Discovery of the oldest oxic granitoids in the Kaapvaal Craton and its implications for the redox evolution of early Earth
Published:January 01, 2006
- PDF LinkChapter PDF
Shunso Ishihara, Hiroshi Ohmoto, Carl R. Anhaeusser, Akira Imai, Laurence J. Robb, 2006. "Discovery of the oldest oxic granitoids in the Kaapvaal Craton and its implications for the redox evolution of early Earth", Evolution of Early Earth's Atmosphere, Hydrosphere, and Biosphere - Constraints from Ore Deposits, Stephen E. Kesler, Hiroshi Ohmoto
Download citation file:
Phanerozoic granitoids have been classified into magnetite and ilmenite series based on the abundance of magnetite, which is related to the Fe2O3/FeO ratio of the rock and the oxygen fugacity (f O2 ) of its parent magma. We have examined the temporal and spatial distributions of both series in Archean granitoids from the Barberton region and the Johannesburg Dome of the Kaapvaal Craton, South Africa. The oldest syntectonic TTG (tonalite-trondhjemite-granodiorite) granitoids (ca. 3450 Ma in age) were found to be ilmenite series, whereas some intermediate-series granitoids occurred locally. Younger and larger syntectonic TTGs (e.g., the 3230 Ma Kaap Valley plutons) comprise nearly equal quantities of magnetite and ilmenite series. The major 3105 Ma calc-alkaline batholiths (e.g., Nelspruit batholith), emplaced during the late-tectonic stage, comprise mostly magnetite-series granitoids, suggesting that an oxidized continental crust already existed by this time.
The rare earth element ratios and δ18O values, as well as the Fe2O3/FeO ratios, of the Archean magnetite-series granitoids suggest that their magmas were generated from the partial melting of subducted oceanic basalts that had been oxidized by interaction with seawater on mid-oceanic ridges; the processes of magma generation were much like those for Phanerozoic magnetite-series granitoids. This further suggests that the concentrations of oxidants (O2 and/or SO4 2−) in the Archean oceans were similar to those in Phanerozoic oceans.
Low concentrations of chlorine in the magmas, as well as deep levels of granite erosion, appear to explain the absence of major mineral deposits associated with the Kaapvaal granitoids.
Granitoids represent an integrated component of the continental lithosphere, hydrosphere, oceanic crust, and upper mantle, which were amalgamated by dynamic plate motion. Ishihara (1977) has suggested that Phanerozoic granitoids can be classified into two types, magnetite series and ilmenite series, based on the presence or absence of magnetite, which is determined by microscopic observation and modal analyses. However, when the magnetite content of a rock is very low, microscopic observation or modal analysis is difficult; in these situations, magnetite can be easily detected by a magnetic susceptibility measurement, even in the field (Ishihara et al., 2000). Therefore, Ishihara (1977) has proposed a magnetic susceptibility value of 100 × 10−6 emu/g (or 3 × 10−3 SI unit) at SiO2 ≈70% as the boundary between ilmenite and magnetite series. Magnetic susceptibility of ilmenite-series granitoids can be as low as ∼0.01 × 10−3 SI, and that of magnetite-series granitoids can be as high as ∼100 × 10−3 SI. Ishihara (1981) has also found that the Fe2O3/FeO ratios of granitoids generally increase with increasing values of magnetic susceptibility, from less than 0.5 in weight ratio (or 0.22 in mole ratio) for ilmenite series to greater than 0.5 in magnetite series.
The magnetite- versus ilmenite-series granitoid classification is primarily based on the oxidation state of magma: magnetite series indicate higher f O2 values and ilmenite series denote lower f O2 values at a given temperature and pressure condition. Ishihara (1977) and Czamanske et al. (1981) suggest that the f O2 boundary for the two series probably lies near the NNO (nickel+nickel oxide) buffer line (Fig. 1).
Ishihara (1977) has also recognized that Cu-Au and Cu-Mo porphyry deposits are typically associated with magnetite-series granitoids, whereas W and Sn deposits are found with ilmenite-series granitoids. Such associations of granitoid and ore deposit types occur because the redox state of magma strongly influences the sulfur chemistry (e.g., SO2, H2S, and SO4 2−), as well as metal chemistry (e.g., Fe, Cu, Mo, Au, W, and Sn), of magmatic fluids (Burnham and Ohmoto, 1980).
Distinct differences also exist between magnetite- and ilmenite-series granitoids in various geochemical parameters (e.g., O and S isotope ratios; Ishihara et al., 2000; Ishihara and Matsuhisa, 2002; Sasaki and Ishihara, 1979; Ishihara and Sasaki, 2002), as well as in their Fe2O3/FeO ratio (Ishihara, 1977, 2004). Such data have been used to suggest that magnetite-series granitoids were generated in subduction zones from the partial melting of hydrated oceanic crust with an Fe2O3/FeO ratio that had been increased by reactions with O2− - and SO4 2−–rich seawater at mid-oceanic ridges, whereas ilmenite-series granitoids were generated from the partial melting of normal mantle amphibolites with some contributions of metasediments from the continental crust (e.g., shales, graywackes) (Burnham and Ohmoto, 1980; Ishihara and Matsuhisa, 1999). Note that by seawater-rock interactions, the Fe2O3/FeO (weight) ratios of oceanic basalts have typically increased from ∼0.08 (normal mantle value) to ∼0.1 − ∼1.0 (average = 0.31) (e.g., Lécuyer and Ricard, 1999).
Compared with Phanerozoic granitoids, very little is known about the petrochemistry and genesis of Archean granitoids. No study has evaluated the redox state of Archean granitoids. The main objective of this study is, therefore, to determine whether magnetite- and/or ilmenite-series granitoids formed during the Archean. The answer will provide important constraints on the chemical evolution of the mantle, crust, oceans, and atmosphere. We have pursued this objective through magnetic, mineralogical, and geochemical investigations of major granite batholiths and plutons (older than 3150 Ma) in the Kaapvaal Craton, South Africa.
FERRIC/FERROUS RATIOS AND OXIDATION STATE OF GRANITOID MAGMAS (THEORETICAL)
Because our study attempts to relate the magnetic susceptibility and Fe2O3/FeO ratios of granitoids to the oxidation state of magmas, it is necessary first to evaluate the relationships among the magnetic susceptibility, relative abundances of Fe-bearing minerals, Fe2O3/FeO ratio of granitoids, and the fugacity of oxygen (f O2 ) in the magmas.
Magnetic Susceptibility and Fe2O3/FeO Ratio of Granitoids
The magnetic susceptibility of granitoids, ranges from 0.01 × 10−3 to 100 × 10−3 SI and is primarily a measure of the amount of magnetite present in the rock, because other magmatic magnetic minerals (e.g., pyrrhotite, ilmenite) are typically much less abundant than magnetite. Magnetic susceptibility of magnetite is 5 × 10−4m3kg1 (Thompson and Oldfield, 1986), and that of pyrrhotite and ilmenite is known to be lower by two orders of magnitude. Mafic silicates, common in granitoids such as hornblende and biotite, are lower by three orders of magnitude in magnetic susceptibility, but higher by one order of magnitude in modal abundance, than magnetite. Therefore, the measured magnetic susceptibility represents a combination of these mafic minerals, but heavily depending on the magnetite contents.
Measurement of the magnetic susceptibility indicates that the f O2 conditions of all (or most) granitoid magmas should fall between the quartz+fayalite+magnetite (QFM) and magnetite+hematite (MH) buffer lines (Fig. 1). Oxygen fugacity conditions above the MH buffer line are unlikely because hematite+ferrous silicates are a non-equilibrium assemblage. We ma y further assume that in granitoids (i) the ferrous component (Fe2+ or FeO) mostly resides in ferrous-rich silicates (mostly hornblende and biotite, rarely pyroxene) and magnetite, although some may be found in pyrrhotite and ilmenite; and (ii) the ferric component (Fe3+ or Fe2O3) resides largely in magnetite, although some may exist in biotite (Czamanske et al., 1981). Certain granitoids contain appreciable amounts of hematite that formed by subsolidus reactions with meteoric water (e.g., Taylor, 1968).
This positive relationship between the Fe2O3/FeO ratio and Xmt in rocks is shown in Figure 2. It illustrates that the maximum Fe2O3/FeO ratio is 1 in mole ratio (or 2.2 in weight ratio) when all the Fe atoms in rocks are in magnetite. If the Fe2O3/FeO ratio exceeds this value, it suggests that a significant amount of secondary hematite crystals occur in the granite.
Relationships between the Fe2O3/FeO Ratio and fO2 of Magma
The activity coefficient ratio, γFe 3+/γFe 2+, most likely depends on the bulk chemistry of the rock or melt (e.g., SiO2 content), but here we will assume unity. K depends on temperature (T) and total pressure (Ptotal). Therefore, C has a unique value at a given set of T, pH, f H2O, and melt (or rock) composition.
Equation (8) indicates that, at a given set of T, Ptotal, PH2O, and bulk rock chemistry, the log f O2 value proportionally increases to four times the value of log(Fe2O3/FeO)mol (Fig. 3). For example, when Ptotal = 1 kb and T = 900 °C, the log f O2 - log(Fe2O3/FeO)mol line extends from log f O2 = −7.25 and log(Fe2O3/FeO)mol = 0 (i.e., MH buffer) to log f O2 = −12.86 and log(Fe2O3/FeO)mol = −1.40 (i.e., QFM buffer). When Ptotal = 1 kb and T = 700 °C, the line extends from log f O2 = −11.70 and log(Fe2O3/FeO)mol = 0 (i.e., MH buffer) to log f O2 = −17.35 and log(Fe2O3/FeO)mol = −1.41 (i.e., QFM).
The above examples of the f O2 versus Fe2O3/FeO ratio relationship illustrate an important point when comparing the redox state of igneous rocks that crystallized at different T-P conditions; the comparison should be made in terms of the Fe2O3/FeO (or Fe3+/Fe2+) ratios of igneous rocks, rather than the f O2 values, because the f O2 value varies depending on temperature and pressure conditions.
GRANITOIDS IN THE KAAPVAAL CRATON
The Kaapvaal Craton (Fig. 4) is one of the oldest and best-preserved Archean continental fragments on Earth. Its assembly during the Archean eon is attributed to a complex combination of processes analogous to modern-day plate tectonics (Poujol et al., 2003). Such processes took place episodically over a 1000 million-year period (ca. 3500–2500 Ma) and involved magmatic arc formation and accretion as well as the tectonic amalgamation of numerous, discrete terranes or blocks (de Wit et al., 1992; Lowe, 1994; Poujol and Robb, 1999).
The oldest rocks so far recognized are located in the Swaziland-Barberton regions on the eastern side of the Kaapvaal Craton (Fig. 4), where ages more than 3600 Ma have been recorded. The early stages of shield development are also best exposed in the Barberton Mountains, where it is now apparent that continent formation took place by magmatic accretion and tectonic amalgamation of small protocontinental blocks. At Barberton, several diachronous blocks that formed between 3600 and 3200 Ma have been identified (Kamo and Davis, 1994), each of which represents a cycle of arc-related magmatism and sedimentation (Lowe, 1999; Poujol et al., 2003).
The Barberton Region
The Archean granitoids in the Barberton region occupy an area of ∼20,000 km2 (Fig. 5). They are divided into three groups based on their intrusive ages in tectonic history: syntectonic, late-tectonic, and post-tectonic (Anhaeusser and Robb, 1980a; Meyer et al., 1994).
Syntectonic tonalite-trondhjemite-granodiorite (TTG). S1 substage: small intrusive bodies with the oldest age (ca. 3450 Ma), e.g., Rooihoogte, Steynsdorp, Stolzburg, Theespruit, and other bodies; and 2 substage: larger bodies with younger ages (ca. 3230 Ma), e.g., Kaap Valley (KV in Fig. 5) and Nelshoogte plutons.
The syntectonic TTGs are closely associated with ultra-mafic to mafic volcanic rocks of the greenstone belt and intrude them concordantly. The largest Kaap Valley pluton is circular in form, and intrudes them discordantly. The S1 substage granitoids are mostly hornblende-biotite tonalites and are free of magnetite and ilmenite, but contain titanite that occurs as euhedral wedges and mostly subhedral crystals associated with mafic silicates, and which could have been converted from ilmenite. The color of the Z-axis of biotite is generally brown and similar to ilmenite-series biotites of Phanerozoic age. The S2 substage granitoids are also mostly magnetite-free, but the Kaap Valley pluton contains magnetite (up to 1.7% by volume), locally. The magnetite forms polygonal to granular crystals and contains small inclusions of ilmenite and chalcopyrite. In the most mafic intrusive phase, hematite blades occur along the 111 cleavage. However, the hematitization along the cleavage and margin of the crystals is not seen in the magnetite of most magnetite-bearing phases, implying lack of oxidation at the latest stage of their crystallization.
Late-tectonic calc-alkaline granitoids (ca. 3105 Ma); e.g., Nelspruit, Mpuluzi and Heerenveen (HV in Fig. 5). The late-tectonic calc-alkaline granites are batholithic in dimension, and biotite granite in composition, consisting largely of unzoned plagioclase, microcline, and/or orthoclase and quartz. Biotite is the dominant mafic silicate mineral and its Z-axis color is generally greenish brown but rarely green, as is typical for fresh magnetite-series granitoids of Phanerozoic age (Ishihara, 1977). The green biotite, which is found locally, could be an alteration product because it is associated with chloritized biotite. Flaky muscovite that is found in a few samples could also be an alteration product. Magnetite is composed of cubic to polygonal crystals associated with rare hematite blades or stringers. Ilmenite is absent, but titianite is abundant and forms both euhedral wedge-shaped crystals and anhedral aggregates associated with biotite.
Post-tectonic low-Ca (ca. 3070 Ma) and high-Ca (ca. 2700 Ma) granitoids. These rocks include an older, low-Ca biotite granite, and younger, high-Ca alkaline granitoids. The older granites are magnetite free, but the younger granitoids contain magnetite. The younger granitoids of the Boesmanskop and Salisburgkop stocks consist of minor quartz, alkaline amphiboles, and greenish-brown biotite, with more abundant zoned plagioclase and microcline. Euhedral titanite and polygonal to rounded magnetite are common, but ilmenite appears to be absent. No hematitization of the magnetite is observed.
In the central part of the Kaapvaal Craton, Paleoarchean granitoids similar to those found in the Barberton region form the Johannesburg Dome (∼50 km in diameter). Evidence has been presented elsewhere showing that two types of granitoid gneisses exist in this region. The oldest variety (ca. 3340 Ma; Poujol and Anhaeusser, 2001) consists of leuco-biotite trondhjemitic gneisses and associated migmatites that developed on the northern half of the dome (TTG in Fig. 6) (Anhaeusser, 1973, 1999). The younger variety, which is on the southern edge of the dome, consists of hornblende-biotite tonalitic gneisses that yielded a multiple zircon age of ca. 3170 Ma (Anhaeusser and Burger, 1982), and a more recent single zircon emplacement age of ca. 3201 Ma (Poujol and Anhaeusser, 2001).
Following the emplacement of trondhjemite-tonalite gneiss, another episode of magmatism took place on the Johannesburg Dome. This produced intrusions of Mesoarchean potassic granodiorites that occupy an area of batholithic dimension, extending across most of the southern portion of the dome (CA [calc-alkaline], Fig. 6). Two granodiorite phases have been distinguished: one on the southern and southeastern parts of the dome consists mainly of medium-grained, homogeneous, gray granodiorites, whereas a second variety, found mainly on the southwestern part of the dome, consists of porphyritic granodiorites (Anhaeusser, 1973). Zircons extracted from the two granodiorite types yielded ages of ca. 3121 Ma for the homogeneous variety and ca. 3114 Ma for the porphyritic variety (Poujol and Anhaeusser, 2001).
Numerous pegmatite dykes and veins crosscutting the granodiorites are younger than 3114 Ma and may represent the final stages of magmatism associated with a batholith emplacement ca. 3000 Ma.
Magnetic susceptibilities of 546 hand specimens from the Barberton region and 346 specimens from the Johannesburg Dome were measured using a Kappameter KT-3. Most of the samples used in this study were collected for earlier projects (Robb and Anhaeusser, 1983; Robb et al., 1983; Robb et al., 1986), although ∼30 samples were collected especially for this study.
Concentrations of major, trace, and rare earth elements (REE) were determined on seven samples by XRF methods for most elements, titrimetry for FeO, and ICP-MS method for REE. Concentrations of Cl and F, as well as the major element compositions of minerals, were determined using an electron microbe (EPMA) on 452 grains of biotite from eight representative granite samples. Concentrations of Cl, F, and S in apatite crystals (258 spot analyses) from two representative granite samples were also determined using an EPMA.
DISCUSSION OF THE ANALYSES
The magnetic susceptibility of the ∼900 samples range from 0.03 × 10−3 to 53 × 10−3 SI, indicating that both ilmenite series (<3 × 10−3 SI) and magnetite series (>3 × 10−3 SI) are present. Approximately 10% of the granitoid samples from the Barberton syntectonic plutons to ∼80% in the late-tectonic plutons belong to magnetite series (Fig. 7). Approximately 30% of the samples from the Johannesburg Dome are magnetite series.
Temporal and Spatial Distributions of Granitoid Types in the Barberton Region
In the Barberton region, all of the earliest S1, as well as S2, substage TTGs of the Nelshoogte pluton possess magnetic susceptibility values less than 3 × 10− 3 SI, indicating that they belong to very reduced ilmenite-series granitoids (Fig. 7). Anhaeusser and Robb (1980b) have shown that many of the TTGs interacted with and assimilated greenstone wall rocks that have Fe2O3/FeO (weight) ratios much less than 0.3 (Hunter, 1974; Hawkesworth and O'nions, 1977), suggesting that the reduced nature of the TTGs may have been acquired (and/or enhanced) during emplacement, rather than at the magma sources.
The S2 substage TTGs of the Kaap Valley tonalite pluton (Robb et al., 1986) likewise possess generally low magnetic susceptibility values, which suggests they are ilmenite series, but contain sporadic values higher than 3 × 10− 3 SI (Ishihara et al., 2002b). The distribution of magnetite-bearing rocks is so erratic that its genetic significance is difficult to evaluate. The post-tectonic, low-Ca biotite granites also appear to be ilmenite series, although the number of samples examined was small.
The late-tectonic calc-alkaline granitoids, which form large batholiths in the Barberton region, intrude the greenstones as well as the older TTGs (Anhaeusser and Robb, 1983). These granitoids, along with the post-tectonic high-Ca granitoids, generally possess magnetic susceptibilities higher than 3.0 × 10− 3 SI, indicating they are mostly magnetite series.
In the ca. 3105 Ma Nelspruit batholith, 86% of the 243 samples exhibit magnetic susceptibility values greater than 3 × 10− 3 SI (Fig. 7), which are characteristic of magnetite-series granitoids. The Nelspruit batholith is the largest granitic body in the region, and consists dominantly of homogeneous K-rich porphyritic granitic phases, associated potassic gneisses and migmatites that occur together with remnants of older Na-rich gneisses and greenstones, and a homogeneous, medium-grained, granodioritic phase (Robb et al., 1983).
The Mpuluzi batholith (Fig. 5) is the second largest granitoid massif in the region and consists of 53% magnetite series (n = 64). In contrast, the smaller Boesmanskop syenite pluton is 100% magnetite series (n = 13).
The post-tectonic high-Ca granitoids, found mainly in Swaziland, occur as episodic, discrete intrusions crosscutting all the other Archean rock types. They are hornblende-biotite alkaline granitoids and have high magnetic susceptibilities similar to alkaline granitoids.
Temporal and Spatial Distributions of Granitoid Types in the Johannesburg Dome
Within the Johannesburg Dome area (Fig. 6), the Archean granitoids comprise older TTGs (ca. 3300–3200 Ma) and younger calc-alkaline granitoids (ca. 3100–2900 Ma). Magnetic susceptibility measurements indicate that the TTGs consist mainly of ilmenite series (78% of the measurements) and partly magnetite series (22%); the calc-alkaline granitoids are also largely composed of ilmenite series (83%) with some local (17%) magnetite series (Ishihara et al., 2002a). Fe2O3/FeO ratios of bulk rock samples range from 0.05 to 0.72 (Ishihara et al., 2002a) and are generally below 0.5, implying a generally reduced nature for the Johannesburg Dome pluton.
The weakly magnetic granitoids tend to occur in the central part of the Johannesburg Dome around an undifferentiated phase (UD) at the interface between the TTG and calc-alkaline granitoids (Fig. 6).
Chemical Characteristics of Representative Granitoids
Partial chemical analyses were available for ∼60 of the samples. Plots of magnetic susceptibility versus SiO2 content (Fig. 8) show that the Nelspruit granitoids fall within the magnetite-series field (12 of 15 samples). The sporadic occurrence of magnetite-bearing rocks in the Kaap Valley TTGs are also illustrated; one and three samples of the 16 specimens fall within the magnetite- and intermediate-series fields, respectively. Thus, the Kaap Valley TTGs belong essentially to the ilmenite-series granitoids. All other TTGs also have very low magnetic susceptibility values, in contrast to the high values of sodic and/or adakitic granitoids of Phanerozoic ages (e.g., Tanzawa pluton) (Fig. 8).
Chemical compositions of the representative magnetite-series granitoids are presented in Table 1. A sample of magnetite-series Kaap Valley tonalite (LKV7) reveals a much higher Fe2O3/FeO ratio (1.35) than the magnetite-free TTGs of the Theespruit pluton (Fe2O3/FeO = 0.30–0.49), indicating a good correlation between magnetic susceptibility and Fe2O3/FeO ratio. Trace amounts of V and Cr, which may be substitutes for Fe2+ and Fe3+ in magnetite, respectively, are higher in the magnetite-bearing rocks than in the magnetite-free rocks.
Other TTG characteristics include high contents of Na2O and Sr, but low contents of K2O, Rb, and Ba (Robb and Anhaeusser, 1983) (Table 1). The LKV7 tonalite and 23-4 trondhjemite samples are high in Sr and low in Y, thus exhibiting adakitic characteristics (Drummond et al., 1996). Compared with the late-tectonic granitoids, the TTGs are rich in MgO, Al2O3, CaO, Ni, and Co (compare 23-4 and C31 in Table 1).
Magnetite-rich rocks of the Nelspruit and Mpuluzi batholiths have similarly high Fe2O3/FeO ratios (0.62–1.49). These batholiths, however, have differing feldspar chemistry; the Nelspruit granites are calcic, whereas the Mpuluzi granites are potassic. The Mpuluzi granites also have higher Rb, Sr, and Ba contents than the Nelspruit granites.
The granitoids generally display enrichments of light rare earth elements (LREE) and depletions of heavy rare earth elements (HREE) with respect to chondrite (Fig. 9). There are almost no Eu anomalies in the studied samples, except for the 23-5 Theespruit-TTG specimen, which, with a higher SiO2content and a higher Rb/Sr ratio compared to the other TTGs, represents a highly fractionated phase of the TTGs. The general REE patterns of the studied samples are similar to those of other TTGs (Robb et al., 1986) and garnet-bearing metamorphic protoliths; the HREE-depleted patterns are attributed to garnet and/or hornblende fractionation. The magnetite-series granitoids, particularly of the Mpuluzi potassic granite, are most enriched in REEs, whereas the ilmenite-series TTGs are generally most depleted in REE.
Magnetite crystals in the calc-alkaline magnetite-series granitoids tend to occur together with mafic silicates (mostly biotite). The Mg/Mg+Fe atomic ratios of biotites from the Nelspruit batholith (Table 2) are similar to those of typical magnetite-series biotite granite in the Sanin District, Japan (Fig. 10). The magnetite crystals contain rare hematite blades, which may have formed during the subsolidus stage, corresponding to a high degree of oxidation. Hematitization is only locally and weakly observed in the batholiths.
SIGNIFICANCE OF THE RESULTS
Genesis of Archean Granitoids in the Kaapvaal Craton
The Kaap Valley pluton contains local magnetite-bearing rocks (Figs. 7 and 8), and has, as a whole, higher Fe2O3/FeO ratios than the other TTGs. The whole-rock δ18O values of the Kaap Valley pluton are on average 2‰ higher than those of Quaternary low-K tholeiite (Faure and Harris, 1991). We interpret these data to indicate that the magmas for the Kaap Valley pluton were generated from partial melting of subducted seafloor basalt, which had increased its δ18O value by low-temperature interaction with seawater.
The Fe2O3/FeO ratios of oceanic basalts probably increased from ∼0.1 to ∼1 by interaction with O2- and/or SO4 2−−bearing seawater, much like the processes of seawater-rock interaction in modern oceanic crust (e.g., Shanks et al., 1981; Lécuyer and Ricard, 1999). The possibility of SO4 2–-rich Archean seawater is also suggested from the abundance of barite beds associated with felsic volcanic rocks in older greenstones (ca. 3500 Ma Onver-wacht Group) and later ferruginous sediments (ca. 3230 Ma Fig Tree Group). The geochemistry of banded iron-formations and associated rocks also led Ohmoto et al. (this volume) to conclude that the Archean oceans were sulfate-rich and generally oxygenated, except in local basins.
The late-tectonic calc-alkaline granites of the Nelspruit and Mpuluzi batholiths are also composed of typical magnetite-series rocks. They are thought to have been generated by the partial melting of earlier TTGs (Anhaeusser and Robb, 1983; Robb et al., 1983). The Fe2O3/FeO weight ratios of earlier TTG magmas may have increased to >0.5 by the same processes as the Kaapvaal plutons (i.e., the partial melting of altered and oxidized subducted mid-oceanic-ridge basalt). The magnetite-series calc-alkaline granites (e.g., Nelspruit batholith) may have been generated from such a TTG lower crust by the heat from the upper mantle and water from dehydration of the altered subducting oceanic crust.
Mineralization Associated with Archean Granitoids
Highly oxidized magnetite-series magmas, rather than ilmenite-series magmas, are more favorable sources for the generation of ore-forming fluids for Cu, Mo, Pb, Zn, Ag, and Au deposits (e.g., Burnham and Ohmoto, 1980). Many such examples are found in the circum-Pacific Rim (Ishihara, 1998). In the Barberton region, orogenic gold deposits that occur in shear zones in the greenstones are the most prominent metallic mineralization type present, but these gold deposits are not considered genetically related to the adjacent TTG and calc-alkaline granitic occurrences.
Along the northern flank of the Barberton greenstone belt, there is a broad temporal overlap between a mineralized, felsic porphyry (ca. 3126 Ma Fairview Mine porphyry; de Ronde et al., 1991) and the ca. 3105 Ma magnetite-series Nelspruit granite batholith (Kamo and Davis, 1994). But the late-tectonic Nelspruit granites may be too large in exposed dimensions, thus exposing deeper parts of the batholiths, to be genetically related to the porphyry that formed at a shallow crustal depth.
Several small cassiterite-bearing pegmatite dikes and veins occur in the Mpuluzi batholith close to the Swaziland border. Although tin mineralization is typically associated with reduced (ilmenite-series) and/or highly fractionated granites (Ishihara, 1981), the redox state of granite related to cassiterite mineralization in the Mpuluzi batholith is unknown.
The TTGs of the Barberton region are adakitic and similar to those in the Au-Cu mineralized regions of the Philippines (Sajona and Maury, 1998), northern Chile (Oyarzun et al., 2001), and Kitakami Mountains (Tsuchiya and Kanisawa, 1994; Ishihara and Murakami, 2004). However, an important difference is their low redox state, which may have limited the concentration of sulfur, and hence S-combined ore metals, to the magmatic fluids (Burnham and Ohmoto, 1980; Ishihara et al., 1988).
Oxidizing conditions of porphyry Cu-related magmas are represented by high SO3 contents of apatites, such as those in the Philippines (Imai, 2001, 2002, 2004). Chlorine is an important metal carrier in magmatic fluids, and the Cl content of rock-forming apatite is a good indicator of the Cl contents of granitic magmas (Nedachi, 1980). EPMA analyses indicate that the SO3 contents of apatite crystals in magnetite-series granitoids from the Barberton region approach 0.36 wt.% (Table 3), which is similar to the value of porphyry Cu-related intrusions from the Philippines (Imai, 2001, 2002). However, Cl contents of apatites from the Barberton granitoids are extremely low (below 0.1%) (Fig. 11). These low Cl concentrations in the magnetite-series granitoids were probably an important reason for the scarcity of S-combined metallic mineralizations in the Barberton region.
Porphyry Mineralization in Earth's History
Porphyry Cu–Mo and Cu-Au deposits are typically associated with oxidized magnetite-series magmatism, whereas Sn-W deposits generally occur with reduced ilmenite-series granitoids (Fig. 1). Because mineral exploration has been carried out much more extensively than regional granitoid-series studies, ore deposit data may provide a clue to the distributions of the two types of granitoids in Earth's history.
A histogram of the total Cu tonnages in all the discovered porphyry-type deposits, plotted against their geologic age, is shown in Figure 12. It may suggest that the abundance of magnetite-series granitoids, as approximated by porphyry Cu deposits, increased with younger geologic age. The Cenozoic peak in Figure 12 comes mostly from young and super-large deposits at Grasberg-Ertzberg (3 Ma, 28 million tons [MT] Cu), El Teniente (4 Ma, 56.8 MT Cu), Andina (4 Ma, 37.2 MT Cu), Los Pelambres (10 Ma, 20.5 MT Cu), Chuquicamata (31–34 Ma, 55.2 MT Cu), Escondida (37–38 Ma, 34.3 MT Cu), and Bingham (39–40 Ma, 23.2 MT Cu) (the age and metal tonnage statistics are from Watanabe, 2003). Quaternary volcanoes of the circum-Pacific belt are composed predominantly of magnetite series, and porphyry-type deposits are thought to be forming below these volcanoes (e.g., Mt. Pinatubo; Imai et al., 1993, 1996; Hattori and Keith, 2001).
Most porphyryrelated hydrothermal deposits are formed at shallow depths, typically less than ∼5 km (Burnham and Ohmoto, 1980). Therefore, the observed trends of decreasing amounts of Cu deposits with older geologic age are likely due to preservation effects. Pillow lavas and high-level granitoids are still locally preserved in the Precambrian terrains. Copper deposits at Malanjkhand, central India (2400 Ma, 6.5 MT Cu), may or may not be porphyry-type (Sarkar et al., 1996; Panigrahi and Mookherjee, 1997). However, the oldest (3300 Ma) Mo-Cu deposit at the Coppin Gap, Pilbara Craton (102 MT; 0.152% Cu, 0.105% Mo; Jones, 1990), is a large porphyry-type deposit. A reconnaissance magnetic survey of the porphyry indicates that it is intermediate between ilmenite and magnetite series. These data, therefore, suggest that the formation of porphyry Cu-Mo and Cu-Au deposits, as well as magnetite-series granitoids, have occurred since at least ∼3.3 billion years ago.
Sn deposits, which are typically associated with ilmenite-series granitoids, are found mostly in the Phanerozoic rock (Meyer, 1985). They are concentrated in the late Paleozoic (southwestern Europe) and Triassic–Jurassic (Southeast Asia), reflecting the predominance of ilmenite-series granitoids in these regions. Cenozoic-age Sn-bearing granitoids are limited to the Bolivian Miocene granitic belt and Miocene Outer Zone granitoids of southwestern Japan. Scarcity of the late Cenozoic Sn deposits is likely due to paucity of ilmenite-series granitoids during this period in the circum-Pacific region.
Sizable Sn deposits are known in Precambrian terrains. The late Proterozoic granites (1.0 Ga), of Rondonia, Brazil (Bettencourt and Dall'angol, 1995), and early Proterozoic granites (2.1 Ga) of the Bushveld complex host many cassiterite-pegmatite-greisen-veins. Archean granites also host cassiterite associated with Li and Ta pegmatites. At Greenbush, Western Australia, Sn-Ta-Li pegmatite sheet (2.5 Ga) produced in 1888–1987 the tin concentrate of 26,000 tons (∼72% Sn), Ta2O5 concentrate of 2300 tons, low iron spodumene concentrate (72% Li2O) of 39,700 tons and crude kaolin ore of 9500 tons (Hatcher and Clynick, 1990). Ta-pegmatite of the Mount Cassiterite orebody (2.9 Ga) in the Pilbara Craton contains 8785 tons Sn metal (Huston et al., 2001). Archean pegmatite in Zimbabwe is said to contain cassiterite with 114,000 tons Sn metal. Cassiterite mineralization therefore appears to be concentrated in Precambrian time, yet the examples are still too small to draw a definite conclusion.
Amitsoq-gray tonalite gneisses from West Greenland (ca. 3700 Ma) have Fe2O3/FeO ratios that average 0.33 (n = 23; Nutman and Bridgwater, 1986), thus being ilmenite series. However, because these tonalites were highly metamorphosed, and metamorphism generally lowers the Fe2O3/FeO ratios of rocks by reactions with graphite (Ohmoto and Kerrick, 1977), the original tonalites may have been intermediate or magnetite series. Archean granitoids from the Kaapvaal Craton that are older than ca. 3230 Ma are mostly reduced-type (ilmenite series) with a Fe2O3/FeO weight ratio of 0.3–0.5. Because their Fe2O3/FeO ratios appear to have been lowered by the assimilation of greenstones, however, their original magmas might also have been more oxidized than normal ilmenite-series magmas. The 3230 Ma Kaap Valley pluton and younger granitoids are mostly magnetite series, suggesting that by ca. 3200 Ma, oxidized protoliths existed in the Kaapvaal continental crust.
Source rocks for the Archean TTGs, as suggested by their REE patterns and geochemical characteristics, appear to have been subducted oceanic crust that was subjected to a garnet-bearing metamorphic grade. The late-tectonic calc-alkaline granitoids may have formed from the partial melting of such older Archean TTGs. Although these rocks form part of the magnetite series, they have low Cl contents. The deep erosion level of the granitoids and low Cl contents of the magmas were probably the primary reasons why no metallic deposits are associated with magnetite-series granitoids in the Barberton region.
Porphyry-type Cu (-Au or -Mo) deposits are the most representative mineralization associated with oxidized, magnetite-series magmatism. The oldest example may be the 3300 Ma Coppin Gap in the Pilbara Craton. The presence of such Cu deposits and the abundance of magnetite-series granitoids in the Kaapvaal Craton suggest that the processes of seawater-rock interaction at mid-oceanic ridges and granite magma generation in subduction zones, as well as the redox chemistry of ocean water, have been basically the same since at least ca. 3200 Ma.
We are grateful to D.A. Singer for providing the world copper tonnage statistics and to S.E. Kesler, B.R. Frost, and E.J. Essene for helpful comments and suggestions on earlier manuscripts. Ohmoto acknowledges support from the National Science Foundation (EAR-9706279, EAR-0229556), the NASA Astrobiology Institute (NCC2-1057, CA#NNA04CC06A), and the NASA Exobiology Program (CA#NNG04GK00G).
Figures & Tables
Evolution of Early Earth's Atmosphere, Hydrosphere, and Biosphere - Constraints from Ore Deposits
- calc-alkalic composition
- chemical composition
- copper ores
- ferric iron
- ferrous iron
- gold ores
- greenstone belts
- igneous rocks
- isotope ratios
- Kaapvaal Craton
- magnetic properties
- magnetic susceptibility
- metal ores
- metamorphic belts
- molybdenum ores
- plutonic rocks
- porphyry copper
- porphyry molybdenum
- rare earths
- Southern Africa
- stable isotopes