| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology | Volume 43 | Number 11 | Pages 2143-2170 | 2002
© Oxford University Press 2002
A Metamorphosed, Early Archaean Chromitite from West Greenland: Implications for the Genesis of Archaean Anorthositic Chromitites
1GEMRU, UNIVERSITY OF GLOUCESTERSHIRE, FRANCIS CLOSE HALL, SWINDON ROAD, CHELTENHAM GL50 4AZ, UK
2GEOLOGICAL SURVEY OF DENMARK AND GREENLAND, THORAVEJ 8, DK-2400 COPENHAGEN NV, DENMARK
3GEOLOGICAL INSTITUTE, UNIVERSITY OF COPENHAGEN, ØSTER VOLDGADE 10, DK-1350 COPENHAGEN K, DENMARK
4DANISH LITHOSPHERE CENTRE, ØSTER VOLDGADE 10, DK-1350, COPENHAGEN K, DENMARK
Received September 6, 2001; Revised typescript accepted May 17, 2002
 |
ABSTRACT |
|---|
An early Archaean (>3·81 Ga) chromitite–ultramafic layered body from the Ujaragssuit nunât area, west Greenland, may represent the Earth’s oldest chromitite. The layered body occurs as a large xenolith (800 m x 100 m) entrained within tonalitic gneisses and preserves primary igneous layering and textures. New Re–Os and Pb–Pb isotope results support the view that it has been metamorphosed twice, in the early and late Archaean at
3·75 Ga and 2·8 Ga. Mineral chemistry and textures indicate that the chromite compositions preserve two different evolutionary trends. There is a main magmatic trend in which Cr/(Cr + Al) ratios remain relatively constant but in which there is strong enrichment in Fe3+, Fe2+ and Ti with progressive differentiation. This trend is a composite of magmatic-liquidus, magmatic-cooling and subsolidus re-equilibration processes. A second trend is defined by chromites from harzburgites in the upper part of the layered body. These chromites show magmatic replacement textures in which Fe-rich chromites are altered to aluminous chromites. Chromites showing magmatic replacement textures are thought to have formed by reaction with a late, interstitial melt during the solidification of the layered body. The close association between the Fe3+–Cr-chromites of the main trend and Al-rich chromites of the type found in other Archaean megacrystic anorthosites suggest a magmatic-genetic relationship between the two types of chromite. We propose that anorthositic chromites form in an Fe-rich basaltic melt derived from a komatiitic, boninitic or basaltic parent magma through reaction between the melt and early-formed Fe-rich chromite. KEY WORDS: chromite; anorthosite; Archaean; Greenland; Re–Os
|
INTRODUCTION |
|---|
Terrestrial anorthosites that formed during the Archaean occur as stratiform bodies that are commonly highly deformed and metamorphosed. Ashwal (1993) has argued that Archaean anorthosites are distinctive in both form and origin from anorthosites formed later in Earth history, and has proposed that they crystallized from a basaltic parent magma. The details of their tectonic setting, however, and the precise mechanism of their crystallization history have still to be worked out (Ashwal, 1993).
A number of Archaean anorthosites contain chromitite horizons and it is the chromite–anorthosite association that is the focus of this study. It has long been known that chromite can be used as a petrogenetic indicator (Irvine, 1965, 1967), because chromite compositions are a rich source of information on the origin and evolution of their parent magmas. Of importance here is the unusual field of chromite compositions from Archaean anorthosites when plotted on a diagram of Cr/(Cr + Al) vs Fe2+/(Fe2+ + Mg) (Rollinson, 1995), defining a compositional range that is different from that of chromites in both Phanerozoic mafic rocks and layered intrusions (Steele et al., 1977). This has recently been confirmed in a study by Barnes & Roeder (2001), who have defined the compositional range of terrestrial chromites using a database of 26 000 chromites from both igneous and metamorphic rocks. Although they did not explicitly describe chromites from anorthosites it is clear from published data that some chromites from anorthosites plot outside the main field for terrestrial chromites.
We describe here the petrology of a small (800 m x 100 m), layered ultramafic–chromitite xenolith, metamorphosed to amphibolite grade entrained in tonalitic orthogneiss in inner Godthåbsfjord, west Greenland. This ultramafic xenolith contains thin leucogabbro–anorthosite horizons in the upper part of the body and all layers contain chromite. Previous studies of Archaean anorthositic chromites have focused only on the anorthositic component of the intrusion (Ghisler, 1976). Here we explore chromitite–anorthosite relationships in the context of part of an evolving magma chamber. Our results provide important insights into the processes by which Archaean anorthositic chromitites have formed.
The ultramafic xenolith described here is very well exposed and displays igneous layering on a centimetre to decimetre scale, indicating that it formed part of a disrupted layered intrusion. It was first described by Chadwick & Crewe (1986), who drew attention to the fact that it may represent one of the oldest known terrestrial chromite deposits. Here we confirm the antiquity of the chromitite with the first direct dating of the xenolith, describe in detail the mineral chemistry of the chromitites and associated dunites, harzburgites and leucogabbros, and seek to identify the nature of the parent magma from which the layered body formed.
|
FIELD RELATIONSHIPS |
|---|
The ultramafic xenolith described here is located in the Ujaragssuit nunât area of west Greenland, about 20 km SE of the Isua Greenstone Belt (Fig. 1a), at 64°56·24'N, 49°59·26'W (World Geodetic System 1984 map datum). The xenolith is about 100 m wide and 800 m long, comprises dunites, chromitites, amphibole-harzburgites and gabbro anorthosites, and is the largest of several hundred ultramafic xenoliths in the Ujaragssuit nunât area. Xenoliths of this type have been traced along strike in an east–west direction over several kilometres and were mapped by Chadwick et al. (1983) as the limb of a large fold (Fig. 1b). Amphibolites and banded iron formations are also found associated with the ultramafic rocks and the presence of ultramafic rocks within amphibolite sheets led Chadwick & Crewe (1982) to suggest that the original ultramafic rocks were intruded as sills into the amphibolites. The xenoliths are surrounded by tonalitic and granodioritic gneisses and intruded by veins of a relatively undeformed biotite-tonalite. Nutman et al. (1996) reported a concordant U–Pb zircon age of 3811 ± 4 Ma for the gneisses enclosing the chromitite-bearing ultramafic rocks, which therefore provides a minimum age for the ultramafic xenoliths in this area.
|
The ultramafic xenolith preserves primary igneous layering, indicating that it probably formed part of a layered igneous intrusion. The stratigraphy of the xenolith is given in Fig. 2. The total preserved thickness is 45 m, although this is a minimum thickness as both the base and top of the layered complex are truncated by intrusive felsic gneisses. We have assumed that each appearance of chromitite bands represents the beginning of a new magmatic cycle, reflecting open-system magmatic processes. In the lower part of the layered succession there are four cycles, each between 6 and 10 m thick. In the upper part three more cycles are recognized on the basis of alternating dunite–amphibole-harzburgite layers, each between 3 and 6 m thick (Fig. 2). In this case the start of a new cycle is thought to be indicated by the presence of a dunite layer. It should be noted that, although the stratigraphic section preserved here is probably incomplete, the transition from dunites to harzburgites over 45 m represents a remarkably compressed section compared with layered igneous intrusions showing a similar range of bulk compositions, such as the Great Dyke of Zimbabwe (Wilson, 1982).
|
There is one main massive chromitite band, which is 3 m thick. In places, within a metre or so of this chromitite layer, there are blocks of chromitite up to 2 m in length enclosed in dunite. It is thought that these blocks are fragments of the main chromitite band, which became detached and descended into the underlying dunite before the final solidification of the layered body (Fig. 3a). Banded chromitites appear four times in the succession and comprise units 1–2 m thick of 1–10 cm alternating layers of chromitite and dunite (Fig. 3b and d). One of these units (banded chromitite-1, Fig. 2) shows bands with a sharp (chromite-rich) base and a gradational (olivine + chromite) top, which indicates ‘younging’ towards the harzburgites (Fig. 3b), i.e. towards the south. Dunites are present in every cycle and are medium- to coarse-grained olivine-rich rocks, forming layers up to 9 m thick. The thicker layers are associated with massive and banded chromitites and the dunites may contain large olivine crystals up to 1 cm in length. Thinner dunite layers are associated with amphibole-harzburgites in the upper part of the sequence. Amphibole-harzburgites are strongly banded with olivine- and orthopyroxene-rich bands. The amphibole-harzburgites also contain layers of leucogabbro (Fig. 3c) up to 20 cm thick, some of which contain small (2–3 mm) grains of ruby corundum. Some leucogabbro layers are sufficiently feldspathic to be anorthosites.
|
|
PETROGRAPHY |
|---|
Massive chromitite
Massive chromitite occurs as a unit of 3 m thickness at the base of lower cycle 3 and is dominated by subhedral grains of chromite, 0·2–2 mm in diameter, showing a cumulus texture (Fig. 4a). The intercumulus mineralogy is now altered to chlorite + phlogopite (mg-number = 97) + carbonate. Many chromite grains are affected by this late alteration and either contain abundant, small chlorite and/or phlogopite inclusions, or show irregular grain boundaries at which chromite is intergrown with chlorite and/or phlogopite.
|
Banded chromitites
There are four banded chromitite horizons (Fig. 2). Chromite forms between 40 and 60% of the banded chromitites, the remainder comprising olivine and/or alteration minerals (chlorite, phlogopite and carbonate). The chromite bands contain subhedral and euhedral grains up to 1 mm in diameter. Alternating olivine bands are made up of irregular to elliptical grains normally 1–4 mm in diameter, sometimes enclosed by trains of small chromite grains (Fig. 4g). Some larger grains are up to 1 cm in length and may represent former phenocrysts. Alteration minerals (chlorite, phlogopite and carbonate) make up as much as 50% of some banded chromitites. Some chromite grains show a trellis pattern of phlogopitic alteration (Fig. 4b), in which the phlogopite has developed along cleavage planes within the chromite.
Dunites
Olivine is the main cumulus phase in the dunites, forming grains 1–3 mm in diameter. In some samples the dunites have a granular texture (Fig. 4f). Less commonly there are larger elliptical grains, up to 10 mm in length in a finer-grained olivine-rich matrix. Chromite forms smaller grains (up to 0·5 mm) at olivine grain boundaries (Fig. 4g). In some samples a small amount of orthopyroxene (up to 5%) is present. This may be present as inclusions in the olivine (Fig. 4d) or as discrete grains. In one sample (462389) orthopyroxene is altered to an Mg-rich amphibole. The later alteration assemblage of chlorite and phlogopite is much less abundant than in the chromite-rich samples. Dunites associated with amphibole-harzburgite in the upper part of the succession contain more orthopyroxene and tend to be harzburgitic. Orthopyroxene grains are up to 4 mm across and contain both olivine and chromite inclusions.
Amphibole-harzburgites
Amphibole-harzburgites are banded, with alternating olivine and orthopyroxene-rich bands. In the orthopyroxene-rich bands some orthopyroxene grains are up to 1 cm long and contain inclusions of olivine. In the olivine-rich bands, some olivine grains contain inclusions of orthopyroxene. This relationship between olivine and orthopyroxene is thought to be a relict, primary igneous feature reflecting a reaction relationship between olivine and melt during the crystallization of the magma (see Fig. 4d). A pale green to colourless edenitic amphibole forms elongate, straight-sided grains up to 1 mm in length. In some instances the amphibole replaces orthopyroxene (Fig. 4e), elsewhere amphibole grains surround elliptical olivine grains (Fig. 4c and h). These relationships imply that the amphibole grew either as a late intercumulus phase, or during a later metamorphic event.
Chromite is only a minor constituent of the amphibole-harzburgites but these grains preserve an important textural history. A scanning electron microscopy investigation has shown that although most chromite grains are small (50–1000 µm), they are composed of two phases, which when viewed as back-scatter images appear as a light-coloured (Fe-rich) phase and a dark Al-rich phase. The details of this study have been given by Appel et al. (2002). The light-coloured phase contains exsolution lamellae of ilmenite, which are normally truncated by the darker phase (Fig. 5), but occasionally are preserved as ghost structures within the darker phase. At the boundary between the two phases there is a reaction rim comprising a symplectite of dark and light-coloured spinels. Fine-scale exsolution of a dark phase in the light and a light phase in the dark is also seen; however, these lamellae, together with the phases in the symplectite, are too small to analyse with a standard electron beam. The patchy, unsystematic distribution of the two phases, together with the truncation of structures and the presence of a reaction rim, suggests that the darker phase has replaced the lighter phase. The observation that the replacement post-dates oxidation–exsolution strongly suggests that cooling had taken place before the replacement.
|
Leucogabbros
Leucogabbros contain the mineral assemblage pargasitic amphibole + plagioclase. Individual bands are relatively enriched in either amphibole or plagioclase. Some bands are anorthositic (>90% plagioclase) in composition, but most are leucogabbros [gabbros with >65% plagioclase (LeMaitre, 1989)]. In the more feldspathic bands the amphibole–plagioclase assemblage is granular and the plagioclase strongly zoned, whereas in the amphibole-rich bands large grains of amphibole up to 8 mm long with small inclusions of plagioclase are more common. Metamorphic biotite (mg-number = 83) is a late-stage mineral in the leucogabbros, frequently surrounding the plagioclase. In one sample (472807) granular aggregates of pink (ruby) corundum surround plagioclase in an anorthositic band within the leucogabbro. Also present in the plagioclase of this sample are very small (10–50 µm) inclusions of aluminous chrome-spinel.
Petrographic summary
Our petrographic observations indicate that these rocks preserve some primary igneous textures. These include primary igneous layering, chromite cumulate textures and mineral reaction textures. In addition, however, there is a later suite of minerals that overgrow the igneous assemblages. The earliest of these is amphibole, thought to be the product of an early phase of metamorphism. In places amphibole is replaced by chlorite (Fig. 4c and h) and, in the chromitites, chromite is replaced by phlogopite. The chlorite and phlogopite have a random orientation in the rock and are thought to have formed during either a later static metamorphism or retrograde cooling. A further metamorphic assemblage talc ± carbonate replaces chlorite and phlogopite.
The metamorphic rocks of this region are known to have experienced a complex history. In the Isua Greenstone Belt, 20 km to the north, a detailed study of garnet textures has revealed evidence of three distinct garnet-growth events, which are thought to represent at least two distinct metamorphic events (Rollinson, 2002). Frei & Rosing (2001) have obtained a Pb–Pb step-leaching age of 3739 ± 21 Ma on Isua garnets, but also there are well-established late Archaean mineral ages from Isua (Gruau et al., 1996; Frei et al., 1999) implying both early and late Archaean amphibolite-facies metamorphic events in the greenstone belt. It is probable that these metamorphic events are also recorded in the region south of Isua, in the study area described here.
Thus we recognize textures that we attribute to four discrete events:
- the magmatic crystallization of the layered body;
- an amphibolite-facies metamorphism;
- a lower-temperature fluid influx event, during which chlorite and phlogopite grew;
- a low-temperature event during which an H2O–CO2-rich fluid was introduced and the talc–carbonate assemblage grew.
We present evidence below to suggest that the amphibolite-facies event was early Archaean and that the low-temperature events were late Archaean in age.
|
ANALYTICAL TECHNIQUES |
|---|
Samples were collected up the stratigraphy of the intrusion so as to sample all the major chromite horizons and a representative selection of the major lithologies. Mineral analyses for the phases chromite, olivine, orthopyroxene, hornblende, chlorite were made at the University of Bristol, using a JEOL JXA 8600 wavelength-dispersive electron microprobe. The operating conditions were 15 kV with an initial beam current of 15 nA. A range of natural silicate and oxide standards was used. Fe3+ in spinels was calculated according to the charge balance equation of Droop (1978). Plagioclase and chromite analyses in the leucogabbro were made by energy-dispersive X-ray spectrometry on a Philips XL40 scanning electron microscope at the Geological Survey of Denmark and Greenland in Copenhagen.
Re–Os isotopic analyses were performed on chromite separates from both massive and banded chromitites. The sample locations are given in Fig. 2. We also analysed two whole-rock samples and two olivine separates. Because the focus of the Re–Os work was the chromitite, samples were collected over an interval of 12 m from the lower part of the stratigraphic section. Mineral separates (chromite and olivine) were made from crushed samples using magnetic and heavy liquid concentration techniques, and sieve fractions <200 µm, purified by hand-picking under a binocular microscope to sample amounts of 200 mg. The samples were attacked in Carius tubes for 3 days at 210°C after addition of a 188Os and 185Re spike, following the method of Shirey & Walker (1995). Os was distilled directly into HBr following the technique of Naegler & Frei (1997), and Re was recovered by liquid separation. Final purification of Os fractions was achieved by applying the mini-distillation procedure of Roy-Barman & Allègre (1995). Os isotopic compositions were measured in negative mode on a VG Sector 54 mass spectrometer at the University of Copenhagen (Denmark), both in static multicollection and single electron multiplier modes, depending on the intensity of the Os beam. Os was loaded with HBr and Ba(OH) and ionized from Re-free Pt filaments at temperatures <800°C. Re concentrations were measured by multi-collector inductively coupled plasma mass spectrometry (MC-ICP-MS) on an AXIOM system at the Danish Lithosphere Centre. On-line mass fractionation was achieved by tracing the 191Ir/194Ir ratio from an iridium standard solution added to the sample solutions before the aspiration into the plasma. Procedural blanks for Os remained below 1·2 pg, and Re total blanks ranged between 3 and 6 pg.
Pb isotopic analyses were performed on 200 mg purified chromite dry split aliquots and on one whole-rock sample. The chromite samples were powdered by hand in a agate mortar and subsequently attacked in aqua regia for 7 days in Savillex Teflon beakers. Pb was conventionally separated in 0·5 ml glass columns charged with anion exchange resin, followed by a clean-up on 300 µl Teflon columns and using a standard HBr–HCl elution recipe. Total procedural blanks for Pb amounted to <83 pg. Judged from relative lead beam intensities on the mass spectrometer, this level is considered insignificant with respect to contamination of the sample leads in every loading. Isotope analyses were performed by thermal ionization mass spectrometry (TIMS) on a VG Sector 54-IT system at the University of Copenhagen (Denmark). Fractionation for Pb was controlled by repetitive analysis of the NBS SRM-981 standard [values of Todt et al. (1993)] and amounted to 0·110 ± 0·002%/a.m.u (2
; n = 8). Procedural Pb blanks remained below 80 pg.
|
MINERAL CHEMISTRY |
|---|
Compositional variations in chromite and olivine from 10 samples at different positions in the stratigraphic sequence are plotted against stratigraphic height to investigate cryptic variation (Fig. 2). These data are for mineral core compositions only. Only cycle 3, in the lower part of the layered body, has enough data points to show any systematic variation. Here there is an upwards increase in cr-number in chromite and an upwards decrease in mg-number in both olivine and chromite. These observations support but cannot prove the suggestion made above on the basis of the field evidence, that the appearance of chromitite bands indicates the beginning of a new magmatic cycle, as, for example, in the Great Dyke (Wilson, 1982). Thus our working hypothesis is that the layered complex may preserve as many as seven cycles and that each of these represents a fresh pulse of magma into a crystallizing magma chamber.
Chromite mineral chemistry
Chromite compositions are presented on a Cr/(Cr + Al) vs Fe/(Fe + Mg) (cr-number–fe-number) diagram (Fig. 6). This is a projection from the spinel prism onto the plane chromite (FeCr2O4)–picrochromite (MgCr2O4)–hercynite (FeAl2O4)–spinel (MgAl2O4). This diagram shows the range of measured chromite compositions for each sample in the shaded fields. Individual point analyses for core compositions (samples 389–396) and, for grains that show replacement textures, the earliest-formed chromite (samples 389–400), are shown as filled symbols. The core compositions are thought to represent magmatic chromite compositions in this metamorphosed igneous layered body. Representative analyses are given in Table 1.
|
, core compositions (cores only). Samples from the replacement trend are shown as darker grey fields (full range of compositions); •, core compositions. The field for chromites in anorthosite is shown separately; +, individual analyses.
|
Chromite core compositions fall into two groups. In the cr-number vs fe-number diagram (Fig. 6) samples from dunites and chromitites in the lower part of the layered body (sample numbers 389–392, 394, 396) and harzburgites from the upper part (samples 398–400) define a flat trend in which cr-number varies over a small range, but for which fe-number shows wide variation (Fig. 6). On a trivalent cation plot (Fig. 7e) the chromite core compositions define a trend of increasing Fe3+ with near-constant Cr/Al ratio. This chromite trend is referred to below as the ‘main chromite trend’. In contrast, the darker phase of the two-phase spinels in the harzburgites in the upper part of the layered body (samples 398–400), described above under the petrography of the amphibole-harzburgites (see Fig. 5), defines a trend on the cr-number vs fe-number diagram in which there is a wide variation in cr-number and a positive correlation between cr-number and fe-number (Fig. 6). On the trivalent cation plot (Fig. 7e) this trend is seen as one of increasing Al with a near-constant Fe3+/Cr ratio. We interpret these compositions as the result of replacement (see a full discussion below in the section on petrogenesis) and this trend is referred to as the ‘chromite replacement trend’. The two trends converge on the cr-number vs fe-number diagram at cr-number 0·85 and fe-number 0·90. Chromites from anorthosites define a compositional trend sub-parallel to that of the replacement trend.
|
), from the replacement trend (
) and from the anorthosites (+). (a) Cr3+–Fe3+, calculated on the basis of 16 cations on the trivalent site; the diagonal line represents a 1:1 replacement. (b) Al3+–Fe3+, calculated on the basis of 16 cations on the trivalent site; the diagonal line represents a 1:1 replacement. (c) Ti4+–Fe3+ plot. (d) TiO2–Cr2O3 concentrations compared with the fields for arc and MORB spinels (after
Chromite compositions from the main chromite trend
Chromite core compositions show a small increase in cr-number and fe-number with increasing stratigraphic height in the intrusion (Fig. 2). Cation plots for the trivalent ions in spinel show negative correlations between Cr and Fe3+ and Al and Fe3+ (Fig. 7a and b). A cation plot for Fe3+ vs (Cr + Al) shows a one-for-one replacement of (Cr + Al) by Fe3+, indicating a progressive enrichment in Fe3+ with magmatic evolution from chromitite to harzburgite. There is also a small but systematic increase in Ti with Fe3+ (Fig. 7c). A cation plot for the divalent cations Mg and Fe2+ shows a linear correlation, implying the substitution of Mg by Fe2+ with increasing height. There is also a positive linear relationship between Fe2+ and Fe3+, implying that chromites are progressively enriched in total iron with increasing stratigraphic height. End-member compositions calculated on the basis of 24 cations vary from (Fe4·5 Mg3·5)8 (Cr10·6 Al3·3 Fe3+2·0)16 (most primitive) to (Fe7·76 Mg0·44)8 (Cr3·74 Al0·55 Fe3+11·1)16 (most evolved). This compositional range implies an increase in molar proportion of the Fe3O4 end-member from 13% in the most primitive composition to 72% in the most evolved composition. Thus the most evolved spinels are more properly termed chrome-magnetites.
The main chromite trend identified here is similar to the Fe–Ti trend identified by Barnes & Roeder (2001) in their study of chromites in terrestrial mafic and ultramafic rocks, although on a cr-number–fe-number plot the results presented here suggest that at high fe-number chromites trend towards high Cr/Al ratios, whereas in the Barnes & Roeder dataset they trend towards lower Cr/Al ratios.
Chromites from the main chromite trend are compositionally zoned. Compositional trends for individual samples on a cr-number–fe-number diagram (Fig. 6) show a strong increase from core to rim in cr-number, with a small increase in fe-number. This is thought to reflect the alteration of chromite in the presence of a hydrothermal fluid to chlorite, in which Al, and to a lesser extent Mg, are removed from the chromite to create magnesian chlorite.
Chromite compositions falling on the chromite replacement trend
Chromites from the harzburgites in the upper part of the layered body show a range of replacement textures, varying from incipient replacement to almost total replacement of Fe-chromite by Al-chromite (Appel et al., 2002). Compositional variation within the replacement chromites is from (Fe6·2 Mg1·8)8 (Cr6·1 Al4·9 Fe3+4·8)16 (least aluminous) to (Fe5·0 Mg3·0)8 (Cr5·0 Al8·6 Fe3+2·3)16 (most aluminous).
Chromite in anorthosite
Chromite in the anorthosites occurs as very small (20 µm) grains enclosed in plagioclase. There is a wide compositional variation between grains. On the cr-number vs fe-number diagram (Fig. 6) the chromites define a broad field with a trend from low fe-number, low cr-number to higher fe-number and cr-number (Fig. 6). The cation plots in Fig. 7a and b emphasize the aluminous nature of these spinels and their low Fe3+ and Ti contents. These spinels also contain up to 1·0 wt % MnO and up to 2 wt % ZnO. Compositions range from (Fe5·4 Mg2·1)8 (Cr5·5 Al10·2 Fe3+0·3)16 (least aluminous) to (Fe3·6 Mg4·0)8 (Cr2·8 Al13·3 Fe3+0·0)16 (most aluminous) and show a progressive increase in the spinel end-member (MgAl2O4) relative to chromite (FeCr2O4). Thermodynamic calculations reported by Roeder et al. (1979) show that Al in the chromite structure stabilizes Mg relative to Fe2+. It is likely, in view of the very small size of these grains and the association with metamorphic corundum, that the range of spinel compositions in the anorthosites represents the effects of the later metamorphism. We propose that this has enhanced the Al content of the chromites through the breakdown of plagioclase and an aluminous amphibole. If this proposal is correct, then the more Cr-rich spinels in the anorthosites approach the original igneous chromite compositions.
The mineral chemistry of the main silicate phases
Olivine
Olivine core and rim compositions are shown in a NiO–Fo content plot in Fig. 8 (data in Table 2). In most samples there is no significant compositional difference between rim compositions (shaded areas in Fig. 8) and core compositions (filled symbols in Fig. 8). Olivine core compositions range from Fo85·5 to Fo91·5 (see Fig. 8). Harzburgitic olivines are the most Fe rich and olivines from banded chromitites the most magnesian. Olivines become more Fe rich through individual magmatic cycles and through the evolution of the layered body as a whole (Fig. 2). The compositional variation in lower cycle 3 is similar in range to that found in individual cycles in the Great Dyke (Wilson, 1982), although the total range of olivine compositions in the Ujaragssuit nunât layered body is less than that observed in the Great Dyke intrusion as a whole. Compositional variation in olivines correlates with that of the coexisting chromites from the main trend and core compositions define a trend in which increasing mg-number in chromite correlates with increasing mg-number in olivine. Replacement chromites do not, however, lie on this trend. The co-variation between mg-number in chromite and olivine in the main trend is well known and is a complex function of equilibration temperature (Irvine, 1965) and the bulk composition of the host rock (Roeder et al., 1979).
|
|
Orthopyroxene
Orthopyroxene is most abundant in the higher-level rocks in the layered body. Core compositions have mg-number ranging from 88·5 to 85·5 (Table 3), almost identical to those of coexisting olivines. There is a negative correlation between mg-number and Al2O3 in orthopyroxene cores. The less magnesian, more aluminous orthopyroxenes occur in the harzburgites in the upper part of the intrusion where the more magnesian Al-chromites of replacement origin are found.
|
Other phases present
Plagioclase. Plagioclase is present only in the leucogabbro layers in the harzburgite unit in the upper part of the layered body. Plagioclase compositions vary from An22 to An76. Plagioclase rims tend to be more sodic (An22–46), whereas cores are more calcic (An49–76). The zonation is thought to reflect re-equilibration between an originally igneous plagioclase and metamorphic hornblende.
Amphibole. Calcic amphiboles present in the more evolved samples are edenites (Leake et al., 1997). Mg-number ranges between 85 and 88, except in the leucogabbros where mg-number is between 74 and 78 (Table 3). In sample 396 the edenitic amphibole is zoned from edenite (in contact with olivine) to cummingtonite (in contact with orthopyroxene). In sample 398 in addition to edenite there is also a small amount of tremolite present associated with carbonate.
Chlorite. Chlorites are magnesian and have mg-number between 94 and 97. They also contain small amounts of Cr2O3—normally between 1·5 and 2·5 wt %. The Cr2O3 content of chromites varies with the proportion of chromite in the rock, and in the chromitites chlorites contain >4 wt % Cr2O3.
Phlogopite. Phlogopites have mg-number between 93·5 and 98 (Table 3), values that are almost identical to those of coexisting chlorites. The composition varies with the Mg content of the host rock. Phlogopites contain a small amount of Cr2O3—normally between 0·8 and 1·1 wt %. Metamorphic biotite is present in leucograbbro 472803 and this has a lower mg-number of 83.
Carbonate. Carbonate is magnesite with 0·05 mole fraction Fe and 0·005 mole fraction Ca.
Sulphides. Locally there are high concentrations of pentlandite with associated pyrrhotite and chalcopyrite. These were found close to the margins of the layered body and some distance away from the samples used for our mineral-chemical studies.
|
THERMOMETRY AND OXYGEN BAROMETRY |
|---|
 TOPABSTRACTINTRODUCTIONFIELD RELATIONSHIPSPETROGRAPHYANALYTICAL TECHNIQUESMINERAL CHEMISTRY THERMOMETRY AND OXYGEN BAROMETRY GEOCHRONOLOGY AND ISOTOPE...PETROGENESISDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Olivine–spinel thermometry
Equilibration temperatures for coexisting olivine–chromite pairs were calculated using the olivine–spinel Fe–Mg exchange thermometer of Sack & Ghiorso (1991) and are given in Table 4. Temperatures for olivine–chromite pairs from the main trend are in the range 514–646°C whereas olivine–chromite pairs from the replacement trend chromites record higher temperatures, in the range 621–685°C. Temperatures calculated for an olivine–chromite pair, not in contact and representing compositions closest to the original igneous compositions, are
810°C. Analytical errors on mineral analyses propagate to ±10°C errors in the calculated temperatures. The calculated temperatures are lower than those found for olivine–chromite pairs in rocks of similar bulk composition (dunite and chromitite) from the Great Dyke (Wilson, 1982), suggesting that they represent re-equilibration during subsequent cooling of the layered body.
|
Similar results are obtained for amphibole–plagioclase (rim) pairs in leucogabbro 472803, which yield temperatures in the range 610–670°C using the thermometer of Holland & Blundy (1994).
Olivine–spinel–orthopyroxene oxygen barometry
The olivine–spinel–orthopyroxene oxygen barometer of Ballhaus et al. (1991) is based upon the fayalite–ferrosilite–magnetite buffer reaction
 |
|
Delano (2001) has recently argued that chromites from the Ujaragssuit nunât layered body plot within the same field as modern mid-ocean ridge basalt (MORB) and Hawaiian volcanic rocks, implying that they crystallized under similar oxidizing conditions, close to the QFM buffer. The argument is based upon the similarity of chromite compositions in Fe2+–Fe3+–Mg cation space. The results presented here for chromites from the chromitite support this argument (Fig. 7f). These are chromites which are probably close to their original igneous compositions. Chromites from the main chromite trend plot away from the compositional space of MORB towards more oxidizing conditions. These results could suggest that the oxidation state of the original melt was close to that of the QFM buffer, whereas the interstitial melt evolved to QFM + 5.
|
GEOCHRONOLOGY AND ISOTOPE GEOLOGY |
|---|
Re–Os isotopic results
Chromite has low Re/Os ratios and has been used to image the Os isotopic composition of the mantle reservoir from which its parent magma was derived. Os isotopic data from Archaean chromitites have been used to constrain the evolution of the sub-continental lithospheric mantle in southern Africa (Naegler et al., 1997). Thus we examined what we thought might be the oldest terrestrial chromites with a view to constraining the nature of the subcontinental lithospheric mantle beneath SW Greenland. The Re–Os isotope data are presented in Table 5a and are plotted in an age vs Os isotopic evolution diagram in Fig. 10. The Os concentrations of the chromite separates vary significantly between 6·9 and 802 ppb. In part this variation can be attributed to a dilution effect caused by the presence of olivine and our inability to obtain pure mineral separates. However, the very large range, coupled with very low Os concentrations in olivine, suggests that the measured concentrations reflect original compositional differences preserved in different layers of the intrusion. Three chromite separates from sample 422390b taken from individual chromite bands a few centimetres apart have Os concentrations between 55 and 101 ppb (Table 5a). These variations are thought to reflect differences in the Os concentration of the melt from which the chromite grew (e.g. McCandless et al., 1999). The measured Os concentrations are within the range reported for Archaean chromites elsewhere (Naegler et al., 1997) apart from the very high Os concentration of 802 ppb in the massive chromitite (sample 460022).
|
|
|
Olivine from samples 460020 and 460021 contains Os concentrations between 1·9 and 2·5 ppb, which is consistent with the many results that show that Os is incompatible in the olivine structure (e.g. Puchtel et al., 1999). The three measurements for sample 460021 are self-consistent, and show that Os concentrations in the whole rocks lie between those for chromite and olivine.
The oldest model ages (Tma)—the time of separation from mantle, and calculated by assuming that the Re/Os of the sample is representative of its long-term history in the mantle or in the crust—calculated here are for the chromites, and are between 3514 and 3714 Ma. The heterogeneity in Tma probably reflects partial open-system behaviour in the Re–Os system. This is thought to be principally due to the mobility of Re (e.g. Frei et al., 1998c) during post-magmatic processes that are thought to have affected this region. Late Re mobility is particularly apparent in the massive chromite (sample 460022), which has a very high Re concentration relative to other chromite separates (2409 ppt, compared with 80–617 ppt, Table 5a). High Re in the massive chromite may reflect the altered intercumulus mineral paragenesis (chlorite, phlogopite–carbonate) of this sample and the fact that chromite grains are commonly fractured and infilled with late chlorite and phlogopite. We propose that the fluids responsible for the chlorite–phlogopite alteration were derived from the enclosing gneisses and that they contaminated the ultramafic rocks with Re when they percolated through the layered sequence. The observation that chromite separates contain higher Re concentrations than olivine separates (Table 2a) is in part due to differences in the partitioning of Re between the two phases but may also be due to the presence of alteration minerals containing Re in the chromite, for it is more difficult to obtain pure separates of the chromite than of the olivine. The olivine separates yield mid-Archaean Tma (3·634–3·639 Ga, Table 5a).
We have also calculated Re depletion ages (Trd, Table 5a). These are used to assess the minimum age for the Re depletion and are calculated by estimating the time of intersection with chondritic mantle growth line, assuming that a melt depletion event previously removed all Re from the sample, and hence, growth of 187Os was terminated at that time. The Trd ages indicate that the minimum age of the chromites is between 2·73 and 3·36 Ga, strongly suggesting that they are not late Archaean in age.
If the 3·81 Ga U–Pb zircon age for a felsic gneiss enclosing the ultramafic rocks (Nutman et al., 1996) represents a minimum emplacement age, and there is no evidence that the zircons have inherited cores, then the Tma values obtained for this layered body are clearly too young. This we attribute to the addition of Re to the rocks after their igneous crystallization. Alternatively, the suprachondritic 187Os/186Os ratios at 3·81 Ga could reflect a real enrichment of the mantle very early in Earth’s history, possibly by recycling of enriched oceanic crustal material back into Hadean mantle. Figure 10 depicts the position of data points relative to the mantle evolution line. On the basis of independent geological constraints, including geochronological ones, and the knowledge of an Early Archean 3·65–3·74 Ga metamorphic overprinting superimposed on the Isua Greenstone Belt, we prefer to interpret the suprachondritic ratios at 3·81 Ga as resulting from post-formational disturbances of the Re/Os ratio in these rocks. Because of such post-formational open-system behaviour of the Re–Os system, it is not possible to infer models for the character of the subcontinental lithospheric mantle (e.g. Naegler et al., 1997) beneath SW Greenland from these data. The timing of the Re mobility is not certain. There are two possibilities. First, a 3·65–3·74 Ga amphibolite-facies metamorphic overprint is recorded in the gneiss area south of the Isua Greenstone Belt and within the supracrustals themselves (Frei et al., 1999, 2002; Frei & Rosing, 2001). Second, a late Archaean amphibolite-facies metamorphism (2·65–2·8 Ga) is also recorded from the Itsaq gneiss terrane and the Isua supracrustals (Baadsgaard et al., 1986; Gruau et al., 1996; Frei et al., 1998a, 1998b). The possibility of such a late Archaean disturbance is supported by the tendency of individual sample evolution lines in Fig. 10 to converge in this age interval.
Pb isotopic results
Pb isotopic data from five chromite separates and one whole-rock sample are reported in Table 5b and plotted on a conventional Pb isotope diagram in Fig. 11a and b. The early Archaean Pb isotope evolution curves of Frei & Rosing (2001), valid for this part of Greenland, are plotted for reference. Unlike most ultramafic rocks, which have very low U/Pb ratios, the analysed samples are fairly radiogenic with 206Pb/204Pb in a narrow range from 12·0 to 12·4. The narrow spread in the data is consistent with the fact that the samples originate from a relatively small sample area and from within a single layered body. In the 207Pb/204Pb vs 206Pb/204Pb plot (Fig. 11a), the data define a linear correlation with a slope corresponding to an age of 2·79 ± 0·69 Ga. This line intersects the early Archaean growth curve at 3·43 Ga. In the plot of 208Pb/204Pb vs 206Pb/204Pb (Fig. 11b), the data points lie consistently above the growth curve for Pb and indicate an elevated Th/U in the source of the Pb relative to that of the bulk crust at this time. This characteristic is encountered in many supracrustal rocks in the western sector of the Isua belt (e.g. Blichert-Toft & Frei, 2001; Frei et al., 2002). Although the growth curve intersection age in Fig. 11a appears geologically meaningless, the late Archaean age defined by the reference line is interpreted as caused by partial resetting during a a late Archaean metamorphism. The Pb isotopic compositions recovered from the Ujaragssuit nunât samples require both elevated U/Pb and Th/Pb ratios and probably represent Pb derived from outside the ultramafic xenolith. This source is likely to be the enclosing gneisses. The high Th/U relative to the average crustal growth may suggest that the lead was derived from a Th–U-rich phase that was introduced into the layered sequence during metamorphism. In our work on the supracrustal rock sequences from Isua, we have proposed that secondary allanite is the host to the elevated 208Pb/204Pb. Metamorphosed pillow basalts from the western Isua supracrustal belt contain very rare allanite, but the radiogenic Pb in these rare grains dominates the radiogenic Pb budget in such samples (Frei et al., 2002). Here we suggest that this Pb probably originated in the enclosing gneisses and was introduced into the ultramafic xenolith during the early Archaean metamorphism at 3·74 Ga (Frei & Rosing, 2001; Frei et al., 2002). We assume that increased metamictization of this accessory mineral facilitated Pb loss during late Archaean metamorphism, similar to the basalts from Isua described by Frei et al. (2002).
Our Re–Os and Pb–Pb isotopic studies are consistent with data from the Isua Greenstone Belt, 20 km to the north, which show that there was a metamorphic event or events between 3·65 and 3·74 Ga. Our Pb–Pb data show that the U–Th–Pb systematics of these rocks were disturbed during the late Archaean at 2·79 Ga, and we infer that this is probably the event during which K was mobile and the phlogopite–chlorite alteration assemblage formed. The best age for the ultramafic xenolith is the U–Pb zircon age of 3·81 Ga from the enclosing gneisses (Nutman et al., 1996) and is consistent with our oldest measured chromite model age in the light of the subsequent Re mobility.
|
PETROGENESIS |
|---|
It is apparent from our petrographic descriptions and the geochronological data presented above that the layered body from the Ujaragssuit nunât area of west Greenland has had a complex, multi-stage history. We suggest that four separate stages can be identified:
- a primary magmatic stage; our observations suggest that there were a number of separate magma pulses of slightly different composition, each of which experienced some fractional crystallization;
- a late magmatic stage, during which solid mineral phases re-equilibrated with residual, interstitial melt, during the solidification of the magma;
- amphibolite-facies metamorphism, during which solid phases further re-equilibrated through solid-state diffusion and/or via a fluid phase;
- a low-grade, fluid-present metamorphic event during which chlorite and phlogopite grew.
We tentatively assign the magmatic events to >3·81 Ga on the basis of the zircon geochronology of Nutman et al. (1996). The amphibolite-facies event is thought to be early Archaean (3·75 Ga) from a similar event observed in the nearby Isua Greenstone Belt. The low-grade metamorphic event could be either the retrograde stage of the early Archaean metamorphism, or, as the isotopic results presented here suggest, a separate event in the late Archaean. Here we seek to interpret the chemistry of the chromites in the light of this multi-stage history.
Origin of the ‘main trend’ chromites
It is assumed here that the least modified chromite compositions are those preserved in the chromitites, for in these lithologies there are few phases with which the chromite may undergo chemical exchange. Thus the main chromite trend can be described in terms of the changes that have taken place away from the compositions of the chromitites. The principal compositional changes are a large increase in the Fe2+/(Fe2+ + Mg) ratio and in the Fe3+ concentration, and smaller increases in the Cr/(Cr + Al) ratio and Ti concentration (Figs 6 and 7c). We propose that a number of different processes have contributed to these changes.
First, some compositional changes reflect the magmatic fractional crystallization history of the melt. For example, cumulates in the lower part of the layered body are dunitic and dominated by Mg-rich olivine, whereas in the upper part of the layered body they are harzburgitic and contain slightly more Fe-rich olivine. Experimental studies show that Ti, Al and Cr diffuse slowly at 1100°C in the solid state in ultramafic mineral assemblages, indicating that Ti concentrations and Cr/(Cr + Al) ratios may also be little modified from the magmatic stage of these chromites (Scowen et al., 1991; Kamenetsky et al., 2001). Second, other compositional changes are the result of re-equilibration between chromite and residual melt. This process reflects the evolving compositional evolution of the interstitial melt and leads to iron enrichment in both chromite and of olivine (Roeder & Campbell, 1985). Scowen et al. (1991) showed that, in volcanic rocks, the oxidation of iron in chromite may be the product of a magmatic reaction in which olivine is oxidized to orthopyroxene:
 |
3·65–3·74 Ga amphibolite-facies metamorphic event is indicated by the relatively low (subsolidus) temperatures calculated from olivine–spinel Fe–Mg exchange thermometry and hornblende–plagioclase thermometry. The effects of this exchange were to shift the compositions of main trend chromites coexisting with silicates to the right in Fig. 6. However, this process will not affect the composition of chromites in the chromitites (the most primitive, and most magnesian chromites described here), for they have no other phase with which to exchange Fe–Mg. The net effect of re-equilibration during the amphibolite-facies metamorphism therefore is to ‘stretch’ the main chromite trend towards the right. The Cr/Al ratio will be unaffected.
Origin of the ‘replacement trend’ chromites
Textural evidence given above and presented by Appel et al. (2002) is used to argue that the darker, aluminous spinel phase, found as part of two-phase spinels in the amphibole harzburgites, is of replacement origin. These spinels are more aluminous and magnesian than the ‘main trend’ chromites. Single-phase spinels in anorthosites associated with the harzburgites are even more enriched in Mg and Al. The replacement of what are now two-phase spinels took place either during the late-magmatic evolution of the layered body, or during the later metamorphism. Here we argue that the replacement took place during the late magmatic history of the layered body. Our reasons are threefold.
First, there is no petrographic evidence to unequivocally link the two-phase spinels to the growth of metamorphic phases in these rocks. Assuming that the amphibole in these rocks grew during the amphibolite-facies metamorphism we find in sample 462398 no association between hornblende and the distribution of two-phase spinels. We also note that two-phase spinels are enclosed within olivine grains, protected from the effects of later metamorphism. Further, in sample 472805 there are abundant two-phase spinels, showing a range of grain sizes in a variety of textural associations, both within and outside of olivine, and yet there are no hydrous minerals present in this sample. In contrast, in samples 472806 and 472804 there are abundant grains of hydrous phases and yet only a single-phase spinel comprising the (early) iron-rich component.
Second, experimental and empirical studies show that trivalent cations in spinel diffuse rapidly only in a melt (Scowen et al., 1991; Kamenetsky et al., 2001). Thus a replacement reaction that involves the diffusion of trivalent cations most reasonably occurs through reaction with a melt. This is supported by the observation that the compositions of the replacement spinels lie to the high-temperature side of the 600°C isotherm of Sack & Ghiorso (1991) for compositions in equilibrium with Fo90 (see Fig. 13b, below). Although this does not eliminate the possibility of a metamorphic origin, it does indicate their formation at high temperature.
|
Third, there is no similarity between the textural relationships illustrated by Barnes (2000) for the amphibolite-facies alteration of komatiitic chromites and the replacement spinels described here. The metamorphic alteration described by Barnes (2000) is typified by concentric zoning, whereas the replacement texture described here is extremely irregular.
Thus on the grounds of the observed textures and on the basis of experimental studies we argue that the replacement textures in the two-phase spinels were produced by interaction with a melt. As we shall argue further below, we prefer a model in which Fe-rich chromites from the main chromite trend reacted with interstitial melt with an evolved composition in the cooling magma body to produce the Mg–Al-rich spinels of the replacement trend. Similar melt-reaction textures are observed between olivine and orthopyroxene (Fig. 4d).
|
DISCUSSION |
|---|
The nature of the parent magma
Because of the cumulate nature of the Ujaragssuit nunât layered body we attempt to estimate the composition of the parent magma from the composition of phases that crystallized from the original melt. Olivine cores from dunitic adcumulates are in the range Fo87·2–88·3 and are typical of olivines in equilibrium with a basaltic melt (Kamenetksy et al., 2001). The magmatic parentage of chromites may be determined by using the cr-number, trivalent cation ratios and the TiO2 content from the database of Barnes & Roeder (2001). The most primitive chromites on the main chromite trend have Cr/(Cr + Al) ratios that are higher than those found in chromites in MORB, but similar to those of chromites in oceanic plateau basalts, boninites and komatiites (Rollinson, 1997). It should be noted that although the fe-number for the main chromite trend (Fig. 12) is higher than that observed in chromites in oceanic plateau basalts, boninites and komatiites, the fe-number has probably been enhanced through subsolidus processes. A comparison of trivalent cation plots and TiO2–fe-number plots for the most primitive chromites with the dataset of Barnes & Roeder (2001) also supports an oceanic plateau basalt, boninitic or komatiitic origin. On the spinel TiO2–Al2O3 diagram of Kamenetsky et al. (2001), the most primitive of the main trend chromites plot in the arc-boninite field (Fig. 7d).
|
These observations do not permit an unequivocal statement about the nature of parent magma for the Ujaragssuit nunât area layered body. The high chromite Cr/(Cr + Al) ratios imply that the parental melt was either a high-melt-fraction, high-temperature melt or was derived from a Cr-enriched source (Roberts, 1988; Klingenberg & Kushiro, 1996). The chromite compositions we have observed could have evolved from a basaltic, boninitic or komatiitic parent. We note that Parman et al. (2001) have recently drawn attention to the geochemical similarity between Archaean basaltic komatiites and modern boninites.
Comparison with other layered intrusions
Layered intrusions containing chromites that show two distinct evolutionary trends—one of Fe enrichment and the other of Al enrichment—are known from the British Tertiary Igneous Province (BTIP) in Scotland. Here chromite is commonly found in olivine–plagioclase cumulates, and the comparison with anorthosites is pertinent. Dunham & Wilkinson (1985) and Bell & Claydon (1992) showed that the first-formed chromites in the BTIP layered intrusions were aluminous and that they became progressively enriched in Cr through reaction with trapped intercumulus liquid, plagioclase and olivine (Fig. 13a, trend 1). This trend, identified by Barnes & Roeder (2001) as their ‘Rum trend’, is uncommon in mafic magmatic rocks. Disseminated spinels in the olivine–plagioclase cumulates of the BTIP formed from the Cr-enriched phase by the progressive enrichment in Fe3+ through exchange with interstitial melt and/or adjacent cumulus phases (Fig. 13a, trend 2). A possible model for chromite genesis in the BTIP, adapted from Henderson (1975) is as follows:
- Mg–Al-chromite crystallizes from an aluminous melt (Fig. 13a, most primitive compositions);
- Mg–Al-spinel reacts with trapped intercumulus melt leading to the reaction
(Fig. 13a, trend 1);
- Fe–Cr-spinel continues to exchange with interstitial melt to become enriched in Fe3+ and Ti (Fig. 13a, trend 2). This change in composition reflects the change in composition of the intercumulus liquid before solidification.
Alternatively, it is possible that the data for the BTIP do not represent evolutionary trends, but rather reflect a wide range of differing localized conditions under which chromite equilibrated with the interstitial melt and silicate phases (Roeder & Campbell, 1985).
The magmatic evolution of the Ujaragssuit nunât layered body
A comparison between BTIP chromites and those Ujaragssuit nunât layered body highlights two important points in understanding the evolution of the Ujaragssuit nunât magma. First, in the case of Ujaragssuit nunât there is an obvious chronology of chromite evolution, based upon textural evidence. Second, localized processes controlled by the interstitial melt composition and the silicate environment are very important in understanding chromite compositional evolution.
We propose the following sequence of events:
- Mg–Cr-chromite crystallized from a relatively Al-poor high-temperature melt; as indicated above, the parent magma from which this melt was derived could be basaltic, boninitic or komatiitic (Fig. 13b, most primitive compositions).
- Chromite compositions became Fe enriched during magmatic evolution and this was enhanced by extensive exchange between chromite and residual interstitial melt. These are the chromites of the main trend (Fig. 13b, trend 1). The observed Fe enrichment has been enhanced by later subsolidus equilibration.
- Fe–Cr-rich chromites in the upper part of the intrusion came into contact with a Mg–Al-rich melt. This melt was less oxidizing that the original basaltic melt composition and was in equilibrium with both orthopyroxene and olivine. The following reaction took place:
These are the chromites of the replacement trend (Fig. 13b, trend 2).
We propose that the Mg–Al-rich melt is a residual, interstitial melt that has developed during the cooling and fractionation of the magma body. During the development of the chromites and associated silicates of the main trend the main phases became progressively enriched in Fe and the interstitial melt became depleted in Fe. This melt reacted with Fe-rich spinel to form the two-phase spinels, and with olivine to produce orthopyroxene. It is possible that the processes described here took place on a variety of scales. There is evidence from the stratigraphy of the layered body that there were a number of magma pulses, and these magmas evolved towards more silica- and iron-rich compositions. It may be, therefore, that the magmas in upper cycles of the layered body were slightly different from those in the lower part of the cycle and that the interstitial melt compositions in the upper part of the layered body were even more evolved still.
The results of Delano (2001) suggest that the oxygen fugacity of the initial magma was close to that of the QFM buffer (Fig. 7f), implying a mantle oxidation state at >3·81 Ga similar to that of the present-day mantle. The difference in the oxygen fugacity between the two chromite trends (Fig. 9) is thought to reflect the changing oxygen fugacity of the evolving interstitial melt.
The origin of anorthositic chromitites
The layered body from the Ujaragssuit nunât area of west Greenland contains two distinct chromite compositions, one of which plots on a cr-number vs fe-number diagram in the fields of evolved komatiitic or basaltic chromites, and the other of which plots near those of anorthositic chromites (Fig. 12b). Although a link has been suggested between komatiitic and anorthositic rocks (Phinney et al., 1988) this is not the consensus view (Ashwal, 1993). Nevertheless, the association here between basaltic, boninitic or komatiitic chromites and anorthositic chromites in the same layered body strongly suggests that the two were derived from the same mafic magma. Here we propose that our observations made at Ujaragssuit nunât may be used to explain the origin of anorthositic chromites.
The model we have developed here suggests that a mafic melt derived from a basaltic, boninitic or komatiitic parental magma crystallized chromites that evolved along an Fe-enrichment trend through reaction between chromite and interstitial melt. When this interstitial melt was sufficiently exhausted of Fe, the resultant Mg–Al-rich interstitial melt reacted with Fe-rich chromite to produce a chromite composition very similar to that found in anorthosites. Our model is supported by the observation that during fractional crystallization of the Ujaragssuit nunât intrusion, the magma became more silica rich (orthopyroxene replaces olivine in the stratigraphy), more aluminous (as seen in the orthopyroxene compositions), more Fe rich (as seen in all ferromagnesian phases) and depleted in Cr. This gave rise to the following reaction in the harzburgites in which an evolved, interstitial melt reacted with chromite and olivine:
 |
 |
|
CONCLUSIONS |
|---|
- The chromitite–ultramafic layered body from the Ujaragssuit nunât area is early Archaean in age (>3·81 Ga) and may represent the Earth’s oldest chromitite.
- The layered body has been metamorphosed twice. These events took place at 3·75 Ga and during the late Archaean at 2·8 Ga.
- Chromite compositions preserve features that define two evolutionary trends. There is a main magmatic trend that has been modified during cooling to show strong iron enrichment.
- Chromites from the upper part of the intrusion in harzburgites show magmatic replacement textures in which Fe–Cr-spinels are altered to Mg–Al-spinels. These are thought to have formed through reaction with the interstitial melt of the cooling magma-body.
- Chromite compositions of this type are typical of those found in association with Archaean anorthosites.
- The association between ‘basaltic chromites’ and ‘anorthositic chromites’ in the same small layered body strongly supports the view that anorthositic chromites are formed during the late magmatic evolution of an Fe-rich melt originally derived from a basaltic, boninitic or komatiitic parent.
|
ACKNOWLEDGEMENTS |
|---|
We thank Brian Chadwick, Lew Ashwal, Steve Barnes and David Lowry for their comments on the manuscript, and Charlotte Appel for the scanning electron microscopy work and many valuable discussions. H.R.R. and P.A were supported by the Danish Natural Science Research Council, the Commission for Scientific Research in Greenland, the Bureau of Minerals and Petroleum of the Government of Greenland and a Royal Society European Exchange Programme Grant. H.R.R. acknowledges a grant from the Leverhulme Trust. R.F. was supported by the Danish Lithosphere Centre. H.R.R. thanks Stuart Kearns for assistance and access to the electron microprobe at the University of Bristol, via their EU Geochemical Facility. The paper is published with the permission of the Geological Survey of Denmark and Greenland.
|
FOOTNOTES |
|---|
*To whom Correspondence should be addressed at present address: Sultan Qaboos University, Department of Earth Science, PO Box 36, Postal Code Al Khodh 123, Omen. Telephone: +968 515403. E-mail: hrollin{at}squ.edu.om
|
REFERENCES |
|---|
Appel, C. C., Appel, P. W. U. & Rollinson, H. R. (2002). Complex chromite textures unravel the history of an early Archaean layered ultramafic body in west Greenland. Mineralogical Magazine (in press).
Ashwal, L. D. (1993). Anorthosites. Berlin: Springer, 422 pp.
Baadsgaard, H., Nutman, A. P., Rosing, M., Bridgwater, D. & Longstaffe, F. J. (1986). Alteration and metamorphism of Amîtsoq gneisses from the Isukasia area, West Greenland: recommendations for isotope studies of the early crust. Geochimica et Cosmochimica Acta 50, 2165–2172.[Web of Science]
Ballhaus, C., Berry, R. F. & Green, D. H. (1991). High pressure experimental calibration of the olivine–orthopyroxene–spinel oxygen barometer: implications for the oxidation state of the upper mantle. Contributions to Mineralogy and Petrology 107, 27–40.[Web of Science]
Barnes, S. J. (1998). Chromite in komatiites, I. Magmatic controls on crystallization and composition. Journal of Petrology 39, 1689–1720.[Web of Science]
Barnes, S. J. (2000). Chromite in komatiites, II. Modification during greenschist to mid-amphibolite facies metamorphism. Journal of Petrology 41, 387–409.
Barnes, S. J. & Roeder, P. L. (2001). The range of spinel compositions in terrestrial mafic and ultramafic rocks. Journal of Petrology 42, 2279–2302.
Bell, B. R. & Claydon, R. V. (1992). The cumulus and post-cumulus evolution of chrome-spinels in ultrabasic layered intrusions: evidence from the Cuillin Igneous Complex, Isle of Skye, Scotland. Contributions to Mineralogy and Petrology 112, 242–253.[Web of Science]
Blichert-Toft, J. & Frei, R. (2001). Complex Sm–Nd and Lu–Hf isotope systematics of metamorphic garnets from the Isua Supracrustal Belt, West Greenland. Geochimica et Cosmochimica Acta 65, 3177–3189.[Web of Science]
Chadwick, B. & Crewe, M. A. (1982). Field work in the Ujaragsssuit region, inner Godhabsfjord, southern west Greenland. Rapport Grønlands Geologiske Undersøgelse 110, 57–63.
Chadwick, B. & Crewe, M. A. (1986). Chromite in the early Archaean Akilia association (ca. 3,800 m.y.), Ivisartoq region, inner Godthabsfjord, southern west Greenland. Economic Geology 81, 184–191.
Chadwick, B., Crewe, M. A. & Park, J. W. (1983). Field work in the north of the Ivisartoq region, inner Godhabsfjord, southern west Greenland. Rapport Grønlands Geologiske Undersøgelse 115, 49–56.
Delano, J. W. (2001) Redox history of the Earth’s interior since -3900 Ma: implications for prebiotic molecules. Origins of Life and Evolution of the Biosphere 31, 311–341.[Web of Science]
Droop, G. T. R. (1978). A general equation for estimating Fe3+ concentrations in ferromagnesian silicates and oxides from microprobe analysis, using stoichiometric criteria. Mineralogical Magazine 51, 431–435.
Dunham, A. C. & Wilkinson, F. C. F. (1985). Sulphide droplets and the Unit 11/12 chromite band, Rhum: a mineralogical study. Geological Magazine 122, 539–548.[Abstract]
Frei, R. & Rosing, M. T. (2001). The least radiogenic terrestrial leads; implications for the early Archean crustal evolution and hydrothermal–metasomatic processes in the Isua supracrustal Belt (West Greenland). Chemical Geology 181, 47–66.[Web of Science]
Frei, R., Rosing, M. T., Krogstad, E. J., Storey, M., & Albarède, F. (1998a). Unraveling the tectono-metamorphic history in the Isua supracrustals (west Greenland): mineral Pb stepwise leaching evidence for early- and late Archean metasomatic–metamorphic events. In: Ninth Annual V. M. Goldschmidt Conference. Lunar and Planetary Institute Contribution 971, 91.
Frei, R., Rosing, M. T., Krogstad, E. J., Storey, M., Ulfbeck, D. G. & Albarède, F. (1998b). A pre-3·74 Ga old tectono-metamorphic event in the outer Isua (West Greenland) supracrustal belt: evidence from Pb stepwise leaching of sulfide-bearing quartz–tourmaline veins. EOS Transactions, American Geophysical Union, Abstract Supplement 79(45), F932.
Frei, R., Naegler, T. F., Schonberg, R. & Kramers, J. D. (1998c). Re–Os, Sm–Nd, U–Pb, and stepwise lead leaching isotope systematics in shear-zone hosted gold mineralization: genetic tracing and age constraints of crustal hydrothermal activity. Geochimica et Cosmochimica Acta 62, 1925–1936.[Web of Science]
Frei, R., Bridgwater, D., Rosing, M. & Stecher, O. (1999). Controversial Pb–Pb and Sm–Nd isotope results in the early Archean Isua (West Greenland) oxide iron formation: preservation of primary signatures versus secondary disturbances. Geochimica et Cosmochimica Acta 63, 473–488.[Web of Science]
Frei, R., Rosing, M., Waight, T. E., Krogstad, E., Storey, M., Ulfeck, D. G. & Albarède, F. (2002). Hydrothermal–metasomatic and tectono-metamorphic processes in the Isua supracrustal belt (West Greenland): a multi-isotopic investigation of their effects on the Earth’s oldest oceanic crustal sequence. Geochimica et Cosmochimica Acta 66, 467–486.[Web of Science]
Ghisler, M. (1976). The geology, mineralogy and geochemistry of the pre-orogenic Archaean stratiform chromite deposits at Fiskenaesset, west Greenland. Mineral Deposits Monthly Series 14, 156 pp.
Gruau, G., Rosing, M., Bridgwater, D. & Gill, R. C. O. (1996). Resetting of Sm–Nd systematics during metamorphism of >3·7-Ga rocks: implications for isotopic models of early Earth differentiation. Chemical Geology 133, 225–240.[Web of Science]
Henderson, P. (1975). Reaction trends shown by chrome spinels of the Rhum layered intrusion. Geochimica et Cosmochimica Acta 39, 1035–1044.[Web of Science]
Holland, T. & Blundy, J. (1994). Non-ideal interactions in calcic amphiboles and their bearing on amphibole–plagioclase thermometry. Contributions to Mineralogy and Petrology 116, 433–447.[Web of Science]
Irvine, T. N. (1965). Chromium spinel as a petrogenetic indicator. Part I. Theory. Canadian Journal of Earth Sciences 2, 648–672.
Irvine, T. N. (1967). Chromium spinel as a petrogenetic indicator. Part II. Petrological applications. Canadian Journal of Earth Sciences 4, 71–103.
Kamenetsky, V. S., Crawford, A. J. & Meffre, S. (2001). Factors controlling chemistry of magmatic spinel: an empirical study of associated olivine, Cr-spinel and melt inclusions from primitive rocks. Journal of Petrology 42, 655–671.
Klingenberg, B. M. E. T. & Kushiro, I. (1996). Melting of a chromite-bearing harzburgite and generation of boninitic melts at low pressures under controlled oxygen fugacity. Lithos 37, 1–13.[Web of Science]
Leake, B. E., et al. (1997). Nomenclature of amphiboles: report of the subcommittee on amphiboles of the international mineralogical association commission on new mineral names. Mineralogical Magazine 61, 295–321.[Abstract]
LeMaitre, R. W. (ed.) (1989). A Classification of Igneous Rocks and Glossary of Terms. Oxford: Blackwell Scientific.
Ludwig, K. R. (1988) A computer program to convert raw U–Th–Pb isotope ratios to blank corrected isotope ratios and concentrations with associated error correlations. US Geological Survey Open File Report. OF-82-820.
McCandless, T. E., Ruiz, J., Adair, B. I. & Freydier, C. (1999). Re–Os isotope and Pd/Ru variations in chromitites from the critical zone, Bushveld complex, South Africa. Economic Geology 63, 911–923.
Naegler, T. F. & Frei, R. (1997). ‘Plug in plug’ Os distillation. Schweizerische Mineralogische und Petrographische Mitteilungen 77, 123–127.
Naegler, T. F., Kramers, J. D., Kamber, B. S., Frei, R. & Prendergast, M. D. A. (1997). Growth of subcontinental lithospheric mantle beneath Zimbabwe started at 3·8 Ga: a Re–Os study on chromites. Geology 92, 747–750.
Nutman, A. P., McGregor, V. R., Friend, C. R. L., Bennett, V. C. & Kinney, P. D. (1996). The Itsaq Gneiss Complex of southwest Greenland: the world’s most extensive record of early crustal evolution (3900–3600 Ma). Precambrian Research 78, 1–39.[Web of Science]
Parman, S. W., Grove, T. L. & Dann, J. C. (2001). The production of Barberton komatiites in an Archaean subduction zone. Geophysical Research Letters 28, 2513–2516.[Web of Science]
Phinney, W. C., Morrison, D. A. & Maczuga, D. E. (1988). Anorthosites and related megacrystic units in the evolution of the Archaean crust. Journal of Petrology 29, 1283–1323.
Puchtel, I. S., Brugmann, G. E. & Hofmann, A. W. (1999). Precise Re–Os mineral isochron and Pb–Nd–Os isotope systematics of a mafic–ultramafic sill in the 2·0 Ga Onega plateau (Baltic Shield). Earth and Planetary Science Letters 170, 447–461.[Web of Science]
Renner, R., Nisbet, E. G., Cheadle, M. J., Arndt, N. T., Bickle, M. J. & Cameron, W. E. (1994). Komatiite flows from the Reliance Formation, Belingwe Belt, Zimbabwe: I. Petrography and mineralogy. Journal of Petrology 35, 361–400.
Roberts, S. (1988). Ophiolitic chromite formation: a marginal basin phenomenon? Economic Geology 83, 1034–1036.
Roeder, P. L. & Campbell, I. H. (1985). The effect of postcumulus reactions on composition of chrome-spinels from the Jimberlana intrusion. Journal of Petrology 26, 763–786.
Roeder, P. L. & Reynolds, I. (1991). Crystallisation of chromite and chromium solubility in basaltic melts. Journal of Petrology 32, 909–934.
Roeder, P. L., Campbell, I. H. & Jamieson, H. E. (1979). Re-evaluation of the olivine–spinel geothermometer. Contributions to Mineralogy and Petrology 68, 325–334.[Web of Science]
Rollinson, H. R. (1995). Composition and tectonic settings of chromite deposits through time—a discussion. Economic Geology 90, 2091–2092.
Rollinson, H. R. (1997). The Archaean komatiite-related Inyala chromitite, southern Zimbabwe. Economic Geology 92, 98–107.
Rollinson, H. R. (2002). The metamorphic history of the Isua Greenstone Belt, West Greenland. In: Fowler, C. M. R., Ebinger, C. J. & Hawkesworth, C. J. (eds) The Early Earth: Physical, Chemical and Biological Development. Geological Society, London, Special Publications (in press).
Roy-Barman, M. & Allègre, C.-J. (1995). 187Os–186Os ratios of mid-ocean ridge basalts and abyssal peridotites. Earth and Planetary Science Letters 129, 145–161.[Web of Science]
Sack, R. O. & Ghiorso, M. S. (1991). Chromian spinels as petrogenetic indicators: thermodynamics and petrological applications. American Mineralogist 76, 827–847.[Abstract]
Scowen, P. A. H., Roeder, P. L. & Helz, R. T. (1991). Reequilibration of chromite within Kilauea Iki lava lake, Hawaii. Contributions to Mineralogy and Petrology 107, 8–20.[Web of Science]
Shirey, S. B. & Walker, R. J. (1995). Carius tube digestions for low-blank rhenium–osmium analysis. Analytical Chemistry 67, 2136–2141.
Steele, I. M., Bishop, F. C., Smith, J. V. & Windley, B. F. (1977). The Fiskenaesset complex, west Greenland. Part III. The chemistry of silicate and oxide minerals from oxide-bearing rocks, mostly from Qeqertarssuatsiaq. Bulletin Grønlands Geologiske Undersøgelse 124, 38 pp.
Todt, W., Cliff, R. A., Hanser, A. & Hofmann, A. W. (1993). Re-calibration of NBS lead standards using a 202Pb + 205Pb double spike. Terra Abstracts 5(Supplement 1), 396.
Wilson, A. H. (1982). The geology of the Great ‘Dyke’, Zimbabwe: the ultramafic rocks. Journal of Petrology 23, 240–292.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. D. Prendergast Archean Komatiitic Sill-hosted Chromite Deposits in the Zimbabwe Craton Economic Geology, August 1, 2008; 103(5): 981 - 1004. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||