Journal of Petrology Advance Access originally published online on June 17, 2006
Journal of Petrology 2006 47(10):1973-1996; doi:10.1093/petrology/egl033
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Temperature and Bulk Composition Control on the Growth of Monazite and Zircon During Low-pressure Anatexis (Mount Stafford, Central Australia)
1 RESEARCH SCHOOL OF EARTH SCIENCES, THE AUSTRALIAN NATIONAL UNIVERSITY CANBERRA 0200 ACT, AUSTRALIA
2 SCHOOL OF GEOSCIENCES, MONASH UNIVERSITY MELBOURNE 3800 VIC., AUSTRALIA
RECEIVED OCTOBER 4, 2005; ACCEPTED MAY 15, 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONREGIONAL SETTING AND...METHODSSAMPLE DESCRIPTIONZIRCON AND MONAZITE PETROGRAPHY...TRACE ELEMENT COMPOSITIONDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The formation, age and trace element composition of zircon and monazite were investigated across the prograde, low-pressure metamorphic sequence at Mount Stafford (central Australia). Three pairs of inter-layered metapelites and metapsammites were sampled in migmatites from amphibolite-facies (T
600°C) to granulite-facies conditions (T 800°C). Sensitive high-resolution ion microprobe U–Pb dating on metamorphic zircon rims and on monazite indicates that granulite-facies metamorphism occurred between 1795 and 1805 Ma. The intrusion of an associated granite was coeval with metamorphism at 1802 ± 3 Ma and is unlikely to be the heat source for the prograde metamorphism. Metamorphic growth of zircon started at T 750°C, well above the pelite solidus. Zircon is more abundant in the metapelites, which experienced higher degrees of partial melting compared with the associated metapsammites. In contrast, monazite growth initiated under sub-solidus prograde conditions. At granulite-facies conditions two distinct metamorphic domains were observed in monazite. Textural observations, petrology and the trace element composition of monazite and garnet provide evidence that the first metamorphic monazite domain grew prior to garnet during prograde conditions and the second in equilibrium with garnet and zircon close to the metamorphic peak. Ages from sub-solidus, prograde and peak metamorphic monazite and zircon are not distinguishable within error, indicating that heating took place in less than 20 Myr. KEY WORDS: accessory phases; anatexis; trace element partitioning; U–Pb dating
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL SETTING AND...METHODSSAMPLE DESCRIPTIONZIRCON AND MONAZITE PETROGRAPHY...TRACE ELEMENT COMPOSITIONDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The geochronology of amphibolite- to granulite-facies rocks is commonly dependent on the U–Pb dating of zircon and monazite. Thus the understanding of the behaviour of these two accessory minerals during medium- to high-temperature metamorphism is crucial for age interpretation. The processes by which zircon and monazite (re)crystallize (as a result of melting reactions or melt crystallization, solid-state metamorphic reactions or fluid infiltration), the conditions in which they form, the events that they record, and how the P–T conditions of their formation can be established have recently been under intense scrutiny. Significant work has focused on characterizing the appearance (zoning patterns, textures and inclusion assemblages) and trace element composition of metamorphic zircon (Williams & Claesson, 1987
; Heaman et al., 1990; Hinton & Upton, 1991; Williams et al., 1996; Williams, 2001; Rubatto, 2002; Corfu et al., 2003; Hoskin & Schaltegger, 2003) and on the processes leading to its formation (Fraser et al., 1997; Roberts & Finger, 1997; Degeling et al., 2001; Williams, 2001). The way that monazite is affected by metamorphism has also been extensively investigated (e.g. Rubatto et al., 2001; Williams, 2001; Wing et al., 2003; Fitzsimons et al., 2005): monazite will first form at lower metamorphic grade (lower to mid-amphibolite-facies conditions) than zircon and may lose its Pb memory as a result of solid-state resorption or re-precipitation reactions earlier during prograde metamorphism. In contrast, zircon generally, but not always (Degeling et al., 2001), forms in the presence of melt and once formed retains its isotopic memory up to extremely high temperatures (Vavra et al., 1996; Oliver et al., 1999; Rubatto et al., 2001; Williams, 2001).
Most recently, the focus has shifted toward the petrology of zircon and monazite, their equilibrium with major rock-forming metamorphic minerals and their trace element compositions (e.g. Pyle & Spear, 1999; Degeling et al., 2001; Foster et al., 2002; Rubatto, 2002; Hermann & Rubatto, 2003; Hoskin & Schaltegger, 2003; Whitehouse & Platt, 2003; Wing et al., 2003; Hokada & Harley, 2004; Kelly & Harley, 2005; Watson & Harrison, 2005; Buick et al., 2006). Trace elements have been used for zircon thermometry (Watson & Harrison, 2005) and for assessing equilibrium between zircon and monazite and other minerals (Foster et al., 2002; Rubatto, 2002; Hermann & Rubatto, 2003; Whitehouse & Platt, 2003; Hokada & Harley, 2004; Kelly & Harley, 2005; Buick et al., 2006).
From the large number of studies of metamorphic U–Pb minerals it also appears that there might be a difference in metamorphic behaviour of zircon and monazite according to host-rock composition; for example, the formation of monazite in metasediments depends on the Ca and/or Al content of the rock (Wing et al., 2003; Fitzsimons et al., 2005). Differences are also found in the distribution of trace elements between zircon and garnet at different temperatures (see Rubatto et al., 2001; Hokada & Harley, 2004).
We address the issue of the response of zircon and monazite to prograde metamorphism and partial melting using petrography, sensitive high-resolution ion microprobe (SHRIMP) U–Pb dating and trace element analysis for a representative suite of rocks from Mount Stafford (central Australia). Mount Stafford is host to a well-studied Palaeoproterozoic low-pressure metamorphic sequence from greenschist to granulite facies, which reached melting conditions below 4 kbar, partly in the andalusite field (Greenfield et al., 1998; White et al., 2003). The detailed work of White et al. (2003) provides temperature estimates, as well as melting reactions and calculated melt proportions in different rock types (pelitic and psammitic protoliths, respectively). This provides an excellent basis to study the behaviour of zircon and monazite as a function of temperature, bulk-rock composition and degree of partial melting in rocks. We address these issues using zircon and monazite petrography, SHRIMP U–Pb ages and trace element contents for a representative suite of rocks.
|
REGIONAL SETTING AND METAMORPHISM |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL SETTING AND...METHODSSAMPLE DESCRIPTIONZIRCON AND MONAZITE PETROGRAPHY...TRACE ELEMENT COMPOSITIONDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Mount Stafford is located
170 km NW of Alice Springs in central Australia. It constitutes the NW end of the Anmatjira–Reynolds Range area, within the Proterozoic Arunta Inlier (Fig. 1a). The Mount Stafford area is dominated by metamorphosed pelitic to psammitic sediments (Mount Stafford Beds) and large deformed granitoids (Mount Stafford granite). The majority of the area studied consists of aluminous metapelites, interbedded with metapsammite layers on a centimetre to metre scale (Vernon et al., 1990; Greenfield et al., 1998; White et al., 2003). Rare metamorphosed diorite sills cross-cut the sedimentary sequence. According to Collins et al. (1991), the Mont Stafford metasediments and the intruding diorite sills underwent three phases of deformation ascribed to the main tectonic cycle. Deformation mainly translated into folding from tight, metre-scale folds to open, isoclinal and kilometre-scale folds. Opposite to folding, foliation and lineation are generally weak and localized.
|
The metamorphic conditions at Mount Stafford increase from greenschist to granulite grade from SW to NE and towards the contact with the granite over a distance of
10 km (Clarke et al., 1990; Vernon et al., 1990; Collins & Vernon, 1991; Greenfield et al., 1998). This reflects a high geothermal gradient of 75°C per vertical kilometre (Vernon et al., 1990). The strong temperature increase across the sequence (500–810°C, Greenfield et al., 1998; White et al., 2003) is accompanied by a small pressure increase (from 2·3–2·8 to 3·3–4·0 kbar, White et al., 2003). The metamorphic sequence has been classically subdivided into a number of zones (Zones 1–5, Fig. 1a) of progressively increasing grade and degree of partial melting (Vernon et al., 1990; Greenfield et al., 1996, 1998). Zone 1 lacks evidence of melting and is characterized by low-grade muscovite quartz schists. Zone 2 is defined by the sub-solidus breakdown of muscovite to andalusite and K-feldspar (Zone 2a) and then by localized vapour-present melting (Zone 2b). The main rock assemblage is characterized by andalusite, cordierite and K-feldspar, with sillimanite appearing in the high-grade portion of the zone (Zone 2c). Metasediments in Zone 3 are migmatites that contain spinel, sillimanite, cordierite and K-feldspar, with occasional garnet. In Zone 4 the migmatites may additionally contain garnet and/or orthopyroxene. In Zone 5 the main rock type is a diatexite with biotite, cordierite and plagioclase, which is a physical mixture of migmatitic pelitic and psammitic rocks injected by magma from the granite. This simplified classification in zones is, however, complicated by the layered nature of the sequence and the reactions occurring in the different bulk rocks (White et al., 2003). A comprehensive documentation of the metamorphic reactions and conditions has been given by White et al. (2003), who estimated peak metamorphism at around 800°C and 3·3–4·0 kbar. Whereas in the lower part of Zone 2 muscovite + quartz broke down to form K-feldspar-bearing assemblages via sub-solidus dehydration reactions, a slightly higher grade (Zone 2b) melting began in the andalusite stability field, an unusual occurrence (Clarke et al., 2005). The wet pelite solidus can intersect the andalusite–sillimanite transition if the Pattison (1992) Al2SiO5 polymorph triple point is used; in this case, andalusite and silicate melt can coexist at pressures of 2–2·5 kbar (Cesare et al., 2003). Overall, the melting history at Mount Stafford is controlled by a series of biotite breakdown reactions experienced by the different rock types (metapsammites vs metapelites) at slightly different conditions over a P–T range and according to bulk-rock composition (White et al., 2003). Thus, the banded nature of the protolith controls the rock's texture and the degree of melting at any given metamorphic grade. Compositional heterogeneity also controls the assemblages and the presence or absence of key minerals (e.g. garnet, orthopyroxene, aluminium silicate), and can produce micro-domains with different parageneses (White et al., 2003). Melt mobilization over large distances appears to be limited, with most rocks containing the expected proportion of leucosome (White et al., 2003). This is confirmed by the relatively constant bulk-rock composition of metapelites and metapsammites across the various zones (Greenfield et al., 1996).
The Mount Stafford Beds appear to be equivalent to the more geographically widespread Lander Rock Beds, which occur in the south and NW of Mount Stafford. The Lander Rocks Beds have long been thought to have been deposited at 1870 Ma (Blake & Page, 1988). However, more recent detrital zircon analysis suggested deposition after 1838 Ma (Vry et al., 1996). The intrusion of the Mount Stafford granite has previously been estimated at 1818 ± 15 Ma through dating of magmatic zircon overgrowths (Collins & Williams, 1995). Relationships between granite intrusion and high-grade metamorphism are complex: (1) the granite cuts the mapped metamorphic isograds (Clarke et al., 1990; Collins & Williams, 1995; Greenfield et al., 1996), and itself has as a metamorphic aureole (Collins & Vernon, 1991; Collins & Williams, 1995); (2) the diatexite formed during metamorphism has a gradational contact with the northern part of the granite (Clarke et al., 1990; Greenfield et al., 1996); (3) xenoliths of metamorphic country rocks are contained in the granite (Collins & Williams, 1995). In summary, the overall picture is one of interrelated metamorphism and magmatism (Vernon et al., 1990; Collins & Vernon, 1991).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL SETTING AND...METHODSSAMPLE DESCRIPTIONZIRCON AND MONAZITE PETROGRAPHY...TRACE ELEMENT COMPOSITIONDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
The major element composition of garnet was determined on a JEOL 6400 SEM (Electron Microscopy Unit, ANU) and on a Cameca SX100 electron microprobe (EMP) using an acceleration voltage of 15 kV and a beam current of 1 and 10 nA, respectively. Garnet element mapping was performed on the Cameca SX100 with a voltage of 15 kV, current of 150 nA, a focused beam, 5 µm steps and a dwell time of 100 ms. Ca K
was measured on a PET crystal and Y L on a TAP crystal.
Zircon and monazite grains were prepared as mineral separates mounted in epoxy and polished to expose the grain centres. Garnet was analysed in polished thin sections. Trace element analyses were performed by laser ablation–inductively coupled plasma mass spectrometry (LA-ICPMS) at the Research School of Earth Sciences (RSES) using a pulsed 193 nm ArF excimer laser with 100 or 70 mJ energy at a repetition rate of 5 Hz (Eggins et al., 1998) coupled to an Agilent 7500 quadrupole ICPMS system. During the time-resolved analysis of minerals, the contamination from inclusions, fractures and zones of different composition was detected by monitoring several elements and integrating only the relevant part of the signal. A spot size of 23 µm (zircon and monazite), and 83 or 142 µm (garnet) was used. External calibration was performed relative to NIST 612 glass using the concentrations given by Pearce et al. (1997). Internal standards were Ca and Si for garnet, as measured by EMP, and stoichiometric Si and Ce for zircon and monazite, respectively.
Cathodoluminescence (CL) investigation of zircon was carried out at the Electron Microscope Unit, Australian National University, with an Hitachi S2250-N scanning electron microscope working at 15 kV, 60 µA and 20 mm working distance. Back-scattered electron (BSE) images were obtained with a Cambridge S360 scanning electron microscope using a voltage of 20 kV, current of 2 nA and a working distance of 15–20 mm.
U–Th–Pb analyses were performed using SHRIMP I and II at the RSES. Instrumental conditions and data acquisition for zircon analysis were generally as described by Compston et al. (1984). For monazite, energy filtering was used to eliminate the interference on 204Pb, as described by Rubatto et al. (2001). The data were collected in sets of six or seven scans throughout the masses. The measured 206Pb/238U ratio was corrected using reference zircon from a gabbro of the Duluth Complex in Minnesota (AS3, 1099 Ma, Paces & Miller, 1993) and reference monazite from the Thompson mine (1766 Ma). A zircon of known composition (SL 13) was used to determine the U content of zircon. The data were corrected for common Pb on the basis of the measured 204Pb, as described by Compston et al. (1984). The low common Pb content of both standard and unknowns indicated that the common Pb was mainly surface-derived and instrumental background. Therefore, the common Pb composition was assumed to be that of Broken Hill galena, which approximates the laboratory common Pb background at the RSES. Ages were calculated using the software Isoplot/Ex (Ludwig, 2000). Isotopic ratios and single ages are reported with 1
error, whereas mean ages are given at the 95% confidence level (c.l.).
|
SAMPLE DESCRIPTION |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL SETTING AND...METHODSSAMPLE DESCRIPTIONZIRCON AND MONAZITE PETROGRAPHY...TRACE ELEMENT COMPOSITIONDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Several pairs of inter-layered samples were collected at different metamorphic grades in the prograde sequence. Across the Mount Stafford terrane, melting did not lead to the formation of discrete melanosomes and leucosomes at the sample scale (Greenfield et al., 1996
; White et al., 2003). Instead, the banded nature of the rocks at all grades mostly reflects the original sedimentary layering [the ‘bedded migmatites’ of Greenfield et al. (1998)]. On an outcrop scale, at grades above the solidus, metapelites have generally melted to a greater extent than inter-layered metapsammites (e.g. White et al., 2003). Three pairs of samples were investigated in detail; each comprises a metapelite, in which a higher degree of melting was reached, and the associated metapsammite, in which generally melting was less pronounced. Sample locations are shown in Fig. 1a and assemblages are given in Table 1.
|
SGL1 is a metapelite and SGP1 a nearby (<1 m) metapsammite, which form decimetre-scale bands that reflect the original layering of the sedimentary precursor. SGLI has a weak foliation defined by biotite and sillimanite, and also contains deformed quartz and cordierite. The rock also contains skeletal garnet with inclusions of biotite, sillimanite and ilmenite. The presence of garnet at this low metamorphic grade [Zone 2 of Greenfield et al. (1998)
] is unexpected. The crystals appear to be corroded, are deformed along the foliation, and thus are in a different textural setting from the late symplectitic garnet described by Clarke et al. (1990). The composition of this garnet is also different from the garnet found in higher-grade samples: it is poorer in CaO (0·3–0·5 vs 1·0–1·1 wt %) and in MnO (0·35–0·75 vs 1·0–1·6 wt %). The same rock contains large sillimanite blasts wrapped by a fibrolite and biotite fabric; although the garnet and the sillimanite blasts are never in contact, both minerals could belong to the same paragenesis. SGP1 is a fine-grained, light-coloured metapsammite poor in biotite (<10%) and rich in quartz and K-feldspar. Other rocks from the same outcrop contain andalusite: this places the sample at the boundary of Zones 2b and 2c (Fig. 1), corresponding to the andalusite–sillimanite transition according to the scheme of Greenfield et al. (1998). Metamorphic conditions at this boundary are estimated at T 650°Cand P 2·5 kbar (Greenfield et al., 1998; White et al., 2003).
SGL3 corresponds to a metapelite and SGP3 is a nearby metapsammite that shows evidence of limited melting. Both samples are sillimanite absent, as are similar rocks described elsewhere in Zones 3 and 4 (Greenfield et al., 1998; White et al., 2003). They were collected in the area where partial melt was widespread (15–25% in metapelites according to previous estimates; Greenfield et al., 1998; White et al., 2003), but still located in bands according to the sedimentary layers (Fig. 1b). Cordierite and spinel symplectites are common in the metapelite. In contrast, garnet and orthopyroxene are present in the metapsammite as peritectic products of biotite dehydration melting reactions. The development of garnet- and orthopyroxene-bearing assemblages at Mount Stafford is consistent with P–T conditions of 750–800°C and 3–4 kbar (White et al., 2003).
SGL5 and SGM5 were collected closer to the granite (Zone 4, Fig. 1), where the peak temperature conditions were 800°C (Greenfield et al., 1998; White et al., 2003), and a higher percentage of partial melting occurred (20–30% in metapelites; Greenfield et al., 1998; White et al., 2003). In this area, the regular banded texture and layers of metapelites and metapsammites seen at lower grades is less evident, and the rocks are schlieren migmatites composed of a diatexite matrix (SGL5) with ribbons or lenses of darker material (SGM5, Fig. 1c). Garnet is present in both SGM5 and SGL5 and contains frequent biotite inclusions, suggesting that it formed at the expense of biotite. Cordierite is significantly more abundant in SGM5 (60% by volume) than in the diatexite SGL5. Rare, interstitial orthopyroxene is present in SGM5.
Sample SGRA represents the Mount Stafford granite and was collected within the eastern part of the body. The rock is primarily composed of K-feldspar and quartz with very minor plagioclase. Biotite and cordierite occur locally, partly corroded or rimmed by feldspar, and appear to be xenocrysts. Spinel occurs as aggregates or trails in the interstices between feldspars.
|
ZIRCON AND MONAZITE PETROGRAPHY AND GEOCHRONOLOGY |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL SETTING AND...METHODSSAMPLE DESCRIPTIONZIRCON AND MONAZITE PETROGRAPHY...TRACE ELEMENT COMPOSITIONDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
SGL1–SGP1
Monazite was separated from both lower-grade samples, SGL1 and SGP1. Monazite grains are yellow, clear and subhedral. In BSE images (Fig. 2), monazite from sample SGL1 generally shows dark cores of irregular shape surrounded by brighter, homogeneous rims. SHRIMP dating did not identify any significant age difference between cores and rims (Electronic Appendix 1, available at http://www.petrology.oxfordjournals.org): 12 analyses out of 13 yielded an average 207Pb/206Pb age of 1805 ± 4 Ma (MSWD = 1·8; 95% c.l., Fig. 3). One discordant analysis yielded a slightly older age and was thus excluded from the calculation.
|
|
In the metapsammite SGP1, the BSE imaging revealed a more complex texture, with different zoning patterns varying from weak oscillatory, to patchy, to homogeneous rims on dark cores (Fig. 3). The complexity in the zoning is partly confirmed by U–Pb dating; the ages scatter between 1797 and 1883 Ma (Electronic Appendix 1, http://www.petrology.oxfordjournals.org). However, zoning and ages do not necessarily correlate. Two of the older ages (1878 and 1871 Ma) were measured on the only monazite grain with oscillatory zoning that was dated, which also has significantly higher Th/U. Another age of 1883 Ma was measured on a patchy-zoned domain; however, similarly zoned domains in other grains yielded younger ages. Of the 17 analyses, seven were excluded because they were significantly older, high in common Pb and/or discordant. The remaining 10 analyses yielded an average 207Pb/206Pb age of 1805 ± 3 Ma (MSWD = 1·6; 95% c.l., Fig. 3). The other analyses scatter in age between 1819 and 1883 Ma (207Pb/206Pb ages within ±10% discordance) and correspond to domains with commonly lower U and higher Th/U.
Zircon was recovered only from metapsammite SGP1. The crystals are mainly euhedral with rounded edges, but a number of crystals are fragments. They are brownish with abundant inclusions. The internal zoning in CL is highly variable, from bright oscillatory to dark domains where no zoning pattern can be distinguished. Commonly, a core–rim structure was observed. SHRIMP analyses were performed exclusively on dark rims, which were considered more likely to have formed during anatexis. Some of these rims have Th/U as low as 0·01, which, together with the poor zoning, might suggest a metamorphic origin (e.g. Williams et al., 1996; Rubatto et al., 2001). The 207Pb/206Pb ages range between 1853 and 2063 Ma (Electronic Appendix 1, http://www.petrology.oxfordjournals.org). The major components of this zircon population are five analyses on three grains at 1856 ± 3 Ma and five others at 1873 ± 4 Ma. None of the zircon appears to have grown during the event that produced the 1805 Ma monazite.
SGL3–SGP3
Both samples contain yellow, clear monazite grains that preserve some crystal faces, but are generally not euhedral. In BSE images monazite displays a core–rim texture with dark cores with patchy zoning, and homogeneous brighter rims (Fig. 2). The zoning correlates with the U content of the monazite (zones with higher BSE emission have higher U content) and with Th/U (core = 2, rims >10; see also the trace element section). The analyses (except for two) were performed on rims because cores are generally smaller than 25 µm. All 10 analyses, including the two rim analyses, form a tight group with an average 207Pb/206Pb age of 1797 ± 5 Ma (MSWD = 0·95; 95% c.l., Fig. 3).
In metapsammite SGP3, monazite is abundant in the matrix and occasionally included in garnet. Garnet in this rock is highly skeletal and contains few inclusions. In BSE images monazite displays a strong core–rim structure, with large patchy-zoned cores and unzoned rims (Fig. 2). As inheritance was suspected, most of the analyses were performed on the rims and yielded ages with an average 207Pb/206Pb of 1790 ± 7 Ma (MSWD = 0·55; 95% c.l., Fig. 3). One analysis has a significantly younger age and was excluded from the calculation. The analyses on cores or overlapping core and rim (some of the rims were too small) have, compared with the rims, higher U, lower Th/U and older ages, up to 2448 ± 4 Ma for analysis 1.2, which corresponds to a core (Electronic Appendix 1, http://www.petrology.oxfordjournals.org). Monazite contains inclusions of metamorphic minerals such as cordierite and biotite.
In metapelite SGL3 the majority of zircon crystals are clear, pink and euhedral. In sample SGP3, a large number of crystals are rounded and with a rough surface. In this sample a large fraction (70% of grains) is dark, non-transparent, red to brown and rounded. CL imaging revealed that in both samples the internal oscillatory zoning pattern is truncated at the surface (Fig. 4), indicating that abrasion or corrosion of the zircon occurred after formation. The zircons commonly show a core–rim structure in CL images. A significant number of crystals, more abundant in sample SGP3, are dark in CL, and hardly any zoning can be distinguished in these grains. They probably correspond to the dark, non-transparent, coloured crystals observed in transmitted light. This zircon type contains inclusions of major rock-forming minerals (K-feldspar, biotite, quartz, plagioclase and ilmenite) as well as uncommon accessory minerals such as xenotime, U and Pb oxides, and small grains of light rare earth element (LREE) phosphate(s).
|
SHRIMP analyses were concentrated in the outer rims of the crystals and in the dark domains, which, because of their outer position and/or poor zoning were considered more likely to have formed during metamorphism. The pattern of ages obtained is similar in both samples (Electronic Appendix 1, http://www.petrology.oxfordjournals.org). They scatter between 1802 Ma (1818 Ma in SGP3) and
2500 Ma with some clustering at around 1820–1840–1870 Ma (average of two samples: 1823 ± 5 Ma, n = 8; 1842 ± 6 Ma, n = 6; 1871 ± 5 Ma, n = 5). Low Th/U values were measured for one of the domains in each age group. Sample SGL3 has a younger component represented by three analyses between 1802 and 1805 Ma. The analyses around 1805 Ma all have low Th/U (0·05–0·01) whereas in both samples the majority, but not all, of the older ages correspond to domains with Th/U >0·1. In sample SGP3 zircon domains dated at 1820 Ma (207Pb/206Pb age more than 95% concordant) contain inclusions of biotite, whereas cordierite is present at the boundary between a 1820 and 1860 Ma domain, both domains having Th/U >0·1.
SGL5–SGM5
Monazite in sample SGL5 has been observed as inclusions in garnet and in the matrix (Fig. 5). Monazite from both samples has similar characteristics: they are clear, light yellow, and mainly euhedral. The crystals display a patchy-zoned core surrounded by a more homogeneous or oscillatory-zoned rim (Fig. 2). Monazite cores are also distinguished by a Th/U ratio below 10, whereas the rims have Th/U of the order of 30–50 (see trace element section, below). In both samples, there is a tendency for the monazite rims to yield lower ages than the cores, but the difference cannot be resolved with the present analytical precision. However, on the basis of the BSE information the 15 analyses from SGL5 monazite yielded an age of 1803 ± 8 Ma for the cores, and 1798 ± 6 Ma for the rims (mean age 1800 ± 5 Ma). Monazite in sample SGM5 yielded a core age of 1811 ± 7 Ma and a rim age of 1801 ± 10 (mean age 1806 ± 6 Ma).
|
Zircon crystals have similar features in both samples: they are clear, pink to brownish, elongated, and mainly euhedral. The CL investigation confirmed this similarity between the two samples: the zircon crystals generally have a composite core with various oscillatory-zoned patterns that commonly cross-cut each other (Fig. 4). The core is commonly surrounded by a homogeneous overgrowth that can range in thickness from a few microns to 50 µm on some crystal tips. The main difference between the samples is that in the diatexite SGL5 the overgrowths are more abundant and generally thicker. Isotopic measurements were mainly carried out on the overgrowths. Of the 17 analyses on overgrowths in sample SGL5, one with particularly high common Pb yielded a younger age (1754 ± 16 Ma). The remaining 15 analyses are from domains with low Th/U (0·12–0·07) and define a tight cluster with an average 207Pb/206Pb age of 1797 ± 6 Ma (Fig. 6). Two analyses on zircon cores yielded 1854 ± 11 Ma and 2167 ± 11 Ma and are characterized by a higher Th/U of 0·9 and 0·5, respectively.
|
For sample SGM5, only seven analyses were possible on the overgrowths because of their small size and low abundance. U, Th and Th/U compositions are similar to those of the overgrowths in sample SGL5 (Electronic Appendix 1, http://www.petrology.oxfordjournals.org). The analyses cluster together and define a 207Pb/206Pb age of 1792 ± 9 Ma (Fig. 6), which is indistinguishable from that of the zircon overgrowths in SGL5.
SGRA
The granite contains large monazite crystals that are clear, yellow and mainly euhedral. In BSE images, they have a zoning pattern characterized by a somewhat chaotic or homogeneous core, which is surrounded by a rim with oscillatory growth zoning (Fig. 2). This kind of zoning is consistent with a magmatic origin for the monazite rims. SHRIMP dating revealed no differences in age between monazite cores and rims. Twelve analyses define an average 207Pb/206Pb age of 1802 ± 3 Ma (Fig. 3), which is indistinguishable from the age of the metamorphic zircon rims.
Zircon is very similar in aspect to that of the diatexite SGL5. The CL features are also similar, with large unzoned overgrowths on oscillatory-zoned cores (Fig. 4). Ten dated overgrowths yielded 207Pb/206Pb ages that tightly cluster around an average of 1805 ± 3 Ma (Fig. 6). A single core measured yielded an older age of 1865 ± 5 Ma. This analysis is characterized by a higher Th/U (0·13) than measured for the overgrowths (0·01–0·04).
|
TRACE ELEMENT COMPOSITION |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL SETTING AND...METHODSSAMPLE DESCRIPTIONZIRCON AND MONAZITE PETROGRAPHY...TRACE ELEMENT COMPOSITIONDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Trace element composition of zircon was obtained for sample pairs SGP3–SGL3 and SGL5–SGM5, whereas monazite data were also collected for samples SGL1 and SGP1, to test their homogeneity and correlate the chemical variation with age. For the interpretation of zircon and monazite response to metamorphism it is essential that only domains that formed during metamorphism are considered and not potential inherited domains. The petrography and dating provided evidence that domains with different appearances display indistinguishable metamorphic ages. Here we will investigate whether or not these domains have different trace element characteristics. The trace element data are subdivided into the following three categories: (1) inherited domains (>1820 Ma); (2) metamorphic cores or inner domains (
1800 Ma); (3) metamorphic rims (1800 Ma).
The trace element composition of garnet was also measured because garnet is the major phase rich in trace elements, particularly heavy rare earth elements (HREE), present in this rock. Biotite, cordierite and feldspar are generally poor in REE and Y, although they might contain some large ion lithophile elements (LILE) and LREE (e.g. Bea et al., 1994; Yang et al., 1999; Yang & Rivers, 2002; Buick et al., 2006).
Monazite
The chondrite-normalized (Sun & McDonough, 1989) trace element pattern of monazite is characterized by a strong relative enrichment in LREE (500 000–550 000 times chondrite for La, a major element in monazite) with a progressive and smooth decrease in normalized abundance toward the HREE, and generally a negative Eu anomaly (e.g. Bea et al., 1994; Hermann & Rubatto, 2003; Buick et al., 2006). Monazite trace element data for the Mount Stafford samples are plotted in Fig. 7 and average compositions are given in Electronic Appendix 2 (http://www.petrology.oxfordjournals.org).
|
Trace element analysis of monazite from samples SGL1 and SGP1 revealed a difference between the metamorphic core and rim observed in BSE images, which yield the same U–Pb age of 1805 Ma (no inherited monazite cores were analysed). Monazite rims generally have a more marked negative Eu anomaly and a pattern less depleted in HREE when compared with the dark or patchy zoned cores. The difference in HREE abundances is particularly marked in metapsammite SGP1. Inherited cores older than 1820 Ma have variable compositions with a strong relative HREE depletion.
For monazite in SGL3, which has a uniform U–Pb isotopic composition and hence the observed core–rim zoning is entirely metamorphic, the measured trace elements show differences that correlate with the zoning observed in BSE images. In particular, analyses from the dark cores in BSE images differ from the unzoned rims in having a less pronounced negative Eu anomaly (Eu/Eu* = 0·18–0·25 vs 0·02–0·04 in rims), lower contents of Y, Ti and Th, and higher U contents.
Monazite in the metapsammite at the same metamorphic grade, SGP3, preserves a variety of zoning patterns and inherited U–Pb components in the cores of the crystals. There is also a large variation in the trace element content, particularly in the middle REE (MREE) to HREE. It appears that the composition of the monazite cores dated at 1800 Ma varies over a wider range (Gd/Lu 15–360) than that of the rims (Gd/Lu 62–139); however, the two domains have only a slight difference in Eu/Eu*, and similar Th and U contents. A variety of compositions are also observed in the inherited cores (1947–2448 Ma).
In samples SGL5 and SGM5 some difference in age, although not completely resolvable, was detected between monazite cores and rims. For sample SGL5 monazite trace element analysis revealed variations in Th/U and the extent of the negative Eu anomaly that can be generally correlated to the core–rim zoning. In the associated layer SGM5, the composition of monazite reflects the BSE zoning and the stronger age difference: dark, older cores are richer in U and poorer in Th, have a low Th/U ratio (5–13 vs 44–51 in the rim) and a smaller negative Eu anomaly with respect to the younger rims.
The different monazite domains from the granite are similar in REE contents with limited variations in degree of negative Eu anomaly (larger in the rims) and Th content (higher in the rims; Electronic Appendix 2, http://www.petrology.oxfordjournals.org).
Zircon
It has been widely documented that zircon generally has a chondrite-normalized REE pattern characterized by strong enrichment in HREE with respect to MREE, low abundance of LREE close to or below chondrite values, a negative Eu anomaly and a positive Ce anomaly (e.g. Heaman et al., 1990; Hinton & Upton, 1991; Hoskin & Schaltegger, 2003). This is the case for magmatic zircon, but zircon formed in a metamorphic environment, particularly in sub-solidus conditions, can have different characteristics (e.g. Rubatto, 2002; Hoskin & Schaltegger, 2003).
Trace element analyses of zircon from Mount Stafford were mainly obtained from rims that yielded ages around 1800 Ma (Fig. 8 and Electronic Appendix 2, http://www.petrology.oxfordjournals.org). The composition of zircon rims from the two highest-grade samples, SGL5 and SGM5, is very similar. Both have patterns with strong enrichment in HREE relative to the MREE (Lu/Gd 25–70), negative Eu anomalies (Eu/Eu* 0·02–0·36) and positive Ce anomalies (1·8–8·1). The zircon from SGL5 is slightly richer in HREE than its counterpart in SGM5 (Lu 87–136 vs 54–84 ppm). There is a variation in Eu anomaly from grain to grain, but the average is similar in both samples, as is Th/U. Two analyses were obtained for zircon cores in SGL5 and both are similar and clearly distinguishable from the rim: the former differ from the latter by a stronger positive Ce anomaly and negative Eu anomaly, weaker relative HREE enrichment, and higher Th and Th/U. One of the cores corresponds to an oscillatory zone dated at 1854 ± 11 Ma. For SGM5 the two cores analysed are different and both can be distinguished from the rims because of higher Th/U (1·6 vs 0·03–0·1), and either larger Ce anomaly or higher LREE content.
|
The composition of the 1805 Ma zircon rims from the granite is similar to that measured in zircon rims from the high-grade granulites.
Garnet
Garnet from samples SGP3, SGM5 and SGL5 was analysed for major and trace elements (Electronic Appendix 3 and 4, http://www.petrology.oxfordjournals.org). The garnet is generally homogeneous in major elements; it is rich in FeO (almandine 80–85%) and poor in CaO and MnO (grossular 2·9–3·0%, spessartine 2·2–3·8%). In the highest-grade samples (SGL5 and SGM5), a thin, low-CaO domain (CaO 0·5 wt %) may be present along the grain rim or at the contact with inclusions or fractures (Fig. 7).
The trace element composition of the garnet is more complex. Y mapping (Fig. 9), although close to detection limits, shows that Y zoning is not necessarily related to Ca zoning and is more pronounced than any major element zoning. LA-ICPMS traverses across a number of garnet crystals (8–10 analyses spaced 700–900 µm for each garnet) were also measured (Fig. 10). EMP element mapping coupled with LA-ICPMS spot analysis was used to reconstruct the garnet trace element zoning of garnets from samples SGL5 and SGM5. Zoning mainly involves Ti, Y and the HREE and, to a lesser extent P, Zr and Hf. A zone of irregular shape, and richer in Ti, Y and the HREE, but generally poorer in P is visible in the centre of the garnet (hereafter referred to as the garnet core); its location and dimension can vary from garnet to garnet. It contains 80–100 ppm P, 190–240 ppm Ti, up to 2000 ppm Y, and 250–700 ppm Yb, varying from grain to grain. This Y-rich core gradually changes to an intermediate domain that is volumetrically the most important (hereafter referred to as the intermediate zone): P is around 140–200 ppm, Ti is 140–170 ppm, Y is 500–1200 ppm and Yb is in the range 60–180 ppm. In one of the three garnet grains mapped, a Y-rich annulus of 150 µm width is present in the intermediate part. However, because of its narrowness, no quantitative LA-ICPMS analyses are available for this annulus. A 200–250 µm wide Y-rich zone is present along most of the garnet rim (hereafter referred to as the garnet rim) at the contact with a variety of minerals (quartz, cordierite or feldspars). The rim is similar in composition to the core and has Ti, Y and Yb contents of 150–250 ppm, 1000–1800 ppm, and generally 400–650 ppm, respectively. There is some systematic variation in absolute trace element content between garnets from the two samples: SGP3 garnet is the lowest in trace element content, and SGM5 garnet is slightly lower than SGL5 garnet. Particularly for the two high-grade samples, this difference could partly be an artefact of sampling bias and/or the cut of the garnet investigated. Because of the small dimension of the low-Ca rim, no trace element analysis has been obtained for this domain.
|
In summary, trace element traverses across garnet have a W-shaped trend, with core and rim regions richer in Ti, Y and HREE. This W-shaped pattern is reproducible and is particularly marked in the higher-grade samples where the crystals are larger. Sample SGP3 has a weak W-shaped zoning, but this is more difficult to reconstruct because of the smaller size of the garnet and its highly skeletal nature. SGP3 garnet has rare inclusions of monazite. The largest garnet blasts are observed in sample SGL5, where they include monazite and xenotime in the intermediate zone (xenotime in this zone is mainly found as part of composite inclusions) and in the rim (Fig. 5c and d). Zircon is a common inclusion throughout the garnet in all samples investigated.
Garnet chondrite-normalized REE patterns for the three samples show similar features (Fig. 10c–e) that are common to high-grade garnet in granulites and migmatites (Degeling et al., 2001; Hermann & Rubatto, 2003; Whitehouse & Platt, 2003; Kelly & Harley, 2005): LREE depletion well below chondrite values (La mainly 0·15–0·02), strong negative Eu anomaly (Eu/Eu* 0·014–0·028) and relative HREE enrichment. A significant variation is observed between SGP3 garnet and the higher grade ones: in SGP3 HREE enrichment is limited (Lu/Gd 0·4–12, mostly 1–4) and some of the patterns for the intermediate zone are virtually flat in the HREE (Lu/Gd 0·44–1·8). HREE enrichment is maximum in sample SGL5, where Lu/Gd for the garnet core and rim is 25–43. Regardless of the amount of relative HREE enrichment, in each sample the garnet core has a more enriched pattern than the intermediate zone. Despite differences in the HREE, the negative Eu anomaly and LREE are remarkably similar across domains or grains, with the single exception of some higher LREE content in the SGP3 garnet.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONREGIONAL SETTING AND...METHODSSAMPLE DESCRIPTIONZIRCON AND MONAZITE PETROGRAPHY...TRACE ELEMENT COMPOSITIONDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Age interpretation
Two major types of zircon crystal were found in the samples dated: (1) pre
1800 Ma domains, which appear as rounded grains at low grade or as zircon cores at higher grade; (2) CL-dark, poorly zoned, euhedral overgrowths dated at around 1800 Ma, which overgrow cores in the high-grade samples SGL3, SGL5, SGM5 and SGRA. Monazite separated from seven samples, including the granite, has a similar age pattern to zircon, but with a larger 1800 Ma component already present at lower grade. A range of ages can be defined in this extensive dataset and their interpretation is presented below.
Metamorphic age
The major age component found in monazite (monazite rims at low grade and entire monazite grains at high grade) is in the range 1790 ± 7 to 1806 ± 6 Ma (Fig. 11). This age is interpreted as dating metamorphism, in agreement with the zircon data. Additionally, monazite has been found as inclusions in the intermediate part and the rim of the garnet in the high-grade samples (Fig. 5). In these granulites, garnet formed during high-grade metamorphism.
|
The second zircon type, which occurs in SGP3, the high-grade migmatites (SGL5–SGM5) and the granite, has the same age as most of the monazite (Fig. 11). The characteristics of this zircon type (poor zoning, rims surrounding inherited cores, low Th/U
0·1) are typical of high-grade metamorphic zircon as found in migmatites (e.g. Williams & Claesson, 1987; Williams et al., 1996; Oliver et al., 1999; Rubatto et al., 2001; Corfu et al., 2003). This zircon type is thus interpreted in a general sense as dating the anatectic event and high-grade metamorphism at Mount Stafford. A more rigorous correlation between age and metamorphism is discussed in the following sections.
The range of monazite ages between samples (particularly between SGP3 and the rest of the samples) is larger than can be explained by analytical error. This range and the possibility of age differences between monazite cores and rims deserve consideration. The youngest monazite age was obtained from metapelite SGP3 (1790 ± 7 Ma) and is significantly lower than ages of four of the other monazite samples at higher and lower grade. The analytical session during which monazite from SGP3 and SGL3 was dated featured a significant amount of excess 204Pb, despite the use of energy filtering to reduce molecular interferences (see Rubatto et al., 2001). If 204Pb is overestimated, the calculated ages will be overcorrected and thus younger than the real age. During data evaluation, the analyses were corrected for 204Pb over-counts using the standard. However, it is still possible that the low apparent age of sample SGP3 is an analytical artefact. Therefore, the age of this sample should be treated with caution and we do not place any geological significance on the age difference.
The apparent difference in age between core and rim in monazite from the two highest-grade granulites (sample SGL5 and SGM5) cannot be discarded as an analytical artefact. Data from both samples were collected in the same analytical session, during which cores and rims were analysed alternately. The BSE imaging shows a marked change in zoning between monazite cores and rims, and in sample SGM5 the chemical composition (particularly Th and U contents, Th/U and Eu anomaly, Fig. 7 and Electronic Appendix 2, http://www.petrology.oxfordjournals.org) of cores and rims is distinguishable. A similar compositional difference is present in monazite cores and rims of sample SGL3, although no age difference could be detected or resolved in that case. There is thus an indication that the physical and chemical conditions changed in the course of monazite growth. Although the analytical uncertainties of the ages (core vs rim in each sample, Fig. 11) overlap, the zoning and chemical differences support a two-stage growth for the monazite within a time span of <10 Myr. Multiple monazite growth–recrystallization episodes within a single metamorphic event have been documented in several cases following intra-crystalline age variation and Y zoning (e.g. Fitzsimons et al., 1997; Foster et al., 2002; Hermann & Rubatto, 2003). The possible scenarios for two episodes of monazite growth are discussed below.
Magmatic age
The metamorphic rocks and the granite yielded indistinguishable zircon and monazite ages (Fig. 11). The characteristics (zoning and chemical composition) of these minerals in both rock types are similar. The age of the zircon rims and the monazite from the granite are interpreted as dating the granite crystallization. However, the distinction between high-grade metamorphism and anatectic magmatism is semantic, as both probably occurred simultaneously in the same geological setting. The identical age of granite emplacement and high-grade metamorphism confirms the complex field relationships between granite and granulites described earlier and the hypothesis that magmatism and metamorphism at Mount Stafford are interrelated (Vernon et al., 1990; Collins & Vernon, 1991). It also implies that it is unlikely that the granite was the heat source for the high-temperature metamorphism, and that another heat source has to be found.
A previous ion-microprobe study of zircon from the Mount Stafford granite determined an age of 1818 ± 15 Ma for the formation of the most external zircon overgrowth, and was taken as dating the granite crystallization (Collins & Williams, 1995). This age is within analytical error of our age (1805 ± 3 Ma). The larger analytical spread of the Collins & Williams age is partly attributed to reverse discordant analyses and possibly mixing with older (e.g. 1857 Ma) components found in that zircon population (the analyses were performed without prior CL imaging). The monazite from the granite does not show the complexity of the zircon and its age of 1802 ± 3 Ma is in agreement with the zircon rim age and the best estimate of the granite crystallization. This age estimate indicates that the Mount Stafford granite is coeval with the Yaningidjara Orthogneiss in the adjacent Reynolds Range (SE of Mount Stafford), which has been dated by U–Pb on zircon at 1806 ± 6 Ma (Vry et al., 1996).
Detrital and inherited ages
The pre-1800 Ma zircon is interpreted as detrital because of its rounded aspect, abraded surface (in the low-grade samples), variable zoning pattern, and age. The majority of these domains have oscillatory zoning and Th/U composition >0·1, which suggest an igneous origin (e.g. Williams & Claesson, 1987; Heaman et al., 1990). Some of the pre-1800 Ma domains, however, have features more typical of metamorphic domains with low Th/U values, weak or no zoning, and inclusions of biotite and cordierite (1820 and 1860 Ma domains, Fig. 4). These domains could have been formed during high-grade metamorphism or partial melting in some of the rocks from which the detrital zircon grains originated. The youngest concordant age measured on the likely metamorphic detrital grains is 1823 ± 5 Ma (eight analyses from samples SGP3 and SGL3 mainly on dark zircon rims, MSWD 0·5, 95% c.l.). Similarly, the youngest high Th/U domain with weak oscillatory zoning yielded a concordant age of 1820 Ma (two analyses on the same domain, SGP3-9.1 and 9.2). Another six analyses on metamorphic-looking zircon rims from the same samples yielded slightly older ages averaging at 1842 ± 3 Ma (MSWD 0·85). Samples SGP1 and SGP3 contain a population of six zircon rims averaging at 1856 ± 3 Ma (MSWD 0·41), which is also the age of a concordant zircon core with undoubtedly detrital features (1854 ± 11 Ma). Another six analyses cluster at 1871 ± 4 Ma (MSWD 0·17). From this dataset, we conclude that there is sufficient evidence to suggest deposition of the Mount Stafford Beds at or before 1820 Ma. The deposition of the Lander Rock Beds, to which the Mount Stafford Beds are correlated, has been inferred to be younger than 1838 Ma, the age of the youngest detrital zircon core found in the Lander metasediments (Vry et al., 1996). Similarly, ion microprobe work on the low-grade Mount Stafford Beds indicates deposition after about 1840 Ma (Claoué-Long, 2003