precambrian banded iron-formations

American Mineralogist Volume 90 pages 14731499 2005

0003-004X0500101473$0500DOI 102138am20051871 1473

E-mail ckleinunmedu

PRESIDENTIAL ADDRESS TO THE MINERALOGICAL SOCIETY OF AMRICA BOSTON NOVEMBER 6 2001

Some Precambrian banded iron-formations (BIFs) from around the world Their age geologic setting mineralogy metamorphism geochemistry and origin

CORNELIS KLEIN

Department of Earth and Planetary Sciences University of New Mexico Albuquerque New Mexico 87131 USA

ABSTRACT

Banded iron-formations (BIFs) occur in the Precambrian geologic record over a wide time span Beginning at 38 Ga (Isua West Greenland) they are part of Archean cratons and range in age from about 35 until 25 Ga Their overall volume reaches a maximum at about 25 Ga (iron-formations in the Hamersley Basin of Western Australia) and they disappear from the geologic record at about 18 Ga only to reappear between 08 and 06 Ga

The stratigraphic sequences in which BIFs occur are highly variable Most Archean iron-formations are part of greenstone belts that have been deformed metamorphosed and dismembered This makes reconstruction of the basinal setting of such BIFs very difTHORN cult The general lack of metamorphism and deformation of extensive BIFs of the Hamersley Range of Western Australia and the Transvaal Supergroup of South Africa allow for much better evaluations of original basinal settings Most Archean iron-formations show THORN ne laminations andor microbanding Such microbanding is especially well developed in the Brockman Iron Formation of Western Australia where it has been interpreted as chemical varves or annual layers of sedimentation BIFs ranging in age from 22 Ga to about 18 Ga (eg those of the Lake Superior region USA Labrador Trough Canada and the Nabberu Basin of Western Australia) commonly exhibit granular textures and lack microbanding

The mineralogy of the least metamorphosed BIFs consists of combinations of the following minerals chert magnetite hematite carbonates (most commonly siderite and members of the dolomite-ankerite series) greenalite stilpnomelane and riebeckite and locally pyrite Minnesotaite is a common very low-grade metamorphic reaction product The Eh-pH stability THORN elds of the above minerals (andor their precursors) indicate anoxic conditions for the original depositional environment

The average bulk chemistry of BIFs from 38 through 18 Ga in age is very similar They are rich in total Fe (ranging from about 20 to 40 wt) and SiO2 (ranging from 43 to 56 wt) CaO and MgO contents range from 175 to 90 and from 120 to 67 wt respectively Al2O3 contents are very low ranging from 009 to 18 wt These chemical values show that they are clean chemical sediments devoid of detrital input Only the Neoproterozoic iron-formations (of 08 to 06 Ga in age) have very different mineralogical and chemical make-ups They consist mainly of chert and hematite with minor carbonates

The rare-earth element proTHORN les of almost all BIFswith generally pronounced positive Eu anomalies indicate that the source of Fe and Si was the result of deep ocean hydrothermal activity admixed with sea water

The prograde metamorphism of iron-formations produces sequentially Fe-amphiboles then Fe-pyroxenes and THORN nally (at highest grade) Fe-olivine-containing assemblages Such metamorphic reactions are isochemical except for decarbonation and dehydration

The common THORN ne lamination (andor microbanding) as well as the lack of detrital components in most BIFs suggest that such are the result of deposition below wave base in the deeper parts of ocean basins Those with granular textures are regarded as the result of deposition in shallow water platformal areas Carbon isotope data suggest that for a long period of time (from Archean to Early Proterozoic) the ocean basins were stratiTHORN ed with respect to δ13C (in carbonates) as well as organic carbon content In Middle Proterozoic time (when granular BIFs appear) this stratiTHORN cation diminishes and is lost

The Neoproterozoic BIFs occur in stratigraphic sequences with glaciomarine deposits These BIFs are the result of anoxic condi-tions that resulted from the stagnation in the oceans beneath a near-global ice cover referred to as Snowball Earth

All of the most primary mineral assemblages appear to be the result of chemical precipitation under anoxic conditions There are as yet no data to support that BIF precipitation was linked directly to microbial activity The relative abundance of BIF throughout the Precambrian record is correlated with a possible curve for the evolution of the O2 content in the Precambrian atmosphere

INTRODUCTION

This paper deals with a selection of Precambrian iron-forma-tions ranging in age from 38 Ga to about 07 Ga on which col-leagues and I have done research over almost a 30 year period The term iron-formation as used here is similar to that of James (1954) with some modiTHORN cations as suggested by Trendall (1983) As such iron-formation is deTHORN ned as a chemical sediment typically thin bedded or laminated whose principal chemical characteristic is an anomalously high content of iron commonly but not neces-sarily containing layers of chert In James original deTHORN nition

of 1954 a quantitative lower limit of Fe content (ge15 wt Fe) was incorporated This arbitrary lower limit is commonly too restrictive in the evaluation of rock types that reszlig ect a range in Fe content (from ferruginous to Fe-rich) all of which show the pertinent characteristics of iron-formation and which are part of Fe-rich sedimentary sequences In this paper all iron-formations are referred to as BIF (meaning that they are banded) even though some might be better described as IF (Trendall 2002) Most well-banded IF (BIF) occurrences are generally older than about 20 Ga Even those BIFs that have been highly metamorphosed retain relict banding The iron-formations of the younger Lake Superior region USA and the Labrador Trough Canada are commonly

KLEIN SOME PRECAMBRIAN BANDED IRON-FORMATIONS1474

granular but also show banding (on a scale of tens of centimeters instead of millimeters) and these have been referred to as GIF (granular iron-formation) by Trendall (2002) In this paper all are grouped under the name BIF This overview of iron-formations does not include a discussion of Phanerozoic ironstones although Kimberley (1978) suggested that these Phanerozoic deposits be considered iron-formations Most Phanerozoic ironstones are very different from Precambrian BIF in texture as well as mineralogy They are commonly ooidal in texture and many contain abundant silicates such as berthierine and chamosite (Young and Taylor 1989) that are generally absent from Precambrian iron-formations The overall mineralogy of ironstones consisting of hematite (or goethite) magnetite berthierine siderite-rich types and lacking chert is totally different from that of Precambrian BIF

The global distribution of well-known Precambrian IFs is shown in Figure 1 This map (modiTHORN ed after Trendall 2002) emphasizes their wide global distribution and does not attempt to show all known occurrences

DISTRIBUTION OF IRON-FORMATIONS THROUGHOUT PRECAMBRIAN TIME

A schematic relative abundance curve for Precambrian iron-formations as a function of age is given in Figure 2 This curve which is adapted from that of Gole and Klein (1981a) is in essence a best estimate of total volume of iron-formation (not Fe ore) for all iron-formations tabulated by James (1983) James and Trendall (1982) Walker et al (1983 which lists the age ranges of 54 BIFs in their Table 111) and Fyon et al (1992 with Table 224 listing the ages of several Archean iron-forma-tions in the Superior Province of Ontario) relative to a maximum represented by the total iron-formation volume of the Hamersley Range of Western Australia Figure 2 is similar but not identical to a curve of Trendall (2002 his Fig 6) in which the estimated volume of smaller Archean BIFs (in greenstone belts in Archean Shields) is smaller than that implied by the concave upward part of the curve in Figure 2 between about 35 and 27 Ga Gole and Klein (1981a) concluded that the size and extent of Archean iron-formations have commonly been underestimated Such BIFs

are generally discontinuous commonly tectonically deformed and dismembered and only partially exposed which leads to an underestimation of their original size and volume (see Fig 3) The elevated concave curve in Figure 2 (between 33 and 27 Ga) is the result of compensation for this underestimation in volume There is general agreement that the largest BIFs are those of the Hamersley Range of Western Australia (ranging in age from about 26 to 245 Ga Trendall and Blockley 1970 Trendall 2002 Tren-dall et al 2004) and the Transvaal Supergroup of South Africa with an age of approximately 25 to 23 Ga (Klein and Beukes 1989) There is additional agreement on a sharp decline in major iron-formations in the Precambrian record at about 18 Ga and the return of several BIFs in the geologic record that show ap-proximate ages between 08 to 06 Ga (such as iron-formations in the Rapitan Group Yukon and NWT Canada in the Urucum region Brazil and in the Damara Supergroup of Namibia)

STRATIGRAPHIC SETTING

The stratigraphic sequences in which BIFs occur are highly variable and there appears to be no consistent litho-

FIGURE 1 Global occurrence of Precambrian BIFs (after Trendall 2002) The map emphasizes the wide distribution of BIF and shows only a selection of well-known occurrences of different ages and kinds BIFs that are discussed in this paper are shown in italics

FIGURE 2 Highly schematic diagram showing the relative abundance of Precambrian BIFs vs time with several of the major BIFs or major BIF regions identiTHORN ed Estimated abundances are relative to the Hamersley Group BIF volume taken as a maximum (adapted from Gole and Klein 1981a also based on tabulation of BIF dates in Walker et al 1983 their table 111) The most recent age evaluations for the Hamersley Basin (28 to 22 Ga) are available from Trendall et al (2004) for the Labrador Trough BIFs (188 Ga) from Findlay et al (1995) and for the BIFs in the Frere Formation Western Australia (19 to 18 Ga) from Williams et al (2004)

KLEIN SOME PRECAMBRIAN BANDED IRON-FORMATIONS 1475

FIGURE 3 Illustration of the probable underestimation of the extent and volume of typical Archean BIFs in greenstone belts (left) Map of the outcrop pattern of BIFs in the Archean Yilgarn Block of Western Australia (right) Aeromagnetic map of the same region showing the much greater continuity and extent of the same BIFs (Gole and Klein 1981a unpublished THORN gure)

logical association either before or after iron-formation deposition (Gole and Klein 1981a) Figure 4 illustrates the diversity of lithologies associated with Archean as well as Proterozoic sequences The variations in the stratigraphic sequences shown in Figure 4 are not exhaustive but are sufficient to illustrate the wide diversity in lithological as-sociation Because of the largely volcanogenic nature of many Archean greenstone belts the lithologic associations of Ar-chean BIFs are probably somewhat less variable than those in the Proterozoic Although volcanic rocks dominate many Archean BIF sequences they are also present in Proterozoic iron-formation sequences (Fig 4 nos 6 and 8)

Most Archean BIFs are tectonically deformed and have been metamorphosed to various metamorphic grades which makes reconstruction of their depositional basins and overall geologic setting very difTHORN cult

Several of the largest BIF sequences in the world (in Brazil South Africa India and Australia) occur in well-preserved little-deformed supracrustal sequences rather than in greenstone belts Such sequences are generally only mildly deformed (Trendall 2002) They include in Brazil the Archean Carajaacutes Formation (Klein and Ladeira 2002) Early Proterozoic Kuruman and Griqualand Iron Formations of South Africa (Klein and Beukes 1989 Beukes and Klein 1990) the Mulaingiri Formation in India and the Hamersley Basin of Western Australia (Trendall and Blockley 1970)

Some major somewhat younger BIFs include those of the Lake Superior region in the USA and the Labrador Trough of Canada Both of these regions have been subjected to regional metamorphism The Lake Superior region includes the Negaunee Iron Formation of north-ern Michigan (James 1955 Haase 1982) the Mesabi Range of northern Minne-sota (French 1968) and the Gunflint Iron Formation in Ontario Canada (Floran and Papike 1975) The un-metamorphosed and meta-morphic assemblages of the Sokoman Iron Formation in the Labrador Trough have been reported by Klein (1966 1973 1974 1978 and 1983) and Klein and Fink (1976)

The Neoproterozoic BIFs of the Rapitan Canada and Urucum Brazil regions have totally different stratigraphic settings and are associated with glaciogenic litholo-gies such as diamictites and dropstone layers (Klein and Beukes 1993b Klein and Ladeira 2004)

SEDIMENTARY STRUCTURES Most Precambrian iron-formations especially those of

Archean age are well banded (see Fig 5a) Such banding is best preserved in low-grade metamorphic occurrences but even metamorphosed BIFs retain relict banding (eg at Isua West Greenland at amphibolite-facies metamorphism Dymek and Klein 1988) The various scales of banding that may exist have been deTHORN ned by Trendall and Blockley (1970) as follows macrobands are coarse alternations of contrasting rock types mesobands tend to have an average thickness of less than an inch (= 254 cm) and microbands usually range from 03 to 17 mm Trendall (1973b) and Trendall and Blockley (1970) have deTHORN ned microbanding as the alternation of Fe-rich and chert-rich laminae with one Fe-rich and one chert-rich lamina constituting a single microband (see Fig 5b) Such a microband is interpreted as an annual layer of sedimentationa chemical varve (Trendall 2002) Although small-scale structures such as laminations including microbanding are extremely well preserved in the Hamersley iron-formations similar structures have been noted in many Ar-chean occurrences see Fig 5c (Beukes 1973 Trendall 1973a 1973b Gole 1980) It should be noted however that Krape et al (2003) dispute the interpretation of microbanding in the BIFs of the Hamersley Basin as varves Instead they argue that such rhythmic features are the result of deposition from density currents Furthermore they interpret the precursor sediments to have been granular hydrothermal muds instead of Fe minerals and chert deposited from seawater

Other Precambrian BIFs may contain oolites granules and


FIGURE 4 Stratigraphic proTHORN les of THORN ve Archean and four Proterozoic iron-formation sequences The Archean iron-formations shown are (1) Moodies Group South Africa (2) Tallering Range Western Australia (3) Windarra Western Australia (4) Mount Belches Western Australia And (5) Hamersley Group iron-formations Western Australia The stratigraphic columns for 2 3 and 4 are somewhat schematic The location of the stratigraphic proTHORN les within each basin except for the Nabberu Basin (shown in column 9) is given beneath the supergroup or group name The stratigraphic columns for the Proterozoic sequences are from (6) Kuruman Iron Formation South Africa (7) Biwabik Iron Formation Minnesota (8) Sokoman Iron Formation Labrador Trough Canada (9) Frere Formation (Naberru Basin) Western Australia The stratigraphic columns were compiled from the following sources (1) Viljoen and Viljoen (1969) and Ericksson (1978) (2) Baxter (1971) (3) Roberts (1975) (4) Dunbar and McCall (1971) (5) Daniels and Halligan (1968) and Trendall and Blockley (1970) (6) Beukes (1973) (7) Bayley and James (1973) (8) Zajac (1974) (9) Hall and Goode (1978) Illustration modiTHORN ed from Gole and Klein (1981a)


other fragments embedded in a matrix and they may be banded as well Iron-formations with oolites and granules (see Fig 5d) appear to be mostly restricted to the Proterozoic as these structures have rarely been observed in Archean banded iron-formations (Gross 1972 Beukes 1973 Kimberley 1978) Oolitic and granular iron-formations occur in the Lake Superior region (James 1954 Goodwin 1956 Schmidt 1963) the Labrador Trough (Zajac 1974 Klein 1974 Klein and Fink 1976 Gross and Zajac 1983) and the Nabberu Basin of Western Australia (Hall and Goode 1978) It appears however that oolitic and granular types are either absent or extremely rare in most other Proterozoic sequences The banded and granular parts of iron-formations generally occur in different macrobands although laminated mesobands may be interbedded with mesobands containing granules Lamination is most commonly expressed by alternating 13 mm thick laminae of different Fe-minerals (Fig 5c) and microbanding in the strict sense is uncommon (Gole and Klein 1981a) Generally only quartz- or chert-rich mesobands are microbanded Because the

banded members consist mainly of Fe-silicate- and carbonate-rich assemblages with or without magnetite (French 1968 Floran and Papike 1975 Haase 1982 Klein 1974) it is perhaps not surprising that microbanding is not more common

Diagenetic to very low-grade metamorphic assemblages

Most Archean iron-formations have undergone various grades of metamorphism and their metamorphic assemblages will be discussed in a subsequent section only a few Archean occurrences have low metamorphic-grade assemblages However many of the well-studied Proterozoic BIFs that have undergone metamor-phism preserve sections in which late-diagenetic to very low-grade metamorphic lithologies can be evaluated Furthermore the very large BIFs of the Hamersley Basin Western Australia as well as of the Kaapvaal Craton South Africa contain well-preserved early assemblages because their metamorphic grades are very low The rocks in the Hamersley Basin are interpreted as having been metamorphosed to the sub-greenschist to greenschist

FIGURE 5 Dominant small structures in Precambrian banded iron-formations (a) Alternating massive and microbanded mesobands in magnetite-quartz-minor grunerite-bearing iron-formation metamorphosed to amphibolite grade from the Forrestania area in the Archean Yilgarn Block Western Australia (b) Microbanding in quartz-magnetite mesobands from the Joffre Member of the Hamersley Group Western Australia (c) Laminae in a minnesotaite-stilpnomelane-quartz-minor magnetite assemblage from the Negaunee Iron Formation northern Michigan (d) Oolites and granules composed of magnetite chert minor carbonate and stilpnomelane in a chert-rich matrix from the Sokoman Iron Formation Labrador Trough Canada Photographs from Gole and Klein (1981a)


facies (Klein and Gole 1981a) This implies temperatures during burial of 200 to 300 degC and a maximum pressure of burial of about 12 kbar The Kaapvaal South Africa BIF sequences have been subjected to estimated burial temperatures in the range of 100 to 150 degC (Miyano and Klein 1983a)

Most of the following discussion of diagenetic to low-grade metamorphic assemblages is derived from the petrologic study of the Negaunee Iron-Formation in northern Michigan the Bi-wabik Iron Formation in northern Minnesota the Gunszlig int Iron Formation in southern Ontario Canada and the Sokoman Iron Formation of the Labrador Trough Canada (see Klein 1983 for references to the original studies)

The minerals that are found in late-diagenetic and very low-grade metamorphic BIFs as well as their compositional ranges are listed in Table 1 Chert (and recrystallized quartz) is almost always present in all types of BIF It is most common in oxide-rich iron-formation types in association with magnetite hematite or both (see Figs 5a and 5b)

Magnetite is pervasive in oxide- silicate- carbonate- and carbonate-oxide-rich iron-formations Magnetite and pyrite may occur together in some of the most reduced assemblages that are part of sulTHORN de-rich BIF Magnetite also occurs with all three of the most common iron-silicates (greenalite stilpnomelane and minnesotaite) as well as with siderite and members of the ankerite-dolomite series Almost universally magnetite tends to be medium grained well crystallized and with a sub- to euhedral habit It is commonly coarser grained than coexisting chert (or quartz) hematite and Fe-silicates

Hematite is common in oxide-rich and carbonate-oxide-rich iron-formations occurring as hematite-rich bands as irregular con-centrations and in hematite-magnetite-chert (quartz) assemblages

The only early iron-silicate in such associations is stilpnomelane and minnesotaite is commonly present as a later reaction product of stilpnomelane Although hematite is clearly part of many early quartz-Fe-oxide occurrences in BIF it should be noted that many BIFs with a red color are the result of hematite that is a secondary oxidation product Such oxidation is especially common in BIFs in Brazil where deep lateritic weathering conditions exist (Klein and Ladeira 2000 2002) Members of the dolomite-ankerite series as well as calcite occur in hematite-rich assemblages

Pyrite is a major constituent of sulTHORN de-rich iron formations (eg James 1954 1955 French 1968 Klein and Fink 1976) In such BIF occurrences the pyrite grains are coarser than the other constituents (eg greenalite or chamosite carbonates and chert) and tend to have euhedral outlines (Klein and Fink 1976) Pyrite may be concentrated in thin well-deTHORN ned layers It is likely that mackinawite [Fe(1+x)S] is the sedimentary precursor to pyrite (Berner 1970) and it is for this reason that a mackinawite stability THORN eld is included in Eh-pH stability diagrams for late diagenetic sulTHORN de-rich BIF assemblages (see Fig 8)

Of the minerals in Table 1 it is greenalite that shows the least crystallized and as such probably the most primary tex-tures of any of the minerals in iron-formations It is generally light green microcrystalline and occurs in laminations oolites and granules in small irregular patches and as cements around granules (Figs 6a and 6b) It almost always has a grain size that is many orders THORN ner than that of any of its coexisting minerals (see Fig 6c) It has a restricted compositional range as shown in Figure 7

Many greenalite occurrences are criss-crossed and transected by two other silicates stilpnomelane and minnesotaite (Figs 6c and 6e) Stilpnomelane is most commonly a highly pleochroic

TABLE 1 Iron-formation minerals their compositions and approximate compositional ranges in diagenetic to very low grade metamorphic assemblages

Mineral name Simplifi ed composition Approximate compositional range

Chert (or quartz) SiO2 noneMagnetite Fe3O4 noneHematite Fe2O3 nonePyrite FeS2 noneGreenalite Fe6

2+Si4O10(OH)8 (Fe40Mg10 to Fe53Mg02) Al0-02Si4O10(OH)8Stilpnomelane (Fe Mg Al)27(Si Al)4(O OH)12xH2Odagger with traces of K Na Ca (Fe13Mg15Al01) (Si37Al03) to (Fe25Mg02) (Si36Al03) with K asymp 01 to 02 and Na asymp 005 per formula unitMinnesotaiteDagger Fe3

2+Si4O10(OH)2 Mg17Fe13 to Fe28Mg02Si4O10(OH)2ChamositeDagger (Fe2+ Al)6(Si Al)4O10(OH)8 (Fe33Mg13Al13) (Si30Al10) to (Fe38Mg13Al09) (Si28Al12)O10(OH)8RipidoliteDagger (Fe2+ Mg Al)12(Si Al)8O20(OH)16 Composition in iron-formation (Fe55Mg43Al23) (Si54Al26)O20(OH)16

2

Riebeckite Na2(Fe2+Mg)3Fe23+Si8O22(OH)2 Fe2+(Fe2++Mg) ranges from 064 to 086

Ferri-annite K2(MgFe)6Fe23+Si6O22(OH)4 Fe2+(Fe2++Mg) ranges from 050 to 071

Siderite FeCO3 (Mg03Mn01Fe06) to (Mg02 Mn02Fe06) CO3

Dolomite-ankerite CaMg harr CaFe(CO3)2 Ca10(Mg08Fe01Mn01) to Ca10(Mg05Fe02Mn03) to Ca10(Mg04Fe06) (CO3)2

Calcite CaCO3 Ca09(Fe Mg Mn)01CO3

Notes Compositional data on silicates and carbonates mainly from Klein (1974 1983) Klein and Fink (1976) Floran and Papike (1975) Miyano and Miyano (1982) and Miyano and Klein (1983b) The atomic proportions of the cations in the hydrous silicate formulas in this table are taken from electron microprobe analyses in the literature Such analyses are recalculated on an anhydrous basis only In this table these anhydrous cation recalculation results are used for the hydrous formulas given This procedure introduces slight error only (For comparative anhydrous and hydrous recalculations for the same hydrous mineral see Klein 1974 Tables 2 and 4) Total Fe assumed to be Fe2+onlydagger Eggleton (1972) on the basis of a structure determination suggests the following average formula for stilpnomelane (Ca Na K)4(Ti01Al23Fe355Mn08Mg93) (Si63Al9) (O OH)206nH2O This can be divided by 18 and reduced to (Fe Mg Al)27 (Si Al)4 (O OH)12xH2O The average number of (OH) in the (O OH)12 group is approximately 15 The additional (n or x) H2O values are highly variable Analyses of natural stilpnomelane show a considerable variation in Fe3+Fe2+ ratio Dagger Probably of later origin than the other minerals listed May be the result of reactions in late diagenetic to very low grade metamorphic conditions Therefore probably not part of the early diagenetic assemblages


FIGURE 6 Photomicrographs of diagenetic to very low-grade metamorphic BIF assemblages (a) Finely banded greenalite White patches at the top are minnesotaite (b) Greenalite oolite with THORN ne internal banding Lighter-colored greenalite cement at the top White patch in oolite is minnesotaite (c) Dark brown highly pleochroic stilpnomelane (St) sprays inside greenalite (G) Minnesotaite (m) cross-cutting stilpnomelane Opaque is magnetite (d) Finely banded magnetite(black)-siderite-stilpnomelane iron-formation Siderite is light colored stilpnomelane in randomly oriented sheaves (e) Minnesotaite masses in what originally had been greenalite granules Relict edges of granules and some greenalite cement (at bottom center) still remain [Photos a b c d and e all from the Sokoman Iron Formation Howells River area Newfoundland (Labrador Trough from Klein 1974)] (f) Siderite-rich iron-formation with dark bands of siderite and light bands of chert Kuruman Iron Formation South Africa From Beukes et al (1990)


(yellowish brown to dark brown light to dark green in some cases) sheaf-like silicate that occurs as thin continuous lamina-tions as very THORN ne-grained mattes of sheaves and needles and as coarse-grained sprays and patches Stilpnomelane is probably the most common Fe-silicate in most very low-grade metamor-phic iron-formations (Klein 1983) When it is associated with greenalite it commonly has a cross- cutting textural relationship with it THORN ne- to medium-grained sheaves of stilpnomelane criss-cross microcrystalline almost amorphous appearing masses of greenalite (Fig 6c) It can also occur in well-banded magnetite-stilpnomelane-carbonate assemblages (see Fig 6d) The general compositional range of stilpnomelane in terms of Al Mg and (Fe + Mn) is shown in Figure 7 Considerable amounts of K2O (1 to 92 wt) and lesser amounts of Na2O (0 to 08 wt) are present in this layer silicate structure

Of the three most common Fe-rich silicates in diagenetic to low-grade metamorphic BIF assemblages minnesotaite is generally not as abundant as stilpnomelane but it is much more common than greenalite It commonly occurs as THORN ne- to medi-um-grained needles arranged in sprays bow-ties and irregular patches (Figs 6c and 6e) Such textures cut across chert THORN ne- to medium-grained quartz carbonates microcrystalline greenalite as well as medium- to coarse-grained stilpnomelane sheaves In all such occurrences minnesotaite is a low-grade metamorphic reaction product of greenalite of stilpnomelane and of chert (or quartz) + Fe-carbonate Examples of such reactions are

Fe6Si4O10(OH)8 + 4SiO2 rarr 2 Fe3Si4O10(OH)2 + 2 H2Ogreenalite chert minnesotaite

2 Fe6Si4O10(OH)8 + O2 rarr 2 Fe3Si4O10(OH)2 + 2 Fe3O4 + 3 H2Ogreenalite minnesotaite magnetite

FeCO3 + 4 SiO2 + H2O rarr Fe3Si4O10(OH)2 + 3 CO2

siderite chert minnesotaite

Fe2middot7(SiAl)4(OOH)12middotxH2O + 033 Fe2+ rarr Fe3Si4O10(OH)2

stilpnomelane-like minnesotaite+H2O + Al + minor (NaK)

The compositional extent of the talc-minnesotaite series is shown in Figure 7

Chamosite and ripidolite are generally sporadic mineral components of some BIFs (Klein and Fink 1976) Chamosite an Fe-rich chlorite that contains considerable Al is common in Phanerozoic ironstones but is only a very minor constituent in BIF It is light green in color virtually nonpleochroic and extremely THORN ne-grained (Klein and Fink 1976)

Ripidolite an Fe-rich chlorite is a minor constituent of some low-grade metamorphic BIF assemblages but is a common con-stituent of the amphibolite-grade BIF assemblages of Isua West Greenland (Dymek and Klein 1988) There ripidolite occurs as part of the high-temperature equilibrium assemblages as well as a retrograde product in other assemblages

Riebeckite is a minor constituent of most BIFs but it and its THORN brous variety crocidolite are major constituents of some parts of the Brockman Iron Formation of the Hamersley Basin (Trendall and Blockley 1970 Klein and Gole 1981) and of iron-formations of the Kuruman and Griqualand Groups of the Transvaal Super-group of South Africa (Beukes 1973) These enormously large iron-formation sequences contain large amounts of riebeckite and its THORN brous variety crocidolite (blue asbestos) The Brockman Iron Formation of Western Australia may well contain one of the largest masses of alkali amphibole on Earth Trendall and Block-ley (1970) estimated the total reserves and resources of crocidolite

FIGURE 7 Compositional ranges of the talc-minnesotaite series = (Fe2+ Mg) Si4O10(OH)2 greenalite = (Fe6

2+Si4O10(OH)8 and stilpnomelane = K06(Mg Fe2+ Fe3+)6 Si8Al(OOH)27middot2-4H2O in late diagenetic to very-low-grade metamorphic iron-formations From Klein and Beukes (1993a)


(excluding non-THORN brous riebeckite) to be approximately 2 410 000 tons in the Brockman Iron Formation The last crocidolite mine in the Hamersley region was closed in 1966 on account of health effects (Klein 1993) However in most other late-diagenetic to low-grade metamorphic BIF occurrences (eg the Sokoman Iron Formation Labrador Canada Dimroth and Chauvel 1973 Zajac 1974 Klein 1974) it is a minor and sporadic component Crocidolite was mined in South Africa until 1997 and amosite (the THORN brous variety of grunerite also known as brown asbestos) was mined in South Africa until 1992

Ferri-annite is a type of mica that is commonly present as szlig aky or tabular grains and as massive aggregates near or within riebeckite-rich zones of banded iron-formation It commonly co-exists with hematite magnetite quartz ankerite stilpnomelane and riebeckite It was THORN rst described by Miyano (1982) from the Dales Gorge Member of the Hamersley Group A mineralogic characterization is given by Miyano and Miyano (1982)

Of the carbonates in unmetamorphosed iron-formation sid-erite and members of the dolomite-ankerite series are probably most common with calcite a lesser constituent Carbonates may make up a very large percentage (50 vol or more) of carbonate-rich (mainly siderite) iron-formation but they may also be major constituents of silicate-rich and oxide-rich iron-formation A very well-banded example of siderite-rich BIF is shown in Figure 6f

PHYSICAL AND CHEMICAL CONDITIONS OF IRON- FORMATION DIAGENESIS AND VERY LOW-GRADE

METAMORPHISM

The above mineralogical and petrological discussion of diage-netic and very low-grade metamorphic BIF assemblages leads to the question of what are the most likely precursor materials that were the original sedimentary products These were probably hydrous Fe-silicate gels of greenalite-type composition hydrous Na- K- and Al-containing gels approximating stilpnomelane compositions SiO2 gels Fe(OH)2 and Fe(OH)3 precipitates and very THORN ne-grained carbonate oozes of variable composition The possible products of sedimentation are listed in Table 2 During diagenesis H2O is lost from the more open gel-type compositions due to compaction Fur-thermore the amorphous silicate gels probably become somewhat less disordered structures This leads to the possibility of calculated stability diagrams for primary andor diagenetic Fe-rich phases at low pressure and temperature It is very probable that Fe(OH)3 was the precursor to hematite (Fe2O3) The precursor to magnetite (Fe3O4) is less well deTHORN ned it may have been a mixture of Fe(OH)2 and Fe(OH)3 a hydro-magnetite (Fe3O4middotnH2O) or perhaps even a very magnetite-like material The choice of precursor material will alter often signiTHORN cantly its Eh-pH stability THORN eld (Klein and Bricker 1977) Figures 8a and 8b show Eh-pH diagrams for several of the common mineral assemblages in BIF at 25 degC and 1 atm pressure superimposed on the diagram are calculated values of PO2

(Walker et al 1983) In Figures 8a and 8b Fe3O4 was chosen to represent the magnetite THORN eld but Fe(OH)3 was chosen as the precursor to hematite This choice of oxides allows for the Eh-pH delineation of very common associations such as magnetite-siderite magnetite-greenalite and magnetite-stilpnomelane (but makes impossible the coexistence of hematite and greenalite for example a mineral pair that in fact is extremely rare or absent in BIFs) As seen in Figures 8a and 8b the hematite precursor THORN eld [ie that of Fe(OH)3] is

very small and exists only at relatively high PO2 values Figures

8c and 8d show the extent of only the stability THORN elds of hematite and magnetite and Fe(OH)3 and Fe(OH)2 (as possible precursor phases) respectively These two illustrations show hematite and Fe(OH)3THORN elds that are much larger than those shown in Figures 8a and 8b for Fe(OH)3 and they illustrate the stability of hematite or Fe(OH)3 to very low PO2

values This shows that hematite (or its precursor) can be formed over a wide range of PO2

conditions (see also Walker et al 1983) and that hematite (which is abundant in many Early Proterozoic iron-formations) can be precipitated under anoxic to highly reducing conditions

The magnetite THORN eld is stable only at substantially lower values of PO2

and the sulTHORN de-rich and carbonate-silicate-rich mineral as-semblages observed commonly in BIFs (as discussed in the prior section) reszlig ect even lower PO2

values In short Figure 8 indicates that under equilibrium conditions a very common range of BIF assemblages reszlig ects anoxic conditions of deposition This is in agreement with calculated values for O2 in the Precambrian atmo-sphere (Pavlov and Kasting 2002) prior to 23 Ga These authors conclude that the pre-23 Ga atmosphere was anoxic The low redox state of such common BIF assemblages is corroborated by the bulk chemistry of a very large number of iron-formation analyses [see Fig 21 as well as the accompanying discussion of low Fe3+(Fe2+ + Fe3+) values] Such highly reducing (anoxic) environmental condi-tions for iron-formation precipitation are to be expected because large amounts of Fe2+ must have been transported in solution Iron solubility is possible only under highly reducing conditions

One mineral that is relatively uncommon in most BIFs but which is a common constituent of iron-formations in the Hamer-sley Range of Western Australia and of the Kuruman-Griqualand iron-formation sequence in South Africa is riebeckite This Na-amphibole is part of diagenetic to very low-grade metamorphic assemblages The temperature of formation of riebeckite in such assemblages has been estimated to range from 100 to 150 degC (Miyano and Klein 1983a) Thermodynamic calculations of the stability THORN eld of riebeckite by these authors resulted in the log fO2

-pH stability diagram (for 130 degC) shown in Figure 9 for carbon-ate-containing assemblages which are common in the riebeckite assemblages of the Dales Gorge Member of the Hamersley Range This shows that riebeckite formation is a likely diagenetic process in which alkali-bearing solutions have interacted with the Fe-rich bulk chemistry of Fe-formations

METAMORPHISM OF IRON-FORMATIONS

Many iron-formation sequences especially those of Archean age are the metamorphic products of the minerals discussed in the

TABLE 2 Inferred initial compositions and their sedimentary products in primary iron-formation assemblages (from Klein 1974)

Colloidal SiO2 rarr chertDissolved SiO2 + Fe2+ (Mg) rarr amorphous Fe-Si-O-OH-gel darr

+ of greenalite type Locally Na+ K+ Al3+ rarr amorphous Fe-Si-O-OH(NaKAl) gel of stilpnomelane typeDissolved Fe3+ rarr Fe(OH)3 becomes hematiteDissolved Fe2+ rarr Fe(OH)2 + Fe(OH)3 or hydromagnetite (Fe3O4middotH2O) becomes magnetiteDissolved CO2 Fe2+ Mg Ca rarr very fi ne grained mixed carbonates

Chert may form from Na-silicate (Na Si7O13(OH)3-magadiite) during diagenesis (Eugster and Chou 1973)


prior section of this paper Greenalite and stilpnomelane give way to minnesotaite and at increasing metamorphic grades amphi-boles pyroxenes and fayalite become high-temperature reaction products A schematic diagram showing relative mineral stabilities

in BIF ranging from very low to the highest metamorphic grade is given in Figure 10

The most conspicuous mineral group in medium-grade meta-morphosed BIF is that of Fe-rich amphiboles of the cumming-

FIGURE 8 Eh-pH diagrams depicting the stability THORN elds of some minerals (or their precursors) in the system Fe-H2O-O2-SiO2-dissolved C at 25 degC and 1 atm total pressure Boundaries between aqueous species and solids at A[aqueous species] = 106 (a) With stilpnomelane-like composition with Fe2+Fe3+ = 21 (b) Stilpnomelane not considered (c) Stability THORN elds of hematite and magnetite in water (d) Stability THORN elds of ferrous and ferric hydroxides in water Both c and d after Garrels and Christ (1965) Contours show partial pressure of oxygen (from Klein and Bricker 1977 and Walker et al 1983b)


tonite-grunerite series as a result of the reaction between Fe-rich carbonates and quartz and as a reaction product of minnesotaite Such reactions are

7Ca(FeMg)(CO3)2 + 8SiO2 + H2O rarr ferrodolomite quartz

rarr (FeMg)7Si8O22(OH)2 + 7CaCO3 + 7CO2

grunerite calcite

8(FeMg)CO3 + 8SiO2 + H2O rarr (FeMg)7Si8O22(OH)2 + 7 CO2

siderite quartz grunerite

and

7Fe3Si4O10(OH)2 rarr 3Fe7Si8O22(OH)2 + 4SiO2 + 4H2Ominnesotaite grunerite

An example of the THORN rst reaction is illustrated in Figure 11a Figures 11c and 11d show a progressive increase in grunerite and decreasing carbonate content However quartz (or chert) Fe oxide iron-formation devoid of carbonates andor silicates will not develop any prograde silicates and will persist with the same assemblage throughout all metamorphic grades This is shown in Figure 11b where relict magnetite granules and quartz (recrystallized from original chert) show no reaction features at

staurolite-kyanite zone metamorphism Magnetite and hematite occur as medium- to coarse-grained well-crystallized grains in most metamorphic BIF assemblages These are the result most commonly of recrystallization of earlier THORN ner-grained precursor grains of magnetite and hematite composition respectively There is little to no evidence of a possible reaction of siderite and quartz to produce magnetite during prograde metamorphic conditions (Klein 1973 1978 1983) The non-reactive persistence of mag-netite hematite and much quartz (in quartz-oxide BIF) is shown by the solid lines for these minerals in Figure 10

The Fe-Mg clinoamphiboles are commonly intergrown with Ca-clinoamphiboles such as actinolite and hornblende (Klein 1968) Such coexistences are the result of the following reac-tion

14Ca(Mg05Fe05)(CO3)2 + 16SiO2 + 2H2O rarr ferrodolomite quartz

Ca2Mg5Si8O22(OH)2 + Fe7Si8O22(OH)2 + 14(Ca09Mg01)CO3 tremolite grunerite calcite

+ 14CO2

Figure 12 illustrates the compositional ranges and coexistences for members of the cummingtonite-grunerite and tremolite-ac-tinolite series in medium-grade metamorphic iron-formations Figure 13 shows the range of amphibole compositions as well as coexisting amphibole pairs in the oldest medium-grade meta-morphosed iron-formation from Isua West Greenland (of 38 Ga age) All these amphiboles are part of quartz-magnetite-grune-rite-actinolite(hornblende)-ripidolite assemblages with traces of carbonate (Dymek and Klein 1988)

Although Fe-rich amphiboles are the most common silicates

FIGURE 9 Riebeckite-siderite-magnetite-hematite stability relations in terms of logfO2

-pH at a total CO2 = 101(10264) at 130 degC In this THORN gure aNa+ is THORN xed at 102 to 104 and at atotal Fe at 104 The activity of carbonate is THORN xed on the basis of thermodynamic data for siderite taken from Robie et al (1978) The equivalent activity of total carbonate for the siderite data of Helgeson et al (1978) is shown in parentheses From Miyano and Klein (1983)

FIGURE 10 Relative stabilities of minerals in metamorphosed iron-formations as a function of metamorphic zones From Klein (1983)


FIGURE 11 Photomicrographs of grunerite-rich BIF assemblages (a) Incipient formation of THORN ne needles of grunerite around coarse patches of ferrodolomite (fd) and quartz (Q) Biotite zone metamorphism Labrador Trough Canada Plane polarized light (b) Relict granules composed of magnetite and lesser quartz in a quartz matrix Staurolite-kyanite zone metamorphism Labrador City area Newfoundland Plane-polarized light (c) Coarse-grained grunerite (g) schist with patches of ankerite (ank) Staurolite-kyanite zone metamorphism Labrador City area Newfoundland (d) Medium-grained grunerite (g) quartz (Q) schist No pre-existing carbonate remains Staurolite-kyanite zone metamorphism Labrador City area Newfoundland From Klein (1983)


in medium-grade metamorphosed BIF some garnet may be pres-ent as well Garnet (of almandine composition) is generally rare because of the inherently low Al2O3 content of the bulk chemistry of BIF (see Fig 21) Some andradite garnets (CaFe2

3+Si3O12) have been reported

The carbonates that were present in late-diagenetic to very low-grade metamorphic assemblages tend to persist through me-dium-grade metamorphic conditions even though carbonates are consumed in the production of metamorphic amphiboles as well as pyroxenes (see some of the above reactions)

Iron-formations that have undergone the highest metamorphic grade are characterized by essentially anhydrous assemblages in which variable amounts of ortho- and clinopyroxene predominate Fayalite may be present as well as carbonates and garnet and lesser

amounts of amphiboles Quartz magnetite andor hematite are still the major constituents of oxide-rich iron-formations Such high-grade assemblages are the result of metamorphic conditions that straddle the sillimanite isograd of pelitic rocks or that range from upper-amphibolite to granulite facies (see Fig 10) Pyroxenes are the result of the following types of decarbonation reactions

Ca(FeMg)(CO3)2 + 2SiO2 rarr Ca(FeMg)Si2O6 + 2CO2

ankerite quartz clinopyroxene

and

(FeMg)CO3 + SiO2 rarr (FeMg)SiO3 + CO2

siderite quartz orthopyroxene

Amphibole decomposition reactions are also responsible for pyroxene formation eg

Fe7Si8O22(OH)2 rarr 7FeSiO3 + SiO2 + H2Ogrunerite orthopyroxene

These high-grade rocks tend to show equigranular textures in which it is generally impossible to distinguish relict minerals or textures Photomicrographs of pyroxene- and fayalite-containing BIF assemblages are shown in Figure 14 The range of pyroxene compositions and pyroxene pairs are shown in Figure 15

Fayalite is a major constituent of parts of the Biwabik and Gunszlig int Iron Formations that have been contact metamorphosed by the Duluth Gabbro Complex (Bonnichsen 1969) Regionally metamorphosed Archean iron-formations in the Yilgarn Block of Western Australia also contain fayalite (Gole and Klein 1981b) Fayalite-pyroxene (with lesser grunerite) compositional ranges are illustrated in Figure 16

FIGURE 12 Compilation of compositional ranges and major-element fractionation data for amphiboles from several medium-grade metamorphic iron-formations (a b and c) are for Proterozoic iron-formations (d) is for Archean From Klein (1983 and 1964)

FIGURE 13 Compositions of all analyzed amphiboles in iron-formation assemblages from Isua West Greenland (from Dymek and Klein 1988) (a) Overall ranges of composition in actinolite-hornblende and cummingtonite-grunerite (b) Compositions of all amphibole pairs


FIGURE 14 Photomicrographs of some high-grade metamorphic BIF assemblages (a) Granular iron-formation consisting of hypersthene (hyp)-ferrosalite (fsl) almandine (alm) quartz (qtz) and minor hornblende (hbl) from Archean BIF in the Tobacco Root Mountains Southwestern Montana (b) Coexisting ferroaugite (cpx) eulite (opx) grunerite (grun) quartz (qtz) and magnetite (mag) (c) Fayalite (fay) eulite (opx) ferroaugite (cpx) grunerite (grun) quartz (qtz) Smooth curved grain boundaries occur between all minerals suggestive of equilibrium crystallization (d) Fayalite (fay) grunerite (grun) quartz (qtz) assemblage in which grunerite mantles fayalite grains such that quartz and fayalite do not occur in contact Photographs b c and d from Archean iron-formations in the Yilgarn Block Western Australia (Klein 1983)

Carbonates may still be present in the highest metamorphic grade assemblages It appears that members of the dolomite-ankerite series as well as calcite may be reasonably abundant species throughout the complete metamorphic range discussed

here but siderite becomes less abundant at the highest grades (Klein 1983) This may indicate that the FeCO3 component is most involved in the production of metamorphic silicates Some almandine-rich garnets may also be present in small amounts


Among the few Al-containing silicates in BIF stilpnomelane is stable to higher metamorphic grades than silicates such as greenalite chamosite and minnesotaite However it is absent from higher-grade rocks in which Fe-rich amphiboles pyroxenes andor olivine predominate Stilpnomelane therefore is indicative of low- to medium-grade metamorphism of iron-formation The general stability THORN eld of stilpnomelane in the system K2O-FeO-Al2O3-SiO2-H2O has been evaluated by Miyano and Klein (1989) Their calculated stability diagram for stilpnomelane is given in Figure 18 In this THORN gure curve 1 shows the upper stability limit of minnesotaite reacting to form grunerite The upper stability limit for stilpnomelane in iron-formations is about 430470 degC and 56 kilobars The reactions in this diagram that show the presence of zussmanite (zus) apply to Al-bearing silicate assemblages that have been reported from blueschist-facies metamorphism

At medium-grade metamorphism of BIF Fe-rich amphiboles (cummingtonite-grunerite series) become very abundant and at even higher grade pyroxenes and subsequently olivine (fayalite) occurs A calculated P-T stability THORN eld for grunerite orthopyrox-ene (XFe

opx = 08) olivine (fayalite) and siderite in the presence of quartz is given in Figure 19

An overall evaluation of these silicate stability THORN elds as de-duced from the temperature-pressure estimates of various petro-logic studies of metamorphosed BIFs is given in Figure 20

AVERAGE MAJOR ELEMENT CHEMISTRY OF BIFIron-formations are unusual chemical sediments because of

their high contents of total Fe (ranging from about 20 to 40 wt see Fig 21) and SiO2 (ranging from 34 to 56 wt Fig 21) and

FIGURE 15 Compilation of compositional ranges and major-element fractionation data of pyroxenes from various high-grade metamorphic iron-formations From Klein (1983)

FIGURE 16 Compositional ranges of orthopyroxene clinopyroxene and fayalite (and lesser grunerite) in some high-grade metamorphic iron-formation occurrences (a) Assemblages in the highest grade metamorphic zone of the Biwabik Iron Formation (b) Mineral compositions and assemblages from high-grade iron-formations in the Yilgarn Block of Western Australia From Klein (1983)

in the highest grade assemblages All of the above discussed metamorphic reactions are essentially isochemical except for prevalent dehydration and decarbonation

THEORETICAL EVALUATIONS OF SOME OF THE CONDITIONS OF PROGRADE METAMORPHISM OF IRON-

FORMATION

Some of the most common primary mineral constituents of silicate-rich iron-formation are greenalite stilpnomelane and carbonates (in addition to chert and most commonly magnetite) These silicates and carbonates are almost always overprinted by minnesotaite in very low-grade metamorphic assemblages (see Fig 6e) This common occurrence of minnesotaite having formed at the expense of earlier greenalite stilpnomelane or Fe-carbonates is explained by the minnesotaite stability THORN eld as shown in Figure 17 The minnesotaite THORN eld completely eliminates that of greenalite (of end-member composition) and reduces the stilpnomelane and siderite stability THORN elds as well At increasing temperatures the minnesotaite THORN eld would expand further into the stilpnomelane THORN eld As such greenalite and minnesotaite are index minerals in the lowest metamorphic range of iron-forma-tions They react with the assemblage in which they occur to from grunerite at higher grade


much lesser amounts of CaO MgO MnO Al2O3 Na2O K2O and P2O5 The CaO MgO and MnO contents reszlig ect the common pres-ence of carbonates in BIF (siderite ankerite minor calcite) and Al2O3 Na2O and K2O are housed mainly in silicates (riebeckite greenalite and stilpnomelane) The CaO and MgO values range from 175 to 90 wt and 120 to 67 wt respectively MnO con-tent is generally very small ranging from 01 to 115 wt Al2O3 shows a range from 009 to 18 wt A larger value of 241 wt is shown in Figure 21 for S bands (macrobands interbedded with the various BIFs) in the Hamersley Group of Western Australia These S bands consist mineralogically mainly of sheet silicates (stilpnomelane biotite and chlorite) and iron-rich carbonates (such as siderite and ankerite) with lesser quartz and feldspar (Trendall and Blockley 1970) The Na2O and K2O contents are both low Na2O ranges from 0 to 08 wt K2O from 0 to 115 wt These ranges are shown in Figure 21 which is based on a large analytical database reported in Klein and Beukes (1992) with additional data for the Archean Nova Lima Group in Brazil (C Klein and EA Ladeira unpublished manuscript) and the Urucum BIFs also Brazil (Klein and Ladeira 2004) Figure 21 shows clearly that there is a strong similarity to all of the chemical averages except for the curves shown by both the Rapitan and the Urucum BIFs In the selection of the analyses that are part of this THORN gure every effort was made to exclude iron-formation samples that had undergone any type of secondary alteration such as oxidation or leaching thus excluding any materials that might be considered ore or in the process of becoming ore It is

FIGURE 18 Phase relations of Al-bearing silicates in low-grade metamorphosed Fe-rich rocks The shaded THORN eld represents the overall stability of stilpnomelane Abbreviations used are as follows Alm = almandine Bio = biotite Chl = chlorite Gru = grunerite Stil = stilpnomelane and Zus = zussmanite Curve 1 is the upper stability limit of minnesotaite reacting to form grunerite From Miyano and Klein (1989)

FIGURE 19 Calculated P-T diagram showing phase relations of orthopyroxene olivine grunerite quartz and siderite at given XFe

opx and XCO2 and Ps = PH2O + PCO2 From Miyano and Klein (1986)

FIGURE 17 Eh-pH diagram depicting the stability relations among phases in the system Fe-H2O-O2-C at 25 degC and one atmosphere total pressure Boundaries between aqueous species and solids at A[aqueous spaces] = 106 This diagram illustrates that at 25 degC minnesotaite (the shaded THORN eld) is the stable and greenalite the metastable phase From Klein and Bricker (1977) Compare with Figure 8a


FIGURE 20 Evaluations of the stability fields of minnesotaite and grunerite from various petrologically calculated P-T ranges Curve a represents the apparent upper stability limit of grunerite and curve c is the adjusted upper stability limit for minnesotaite From Klein (1983)

FIGURE 21 Plot of the major chemical oxide components of iron-formations with the original analytic results recalculated to 100 on an H2O-CO2-free basis The shaded area enclosed by thin dashed lines brackets the overall range of all average values (except for the Rapitan and Urucum iron-formations) as based on at least 215 reported whole-rock analyses on bulk samples of unaltered and unleached iron-formations (all analytic data except for the Nova Lima Group and Urucum are from Klein and Beukes 1992 Nova Lima Group data from C Klein and EA Ladeira (unpublished manuscript) Urucum data from Klein and Ladeira 2004) The number of samples represented by speciTHORN c points on the graph is given as n Note that data for the various components are presented relative to three different (vertical) weight percentage scales

for this reason that even the least altered BIF types of eg the Carajaacutes region Brazil (Klein and Ladeira 2002) are not included here all BIF materials from that region have undergone secondary oxidation and considerable supergene enrichment

The total Fe content of samples in Figure 21 ranges from 20 to a maximum of 40 wt which is much less than that of commercial iron ores The ranges of Fe2O3 and FeO shown

graphically in Figure 21 can be expressed as Fe3+(Fe2+ + Fe3+) as obtained from the original analytical data referenced in the legend of the THORN gure This range is from 005 (for the Kuruman Iron-Formation rich in siderite photograph given in Fig 6f) to 058 (for metamorphosed Archean iron-formations in Montana) The only two iron-formations with much larger Fe3+(Fe2+ + Fe3+) values are the Rapitan and Urucum occurrences both with values


of 097 It is instructive to compare the 005 to 058 range with the values for two of the common iron oxides magnetite [Fe3+(Fe2++ Fe3+) = 067] and hematite [Fe3+ (Fe2++ Fe3+) = 1] These two numbers illustrate that the Fe in all of the iron-formations shown in Figure 21 (except for the Rapitan and Urucum occur-rences) is in an average oxidation state between that of wuumlstite (FeO) and that of magnetite (Fe3O4) reszlig ecting the very common association of magnetite with Fe2+-containing minerals such as carbonates (siderite and ankerite) silicates (such as greenalite in essentially unmetamorphosed iron-formation minnesotaite members of the cummingtonite-grunerite series and members of the orthopyroxene series in metamorphosed assemblages) and locally pyrite These low Fe3+(Fe2+ + Fe3+) ratios are corroborated by the low redox state of common BIF assemblages as depicted in Figure 8 These data suggest that any models for iron-forma-tion precipitation require much less oxygen input than might be needed if iron-formations are assumed to consists mainly of hematite Fe2O3 (and quartz) In this regard however even Fe3+ oxide or Fe3+ hydroxide precipitates can originate under anoxic to very highly reducing environments as seen in Figures 8c and 8d The only iron-formations that are radically different from all of the others are the Rapitan and Urucum occurrences as shown by the curves in Figure 21 It should be noted here that Fe-rich sedimentary rocks that may be associated with volcanic-hosted base metal deposits (these are referred to as iron-formations Peter 2003) and which may contain considerable clastic detrital components are not the same as the BIFs discussed here Their bulk chemistry and mineralogy are distinctly different from that of the iron-formations in this paper Their bulk chemistry shows large amounts of Al2O3 (averaging about 67 wt with maxima of up to 208 167 and 204 wt for different Fe-rich lithologies Peter 2003) as well as highly elevated trace element levels for metals such as Co Cr Cu Pb and Zn as well as S All of these trace elements are extremely low in the BIFs discussed here (see Klein and Beukes 1992 their table 422) The overall mineralogy of these sulTHORN de-rich and Fe-rich rocks is very different as well from that presented in a prior section of this paper Peter (2003) describes these iron-rich rocks as exhalites that were precipitated in very close proximity to submarine hydrothermal vents

RARE EARTH ELEMENT CHEMISTRY OF SELECTED IRON FORMATIONS

The question of what was the ultimate source of the Fe as well as the silica in Precambrian iron-formations was debated for several decades prior to about 1973 In that year Holland (1973) wrote an article that questioned the possibility of forming Fe deposits as a result of the derivation of Fe from weathering and transport by rivers into a sedimentary basin He furthermore questioned the possibility of the derivation of Fe from volcanic sources Instead he postulated that Fe from deep ocean water would be the most likely source for banded iron-formations worldwide Since 1973 there have been extensive studies of rare earth elements (REE) in BIFs (for an early overview see Fryer 1983) that have elucidated the source of iron (and silica) as hy-drothermal input into the deep ocean In such REE studies of BIF it was concluded that hydrothermal waters (containing abundant Fe and SiO2) are released to the deep sea in a density-stratiTHORN ed ocean This hydrothermal signature was transmitted into shallower

ocean water by upwelling The discussion of REE signatures that follows is based on this model It should be noted however that Isley (1995) has proposed an alternate mechanism of Fe delivery to BIFs through hydrothermal plumes that had a dominant source in the upper water column

Rare earth proTHORN les determined by Instrumental Neutron Acti-vation Analysis (INAA) normalized to the North American Shale Composite (NASC Gromet et al 1984) for BIFs over a range of ages are given in Figure 22 The THORN rst diagram (Fig 22a) that for Isua BIF shows a clearly deTHORN ned positive Eu anomaly on an REE proTHORN le with an overall background slope that reszlig ects depletion of light REE and enrichment of heavy REE (The apparent negative Ce anomalies that appear in many of the patterns in Fig 22 are not interpreted here as true negative anomalies because of the large variation in analytical accuracy in Ce determinations by INAA in Fe-rich samples see Bau and Moumlller 1993 The apparent (false) negative Ce anomalies may be the result of positive La anomalies as obtained for REE by Inductively Coupled Plasma Mass Spectrometry (ICP-MS) see Yamaguchi et al 2000 Real negative Ce anomalies are interpreted as reszlig ecting an oxidizing environment which is incompatible with all the mineralogical and bulk chemical data presented in this BIF overview) The Isua REE proTHORN le can be reproduced by mixing the REE signature of modern high-temperature hydrothermal solutions with North At-lantic seawater using a seawater to hydrothermal ratio of 1001 This was done by Dymek and Klein (1988) and their resultant 1001 dashed mixing curve is shown in Figure 22a as well This led to the conclusion that the REE pattern of the Isua BIFs is the result of chemical precipitation from solutions that represent mixtures of seawater and hydrothermal szlig uid Fryer et al (1979) had suggested that Archean oceans may have had their REE (and other trace element) characteristics controlled dominantly if not exclusively by hydrothermal input Figure 22 shows REE proTHORN les of other BIFs as well in order of decreasing age All of the proTHORN les shown in Figures 22a through 22c show distinct similarities and pronounced positive Eu anomalies on similarly sloping back-grounds There appears to be a fairly consistent decrease in the overall concentration of REE as well as in the size of the positive Eu anomaly with decreasing BIF age except for two of the upper proTHORN les shown in Figure 22c This decrease suggests a declining hydrothermal input into a deep ocean basin from Early Archean to Early Proterozoic time Bau and Moumlller (1993) concluded that this decrease in the positive Eu anomaly is the result of the lowering of the temperature of the hydrothermal solutions as a reszlig ection of decreasing upper mantle temperature

The REE proTHORN les in Figures 22f and 22g for the Neoprotero-zoic iron-formations of Urucum and Rapitan are distinctly dif-ferent from the ones shown above In both THORN gures the BIFs show very similar and distinctive patterns All samples are depleted in light REE and the majority of samples (except for two proTHORN les in Fig 22g) lack a clear positive Eu anomaly In general these plots are similar to the REE patterns for shales All of the proTHORN les in Figures 22f and 22g are similar to the REE proTHORN le of modern ocean water (see dashed proTHORN le in Fig 22g) From this it is concluded that the source of the metals (Fe Mn and Si) is most likely of deep hydrothermal origin (as it was in older BIF sequences) but that the hydrothermal component appears to have been much more dilute than in the older Precambrian BIF sequences


ORGANIC CARBON CONTENT CAR-BON SULFUR AND IRON ISOTOPES

Extensive analytical data on the or-ganic carbon content of iron-formations are available from studies of the very low-grade metamorphosed iron-forma-tions of the Transvaal Supergroup South Africa (Klein and Beukes 1989 Beukes and Klein 1990) and the Dales Gorge Member of the Brockman Iron Formation (Kaufman et al 1990) Siderite-rich BIFs from the Kuruman Iron Formation (Klein and Beukes 1989) show organic carbon values that range from 0047 to 020 wt with an average of 008 wt Mag-netite-rich iron-formations from the same sequence show values of organic carbon

FIGURE 22 Comparison of North American shale composite (NASC)normalized REE patterns of Precambrian iron-formations of various ages The speciTHORN c iron-formations and their ages are given along the REE proTHORN les (a) The Isua BIF illustration also shows a seawater to hydrothermal szlig uid mixing curve of 1001 where a multiplication factor of 106 was used for the original REE values of seawater (see Dymek and Klein 1988) (g) The REE proTHORN les of the Rapitan iron-formation are accompanied by an REE curve for modern seawater at 100 m multiplied by 106 (ElderTHORN eld and Greaves 1982 Klein and Beukes 1993b) Other references are (b) C Klein and EA Ladeira (unpublished manuscript) (c) Klein and Ladeira 2000 (d) Klein and Beukes 1989 (e) Beukes and Klein 1990 (f) Klein and Ladeira 2004


ranging from 0008 to 0017 wt with an average of 0012 wt In contrast closely associated limestones contain 0886 wt do-lomites 0523 wt shales 391 wt ferruginous shale 455 wt pyrite-rich shale 267 wt and carbonaceous shale 411 wt organic carbon (see Fig 23) This THORN gure illustrates the organic carbon contents for a facies transition from interbedded carbon-ate-shale (deposited on the Campbellrand carbonate platform) to banded iron-formation (deposited in much deeper waters) The limestone and dolomite lithologies contain macroscopic crystalgal laminae as well as intraclastic textures The Al2O3 values in this THORN gure are highest for the shales (averaging 955 wt) with the two iron-formation types having average Al2O3 values of 0099 wt (siderite-rich) and 0066 wt (magnetite-rich)

Beukes et al (1990) reported similarly low organic carbon values for siderite-rich iron-formation from the Transvaal Super-group with much higher organic carbon contents for associated limestone and shale Oxide-rich BIF values range from 0008 to 0017 wt and siderite-rich BIFs from 0041 to 0203 wt organic carbon Associated limestones (and dolomites) and shales show organic carbon ranges from 0188 to 22 and 279 to 636 wt respectively

Samples from the Dales Gorge Member of the Hamersley Range (Kaufman et al 1990) consisting mainly of quartz-mag-netite-hematite-siderite-greenalite-stilpnomelane show a range of organic carbon values from 0032 to 0108 wt Similarly low values of organic carbon were reported for the essentially unmeta-morphosed Neoproterozoic Rapitan and Urucum iron-formations Organic carbon values in the Rapitan sequence range from 0131 to 0165 wt (Klein and Beukes 1993b) and for the Urucum deposit from 000 to 008 wt (Klein and Ladeira 2004) These studies are ample proof of the essential lack of organic carbon in well-preserved very low-grade metamorphosed iron-forma-tion sequences Such overall low organic carbon values in BIFs suggest that microbial activity has not played a signiTHORN cant part in the depositional environment of iron-formations (see Beukes et al 1990) The almost total lack of bona THORN de microfossils in BIF sequences supports this conclusion (Walter and Hoffman 1983) The only well-documented occurrence of THORN lamentous microbial sheaths in intermeshed mats is in a chert-ferroan do-lomite assemblage in the transition from carbonate lithologies to iron-formation in the Transvaal Supergroup South Africa (Klein et al 1987) Knoll and Simonson (1980) recognized two assemblages of benthic microfossils in cherts of the Sokoman Iron Formation similar to those of the Gunszlig int Formation also in cherts (Barghoorn and Tyler 1965) Walter et al (1976) re-ported on THORN lamentous microfossils in the Frere Formation of the Nabberu Basin Western Australia None of these occurrences however is part of typical Fe-rich assemblages as such there appears to be no genetic relationship between Fe-rich minerals and microfossils Furthermore Walter and Hoffman (1983) noted that neither stromatolites nor convincing microfossils are known from Archean iron-formations

Carbon isotopic compositions of carbonates are available for all of the above essentially unmetamorphosed to very low-grade metamorphic BIFs as well Such data are tabulated in Table 3 The THORN rst entry in that table shows a δ13C range from 50 to 137 for BIFs from South Africa Of this range δ13C for well-banded siderite-rich BIF (a photomicrograph of which is shown in Fig

6f) is from 30 to 79 (see Fig 24) The more depleted val-ues from 50 to 97 are for carbonates from oxide-rich BIF (magnetite andor hematite-quartz) and the most depleted range from 90 to 137 is for minnesotaite-rich BIF that was locally metamorphosed by a diabase sill In contrast the δ13C range for carbonates in closely associated limestone and dolomite litholo-gies is given as 01 to 28 Beukes et al (1990) reasoned that the δ13C values for the limestones reszlig ect the total dissolved carbon isotope composition of the basin water and that the on average 398 depleted δ13C values for the unmetamorphosed and very well-banded siderite-rich BIFs represent a primary car-bon isotope signature of the deeper basinal waters in which the BIFs were precipitated They concluded that both the limestones and the siderite-rich BIFs are primary precipitates from a basin water yet they display different isotopic compositions with the siderite depleted by approximately 4 in 13C over calcite This THORN nding then implies a water column stratiTHORN ed with regard to isotopic composition of total dissolved CO2 (this will be addressed further in a subsequent section on basin evaluation) Kaufman et al (1990) conTHORN rmed that primary siderite (δ13C ~ 5) was precipitated from an anoxic water column depleted in 13C This carbon isotopic depletion of the microbanded siderite BIF is at-tributed to its formation in an isotopically depleted water mass enriched in Fe and depleted in 13C The most likely sources for this are hydrothermal vents in the deep ocean as corroborated by the REE data (see Fig 22 and associated text)

The smaller negative δ13C ranges for carbonates in the Neo-proterozoic Rapitan and Urucum iron-formations (Table 3) reszlig ect

FIGURE 23 Correlation between Al2O3 and organic carbon content (in wt) for THORN ve different lithologies in the transition from limestone to iron-formation (from the Campbellrand carbonate sequence to the overlying Kuruman Iron Formation of the Transvaal Supergroup) from Klein and Beukes (1989) The limestone and dolomite lithologies contain macroscopic cryptalgal laminae and show intraclastic textures These two lithologies as well as the associated shales show high values of Al2O3 and organic carbon The two iron-formation types are low in both these components


FIGURE 24 Plot of organic C content vs carbon isotopic composition of carbonates in various lithofacies as identiTHORN ed in the legend This diagram depicts the same transition as shown in Figure 23 from limestone to iron-formation in the Transvaal Supergroup of South Africa From Beukes et al (1990)

their stratigraphic setting (with diamictites and dropstone layers) that resulted from continental glaciation (Kaufman and Knoll 1995 Des Marais 2001)

Sulfur isotope compositions are available from the sulTHORN de-containing Archean BIFs in the Nova Lima Group Brazil (C Klein and EA Ladeira unpublished manuscript) and the Kuru-man Iron-Formation South Africa (Beukes et al 1990) These δ34S ranges are as follows

Nova Lima Group Brazil 22 to 82 vs CDTKuruman Iron Formation 59 to 210 vs CDT

This variation is within the total spread of δ34S values reported for Precambrian sedimentary sulTHORN des reported by Schidlowski et al (1983) ranging in age from 38 to 10 Ga their published range of values is from about 11 to +30 δ34S values for Proterozoic

sedimentary sulTHORN des show an even larger spread from about 32 to +58 (Hayes et al 1992) These authors envisage that most of the sulTHORN des in Archean sedimentary strata formed directly or indirectly from sulfurous emanations that were abundant because of extremely high levels of igneous activity in the Archean The Proterozoic δ34S range of values may have resulted from reduction of sulfate in hydrothermal systems or from bacterial reduction of marine sulfate enriched in 34S (Hayes et al 1992)

Iron isotope studies on samples of the Early Proterozoic iron-formations of the Transvaal Supergroup South Africa (Johnson et al 2003) show a range in 56Fe54Fe ratios from 25 to +10 These authors concluded that is it not yet clear if low δ56Fe values of magnetite and Fe-rich carbonates reszlig ect biological processing or simply inorganic precipitation

BASINS OF IRON-FORMATION DEPOSITION

Most Archean BIF sequences are part of greenstone belt se-quences in major old cratons (see Fig 4 for some stratigraphic sequences) The iron-formations in these greenstone belts are generally highly deformed metamorphosed and dismembered This makes it essentially impossible to reconstruct the deposi-tional basin setting for such Archean occurrences That said the prevalence of THORN nely laminated even microbanded BIF sequences (see Figs 5a and 5b) throughout the Archean leads to the conclu-sion that all such occurrences were deposited in basins deeper than at least 200 m which is the minimum depth for modern storm wave base (Trendall 2002) The 200 m depth value should probably be considered as a minimum depth of deposition in light of the likely very delicate nature of many of the original gel-like Fe-rich precipitates (see Table 2) There is also the consistent absence in all iron-formation bulk compositions (see Fig 21) of an epiclastic (or detrital) component This is reszlig ected in the very low Al-content of all BIFs as shown in Figure 21 with a range of Al2O3 content from 009 to only 18 wt This lack of detrital inszlig ux suggests BIF deposition beyond the range of effective off-shore epiclastic inszlig ux (Trendall 2002) which goes hand in hand with the conclusion that most BIFs are the result of deep water deposition In contrast to the Archean BIFs the Palaeoproterozoic BIF sequences of the Hamersley Range Western Australia the Kaapvaal Craton of South Africa and of the Labrador Trough Canada occur in well-preserved little-metamorphosed sequences rather than in greenstone belts Both the BIFs of the Hamersley Range as well as of the Transvaal Supergroup in South Africa exhibit well-developed microbanding An exhaustive model for the basinal setting the banding of BIF and chemical aspects of its deposition in the Hamersley Range of Western Australia was given by Morris (1993) Current-generated structures such as cross-bedding and ripple marks are virtually absent from the Hamersley and Transvaal sequences However they are prevalent in the granular iron-formations of the Labrador Trough and the Nabberu Basin of Western Australia Such BIFs suggest shallow-water high-energy environments

The most thorough evaluation of the basinal setting of an iron-formation sequence has been made by Klein and Beukes (1989) and Beukes et al (1990) as based on their study of the transition from interbedded carbonate-shale to banded iron-formation in the Campbellrand carbonate sequence to the overlying Kuruman Iron Formation of the Transvaal Supergroup of South Africa The

TABLE 3 Carbon isotope variations in carbonates from selected essentially unmetamorphosed iron-formations

BIF Sequence δ13C (permil) Reference

Campbellrand ndash ndash50 to ndash137 Beukes et al (1990)Kuruman Iron Formationtransition South Africa

Limestone and ndash01 to ndash28dolomites at same above location

Kuruman ndash Griqualand ndash545 to ndash842 Beukes and Klein (1990)iron-formation transition South Africa

Brockman Iron Formation ndash692 to ndash796 Kaufman et al (1990)Dales Gorge Member Paraburdoo Western Australia

Rapitan Canada ndash083 to ndash337 Klein and Beukes (1993b)

Urucum Brazil ndash442 to ndash702 Klein and Ladeira (2004)


granular iron-formations such as those of the Lake Superior region (Goodwin 1956 Simonson 1985) the Sokoman Iron Formation in the Labrador Trough (Dimroth and Chauvel 1973 Klein and Fink 1976) and of the Nabberu Basin of Western Australia (Goode et al 1983) in platformal (shoal) areas A mechanism for the trans-port of Fe to surface ocean water must have developed This may have been the result of a declining chemical density stratiTHORN cation due to lesser hydrothermal input as indicated by REE results (Fig 22) Following this period (circa 19 Ga) the oceans may have become completely mixed somewhat less reducing and depleted in Fe as indicated by the absence of iron-formations between about 18 Ga and about 08 (or 07) Ga (see Fig 2) This overall change in the paleoceanographic deposition of BIFs is shown in Figure 27

In Neoproterozoic time however iron-formations are again part of the geologic record These iron-formations are intimately associated with glaciomarine deposits and may be interbedded with Mn deposits as well (Klein and Beukes 1993b Klein and Ladeira 2004) In both the Rapitan and Urucum iron-formation sequences the only iron-oxide is hematite The sedimentologic setting of the Rapitan iron-formation is among a thick sequence of glaciogenic materials (diamictites) the iron-formation also contains dropstones and faceted pebbles This iron-formation se-quence appears to have been deposited during a major transgres-sive event with a rapid rate of sea-level rise during an interglacial period The Urucum Brazil BIF (and Mn-formations) sequence contains layers of abundant dropstones

The appearance of these Neoproterozoic BIFs reszlig ects low oxygen (anoxic to highly reducing) conditions that were the re-sults of stagnation in the oceans beneath a near-global ice cover as

FIGURE 25 Schematic depositional environment for iron-formation and that of associated lithofacies in a marine system with a stratiTHORN ed water column in (a) a regressive stage and (b) a transgressive stage In a the photic zone reaches the szlig oor of the deep shelf allowing for cryptalgalaminated limestone deposition In b the photic zone is considerably above the szlig oor of the deep shelf causing the deposition of various iron-formation types and chert The thick arrows labeled C (carbon) in a represent high carbon productivity and supply and the narrow arrows in b represent less carbon productivity and supply From Klein and Beukes (1989)

oldest rocks are limestones and lesser dolomite with cryptalgal laminae and intraclastic textures The carbonates and shales are overlain by meso- and microbanded siderite-chert iron-formation (illustrated in Fig 6f) that grades upward into magnetite- chert- and carbonate-rich BIF The shales are the most aluminous (aver-age 955 wt Al2O see Fig 23) and they also have the highest organic carbon contents (average 391 wt organic C see also Fig 23) The other lithologies have intermediate values of Al2O3 and organic C between the two extremes of shales on the one hand and BIF on the other The two iron-formation types have average Al2O3 values of 0099 wt (siderite-rich) and 0066 wt (magnetite-rich) and corresponding averages of organic carbon of 0080 wt (siderite-rich) and 0012 wt (magnetite-rich) The siderite-rich BIF was concluded to be a primary precipitate On the basis of these geochemical data and a reconstruction of the depositional basin for the carbonate-shale to iron-formation transition Klein and Beukes (1989) concluded that the limestone-dolomite-shale lithologies originated in a water column quite distinct from that in which the iron-formation was precipitated They proposed a model with a stratiTHORN ed water column in which the surface waters (during a regressive stage in the depositional basin) were the site of much organic carbon productivity and the locus of cryptalgal limestones The deeper waters (during a transgressive stage of the basin with the Kaapvaal Craton more deeply sub-merged) was proposed as the site for iron-formation deposition These deeper waters were depleted in organic C and enriched in dissolved FeO (from a hydrothermal deep ocean source) relative to the shallower water mass This model is illustrated in Figure 25 This depositional (basin) model was further enhanced by the carbon isotopic studies of the same above-discussed lithologies As noted in Table 3 the primary siderites (in siderite-rich BIF) and the primary limestone (micritic precipitates) differ significantly in their 13C composition with the siderites depleted in 13C by 46 on average relative to the calc-microsparite This finding made Beukes et al (1990) conclude that the siderites were precipitated from water with a dissolved inorganic carbon depleted in 13C relative to that from which the limestones precipitated This implies an ocean system stratiTHORN ed with regard to total carbonate with the deeper water from which the siderite-rich iron-formations formed depleted in 13C This further modiTHORN cation of the ear-lier model is illustrated in Figure 26 with the iron-formations deposited in areas of very low organic matter supply

In middle Early Proterozoic time the above stratified ocean system may have started to break down as concluded from the de-velopment of abundant oolitic and


FIGURE 26 Schematic depositional environment for iron-formation and associated lithofacies in a marine system with a stratiTHORN ed water column The thick arrows labeled C (organic carbon) represent high productivity (as in Fig 25) whereas the thinner arrows represent lesser carbon productivity and supply Banded siderite iron-formation precipitates along the chemocline where there is some organic carbon supply Magnetite- and hematite-rich iron-formation precipitate where the organic matter supply is low and some oxygen is available The well-mixed and near-shore water masses where limestone deposition took place had a 13C composition similar to that of present-day oceans close to 0 The deeper waters where siderite-rich iron-formation was a primary precipitate (with δ13C on average negative at 53) as a result of signiTHORN cant hydrothermal input is shown as stratiTHORN ed with δ13C (carb) asymp 5 (from Beukes et al 1990)

FIGURE 27 Paleoceanographic models for iron-formation deposition from the Archean to the Middle Proterozoic (a) Archean to Early Proterozoic stratiTHORN ed ocean system with predominantly deep water deposition of microbanded iron-formation (b) Middle Early Proterozoic Breakdown of the stratiTHORN ed ocean system and deposition of oolitic iron-formation in shoal areas (c) Middle Proterozoic Fe-depleted well mixed ocean system that is somewhat oxygenated but depleted in Fe From Beukes and Klein (1992)

was suggested by Kirschvink (1992) and referred to as Snowball Earth The Snowball Earth hypothesis of Kirschvink provides a convincing explanation for many features in the Neoproterozoic (see Hoffman and Schrag 2002 for details) although aspects of the model are still unresolved (Schmidt and Williams 1995) The ice cover allowed for the buildup of dissolved Fe (and Mn as is the case at Urucum) during a glacial period and deposition of these metals during interglacial periods when the ocean and atmosphere were in direct contact At that time only very small amounts of oxygen were necessary for the precipitation of the precursors to hematite (and various manganese oxides at Urucum) as shown in Figures 8c and 8d A paleoceanographic model for BIF deposition in the Neoproterozoic is shown in Figure 28

CONCLUDING REMARKS

Banded iron-formations are uniquely Precambrian chemical precipitates They are present in the Archean cratons of many continents ranging in age from 38 to 25 Ga Most of these Archean BIFs occur in greenstone belts are metamorphosed tectonically deformed and dismembered Reconstruction of their original depositional (basinal) settings is therefore very difTHORN cult However because almost all these Archean BIF occurrences show THORN ne lamination andor microbanding it appears that they were precipitated in a minimum depth of below 200 m which is the wave base for modern storms The much better preserved and only very slightly metamorphosed enormous BIFs of the Hamer-sley Basin of Western Australia (ranging in age from about 26 to 245 Ga) and the somewhat younger nearly identical BIFs of the Transvaal Supergroup of South Africa allow for better assessment of the depositional setting of these iron-formations Both show

well-developed microbanding that (in the case of the BIFs of the Hamersley Basin) is very well documented and is generally interpreted as varves or annual layers

The well-known stratigraphy of the transition from carbonate-shale to iron-formation deposition in the Transvaal Supergroup South Africa has allowed for further evaluation of the deep ocean basin setting of these Late Archean (Hamersley) to Early Pro-terozoic (Kaapvaal Craton) iron-formation sequences Not only are the iron-formations deep ocean basin deposits (microbanded oxide-chert occurrences and well-preserved laminations in sili-cate-rich and carbonate-rich BIFs that originated from original delicate gel and ooze precipitates as well as a general lack of detrital input) but they appear to have formed in a stratiTHORN ed ocean system The deep ocean was the locus of BIF precipitation (with essentially no organic C) and δ13C values in well-banded siderite BIF of asymp 5 and the shallow water the locus of carbonate-shale lithologies (with abundant organic C) and δ13C in carbonates of asymp 0

The younger Early Proterozoic BIFs of the Late Superior region USA the Labrador Trough Canada and the Nabberu Basin Western Australia are commonly granular and oolitic in nature with mesobanding but not microbanding This represents a much shallower deposition of iron-formation in essentially surface waters (in platformal shoal regions)

The Neoproterozoic BIFs of the Rapitan sequence Canada


and the Urucum region Brazil are the result of Fe precipitation in ice-covered basins locally causing stagnant anoxic conditions

The source of the major component of BIFs (as deduced from REE data) such as Si Fe and Mn appears to be hydrothermal input into deep ocean basins during Archean to Early Protero-zoic time and further upwelling of such hydrothermal solutions to shallower ocean basin settings in Early Proterozoic time (for granular BIF formation) It is likely that the hydrothermal com-ponent introduced into the ocean system may have diminished with time (as shown in Fig 22) or that the temperature of the hydrothermal input decreased over time The REE proTHORN les of Neoproterozoic BIFs suggest relatively little hydrothermal input and possibly dissolution of material from basin szlig oors during es-sentially anoxic conditions

There is no available evidence that the precipitation of Fe-rich phases was the result of direct microbial interaction Iron-formations essentially lack organic C as well as reports of well-documented microfossils that are part of the BIF lithologies Iron isotope studies of BIF are inconclusive as to any possible biological processes responsible for their precipitation As such it would appear that iron-formations are purely chemical pre-cipitates This is not to exclude the connection between biologic activity and iron-formation deposition Cloud (1968 1972 1973 1983) developed an elegant hypothesis for the interrelationship among iron-formation the biochemical evolution of terrestrial life and the chemical evolution of the atmosphere and oceans In his model the Fe in BIFs was oxidized and subsequently pre-cipitated in the oceans by oxygen produced by photosynthesizing organisms This concept is still very much part of the question where did the necessary oxygen ultimately come from How-

ever the question of whether microbial activity was directly responsible for the precipitation of BIF mineral assemblages is a very different one As noted above there is much evidence that BIFs were the products of purely chemical precipitation The question of direct microbial involvement however is still much alive Konhauser et al (2002) suggested that direct microbial oxidation of Fe-rich solutions has the potential to generate the bulk if not all of the Fe2O3 in BIFs Beukes (2004) reviewed three models of Fe-mineral deposition in the Archean ocean (1) one in which free oxygen was derived from microbial photosynthesis (2) a second in which Fe-carbonate was precipitated (without accompanying Fe oxides) by anoxygenic photosynthesis and (3) the model proposed by Konhauser et al (2002) which involves direct microbial activity in BIF precipitation

The mineral reactions that occur during prograde metamor-phism of BIF are essentially isochemical except for dehydration and decarbonation

The mineralogy as well as average bulk chemistry of dia-genetic to low-grade metamorphic iron-formations ranging in age from Archean to Early Proterozoic are essentially the same throughout this long time period The mineral assemblages reszlig ect anoxic conditions of sedimentation and these are corroborated by the low average redox state of Fe in bulk-chemical averages (between that of wuumlstite and magnetite) As such the early mineralogy as well as the average bulk chemistry of almost all BIFs (except for the Neoproterozoic occurrences) point toward anoxic conditions of deposition This is what would be expected if large amounts of Fe2+ were in solution in ocean basins from hydrothermal sources Iron solubility is possible only under highly reducing conditions

In conclusion it is useful to combine the curve of relative iron-formation abundance in the Precambrian (Fig 2) with some curves that reszlig ect the O2 and CO2 concentrations in the Precambrian atmosphere This is done in Figure 29 which shows a fairly complex curve for the evolution of CO2 concentrations in the atmosphere over time (Kasting 2004 Kasting and Catling 2003) from very high values in the Early Archean to the present modern value Also shown is a calculated curve for the evolution of O2 in the atmosphere over time (Kasting 2004 2001 Kasting and Catling 2003) O2 concentration levels are extremely low for all of Archean time until 23 Ga when a transition is postulated from an anoxic to a less reducing atmosphere With regard to the oxygen content of the oceans this means that all ocean waters were anoxic until 23 Ga and that after that time the deep ocean remained anoxic (keeping Fe2+ soluble) but surface waters may have become somewhat less reducing These observations are in accordance with the mineralogic and bulk-chemical data for iron-formations All BIFs before 23 Ga resulted from deep ocean (anoxic) precipitation whereas the granular iron-forma-tions ranging from about 22 to 18 Ga were the result of transport of dissolved Fe2+ (under anoxic conditions) to the higher energy environment of surface waters (as reszlig ected in their granular and oolitic textures) The unmetamorphosed parts of the Sokoman Iron Formation Labrador Trough of about 188 Ga in age (Findlay et al 1995) contain laminated as well as granular (and oolitic) mineral assemblages of greenalite (see Figs 6a and 6b) pyrite magnetite stilpnomelane hematite chert and carbonates (Klein and Fink 1976) This suggests that such assemblages although

FIGURE 28 Paleoceanographic model for Neoproterozoic iron-formation deposition During Snowball Earth conditions sea level stand would have been very low and the oceans stagnant such that highly reducing conditions would have developed for the accumulation of dissolved Fe andor Mn either from hydrothermal sources or dissolution of material from basin szlig oors The onset of interglacial or postglacial stages would have resulted in transgressions restoration of ocean circulation and precipitation of Fe3+-rich iron-formation and manganese-formations From Beukes and Klein (1992)


FIGURE 29 Relative abundance curve for BIF in the Precambrian (taken from Fig 2) as well as calculated curves for the atmospheric evolution of oxygen and carbon dioxide from Kasting (2001 2004) Kasting and Catling (2003) and Pavolv and Kasting (2002)

formed in shallower waters than the older microbanded BIFs were still the result of chemical precipitation (in surface waters) that was highly reducing The granular and oolitic iron-forma-tions of the Frere Formation (Naberru Basin Western Australia) of about 19 to 18 Ga (Williams et al 2004) were originally reported (Hall and Goode 1978) to consist exclusively of chert hematite and magnetite as based on assemblage studies of BIF from outcrops Williams et al (2004) reported on assemblages (from two unweathered drill cores) that contain hematite and magnetite but also siderite ankerite stilpnomelane and possibly greenalite (as determined by X-ray diffraction) The absence of these Fe carbonates and silicates in earlier studies of the BIF in the Frere Formation is ascribed to deep surface oxidation and prolonged weathering As such the two almost coeval granular BIFs (the Sokoman Iron Formation in Labrador and Newfound-land and the Frere Formation in Western Australia) have very similar primary mineral assemblages The oxidizing conditions for the atmosphere as implied by the sharp rise in the O2 curve (in Fig 29) at about 23 Ga are thus not reszlig ected in the Soko-man and Frere BIF assemblages This discrepancy is likely the result of the disequilibrium between the atmosphere and oceans as described by Kasting (1992) The Neoproterozoic iron-formation occurrences are the result of anoxic conditions in stagnant ocean basins created by an ice cover during the period of Snowball Earth and subsequent hematite precipitation during interglacial or post glacial periods

ACKNOWLEDGMENTSMy interest in iron-formations was kindled while I was THORN rst employed during

the summer season of 1959 as a THORN eld geologist with the Iron Ore Company of Canada at Labrador City Newfoundland while I was a graduate student at McGill University That led to my Master s thesis at McGill entitled Amphiboles and as-sociated minerals in the Wabush Iron Formation Labrador 1960 Subsequently during my PhD years at Harvard University I returned to the Iron Ore Company of Canada in Labrador City as a research geologist which allowed me to collect and document all of the materials for my PhD dissertation at Harvard entitled Mineralogy of petrology of the Wabush Iron Formation Labrador City area Newfoundland 1965 In 1973 I had the opportunity to visit some of the Archean iron-formations in the Ruby Mountains of Montana under the expert guidance of the late Hal James And in 1978 as a Guggenheim Fellowship recipient residing for six months in Perth Western Australia I received much encouragement and aid from Alec Trendall toward my THORN eld research in and the sampling of the iron-formations in the Hamersley Range

My iron-formation research from 1972 until the present would have been impossible without the many cooperative research projects with colleagues all over the world Much time in Western Australia was in company with Martin Gole in South Africa with Nic Beukes in Brazil with Eduardo Ladeira and in Canada with

Richard Fink All of these THORN eld-based studies have led to joint publications The late Takashi Miyano provided much thermodynamic expertise wonderful inspiration and much hard work in our joint evaluations of Fe-rich mineral stabilities Others with whom I have had successful joint BIF studies are Bob Dymek Owen Bricker John Hayes Jim Walker Clark Johnson Alan Kaufman and Bill Schopf I much enjoyed my long-term educational experience (19791990) as a member of the Precambrian Paleobiology Research Group (PPRG) under the expert enthusiastic and generous directorship of Bill Schopf at UCLA Several of my graduate students at Harvard and Indiana University Bloomington chose thesis and dissertation subjects related to iron-formations These are Norm Gillmeister Inda Immega Pete Dahl Mike Lesher Steve Haase and Alan Kaufman Clearly the present overview paper on iron-formations owes much to all of the above colleagues Critical reviews by Alec Trendall and Nic Beukes helped to improve the manuscript

I am grateful to Jim Kasting for his input on the evaluation of the evolution of the content of gases such as O2 and CO2 in the Precambrian atmosphere And last but not least none of this work would have been possible had it not been for the research grant support provided me by the National Science Foundation from 1972 until 2000

REFERENCES CITED

Barghoorn ES and Tyler SA (1965) Microorganisms from the Gunszlig int chert Science 147 563577

Bau M and Moumlller P (1993) Rare earth element systematics of the chemically precipitated component in Early Precambrian iron-formations and the evolution of the terrestrial atmosphere-hydrosphere-lithosphere system Geochimica et Cosmochimica Acta 57 22392249

Baxter JL (1971) The Archean stratigraphy of the Tallering Range and Nunierra Hill area Western Australia Geological Survey Annual Report 1970 4345

Bayley RW and James HL (1973) Precambrian Iron-Formation of the United States Economic Geology 68 934959

Berner RA (1970) Sedimentary pyrite formation American Journal of Science 268 123

Beukes NJ (1973) Precambrian Iron-Formations of Southern Africa Economic Geology 68 9601004

(2004) Early options in photosynthesis Nature 431 522523Beukes NJ and Klein C (1990) Geochemistry and sedimentology of a facies

transition from microbanded to granular iron-formation in the early Proterozoic Transvaal Supergroup South Africa Precambrian Research 47 99139

(1992) Models for iron-formation deposition In JW Schopf and C Klein Eds The Proterozoic Biosphere a multidisciplinary study p 147151 Cam-bridge University Press New York

Beukes NJ Klein C Kaufmann AJ and Hayes JM (1990) Carbonate petrogra-phy kerogen distribution and carbon and oxygen isotopic variations in an Early Proterozoic transition from limestone to iron-formation deposition Transvaal Supergroup South Africa Economic Geology 85 663690

Bonnichsen B (1969) Metamorphic pyroxenes and amphiboles in the Biwabik Iron Formation Dunka River area Minnesota Mineralogical Society of America Special Paper 2 217241

(1975) Geology of the Biwabik Iron Formation Dunka River area Min-nesota Economic Geology 70 319340

Butler P Jr (1969) Mineral compositions and equilibria in the metamorphosed iron-formation of the Gagnon region Quebec Canada Journal of Petrology 10 56101

Cloud PE (1968) Atmospheric and hydrospheric evolution on the primitive Earth Science 160 729736

(1972) A working model for the primitive Earth American Journal of


Science 272 537548 (1973) Paleoecological signiTHORN cance of banded iron-formation Economic

Geology 68 11351143 (1983) Early biogeologic history The emergence of a paradigm In JW

Schopf Ed Earth s Earliest Biosphere Its origin and evolution 1431 Princ-eton University Press Princeton NJ

Dahl PS (1979) Comparative geothermometry based on major-element and oxygen isotope distributions in Precambrian metamorphic rocks from southwestern Montana American Mineralogist 64 12801293

Daniels JL and Halligan R (1968) Explanatory notes on the Wyloo 1250000 geo-logical sheet Western Australia Geological Survey 196869 (unpublished)

Des Marais DJ (2001) Isotopic evolution of the biogeochemical carbon cycle during the Precambrian In JW Valley and DR Cole Eds Stable Isotope Geochemistry 43 555578 Reviews in Mineralogy and Geochemistry Min-eralogical Society of America Washington DC

Dimroth E and Chauvel JJ (1973) Petrography of the Sokoman Iron Formation in part of the central Labrador Trough Quebec Canada Geological Society of America Bulletin 84 111134

Dunbar GJ and McCall GJH (1971) Archean turbidites and banded ironstones of the Mt Belches area (Western Australia) Sedimentary Geology 5 93113

Dymek RF and Klein C (1988) Chemistry petrology and origin of banded iron-formation lithologies from the 3800 Ma Isua Supracrustal Belt West Greenland Precambrian Research 39 247302

Eggleton RA (1972) The crystal structure of stilpnomelane Part II The full cell Mineralogical Magazine 38 693711

ElderTHORN eld H and Greaves MJ (1982) The rare earth elements in seawater Nature 296 214219

Eriksson KA (1978) Marginal marine depositional processes from the Archean Moodies Group Barberton Mountain Land South Africa Evidence and sig-niTHORN cance Precambrian Research 8 153182

Eugster HP and Chou I-Ming (1973) The depositional environments of Precam-brian banded iron-formations Economic Geology 68 11441169

Findlay JM Parrish RR Birkett TC and Watanabe DH (1995) U-Pb ages from the Nimish Formation and the Montagnais glomerophitic gabbro of the central New Queacutebec Orogen Canada Canadian Journal of Earth Science 32 12081220

Floran RJ (1975) Mineralogy and petrology of the sedimentary and contact metamorphosed Gunszlig int Iron-Formation Ontario-Minnesota PhD thesis 342 p State University of New York Stony Brook NY

Floran RJ and Papike JJ (1975) Petrology of the Gunszlig int Iron-Formation Ontario-Minnesota the low-grade rocks Geological Society of America Bul-letin 86 11691190

French BM (1968) Progressive contact metamorphism of the Biwabik Iron-For-mation Mesabi Range Minnesota Minnesota Geological Survey Bulletin 45 103 p

Fryer BJ (1983) Rare earth elements in iron-formation In AF Trendall and RC Morris Eds Iron-Formation Facts and Problems p 345358 Elsevier Amsterdam

Fryer BJ Fyfe WS and Kerrich R (1979) Archean volcanogenic oceans Chemical Geology 24 2533

Fyon JA Breaks FW Heather KB Jackson SL Muir TL Scott GM and Thurston PC (1992) In PC Thurston HR Williams RH Sutcliffe and GM Scott Eds Geology of Ontario Ontario Geological Survey Special Volume 4 Part 2 10911174

Garrels RM and Christ CL (1965) Solutions Minerals and Equilibria Freeman Cooper San Francisco 765 p

Gillmeister N (1971) The petrology stratigraphy and structure of the Precambrian metamorphosed rocks of the central Tobacco Root Mountains with special em-phasis on the electron microprobe analysis of high grade mineral assemblages PhD thesis Harvard University Cambridge Mass 201 p

Gole MJ (1980) Mineralogy and petrology of very-low-grade metamorphic grade Archean banded iron-formations Weld Range Western Australia American Mineralogist 65 825

Gole MJ and Klein C (1981a) Banded Iron-Formation through much of Precam-brian time Journal of Geology 89 169183

(1981b) High-grade metamorphic Archean banded iron-formations West-ern Australia Assemblages with coexisting pyroxenes plusmn fayalite American Mineralogist 66 8799

Goode ADT Hall WDM and Bunting JA (1983) The Naberru basin of Western Australia In AF Trendall and RC Morris Eds Iron-Formation Facts and Problems p 295323 Elsevier Amsterdam

Goodwin AM (1956) Facies relations in the Gunszlig int iron-formation Economic Geology 51 565595

Gromet LP Dymek RF Haskin LA and Korotev RL (1984) The North American shale composite its compilation major and trace element charac-teristics Geochimica et Cosmochimica Acta 48 24692482

Gross GA (1972) Primary features in cherty iron-formations Sedimentary Geol-ogy 7 241261

Gross GA and Zajac IS (1983) Iron-formation in fold belts marginal to the

Ungava craton In AF Trendall and RC Morris Eds Iron Formation Facts and Problems p 253294 Elsevier Amsterdam

Haase CS (1982) Metamorphic petrology of the Negaunee Iron-Formation Marquette District Northern Michigan Mineralogy metamorphic reactions and phase equilibria Economic Geology 77 6081

Hall WDM and Goode ADT (1978) The early Proterozoic Nabberu Basin and associated iron-formations of Western Australia Precambrian Research 7 129184

Hayes JM Lambert IB and Strauss H (1992) The sulfur-isotopic record In JW Schopf and C Klein Eds The Proterozoic Biosphere a multidisciplinary study p 129132 Cambridge University Press New York

Helgeson HC Delany JM Nesbitt HW and Bird DK (1978) Summary and critique of the thermodynamic properties of rock-forming minerals American Journal of Science 178A 1229

Hoffman PF and Schrag DP (2002) The Snowball Earth hypothesis Testing the limits of global change Terra Nova 14 129155

Holland HD (1973) The Oceans a possible source for iron in iron-formation Economic Geology 68 11691172

Immega IP and Klein C (1976) Mineralogy and petrology of some Precam-brian iron-formations in southwestern Montana American Mineralogist 61 11171144

Isley AE (1995) Hydrothermal plumes and the delivery of iron to banded iron formation Journal of Geology 103 169185

James HL (1954) Sedimentary facies of iron-formation Economic Geology 49 235293

(1955) Zones of regional metamorphism in the Precambrian of Northern Michigan Geological Society of America Bulletin 66 14551487

(1983) Distribution of Banded Iron-Formations in space and time In AF Trendall and RC Morris Eds Iron Formation Facts and Problems p 471490 Elsevier Amsterdam

James HL and Trendall AF (1982) Banded Iron-Formation Distribution in time and paleoenvironmental signiTHORN cance In HD Holland and M Schidlowski Eds Mineral Deposits and the Evolution of the Biosphere p 199218 Springer Verlag NY

Johnson CM Beard BL Beukes NJ Klein C and OʼLeary J (2003) Ancient geochemical cycling in the Earth as inferred from Fe isotope studies of banded iron formations from the Kaapvaal Craton Contributions to Mineralogy and Petrology 144 523547

Kasting JF (1992) Models relating to Proterozoic atmospheric and ocean chemistry In JW Schopf and C Klein Eds The Proterozoic Biosphere A multidisci-plinary study p 11851187 Cambridge University Press New York

(2004) When methane made climate ScientiTHORN c American 291 7885 (2001) The rise of atmospheric oxygen Science 293 819820Kasting JF and Catling D (2003) Evolution of a habitable planet Annual Review

of Astronomy and Astrophysics 41 429463Kaufman AJ Hayes JM and Klein C (1990) Primary and diagenetic controls

of isotopic compositions of iron-formation carbonates Geochimica et Cosmo-chimica Acta 54 34613473

Kaufman AJ and Knoll AH (1995) Neoproterozoic variations in the C-isotopic composition of seawater Stratigraphic and biogeochemical implications Pre-cambrian Research 73 2449

Kimberley MM (1978) Paleoenvironmental classiTHORN cation of Iron-Formations Economic Geology 73 p 215229

Kirschvink JL (1992) Late Proterozoic low-latitude global glaciation The Snow-ball Earth In JW Schopf and C Klein Eds The Proterozoic Biosphere a multi-disciplinary study 5152 Cambridge University Press New York

Klein C (1964) Cummingtonite-grunerite series a chemical optical and X-ray study American Mineralogist 46 963982

(1966) Mineralogy and petrology of the metamorphosed Wabash Iron-For-mation southwestern Labrador Journal of Petrology 7 246305

(1968) Coexisting amphiboles Journal of Petrology 9 281330 (1973) Changes in mineral assemblages with metamorphism of some banded

iron-formations Economic Geology 68 10751088 (1974) Greenalite stilpnomelane minnesotaite crocidolite and carbon-

ates in very low-grade metamorphic Precambrian iron-formation Canadian Mineralogist 12 475498

(1978) Regional metamorphism of Proterozoic iron-formation Labrador Trough Canada American Mineralogist 63 898912

(1983) Diagenesis and metamorphism of Precambrian iron-formations In AF Trendall and RC Morris Eds Iron-Formation Facts and Problems p 417469 Elsevier Amsterdam

(1993) Rocks Minerals and a Dusty World In GD Guthrie and BT Moss-man Eds Health Effects of Mineral Dusts 28 759 Reviews in Mineralogy and Geochemistry Mineralogical Society of America Washington DC

(2001) Precambrian banded iron-formations (BIFs) worldwide Their setting mineralogy metamorphism and origin Geological Society of America Annual Meetings Abstracts with Programs A-195 GSA Boulder Colorado

Klein C and Beukes NJ (1989) Geochemistry and sedimentology of a facies transition from limestone to iron-formation deposition in the Early Proterozoic


Transvaal Supergroup South Africa Economic Geology 84 17331774 (1992) Time distribution stratigraphy and sedimentologic setting and geo-

chemistry of Precambrian Iron Formation In JW Schopf and C Klein Eds The Proterozoic Biosphere A multidisciplinary study p 139146 Cambridge University Press New York

(1993a) Proterozoic iron-formation In KC Condie Ed Proterozoic Crustal Evolution p 383418 Elsevier Amsterdam

(1993b) Sedimentology and geochemistry of the glaciogenic Late Protero-zoic Rapitan Iron-Formation in Canada Economic Geology 88 542565

Klein C and Bricker OP (1977) Some aspects of the sedimentary and diagenetic environment of Proterozoic banded iron-formations Economic Geology 72 14571470

Klein C and Fink RP (1976) Petrology of the Sokoman Iron-Formation in the Howells River area at the western edge of the Labrador Trough Economic Geology 71 453487

Klein C and Gole MJ (1981) Mineralogy and petrology of parts of the Marra Mamba iron-formation Hamersley Basin Western Australia American Min-eralogist 66 507525

Klein C and Ladeira EA (2000) Geochemistry and petrology of some Proterozoic banded iron-formations of the Quadrilaacutetero Ferriacutefero Minas Gerais Brazil Economic Geology 95 405428

(2002) Petrography and geochemistry of the least altered banded iron-formations of the Archean Carajaacutes Formation Brazil Economic Geology 97 643651

(2004) Geochemistry and mineralogy of Neoproterozoic banded Iron-For-mations and some selected siliceous manganese formations from the Urucum district Mato Grosso do Sul Brazil Economic Geology 99 12331244

Klein C Beukes NJ and Schopf JW (1987) Filamentous microfossils in the Early Proterozoic Transvaal Supergroup Their morphology signiTHORN cance and paleoenvironmental setting Precambrian Research 36 8194

Konhauser KO Hamade T Raiswell R Morris RC Ferris FG Southam G and CanTHORN eld DE (2002) Could bacteria have formed the Precambrian banded iron-formations Geology 30 10791082

Knoll AH and Simonson B (1980) Early Proterozoic microfossils and penecon-temporaneous quartz cementation in the Sokoman Iron Formation Science 211 478480

Kranck SH (1961) A study of phase equilibria in metamorphic iron-formation Journal of Petrology 2 137184

Krape B Barley ME and Pickard AL (2003) Hydrothermal and resedimented origins of the precursor sediments to banded iron-formation Sedimentological evidence from the Early Palaeoproterozoic Brockman Supersequence of Western Australia Sedimentology 50 9791011

Lesher CM (1978) Mineralogy and petrology of the Sokoman Iron Formation near Ardua Lake Quebec Canadian Journal of Earth Science 15 480500

Miyano T (1978) Effects of CO2 on mineralogical differences in some low-grade metamorphic iron-formations Geochemical Journal 12 201211

(1982) Stilpnomelane Fe-rich mica K-feldspar and hornblende in banded iron-formation assemblages of the Dales Gorge Member Hamersley Group Western Australia Canadian Mineralogist 20 189202

Miyano T and Klein C (1983a) Evaluation of the stability relations of amphibole asbestos in metamorphosed iron-formations Mining Geology 33 213222

(1983b) Conditions of riebeckite formation in the iron-formation of the Dales Gorge Member Hamersley Group Western Australia American Min-eralogist 68 517529

(1986) Fluid behavior and phase relations in the system Fe-Mg-Si-C-O-H application to high grade metamorphism of iron-formations American Journal of Science 286 540575

(1989) Phase equilibria in the system K2O-FeO-MgO-Al2O3-SiO2-H2O and the stability limit of stilpnomelane in metamorphosed Precambrian iron-forma-tions Contributions to Mineralogy and Petrology 102 478491

Miyano T and Miyano S (1982) Ferri-annite from the Dales Gorge Member iron-formation Wittenoom area Western Australia American Mineralogist 67 11791195

Morey GB Papike JJ Smith RW and Weiblen PW (1972) Observations on the contact metamorphism of the Biwabik Iron-Formation East Mesabi district Minnesota Geological Society of America Memoir 135 225264

Morris RC (1993) Genetic modeling for banded iron-formations of the Ham-ersley Group Pilbara Craton Western Australia Precambrian Research 60 243286

Mueller RF (1960) Compositional characteristics and equilibrium relations in mineral assemblages of a metamorphosed iron-formation American Journal of Science 258 449497

Papike JJ Cameron K and Shaw KW (1973) Chemistry of coexisting ac-tinolite-cummingtonite and hornblende-cummingtonite from metamorphosed iron-formation (abstract) Geological Society of America Abstracts with Programs 5 763764

Pavlov AA and Kasting JJ (2002) Mass-independent fractionation of sulfur isotopes in Archean sediments Strong evidence for an anoxic Archean atmo-sphere Astrobiology 2 27410

Peter JM (2003) Ancient iron-formations their genesis and use in the exploration for stratiform base metal sulTHORN de deposits with examples from the Bathurst mining Camp In DR Lentz Ed Geochemistry of sediments and sedimentary rocks Evolutionary considerations to mineral-deposit-forming environments Geological Association of Canada GeoText 4 145176

Roberts JB (1975) Windarra nickel deposits In CL Knight ed Economic Geol-ogy of Australia and Papua New Guinea vol 1 Metals Australian Institute of Mining and Metallurgy Melbourne 883892

Robie RA Hemingway BS and Fischer JR (1978) Thermodynamic properties of minerals and related substances at 29815 K and 1 bar (105 Pascals) pressure and at higher temperatures US Geological Survey Bulletin 1452 456 p

Ross M Papike JJ and Shaw KW (1969) Exsolution textures in amphiboles as indicators of subsolidus thermal histories Mineralogical Society of America Special Paper 2 275299

Schidlowski M Hayes JM and Kaplan IR (1983) Isotopic inferences of an-cient biochemistries Carbon sulfur hydrogen and nitrogen In JW Schopf Ed Earth s Earliest Biosphere its origin and evolution p 149186 Princeton University Press New Jersey

Schmidt PW and Williams GE (1995) The Neoproterozoic climatic paradox Equatorial paleolatitude for Marinoan glaciation near sea level in South Aus-tralia Earth and Planetary Science Letters 134 107124

Schmidt RG (1963) Geology and ore deposits of the Cuyuna North Range Min-nesota US Geological Survey Professional Paper 407 96 p

Simmons EC Lindsley DH and Papike JJ (1974) Phase relations and crys-tallization sequence in a contact metamorphosed rock from the Gunszlig int Iron-Formation Minnesota Journal of Petrology 15 539565

Simonson BM (1985) Sedimentological constraints on the origins of Precambrian iron-formations Geological Society of America Bulletin 96 244252

Trendall AF (1973a) Precambrian iron-formations of Australia Economic Geol-ogy 68 10231034

(1973b) Iron-formations of the Hamersley Group of Western Australia type examples of varved Precambrian evaporates In Genesis of Precambrian iron and manganese deposits Proceedings Kiev Symposium 1970 Unesco Paris 257270

(1983) Introduction In AF Trendall and RC Morris Eds Iron-Formation Facts and Problems 111 Elsevier Amsterdam

(2002) The signiTHORN cance of iron-formation in the Precambrian stratigraphic record Special Publication International Association of Sedimentologists 33 3366

Trendall AF and Blockley JG (1970) The Iron-Formations of the Precambrian Hamersley Group Western Australia Geological Survey Western Australia Bulletin 119 366 p

Trendall AF Compston W Nelson DR De Laeter JR and Bennett VC (2004) SHRIMP zircon ages constraining the depositional chronology of the Hamersley Group Western Australia Australian Journal of Earth Sciences 51 621644

Vaniman DT Papike JJ and Labotka T (1980) Contact metamorphic effects of the Stillwater Complex Montana the concordant iron-formation American Mineralogist 65 10871102

Viljoen MJ and Viljoen RP (1969) An introduction to the geology of the Bar-berton granite-greenstone terrain Geological Society South Africa Special Publication 2 928

Walker JCG Klein C Schidlowski M Schopf JW Stevenson DJ and Walter MR (1983) Environmental evolution of the Archean-Early Proterozoic Earth In JW Schopf Ed Earth s Earliest Biosphere its Origin and Evolution 260290 Princeton University Press New Jersey

Walter MR and Hoffman HJ (1983) The palaeontology and palaeoecology of Precambrian iron-formations In AF Trendall and RC Morris Eds Iron-Formation Facts and Problems p 373400 Elsevier Amsterdam

Walter MR Goode ADT and Hall WDM (1976) Microfossils from a newly discovered Precambrian stromatolitic iron-formation in Western Australia Nature 261 221223

Williams GE Schmidt PW and Clark DA (2004) Paleomagnetism of iron-formation from the late Palaeoproterozoic Frere Formation Earaheedy Basin Western Australia Paleogeographic and tectonic implications Precambrian Research 128 367383

Yamaguchi KE Bau M and Ohmoto H (2000) Geochemistry of rare earth ele-ments in Precambrian banded iron-formations Are the Ce anomalies real (ab-stract) First Astrobiology Science Conference Ames Research Cener 296

Young TP and Taylor WEG Eds (1989) Phanerozoic Ironstones Geological Society Special Publication 46 251 p

Zajac IS (1974) The stratigraphy and mineralogy of the Sokoman Formation in the Knob Lake area Quebec and Newfoundland Geological Survey of Canada Bulletin 220 159 p

MANUSCRIPT RECEIVED MARCH 28 2005MANUSCRIPT ACCEPTED DECEMBER 3 2004MANUSCRIPT HANDLED BY ROBERT F DYMEK


granular but also show banding (on a scale of tens of centimeters instead of millimeters) and these have been referred to as GIF (granular iron-formation) by Trendall (2002) In this paper all are grouped under the name BIF This overview of iron-formations does not include a discussion of Phanerozoic ironstones although Kimberley (1978) suggested that these Phanerozoic deposits be considered iron-formations Most Phanerozoic ironstones are very different from Precambrian BIF in texture as well as mineralogy They are commonly ooidal in texture and many contain abundant silicates such as berthierine and chamosite (Young and Taylor 1989) that are generally absent from Precambrian iron-formations The overall mineralogy of ironstones consisting of hematite (or goethite) magnetite berthierine siderite-rich types and lacking chert is totally different from that of Precambrian BIF

The global distribution of well-known Precambrian IFs is shown in Figure 1 This map (modiTHORN ed after Trendall 2002) emphasizes their wide global distribution and does not attempt to show all known occurrences

DISTRIBUTION OF IRON-FORMATIONS THROUGHOUT PRECAMBRIAN TIME

A schematic relative abundance curve for Precambrian iron-formations as a function of age is given in Figure 2 This curve which is adapted from that of Gole and Klein (1981a) is in essence a best estimate of total volume of iron-formation (not Fe ore) for all iron-formations tabulated by James (1983) James and Trendall (1982) Walker et al (1983 which lists the age ranges of 54 BIFs in their Table 111) and Fyon et al (1992 with Table 224 listing the ages of several Archean iron-forma-tions in the Superior Province of Ontario) relative to a maximum represented by the total iron-formation volume of the Hamersley Range of Western Australia Figure 2 is similar but not identical to a curve of Trendall (2002 his Fig 6) in which the estimated volume of smaller Archean BIFs (in greenstone belts in Archean Shields) is smaller than that implied by the concave upward part of the curve in Figure 2 between about 35 and 27 Ga Gole and Klein (1981a) concluded that the size and extent of Archean iron-formations have commonly been underestimated Such BIFs

are generally discontinuous commonly tectonically deformed and dismembered and only partially exposed which leads to an underestimation of their original size and volume (see Fig 3) The elevated concave curve in Figure 2 (between 33 and 27 Ga) is the result of compensation for this underestimation in volume There is general agreement that the largest BIFs are those of the Hamersley Range of Western Australia (ranging in age from about 26 to 245 Ga Trendall and Blockley 1970 Trendall 2002 Tren-dall et al 2004) and the Transvaal Supergroup of South Africa with an age of approximately 25 to 23 Ga (Klein and Beukes 1989) There is additional agreement on a sharp decline in major iron-formations in the Precambrian record at about 18 Ga and the return of several BIFs in the geologic record that show ap-proximate ages between 08 to 06 Ga (such as iron-formations in the Rapitan Group Yukon and NWT Canada in the Urucum region Brazil and in the Damara Supergroup of Namibia)

STRATIGRAPHIC SETTING

The stratigraphic sequences in which BIFs occur are highly variable and there appears to be no consistent litho-

FIGURE 1 Global occurrence of Precambrian BIFs (after Trendall 2002) The map emphasizes the wide distribution of BIF and shows only a selection of well-known occurrences of different ages and kinds BIFs that are discussed in this paper are shown in italics

FIGURE 2 Highly schematic diagram showing the relative abundance of Precambrian BIFs vs time with several of the major BIFs or major BIF regions identiTHORN ed Estimated abundances are relative to the Hamersley Group BIF volume taken as a maximum (adapted from Gole and Klein 1981a also based on tabulation of BIF dates in Walker et al 1983 their table 111) The most recent age evaluations for the Hamersley Basin (28 to 22 Ga) are available from Trendall et al (2004) for the Labrador Trough BIFs (188 Ga) from Findlay et al (1995) and for the BIFs in the Frere Formation Western Australia (19 to 18 Ga) from Williams et al (2004)


































2





























METAMORPHISM



























grunerite calcite



and







+ 14CO2
















and
























FORMATION































































CONCLUDING REMARKS























REFERENCES CITED





































































































































































2





























METAMORPHISM



























grunerite calcite



and







+ 14CO2
















and
























FORMATION































































CONCLUDING REMARKS























REFERENCES CITED


























































































































































METAMORPHISM



























grunerite calcite



and







+ 14CO2
















and
























FORMATION































































CONCLUDING REMARKS























REFERENCES CITED













































































































































grunerite calcite



and







+ 14CO2
















and
























FORMATION































































CONCLUDING REMARKS























REFERENCES CITED














































































































































and
























FORMATION































































CONCLUDING REMARKS























REFERENCES CITED



















































































































































FORMATION































































CONCLUDING REMARKS























REFERENCES CITED

































































































































































































CONCLUDING REMARKS























REFERENCES CITED





















































































































































REFERENCES CITED




































































































































precambrian banded iron-formations

Documents

ga ironformations

metamorphosed bifs

archean ironformations

neoproterozoic ironformations

ga isua

deformation of extensive

average bulk chemistry

geologic setting