfide rather than dissolved oxygen to limit iron solubility in the deep midProterowic ocean (Canfield, 1998; Anbar and Knoll, 2002; Arnold et al. , 2004) is gaining adherents. Whatever the change was, it clearly prevented tl,e deep ocean from storing and transporting dissolved iron over long distances, thereby severing the connection between sea-floor hydrotllerrnal systems and continental shelf environments, thus putting a stop

to the depoSition of iron formations. The (111)' sigllitll'ulIl reappearance of iron fonnations happened in till' N,'oproteroroic. The source of iron for the Neoproterozoic iron lill" mations again appears to have been hydrothe rr"al (Bn·ilko],!". 1988; Young, 1988), but their intimate association willi gladogenic sediments may also be important. The Neoproterozoic glaciations were probably the most severe in Earth hi stol~ ' (Hoffman et al .. 1998). It is possible that global ocealls emered by ice became highly stratined for the first time in o""r an eon, thereby reactivating some of the mechanisms at work in tl,e Paleoproterowic (Klein and Beukes, 1992; Trendal!. 2002; Klein and Ladeira, 2004). Another possibility is that SIII

fide concentrations began to decrease in deep ocean waters ill the Neoproterow ic, leading to greater iron mobility. Either way, the deep ocean definitely became ventilated as th" Phane row ic approached, redUCing the mobility of dissolved iron for good (Knoll, 2003; Canfield, 2005; Kerr, 2005).

In contrast to the \videspread agreement on the points raised above, the re are still at least two competing explanations that relate iron phases with different oxidation states to differences in water depths. One explanation is based primarilyon detailed studies of iron formations and associated strata of the Transvaal Supergroup in Soutll Africa. where Beukes and Klein (1992) concluded that iron minerals became progreSSively more oxidized in progreSSively greater water depths. They envision sideri tic sediment being depoSited in the shallowest, highest energy environments, and fully oxidized hematitic sediment (or a suitable precursor) in deeper water. This is consistent with the fact that siderite is the main iron-bearing mineral in the prinCipal GIF of tl,e Transvaal basin, the Griquatown Iron Fonnation. They attribute this relationship to a Ilux of organic carbon being transported from shallow to deeper environments, thereby creating a gradient from more anoxic shallow water to more oxic deeper water. However, extending this explanation to other basins is problematic because the Transvaal situation is anomalous. In almost all other basins. GIF are dominated by oxide iron phases with an abundance of hematitic minerals, whereas siderite and other reduced iron phases are much more abundant in deeper water BIF (e.g., James, 1954; Zajac, 1974; Simonson, 1985).

According to the second explanation, iron phases become progreSSively more reduced (rather than oxidized) with increasing water depth . GIF with fully oxidized iron phases suggest an atmosphere that was sufficiently oxic to keep concentrations of dissolved ferrous iron in an oceanic surface layer at Vanishingly low levels, thereby restricting depoSition of reduced phases to deeper water. Support for this model comes from the fact that many of the cherty oolitic and stromatolitic layers in GIF are among the reddest layers, indicating all abundance of finely disseminated hematite . The depositional textures are exquiSitely preseJVed in these oolitic and stroma· tolitic layers (as is the famous Gunflint microbiota; Walter and Hofmann, 1983). suggesting they are closest to their primaI)' mineralogy. Moreover, tl,ey are most likely to have eqlli Iibrated \vitll the atmosphere chemically as they were d,,posited in some of the shallowest. highesl C11(>r",' environments of any iron formations . DepoSition of the most oxidized iron formations in the sha110west water t'llviro ll · ments would also be consistent with the low L'oneent nit ill iiS I If


iron found in platformal carbonates deposited coevally (Veizer et al., 1990, 1992). Had the level of dissolved ferrous iron been uniformly high in ocean waters both deep and shallow, much greater quantities of iron would surely have substituted for calcium while shallow-water carbonates were precipitating. The relatively low iron content is observed in both carbonates that grade into iron formations stratigraphically (e.g., Klein and Beukes, 1989) and others that accumulated in shallow water at the same time that BIF were being deposited in deeper parts of the same basin (Simonson and Hassler, 1996; Kepert, 2001). Recent work on iron isotope ratios provides support for ocean stratification during deposition of the younger Superior-type iron formations but not those of Archean and earliest Paleoproterowic (Rouxel et al., 2005).

In summary, the evidence is mounting that large Superiortype iron formations owe their existence to a unique confluence of three main circumstances in the N eoarchean to Paleoproterozoic: (1) the presence of large hydrothermal systems on the open ocean floor, (2) a dramatic expansion in the total area of continental shelves, and (3) a stratified ocean with intermediate and/or deep water masses through which large fluxes of dissolved ferrous iron could travel from sea-floor hydrothermal.ystems to distant depocenters.

The fact that large iron formations occur in many different tectonic settings and are associated with many different rock types (Gross, 1983; Fralick and Barrett, 1995) suggests that these circumstances were met in a variety of different settings. If so, the first-order cause of large iron formations could simply be unusually vigorous hydrothermal activity (e.g., Morris, 1993; Barley et al., 1997), espeCially when connections between hydrothermal sources and continental-margin sinks were enhanced by sea-level highstands. Highstands are probably linked to the increased growth of continental crust and have coincided with periods of heightened volcanism, e.g., increased activity at hot spots and/or spreading ridges, throughout Earth history. At such times, precipitates formed on a chemocline could overwhelm clastic input and accumulate relatively undiluted (Simonson and Hassler, 1996), consistent with depositional rates calculated by Picard (2002, 2003) and Trendall et al. (2004). Isley and Abbott (1999) believe there is a statistically Significant correlation in age between iron formations and proxies for mantle plume activity, such as komatiites and flood basalts. A connection between deposition of iron formations and mantle superplumes could help explain why Superior-type iron formations do not appear to be evenly distributed in either time or space.

The existence of a hypsometry unlike any before or since may have been a contributing factor in the formation of large iron fonnations during the Archean-Proterozoic transition. While it is commonly assumed that continental freeboard has remained constant through geolOgiC time, this is not necessarily the case (Eriksson, 1999). Arndt (1999) has pinpointed unusual aspects of Archean and Proterozoic volcanic rocks that suggest the existence of broad, submerged continental platforms unlike any later in Earth history. Widespread evidence of basaltic hydrovolcanism in large iron formation basins (Hassler and Simonson, 1989; Hassler, 1993; Altermann, 1996) prOvides support for this scenario, which may reflect secular differences in the thickness and buoyancy of Archean crust (Groves et al., 2005). Replaced shards appear

in some tufTs found in iron formations or associated units (LaBerge, 1966a, b; Ewers and Morris, 1981; Pickard, 2002), but signs of explOSive felsic volcanic activity are scarcer than one might expect if these iron formations had been deposited close to a convergent margin (e.g., in a backarc setting, as suggested by Blake and Barley, 1992). The existence of Uniquely large areas of stable flooded continental crust could help explain the accumulation of uniquely large and well-preserved iron formations in the Neoarchean to Paleoproterozoic and perhaps the exceptional lateral continuity of depoSitional bands in the Hamersley and Transvaal BIF.

Review of Iron Formation-Hosted Iron Ore Deposits

Iron formation-hosted iron ore depoSits account for the majority of current world production and resources of iron ore, followed by the important channel iron deposits which filled Tertiary river channels, iron-apatite ores (Williams et al., 2005) that are reCOgnized by most as hydrothermal and argued by some as magmatic (e.g., Kiruna and Malberget in Sweden and iron ore depOSits in coastal Peru and nothern Chile), and finally certain types of copper skarn and rare earth depoSits (e.g., Da Ye and Bayan Obo, China). There are many other types of iron ore deposits that have been histOrically important but not as Significant as the iron formation iron ores (e.g., oolitic goethite deposits of Minette and/or Salzgitter type, contact metamorphic ores formed by replacement of carbonate rocks in the aureoles of granitoid intrusions, and detrital marine placer deposits ).

BIF - and GIF -hosted iron ores can be subdivided into three classes: (1) unenriched primary iron formation with typically 30 to 45 wt percent Fe; (2) martite-goethite ore formed by supergene processes, with abundant hydrous iron oxides containing 56 to 63 wt percent Fe; and (3) high-grade hematite ores thought to be of supergene modified hypogene or metamorphic origin with 60 to 68 wt percent Fe (Table 2; Morris, 1985; Beukes et al. , 2002). Martite is a commonly used textural term to denote hematite pseudomorphs after primary magnetite where the octagonal outlines of much of the original magnetite are preserved.

The high-grade hematite ores can be further subdivided into hematite, including itabirite-derived residual ore and microplaty hematite replacement ore. Itabirite is oxidized, metamorphosed, and heterogeneously deformed BIF that contains iron ore deposits formed by supergene leaching of gangue minerals and residual accumulation of hematite (Dorr, 1969; Spier et al., 2003). Microplaty hematite replacement ore consists of a three-dimensional network of 10- to 200-l'm plates of hematite with interstitial voids, formed from replacement of silicate and carbonate bands in the iron formation (Morris, 1985). Individual high-grade hematite iron ore depOSits range from a few million tons to over 2 billion tons at >64 wt percent Fe, although most fall within the range of 200 to 500 M t.

Many of the primary mesobands and microbands of the parent BIF have been preserved during ore formation of the martite-goethite and high-grade hematite ores (Fig. 4; Morris , 1985). This preservation has resulted from replacement of chert and carbonate bands by hematite (in the case of highgrade microplaty hematite ores) or goethite (in martitegoethite ores), or residual accumulation of martite as the


TABLE 2. Iron Ore Deposit Classifications for Current and Some Future Mines with Either Published Resources or Resenres

Classification Subtype Current and future mines

Iron Fonnation, BIF and GIF Magnetite China: Diao }untai, Gong Chang Ling. Chita Shan , Dashihe United States: Empire. Hibbing, Northshore

Martite~goethite

High-grade hematite

Banded

H ematite

Microplaty Hematite

Hematite

Australia: Mount Gibson, Koolanooka, Tallering Peak (Yilgam craton ), Balmoral (Pilbara cm.ton )

Canada: Wabush, Mont Wright, Carol Lake United States: Tilden China: Dong Anshan, Hainan Island

Australia: Marandoo, West Angelas, Orebody 29, Mining Area C, Hope Downs. Christmas Creek, Ophthalmia Range, Rhodes Ridge, Paraburdoo Eastern Hanges, Sectio n 6 and 7 (Pilbara craton), Koolyanobhing, Mount Jackson and Mount Windarling (Yilgalll craton )

Australia: Mount Whaleback, Mount Tom Price, Pat'dburdoo, Channar, Yame, Ciles Mini (Pilbara craton); Iron Duke, Iron Knight, Iron Duchess (Gawler craton)

Brazil: Caraj1s district: NI to N9, including N4E, Sl1-S45 India: Goa, Noamundi, Aridongri district Guinea: Simandou, Mount Nimba South Africa: Thabazimbi Ukraine: K.rivoy Rog district

Brazil: Quadrilatero Ferrifero district: Aguas Clams, Alegria, Andrade, Bali , Brucuru, Caue, C6rrego do Feijao, C6rrego do Meio, Capenema, Concei~, Casa de Pedra, Fabrica, Fazendao, Morro Agudo, Maruca, Ouro Fino, Pico, Pires, Retiro das Almas, Tamandua and TImbopeba

China: Hainan Island South Africa: Sishen-Beeshoek Sierra Leone: Marampa

Microplaty Hematite Ore Iron Formation ><'>8<'QXI ___ K Microplaty Hematite

result of leaching out of chert and carbonate from bands which once contained disseminated magnetite. Magnetiterich bands in the iron formation were oxidized to martite andlor replaced by secondary hematite in ore, whereas Alrich silicate bands have been leached and partly replaced by clays to form shalelike bands. The shalelike mesobands should not be considered as epiclastic sedimentary rocks but rather the Al-rich residue of supergene leaching and weathering of Al-rich silicate mesobands that originated as a mixture of fine volcaniclastic material and chemical· precipitates (Ewers and Morris, 1981; Picard et al ., 2004). In martitegoethite ores, chert- and carbonate-dominant bands have been leached out and replaced by goethite so that the overall iron content has been enriched. Disseminated magnetite in chert and carbonate bands is oxidized to martite in martiteg~thite ores. In the microplaty hematite ores, a network of fine (O.2-D.Ol mm) microplates of hematite has completely replaced chert and carbonate-dominant bands. For the itabirite-derived ores from the Quadrilatero Ferrifero, residual concentration of hematite and martite has been achieved by leaching out of silicate- or carbonate-rich bands. In the less strongly leached itabirite-derived ore, some remnant friable quartz may be present with residual hematite and martite.

'.' ~ Martite/ Hematite

_. -

cheril Ore E=3 Shale/Clay Carbonate Format ion Martite-Goethite Ore

V'-I '~ilii;il'e~~ Martite .liiiUI G~~ite-Martite ~ Shale/Clay

Hematite & Itabirite Ores '11 Hematite/Quartz P: Hematite

= == Friable quartz/day

FIG. 4. Simplified summary of ore relationships to primary banding in BIF for high-grade hematite and martite-goethite iron ores. In microplaty hematite ores, BIF chert or carbonate bands are replaced by microplaty hematite (cross-hatching), magnetite bands by martite-hematite (black bands), and Al silicate bands by shale (clays), For mamte-goethite ores, BIF chert or carbonate bands are replaced by goethite-disseminated martite (black squares), magnetite bands by martite (black bands), and AI silicate hands by shale (clays). In the case of itabirite and othe r hematite o res, leaching of BlF chert o r carbonate bands has resulted in residual accumulation of hematite (Stipple) or friable hematite-residual quartz, magnetite bands by hematite and/or martite (black bands), and AI silicate bands by shale (clays). Note that chert and carbonate bands in parent BIF commonly contain disseminated magnetite. Note similar replacement relationships also apply to GIF.

Although the main ore minerals are hematite and g~thite and the overall mineralogy is quite simple, are textures and their spatial distribution and modification across individual deposits are typically complex (Clout, 2002).

Iron ore tenninology

There are a number of terms used frequently to .describe the physical properties of iron ores not commonly used in other commodities, especially relative physical hardness. The relative physical hardness of high-grade hematite and


martite-goethite ores varies conSiderably within a deposit and is a function of ore texture and the extent of secondary leaching of martite. goethite. and microplaty hematite (Clout. 2002). Iron ores are commonly described in terms of decreasing physical hardness from hard to medium. friable. soft. and finally powdery or dust. Ores that are hard to medium have low porosity with physically strong interlocking textures of martite. hematite. or martite-microplaty hematite. Friable is used to describe ore that can be easily broken up by hand. commonly into centimeter-scale prisms defined by joint planes and fissile bedding. Friable ore is more porous than medium-hardness ore. Soft is ti,e term used where the ore can be dug in situ by hand or with a shovel. is composed of 0.05- to 1.0-mm particles. and is very porous and typically not dusty. Powdery or dusty ore contains appreciable particles less than O.Ol-mm diameter. Blue dust is a term used in Australia and India to describe distinctly blue-gray fme-grained (0.005-0.2 mm) powdery ore composed of leached martite and or microplaty hematite. The hardness and related characteristics of ore are obviously important factors in its grinding and beneficiation. as well as the amount of lump ore obtained as it attracts a premium price over the fines. Lump ore is defined as having a particle size of 31.5 to 6.3 mm. whereas particles less than 6.3 mm are known as sinter fmes.

Unenriched iron formation ores

Unenriched primary iron formation is a major source of iron ore in many parts of the world. especially China and North America. and includes both magnetite- and hematiterich iron formation. In North America. the term taconite is used to describe magnetite- and hematite-rich BIF and/or GIF ore with >30 wt percent Fe (James. 1954; Neal. 2000). In Australia and Brazil , little unenriched iron formation are has been mined to date due to the presence of Significant resources of direct-ship or easy to beneficiate high-grade hematite and martite-goethite ore.

Where observed. the geology and mineralogy of the iron formation ores are similar to the surrounding uneconomic iron formation, except that the are zone contains more abundant and coarser grained magnetite (Fig. 5A) or hematite. less gangue inclusions in magnetite. or a higher percentage of iron oxide meso- and/or microbands compared to silicate- or carbonate-rich mesobands (Fig. 5B). Iron formation ores are commonly located in greenschist- to amphibolite-facies terrane. with the higher metamorphic grade associated with coarser magnetite grain size and more discrete grains of gangue (Neal. 2000). Sub- to lower greenschist-facies iron formations such as those in Australia and Brazil are commonly uneconomic, since the magnetite is either very fine (Fig. 5C. D) or contains very fine inclusions of gangue which require expensive fine grinding to liberate prior to beneficiation. Many iron formation ores contain a weathered cap of hard hematite. hematite-goethite. or friable hematite-quartz (Fig. 6 ). which in some cases may be economic to process (e.g .• Tallering Peak mine. Western Australia. and the now depleted high-grade supergene-upgraded ore above taconite ore in the Great Lakes region).

In the United States. Significant resources of Lake Superior-type BIF and/or GIF taconite have been mined from Michigan (Empire and Tilden) and Minnesota (Hibbing and

North Shore). with the Fe grade of ores as low as 30 to 35 wt percent (Graber and Sundberg. 2002). Although magnetite is commonly the main ore mineral in the iron formation, other ore minerals include hematite-siderite (Tilden mine) and minor goethite near the surface. Key gangue minerals include quartz. iron carbonates (siderite. ankerite). and iron silicates (minnesotaite, greenalite. stilpnomelane. cummingtonite).

The Anshan district. located in the Archean Anshan-Liaoning granite-greenstone belt in the southern part of the Liaoning province in northeast China. has a history of iron formation mining dating back to the middle Tang Dynasty of the 7" century. The Gongchangling deposit is typical of a number of >500 Mt BIF ore deposits (25-35 wt % Fe) in the Anshan district. with Significant production of 35 Mt of magnetite concentrates in 2004 (Bofei. 2005). The Gongchangling deposit is located within a steeply dipping greenschist- to amphibolitefacies metavolcano-serumentary sequence. The main ore mineral is magnetite. although it has been oxidized to martite near surface. The gangue mineralogy is quite complex and includes quartz. stilpnomelane. chlorite. muscovite. eastonite. alumino-greenalite, almandine, grunerite, cummingtonite. calcite. albite. epidote. or hornblende as independent or composite bands. Cummingtonite. grunerite. riebeckite. and arfvedsonite are indicators of high-grade ore. whereas different varieties of hornblende are indicators of low-grade ore.

The Mount Gibson (Extension Hill) deposit (200 Mt) is a typical iron fannation ore located in the Murchison province of the Archean Yilgarn craton of Western Australia (Western Australia Department of Industry and Resources. unpub. data; Lascelles. 2002). The deposit is located within steeply dipping and tightly folded magnetite BIF of the Windanning Formation. which has been metamorphosed to lower greenschist facies . and has a hematite-goethite oxidized cap (Fig. 7; Lascelles. 2002). Magnetite is present as 0.05- to 0.2-mm-size grains in typically lO-mm-thick (<1 mm->20 cm range) oxide-rich (>80%) mesobands or as a matrix surrounding silicate gangue. as well as in disseminated fine (0.001-0.01 mm) grains within chert-rich mesobands (Lascelles. 2002). Separating the magnetite-rich mesobands are bands ranging in composition from hydrous iron silicates to carbonate or chert. Although magnetite is the main ore mineral. like in many other BlF ores, iron is also present in iron silicates (grunerite, minnesotaite, chlorite, and stilpnomelane) and carbonates (siderite. ankerite. and ferroan dolomite). which are usually either not recoverable or are undesirable in downstream ironmaking processes. The ore has been subdivided into four ore types based on the percent of chert and other gangue between the magnetite-rich layers. with ore types 1 and 2 containing massive magnetite, whereas are types 3 and 4 contain more disseminated. fine-grained magnetite and a greater abundance of chert and other gangue (Povey and Leather. 1997).

Martite-goethite supergene ores

The martite-goethite ores are generally accepted as the products of recent supergene leaching and replacement of BIF (Morris. 1980. 1985). Martite-goethite ores are characterized by a predominance (>50%) of goethite over martite and well-preserved bedding from the primary iron formation. About 90 percent of the premining iron formation-hosted

1

IRON FORMATIONS AND ASSOCIATED IRON ORE DEPOSITS

FIG. 5. Photomicrographs of typical ore textures and microstructures for BIF, mamte-goethite. and rnicroplaty hematite ore. Plane-polarized reflected light images. E '" epoK)' resin. G "" goethite, Ceo;: goethite replacing carbonate. H '" hematite, M ., mamte. mpl H :: microplaty hematite, Mt = magnetite, OC = ochreous goethite, P = micropore.s-are black, Si '"' silicates-mainly quartz. and SVC .. silicates and carbonates. A. Coarse magnetite with inclusions of quartz, unenriched. Nammuldi Member BIF, Marra Mamba Iron Fannatian . Chichester Ranges, Western Australia; same location for (8 )·(D). B. BIF composed of alternating microbands of magnetite and chert-disseminated dolomite. C. BlF comprising disseminated magnetite in chert wUh disseminated dolomite. D . Fine-grained disseminated magnetite in chert BIF-particles mounted in epoxy resin. E. Martite-goethite ore, Marandoo deposit, Mount Newman Member, Marra Mamba [ron Formation . F. Martite--ochreous goethite ore, West Angelas A deposit, Mount Newman Member. Note the microbanding is similar 10 Ihal in typicaJ Marra Mamba BIF in (B). C. High phosphorous martite.goethite ore, Mount Tom Price section 6 , Brockman Iron Formation. H. Silicates psudomorphouse after ochreous goethite, Marandoo deposit, Mount Newman Member. I. Interocking network of hematite microplates, Mount Tom Price, DaJes Gorge Member, Brockman Iron Formation . J. Microplaty hematite-goethite ore, Iron Duke deposit, South Austraha. K. Martite-microplaty hematite ore, N4E depoSit, Carajas, Brazil. L. ltabirile-derived foliated hematite ore, Aguas Claras deposit, Quadrilatero Fenifero, Brazil .

657

resource for the Hamersley province is of Phanerowic supergene martite-goethite ore (Morris, 2002b). Some of the best examples of martite-goethite ores (Table 2) are hosted in the 2.60 Ga Marra Mamba Iron Formation (Fig. 8) in the Pilbara craton of West em Australia (Trendall et al ., 1998) and include

the Marandoo, Area C, and West Angelas deposits (Harmsworth et al. , 1990). Significant martite-goethite mineralization is also well developed in the stratigraphically higher Brockman Iron Formation (Fig. 8; e.g., Ophthalmia Range, Rhodes Ridge, Paraburdoo Eastern Ranges, Section 6 and 7 deposits).

658 cwur AND SIMONSON

Surlace

Depth of We,alherina

FIG. 6. Stylized cross section through the magnetite ore deposit of a t)pical Australian iron fonnation, showing depth of weathering and hematite cap to the deposit. In more intensely weathered areas, low-grade goethite may dominate over hematite in the deposit cap. Vertical scale ,. horizontaJ scale and depth of weathering is 5 to 50 m.

Hematile and cherl-free BIF Quartz mica schist Felsic volcanics

SOOm

- - - - Shear zone

r~~r~' m;" ~h;M DIF ,','. Dolerite

B

100m

-.----. Geological boundary (inferred) . . . Bedding trace

- . _ Base of weathering ~ Drillhole - Shear zone

Flc.7. Geologic block model and cross section of Mount Gibson (Extension Hill) deposit, Western Australia. Adapted from Lascelles (2002).

220 BOOLGEEDA IRON FORMATION

200 WOONGARRA RHYOLITE

400 WEELI WOLL! FORMATION

g 40=

~ 340 ~ u c

35 = -" .!l 120 "" 100-f-

45=

Yandicoogma Shale }

Joffre Member ~R~;KMAN Whaleback Sbale FORMATION Dales Gorge Member MT McRAE SHALE MT SYLVIA FORMATION

500 WITTENOOM FORMATION

235 Mt Newman Member } MARRA MAMBA MacLeod. Member IRON FORMATION Nammuldl Member

Flc. 8. Stratigraphic column for the Hamersley Group, Pilbara craton, Westem Australia. Adapted from Hannsworth et aJ . (1990).

In Marra Mamba and high phosphorous Brockman martite-goethite ores, martite has replaced magnetite micro bands (Fig. 5A-B) and magnetite disseminated in silicate andlor carbonate microbands (Fig. 5C), preserving the original outline of banding in the BIF (Fig. 5E-F). Preservation of banding also occurred due to iron enrichment where silicate and/or carbonate microbands are replaced by either brown goethite (Fig. 5A, G) or yellow ochreous goethite (Fig. 5B). Delicate brushlike silicate textures (Fig. 5F) and carbonate rhombs (Fig. 5F) from silicate-earbonate BIF are pseudomorphous after ochreous goethite.

The West Angelas A deposit (reseIVes in excess of 418 Mt at 62 wt % Fe; Rio Tmto, 2(04) is hosted within the Mount Newman Member at the top of the Marra Mamba Iron Formation, immediately beneath the Wittenoom Formation (Fig. 8; Harmsworth et al. , 1990; Bergstrand et al., 2(03). The deposit is located in a synclinal structure on the southern flank of the east-west-trending and west-plunging Won Munna anticline. Bergstrand et al. (2003) provided deSCriptive details of the West Angelas A deposit from which the follow account is given. The main ore minerals in this deposit and many other Marra Mamba ores are martite, hard brown goethite, and powdery yellow ochreous goethite that has histOrically been referred to as limonite. In the flat-lying section of the West Angelas A deposit (Fig. 9), the top of the mineralization is overprinted by areas of vitreous goethite hardcap. Hardc.p is a common term used in iron are geology to describe recent intense weathering of iron ore, up to 60 m below surface, consisting of highly porous or coarse cellular-textured vitreous goethite with high concentrations of Si and AI either substituted into the crystal lattice or occurring as inclusions of clay. Hardcap contains minor visible colloform-banded quartz. The hardcap at West Angelas is immediately underlain by hard martite-goethite hematite ore where extensive dehydration of goethite inftll has resulted in hematite fonnation (A, Fig. 9). The lumpy hard martite-goethite-(hematite) ore (>50 wt % hematite) passes stratigraphically down into underlying medium to friable goethite-martite are « 50 wt % hematite;


S ~----------------~

Water Table 600 m RL

o 100 m ---

Mineralized West Angela Member

Bedded martite-goethite

Bedded friable goethite

D Alluvium, colluvium

o West Angela Member

!1 Mount Newman Member

Wil Macleod Member rool •• ~ Hardcap

Flc. 9. GeologiC cross section looking west through the West Angelas A deposit, Western Australia. Letters A-O denote stratigraphic zones separated by thin shale units within the Mount Newman Member of the Marra Mamba Iron Formation. Adapted from Bergstrand et aI. (2003).

B, Fig. 9) and then into leached, friable to powdery, bedded ochreous goethite-martite ore «30 wi % hematite; C, Fig. 9), reflecting original variations in stratigraphic composition. Ochreous goethite-martite ore when friable contains denser brown goethite, either as cement between martite grains or as discrete pods, whereas powdery examples have alternating meso bands of leached yellow ochreous goethite with highly leached martite "blue dust."

The steeply dipping sections of the ore tend to contain more friable to powdery ochreous goethite-martite ore, due to secondary leaching by ground waters, compared to the less leached and dehydrated flat hard ore, whereas the moderately dipping ore is intermediate in hardness and mineralogy between the flat and steep ore (Bergstrand et al. , 2003).

Preservation of primary BIF layering is common, with martite pseudomorphous after magnetite micro- to mesobands, and silicate and/or carbonate micro- to mesobands replaced by goethite (Fig. 10). In the hard martite-goethite ore, coarse-grained (1-4 mm) martite is intergrown with goethite and few pores remain (Fig. lOA, I). Moderate leaching of martite-goethite flat hard ore has resulted in either porous martite (M in Fig. lOB) or martite-ochreous goethite ore (MOG in Fig. lOC). Friable ochreous goethite-martite ore in zone C consists of alternating bands of ochreous goethite replacing ex-silicate and/or carbonate microbands and martite grains replacing original magnetite microbands (Fig. 100). The gangue occurs as thin « 2 m) kaolinite-rich shale mesobands, whereas quartz is largely secondary and restricted to the weathered surface hardcap.

Circulation of ground water in the upper portion of the deposit may have been influenced by the presence of aquicludes such as shalelike bands (Bergstrand et al. , 2003). Iron-rich ground water is interpreted to have ponded above near-horizontal aquicludes, encouraging abundant secondary goethite to deposit in localized zones, and subsequently to partly dehydrate to hematite, thus causing the denSity of the host ore to increase (Clout, 2002; Bergstrand et al ., 2003).

Many high-grade hematite deposits in the QuadriIatero Ferrifero and Carajas districts of Brazil also contain a recent thin «30 m) goethite-martite hardcap and goethitic "canga" (Fig. 11), a term to describe loose detrital material and cemented martite-hematite conglomerate at the surface. However, the hardcap and canga may derive from goethite replacement of gangue in BIF or supergene ore during recent weathering (Spier et al., 2003).

High-grade hematite ores

The high-grade (>60 wi % Fe) hematite ores, which include martite-hematite and microplaty hematite-martite replacement ores, have quite variable characteristics (Table 3).

Hematite ores: The hematite ores are composed of residual martite and/or hematite thought to be derived from iron formation by leaching of gangue, leading to residual concentration of the iron minerals . For the most part, hematite ores contain very little (<15%) goethite, except in the surface hardcap, and may include some hematite of interpreted hypogene origin (e.g., Quadrilatero Ferrifero hard high-grade ores).


Flc. 10. T)1)ical hand specimen examples of marlite-goethite from West Angelus A deposit (A-D and I) and variOllS high grade mi croplaty hematite ores (E-H). Samples are (A) hard martite-goethite ore with macropores; (B) hard tlHlrtite-gocthite ore with low porosity (venter) and oute r leached (darker) fdab le martite ore; (C) contact between hard martite-goethile ore (dark, le ft ) and leached friable ochreous gocthitc-martite are; (D) laminations of martite (dark, after magnetite) with ochreous goethite (yellow, after silicate). (El to (Gl arc from Mount Tom Price, Dales Gorge Mem be r, Brockman I ron Formation; (I::) thin laminations of microplaty hematite (after chert ) in massive bands of dense and very hard interlocking-textured hematite (after magnetite); (F) hard martite-hematite ore with macropores (red-omnge); (G) alternating bands of tnicroplaty hematite-goethite (darker) with martite-microplaty hematite (l ight gray); and (1-1 ) porous mieroplaty hematite (slightl)' darker) alternating with well-jointed hard martite-microplaty hematite (lighter) b,mds, Channar mine, Joffre ~ll embcr; (I ) martite-rich hands after BIF magnetite-rich bands, whereas goethite- rich bands are after BIF sil icate an(Vor carbonate bands. C '" goeth ite. \ '1 == martite, mpll l :: microp[aty hematite, OC '" ochreous goethite, P = macropore.

The high-grade Sishen hematite deposit (1690 Mt at 64.8 wt % Fe; Carney and Mienie, 2002) in South Africa occurs immediately beneath a major regional erosional unconformity and grades downward into unmineralized BlF (Be ukes, 1986;

Beukes et aI. , 2002). High-grade hematite ore is only developed in the Asbestos Hills Subgroup where the unconformity transects BlF and the ore is inteqJreted to be of pre-Tertiaty residual supergene origin. The Sishen-type high-grade


S

........... "'c.. . . . . . .~: . . '.':' .. .. . .. . ..... . .

...... ...... , ......... ... . ....... . .. . .««.:-:-: . ".-:-: -: .' ,

.... . . . .. . ... ............. .

N

Itabirite

Dolomitic Itabirite Shale

• Goethite ore

D Soft high grade ore

_ Hard high grade ore

I t I Quartz ltabirite

hematite deposits are best developed where the host BIF has slumped into karstic structures in the underlying Campbellrand Dolomite (Fig. 12). A siliceous chert breccia (Wolhaarkop Breccia) marks the dissolution surface between the dolomite and overlying ore-bearing BIF. Bedding in the iron formation precursor is commonly highly contorted andlor brecciated due to slumping. Iron ores in the karstic slump structures are typically overlain by reworked conglomeratic iron ores and highly aluminous diaspore-rich shales and associated pisolitic lateritic profIles (Gutzmer and Beukes, 1998). Thin (1-2 m), high-grade hematite ore beds are locally presetved below the unconformity away from the karst slump structures .

FIG. 11. Schematic geologic cross section through the Aguas Clams deposit, QuadriMtero Femfero. Brazil. Adapted from Beukes et aI. (2002).

Beukes et aI. (2002) and Carney and Mienie (2002) described two major supergene ore types and their relative abundances: hard microcrystalline massive (58% of the total resource) and laminated (18%) martite ores thought to be derived from supergene residual enrichment of BIF below the Gamagara unconformity; and conglomeratic (detrital) ores (16%) derived from erosion of underlying laminated martite ore and concentrated in the lower part of the overlying Gamagara Formation. Carney and Mienie (2002) also described breccia ores (8%) comprising very angular and poorly sorted fragments of laminated and massive ores that fill sinkholes in the Campbellrand Dolomite. The microcrystalline laminated martite ore preserves original textures and banding of the precursor BIF and are interpreted as supergene residual concentration of martite. This is supported by oxygen isotope data for hematite that vary between --3 and +3

TABLE 3. Summary of Characteristics of Key Hi~h-Grade Hematite Del::itsl

Reserves Gt Interpreted timing D istrict (1.000 MI) Host rock Main ore type Accessory ore types of mineralization

Sishen. South Africa 0.17 Asbestos H;jj, Subgroup Hard massive hematite. Specular hematite Post metamorphism laminated hematite (60-66% Fe)

QuadriU.tero 3.3 CaUl~ Itabirite Friable hematite Specular hematite. magnetite. Premetamorphism Ferrlfero. Brazil (35-41% Fe) (64-68% Fe) dolomite·itabirite (32% Fe)

Carajas. Brazil 18.0 Carajas Formation Friable hematite Brecciated hematite-<lolomite. Synmagmatic (3.5-38% Fe) (66-08% Fe) laminated hematite-dolomite

(45% Fe)

Hamersiey, Australia 3.5 Brockman Iron Hard·friable hematite Magnetite.siderite.apatite. Postmetamorphism Fonllation (34% Fe) (64-68% Fe) hematite-ankerite-apatite

(44-68% Fe)

Krivoy Rog. Ukrame 4.7 Saksagan Suite Hematite Magnetite/magnetite-specular Postmetamorphism (36% Fe) (64% Fe) hematite. magnetite-earbonate.

magnetite-amphibole (50-57% Fe)

B.uadila, India 1.5 Bailadila Group Hard hematite Synmagmatic (35% Fe) (64-68% Fe)

Thabazimbi. 0.3 Penge Iron Fonnation Hard hematite Brecciated hematite-calcite Premetamorphism South Africa (36% Fe) (65% Fe) (45% Fe). laminated

hematite-<lolomite (42% Fe). hematite-talc

All percent values on v.rt basis I Modified from Dalstra and Cuedes (2004), Carney and Mienie (2002). and Hagemann et al. (in press)


--------- - --- - ------ -- ------ - -- --- -- --- ------- - --- --- - - ----------- - ----------- -- -- -- --- - ----- - -- -- - ----- --------- - - - -- - ---------- - ------- - ----- - --- -- -------- -- --- - - - --- - ----............ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

........ ~ ... ~. ~ T E is N 0 .,..,

I

1 -200-500m ,I

c

t::::::::~ Red and cream shale .§ ~. Unconformity -;;; Laminated Ore E

0 ~

r '" ~ c g .9 1» 1 Transgressive quartzite

I I AI-rich shale

~ I z ,.1 Cherty iron fonnation

'" a IT ... TI Wolhaarkop Breccia

'" -.. '" c E '" ~ ::E& '" 8 ~ Dissolution surface

:lIi-; Conglomeratic ore 0 ~ Campbellrand Dolomite

Flc. 12. Schematic geologic cross section through the 5ishen deposit. South Africa, showing karstic solution collapse struchues. Adapted from Beukes et aJ. (2002).

per mil (relative to SMOW), suggesting precipitation from surface waters at low temperature (Beukes et al., 2(02). However, localized coarse specular hematite infill of secondary pores and veinlets crosscuts both laminated and conglomeratic ores and suggests some secondary post-Camagara iron remobilization.

The Quadrilatero Fenifero hosts Significant (-17,500 Mt at >64 wi % Fe) soft, high-grade hematite and/or martite deposits, with resources of -10,000 Mt averaging 66 to 68 wi percent Fe. They are thought to have formed in part by residual concentration due to supergene leaching of carbonate and quartz from hard itabirite (meta-BlF) protore with about 40 wi percent Fe (Cuild, 1953, 1957; Dorr, 1965; Rosiere and Chemale, 1991; Pires, 2002; Beukes et al. , 2002; Cuedes et al., 2002; Ribeiro et al., 2002; Spier et al. , 2003; Rosi~re and Rios, 2004). However, only 15 percent of the total resource is of hard high-grade ore, thought to be of hypogene origin, which form smaller pods and lenses within the dominant (85% of total resource) soft, high-grade supergene residual ores (Fig. 11; Spier et al., 2003). The high hematite-content ores are hosted in the Caue Itabirite Formation of the Proteromic sedimentary Minas Supergroup, which unconformably overlies Archean greenstones. Itabirite in the Quadrilatero Fenifero has been deeply weathered to depths of up to 500 m beneath the surface. The district consists of

two structural domains: an eastern high-strain domain dominated by thrusts and mylonitic shear zones, with tight to isoclinal folds, and a western low-strain domain with well-preseIVed megasynclines, discontinuous shear zones, and thrusts (Rosiere and Chemale, 1991; Chemale et al., 1994; Rosiere et al., 2001, 2(02). Both the iron formations and hard, highgrade ores have been regionally metamorphosed and are quite structurally complex (Rosiere and Chemale, 1991; Rosiere et, al., 2001, 2003, 2004; Hagemann et al., in press).

At the Aguas Claras mine (-288 Mt), in the Quadrilatero Femfero district, high-grade (>64 wt % Fe) ore sits between hard itabirite and black phyllite (shale) and is interpreted to grade at depth into dolomitic itabirite and laterally into soft, hematite-rich itabirite in an overturned sequence (Fig. 11; Pires, 2002; Spier et al. 2003). The dolomitic itabirite is strongly banded with characteristic centimeter-scale mesobanding of carbonate and/or oxide layers and dominates in the mine area over siliceous itabirite found in the north wall of the mine (Spier et al. , 2(03). However, the origin of the dolomitic itabirite is controversial; proposed alternatives include a sedimentary facies variation of the Minas sediments (Dorr, 1965), diagenetic replacement of chert by carbonate (Spier et al., 2(03), and hypogene hydrothermal replacement C!f chert by carbonates (Dalstra and Cuedes, 2004). The Aguas Claras ore consists of pods of hard, high-grade

1

I


hematite are, of interpreted hypogene origin, within dominant soft, high-grade supergene are, with a thin goethitic are hardcap (Fig. 11; Spier et al., 2003). The soft, high-grade are mainly consists of residual martite, granular hematite, and locally foliated tabular hematite crystals (specularite; Fig. 5L) with rare gangue consisting of dolomite, chlorite, talc, and apatite (Spier et aI., 2003).

At least four principal textural types of high-grade hematite ore are recognized in the Quadrilatero Ferrifero, including the follOwing: (1) thin-bedded, laminated, and banded; (2) foliated, micaceous, and schistose, from high-strain domains; (3) minor brecciated; and, (4) massive andlor compact (Rosiere et aI. 2001; Pires, 2002; Rosiere and Rios, 2004). Although the ores are best known for a foliated hematite speculante texture (Fig. 5L), the most common texture is martite after magnetite. Magnetite porphyroclasts occur in schistose (mylonitic) are types in high-strain domains (Rosiere et aI., 2001, 2002; Rosiere and Rios, 2004). The friable hematiterich low-grade (40-58 wt % Fe) itabirite ore is composed of liberated hematite and quartz gangue and so is widely used as low-grade concentrator feed to produce a high-grade (>65 wt % Fe) concentrate (Spier et aI., 2003).

In the Pico mine, south of Aguas Claras, high-grade are and iron-rich itabirite are hosted within siliceous itabilite (Spier et aI., 2003). The soft, high-grade ores are well laminated and highly porous (30-45 vol %) and consist of microbands formed by aggregates of martite and hematite that alternate with highly porous martite-hematite microbands (Spier et aI., 2003). Ribeiro et aI. (2002) reported evidence of a collapSing process with subsidence of the soft, friable hematite ore follOwing dissolution and volume loss, including kink-bands and chevron structures.

Roserie and Rios (2004) present detailed fluid inclusion (infrared microthermometry), petrographic, and textural results on difTerent generations of hematite from the Concei~ao iron are deposit in the northeastern part of the Quadril~tero Ferrffero. The authors define three generations of hematite and specularite related to different deformational phases. Hematite I is composed of porous martite and new hematite crystals interpreted to have formed from oxidation of magnetite by low-temperature, low- to medium-salinity fluids of possible modified surface water origin, follOwing the collapse phase of the Transamazonian orogeny (2.1-2.0 Ga). However, no measurements cou1d be made on fluid inclusions in hematite I. Hematite II defines a granoblastic fabric in iron formations and high-grade ores representing a second episode of mineralization and subsequent recrystalization during regional metamorphism. Hematite II crystals grew from low-temperature and low- to medium-salinity hydrothermal fluids (based on large two-phase fluid inclusions with Th 115°_145°C and salinities eqUivalent to 4-10 wt % NaCI). Tabular hematite III is syndeformational and formed above 120°C (based on two-phase fluid inclusions with Th 120°-140°C) during the Brasiliano-Pan-African orogeny (0 .~.6 Ga). Finally, speculante that is composed of platy hematite crystals contains two- and three-phase fluid inclusions with Th of 140° to 205°C and dissolution of daughter crystals at _350°C; this formed in ductile shear zone-related schistose high-grade orebodies. Hematite III and speculante are both associated with high salinity (> 20 wt % NaCI equiv)

fluids, although unequivocal evidence for their origin is not presented.

Microplaty hematite ore: The high-grade replacement microplaty hematite ores are characterized by ubiquitous microplates of hematite (Fig. 51) and vanable hardness and porosity; and they occur with or without martite. The Mount Tom Price deposit in the Pilbara craton of Western Australia is an example of high-grade microplaty hematite ore, with the Original resource being 900 M t at 63.9 wt percent Fe of low phosphorous (0.053 wt % P) ore (Harmsworth et aI. , 1990; Taylor et al., 2001; Bitencourt et al ., 2002; Dalstra et al., 2002). The deposit is located in a synclinOrium along tlle northern limb of the regional Turner syncline on its eastern closure (Figs. 13, 14). The structure of the Mount Tom Price area is characterized by major thrusts and faults that parallel fold axes, with many open synclines and anticlines plus faults persisting along strike for 20 to 40 km (Harmsworth et aI., 1990; Bitencourt et aI., 2002). The Mount Tom Price mineralization is largely restricted to the Dales Gorge Member of the Brockman Iron Formation and the underlying Colonial Chert Member, altllOugh there is minor iron enrichment of the overlying Mount Whaleback Shale Member and the Joffre Member. Subvertical dolerite dikes cross the deposit subparallel to the major axis of the orebody (Fig. 14) and show locally intense chlorite-hematite-talc alteration. Magnesitedolomite veins and intense talc alteration in shale, ore, and BIF characterize the hydrothermal alteration along the Southern Batter fault and in the North deposit (Dalstra and Guedes, 2004; Thorne et aI ., 2004). Thorne et aI. (2004) have documented a complete hydrothermal alteration zone across mineralization in the North deposit (Fig. 13), with a distal zone of magnetite-siderite-iron silicate that grades into an intermediate zone of hematite-ankerite-magnetite, and finally into a proximal zone of martite-microplaty hematite-apatite which represents the main ore mineralization . Fluid inclusion studies on ankerite in hematite-ankerite veins from the distal alteration zone revealed mostly high salinity H,O-CaC!, pseudosecondary (23.9 wt % CaC!,equiv) and secondary (24.4 wt % CaCI, equiv) inclusions with mean homogenization temperatures of 253° and 117°C, respectively (Thorne et aI. , 2004). The carbon isotope signature of the carbonates is increasingly heavy from distal magnetite-siderite-iron silicate alteration (OIOC -8.8 ± 0.7%0) to the intermediate microplaty hematite-ankerite-magnetite alteration (OI'C -4.9 ± 0.70/00; Thorne et aI., 2004).

The microplaty hematite ore at Mount Tom Price varies from hard to medium massive hematite andlor martite-microplaty hematite, to friable ores, and to powdery and highly leached blue dust ore (Box et aI. , 2002; Clout, 2002). Alternating hard, medium, and friable hematite andlor martite micro- and mesobands persist laterally over a few meters (Fig. 10E, G, H ). Porous microplaty hematite mesobands (Fig. 10E) commonly alternate with hard and dense, interlocking mosaic-textured hematite (Fig. lOE) or residual martite (Fig. lOF, H). Although the microplates of hematite interlock in the harder ore types (Fig. 51), they just touch at their tips in the friable porous ore types. At shallow (0-40 m) depths below the present land surtace, the pores between hematite microplates may be filled with secondary goethite (Fig. lOG), which may be partly dehydrated back to hematite;

664 CT,Of.JT AI\'D SlMON SON

§u~.:.::;~~=

.. _-""""'-..........

""'" c::::J ".,. .~ sw.u;

C=:J WT.l'L\"" FORMAl1(f;

~ " T\"T1;.'<0CIM .OR. ...... nOf'

c::J =,:~'o:<)s c:::::J JEERIl'WI ......... n<;;.:

c:::J-.~ _H .. <; .... ' 11<_ _ ' _ _ ' ARs

~s,...-Am ---

FIG. 13. Geologic plan of the ~Iount Tom Price deposit, Western A%tral ia. After Taylor e\ at (200l).

Microplaty Hematite - Low P Microplaty Hematite - High P

c::J Magnetite - High P

SOUTHERN

BROCKMAN IRON FORMATION

§ Joffre Me mber WhaJeback Shale Me mber Dales Gorge Member

El Footwall Zone MT MCRAE SHALE

CJMTSYLVIA FORMATION

~WflTENOOMFORMATION

Bee Gorge Member Paraburdoo Member

FIe. 14. Geologic cross section of the Mount Tom Price de posit at 13002E, looking northwest. After Taylor , ~t al. (ZOO]).


similar textures occur in the Iron Duke iron ore deposit in South Australia (Fig. 5J; Clout, 2(02). The blue dust ore consists of largely 0.2- to 0.03-mm corroded and leached plates of hematite and minor porous martite. At Mount Tom Price, Taylor et al. (2001) and Dalstra et al. (2002) documented magnetite-carbonate mineralization with high phosphorous concentrations at greater than 200 m below surface (Fig. 14), marginal to the microplaty hematite mineralization that they interpret as protore. Magnesite-dolomite veining and intense talc alteration in shale, ore, and BIF characterize the strong Mg-Fe metasomatic alteration along the Southern Batte r fault (Dalstra and Guedes, 2004).

The Mount Whaleback deposit is the largest iron ore accumulation in Australia, originally having in excess of 1,800 Mt of resources at 64 wt percent Fe. It lies in a faulted outlier of the Hamersley Group (Harmsworth et al., 1990; Brown et al., 2004). The deposit is structurally complex with the orebody defined by the westerly plunging overturned East and South synclines. The northern limit of the ore is truncated by the southeast-dipping Mount Whaleback fault that juxtaposes the Brockman Iron Formation to the south against the older Jeerinah Formation to the north. The Mount Whaleback fault has a normal sense of movement, and several low-angle normal faults branch off of it and cut across the orebody. As with Mount Tom Price, the ore is largely developed in the Dales Gorge Member as medium to hard microplaty hematite-martite. Ore is developed to a lesser extent in the Joffre Member as softer fissile microplaty hematite ore with locally more goethite or as highly leached blue dust microplaty hematite ore. Minor iron enrichment also occurs in the upper section of the Colonial Chert Member of the Mount McRae Shale.

A third example of high-grade microplaty hematite ore is the deposits of the Caraja. district, Brazil, which contain -17,500 Mt with >64 wt percent Fe hosted by the Caraja. Formation (Gibbs et al ., 1986; Beukes et al. , 2002; Guedes et al ., 2002; Lobato et al., 2004; Rios et al., 2004; Rosiere et al.,

v v v v

h:~·.1 Surface laterite and canga

~ Fresh lava

_ Fresh iron formation

v

v v

~ Fresh carbonate-hematite rock

2004; Silva et al., 2004). The Caraja. Formation comprises discontinuous sedimentary layers and lenses of partial to completely dolomitized BIF and lenses of high-grade hematite ore, cut by mafic sills and dikes. In the N4E mine area, dolomite has locally replaced BIF chert along the banding and also occurs as irregular veins and hydraulic breccias (Guedes et aI., 2002; C.A. Rosiere, pers. commun., 2005). Near the contact with hard ore, the underlying volcanic rocks are typically altered and partially mine ralized with dilational breG"Cias and vugs filled with carbonate, quartz, kaolinite, and microplaty he matite, quartz-hematite veins, and fibrous aggregates of chlorite (Guedes et al. , 2(02). High-grade ore occurs as tabular bodies of friable to soft hematite that contain smaller lenses of hard hematite (Fig. 15). The friable hematite ore occurs both as powdery hematite, almost devoid of internal structure, and as millimeter-thick bands of finegrained hematite. In the N4E mine, Guedes et aI. (2002) have documented idioblastic martite surrounded by very finegrained (-10 /lm) microplaty hematite (Fig. 5K). The hard hematite orebodies with >66 wt percent Fe contain microplaty hematite and occur mainly near the contact with the underlying metavolcanic rocks, where they are surrounded by an aureole of hydrothermal carbonate alteration. Relict BIF bedding is generally preserved in the hard ore with dense hematite alternating with porous hematite.

A number of studies have documented an earlier carbonate protore for high-grade hematite ore. For example, Beukes et aI. (2002) recognized an early phase of metasomatic carbonate-bearing ores at the Thabazimbi high-grade hematite deposit of South Africa, hosted by the Penge Iron Formation of the Transvaal Supergroup. Fluid inclusion studies on carbonates and quartz from Thabazimbi indicate mixing of two distinct hydrothermal fluids; one is a high-salinity fluid responsible for deposition of early dolomite at lSO° to 100°C, the other is a low-salinity fluid that led to precipitation of quartz at 120° to 140'C (Netshiozwi, 2(02). Beukes et aI. (2002)

v

v

v v v v

v v v v v v v v v v

v 200m v v

I } I Goethitic ore

I MiMI Friable ore

• Hard ore

v

v v

f::::=::::j Friable iron foonalion

I'v, I Weathered lava

v

v

Flc. 15. Schematic cross section through the N4E deposit, CarajAs district. Brazil. Adapted from Beukes et at. (2002).

---------------------- --~------------------------------------~1


used fluid inclusion studies and oxygen isotope data on hematite and calcite to suggest that the primary hematite ore formed at 160°C from a hydrothermal fluid of -2 per mil (relative to SMOW). Although they speculate this fluid was surface water that had not exchanged with silicate rocks, hydrogen isotope data are required to unambiguously define fluid provenance.

Ore Genesis

Genetic models for iron ores related to iron formations continue to attract controversy, especially for the high-grade hematite iron ore depoSits, even though many core concepts of current models were established as far back as the late 1800s to early 1900s (Dorr, 1965, 1969; Morris, 1985, 2002b). Recent reviews of and contributions to iron ore genesis models include those by Taylor et al. (2001), Beukes et al. (2002), Morris (2002), Spier et al. (2003), Rosiere and Rios (2004), and Hagemann et al. (in press). The controversy is not surprising, given the wide variety of ore types and structural settings (described above), the complex textures of ore and gangue, the lack of data to constrain ore fluid compositions (e.g., limited fluid inclusion work), and the large (-5-20 km') size of the orebodies.

The key genepc models for high-grade hematite and martite-goethite fall into one of three 'categories: (1) early syngenetic and diagenetic processes (King, 1989; Lascelles, 2(02); (2) hypogene alteration and replacement involving deep-seated hydrothermal andlor magmatic fluids (Dorr, 1965; Brandt, 1966; Kneeshaw, 1975; Gutzmer et al., 2(02) or shallow meteoric waters and basinal brines (Hagemann et al., 1999; Powell et al., 1999; Taylor et aI., 2001 , Buekes at aI., 2(02); and (3) supergene events ranging from pre-Tertiary with (Morris 1985, 1993, 2(02) or without (Van Schalkwyk and Beukes, 1986) subsequent burial metamorphism in the Mesozoic or Cenozoic (Dorr, 1964; MacLeod, 1966; Morris, 1985; Harmsworth et al., 1990). Despite Significant differences among models for the origin of the high-grade hematite ores, there is a general consensus that the martite-goethite ores have fonned as a result of recent supergene enrichment of iron formation beneath Cretaceous to Tertiary weathering profiles (Morris, 1980; Harmsworth et al ., 1990; Beukes et al. , 2(02), and that high-grade hematite ores have been further upgraded by recent supergene enrichment (Taylor et al., 2001; Ribeiro et al ., 2002; Dalstra and Guedes, 2004). Present debate is centered on the various composite hypogene-supergene models for different high-grade hematite deposits.

Syngenetic

Syngenetic models assume a clastic origin for iron formation-hosted are, with or without diagenetic concentration, whereas composite syngenetic models include later modification by metamorphism, igneous activity, or supergene processes (King, 1989; Lascelles, 2002). Syngenetic processes are thought to prOvide the initial magnetite-rich iron formation and are discussed earlier under models for the depOSition of large iron formations. However, it is generally accepted that the syngenetic model is unable to account for the speCific location of martite-goethite ore beneath recent weathering surfaces, or the structurally controlled microplaty hematite ores that are commonly developed in fold structures, around normal faults andlor back thrusts, or in brittle shear zones.

Supergene

Supergene models interpret the ores to be residual concentrates formed from leaching of gangue in the iron formation by deep circulating ground water below either current or past erosion surfaces but do not imply that it is a lateritization process (Morris, 1993). Morris (1980, 1985, 1986, 1987, 1993) presented the concept that the martite-goethite ore bodies grew toward the surface via supergene metasomatic replacement of BIF gangue minerals by hydrous iron oxides, driven by a massive hydrodynamiC electrochemical cell. Magnetite is oxidized to hematite (martite), preserving the original magnetite crystal outlines. For enrichment to occur, the model requires a folded, fractured, or faulted BIF structure that forms an open-€nded artesian system, allowing ground-water access to BIF well below the surface (Fig. 16). In this model, the magnetic layers in the BIF are thought to have acted as electron conductors, whereas ground water seIVed as an ionic transfer agent driven by cathodic reactions in the upper BIF zone during wet seasons (4e- + 0, + 2H,O -+ 40H-). Iron leached from the friable silicate facies BIF at the surface was mobilized as ferrous iron by biogenic reactions in the vadose zone and transported deeper into the system. fu; water flowed through the fold or fault structure, chert, silicates, carbonates, and other gangue minerals were gradually replaced andlor leached from the BIF, locally resulting in substantial stratigraphic thinning. The ore-forming process generated significant macro- and microporosity, thereby helping to create its own fluid pathways. Pseudomorphic replacement of gangue at depth by goethite was achieved through anodic oxidation of ferric iron (Fe" -+ Fe" + e-) followed by ferrolysis (Fe" + 3H,O -+ Fe(OH), + 3H·). Silica is more rapidly released into solution at depth as a result of seasonal cyclic iron redox reactions with quartz. However, Ohmoto (2003) has demonstrated that the transfonnation of magnetite to hematite or vice versa can also be achieved via a pH shift without a redox reaction.

Supergene processes are generally accepted by many researchers (e.g., Morris 1980; Harmsworth et al., 1990; Taylor et al., 2001; Clout, 2002; Dalstra and Guedes, 2004; Thome et al. , 2004) to be responsible for the typical Pilbara Marra Mamba martite-goethite ore and high phosphorous Brockman martite-goethite ore, as well as final upgrading of high-grade hematite ores. The Pilbara martite-goethite ores are interpreted to have an origin related to supergene processes beneath the Mesozoic-Tertiary weathering surface. They have limited downdip extension below current outcrop, although a few deposits (e.g. , West Angelas) extend downdip >250 m (Harmsworth et al., 1990; Morris, 2002b). Harmsworth et al. (1990) and Morris (2002b) suggested that simple supergene leaching of silicates is locally responSible in high rainfall areas for leaving a residue of both (martite) blue-dust ore in India and friable quartz-hematite itabirite iron ores in the Quadrilatero Ferrffero. However, most workers consider that only the soft, high-grade hematite ore from the Quadrilatero Ferrifero deposits formed by residual concentration due to recent supergene leaching of carbonate and quartz from hard itabirite protore (Cuedes et al., 2002; Pires, 2002; Ribeiro et al., 2002; Spier et al., 2003; Rosiere and Rios, 2004). Another example are the Sis hen-type deposits in South Africa, which occur immediately below a major erosional unconformity and


BANDED IRON FORMAnON

HEMATITE CARBONATES SlUCATES leachino BLUE DUST I + + ORES

MAGNETITE SIUCA

T • r wpe<geoe enrichment - ~ r oxidation I metasomatic replacement . by goethite , I burial I -..r HEMATITE-GOETHITE ORE regional metamorphism I

+ r mod;ty by~:",ch;ng l..-r ,...~u~ ~ HEMATITE-RICH ORES ordeh rahon , I contact l me

• I VARIOUS ORE TYPES I Y MAGNETITE-RI HORES I

• I surficial processes - eluvial concentration 1 t

"HARD CAP", "HYDRATED ZONE". "CRUSTAL ORE". "CANGA

A. System viewed in 3D

C. Transfer of Fe to anode

E. Transfer of Fe to anode

B. Electrochemical cell 4.- + 0, +? I~"rl "'~ 4{OH)-

Electrical conductor (magnetite)

D. Transfer of Fe to anode

F. Final supergene ore body now subject to leaching by groundwater

Flc. 16. Block diagram to explain the fonnation of supergene iron ore developed from BIF. From Morris (1998).

grade downward into unmineralized BIF. interpreted as preTe rtiary supergene ores (Beukes at aI., 2(02).

Supergene ores and subsequent burial metamorphism Supergene are subsequently overprinted by burial meta

morphism was used by Morris (1980, 1985, 2(02) and

Hannsworth et aI. (1990) to explain the genesis of high-grade microplaty hematite iron ores worldwide. Morris (1980) suggested that some old martite-goethite supergene ores were subjected to burial metamorphism to diagenetic levels (-I00"C) and tl,at dehydration converted supergene metasomatic goethite either partially or totally to microplaty

668 CLOUT AND SiMONSON

hematite (Fig. 16). Exposure of microplaty hematite ore bodies during the Mesozoic then resulted in the dissolut ion of most of the unconvC licd goethite by ground water and partial dissolution of hematite to leave a compact to highly porous high-grade hematite are.

S1lpergelle-modified hypogene ores

There is currently widespread support for a h)lmgene-hydrothcrrnal origin for upgrading of iron formation to highgrade hematite ore, especially for microplaty hematite deposits from Australia and hard, high-grade hematite deposits from Brazil (Barley et ai. , 1989; H agemann et aI., 1999; Oliver and Dickens, 1999; Powell et aI. , 1999; Taylor et al., 2001; Beukes et ai., 2002, Webb et aI., 2002, Spier ct aI. , 2003; DalslTa and Cuedes, 2004; Hosiere and n ios, 2004; Thorne et aI. , 2004 ). These models also include later modillcation of highgrade hypogene hematite ore by recent ~upergelle upgrading. Despite wide support for a h)11ogene Oiigin, there are considerable differences between the various h}1)ogene models proposed; these include combinations of ascending and/or desccnding hydrothermal fluids that include warm basilMI

Soutnern

-S"l<f F ... i!

r: 0 Mag''''tite-<ideri!C-iroo si l i~a[e alteration

Ascending basinal brmcs (150-25ifCj I dolerite

Stage Ie Late hypogene alteration

Wanin.,:. a.cending Nt,in:1l brine(-120' C) I NW.trerlding dolerite

d ike

brines and/or heated meteoric water (Barley et aI. , 1999; Hagemann et al., 1999; Taylor et aI., 2001 ). The models differ in the relative timing ofh)1Xlgene alteration and deformation, the importance or universality of a carbonate protore, and the t)1)e(s) of fluids responsible. Some of the key proponent models are presented below.

The Ilrst detailed model, presented by Barley et al. (1999), Hagemann et al. (1999), and recently updated by Thorne et a1. (2004), involves two-stage hydrothermal and supergene processes; it is based on fluid inclusion, hydrothermal alteration, and stable isotope studies at the North deposit in the Pilbara. The earliest stage (la), hypogene altemtion involving upward movement of hydrothermal brines (IS0"'- 250"'C), tnmsformhl 35 wt percent Fe BIF to a magnetite-sideriteiron silicate ElF with desilieifleation of the chert bands (Fig. 17). Stage Ib hypogene alteration resulted from ascending basinal brines \vith higher temperatures (possibly up to 400"'C ) that imhlCed hematite-ankerite-magnelite alteration and finally the formation of microplaty hematite. The 300'" to 350°C trapping temperatures proposed by Thome et a1. (2004) for pseudosecondary fluid inclusions in stage Ib are

Ib Earl~' hypogene alteration

Sou{h~m

r =nding Ntsin~l brines (200-300' C)

Stage 2 Supergene alteration

Sou thern Ridge

I Descending <hallOw mctcroic waters «100' C)

o Hema{i{e-ankeri{cmagnetite alteration

I ' .... :W. trending doleri{e

• Manite-micropl,uy 'goe{hi{e al'cmtion

I NW,uending dolerite ,

FIG. I i. Schematic block diagram from Thome et aJ. (2004) to explain the stages of hypogene ~nd supergene alte ration for high.grade hematite ore fonnation at the North and Southern Ridge deposits. !l"lount Tom Price. A. Stage la, early hy. pogene magnetite .sideri te-iron silieate formed by ~~~-.;ndi ng 150° to 251YC basinal brines. B. Stage Ib, ea rly hYP3gene hematite-ankerite-magnetite aheration formed by ascending 300° to 400"C hasinal hrines. C. Stage le, ble martite-rnieroplaty hemat ite -apat ite alteration formed by ascending _ 120°C basinal brines. D. Stage 2, supergene mart ite-microplaty hemati te.goethite alteration fonned by de>l:ending metL"Qrie waters « l OO~C).


too high for basinal fluids alone and are more likely to involve a high-temperature magmatic component, although further isotopic studies are required to demonstrate this. Stage 1c hypogene alteration involved the interaction of low-temperature (-120°C) ascending basinal brines that formed a hematite-ankerite-magnetite assemblage; this resulted in dissolution of ankerite to leave a porous martite-microplaty hematite-apatite assemblage. Finally, stage 2 involved supergene enrichment by descending meteoric waters (dOO°C) during the Tertiary, resulting in removal of residual ankerite and apatite, goethite alteration, and the weathering of shale bands (AI silicate BIF) to clays. Ion chromatography investigations on inclusion fluids revealed that quartz-hematite veins contain Na > Mg > Ca > K as major cations and that anionic ratios such as Br/CI, IICI, and CIISO, show a close affinity to Canadian Shield brines but are incompatible with ratios for typical igneous and metamorphic fluids or seawater (Hagemann et al., 1999; Taylor et al., 2001). Data on hydrogen and oxygen isotopes of inclusion fluids and oxygen isotopes of vein quartz at Southern Ridge suggest the involvement of basinal brines and only a minor amount of meteoric wate r late in the hydrothermal history (Hagemann et al. , 1999). Hagemann et al. (in press) extended the work of Thorne et al . (2004) by proposing a discrete model for the genesis of worldwide high-grade hematite deposits involving three end members based on the diversity of geolOgical and geochemical features, tectonic setting, distinct hydrothermal fluid source(s), and processes.

Taylor et al. (2001) presented a four-stage model for highgrade hematite but with microplaty hematite formation from meteoric water, as follows. Stage 1 involved initial upward migration of reduced basinal brines that resulted in hydrothermal replacement of primary BIF silicates with siderite to produce a magnetite-carbonate-apatite protore. This was followed by deep circulation of oxygenated low-salinity meteoric water, stage 2, that oxidized the siderite to rnicroplaty hematite and magnetite to martite, to form microplaty hematite-martite-apatite-ankerite mineralization. However. since meteoric water has a very low concentration of dissolved oxygen, this would require Significant volumes of fluid. Stage 3 involved leaching of carbonate and remaining silicates to microplaty hematite-martite-apatite . Finally, apatite was leached during stage 4 by supergene processes to form high-grade microplaty hematite ore (Taylor et al. 2001; Dalstra et al. 2002). Taylor et al. (2001 ) believe that the mineralizing process took place during a period of uplift and extension that postdated the Ophthalmian orogeny but before the end of the Proterowic, since microplaty hematite ore at Channar was contact metamorphosed by a dolerite dike dated at 752 ± 10 Ma.

Li et al. (1993), Martin et al. (1998), and Powell et al. (1999) proposed a hypogene synorogenic hydrothermal model for Pilbara microplaty hematite ore, envisioning mineralizing fluids as being derived from mixing of oxygenated meteoric water with basinal fluid expelled from deeper levels of a foreland basin during the regional compression phase of the Ophthalmian orogeny (2.20--2.45 Ga). Oliver et al. (1998) also invoked synorogenic interaction of deep-seated orogenic fluids with descending supergene waters in Proterowic times. Muller et al. (2005) suggest a maximum age for hypogene iron ore mineralization in the Hamersley province of

between -2050 and 2000 Ma, eonstntined by PblPb dating of baddeleyite which yielded 2008 ± 16 Ma for a mafic dike swarm that intrudes the Lower Wyloo Group, but older than the Mount McGrath Formation which contains clasts of microplaty hematite mineralization .

Dalstra and Guedes (2004) proposed that all high-grade hematite deposits form a coherent genetic group, and they presented a model in which an early magnetite-carbonate-apatite protore formed by hydrothermal depletion of silica in the BIF and introduction of Ca-Fe-Mg carbonates by heated alkaline brines, with subsequent supergene upgrading. From mineral assemblages for hydrothermal carbonate protore, Dalstra and Guedes (2004) suggested that the temperature of ore form ation varied over a wide range from high (>400°...,s00°C) for magnetite-cummingtonite-siderite (e.g., Kivroy Rog in the Ukraine) to medium (300o_<400°C) for chlorite-talc (e.g., Mount Tom Price) to low « 300°C) for hematite-dolomite (calcite; e.g., Caraji<).

Guedes et al. (2002) suggested that high-grade hematite ores at the Caraji< N4E mine were derived from supergene leaching and residual concentration of hematite during weathering of hypogene hydrothermal carbonate-hematite rock derived from siliceous itabirite. In contrast, Spier et al . (2003) interpreted the carbonate as a primary constituent in the sense that the carbonate-hematite rock originated as dolomitic iron formation interbedded with siliceous itabirite. This interpretation is supported by the gradational contact between itabirite and the overlying Gandarela Formation, which contains shallow-water carbonate sediments. Spier et al . (2003) a1<o pointed ou! that, although the re is evidence for dolomitic protore at the Aguas Claras depoSit, there is no evidence for a carbonate protore at the Pico depoSit, only siliceous itabirite. Spier et al . (2003) further suggested that supergene leaching produced both the soft high-grade hematite ores and iron-rich itabirite from primary dolomitic itabirite and siliceous itabirite, respectively, whereas the hard high-grade hematite ores (<20% of reserves) -are of hydrothermal origin.

Rosiere and Rios (2004) presented a detailed study (summarized above) of fluid inclusions in hematite and petrographic evidence to support synorogenic formation of hard massive and schistose high-grade hematite ores from the Quadriliitero Fenifero. Rosiere and Rios (2004) defined four stages of recurrent hypogene mineralization characterized by three generations of hematite and a final schistose specularite stage. The mineralizing fluids are thought to have evolved over time from low-temperature, low- to medium-salinity fluids that may represent meteoric water and were modified to become high-salinity (4--10 wt % NaCI equiv) hydrothermal fluids. The deformed and metamorphosed nature of the QuadrilMero Fenife ro deposits is in contrast to postmetamorphic un deformed high-grade hematite ores from the Pilbara (post-Ophthalmia orogeny).

In summary, distinct differences between ore deposits make it very difficult to formulate a single unifying model. For example, some deposits G'Ontain carbonate protore (e.g., Mount Tom Price, Caraji<, Thabazimbi), whereas it is absent in others (e.g., Mount Whaleback, Pico). Likewise, hypogene mineralization is metamotphosed in some deposits or districts (e.g., QuadriJatero Fenifero, Thabazimbi) but not others


(e.g., Pilbara andlor Hamer-dey ores, Carajas, and Krivoy Rog). Moreover, a wide variety of fluid types, volumes, temperatures, and sources have been invoked to explain diffe rent hypogene mineralizations, e.g., basinal brines versus meteoric water (Hagemann, 1999; Beukes et al ., 2002; Spier et al., 2003; McLellan et al. , 2004; Rosiere and Rios, 2004; Hagemann et al., in press) or even magmatic fluids for Carajas (Silva et al. , 2004; Lobato et al ., in press). The geochemical processes thought to be involved in upgrading BIF to highgrade hematite ore include early desilicification and carbonation, followed by decarbonatization and hydrothermal replacement of magnetite to martite by oxidation (Hagemann et al. , 1999; Powell et al ., 1999; Taylor et al. , 2001), or a pH shift due to leaching of Fe" from magnetite (Ohmoto, 2003). Late supergene leaching of residual gangue and goethite overprinting are also key parts of the hypogene models.

Hypogene versus supergene origin with subsequent burial metamorphism

Morris (2oo2a, b ) interpreted the residual carbonates beneath the main Mount Tom Price deposit and associated Southern Batter deposit to have been localized by postore metasomatism of BIF and ore, whereas Kneeshaw and Kepert (2002) and Kneeshaw et al. (2002) contend that carbonate protore is ahsent at Mount Whaleback, thus not supporting the Taylor et al . (2001) model. Hagemann et al. (1999) and Thome et al. (2004) suggest that the BIF at Mount Tom Price underwent initial carbonate replacement of chert by a basinal fluid, followed by conversion of iron si~cates to proximal microplaty hematite are by hot oxidized basinal brine. In contrast, Webb et al. (2002) suggested that either BIF was affected only by the latter processes or the carhonate alteration has yet to be found at Mount Whaleback. Proponents of hypogene models cite the absence of carbonate alteration from major deposits (e.g., Mount Whaleback, Pico) as evidence that carbonatization is not a necessary precursor for the formation of large, high-grade hematite deposits (Spier et al., 2003; Hagemann et al ., in press).

Morris (2oo2a, b) cited a number of other problems with the carbonate protore model of Taylor et al. (2001). These include the large amount of basinal fluid that would be required to desiliciJY the BIF at Mount Tom Price, a time lag of some 600 m.y. between basinal fluid generation and BIF enrichment, and the absence of microplaty hematite from the Marra Mamba Iron Formation, which sits below a potential dolomite aquifer, the Paraburdoo Member. These concerns are in part countered by Taylor et al. (2002), who argued that hypogene mineralization is later than that suggested by Morris (2oo2a), and the specific role of the Paraburdoo Member in chanelling fluids into the overlying Brockman Iron Formation.

Morris' (1980) model requires the chert in BIF to be replaced by goethite and subsequently metamorphosed to microplaty hematite. In arguing against the Morris (1980) supergene-metamorphic model, Taylor et al. (2001) use simple volume and assay calculations to suggest that iron has not been added overall during the mineraliZing process, although this assessment would be more definitive if immobile element pairs and mass-transfer calculations were used. Taylor et al. (2001) also argue that, according to the Morris (1980) model, the magnetite-carbonate-apatite mineralization at Mount

Tom Price should not show a spatial relationship to highgrade hematite mineralization, and that the high-phosphorus carbonate-microplaty hematite mineralization should contain remnant goethite, which is not observed. Moreover, fluid inclusion studies (Hagemann et al ., 1999; Spier et al ., 2003; Webb et al. , 2003; Thome et al., 2004) indicate temperatures of > 100°C for ore formation and the presence of basinal brines and modified meteoric water, neither of which fit with the burial metamorphic origin for high-grade hematite ores at _100°C, as proposed by Morris (1980, 1985, 1993, 2oo2b).

Structural and hydrodynamic controls on are jomwtion

Many authors suggest that other relationships are also important in ore formation, including the presence of early, lowangle normal Iystric faults (Taylor et al., 2001 ), fold hinges with enhanced permeability and deep faults (Rosiere and Rios, 2004), or other favorable structures able to serve as hydrothermal fluid conduits (Spier et al. , 2003); location of the main ore bodies near the base of an iron fannation succession in contact with black shales which cap underlying dolomitic carbonates (Beukes et al., 2002); and the presence of impermeable shales and dolerites that acted as hydrolOgical seals to focus ore formation (Beukes et al., 2002).

Processing and Products

are mineralogy and beneficiation

Iron jOn1Ultion ores: The majority of BIF and GIF ores require expensive fine grinding to 20 to 75 I'm in order to liberate the iron oxides (magnetite or hematite) from silicate (quartz, stilpnomelane, amphibole, chlOrite) or carbonate (siderite, dolomite, ankerite) gangue. Taconites from North America require extensive beneficiation of 30 to 35 wt percent Fe feed to make fines (65-67 wt % Fe), blast furnace pellet, or direct reduced iron feedstock grades (>68 wt % Fe; Coyle, 1965; McKim, 1970; DeVaney, 1985). In taconites that are easier tOJrocess. the ore and gangue minerals may be coarse graine (0.05-2.0 mm) and low in porosity, and the ore minerals may be relatively free of very fine « 5 I'm) gangue inclusions. They are primarily tl,e result of metamorphic recrystallization to relatively coarse grain size (Neal, 2000).

A measure of the ability of magnetite BIF and GIF to be upgraded by simple grinding and magnetic separation is just as essential at the evaluation stage of exploration as an assay. This is because iron can also be tied up in silicates that are not of economic value and the magnetite may be so fine grained that it is uneconomic to grind and separate from gangue (Fig. 5D). Iron formations that lack coarse iron oxide grains (Fig. 5D) or contain very fine «20 I'm) gangue inclusions (Fig. 5A) will e ither be subeconomic or marginal, even though the in situ resource grade may exceed what is typically an attractive 45 wt percent Fe. In Figure 18, ore types 1 and 2 contain massive magnetite micro bands and so easily reach >68 wt percent Fe product grade after coarse grinding and magnetic separation, whereas ore types 3 and 4 contain Significant finegrained disseminated magnetite in chert andlor silicate microbands, thus requiring much finer grinding to reach the same product grade.

To date, few geolOgiC criteria have been published that help to target exploration toward iron formation deposits with


72

70 Ore Type 1 & 2

68 CD

68 1ii L -c

II.!'

.~.

.4 64 (I) 0 c 62 /Y' 41 0 0 c 60 (I) Ii.

oreTypeV / • Ore Type 1 and 2

r/, • Ore Type 3 58 ;ii:

56 / Ore Type 3 ... Ore Type 4 •

54

60 70 80 % -45 microns

90 100

FIG. 18. Change of Fe grade of concentrate with increasingly fine grinding for two-stage fme grinding of magnetite BIF ore types. Mount Gibson deposit, Western Australia. Note that the Him shows increasing fineness of the grinding. Ore types I and 2 contain massive magnetite and easily achieve >69 wt percent Fe grade with minimaJ grinding. whereas ore types 3 and 4 contain more disseminated fine-grained magnetite and require finer grinding. Davis tube magnetic separation test. Adapted from Povey and Leather (1997),

coarse-grained or more abundant magnetite that will improve economic processing characteristics. Further assessment of iron formation genesis is required to better understand basinwide controls on thick iron oxide versus thinner gangue (silicate and/or carbonate) deposition as well as the impact of higher regional metamorphic grade.

For magnetite iron formation. the most low cost and effective method used for separation of magnetite from silicate and/or carbonate gangue is low-intensity magnetic separation. using cheap rotary permanent-magnet drums. At the Empire. Hibbing. and Northshore mines in North America. the principal separation techniques are magnetic separation and minor reverse flotation that uses a cationic collector to float liberated quartz and locked quartz-magnetite particles (Coyle. 1965; Graber and Sundberg. 2(02). Size classification using hydrocyclones is also required to remove the ultrafines «0.01 mm) that are rich in silicate gangue. At the Diao Juntai magnetite mine in northeast China, a combination of grinding. low-intensity magnetic separation to recover magnetite. wet high-intensity magnetite separation to recover minor hematite. and reverse flotation is used to produce a 67.5 wt percent Fe concentrate from 29 to 30 wt percent Fe feed.

In contrast, hematitic iron formation often requires more extensive separation techniques. These mainly include wet high-intensity magnetite separation to concentrate fine (0.034l.08 mm ) paramagnetic minerals including hematite; spirals for gravity concentration of hematite; and hydrocyclones to remove gangue-rich ultrafmes. as well as reverse flotation. Overall, hematite iron formation is more expensive to beneficiate than magnetite iron formation. Cheaper wet gravity separation techniques including spirals are an important means of separating 0.05 to 1.0 mm low specific-gravity quartz from well-liberated hematite at a number of mines. including Wabush. Mont Wright. and Humphrey in Canada and

Dong Anshan in northeast China. These three hematite mines from Canada are examples of intensely metamorphosed coarser grained taconites (BIF-GIF). where coarse (specular) hematite and minor magnetite are the dominant iron oxide mine rals; hematite is relatively free of gangue inclusions. and thus high (>67 wt %) Fe grade concentrates can be produced (Neal. 2(00).

High-grade hematite and fJUlrtite-goethite ores: Although high-grade hematite and martite-goethite ores contain relatively few iron-bearing minerals other than hematite and goethite. complex ore and gangue textures and a large range in porosity result in quite variable requirements for beneficiation compared to iron formation ores (Clout et aI .• 1997). The first stage of processing for high-grade hematite and mamte-goethite ores is crushing and screening. The run-ofmine ore is crushed then screened into lump and sinter fines. After blasting. crushing. and screening. ores that are hard to medium in relative physical strength produce about 40 to 60 wt percent lump and the remainder is fines. In contrast. friable ores typically produce less than 30 wt percent lump. Many high-grade microplaty hematite (e.g .. Mount Tom Price. Mount Whaleback, South Middleback Ranges. Carajas) and martite-goethite (e.g .• Marandoo. West Angelas. Area C. Koolyanobbing) deposits are of sufficient iron grade to require only crushing and dry screening before direct shipping of lump and fine ores to customer steelworks. However. wet beneficiation plants are required at some deposits in Brazil. Australia. India. and South Africa to upgrade the ore to produce blast furnace-grade lump and sinter fines.

In martite-goethite and microplaty hematite deposits of Australia. a high percentage of gangue occurs as thick (0.1-6 m) bands of soft and porous kaolinite-rich shale that are easily separated from hematite ore by selective mining. gravity separation. or washing away of fme clay particles (Harmsworth et aI .. 1990). Kaolinite-rich shale. an aluminous residue


from the breakdown of fine volcanic ash layers in the BIF host, forms discrete bands intercalated within the ore.

In contrast, primary BIF carbonate is largely replaced by goethite and is therefore much lower in alumina content (Harmsworth et al., 1990). Gangue also occurs as either coarse-grained (+100 /lm) kaolinite, traces of gibbsite, or minor quartz (Clout, 2(02). There may also be appreciable fine-grained «5/lm) kaolinite gangue derived from fine ash layers interbedded with the ore minerals, as well as alumina, silicon, or phosphorus interpreted to be either substituted into the goethite crystal structure (Morris, 1985) or present as submicron inclusions of as yet unidentified phases. Alumina and silica substitution in vitreous goethite and especially ochreous goethite is more common in the near-surface hydration wne, especially surface hardcap, and is uneconomic to upgrade (Clout, 2(02). The geolOgiC controls on hardcap and its distribution are not well understood, and interpretation of shallow drill hole assay data is further complicated where shales may be present with ore (Clout, 2(02).

Low-grade ores that require beneficiation vary from friable quartz-rich supergene itabirite from the Quadrilatero Ferrifero in Brazil to ores with kaolinite- or gibbsite-rich shale bands in Australia and India, respectively. Some martitegoethite mineralization is uneconomic to upgrade due to the presence of Significant alumina and silica locked within goethite, or the presence of very fine « 5/lm) intergrowths of clay or quartz within goethite or hematite (Clout, 2002; Silva et al., 2(02). Despite high (>50 wt %) Fe grades, some of these ores cannot be upgraded even with grinding as fine as that used for taconites. In contrast, the well-metamorphosed itabirite-derived supergene hematite deposits in the Quadrilatero F errifero have good beneficiation characteristics because they are soft, contain very dense, liberated hematite particles with low porosity and liberated quartz particles (Silva et al., 2002; Spier et al. , 2(03).

For fme ores (nominally <6.3 or 8.0 mm), jigs or dense medium cyclones (DMC) using ferrosilicon suspensions are used to treat the coarser (>1 mm) size fractions; spirals have widespread application for the intermediate to finer fractions (between 0.075-1 mm); wet high-intensity magnetite separation treat the intermediate to fine fractions (0.03-1 mm); hydrocyclones are most commonly used to remove very fine «0.02 mm) clay or quartz-rich ultrafines (Box et al., 1996; Clout et al. , 1997; Bensley et al., 1999; Mason and McSpadden, 2002; Miller, 2(02). Jigs, dense medium cyclones, and spirals are wet processes that separate on the basis of specific gravity and remove quartz or shale from denser hematite and goethite. Low-grade itabirite from the Quadrilatero Ferrifero contains incompletely leached friable chert with residual hematite and requires extensive beneficiation to make sinter fines , blast furnace pellet feed, or direct reduced iron feedstock grades (de Araujo and Peres, 1995; Silva et al., 2(02). Spirals, jigs, reverse flotation, and wet high-intensity magnetite separation are commonly used here as well to remove well-liberated quartz, although some ore types do not respond well to certain concentration methods (Silva et al. , 2002). For example, in ore types containing quartz and gibbSite gangue, gibbSite is not separated from hematite using reverse flotation.

For lump ores (6.3/8.0--31.5 mm) in Australia and South Africa, cheap gravity techniques including jigs are commonly

used for porous ores, whereas more expensive wet densemedium separators using ferrosilicon suspensions in a rotary drum work well with low-porosity feed (Warnock and Bensley, 1996; Clout et al., 1997; Mason and McSpadden, 2(02). These types of processes are ideally suited where thick (>0.05-2 m) mesobands of high specific gravity hematite are interbedded with low specific gravity shales (e.g., Dales Gorge Member, Mount Tom Price mine, Asbestos Hills Subgroup, Sishen deposit; Bitencourt et al. , 2002; Carney and Mienie, 2(02). The distribution of shale (formerly fine ash) and hematite in BIF is directly related back to the primary iron formation and thus genetic controls on dominant ironoxide depoSition versus carbonate and/or silicate. The degree of upgrading is dependent upon the presence of coarse (>10 mm) liberated gangue of lower specific gravity with dense liberated hematite (limited supergene leaching), whereas separation efficiency and recovery of expensive ferrosilicon media may be adversely affected by high porosity (caused byextensive supergene leaching; Clout et al. , 1997; Bitencourt et al., 2(02). Separation efficiency is poor where supergene leaching has resulted in highly porous hematite and goethite that has a similar or lower specific gravity than either the BIF or subgrade goethite. Alternatively, porous hematite results from incomplete replacement of gangue by hematite during ore formation prior to supergene leaching. Media recovery may be low because the fine heavy media becomes trapped in pores and so is lost to the process. This means that many porous ores are unsuitable for processing by heavy media (Clout et al., 1997).

Iron ore products and their uses

The principal use of iron ores is for the production of steel from either a conventional blast furnace pig-iron route or more directly from an electric arc furnace. Although there are other routes for making steel from iron ores, blast furnaces still account for >80 percent of world crude steel production (Astier, 2003).

Beneficiated and high-grade lump iron ore, typically >62 wt percent Fe and between 6.3 and 31.5 mm in size, can be directly added to a blast furnace . In contrast, high (>68 wt % Fe) grade iron ore concentrates, pellets, or lump are reqUired to undergo heating and direct reduction steps to convert hematite to metallic iron before adding to an electric arc furnace to make steel. MineralOgical and metallurgical studies by Clout (2002), together with collaborative industry studies (e.g., Box et al. , 2(02), have demonstrated that ore mineralogy, texture, hardness, porosity, and petrolOgical characteristics directly control lump physical and metallurgical quality in the blast furnace.

High-grade and beneficiated martite-goethite and microplaty hematite lump iron ores can be fed directly to the blast furnace, whereas fine « 0.10 mm) concentrates or fine ores must Hrst be agglomerated into pellets or sinter, respectively, before they can be fed to the blast furnace. This is because Significant amounts of fine particles would simply block the vital upflow of gases and/or be ejected from the top of the blast furnace as dust. Iron ore sinter is produced by mixing of Hne ores or concentrates, fluxes (limestone, burnt or hydrated lime, dolomite, or serpentine), and fuel (coke, anthracite), then granulated with water in a rotating drum and building a


layer 550 to 800 mm deep onto a horiwntal sinter machine grate. The fuel is then ignited at the top of the layer using burners and air is drawn down through the bed under suction, with solids finer than 1 mm melting at _1,300°C and gluing together the coarse> I-mm hematite. The sinter is subsequently cooled to form a porous hard solid, composed of calcium ferrites and un melted hematite ore particles, which physically resembles volcanic scoria and is crushed to 5 to 40 mm in size and fed directly to the blast furnace.

Magnetite concentrates produced from beneficiation of iron formation ore are typically used to make blast furnace pellets (e.g., T~den, United States). Magnetite concentrates are ideally suited to making pellets that require fme « O.045-mm) particle size to agglomerate with water, binders, and fine fluxes in a r0-

tating drum or disk to form nOminally 10- to 15-mm-diameter spherical green balls. The green balls are then heated up to about 1,300°C in a grate or kiln, with the exothermic oxidation of ore magnetite to hematite during induration of pellets generating useable heat to help drive the process. The hematite

pellet product is much more easily reduced to metallic iron in the blast furnace than magnetite. However, although not ideal, Significant (>100 MtIyr) quantities of domestic magnetite concentrates produced in China are also used in sintering. In Canada (e.g., Mont-Wright and Humphrey mines) and elsewhere, hematite concentrates produced from beneficiation of BIF andlor CIF ores are also used to make blast furnace and direct reduced iron-grade pellets.

Although the majority of martite-goethite and high-grade hematite fine iron ores (e.g., Pilbara, Australia; Coa, India; Carajas and Quadrihltero Ferrifero, Brazil) are used for sintering, smaller quantities are also used for blast furnace or direct reduced iron-grade pellets. Some high-grade beneficiated lump is also used in direct reduced iron processes.

Concentrate, lump, and fine ore gangue mineralogy, as well as minor and trace elements. can have an adverse effect on their acceptability for making pellets or sinter and in the blast furnace (Table 4). Elevated concentrations of alkali elements reduce blast furnace refractory life, whereas elevated

TABLE 4. Effect of Deleterious Gangue and Minor andlor Trace Elements on Downstream Process Performance (modified from Clout, 1998)

DeleteriOUs phase/element

Sideritic carbonates

ClaY' (>5%)

Alkalis (e.g., K,o >0.09%, Na)

P(>O.08%)

Base metals (e.g .. Zn. Ph >100 ppm) and heavy metals

S'o.(>5%)

Mn (>0.9%)

Cu (> 100 ppm)

CI (>5OOppm)

S (>0.08%)

BF • blast furnace

GeologiC control

Carbonate SIF or hydrothermal carbonate alteration. extent of replacement andlor leaching

Shale bands and AI suicate content of BlF host

KtQ-BIF micas and rulpnomelane Na hypersaline ground water

Apatite in 8IF. extent of supergene leaching of P

BIF host. hydrothennal source

Shale bands and Al silicate content of BIF host

BIF and shale, extent of supergene leaching

Remobilization of Mn from impure dolomite above or below mineralization. carbonate SIF

BIF host. hydrothennal source

Hypersaline ground water

SIF host. organiC S in hardcap or near-surface mamte-goethite

Shale bands. ilmenite from crosscutting intrusion

Process stage

Sintering and pelletizing


Sintering, pelletizing. BF

Meta] production

BF. sintering

Sintering. SF

Sintering. SF

Steelmaking

Sintering. steelmaking production

Sintering. SF

Sintering, SF


Effects

Lowen strength of sinter and pellets due to increased porosity once carbonates are calcined (CO. driven off above I,OOO'C)

Lowers strength of sinter and pellets due to increased. meTt viscosity due to Alt03

Lowen melt temperature, corrosion of SF refractory bricks

Removal cost

Removal oosts, especially reprocessing of base and heavy metal-rich dusts

Higher levels progressively increase melt viscosity hence increase fuel rates

Increases the amount of slag; increased use of limestone since the ratio of CaOlSiOi must be fIXed

Although some types of steel require Mn, excess levels require dilution with low Mn-bearing ores to maintain steel properties

Catalyses dioxin fonnation during sintering. must be diluted with low Cu ores to maintain steel quality

Increased dioxin (toxic) and N01 emissions, reduces efficiency of electrostatic dust precipitators and increases SF refractory wear

Increased SO, emissions and higher MgO levels required to partition 5 into SF slag

Lower physical strength of pellets and sinter

674 CWUT AND SIMONSON

phosphorus increases steelmaking costs. TIght restrictions are generally placed on maximum concentrations of these and other elements in sinter fines, pellet feed, and hlast furnace lump. Although blending with other ores can lower the concentration of certain deleterious elements, beneficiation is the preferred method to reduce minor and trace elements to a level acceptable to the market.

Discussion

Exploration significance of the hypogene-hydrothemwl models

Models of supergene-modified hypogene mineralization provide the iron are explorationist with new criteria to locate high-grade iron ores. Favorable situations to target now include: (1) former conduits of fluid flow, e.g., where major or secondary splay faults cut across iron formations (Taylor et a1., 2(01), (2) fold hinges or other dilatant sites in iron formations (Roserie and Rios, 2004), (3) iron formations above carbonate aquifers (Taylor et a1., 2(01), (4) areas of carbonate alteration of iron formations (Hagemann et a1., 1999, Taylor et a1., 2002, Dalstra and Guedes, 2004), and (5) geochemical vectors applied to distal carbonate alteration that help to locate highgrade are associated with proximal ·a1teration (Hagemann et a1., 1999, Thome et a1., 2004). The intersection of structures, especially faults , with iron formations provides specific areas of interest, whereas carbonate alteration provides vectors toward possible proximal high-grade mineralization.

Directions for future work

With the increasingly short supply for iron are, especially for China, there is a need to improve our understanding of the formation of iron ores to achieve better exploration targeting, especially iron formation-hosted high-grade iron ores. Five areas for future research related to the genesis of highgrade are include studies to (1) improve our understanding of the nature of the are fluids and their timing with respect to deformation (e.g., Rosiere and Rios, 2004, Thome et a1., 2004) across a larger number of deposits, (2) better understand the role of regional and local structures in are formation, (3) differentiate the geochemical and O-H isotope signatures of proximal versus distal mineralization and alteration, (4) recognize evidence for the presence or absence of substantial carbonate protore within proximal high-grade are, and (5) elucidate the are formation process(es) where carbonate protore is absent.

Recent work has highlighted the importance of major fault structures and deformational history in the origin of the highgrade hematite deposits, although their exact role in hypogene are genesis has received less detailed attention. However, Rosiere and Rios (2004) have set a new benchmark for future studies by successfully plaCing fluid inclusion and alteration work into the complex deformational history of metamorphosed hard high-grade hematite ores in the Quadrilatero Fenifero. Further studies are also required on the role of structures in are formation in the Pilbara and elsewhere. For example, in the Pilbara, the Mount Tom Price and Mount Whaleback are bodies are localized around the Southern Batter and Mount Whaleback faults , respectively. Although these faults have a normal sense of movement, they

have been interpreted variously as extensional normal faults (Harmsworth et aI 1990), extensional faults, or reactivated thrusts (Dalstra et a1., 2002, Taylor et a1 ., 2002, McLellan et a1. , 2004) in a foreland fold-thrust belt or a compresSional setting in a foreland basin (Powell et a1., 1999). Alternatively, these normal faults could represent back thrusts in the hanging wall of major regional thrusts, with are fluids being preferentially focused through the back thrusts since they are most likely to form more dilatant zones. The fluid fOCUSing mechanism and the role of structures in the alteration of very large volumes of iron formation need to be investigated further, with an eye to quantifying the pathways and volumes of the fluids responsible for (orming specific types of iron are deposits.

For the supergene ores, including the soft hematite ores of Brazil and the martite-goethite ores in Australia, new research directions are required to extend the models developed by Morris (1980, 1985, 2002b). The most promising approaches will probably involve fluid inclusion studies and detailed modeling to better understand the key influences that drove fluid flow and factors that resulted in economic versus subgrade mineralization.

Future exploration for iron formation ores needs to carefully consider the economic benefits that higher grades of metamorphism present to their improved economic extraction. At the regional exploration phase, areas need to be ranked on the basis of metamorphic grade, magnetite andlor hematite grain size, and amount of gangue inclusions they contain. Locally, sections of the iron formation stratigraphy where magnetite andlor hematite meso- and micro bands are more abundant are likely to be higher in iron content and so be more attractive as exploration targets. Therefore, there needs to be a greater research emphasiS on basin-wide analysis of the depoSitional features of iron formations to better predict the occurrences of more iron oxide-rich iron formation with coarse magnetite andlor hematite grain size.

Conclusions

Large iron are depoSits are all associated with stratigraphic occurrences of a chemical sedimentary rock known as iron formation, almost all of which were depoSited before 1.8 Ga. Iron minerals in unenriched iron formations vary widely in composition and abundance, reflecting nonrandom sedimentary variations through geolOgiC time. Some magnetite-rich iron formations are mineable (taconite-type are deposits ), but high-grade orebodies only occur where subsequent events have upgraded iron formations , typically from 30 wt percent Fe in an iron formation to 60 to 68 wt percent Fe in high-grade ore. The two main types of high-grade iron are depoSits, martite-goethite and high-grade hematite deposits, both exhibit a diverse range in deposit characteristics and genesis, but the high-grade hematite deposits show the greatest diversity, including differences in ore textures, the presence or absence of carbonate alteration, timing of hypogene mineralization with respect to regional metamorphism, hypogene fluid mineralization temperatures, and amounts of basinal brines versus meteoric waters or magmatic fluids. In the final analysiS, high-grade hematite deposits will only form where an iron formation with the right textures and iron-rich composition experiences the right sort of fluid flow,


e.g .. through a major fault. usually in conjunction with some sort of regional deformation. In contrast. the genesis of supergene martite-goethite and hematite deposits is more consistent and less controversial; they are believed to involve replacement of gangue by goethite versus silicate and/or carbonate leaching and residual concentration of hematite. respectively.

Acknowledgments

The authors thank J. Hedenquist and J. Thompson (editors) and r.articularly several reviewers. including D. Kepert. S. Hass er, and S. Hagemann, for their constructive comments that have greatly improved this paper. and C. Roserie for educating JC on current details of Brazilian iron ore geology. Fieldwork on iron formations and associated units by BMS was funded by grants from the National Geographic Society. National Science Foundation. and Oberlin College.

REFERENCES Aitennann, W .• 1996, Sedimentology, geochemistry and palaeogeographiC

implications of volcanic rocks in the Upper Archaean Campbell Croup. western Kaapvaa1 craton, South Africa: Precambrian Research, v. 79, p. 73-100.

Anbar, A.D ., and Knoll, A.H., 2002, ProteTO'.t.Oic ocean chemistry and evolution: A bioinorganic bridge?: Science, v. 297, p. U37-1142.

Arndt, N., 1999, Why was nood volcanism on submerged continental platfanns so common in the Precambrian?: Precambrian Research, v. 97, p. 155-164.

Arnold, C.L., Anbar, A.D., Barling. J., and Lyons, T.W. , 2004, Molybdenum isotope evidence for widespread anoxia in mid-Proterozoic oceans: Science, v. 304, p. 87-90.

Astier, J.E., 2003, Main developments in iron ore preparation: The Iron Ore Industry Today and Tomorrow (TIOITT-2002), Kiruna, Lulea University, Proceedings. p. 1-9.

Barley, M.E ., Pickard, A.L.. and Sylvester, P.J., 1997, Emplacement of a large igneous province as a possible cause of banded iron fonnation 2.45 billion years ago: Nature, v. 385, p. 55-58.

Barley, M.E., Pickan:l, A.L., Hagemann, S.C., and Folkert, S.L. , 1999, Hydrothennal origin for the 2 billion year old Mount Tom Price giant iron ore deposit, Hamersley province, Western Australia: Mineralium Deposita, v. 34. p. 784-789.

Barrett, T.J., and Fralick, P.W. , I985, Sediment redepoSition in Archean iron fonnation: Examples from the Beardmore-Geraldton greenstone belt, Ontario: Joumal of Sedimentary Petrology, v. 55, p. 205--212.

--1989, Turbidites and iron formations , Beardmore-Geraldton, Ontario: Application of a combined ramplfan model to Archaean clastic and chemical sedimentation: Sedimentology, v. 36, p. 2.21-234.

Beard, B.1..., and Johnson, C.M., 2004, Fe isotope variations in the modem and ancient Earth and other planetary bodies: Reviews in Mineralogy and Geochemistry, v. 55, p. 319-357.

Beard, B.L., Johnson, C.M., Cox, L. , Sun, H. , Nealson, K.H., and Aguilar, C., 1999, Iron isotope hiosignatures: Science, v. 285, p. 1889-1892.

Bekker, A., HoUand, RD., Wang, P.-L. , Rumble Ill , D ., Stein, H.J., Hannah. J.1..., Coetzee, 1...1..., and Beukes, N.J., 2004, Dating the rise of atmospheriC oxygen: Nature, v. 427, p. 1l7-120.

Bensley, C.N., Fe nton, K.1... and Turner, J., 1999, Iron ore beneficiation by BHP Iron Ore in Westem Australia: Australasian Institute of Mining and Metallmgy Monogrnph J9. p. 276-280.

Bergstrand. R.e., Waters, A.C ., and Clout, J.M.F. , 2003, The mineralogy of West Angelas iron ore--the key to unlocking its potential [abs.] : International Conference on Science and Technology of lronmaking, Dusseldorf, 2003. 3 .... Abstracts. p. 6.

Benkes, N.J., 1980, Suggestions towards a classification of and nomenclature for iron-fonnation: Transactions of the Geological Society of South Africa. v. 83. p. 285-290.

--1983, Palaeoenvironmental setting of iron-fonnations in the depositional basin of the Transvaal Supergroup, South Africa, in Trendall, A.F., and Morris, R.C., eds., Iron-fonnation: Facts and problems: Amsterdam, Elsevier. p. 131-209.

--1984. Sedimentology of the Kuruman and Griquatown lron-Fonnabons. Transvaal Supergroup, Griqualand West, South Africa: Precambrian Research , v. 24, p . 47~.

--1986, The TranSV'clai sequence in Griqualand West, in Anhaeusser, C. R., and Maske. S., eds., Mineral deposits of Southern Africa: Johannesburg, Geological Society of South Africa, v. I, p. 819-828.

Beukes, N.J., and Klein, C., 1990. Geochemistry and sedimentology of a facies Imnsition-from microbanded 10 gmllular iron-fonnation-in the early Protero7.oic Tmnsvaal Supergroup. South AfriC'd: Precambrian Research , v. 47, p. 99-139.

--1992. Models for iron-fonnation deposition , in Schopf, j .W., and Klein, C .. eds., nle Proterozoic biosphere: A multidiSCiplinary study: Cambridge University Press, p. 147-151.

Be ukes, N.J., Gutzmer, j ., and Mukhopadhyay, J., 2002, The geology and geneSis of high-grade he matite iron ore deposi ts: Austnuasian Institute of Mining and Metallurgy, Publication Series 712002, p. 23--29.

Bitencourt, R., Mackenzie. p' , Gordon, J. and Morey, 8., 2002, High-grade optimisation and improved grade control practices in Mount Tom Price: Australasian Institute of Mining and Metallurgy Publication Series 712002, p .261-277.

Blake, RS., and Barley. M.E .• 1992, Tectonic evolution of the late Archaean to early Proterozoic Mount Bruce megasequence set, Western Australia.: Tectonics, v. 11, p. 141~1425.

Bofei. J., 2005. Outputs of China major iron ore mines in January-December 2004: China MetallUrgical Newsletter, MetallUrgical Counci1 of China Council for the Promotion of International Trade, v. 7, 20 p.

Box. J.c., Morgan , A.E ., and 4'nch, M.L., 1996, Developments in iron ore processing: Mining Technology Conference, 1996, Cooperative Research Centre for Mining Technology and Equipment (CMTE), Fremant1e, Western Australia. Proceedings, p. 229-237.

Box. J.. Phillips. J. and Clout J.M.F .• 2002. Use of geological material types for predicting iron ore product characteristics: Japan-Australia Symposium on Iron and Steelmaking, Fuwa-Ward Symposium, I"', Kyoto University, Iron and Steel Institute of Japan, April 4-5, 2002, Proceedings, p. 36-41.

Brandt, R.T., 1966, The genesis of the Mount Goldsworthy iron ore deposits of Northwest Australia: ECONOMIC CEOLOGY, v. 61, p. 999-1009.

Breltkopf, j .H ., 1988, Iron formations related to mafic volcanism and eruialic rifting in the Southern Margin zone of the Damara orogen, Namibia: Precambrian Research. v. 38. p. Ill-I30.

Brown, M.C., Oliver, N.H.S., and Dickens, C., 2004, Veins and hydrothermal fluid nOYI in the Mt. Whaleback iron ore district, eastern Hamersley province, Western Australia: Precambrian Research, v. 128, p. 441-474.

Button, A., 1976, Transvaal and Hamersley basins-review of basin development and mineral deposits: Minerals Science and Engineering. v. 8, p. 262-293.

Cameron, E.M., 1983, Genesis of Proterozoic iron-fonnation: Sulphur is0-tope evidence: Ceochimica et Cosmochimica Acta, v. 47, p. 1069-1074.

CanAeld, D.E.. 1998, A new model for ProteroZOic ocean chemistry: Nature, v. 396. p. 450-453.

--2005, The early history of atmospheriC oxygen: Annual Review of Earth and Planetary Sciences, v. 33. p. 1--36.

Camey, M.D., and Miene, P.J., 2002, A geological comparison of the Sishen and Sishen South (Welgevonden) iron ore deposits, Northern Cape province, South Africa: Australasian Institute of Mining and Metallurgy Publication Series 712002, p. 145-151.

Canigan, W.J., and Cameron, E.M., 1991, Petrological and stable isotope studies of carlxmate and sulfide minerals from the Gunflint Formation, Ontario: Evidence for the origin of early Proterozoic iron-fonnation: Precambrian Research, v. 52, p. 347-380.

Chemale, F., Jr., Rosi~re, C.A., and Enda, I. , 1994, The tectonic evolution of the Quadrillitero Ferrifero, Minas Gerais, Bra7jl : Precambrian Research, v. 65. p. 25-54.

Cheney, E.S., 1996, Sequence stratigrdphy and plate tectonic significance of the Transvaal succession of southern Africa and its equivalent in Western Australia: Precambrian Research, v. 79, p. 3-24.

Cloud, P.E., 1968, Atmospheric and hydrospheriC evolution on the primitive Earth: Science, v. 160, p. 729-736.

Clout, J.M.F., 1998, The effects of ore petrology on downstream processing perfonnance: A review, in Mine to Mill 1998: Australasian Institute of Mining and Metallurgy, Brisbane, Proceedings, p. 43-50.

--2002, Upgrading processes in BIF-derived iron ore deposits-implica.tions for ore genesis and downstream processing: Australasian Institute of Mining and Metallurgy Publication Series 712002, p. 237-242.


Clout, J.M.F., Firth, B.A., and O'Connor, M.T., 1997. New advances in beneficiation of iron ore, in Misra, V.N., and Dunlop, J.s., eds .• Towards 2000-National Conference on Ironmaking Resources and Reserves Estimation: AustraJasian Institute of Mining and Meta11urgy Spectrum Series, Proceed· ings, p. 177-183.

Cook. H.E., and Egbert, R.M ., 1983, Diagenesis of deep-sea carbonates, in Larsen, C., and Chilingar. C.v.. eds., Diagenesis in sediments and sedimentary rocks. 2: Amsterdam, Elsevier, p. 213-288.

Coyle, D.S. , 1965, The Empire Concentrator-iron ore's first two stage autogenous grinding Oowsheet: AnnuaJ Mining SymJX>Sium, 26th , Duluth , Minnesota, Proceedings. p. 9--13.

Dalstra, H.J., and Guedes, S., 2004, Giant hydrothennaJ hematite deposits with Mg-Fe metasomatism: A comparison of the Carajas, Hamersley. and other iron ores: EcoNOMIC GEOLOGY, v. 99, p. 1793--1800.

Dalstra. H.J.. Harding. A.E., Riggs. T., and Taylor, D., 2002. Banded iron formation hosted high-grade hematite deposits, a coherent group: Australasian Institute of Mining and Metallurgy Publication Series 712002, p. 57-61.

de Araujo, A.C., and Peres, A.E.C., 1995, Froth flotation: Relevance and the Brazilian case, In Uns. F.F., ed., Tecnologia Mineral 70: Rio de Janerio, MCf, CNPq, CETEM, p. 22-28.

Deer, W.A., Howie, R.A., and Zussman, J., 1992, An introduction to the rockfonning minerals, 2nd ed.: l..Lmdon, Longman Scie ntific and Technical, 696 p.

DeVaney, F.D., 1985, Iron ore, In Weiss, N.L , ed., Society of Mining Engineers Mine ral Processing Handbook, v. 2, p. 2~25.

Dimroth, E ., 1976, Aspects of the sedimentary petrology of cherty iron formation, in Wolf, K.H ., ed., Handbook of strata-bound and stratifonn ore deposits, v. 7: Amsterdam, Elsevier, p. 203-254.

--1986, Depositional environments and tectonic settings of the cherty iron fonnations of the Canadian Shield: Journal of the Geologica1 Society of Inilia, v. 28, p. 23~250.

Dimroth, E., and Chauvel, J.-J., 1973, Petrography of the Sokoman Iron Formation in part of the centraJ Labrador trough: Geological Society of America Bulletin, v. 84, p. 111-134.

Dorr, J. van N .. H, 1964, Supergene iron ores of Minas Gerais, Brazil: EcoNOMIC GEOUX:Y, v. 59, p. 1203-1240.

--1965, Nature and origin of the high-grade hematite ores of Minas Gerais, Brazil: ECONOMIC CEOLOGY, v. 50, p. 1-46.

--1969, Physiographic, stratigraphic and stnlCturaJ development of the Quadri.J4tero Fenifero, Minas Gerais: U.S. Geological Survey Professional Paper 641-A, 110 p.

Dunbar, C.J., and McCall , C.J.H .• 1971. Archaean turbidites and banded ironstones of the Mt. Belches area (Western Australia): Sedimentary Ceology, v. 5, p. 93-133.

Eriksson, K.A., 1983, Siliciclastic-hosted iron-fonnations in the early Archean Barberton and Pilbara sequences: Joum31 of the Ceological Society of Australia. v. 30, p. 473-482.

Eriksson , K.A. , and Don31dson, J.A., 1986, Basinal and shelf sedimentation in relation to the Archaean-Proterozoic boundary: Precambrian Research. v. 33, p. 103-12 1.

Eriksson, P.C., 1999, Sea level changes and the continental freeboard concept: General principles and application to Precambrian: Precambrian Research, v. 97, p. 14~154.

Ewers, W.E. , 1983, Chemical factors in the deposition and diagenesis of banded iron-formation , in Trend31l, A.F. , and Morris, R.C., eds., Iron-formation: Facts and problems: Amsterdam. ElSevier, p. 491-512.

Ewers. W.E., and Morris, R. C .. 1981, Studies of the Dales Corge Member of the Brockman Iron Fonnation, Weste rn Australia: ECONOM IC GEOLOCl'. v. 76, p. 1929-1953.

Farquhar, J., Bao, H., and Thiemens, M., 2000, Atmospheric influence of Earth's earliest sulfur cycle: Science, v. 289. p. 756-758.

Fedo, C.M., and Eriksson, K.A., 1996, Stratigraphic framework of the -3.0 Ca Buhwa greenstone belt: A unique stable-shelf succession in the Zimbabwe Archean c raton: Precambrian Research, v. 77, p. 161-178.

Fedo, C.M., and Whitehouse, M.J., 2002. Metasomatic origin of quartz-pyroxene rock, AkiJia. Greenland. and implications for Earth's earliest life: Science. v. 296, p . 1448--1452.

Force. E.R., and Maynard. J.B., 1991, Manganese: Syngenetic deposits on the margins of anoxic basins: Reviews in Economic Geology, v. 5, p. 147-157.

Fralick. P.W .. 1988, Microbial biohe nns, lower Proterozoic Cunnint Fonnation, Thunder Bay. Ontario: Canadian Society of Petroleum Geologists Memoir 13. p. 24-29.

Fralick, P.W .. and Barrett, T.J .. 1995. DepoSitional controls on iron formation associations in Canada: International Association of Sedimentologist Special Publication 22, p. 137-156.

Fralick, P.w. , Davis, D .W, and Kissin, S.A., 2002, The age of the Gunnint Fonnation. Ont"cUio, Canada: Single zircon U-Pb age detenninations from reworked volc-anic ash: Canadian Journal of Earth Sciences, v. 39, p. 1085-1091.

Fryer, B., Fyfe. W.S., and Kerrich , R., 1979, Archaean volcanogeniC oceans: Chemical Geology, v. 24, p. 25-33.

Gibbs, A.K., Wirth, K.R. . Hirata, W.K .• and Olszewski, w.J ., 1986, Age and composition of the Coo Pant Croup volcanics, Serra dos Carnjas, Brazil: Revista Brasileira de Ceociencias, v. 16. p. 201-211.

Cole, M.J., and Kle in . C., 1981, Banded iron-fonnations through much of Precambrian time: Journal of Geology, v. 89. p. 169-183.

Goode, A.D.T .. Hall, W.D.M. , and Bunting, J.A .. I983, The Nabberu basin of Western Australia, in Trendall , A.F., and Morris, R.C., eds., lron-formation: Facts and problems: Amsterdam, Elsevier, p. 295-323.

Goodwin, A. M., 1991, Precambrian geology---the dynamic evolution of the continental crust: London, Academic Press, 666 p.

Goodwin, A.M .. Thode, H.C ., Chou, C.-L.. and Karkhansis, S.N., 1985, Chemostratigraphy and origin of the late Archean siderite-pyrite--rich Helen Iron Fonnation, Michipicoten belt, Canada: Canadian Journal of Earth Sciences, v. 22. p. 72--84.

Graber, R. G., and Sundberg, C. R., 2002, Evolution of iron ore characterisation techniques-a North American perspective: AustraJasian Institute of Mining and Metallurgy Publicntion Series 712002, p. 195-199.

Gross, G.A., 1961, Metamorphism of iron-formations and its bearing on their beneficiation: Canadian Mining and Metallurgical Bulletin, v. 64, p. 24-31.

--1965, Geology of iron deposits of Canada, v. I , General geology and evaluation of iron deposits: GeolOgical Survey of Canada Economic Geology Report 22, l SI p.

--1972, Primary features in cherty iron fonnations: Sedimentary Geology. v. 2, p. 241-26\.

--1983, Tectonic systems and the deposition of iron-fonnation: Precambrian Research, v. 20, p. 171-187.

Gross. G.A., and Zajac. I.S., 1983. lron-fonnation in fold belts marginal to the Ungava cratoJl , I,. Trendall, A.F., and Morris, R.C .• Iron-fo rmation: Facts and problems: Amsterdam, Elsevier, p. 253-294.

Crotzinger, J.P., 1994, Trends in Precambrian carbonate sediments and their implications for understanding evolution, in Bengston. S., ed., Early life on Earth: Nobel SympoSium 84: New York. NY. Colu mbia University Press. p. 245-258.

Croves, D.l .• Condie, K.C., Coldfarb. R.J ., Hronsk)', I.M.A., and Vie lreicher, R.M .. 2005, Secular changes in global tectonic processes and their influence on the temporal distribution of gold-bearing minerals deposits: EcoNOMIC GEOUX:Y. V. 100, p. ~224.

Guedes, S.C., Rosi~re . C.A., Barley, M. , and Lobato. L. 2002. Carbonate 31-teration associated with Carajas high-grade hematite deposits. Brazil: Australasian Institute of Mining and Metallurgy Publication Series 712002, p. 63-66.

Cuild, P.W, 1953, Iron deposits of the Congonhas district , Minas Gerais. Brazil : EcoNOM IC GEOLOGY v. 48, p. 639-676.

- -1957, Geology and mineral resources of the Congonhas District Minas Gerais, Brazil: U.S. GeolOgical Survey Professional Paper 290, 90 p.

Gutzmer, J., and Beukes, N.J, 1998, Earliest laterites and possible evidence for terrestrial vegetation in the Early Proterozoic: Geology, v. 26. p. 263-266.

Gutzmer, J., Netshiozwi, S., and Beukes , N.J., 2002, Hydrothennal origin of high-grade iron orebodies hosted by the Paleoproterozoic Penge Iron Formation, Thabazimbi. South Africa [abs.J: InternationaJ Sed.imentological Congress, 16th, Auckland Park. South Africa, Abstract Volume 2002, CDROM.

Hagemann. S.C., Barley, M.E., Folkert, S.L., Yardley. 8.W .• and Banks, D.A., 1999, A hydrothennaJ origln fo r the giant Tom Price iron ore depoSit, ill Stanley, C.J., et aI., eels . Mineral deposits: Processes to processing: Rotterdam, A.A. Balkema, p. 41-44.

Hagemann, S.C., Rosiere. C.A., Lobato, L.. Baars, F., Zucchetti, M .• Figueiredo e Silva, R.C, in press, Controversy in genetic mooels for highgrade BIF related Fe deposits: Unifying or discrete model(s)?: Australasian Institute of Mining and Metallurgy. I ron Ore 2005 Conference, Perth, Australia. Proceedings, 4 p.

Hall , W.D.M .• and Goode. A.D.T., 1978, The Early Proterozoic Nabberu basin and associated iron fonnations of Weste m Australia: Precambrian Research. v. 7, p. 129-184.


Hamade, T. . Konhauser, K.O., Raiswell, R.. Goldsmith . S. , and Morris, R.c. t

2003, Using Ge/Si to decouple iron and silica fluxes in Precambrian banded iron fonnations: Geology, v. 31, p. 35-38.

Han, T.·M., 1978, Microstructures of magnetite as guides to its origin in some Precambrian iron-fonnations: Fortschritte cler Mineralogie, v. 56, p. 105-142.

--1988, Origin of magnetite in Prec-dmbrian iron-fonnations of low metamorphic grade: Proceedings of the Seventh QuadrenniallAGOD Symposium: Stuttgart, E. Schweizerbart'sche Verlagsbuchhand1ung (Nagele u. Obennille<), p. 641-656.

Han. T-M., and Runnegar. B., 199-2. Megascopic eukruyotic algae from the 2.1- billion-year-old Negaunee lron-Fonnation, Michigan : Science, v. 257. p. 232-235.

Hannsworth. R.A. , Kneeshaw, M., Morris, R.C., Robinson. c.). and Shrivastava, P.K., 1990, BIF -derived iron ores of the Hamersley province: Australasian Instihlte of Mining and Metallurgy Monograph 14, v. 1, p. 617-642.

Hassler, S.W , 1993, Depositional history of the Main tuff interval of the Wittenoom Formation, Late Archean-Early Prote rozoic Hamersley Group. Western Australia: Precambrian Research, v. 50, p. 337-359.

Hassler, S.W, and Simonson, B.M., 1989, Deposition and alteration of volcaniclastic strata in two large. early Proterozoic iron-formations in Canada: Canadian Journal of Earth Sciences, v. 26, p. 1574-1585.

--2001, The sedimentary record of extraterrestrial impacts in deep-shelf environments: Evidence from the early Precambrian: Journal of Geology, v. 109, p. 1-19.

Hoffman, P.F .• 1987, Early Proterozoic foredeeps , foredeep magmatism, and Superior-type iron-formations of the Canadian Shield: American Geophysical Union Geodynamics Series, v. 17, p. &5-98.

Hoffman. P.F., Kaufman. A.J ., Halverson, G .P' , and Schrag. D.P.. 1998, A Neoproterozoic snowball Earth: Science. v. 281 . p. 1342-1346.

Huston. D.L.. and Logan, G.A .• 2004. Barite, BIF and bugs: Evidence for the evolution of the Earth·s early hydrosphere: Earth and Planetary Science Letters. v. 220, p. 41-55.

Isley. A.E. , 1995, Hydrothermal plumes and the delivery of iron to banded iron fonnation: Journal of Geology, v. 103, p. 169-185.

Isley, A.E., and Abbott, D .H ., 1999, Plume-related mafic volcanism and the deposition of banded iron formation: Journal of Geophysical Research, v. 104, p. 15,461-15,477.

Jacobsen, S.B .• and Pimentel·K1ose, M.R., 1988, A Nd isotopic study of the Hame rs1ey and Michipiooten banded iron formations: The source of REE and Fe in Archean oceans: Earth and Planetary Science Letters, v. 87, p. 856-867.

James, H.L. , 1954, Sedimentary facies of iron-formation: ECONOMIC GEOLOCY, v. 49, p. 235-293.

--1966, Chemistry of the iron-rich sedimentary rocks: U.S. Geological Survey Professional Paper 44O-W, 61 p.

James, H.L., and Trendall, A.F., 1982, Banded iron formation: Distribution in time and paleoenvironmental significance, ill Holland, H.D., and Schidlowski , M., cds., Mineral deposits and the evolution of the biosphere: Berlin. Springer-Verlag. p. 199--217.

Johnson, C.M., Bean!, B.L., Beukes, N.J., Klein, C, and O'Leary, J.M., 2003, Ancient geochemical cycling in the Earth as inferred from Fe isotope studies of banded iron formations from the Transvaal craton: Contributions to Mine ralogy and Petrology, v. 144, p. 52J..-.54 7.

Johnson, GM .• Beard.. B.L., Roden, E.E., Newman, O.K., and Nealson , K.H., 2004, IsotopiC constraints on biogeochemical cycling of Fe: Reviews in Mineralogy and Geochemistry, v. 55, p. 359-408.

Kepert, D.A. , 2001, Carbonate debris flows of the Joffre Member, Hamersley province, Weste rn Australia [ext. abs]: Geoscience Australia Record, v. 2001/37, p. =-240.

Kerr, R.A., 2005, The story of O2: Scie nce, v. 308, p. 1730-1732. Kenich. R., Goldfarb, R., and Richards, J.R. , 2005, Metallogenie provinces

in an evolving geodynamiC framework: ECONOMIC GEOLOGY l(X)'"I ANNlVERSARY VOLUME, p . 1097- 1136.

Kimberley. M.M. , 1989, Exhalative origins of iron fo rmations: Ore Geology Reviews, v. 5, p. 13-145.

King, H.E., 1989, The rocks speak: Australasian Institute of Mining and MetaUu'llY Monograph 15, 316 p.

Klein, G , 1983, Diage nesis and metamorphism of banded iron-formations, ill Trendall, A.F., and Monis, R.C .• eds., Iron-formation: Facts and problems: Amsterdam, Elsevier, p. 417-469.

Klein. C. , and Beukes, N.J., 1989, Geochemistry and sedimentology of a facies transition from limestone to iron-formation deposition in the early Proterozoic Transvaal Supergroup, South Africa: ECONOMIC GEOLOGY, v. 84, p . 1733-1774.

--1992, Proterozoic iron formations , in Condie, K.C .• ed., Proterozoic crustal evolution: Amste rdam , Elsevier, p. 383-418.

Klein , G , and Bricker, O.P. , 1977, Some aspects of the sedimentary and diagenetic environment of Proterozoic banded iron-formation: ECONOMIC GEOLOGY, v. 72, p. 1457-1470.

KJein . c., and Fink, RP. , 1976, Pe trology of the Sokoman Iron Formation in the Howells River area, at the weste rn edge of the Labrador trough: EcoNOMIC GEOLOC)', v. 71 , p. 453-487.

Klein , c., and Lade ira, E.A., 2004, Geoche mist ry and mineralogy of Neoprote rozoic banded iron -fonnations and some selected, siliceous man· ganese formations from the Uruc um district , Mato Grosso do Sui. Brazil: ECONOMIC GEOLOGY, v. 99, p. 1233- 1244.

Kneeshaw, M., 1975, Mount Whaleback orebody: Australasian Institute of Mining and MetalJurgy Monograph 5, p. I90-916.

Kneeshaw, M .• Kepert , D.A., Tehnas, I.J., and Pudovskis, M.A., 2002, From Mt Goldsworthy to Area C-reflections on forty years of iron ore exploration in the Pilbara: Australasian Institute of Mining and Metallurgy Publication Series 71'lJX)2, p. 41-56.

Kneeshaw, M., and Kepert, D .A., 2002, Genesis of high-grade he matite orebodies of the Hamersley province, Western Australia--a discussion: EcoNOMIC GEOLOGY, v. 97, p. 173.

Knoll, A.H., 2003, Life on a young planet-the first three billion years of ev0-

lution on Earth: Princeton, NJ, Princeton University Press, 277 p . Konhauser, K.O., Hamade, T., Raiswell , R., Monis, R.C ., Ferris, F.G.,

Southam, C., and Canfield, D., 2002, Could bacteria have formed the Precambrian banded iron fonnations?: Geology, v. 30, p. 1079-1082.

Krapez, B., and Martin, D.McB., 1999, Sequence stratigraphy of the Palaeoproterozoic Nabberu province of Western Australia: Australian Journal of Earth Sciences, v. 46, p. 89-103.

Kump, L.R., and Seyfried, W.E ., Jr., 2005, Hydrothennal Fe fluxes during the Precambrian: Effect of low oceanic sulfate concentrations and low hydrostatic pressure on the composition of black smokers: Earth and Planetary Science Letters, v. 235, p. 654-662.

LaBerge, C .L., 1966a, Altered pynx:lastic rocks in iron-fonnation in the Hamersley Range. Western Australia: ECONOMIC GEOLOGY, v. 61, p. 147-161.

--I966b, Pyroclastic rocks in South African iron-fonnations: ECONOMIC GEOLOGY, v. 61. p. 572-58J..

LaBerge, G.L., Robbins. E.I.. and Han, T.-M. , 1987, A model for the biological precipitation of Precambrian iron-formations. A: Geological evidence, in Appel, P.W.U., and LaBerge , G .L., ed.s ., Precambrian iron· fonnations: Athens, Greece, Theophrastus Publications, p. 69-96.

Larue, D.K., 1981, The early Proterozoic pre-iron-fonnation Menominee Croup siliciclastic sediments of the southern Lalce Superior region: Evidence for sedimentation in platfonn and basinal settings: Journal of Sedimentary Petrology, v. 51, p. 397--414.

Lascelles, D ., 2002, A new look at old rocks-an alternative model for the origin of in situ iron ore deposits derived from banded iron-formation: Australasian Institute o f Mining and Metallurgy Publication Series 712002, p. 107-126.

Lepp, H., 1987, Chemistry and origin of Precambrian iron-fonnations, in Appel, P.W.U., and LaBerge, C .L., cds., Precambrian iron-fonnations: Athens, Greece, Theophrastus Publications, p. 3-30.

Li, Z.x. , PoweU, C .McA, and Bowman, R , 1993, liming and genesis of Hamersley iron-ore depoSits: Exploration GeophysiCS, v. 24, p. 631--636.

Lobato, L.M. , Rosi~re, GA., Baars, F.J., Silva, R.C.F., Zuchetti, M .. Rios, F.J., Seoane, J.C.S .• Pimentel, M., Undenmayer, Z.G., and Mendes, G.E ., 2004 , TIming of hydrothermal iron mineralization, Carajas province [ext. abs.]: Simp6sio Brasileiro de Explo~ Mineral, Anais . Ouro Pre to, Minas Gerais, CD-ROM.

Lobato. L.M. , Silva, R.C.F., Rosi~re , GA., Zucchetti, M., Baars, F.J., Pimentel. M., Rios, F.J., Seoane, J.C.S. , and Monte iro, A.M., in press, Hydrothennal origin for the iron mine ralization , Carajas Province, Para State. Brazil: Iron Ore 2005, Perth , Austraha.

Lowe, D.R , 1992, Major events in the geological development of the Precambrian earth, in Schopf. J.w. , and Klein, c., cds., 1992, The Proterozoic biosphere-a multidisciplinary study: Cambridge, Cambridge University Press, p. 67-75.

MacLeod. W.N., 1966, The geology and iron deposits of the Hamersley Range area, Western Australia: Geological Survey of Western Australia Bulletin 117, 170 p .

Maliva, R.C ., Knoll, A.H., and Siever, R., 1989, Secular change in chert distribution : A reflection of evolving biological partiCipation in the silica cycle: PALAIOS, v. 4, p. 51S-532.


MaJiva, R.C., Knoll, A.H., and Simonson, B.M., 2005, Secular change in the Precambrian silica cycle: Insights from chert petrology: Geological Society of America Bulletin, v. 117, p. 835--845.

Manikyamba. C., 1999, Reworking of BIF into CIF in the Sandur schist belt. India: Possible evidence of sea level changes in an Archaean proto-ocean: Journal of the Geolog;ca1 Society of India, v. 53, p. 453-461.

Martin, D.MeS .• Lit ZX. Nemchin, A.A., and Powell, C.McA., 1998. A pre-2.2 Ga age for giant hematite ores of the Hamersley province, Australia?: EcoNOMIC GEOLOGY, Y. 93, p. 1084-1090.

Mason, T., and McSpadden. G., 2002, Perfonnance enhancement at BHP Billiton 's Newman Beneficiation Plant: Metallurgical Plant Design and Operating Strategies, 2002, Sydney, Australasian Institute of Mining and Metallurgy, Proceedings, p. 81-104.

McConchie, D. , 1987, The geology and geochemistry of the Joffre and Whaleback Shale Memben of the Brockman Iron Formation, Western Australia, in Appel, P.W.U. , and LaBerge. G.L., eds., Precambrian iron·for· mations: Athens, Greece, Theophrastus Publications, p. 541-597.

McKim, A.M., 1970, Wabush story. Part III. Ore dressing. Scully mine: An· nual Mining Symposium, 311', Duluth, Minnesota, Proceedings.

McLellan, J.C., Oliver, N.H.S., and Schaubs, P.M. , 2004, Fluid flow in ex· tensional environments: N umencal modeling with an application to Hamersley iron ores: Jouma1 of Structural Geology, v. 26, p. 1157-1161.

Mengel, J.T., 1973. Physical sedimentation in Precambrian cherty iron for· mations of the Lake Superior type. in Amstutz, G.c. , and Bernard, A.j.. eds., Ores in sediments: Berlin, SpringerNerlag. p. 179-193.

Miller, O.j. , 2002. Integration of the beneficiation process into iron ore re· source utilisation: Australasian Institute of Mining and Metallurgy Publica· tion Series 712002, p. 243-250.

Miyano, T., 1987, Diagenetic to low-grade metamorphic conditions of Pre· cambrian iron-formations. in Appel , P.W.U., and LaBerge, G.L, eels., Precambrian iron-fonnations: Athens, Greece, Theophrastus Publications, p. 155-186.

Morey, C.B.. 1983, Animikie basin. Lake Superior region, U.S.A., in Trendall, A.F .• and Morris. R.C., lron·fonnation: Facts and problems: Amster· dam, Elsevier, p. 13-67.

Morey, C .B. , and Southwick, D.L., 1995, Allostratigraphic relationships of Early Proterozoic iron·fonnations in the Lake Superior region: ECONOMIC GEOLOCY, v. 90, p. 1983-1993.

Morris, R.C., 1980, A textural and mineralogical study of the relationship of iron ore to banded iron fonnation in the Hamersley iron province of Westem Australia: ECONOMIC GEOLOGY, v. 75, p. 184-209.

--1985, Genesis of iron ore in banded iron-formation by supergene and supergene.metamorph.ic processes---a conceptual model, in Wolf. K.H ., ed., Handbook of strata-bound and stratifonn ore deposits: Amsterdam, EI· sevier, v. 13, p. 73-235.

--1986, A review of geological research on the iron ores of the Hamersley Iron province. in Berkman, D.A., ed., Publications of the CMMI Congress. 13th , v. 2, Geology and Exploration, Singapore, May 1986: Congress of the Council of Mining and Metallurgical Institutions and the AustraJasian Institute of Mining and Metallurgy, p. 191-210.

--1987, Iron ores derived by the enrichment of banded iron·fonnation, in Hein. J.R., ed., Siliceous sedimentary rock-hosted ores and petroleum: New York, NY, Van Norstrand Reinhold, p. 231-267.

--1993, Genetic modelling for banded iron·fonnation of the Hamersley Group, Pilbara Craton, Western AustraJia: Precambrian Research, v. 60, p. 24.'!-286.

--1998, BIF·hosted iron ore deposits-Hamersley style: Journal of Aus· tralian Geology and Geophysics, v. 17, p. 207-211.

--200Za, Cenesis of high.grade orelxxlies of the Hamersley province. Western AustraJia-a discussion: ECONOMIC CEOUX;Y, v. 97, p. li7-178

--2002b, Iron are genesis and post-ore metasomatism at Mount Tom Price: Australasian Institute of Mining and Metallurgy Publication Series 71'JOO2, p. 3-13.

Muller, S.G .• Krape:t, B .• Barley, M.E., and Fletcher, LR., 2005, Giant ironore deposits of the Hamersley province related to the breakup of Paleo-proterozoic Australia: New insights from in situ SH RI MP dating of baddeleyite from mafic intrusions: Geology, v. 33, p. 577-580.

Neal, H .E., 2000, Iron are deposits of the Labrador trough: Exploration Mining Geology, v. 9, p. 113-121.

Nelson. D.R., Trendall , A.F., and Altennann. W, 1999, Chronological correlations between the Pilbara and Kaapvaal cratons: Precambrian Research, v. 97, p. 16&-189.

Netshiozwi, S.T .. 200-2, Origin of high·grade hematite ores at Thabazimbi mine, Limpopo province. South Africa: Unpublished M.Sc. thesis. Auck· land Park, Rand Afrikaans University, 134 p.

Nutman, A.P., Allaart. J.H., Bridgwater, D., Dimroth, E., and Rosing, M., 1984, Stratigraphic and geochemical evidence for the depositional environment of the Early Archaean Isua supracrustal belt. West Creenland: Precambrian Research. v. 25, p. 365-396.

Ohmoto, H., 2OO'l, Nonredox transformations of magnetite-hematite in hydrothennal systems: ECONOMIC GEOLOGY, v. 98, p. 157-161.

Ojakangas, RW., 1983, Tidal depoSits in the early Proterozoic basin of the Lake Superior region-the Palms and Pokegama Formations: Evidence for subtidal shelf deposition of Superior-type banded iron-fonnation: Geological Society of America Memoir 60, p. 49-66.

Oliver, N.H.S., and Dickens, C.R., 1999, Hematite ores of Australia formed by syntectonic heated meteoric fluids , in Stanley, C.J., et aI., Mineral de· posits: Processes to processing: Rotterdam, A.A. Balkema .. v. 2, p. 889-892.

Oliver. N.H.S., Dickens, C.R., and Powe ll, C.McA., Bons, P.D, and Stewart, L.K., 1998. Hamersley hematite ore genesis by syntectonic infiltration of heated meteoric fluid: Isotopic, geochemical and microstructural constraints [abs.]: Australian Geological Convention, 14th, Australian Geological Society, Proceeding', v. 49, p. 340.

Pickard, A.L., 2ro2, SHRIMP U·Pb zircon ages of tuffaceous mu~ks in the Brockman Iron Fonnation of the Hamersley Range, Western AustraJia: Australian Journal of Earth Sciences, v. 49, p. 491-507.

--2003, SHRIMP U-Pb zircon ages for the Palaeoproterozoic Kuruman Iron Fonnation, Northern Cape province, South Africa: Evidence for simultaneous BIF deposition on Kaapvaal and Pilbara cratons: Precambrian Research, ~ 125, p. 275-315.

Pickard, A.I.... Barley. M.E.. and Krnpez, B. , 2004, Deep-marine depositional setting of banded iron formation: Sedimentological evidence from interbedded clastic sedimentary rocks in the early Palaeoprot'erozoic Dales Gorge Member of Western Australia: Sedimentary Geology, v. 170, p. 37-62.

Pinheiro, R.V.L .• and Holdsworth, RE., 1997, The structure of the Carajas N-4 Ironstone deposit and associated rocks: Relationship to Archaean strike-slip tectonics and basement reactivation in the Amazon region, Brazil: Joumal of South American Earth Sciences, v. 10, p. 305-319.

Pires, F.R.M .• 2002, Distribution of hard hematite are at the Quadrihitero Ferrifero, Minas Gerais, Brazil and its possible genetic Significance: AustraJasian Institute of Mining and Metallurgy Public-cltion Series 712002. p. 71-76.

Pavey, B.C.. and Leather, K.G., 1997, Upgrading of Australian magnetite orelxxlies, III Misra, Y.N., and Dunlop. J.S., eds., Towards 200Q..National Conference on lronmaking Resources and ReselVes Estimation: Australasian Institute of Mining and Metallurgy Spectrum Series, Proceedings, p. [53-156.

Powell, C.McA., Oliver, N.H.S., Li, Z.x.. Martin, D.MeS., and Ronaszecki, J., 1999, SynorogeniC hydrothennal origin for giant Hamersley iron oxide orelxxlies: Geology, v. 27, p. 175-178.

Ribeiro, D.T, Pires. F.R.M .. and Carvalho, R.M., 2002, Supergene iron ore and disorder: Australasian Institute of Mining and Metallurgy Publication Series 7/2002. p. 81-89.

Rio TInto, 2004. Unpublished Annual Report: london, U.K., Rio linto, 155 p. Rios. FJ., Lobato, L.M ., Rosii!re, C.A., Silva, R.C.F., and Souza, A.S., 2004.

Resultados preliminares do estudo metalogenetico do minerio hematftic.'O de alto teor do dep6sito de ferro N5-CarajAs, utilizando microscopia e microtennometria de infravennelho [ext. nbs.]: Simp6sio Brasileiro de Explorar;ao MineraJ, Anais. Ouro Preto, Minas Centis, CD-ROM.

Rosif~re , CA, and Chemale, F., Jr. , 1991, Textural and structural aspects of iTOn ores from Iron Quadrangle, Brazil: Society for Geology Applied to Mineral Deposits Anniversary Meeting, 25th , Nancy, Proceedings, p. 485-489.

Rosil!re. CA., and Rios, F.J .. 2004, The origin of hematite in high.grade iron ores based on infrared microscopy and fluid inclusion studies: The example of the Conceir;do mine, Quadril:\tero Ferrifero, Brazil: ECONOMIC GEOL· OCY, v. 99, p. 611-624.

Rosi~re, C.A., Siemes, B .• Quade, B .• Brokmeier, H .·C., and Jansen, E .• 2001, Microstructures, textures and deformation mechanisms in hematite: Journal of Structurnl Geology, v. 23, p. 1429-1440.

Rosi~re, C.A., Siemes, H .• Rios, F.j. , and Quade, H., 2002, Defonnation-con· trolled high-grade iron ores [ext. abs.); Quadrennial IAGOD Symposium and Geocongress, 11th, Windhoek, Namibia, Geological Survey of Namibia, CD-ROM.

IRON FORMATiONS AND ASSOCIATED IRON ORE DEl'DsrrS 679

Hosiere, C.A., Seoane, J.C.S., Baars, F.J ., and Lobato. L.M., 2004, A estruturacao da Provincia de Carajas e sua influencia na mineraliza\=ao de ferro [ext. abs.] ; Simp6sio Brasileiro de Explora~ao Mineral, Anais. Duro Prete, Minas Gerais, CD· ROM.

Rauxel, 0 .]. , Bekker, A., and Edwards, K.J ., 2005, I ron isotope constraints on the Archean and Paleoproterozoic ocean redox state: Science. v. 307, p. 10SS-109!.

Schneider, D .A., Bickford. M.E ., Cannon, W.F., Schulz. K J .• and Hamilton, M.A., 200-Z, Age of volcanic rocks and syndepositional iron (onnations, Marquette Range Supergroup: Implications for the tectonic setting of Pa· ieoproterozoic iron fonnations of the Lake Superior region: Canadian Journal of Earth Sciences, v. 39, p. 999-1012.

Shegelski , R.J., 1987, The derx)sitional environment of Archean iron fannations, Sturgeon-Savant greenstone belt, Ontario, Canada, it! Appel, P.W.U., and LaBerge. G.L, eds .. Prec-dmbrian iron-fonnations: Athens, Greece, Theophrastus PubliC"dtions, p. 329-344.

Silva. R.R.R, Amamnte, S.C., Souza, C.C., and Araujo, A.C .. 2002, Ore minera.1ogy and its relevance for the selection of concentration methods in the processing of Brazilian iron ores: Australasian Institute of Mining and Metallurgy Publication Series 712002, p. 159-169.

Silva, R.C.F., Lobato. L.M., Ros i~re, C.A., Cued.es, S.C., Monteiro. A., Me ireles, H., and Matias. P.H ., 2004. Estudos petrogn1ficos microsc6picos. geoqufmicos e geocronol6gicos em jaspilitos e min~rio de ferro dos dep6sitos NI , N4W. N4E e N5E, Provincia Mineral Caraja.. [ext. abs.]: Simp6sio Brasile iro de Explom~ao Mineral, Amus . Duro Preto, Minas Gerais, C D-ROM.

Simonson , B.M., 1984, High..energy shelf deposit: Early ProteTOZQic Wishart Fonnation, northeastern Canada: Society of Economic Paleontologists and Minera10gists Special Publication 34, p. 251-268.

--1985, Sedime ntological constraints on the origins of Precambrian ironformations: Geological Society of America Bulletin, v. 96, p. 244-252.

--1987, Early silica cementation and subsequent diagenesis in arenites from four early Proterozoic iron formations of North Ame rica: Journal of Sedimentary Petrology, v. 57, p. 494-511.

--2003, Origin and evolution of large Precambrian iron formations: Ge0-logical Society of America Special Paper 370, p. 23 1-244.

Simonson, B.M., and Carney, K.E., 1999, Roll-up structures: Evidence of in situ microbia1mats in Late Archean deep shelf environments: PAl.Al0S, v. 14, p. 13-24.

Simonson , B.M., and Goode, A.D.T., 1989, Fi rst di..soovery of ferruginous che rt arenites in the early Precambrian of Weste rn Australia: Geology, v. 17, p. 269-272.

Simonson, B.M., and Hassle r, S.w., 1996, Was the depoSition of large Precambrian iron fontlations linked to major marine transgression?: Journal of Geology, v. 104, p. 66:>-676.

Simonson, B. M., Schu bel, K.A., Ilnd Hassle r, S.w. , 1993, Carlxmate sedimento logy of the early Prec-dmbrian Hamersley Croup of Western Aus· tra.1ia: Precambrian Research, v. 50, p. 287-335.

Singer, A., and MUlie r, C., 1983, Diagenesis in argillaceous sedi ments, in Larsen, G., and Chilingar, G. V., eels., Diagenesis in sedime nts and sedimentary rocks, 2: Amsterdam, Elsevier, p. 115-212.

Spier, C.A., Oliveria, S.M .. and Rosiere, C.A .. 2003. Geology and geoche mistry of the Aguas Claras and Pico iron mines, QuadrihUe ro Ferrifero, Minas Gerais, Brazil : Mine ralium DepoSita, v. 38, p . 751- 774.

Taylor, D. , Dalstra, H .J., Harding, A.E., Broadbent, C .C., and Barley, M.E , 2001, Genesis of high-grade he matite orebodies of the Hamersley province, Weste rn Australia: ECONOMIC GEOLOGY, v. 96, p. 837-873.

Taylor, D ., Dalstra, H .j . and Harding, A.E ., 2002, Genesis of high-grade hematite orebodies of the Hamersley province, Western Australia-a reply: ECONOMIC CEOUX;\', v. 97, p. 179-181.

Thome, A.M., and Trenda1I, A.F., 2001 , Geology of the Fortescue Group, Pilbara craton , Western Australia: Geologica1 Survey of Western Australia Bulletin 144, 249 p.

Thome, W.S., Hagemann, S.C., and Barley, M., 2004, Petrographic and geoche mical eviden<.'C for the hydrothermal evolution of the North deposit , Mt. Tom Price, Weste rn Australia: Mineralium DepoSita, v. 39, p. 766-783.

Trendall, A.F., 1972. Revulutiun in ('a l1!' lIiston ': j Ol1rl1ll1 of til(' GeologiCi:,1 Society of Australia, v. 19. p. ZSi.....;) I I . .

--1973, ValVe cycles in tht' Wet·li Wulli Fllnlmlillil (If the Prc...ocamhrian Hamersley Group, \ ,yestern Austmlia: ECll/liOMll CEOUK:'·. v. fiB, p. 1089-1098.

--1983, The Hamersley basin, ill Trcmlall. A. I-'., 111111 ~·l lIrris. H.C .. ells .. l ron-fonnation: Facts and problems: Amsterdam. EISt·v;(·r. p. 6Y-- I2'J.

--2002, The Significance of iron-fornmtion ill thl ' Prt 'l:IIUl h ri:1II stml igraphic record: Inte rnational Association of St!tlilllcnt(lluJ.:ist~ Sllt'Cllil PuhIication 33, p. 33--66.

Trendall , A.F., and Blockley, J.G., 1970, The iron fonnations of 11lC' I'rt .. 'l:IIIII

brian Hamersley Group, Western Australia: Geological SurVl')' of Wl'st f'nl Australia Bulletin 119,366 p.

Tre ndall , A.F., Nelson , D.R., De Lae te r, J.R., and Hassler, S.W .. \ gyM. Pn'eise zircon U·Pb ages from the Marra Mamba Iron Fommtion amltlw Wi! tenoom Fonnation, Hamersley Croup, Western Australia: AlIstrdliilll Journal of Earth Sciences, v. 45, p. 137-142.

Trendall , A.F., Compston , W., Nelson, D.R., De Laeter, J.R., and Bennett . V.C., 2004, SHRIMP zircon ages constraining the depositional chronology of the Hamersley Croup, Western Australia: Australian Jouma1 of Earth Sciences, v. 5 1. p . 621~.

Van Hise, G R , and Leith , C. K. , 1911. The geology of the Lake Superior region: U.S. Geological SUlVey Monograph 52, 641 p.

Van Schalkwyk, J.F., and Beukes, N.J., 1986, The Sishen iron ore depoSit, in Anhaeusser, C. R., and Maskes, S., eels., Mine ral deposits of Southe rn Africa: Johannesburg, GeolOgical Society of South Africa, v. 1-2, p. 93 1-956.

Veizer, J., Clayton. R.N., Hinton, R.w. , von Brunn, v., Mason , T.R. , Buck, S.C ., and Hoefs, J., 1990, GeochemiStry of Precambrian carbonates: Ill. Shelf seas and non-marine environme nts of the Archean: Geochimica Cosmochimica Acta, v. 54, p. 2717-2729.

Veizer, J., Clayton, R.N., and Hinton, R.W., 1992, Geochemistry of Precambrian carbonates: rv. Early Paleoproterozoic (2.25 :t. 0.25 Ca) seawater: Geochimica Cos mochimica Acta, v. 56, p. 875--885.

Wa1te r, M.R., 1972, A hot spring analog for the depoSitional environment of Precamblian iron fonnatious of the Lake Superior region: ECONOMIC Gt:OLOCY, v. 67, p. 965-972.

Wa1te r, M.R., and Hofmann, H.J., 1983, The paJaeontology and palaeoecology of Precambrian iron-fornlations, In Trendall , A.F., and Morris, R e., eels. Iron-fonnation: Facts and problems: Amsterdam, Elsevier, p. 373-400.

Warnock, WW., and Bensley, C.N., 1996, The Port Hedland hot briquetted iron project: Australasian Institute of Mining and Metallurgy Annua1 Confe ren<.'C, Perth , Proceedings, p. 197-202.

Watchorn, M.A., 1980, Fluvial and tidal sedimentation in the 3000 Ma Mozaan basin, South Africa: Precambrian Research , v. 13, p. 27-12.

Webb, A.D., Oliver, N.H.S., and Dickens, D .R., 2002, Multistage evolution of high grade hematite orebodies from the Hamersley province, Western Australia: Australasian Institute of Mining and Metal lurgy Publication Series 712002, p. 91-94.

Webb, A.D., Dicke ns, C ., and Oliver, N. H.S ., 2003, From banded. iron-formation to iron ore: Geochemical and mineralogicaJ constraints from across the Hamersley province, Weste rn Australia: Chemical Ceology, v. 197, p. 21S-25!.

Winter, B.L., and Knauth , L,P., 1992, Stable isotope geoche mistry of cherts and carbonates from the 2.0 C a Gunflint Iron Formation : Implications for the depositiona1 setting, and the effects of diagenesis and metamorphism: Precambrian Research, v. 59, p. 283-313.

Young. G.M., 1976. I ron-formation and glaciogenic rocks of the Rapitan Croup, Northwest Te rritories, Canada: PreC'"dmbrian Research, v. 3, p. 13;-158.

--1988, Proterozoic plate tectonics, glaciation and iron-formations: Sedimentary Geology, v. 58, p. 127-144.

Zajac, I.S., 1974. The stratigraphy and mine .... uogyofthe Sokoman Formation in the Knob Lake area, Quebec and Newfoundland : GeolOgical SUlVeyof Canada Bulletin 220, 159 p.

Genesa Iron - World

Documents

iron carbonates

large iron formations

precambrian iron formations

abstract iron formations

iron formation gangue

onomic iron ore deposits

banded iron formations

sulfides iron oxide