Journal of Petrology Advance Access originally published online on September 20, 2008
Journal of Petrology 2008 49(10):1729-1753; doi:10.1093/petrology/egn044
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The Petrogenesis and Mantle Source of Archaean Ferropicrites from the Western Superior Province, Ontario, Canada
Department of earth and Planetary Sciences, Mcgill University, 3450 University Street, Room 238, Montreal, que., Canada H3A 2A7
RECEIVED JANUARY 11, 2008; ACCEPTED AUGUST 21, 2008
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGEOLOGICAL SETTINGANALYTICAL METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Archaean ferropicrites have been under-appreciated in the past because they have been frequently misidentified as enriched komatiites or Al-depleted komatiites. To investigate the nature of Archaean ferropicrite magmatism, we sampled ferropicrites from the Steep Rock, Lumby Lake, Grassy Portage Bay, and Dayohessarah Lake greenstone belts in the Western Superior Province, Ontario, Canada. Ferropicrite samples that are thought to approximate liquid compositions have
18 wt % Fe2O3* at 19 wt % MgO, and frequently contain less than 5 wt % Al2O3. They are enriched in Ti and high field strength elements relative to komatiites, and have fractionated trace element profiles (La/Yb 11). These distinctive geochemical characteristics require that ferropicrites and komatiites have different mantle sources, with that of the ferropicrites being Fe- and incompatible element-enriched compared with that of komatiites. A consideration of recent 5 GPa melting experiments on pyrolite and Fe-rich Martian mantle compositions indicates that Archaean ferropicrites could be generated by melting of an olivine-dominated mantle source with a Mg-number of 85 at 5 GPa. The high densities calculated for the ferropicrite magmas (e.g. 3·33 g/cm3) suggest that more Fe-rich magmas would have difficulty rising to the Earth's surface and would tend to stagnate or sink within the mantle. KEY WORDS: Archaean; ferropicrite; Fe-rich; komatiite; mantle; petrogenesis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGANALYTICAL METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Primitive mantle-derived magmas in the Archaean include komatiites, ferropicrites, and more contentiously boninites. Komatiites are extrusive rocks containing more than 18 wt % MgO, high concentrations of Ni and Cr (Arndt & Nisbet, 1982
), and remarkably consistent FeO* contents at 11 wt %. Al-depleted komatiites have relatively low Al contents, moderate–low levels of incompatible trace elements, and exhibit slight heavy rare earth element (HREE) depletions (Arndt, 2003). Al-undepleted komatiites have higher Al contents, low concentrations of incompatible trace elements, and exhibit flat to slightly light REE (LREE)-depleted trace element profiles (Arndt, 2003). The combination of high MgO contents and low levels of incompatible elements indicates that komatiites are the result of moderate to large degrees (30–60%) of partial melting of a relatively fertile lherzolite source (Ohtani, 1988, 1990; Herzberg, 1992; Kerr et al., 1996; Sproule et al., 2002; Arndt, 2003; Dostal & Mueller, 2004; Grove & Parman, 2004).
Ferropicrites are a distinct variety of Archaean mantle-derived magma that differ from komatiites in that they have significantly higher FeO* contents (18 wt % vs 11 wt % FeO* for komatiites at similar MgO contents). Ferropicrites have been defined by Hanski & Smolkin (1989) as picrites with FeO* 
14 wt %, and we will adopt this criterion for the purpose of this paper. Ferropicrites are enriched in Ti (TiO2 = 1–2 wt %) and incompatible trace elements (Nb = 10–17 ppm; Nb/La = 0·8–1·3) relative to komatiites, and have fractionated REE profiles (La/Yb = 8–18), which are chemical characteristics reminiscent of modern ocean island basalts (OIB). Ferropicrites have similar MgO contents to komatiites (19 wt %) but significantly lower Al2O3 contents (<10 wt %, often <5 wt %). These geochemical differences require a different petrogenesis for ferropicrites, possibly including a unique mantle source composition, as well as differing conditions of pressure, temperature, and degree of partial melting.
Ferropicrites were first recognized by Hanski & Smolkin (1989, 1995) in the Palaeoproterozoic Pechenga volcanic belt in the Kola Peninsula of the Fennoscandian Shield (Hanski, 1992). Since then, many Archaean primitive volcanic occurrences have been recognized as being ferropicritic in character, including those in the Vermillion Belt, Minnesota, in the Kolar Schist Belt, South India, in the Boston Township, Ontario, and at Lake of the Enemy in the Northwest Territories (Green & Schulz, 1977; Schulz, 1982; Rajamani et al., 1985, 1989; Stone et al., 1995; Francis et al., 1999). Despite the significant major and trace element differences between Archaean ferropicrites and komatiites, some researchers still classify ferropicrites as simply a variety of komatiite (Polat et al., 1999; Tomlinson et al., 1999; Sproule et al., 2002; Arndt et al., 2008). Many more ferropicrite occurrences may go unrecognized, however, because they are commonly assumed to be a variety of komatiite. For example, the ferropicrites of this study have variously been called enriched komatiites and Al-depleted komatiites. Tomlinson et al. (1999) and Schaefer & Morton (1991) referred to the Dismal Ashrock Formation of the Steep Rock greenstone belt as ‘enriched basaltic komatiites’ and a ‘komatiitic pyroclastic unit’ respectively. Similarly, Tomlinson (1996) and Tomlinson et al. (1999), and Wyman & Hollings (1998) termed a pyroclastic unit within the Lumby Lake greenstone belt as ‘enriched basaltic komatiites’ and ‘Al-depleted komatiites’, respectively. Sproule et al. (2002) referred to these lithologies from Steep Rock and Lumby Lake as ‘Ti-enriched–Al-depleted komatiites’. Schaefer & Morton (1991) described the ferropicritic tuff of the Rice Bay Dome succession at Rainy Lake as ‘komatiitic lapilli tuff and volcanic breccia’. Finally, Polat et al. (1999) called the ferropicritic intrusive rocks within the Dayohessarah Lake greenstone belt ‘Al-depleted komatiites’.
In this paper, we report the results of a study of ferropicrites from the Western Superior Province, in Ontario, Canada, to gain insight into the nature of their mantle source. Ferropicrite samples were obtained from the Steep Rock and Lumby Lake greenstone belts located in the southern portion of the Wabigoon subprovince, from the Grassy Portage Bay of Rainy Lake, part of the Rice Bay Dome in the northern Quetico subprovince, and from the Dayohessarah Lake greenstone belt in the northeastern portion of the Wawa subprovince (Fig. 1). As all of these rocks are Archaean in age, they have been metamorphosed to greenschist or amphibolite facies. After a brief description of their metamorphic mineralogies and textures, we will focus our attention on their igneous protoliths and will refer to these rocks by their igneous protolith names. Our results indicate that Archaean ferropicrites cannot be derived from the mantle source that generated komatiites but are derived from a previously unrecognized Archaean mantle reservoir.
|
|
GEOLOGICAL SETTING |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGANALYTICAL METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The Steep Rock greenstone belt
The Dismal Ashrock is the fourth of five formations that make up the Steep Rock Group, which unconformably overlies the 3003 ± 5 Ma Marmion tonalitic batholith (Davis & Jackson, 1988
; Wilks & Nisbet, 1988; Stone et al., 1992). The Dismal Ashrock is, however, relatively low in the stratigraphic succession of the belt, and is overlain by the 1–5 km thick Witch Bay Formation, a dominantly mafic metavolcanic unit (Jolliffe, 1955; Shklanka, 1972; Wilks & Nisbet, 1988; Schaefer & Morton, 1991; Stone et al., 1992; Tomlinson, 1996; Tomlinson et al., 1999). The Dismal Ashrock is underlain by the Wagita Formation at the base of the Steep Rock Group, a 150 m thick clastic sedimentary unit consisting of metaconglomerate, metasandstone, and pelite, followed upward by the Mosher Carbonate, a 0–500 m thick dark bluish grey limestone containing diverse stromatolite morphologies, and finally the Jolliffe Ore Zone, a 100–400 m thick goethite and hematite brecciated ore and an unbrecciated goethite–chert banded iron formation (BIF), which directly underlies the Dismal Ashrock (Jolliffe, 1955; Wilks & Nisbet, 1988; Tomlinson et al., 1999). Ferropicrites in the Steep Rock belt occur as a 50–400 m thick pyroclastic volcanic unit called the Dismal Ashrock Formation. According to Stone et al. (1992), the Dismal Ashrock consists of clast-supported, dark green to black, lapilli-sized aphanitic fragments (composed of chlorite, ferristilpnomelane, and magnetite) in a fine-grained matrix of similar colour (composed of actinolite, chlorite, talc, iron oxides, and calcite) with minor calcite amygdules. Our samples of Dismal Ashrock were taken from outcrops with large clasts, minimal carbonate, and minimal shearing. In addition, the Dismal Ashrock contains numerous olive–green weathering, ovoid nodules (2 cm–7 m) that are very fine-grained and homogeneous with dark grey fresh surfaces. These extremely fine-grained nodules consist of actinolite, chlorite, sericite and other clay minerals, and contain plagioclase microphenocrysts. Nodules greater than 15 cm in diameter and free of surrounding Dismal Ashrock were selected for analysis.
The Lumby Lake greenstone belt
The Steep Rock and Lumby Lake belts are connected by the Finlayson greenstone belt, and the stratigraphies of the Steep Rock and Lumby Lake belts can be correlated (Fralick & King, 1996; Wyman & Hollings, 1998; Tomlinson et al., 1999). Both the Steep Rock and Lumby Lake belts overlie the Marmion tonalitic batholith and have a maximum age of 2973 Ma (Davis & Jackson, 1988; Tomlinson et al., 1999), and both belts have been metamorphosed to greenschist facies conditions (Schaefer & Morton, 1991; Stone et al., 1992). The Lumby Lake belt is dominated by mafic metavolcanic rocks, but also includes felsic metavolcanic rocks, rare spinifex-textured komatiite lava flows, BIF, clastic metasediments, and possibly stromatolitic marble (Jackson, 1985; Jackson & Chevalier, 1985; Davis & Jackson, 1988; Wyman & Hollings, 1998; Tomlinson et al., 1999). Ferropicrites at Lumby Lake occur as a pyroclastic unit that closely resembles the Dismal Ashrock Formation of Steep Rock, except that it is more pervasively sheared; the two occurrences will be collectively referred to as the Dismal Ashrock in this paper. Unlike in the Steep Rock greenstone belt, the Dismal Ashrock occurs relatively high in the Lumby Lake stratigraphy, locally overlying a 10 m thick silica-rich BIF and underlying amygdaloidal mafic lava flows. It has also been reported to occur locally interbedded with thin carbonates and iron formation in other parts of the Lumby Lake belt (Tomlinson et al., 1999).
The Grassy Portage Bay, Rainy Lake area
The Rice Bay Dome is located near Rainy Lake, and its Archaean stratigraphy consists of the Keewatin Volcanics and the Coutchiching metasediments. The Keewatin Volcanics are mafic to intermediate in composition with minor felsic metavolcanic rocks and intrusive rocks (Harris, 1974; Schaefer & Morton, 1991), and the oldest igneous rocks in the area range from 2725 to 2728 Ma (Davis et al., 1989; Schaefer & Morton, 1991). The Keewatin Volcanics are overlain by the Coutchiching metasedimentary rocks (Poulsen et al., 1980; Stone et al., 1992), which include the Rice Bay, Bears Passage, and Southern metasediments in the Rainy Lake area (Harris, 1974; Schaefer & Morton, 1991). Ferropicrites of the Rainy Lake area occur as the Grassy Portage Bay ultramafic pyroclastic unit (GUP) on the outer eastern edge of the Rice Bay Dome (Harris, 1974; Schaefer & Morton, 1991). The GUP is a fine-grained actinolite–chlorite schist with 15% magnetite, and it has a distinctive fragmental nature. In contrast to the black Dismal Ashrock at Steep Rock and Lumby Lake, the GUP has a striking green colour in the field, and it is reported to also comprise a dark grey volcanic breccia that is up to 800 km thick (Schaefer & Morton, 1991). The GUP is interpreted to represent magnetite-bearing lapilli to ash tuffs that have been metamorphosed to amphibolite facies. Most of our samples were selected from the ash tuff facies, which is more abundant and homogeneous than the lapilli tuff.
The Dayohessarah Lake greenstone belt
The stratigraphy of the 2·7 Ga Dayohessarah Lake greenstone belt begins with a basal unit of pillowed and massive basalts (Polat et al., 1999; Stott, 1999). These mafic volcanic rocks are overlain by komatiite flows, followed by intermediate to felsic flows and pyroclastic deposits, and finally conglomerates and greywackes, which form the uppermost unit exposed in the centre of the greenstone belt (Williams et al., 1991; Stott et al., 1994; Stott, 1999). The Dayohessarah Lake belt has been interpreted by some to be a south-plunging syncline (Stott et al., 1995a, 1995b), and it has experienced variable degrees of metamorphism, with both greenschist and amphibolite facies being evident within the belt (Williams et al., 1991; Stott et al., 1994; Polat et al., 1999).
Ferropicrite intrusions occur as a series of small plugs along the synclinal axis of the Dayohessarah Lake belt (Stott, 1999). These rocks consist of serpentine, chlorite, actinolite, opaque minerals, and clay minerals, and contain rare relict olivine crystals. The largest of these intrusions is zoned from a thick olivine pyroxenite to pyroxenite rim enclosing a core of magnetite-gabbro, gabbro, and leucogabbro. This zonation, combined with the coarse equigranular texture of the rocks, indicates that these bodies are intrusions whose rocks are in part or entirely cumulates. These coarse-grained mafic to ultramafic bodies have traditionally been interpreted by Stott (1999) to underlie the komatiite flows, possibly acting as feeder dykes. Others have interpreted these intrusive rocks to be Al-depleted komatiites, in contrast to the Al-undepleted komatiites with spinifex texture (Polat et al., 1999). The temporal relationship between the spinifex-textured komatiites and the ferropicrite intrusions, however, remains uncertain, and the fact that the intrusions crosscut the youngest rocks in the centre of the belt suggests that they are relatively late, post-dating the underlying spinifex-textured komatiites. The samples for this study include coarse-grained peridotites, olivine pyroxenites, and magnetite-gabbros from the small ferropicrite intrusions along the synclinal axis of the belt, as well as samples of chilled margins, spinifex, and basal zones of komatiite flows.
|
ANALYTICAL METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGANALYTICAL METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Samples with minimal or no alteration were crushed in a steel case-hardened jaw crusher and ground in an alumina puck grinder. Major and trace elements were analyzed by X-ray fluorescence (XRF) at the Geochemical Laboratories at McGill University, using a Philips PW2440 4kW automated XRF spectrometer system. Major elements, Ni, Co, Cr, and V were analyzed using 32 mm diameter fused beads, whereas Sc, Rb, Sr, Zr, Nb, and Y were analyzed using 40 mm diameter pressed pellets. The accuracy for silica is within 0·5% absolute. The accuracy for other major and trace elements is estimated to be within 1% relative when the concentration of a particular element exceeds its quantitation limit (defined as 3·33 times the detection limit).
Rare earth elements (REE), Hf, Ta, Th, and U were analyzed by Activation Laboratories; REE were also analyzed by the McGill Geochemical Laboratories. At Activation Laboratories, samples were analyzed by inductively coupled-plasma mass spectrometry (ICP-MS) using a Perkin Elmer SCIEX ELAN 6000 system, and using a lithium metaborate–lithium carbonate fusion technique for digestion. At the McGill Geochemical Laboratories, samples were analyzed using a Perkin Elmer Elan 6100 DRCplus ICP-MS system by lithium metaborate fusion decomposition. Precision for this technique is 1% and accuracy is better than 5%.
To determine the contribution of carbonate material to the whole-rock chemistry of the pyroclastic ferropicrites from the Steep Rock belt, seven samples were hand-picked to select clasts with minimal carbonate, treated with acetic acid to dissolve the carbonate, and then reanalyzed for major, trace, and rare earth elements. Acetic acid, rather than hydrochloric acid, was used because acetic acid dissolves carbonate without leaching any of the trace elements from the igneous material. Grains of crushed sample material with minimal visible carbonate were hand-picked under a microscope and then treated with excess 20% acetic acid at room temperature. After the acidic supernatant was decanted, the samples were washed five times with water and then rinsed three times with denatured alcohol before XRF and ICP-MS analysis.
Samples for Nd analysis were crushed to powder form and dissolved with a HF–HNO3 mixture in high-pressure Teflon containers. A 149Sm–150Sm tracer was added to determine the Nd and Sm concentrations. The REE were concentrated by cation exchange chromatography and the Sm and Nd were extracted using an orthophosphoric acid-coated Teflon powder according to the method of Richard et al. (1976). Sm and Nd isotopic ratios were measured on a VG SECTOR-54 mass spectrometer, using a triple filament assembly, in the GEOTOP Laboratories at Université du Québec à Montréal. Repeated measurements of the La Jolla Nd standard yielded a value of 143Nd/144 Nd = 0·511849 ± 12 (n = 21). The total combined blank for Sm and Nd is less than 150 pg. The reported Sm and Nd concentrations and the 147Sm/144Nd ratios have accuracies of 0·5%, corresponding to an average error of 0·5 
Nd units for the final Nd isotopic composition. 146Nd/144Nd was normalized to 0·7219 for mass fractionation correction. The reference value for 143Nd/144NdCHUR was taken to be 0·512638, whereas that of 147Sm/144NdCHUR was taken to be 0·1967, and the decay constant for 147Sm was assumed to be 6·54 x 10–12 a–1.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGANALYTICAL METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Pyroclastic ferropicrites
The ferropicritic samples collected from the four greenstone belts comprise two types of rocks: pyroclastic volcanic rocks and coarse-grained intrusive rocks. The pyroclastic ferropicrites from Steep Rock, Lumby Lake, and Grassy Portage Bay have similar chemical compositions but different mineralogies. The black lapilli tuffs from Steep Rock and Lumby Lake consist dominantly of talc with some tremolite, and although these rocks are not magnetic they contain
30% opaque minerals, some of which are sulfides, such as pyrite, that are visible in hand sample. The distinctly green ash tuffs from Grassy Portage Bay consist of chlorite and actinolite in roughly equal proportions, and as these rocks are magnetic, the 15% opaque minerals are assumed to be magnetite. These pyroclastic ferropicrites have Mg contents comparable with those of komatiites (17·6 ± 3·2 wt % MgO), but they are enriched in Fe (18·2 ± 1·4 wt % Fe2O3*; Fig. 2) and Ti (1·51 ± 0·30 wt % TiO2), and depleted in Al (5·70 ± 1·4 wt % Al2O3; Fig. 3) relative to komatiites from Dayohessarah Lake (representative samples shown in Table 1). Their Al contents remain essentially constant with increasing Mg content. They have Al2O3/TiO2 ratios of 4, and the compositions of these rocks scatter equally between the picrite and Al-depleted picrite fields in the classification diagram for ultramafic rocks developed by Hanski et al. (2001). They are enriched in high field strength elements (HFSE) and have higher incompatible element ratios, such as Zr/Y (6; Fig. 4), than komatiites, although they display a significant range in Zr contents as well as Zr/Y ratios. Samples from Steep Rock and Lumby Lake have higher Zr/Y ratios than samples from Grassy Portage Bay at comparable Zr contents (Fig. 4). The pyroclastic ferropicrites are also enriched in LREE, and display smooth, fractionated REE profiles (La/Yb 11; Fig. 5) with a distinct HREE depletion (Gd/Yb 3). As such, their trace element profiles are broadly similar to those of mildly alkaline olivine basalts, such as those found at modern hotspots (Fig. 6). The CIPW normative mineralogy (calculated throughout this study with XFe3+ = 0·1 x FeTotal) of the pyroclastic ferropicrites is dominated by olivine (31%) and clinopyroxene (25%) at 1 atm, and although most samples are orthopyroxene-normative, a few are feldspathoid-normative. The presence of a significant amount of normative orthopyroxene indicates that these rocks have subalkaline rather than alkaline affinities.
|
|
|
|
|
|
The results for the samples leached with acetic acid are shown in Table 2. The concentrations of most major element oxides and trace elements increased by about 10%, consistent with the mass of sample lost due to dissolution. However, the concentrations of CaO, K2O, and MnO, as well as Rb, Sr, and Y decreased significantly by 46%, 20%, 15%, 9%, 75%, and 19% respectively. REE concentrations decreased by
20%, and loss on ignition (LOI) values were also significantly lower after the leaching process (Table 2). Despite these losses, however, the shape of their REE profiles before and after acid-leaching are virtually identical (Fig. 7).
|
|
Comparison of the pyroclastic ferropicrites from the Western Superior with other Archaean ferropicrites
Several key geochemical features distinguish the ferropicrites of this study from other Archaean ferropicrites, specifically the Vermillion, Kolar, Boston Township, and Lake of the Enemy occurrences, all of which are
2·7 Ga in age (Schulz, 1982; Rajamani et al., 1989; Stone et al., 1995; Francis et al., 1999). Most significantly, samples of the ferropicrites from Steep Rock, Lumby Lake, and Grassy Portage Bay are significantly higher in Mg (19 wt % MgO) than these other Archaean ferropicrite occurrences (8–17 wt % MgO; Fig. 8). The pyroclastic ferropicrites of this study are also significantly lower in Al and Si than other Archaean ferropicrites (Fig. 9). The Western Superior ferropicrites do, however, have similar enrichments in incompatible trace elements to other Archaean ferropicrites, with similar fractionated REE profiles. The Western Superior ferropicrites have higher Zr/Y ratios than the Boston Township and Kolar occurrences at comparable Zr contents, whereas the Lake of the Enemy rocks have similar Zr/Y ratios, but at significantly higher Zr contents. Thus the Western Superior ferropicrites appear to represent a compositional extreme that is more magnesian than other reported ferropicrite occurrences.
|
|
The green nodules
The ovoid, olive–green weathering nodules hosted by the pyroclastic ferropicrites at Steep Rock are dark grey on fresh surfaces, very fine-grained, massive, and homogeneous, with no observable relict igneous textures in hand sample. Even under the petrographic microscope, these rocks are very fine-grained, and they appear to consist of actinolite, chlorite, sericite and other clay minerals, and they contain some relict plagioclase microphenocrysts which are lath- and diamond-shaped. Their whole-rock compositions are similar to that of their host ferropicrite (Table 1), with 17·9 ± 1·4 wt % MgO at slightly lower Fe contents (Fig. 2), but they have significantly higher Al contents (Al2O3 = 11·7 ± 0·6 wt %; Fig. 3) and lower Ca and Ti contents (CaO = 3·5 ± 0·5 wt %; TiO2 = 0·46 ± 0·08 wt %) relative to their ferropicrite host. The green nodules are enriched in LREE with smooth, convex-up LREE- and HREE-depleted chondrite-normalized REE profiles (La/Yb = 16; Gd/Yb = 4) that are nearly identical to those of their ferropicrite host (Fig. 5). Although they have much lower Nb and Ta contents (Nb/La = 0·2) than their host, which produces negative Nb and Ta anomalies in a chondrite-normalized multi-element diagram (Fig. 6), they are generally enriched in HFSE relative to komatiites (Figs 4 and 6). Like their host, the nodules have a significant negative Sr anomaly. The nodules also have significantly lower Ni (275 ± 35 ppm; Fig. 10) and Cr contents (950 ± 215 ppm) compared with their host, despite having similar MgO and Fe2O3* contents. Their CIPW normative mineralogy is dominated by olivine (47%) and feldspar (43%; Table 1).
|
Ferropicritic intrusions
Samples of the ferropicritic intrusions at Dayohessarah Lake consist of 1–2 mm crystals of serpentine, chlorite, actinolite, clay minerals, and opaque minerals, and contain rare relict olivine crystals. These metamorphic rocks, which represent coarse-grained peridotite, olivine pyroxenite, and pyroxenite, are significantly more magnesian (MgO = 28·0 ± 2·3 wt %; Table 3) than the pyroclastic rocks from Steep Rock, Lumby Lake, and Grassy Portage Bay, but have similar Fe contents (Fe2O3* = 17·5 ± 1·8 wt %; Fig. 2). They have lower Ti and Al contents (TiO2 = 0·63 ± 0·14 wt %; Al2O3 = 3·7 ± 0·68 wt %; Fig. 3) than those of the pyroclastic rocks. They have Al2O3/TiO2 ratios of
5, which is similar to those of the pyroclastic ferropicrites. Their trace element profiles are less fractionated than those of the pyroclastic rocks, with Zr/Y ratios of 5 (Fig. 4) and La/Yb ratios of 4 (Figs 5 and 6). The coarse-grained peridotites have the highest abundances of normative olivine (52%), with lesser orthopyroxene (22%) and clinopyroxene (10%; Table 3).
|
Komatiites
Komatiite flows from Dayohessarah Lake have olivine-rich bases overlain by spinifex zones and fine-grained chilled margins. Rocks from the chilled margins and fine-grained spinifex zones, which consist of tremolite–actinolite, serpentine, and chlorite, with 1–2% opaque minerals, have similar Mg contents (MgO = 17·1 ± 3·3 wt %; Table 1) and low Fe and Ti contents (Fe2O3* = 12·0 ± 1·1 wt %; Fig. 2; TiO2 = 0·60 ± 0·14 wt %) compared with the pyroclastic ferropicrites. The Dayohessarah Lake komatiites are typical Al-undepleted komatiites, with significantly higher Al contents (10·4 ± 1·4 wt % Al2O3; Fig. 3) than the pyroclastic ferropicrites, and Al2O3/TiO2 ratios of
18. These rocks plot, however, in the Al-depleted komatiite and Ti-enriched komatiite fields of Hanski et al.'s (2001) classification diagram, but with a strong preference for the former. In contrast to the ferropicrites, Al decreases steadily with increasing Mg content in the samples of chilled margins and rocks with fine-grained spinifex. These samples have less than half the Ni content of the pyroclastic ferropicrites (570 ± 200 ppm and 1170 ± 200 ppm, respectively), but they contain slightly higher Cr contents (1830 ± 830 ppm and 1660 ± 270 ppm, respectively). These komatiite samples have low concentrations of HFSE and LREE (Zr/Y 2·5; Fig. 4), especially compared with the ferropicrites, and their chondrite-normalized REE profiles are almost flat (La/Yb = 2) with slightly higher HREE abundances than the ferropicrites (Gd/Yb = 1·6; Fig. 5). The normative mineralogy of samples with fine-grained spinifex and samples of chilled margins is dominated by plagioclase (35%) and olivine (25%), with lesser amounts of clinopyroxene (20%) and orthopyroxene (17%). Four relatively coarse-grained komatiite samples with Mg contents typical of the fine-grained spinifex rocks and chilled margin samples have significantly lower Al contents and are more depleted in HREE than the other samples. These samples are interpreted to represent cumulates of pyroxene, rather than rare occurrences of Al-depleted komatiite within an Al-undepleted komatiite unit.
The ‘fish-roe’ textured olivine-rich bases of the komatiite flows, consisting of closely packed 0·5–1 mm rounded to subrounded olivine pseudomorphs, today contain serpentine, actinolite, up to 5% opaque minerals, and relict crystals of both olivine and pyroxene. Samples from the bases of the komatiite flows have significantly higher Mg contents (27·5 ± 4·5 wt % MgO) and normative olivine contents than samples of fine-grained spinifex and chilled margins (Table 3; Fig. 2), but similar Fe contents (10·7 ± 1·4 wt % Fe2O3*). They have lower Al and Ti contents than samples of fine-grained spinifex and chilled margins (Al2O3 = 5·5 ± 2·1 wt %; Fig. 3; TiO2 = 0·29 ± 0·09 wt %), but they have identical Al2O3/TiO2 ratios of 18. They have the lowest concentrations of incompatible trace elements, and their REE profiles are almost flat (La/Yb = 2; Gd/Yb = 1·5; Fig. 5), similar in shape to those of samples of fine-grained spinifex and chilled margins (Zr/Y 2·7; Fig. 4). Samples from basal zones of komatiite flows have lower Ni contents than the ferropicritic intrusive rocks (1600 ± 560 ppm vs 1790 ± 490 ppm), but they contain slightly higher Cr contents (2700 ± 910 ppm vs 2550 ± 260 ppm). These rocks are also dominated by normative olivine (36%) and orthopyroxene (35%), with lesser clinopyroxene (9%) at 1 atm (Table 3).
Nd isotopic analyses
Five ferropicrite samples, including one green nodule, and five komatiite samples have been analyzed for their Nd isotope composition (Table 4). Four komatiite samples yielded Nd values at 2·7 Ga that range from +3·2 to +4·2 (av. = +3·6), but one Nd value was significantly lower (+1·9). The ferropicrites from Dayohessarah Lake have Nd values of +3·3 to +3·5 at 2·7 Ga (Polat et al., 1999). Two ferropicrite samples from Steep Rock have Nd values at 3·0 Ga of +0·3 and +5·6, and the green nodule, SR-14, has an Nd value of +0·5 at 3·0 Ga. Two ferropicrite samples from Grassy Portage Bay have Nd values of +2·3 and +2·5 at 2·7 Ga.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGANALYTICAL METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Assessment of alteration and crustal contamination
A comparison of the compositions of leached and unleached samples of pyroclastic ferropicrite from the Steep Rock belt indicates that most major and trace elements are present in the silicate portion of the samples (Table 2). CaO, K2O, MnO, Sr, Rb, and Y were, however, significantly lower in the leached samples, indicating that a significant proportion of these elements was present in carbonate. Thus the negative Sr anomaly seen in many samples probably reflects the mobility of Sr. The Y and REE concentrations were also lower (
20%) in the leached samples (Table 2), but the fact that the REE profiles for leached and unleached samples are parallel indicates that the REE have been relatively immobile during alteration and metamorphism. This interpretation is supported by the similar trace element profiles of samples of pyroclastic ferropicrite from Grassy Portage Bay, which do not contain carbonate. Overall, the results of the acid-leaching procedure indicate that, with the exception of Sr, the majority of the incompatible trace elements are contained within the silicate portion of the ferropicrite, and thus reflect their magmatic origin rather than a subsequent contribution from secondary carbonate-bearing fluids. The low Na and K concentrations and negative Sr anomalies of all the rocks studied (Fig. 6) indicate that these elements, as well as other fluid-soluble large ion lithophile elements (LILE), were mobile during alteration or metamorphism. Although the trace element enrichment of the ferropicrite samples is broadly similar to that of the upper continental crust, their low silica contents argue against significant contamination by granitoid crust. The ferropicrites are also rich in HFSE and lack the negative Nb, Ta, and Eu anomalies characteristic of crustal granitoids. Furthermore, the HREE profiles of crustal granitoids (from Dy to Lu) are flat and higher than those of the HREE-depleted ferropicrites. These features indicate that the ferropicrites did not gain their trace element signatures from crustal granitoids.
The physical association of the pyroclastic ferropicrite with iron formation at Steep Rock and Lumby Lake raises the possibility that their high Fe contents and incompatible element enrichments reflect interaction with the underlying iron formation. The ferropicrites are, however, also enriched in Ti (Fig. 10), which is not characteristic of iron formations (Klein & Ladeira, 2004; Frei & Polat, 2007; O’Neil et al., 2007). Furthermore, the pyroclastic ferropicrites from Grassy Portage Bay that are not associated with iron formation have nearly identical Fe and incompatible element contents to those of the pyroclastic ferropicrites from Steep Rock. Thus the incompatible element enrichment seen in the ferropicrites is deemed to be magmatic and sourced in the mantle.
Ferropicrite petrogenesis
The fine-grained, homogeneous nature of the clasts in the pyroclastic ferropicrites suggests that they are clasts of a magma, and the compositions of the pyroclastic ferropicrites from Steep Rock, Lumby Lake, and Grassy Portage Bay are interpreted to approximate liquid compositions. Thus we have grouped them together as ferropicrite liquids. The coarse-grained ferropicrite samples from the intrusive bodies at Dayohessarah Lake have significantly higher MgO contents than the pyroclastic rocks from the other three belts and are interpreted to be cumulates of olivine ± pyroxene; they will henceforth be referred to as ferropicrite cumulates. Similarly, the bases of the komatiite flows from Dayohessarah Lake are interpreted to be olivine cumulates, and will be referred to as komatiite cumulates. The chilled margins and fine-grained spinifex zones of komatiite flows are thought to approach liquid compositions and will be referred to as komatiite liquids. The few samples with uncertain field contexts were classified as cumulates or liquids on the basis of their MgO contents compared with those of well-characterized samples.
Pressure–temperature phase diagrams were constructed (Fig. 11) for representative samples of ferropicrite and komatiite liquids using MELTS (Ghiorso & Sack, 1995; Asimow & Ghiorso, 1998; Ghiorso et al., 2002). The calculated liquidus phase for both the ferropicrite and komatiite is olivine at 1 atm, and changes to orthopyroxene and then clinopyroxene with increasing pressure. Clinopyroxene becomes the ferropicrite liquidus phase at 1·5 GPa, compared with 2·5 GPa for the komatiite. The stability field of spinel in the ferropicrite is significantly larger than that in the komatiite, with spinel appearing 150–200°C above the solidus at all pressures (except 3·0 GPa) in the ferropicrite, but only between 0·5 and 1·5 GPa in the komatiite. The larger spinel field in the ferropicrite is probably due to its significantly higher Fe and Ti contents relative to the komatiite. In contrast, the stability field of garnet is much larger for the komatiite than the ferropicrite, consistent with the former's higher Al content. The high Fe and low Al contents of the ferropicrite liquids compared with those of the komatiite liquids (Figs 2 and 3) could reflect a variety of factors, including different melting conditions and different mantle source compositions from those that generated komatiites. Although pMELTS is known to have limitations, especially when applied to relatively exotic compositions such as ferropicrites at high pressures, it is the significant differences between the komatiite and ferropicrite phase diagrams that are emphasized here, rather than the absolute accuracy of the phase diagrams.
|
Liquidus temperatures were calculated for ferropicrites and komatiites with MgO
10 wt % assuming dry bulk compositions. Ferropicrite liquidus temperatures averaged 1452°C with an olivine composition of Fo87 (Table 5), whereas komatiite liquidus temperatures averaged 1407°C, with an olivine composition of Fo91. Although the ferropicrite liquidus olivines have lower Fo contents than those of komatiites, consistent with their higher Fe contents, the ferropicrites have Mg contents and liquidus temperatures that are higher than the komatiites. Even at similar MgO contents, the ferropicrites have higher liquidus temperatures despite the lower Fo content of their liquidus olivine (Table 6). The higher incompatible element contents of the ferropicrites relative to the komatiites may, however, indicate higher volatile contents, which would depress the estimated liquidus temperature.
|
|
The ferropicrites define a Ni vs Mg-number trend in which Ni rises asymptotically at a Mg-number of
83, whereas the Ni content of the komatiites rises asymptotically at a Mg-number of 90 (Fig. 12). This observation indicates that the Mg-number of the ferropicrite mantle source was lower than that of the komatiite source. Furthermore, the ferropicrite liquids have more than double the Ni content of the komatiite liquids, and the ferropicrite cumulates have slightly higher Ni contents than the komatiite cumulates, consistent with the higher normative olivine contents of the ferropicrite liquids and cumulates relative to the komatiite liquids and cumulates. Conversely, the higher Cr contents of the komatiite liquids and cumulates relative to ferropicrites is consistent with their higher normative pyroxene contents, as Cr is compatible in pyroxene. It should be noted, however, that Cr-spinel is an important phase in controlling the Cr content of liquids with 10–20 wt % MgO. The normative mineralogy of the ferropicrite liquids and cumulates is dominated by olivine, whereas feldspar dominates the normative mineralogy of the komatiite liquids, and pyroxenes dominate that of the komatiite cumulates.
|
In summary, the ferropicrite liquids have similar liquidus temperatures, higher Ni contents, and higher normative olivine contents than roughly coeval komatiite liquids, indicating that the ferropicrite liquids are as primitive as the komatiite liquids. The high MgO contents of the ferropicrite liquids require that they were generated by the melting of an olivine-dominated mantle source. A ferropicrite liquid with a Mg-number of 65 would coexist with an olivine-dominated mantle source with a Mg-number of 84, whereas a komatiite liquid with a Mg-number of 73 would coexist with an olivine-dominated source with a Mg-number of 88, assuming an Fe–Mg KD (olivine–liquid) of 0·36 (Agee & Draper, 2004
).
The significantly higher Zr/Y and La/Yb ratios of the ferropicrite liquids compared with the Dayohessarah Lake ferropicrite cumulates (Figs 4 and 5) means that the latter have crystallized from liquids with lower Zr/Y and La/Yb ratios than those from Steep Rock, Lumby Lake, and Grassy Portage Bay because crystal fractionation cannot fractionate incompatible trace elements significantly in mafic magmas. Similarly, the crystal fractionation of olivine and pyroxene cannot produce the range of Zr/Y ratios seen within the ferropicrite liquids (Fig. 13). The Steep Rock and Lumby Lake ferropicrites exhibit higher Zr/Y ratios than those from Grassy Portage Bay at comparable Zr contents, a feature that is inconsistent with the relatively narrow range in major element compositions of all the ferropicrite liquids. Partial melting (5–20%, batch or aggregate fractional) of a mantle source with the Zr and Y contents of Primitive Mantle (10 ppm and 4 ppm, respectively) can reproduce the trend of the Grassy Portage Bay ferropicrite liquids, but at lower Zr/Y ratios (Fig. 13), and thus the variation in Zr/Y of the Grassy Portage Bay ferropicrites could reflect melting of a source with a Zr/Y ratio of 4. Varying degrees of partial melting cannot, however, produce the differences between Grassy Portage Bay and the Steep Rock and Lumby Lake ferropicrites, nor the rapid rise in Zr/Y with increasing Zr within the Steep Rock and Lumby Lake ferropicrites. This indicates that the mantle source(s) for these ferropicrites was variably enriched in trace elements.
|
The trace element and Nd isotopic similarities between the green nodules and their pyroclastic ferropicrite host suggest that they are genetically related. The fact that the normative mineralogy of the green nodules is dominated by olivine and feldspar at 1 atm, however, indicates that they might represent troctolite cumulates from a ferropicritic magma that is more evolved than the ferropicrite liquids. Magnetite-gabbros of ferropicritic affinity from the Dayohessarah Lake belt have Mg-numbers (
45) which would be in equilibrium with the green nodules, assuming that the former approximate liquid compositions and assuming an Fe–Mg KD of 0·34. The significantly lower Ni and Cr contents of the green nodules relative to their ferropicrite host are also consistent with crystallization from such an evolved ferropicritic liquid. An alternate possibility is that the green nodules represent magma clasts, as indicated by their very fine-grained textures, whose compositions reflect interaction between komatiitic and ferropicritic magmas, although the Steep Rock belt does not presently contain komatiites.
Several models have attempted to explain the petrogenesis of ferropicrites (Gibson, 2002). Jakobsen et al. (2005) and Veksler et al. (2006) proposed liquid immiscibility documented in the Skaergaard intrusion as a mechanism of producing high-Fe liquids. Veksler et al. (2006) proposed that such Fe-rich liquids would have high CaO/Al2O3 ratios, high Ti, HFSE, and REE contents, and would be too dense to erupt, remaining in magma chambers to mix with new pulses of parental magma. The unmixing of silicate liquids into Fe-rich and Si-rich liquids, however, has been observed only in evolved basaltic magmas with immiscibility typically occurring between 1000 and 1150°C, after about 80–95% crystallization of a basaltic parental magma (Philpotts, 1982; De, 1999; Jakobsen et al., 2005; Veksler et al., 2006). The ferropicrites have liquidus temperatures of 1450°C, and no evidence exists for liquid immiscibility in silicate systems at such high temperatures. More complicated models in which an Fe-rich immiscible liquid (30 wt % FeO; McBirney & Nakamura, 1974; Philpotts, 1982; Jakobsen et al., 2005) mixes with a komatiitic liquid (10 wt % FeO) fail because the resulting mixture would have significantly less MgO than the ferropicrites because of the low Mg content of the immiscible Fe-rich melt. Furthermore, the Western Superior ferropicrites are not associated with felsic rocks that would represent the immiscible Si-rich liquids, and this argues against silicate liquid immiscibility as a mechanism of producing ferropicritic magmas.
Melting of a two-component mantle source of peridotite hosting enriched domains, such as small-degree partial melts or a recycled eclogite component, has been proposed by several workers as a mechanism for increasing the Fe and incompatible trace element concentrations in melts while maintaining high Mg contents (Stone et al., 1995; Polat et al., 1999; Gibson, 2002; Tuff et al., 2005). Tuff et al. (2005) conducted melting experiments on a ferropicrite from the base of the Parana–Etandeka flood basalt province and proposed a mantle source composed of 75% eclogite and 25% peridotite. The low Al2O3 content (9 wt %), fractionated HREE (Gd/Yb = 3), and high Ni content (660 ppm) of their ferropicrite sample were assumed to indicate that garnet was a residual phase in its mantle source, and that olivine was completely consumed during melting. Their experiments found that garnet is the liquidus phase with clinopyroxene at P 6 GPa, leading Tuff et al. to propose garnet pyroxenite as the mantle source of their ferropicrite. No experimental melts of peridotite–basalt mixtures have reproduced the Fe contents of the Western Superior ferropicrites (Yaxley & Green, 1998; Yaxley, 2000), however, and no combination of KLB-1 and Fe-rich basalt is capable of producing both their Fe and Mg contents (Fig. 14). Thus melts of a two-component peridotite–basalt mixture fail to reproduce the Fe contents of ferropicrites, unless one of the two components is significantly more iron-rich than any modern analogue. Furthermore, a comparison of known garnet pyroxenite xenoliths with ferropicritic liquids (Fig. 15) shows that the majority of garnet pyroxenites have insufficient Mg and Fe to produce both the high Mg and Fe contents of ferropicrites, emphasizing the need for an olivine-dominated mantle source.
|
|
Some workers have proposed melting of an Fe-rich mantle source as a method of producing the high Fe contents of ferropicritic liquids (Rajamani et al., 1985
; Hanski & Smolkin, 1995; Francis et al., 1999; Gibson et al., 2000). Melting experiments on Fe-rich mantle xenolith HK-66 (Mg-number = 85) at 3·0 GPa produced melts with 13 wt % FeO and 12 wt % Al2O3, with FeO increasing and Al2O3 decreasing with increasing pressure (Hirose & Kushiro, 1993). These FeO and Al2O3 contents are significantly lower and higher, respectively, than those of the ferropicrite liquids of this study. Thus, melting of a Mg-number 85 mantle source at 3·0 GPa (Hirose & Kushiro, 1993) cannot reproduce the high Fe contents of ferropicrites, and either higher pressure or a more Fe-rich mantle source is needed. High-pressure melting experiments on ferropicrites (Tuff et al., 2005) and model Martian mantle compositions (Bertka & Holloway, 1994; Agee & Draper, 2004) can be used to constrain the composition of the mantle source of Archaean ferropicrites. In these experiments, the Fe–Mg KD for olivine–liquid ranges from 0·27 to 0·33 at 1 atm and increases to values approaching 0·48 at pressures of 5 GPa (Table 7). The dependence of KD on pressure (Ulmer, 1989) means that constraining the pressure at which ferropicritic magmas form is critical to determining the Fe content of their mantle source. If a mantle source such as KLB-1, with 8 wt % FeO, melts at 1·5 GPa to produce a liquid with 18 wt % MgO, assuming a KD of 0·33, the liquid will have 11 wt % FeO. If the same mantle source melts at 5 GPa to produce an 18 wt % MgO liquid, assuming a KD of 0·36, the liquid will have 10 wt % FeO. Employing the same KD values, the mantle source of the ferropicrite liquids would have 13 and 14 wt % FeO at 1·5 and 5 GPa, respectively, if the MgO content of the mantle source is assumed to be that of KLB-1 (39·2 wt % MgO). No combination of temperature and pressure (up to 20 GPa) in melting experiments on peridotite with a Mg-number of 89 can produce the Fe contents of the ferropicritic liquids of this study (Takahashi, 1986; Herzberg et al., 1990; Zhang & Herzberg, 1994; Herzberg & Zhang, 1996; Walter, 1998).
|
Recent interest in Mars has motivated melting experiments on more Fe-rich peridotites that mimic the estimated Martian mantle composition. Melting experiments on DW-Mars, an iron-rich (FeO = 17·9 wt %; Mg-number 75) spinel lherzolite (Dreibus & Wanke, 1985
) conducted at 1·5 GPa (Bertka & Holloway, 1994) produced melts with FeO contents ranging from 16 to 20 wt %, and Mg-numbers ranging from 48 to 71, similar to those of the Western Superior ferropicrite liquids (Fig. 16). However, these experimental melts contain up to 13 wt % Al2O3, and Al decreases with increasing Mg content, a feature not seen in the ferropicrite data. Furthermore, in Al–Si space, the experimental melts of DW-Mars exhibit a large range in Al content at nearly constant Si content, whereas the ferropicrite liquids have nearly constant Al over a range of Si contents (Fig. 17). These trends in Mg–Al and Al–Si space indicate that ferropicrites cannot be produced by the melting of an Fe-rich source, such as DW-Mars, at 1·5 GPa.
|
|
Agee & Draper (2004
) conducted melting experiments on the ordinary chondrite Homestead L5 (Mg-number 77), as a model Martian mantle composition. Melts with up to 30 wt % FeO were produced at 4·7 and 5 GPa and temperatures at and above 1650°C, depending on the degree of partial melting. Although these FeO contents are significantly higher than those of the Western Superior ferropicrite liquids (Fig. 16), the 5 GPa experimental melts of Homestead have similar low Al contents to the Western Superior ferropicrite liquids, and a similar trend of constant Al with varying Si and Mg (Figs 17 and 18; Agee & Draper, 2004). In contrast, the Al contents of both the experimental melts of KLB-1 and the Dayohessarah Lake komatiite liquids decrease steadily with increasing Mg (Takahashi, 1986; Herzberg & Zhang, 1996). It is possible to estimate the Mg-number of a possible mantle source for the Western Superior ferropicrites at 5 GPa by interpolating between the Fe contents of liquids produced by partial melting of KLB-1 and Homestead at 5 GPa. The position of the ferropicrite liquids in Mg–Fe space (Fig. 16) would indicate a source with a Mg-number of 85, if they were produced at 5 GPa. Although melting experiments on HK-66 (Mg-number 85) have been conducted only up to 3 GPa, the increase in Fe content with increasing pressure in the existing experimental data extrapolates to the Fe contents of the Western Superior ferropicrite liquids at pressures of 5 GPa (Fig. 16).
|
The nature of the Archaean ferropicrite mantle source
The Archaean ferropicrites from the Western Superior Province had an Fe-rich (Mg-number of
85) mantle source that was dominated by olivine and enriched in incompatible trace elements compared with that of coeval komatiites. The Mg, Fe, Al, and Si systematics of these ferropicrite liquids are best reproduced by melting at 5 GPa; melting at lower pressures would require a mantle source with a lower Mg-number. The positive Nd values of the Western Superior ferropicrite samples combined with the positive Nd values typical of Archaean ferropicrites (Stone et al., 1995; Francis et al., 1999; Polat et al., 1999) indicate that this Fe-rich mantle source was isotopically depleted relative to CHUR, but not as depleted as the mid-ocean ridge basalt (MORB) source (Depleted Mantle of DePaolo, 1981) at 2·7 Ga (Nd + 2·2). In comparison, komatiites from the 2·7 Ga Dayohessarah Lake belt were derived from a mantle source that was more isotopically depleted than that of the ferropicrites and Depleted Mantle at 2·7 Ga.
Unless ferropicrites represent very small degrees of partial melting, their fractionated trace element profiles require a mantle source that is enriched in trace elements, in contrast to that of the komatiite liquids. The relatively depleted HREE + Y seen in the ferropicrites would traditionally be interpreted to indicate residual garnet in the mantle source or garnet fractionation prior to eruption. Similarly, the low Al content in magmas such as Al-depleted komatiites is traditionally interpreted to indicate the presence of garnet in the mantle residue (Arndt et al., 2008). A mantle source with abundant garnet, such as a garnet pyroxenite, would be consistent with the low Al and HREE contents of the ferropicrites. Furthermore, in the case of a garnet pyroxenite source, the absence of olivine would explain the high Ni contents of the ferropicrites (Sobolev et al., 2005). However, the Al contents of the ferropicrite liquids are very low (6 wt % Al2O3) and the stability field of garnet is sufficiently small (Fig. 11) that garnet does not appear on the liquidus at pressures up to 3·0 GPa (Fig. 11a). An alternate interpretation is that the high Ni contents of the ferropicrites could reflect a high degree of partial melting of an olivine-dominated source, with their low Al and normative garnet contents, along with their depleted HREE, indicating that their Fe-rich mantle source contained little or no garnet. Such an interpretation would require that the garnet signature of the ferropicrites was a characteristic of their mantle source, predating the melting event that produced the ferropicrites, a conclusion also reached by Hanski & Smolkin (1995) when explaining the origin of the Pechenga ferropicrites. This interpretation is also consistent with high-pressure experimental data on the Etendeka ferropicrite produced by Tuff et al. (2005). One possibility for such a source would be an olivine flotation cumulate that crystallized from a deep-seated Hadean magma ocean that had already experienced significant garnet fractionation. Such an olivine cumulate would have a lower Mg-number than Primitive Mantle and would be low in Al. The interstitial liquid between the olivine crystals would give the cumulate a fractionated trace element profile produced by garnet fractionation. This explanation for the garnet signature in Archaean ferropicrites is analogous to that for the plagioclase signature in the trace element profiles of the lunar mare basalts.
Physical implications of partial melting at 5 GPa in the Archaean mantle
The liquidus temperatures of Archaean komatiites (1500–1650°C) correspond to mantle potential temperatures of up to 1900°C (Campbell et al., 1989; Nisbet et al., 1993). This temperature is 300°C above estimated ambient Archaean mantle temperatures predicted from secular cooling models, which are in turn about 200–300°C higher than the potential temperature of the modern ambient mantle (Campbell et al., 1989; Nisbet et al., 1993). Although a more Fe-rich peridotite will have a lower solidus temperature than KLB-1, a more Al-poor composition will have a higher solidus temperature at any given pressure, and thus the dry KLB-1 peridotite solidus is taken as that of the ferropicrite mantle source for the purpose of constraining the P–T conditions of ferropicrite petrogenesis. The Mg, Fe, Al, and Si systematics of the Western Superior Archaean ferropicrites are best reproduced by experiments at 5 GPa and thus a temperature of at least 1680°C, compared with the modern mantle adiabat (1350°C) that intersects the dry peridotite solidus at 1370°C (Fig. 19). The difference between the estimated Archaean 1630°C mantle adiabat and the modern 1350°C mantle adiabat is 280°C, consistent with previous estimates of 300°C based on komatiites. Adiabatically ascending mantle that intersects the dry peridotite solidus at 5 GPa and 1680°C would reach the surface at a temperature of 1630°C, nearly 200°C hotter than the mean calculated liquidus temperature (1448°C) of the ferropicrite liquids. This temperature difference is attributed to several factors. First, although the KLB-1 solidus at 5 GPa is 1680°C, it is unlikely that liquids produced by melting at 5 GPa are this hot, owing to the latent heat of fusion, and thus the predicted temperature at which these liquids would reach the surface (1630°C) is a maximum. Other factors that would contribute to this temperature difference are the higher Fe–Mg KD (olivine–liquid) and the higher Mg partition coefficient with increasing pressure.
|
The densities of the Western Superior ferropicrite and komatiite liquids can be estimated using the Perple_X software (Connolly, 2006
) at 5 GPa and 1700°C, assuming no volatiles are present. The presence of volatiles would decrease the density of the liquids. The mean density for the ferropicrite liquids is 3·33 ± 0·038 g/cm3, whereas that of the komatiite liquids is significantly lower at 3·18 ± 0·031 g/cm3 because of their lower Fe contents. According to the Preliminary Reference Earth Model (PREM; Dziewonski & Anderson, 1981), the density of the mantle at 150 km depth (5 GPa) is 3·36 g/cm3. At this depth, the ferropicritic melts would be neutrally buoyant compared with the surrounding mantle, whereas the komatiitic melts would be positively buoyant. With increasing depth, the density of liquids increases faster than that of the surrounding mantle, because of the greater compressibility of liquids than solids, such that ferropicritic melts produced at greater depths would be negatively buoyant and sink deeper into the mantle. The densities of the Western Superior ferropicrites approach the maximum density of melts that can rise to the surface; melts with Fe contents greater than those of the Western Superior ferropicrites (21 wt % Fe2O3) will be negatively buoyant with respect to the surrounding mantle, and sink into the mantle, rather than ascend to the Earth's surface. |
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGANALYTICAL METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The significance and abundance of Archaean ferropicrites have been under-appreciated in the past because they have frequently been identified as a variety of komatiite. Our results from four greenstone belts of the Western Superior Province indicate that ferropicrites represent a distinct variety of Archaean magma that is enriched in Fe and incompatible trace elements compared with Archaean komatiites, and that ferropicrites were relatively widespread in the Archaean, possibly as abundant as alkaline basalts in the modern realm. An interpolation of melting experiments on pyrolite and model Martian mantle compositions indicates that Archaean ferropicrite magmas could have been generated by melting of an olivine-dominated, trace element enriched mantle source with a Mg-number of
85, at 5 GPa, and may have contained little to no garnet. The Western Superior ferropicrites may be the most Fe-rich magmas that could reach the surface of the Earth, as more Fe-rich magmas would be negatively buoyant and sink within the mantle. |
SUPPLEMENTARY DATA |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGANALYTICAL METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Supplementary data for this paper are available at Journal of Petrology online.
|
ACKNOWLEDGEMENTS |
|---|
We would like to thank Oliver Fu, Witold Ciolkiewicz, and Alexandra Kirshner for assistance in the field. Whole-rock XRF analyses and acetic acid leaching procedures were performed by Glenna Keating and Tariq Ahmedali at the Geochemical Laboratories at McGill University. REE were analyzed by ICP-MS by Bill Minarik at the Geochemical Laboratories and by Activation Laboratories in Toronto. Thin sections were made by George Panagiotidis at McGill University, Vancouver Petrographics, and Vancouver GeoTech Laboratories. Nd isotopic analyses were conducted by Jonathan O’Neil at GEOTOP-UQAM-McGill in Montreal. This research was funded by the Natural Science and Engineering Research Council of Canada (NSERC) in the form of a CGS-M scholarship to S.B.G. and a Discovery Grant to D.F. (RGPIN 7977-00), by a Richard H. Tomlinson Master's Fellowship from McGill University, and by a Sir James Laugheed Award of Distinction from Alberta Learning.
*Corresponding author. Telephone: 514-398-4885. Fax: 514-398-4680. E-mail: donf{at}eps.mcgill.ca
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL SETTINGANALYTICAL METHODSRESULTSDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Agee CB, Draper DS. Experimental constraints on the origin of Martian meteorites and the composition of the Martian mantle. Earth and Planetary Science Letters (2004) 224:415–429.[CrossRef][Web of Science]
Arndt N. Komatiites, kimberlites, and boninites. Journal of Geophysical Research (2003) 108:2293–2303.[CrossRef]
Arndt NT, Nisbet EG. What is a komatiite? In: Komatiites—Arndt NT, Nisbet EG, eds. (1982) London: Allen & Unwin. 526.
Arndt NT, Barnes SJ, Lesher CM. Petrogenesis of komatiite. In: Komatiite (2008) Cambridge: Cambridge University Press. (in press).
Asimow PD, Ghiorso MS. Algorithmic modifications extending MELTS to calculate subsolidus phase relations. American Mineralogist (1998) 83:1127–1131.[Abstract]
Bertka CM, Holloway JR. Anhydrous partial melting of an iron-rich mantle II: primary melt compositions at 15 kbar. Contributions to Mineralogy and Petrology (1994) 115:323–338.[CrossRef][Web of Science]
Campbell IH, Griffiths RW, Hill RI. Melting in an Archaean mantle plume: heads it's basalts, tails it's komatiites. Nature (1989) 339:697–699.[CrossRef][Web of Science]
Connolly J. Perple_X. Zurich. (2006) http://www.perplex.ethz.ch/.
Davis DW, Jackson MC. Geochronology of the Lumby Lake greenstone belt: a 3 Ga complex within the Wabigoon Subprovince, northwest Ontario. Geological Society of America Bulletin (1988) 100:818–824.
Davis DW, Poulsen KH, Kamo L. New insights into Archean crustal development from geochronology in the Rainy Lake Area, Superior Province, Canada. Journal of Geology (1989) 97:379–398.[Web of Science]
De A. Silicate liquid immiscibility in the Deccan Traps and its petrogenetic significance. Memoir, Geological Society of India (1999) 42:727–734.
DePaolo DJ. A neodymium and strontium isotopic study of the Mesozoic calc-alkaline granitic batholiths of the Sierra Nevada and Peninsular Ranges, California. Journal of Geophysical Research (1981) 86:10470–10488.
Dostal J, Mueller WU. Komatiite geochemistry. In: The Precambrian Earth: Tempos and Events—Erickson PG, Altermann W, Nelson DR, Mueller WU, Catuneanu O, eds. (2004) Amsterdam: Elsevier. 290–298.
Dreibus G, Wanke H. Mars, a volatile-rich planet. Meteoritics (1985) 20:367–382.
Dziewonski AM, Anderson DL. Preliminary Reference Earth Model. Physics of the Earth and Planetary Interiors (1981) 25:297–356.[CrossRef][Web of Science]
Fralick P, King D. Meso-Archean evolution of Western Superior Province: evidence from metasedimentary sequences near Atikokan. (1996) Vancouver, BC: Lithoprobe Secretariat University of British Columbia.
Francis D, Ludden J. The mantle source for olivine nephelinite, basanite, and alkaline olivine basalt at Fort Selkirk, Yukon, Canada. Journal of Petrology (1990) 31:371–400.
Francis D, Ludden J, Johnstone R, Davis W. Picrite evidence for more Fe in Archean mantle reservoirs. Earth and Planetary Science Letters (1999) 167:197–213.[CrossRef][Web of Science]
Frei R, Polat A. Source heterogeneity for the major components of 3·7 Ga banded iron formations (Isua Greenstone Belt, Western Greenland): tracing the nature of interacting water masses in BIF formation. Earth and Planetary Science Letters (2007) 253:266–281.[CrossRef][Web of Science]
Ghiorso MS, Sack RO. Chemical mass transfer in magmatic processes. IV. A revised and internally consistent thermodynamic model for the interpolation and extrapolation of liquid–solid equilibria in magmatic systems at elevated temperatures and pressures. Contributions to Mineralogy and Petrology (1995) 119:197–212.[Web of Science]
Ghiorso MS, Hirschmann MM, Reiners PW, Kress VCI. The pMELTS: A revision of MELTS aimed at improving calculation of phase relations and major element partitioning involved in partial melting of the mantle at pressures up to 3 GPa. Geochemistry, Geophysics, Geosystems (2002) 3. 10.1029/2001GC000217.
Gibson SA. Major element heterogeneity in Archean to Recent mantle plume starting heads. Earth and Planetary Science Letters (2002) 195:59–74.[CrossRef][Web of Science]
Gibson SA, Thompson RN, Dickin AP. Ferropicrites: geochemical evidence for Fe-rich streaks in upwelling mantle plumes. Earth and Planetary Science Letters (2000) 174:355–374.[CrossRef][Web of Science]
Green JC, Schulz KJ. Iron-rich basaltic komatiites in the early Precambrian Vermilion District, Minnesota. Canadian Journal of Earth Sciences (1977) 14:2181–2192.
Grove TL, Parman SW. Thermal evolution of the Earth as recorded by komatiites. Earth and Planetary Science Letters (2004) 219:173–187.[CrossRef][Web of Science]
Hanski EJ. Petrology of the Pechenga ferropicrites and cogenetic, Ni-bearing gabbro–wehrlite intrusions, Kola Peninsula, Russia. Bulletin of the Geological Survey of Finland (1992) 367:192.
Hanski EJ, Smolkin VF. Pechenga ferropicrites and other early Proterozoic picrites in the eastern part of the Baltic Shield. Precambrian Research (1989) 45:63–82.[CrossRef][Web of Science]
Hanski EJ, Smolkin VF. Iron- and LREE-enriched mantle source for early Proterozoic intraplate magmatism as exemplified by the Pechenga ferropicrites, Kola Peninsula, Russia. Lithos (1995) 34:107–125.[CrossRef][Web of Science]
Hanski EJ, Huhma H, Rastas P, Kamenetsky VS. The Palaeoproterozoic komatiite–picrite association of Finnish Lapland. Journal of Petrology (2001) 42:855–876.
Harris FR. Geology of the Rainy Lake Area, District of Rainy River (1974) 94 pp: Toronto: Ontario Division of Mines.
Herzberg C. Depth and degree of melting of komatiites. Journal of Geophysical Research (1992) 97:4521–4540.
Herzberg C, Zhang J. Melting experiments on anhydrous peridotite KLB-1: compositions of magmas in the upper mantle and transition zone. Journal of Geophysical Research (1996) 101:8271–8295.[CrossRef]
Herzberg C, Gasparik T, Sawamoto H. Origin of mantle peridotite: constraints from melting experiments to 16·5 GPa. Journal of Geophysical Research (1990) 95:15779–15803.[CrossRef]
Hirose K, Kushiro I. Partial melting of dry peridotites at high pressures: determination of compositions of melts segregated from peridotite using aggregates of diamond. Earth and Planetary Science Letters (1993) 114:477–489.[CrossRef][Web of Science]
Jackson MC. Precambrian Geology of the Lumby Lake Area, Eastern Part, Kenora and Rainy River Districts. Ontario Geological Survey, Preliminary Map P.2829, scale 1:15 840. Geology, 1982 (1985).
Jackson MC, Chevalier P. Precambrian Geology of the Lumby Lake Area, Western Part, Kenora District. Ontario Geological Survey, Preliminary Map P·2828, scale 1:15 840. Geology, 1983 (1985).
Jakobsen JK, Veksler IV, Tegner C, Brooks CK. Immiscible iron- and silica-rich melts in basalt petrogenesis documented in the Skaergaard intrusion. Geology (2005) 33:885–888.
Jolliffe AW. Geology and iron ores of Steep Rock Lake. Economic Geology (1955) 50:373–398.
Kerr AC, Marriner GF, Arndt NT, Tarney J, Nivia A, Saunders AD, Duncan RA. The petrogenesis of Gorgona komatiites, picrites, and basalts: new field, petrographic, and geochemical constraints. Lithos (1996) 37:245–260.[CrossRef][Web of Science]
Klein C, Ladeira E. Geochemistry and mineralogy of Neoproterozoic banded iron-formations and some selected, siliceous manganese formations from the Urucum District, Mato Grosso Do Sul, Brazil. Economic Geology (2004) 99:1233–1244.
McBirney AR, Nakamura Y. Immiscibility in late-stage magmas of the Skaergaard Intrusion. Carnegie Institution of Washington Yearbook (1974) 73:348–352.
Nisbet EG, Cheadle MJ, Arndt NT, Bickle MJ. Constraining the potential temperature of the Archean mantle: a review of the evidence from komatiites. Lithos (1993) 30:291–307.[CrossRef][Web of Science]
Ohtani E. Chemical stratification of the mantle formed by melting in the early stages of the terrestrial evolution. Tectonophysics (1988) 154:201–210.[CrossRef][Web of Science]
Ohtani E. Majorite fractionation and genesis of komatiites in the deep mantle. Precambrian Research (1990) 48:195–202.[CrossRef][Web of Science]
O’Neil J, Maurice C, Stevenson RK, Larocque J, Cloquet C, David J, Francis D. The geology of the 3·8 Ga Nuvvuagittuq (Porpoise Cove) greenstone belt, Northeastern Superior Province, Canada. In: Earth's Oldest Rocks—van Kranendonk MJ, Smithies RH, Bennett VC, eds. (2007) Amsterdam: Elsevier. 219–250.
Palme H, Jones A. Treatise on geochemistry. In: Meteorites, Comets, and Planets—Davis AM, ed. (2004) Volume 1. Elsevier Pergamon. Chapter 1.03, p. 41.
Philpotts AR. Compositions of immiscible liquids in volcanic rocks. Contributions to Mineralogy and Petrology (1982) 80:201–218.[CrossRef][Web of Science]
Polat A, Kerrich R, Wyman DA. Geochemical diversity in oceanic komatiites and basalts from the late Archean Wawa greenstone belts, Superior Province, Canada: trace element and Nd isotope evidence for a heterogeneous mantle. Precambrian Research (1999) 94:139–173.[CrossRef][Web of Science]
Poulsen KH, Borradaile GJ, Kehlenbeck MM. An inverted Archean succession at Rainy Lake, Ontario. Canadian Journal of Earth Sciences (1980) 17:1358–1369.
Rajamani V, Shivkumar K, Hanson GN, Shirey SB. Geochemistry and petrogenesis of amphibolites, Kolar Schist belt, South India: evidence for komatiitic magma derived by low percentages of melting in the mantle. Journal of Petrology (1985) 26:92–123.
Rajamani V, Shirey SB, Hanson GN. Fe-rich Archean tholeiites derived from melt-enriched mantle sources: evidence from the Kolar Schist belt, South India. Journal of Geology (1989) 97:487–501.[Web of Science]
Richard P, Shimizu N, Allègre JJ. A natural tracer: an application to oceanic basalts. Earth & Planetary Science Letters (1976) 31:269–278.[CrossRef]
Schaefer SJ, Morton P. Two komatiitic pyroclastic units, Superior Province, northwestern Ontario: their geology, petrography, and correlation. Canadian Journal of Earth Science (1991) 28:1455–1470.
Schulz KJ. Magnesian basalts from the Archaean terrains of Minnesota. In: Komatiites—Arndt NT, Nisbet EG, eds. (1982) London: Allen & Unwin. 171–186.
Shklanka R. Geology of the Steep Rock Lake Area, District of Rainy River (1972) 114 pp: Toronto: Ontario Department of Mines and Northern Affairs.
Sobolev AV, Hofmann AW, Sobolev SV, Nikogosian IK. An olivine-free mantle source of Hawaiian shield basalts. Nature (2005) 434:590–597.[CrossRef][Web of Science][Medline]
Sproule RA, Lesher CM, Ayer JA, Thurston PC, Herzberg CT. Spatial and temporal variations in the geochemistry of komatiites and komatiitic basalts in the Abitibi greenstone belt. Precambrian Research (2002) 115:153–186.[CrossRef][Web of Science]
Stone D, Kamineni DC, Jackson MC. Precambrian geology of the Atikokan area, northwestern Ontario. Geological Survey of Canada Bulletin (1992) 405:106.
Stone WE, Crocket JH, Dickin AP, Fleet ME. Origin of Archean ferropicrites: geochemical constraints from the Boston Creek Flow, Abitibi greenstone belt, Ontario, Canada. Chemical Geology (1995) 121:51–71.[CrossRef][Web of Science]
Stott GM. Precambrian Geology of the Dayohessarah Lake Area, White River, Ontario (1999) 59 pp: Sudbury: Ontario Geological Survey.
Stott GM, Mahoney KL, Zwiers WG. Summary of Field Work and Other Activities, 1994: Geology of the Dayohessarah Lake Area (1994) Sudbury: Ontario Geological Survey. 48–51.
Stott GM, Mahoney KL, Zwiers WG. Precambrian Geology, Dayohessarah Lake Area (Central). Ontario Geological Survey, Preliminary Map P.3310, scale 1:20 000 (1995a).
Stott GM, Mahoney KL, Zwiers WG. Precambrian Geology, Dayohessarah Lake Area (North). Ontario Geological Survey, Preliminary Map P.3309, scale 1:20 000 (1995b).
Takahashi E. Melting of a dry peridotite KLB-1 up to 14 GPa: implications on the origin of peridotitic upper mantle. Journal of Geophysical Research (1986) 91:9367–9382.
Tomlinson KY. The geochemistry and tectonic setting of early Precambrian greenstone belts, northern Ontario, Canada. In: Ph.D. thesis (1996) University of Portsmouth.
Tomlinson KY, Hughes DJ, Thurston PC, Hall RP. Plume magmatism and crustal growth at 2·9 to 3·0 Ga in the Steep Rock and Lumby Lake area, Western Superior Province. Lithos (1999) 46:103–136.[CrossRef][Web of Science]
Tuff J, Takahashi E, Gibson SA. Experimental constraints on the role of garnet pyroxenite in the genesis of high-Fe mantle plume derived melts. Journal of Petrology (2005) 46:2023–2058.
Ulmer P. The dependence of the Fe2+–Mg cation partitioning between olivine and basaltic liquid on pressure, temperature, and composition. Contributions to Mineralogy and Petrology (1989) 101:261–273.[CrossRef][Web of Science]
Veksler IV, Dorfman AM, Danyushevsky LV, Jakobsen JK, Dingwell DB. Immiscible silicate liquid partition coefficients: implications for crystal–melt element partitioning and basalt petrogenesis. Contributions to Mineralogy and Petrology (2006) 152:685–702.[CrossRef][Web of Science]
Walter MJ. Melting of garnet peridotite and the origin of komatiite and depleted lithosphere. Journal of Petrology (1998) 39:29–60.[CrossRef][Web of Science]
Wilks ME, Nisbet EG. Stratigraphy of the Steep Rock Group, northwestern Ontario: a major Archean unconformity and Archean stromatolites. Canadian Journal of Earth Sciences (1988) 25:370–391.
Thurston PC, Williams HR, Sutcliffe RH, Stott GM. Wawa Subprovince. Geology of Ontario, Ontario Geological Survey, Special Volume 4, Part 1 (1991) 485–539.
Wyman D, Hollings P. Long-lived mantle plume influence on an Archean protocontinent: geochemical evidence from the 3 Ga Lumby Lake greenstone belt, Ontario, Canada. Geology (1998) 26:719–722.
Yaxley GM. Experimental study of the phase and melting relations of homogeneous basalt + peridotite mixtures and implications for the petrogenesis of flood basalts. Contributions to Mineralogy and Petrology (2000) 139:326–338.[CrossRef][Web of Science]
Yaxley GM, Green DH. Reactions between eclogite and peridotite: mantle refertilisation by subduction of oceanic crust. Schweizerische Mineralogische und Petrographische Mitteilungen (1998) 78:243–255.
Zhang J, Herzberg C. Melting experiments on anhydrous peridotite KLB-1 from 5·0 to 22·5 GPa. Journal of Geophysical Research (1994) 99:17729–17742.[CrossRef]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||