| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology Volume 42 Number 8 Pages 1547-1565 2001
© Oxford University Press 2001
Partial Melting of High-P–T Metapelites from the Tshenukutish Terrane (Grenville Province): Petrography and U–Pb Geochronology
DEPARTMENT OF EARTH SCIENCES, CENTRE FOR EARTH RESOURCES RESEARCH, MEMORIAL UNIVERSITY OF NEWFOUNDLAND, ST. JOHN’S, NFLD., A1B 3X5, CANADA
Received November 12, 1999; Revised typescript accepted February 21, 2001
 |
ABSTRACT |
|---|
Kyanite-bearing metapelites from the Tshenukutish terrane (Manicouagan Imbricate zone, Grenville Province) display evidence of anatexis by means of dehydration melting of micas. These rocks were metamorphosed during a Grenvillian high-P–T crustal thickening event with monazite ages ranging from 1040 to 1017 Ma. Samples that best preserve textural differences between former melt and restitic domains provide evidence for dehydration melting of white mica at
1400–1600 MPa, followed by extensive to complete dehydration melting of biotite up to temperatures in excess of 850°C, and subsequent crystallization of the melt at lower pressures (1100 MPa) during cooling to 750°C. Dehydration melting of biotite was accompanied by growth of garnet with distinctive Ca, Y and Cr patterns, locally around subsolidus garnet. In addition, garnet in one sample displays evidence of partial consumption before the latest stage of growth. This is consistent with dehydration melting of phengite instead of muscovite, according to a theoretically defined reaction: Phe + Grt + Qtz = Bt + Ky + Kfs + L. In all samples, melt crystallization was accompanied by growth of retrograde biotite and was completed at temperatures above the stability field of white mica. In samples that achieved textural equilibrium during or after melt crystallization only the composition of garnet provides some hints about the partial melting history. KEY WORDS: metapelite; high-P granulite; partial melting; Grenville Province
|
INTRODUCTION |
|---|
Partial melting of metapelites has been extensively investigated at low- to mid-P granulite-facies conditions, in the stability field of sillimanite, but studies of these rocks at higher pressures are relatively few (Barbey et al., 1990
; Godard et al., 1996). However, high-P granulites occur in several continental collision belts (e.g. European Variscides: Vielzeuf & Pin, 1989; Carswell & O’Brien, 1993; O’Brien et al., 1997; Willner et al., 1997; Grenville Province: Indares, 1995; Indares et al., 1998; eastern Himalayas: Liu & Jong, 1997) suggesting that crustal rocks involved in orogenesis may attain metamorphic temperatures well in excess of 800°C at a wide range of pressures within the kyanite field. Under these conditions, metapelites are expected to undergo extensive dehydration melting of micas, and study of such rocks has the potential to provide new constraints on partial melting in deeply buried crustal segments.
A P–T framework of melting reactions involving muscovite and biotite in the kyanite field can be established by using available experimental data (LeBreton & Thompson, 1988; Vielzeuf & Holloway, 1988; Carrington & Harley, 1995) extrapolated to high pressures (see also Spear et al., 1999). Alternative melting reactions involving phengite, which may occur at high pressure, can be illustrated in theoretical phase diagrams (see Thompson, 1982). On this basis, natural high-P–T metapelites may be interpreted in terms of reaction and P–T history, provided that a textural record of the melting reactions is preserved to some extent. Further constraints may be placed by interpreting chemical zoning of major and selected trace elements in refractory minerals such as garnet (Spear & Kohn, 1996). This contribution documents evidence of partial melting in a set of newly discovered occurrences of kyanite-bearing migmatized metapelites from the Tshenukutish terrane of the Manicouagan Imbricate zone (NE Grenville Province). This study is complemented by U–Pb data on accessory phases to place time constraints on the partial melting event.
|
REGIONAL CONTEXT |
|---|
The Manicouagan Imbricate zone and the Tshenukutish terrane
The Manicouagan Imbricate zone (MIZ; Fig. 1a; Indares et al., 1998
, 2000) in the NE Grenville Province, is a 1500 km2 belt of Mesoproterozoic igneous rocks (gabbro, anorthosite and granitoids) and subordinate supracrustal units, which were metamorphosed under high-P–T conditions during the Grenvillian orogeny (Indares et al., 1998). The MIZ is divided into two main lithotectonic packages, the structurally lower Lelukuau terrane (LT) and the overlying Tshenukutish terrane (TT; Fig. 1b). U–Pb data from mafic rocks constrain the metamorphic peak at 1·05–1·04 Ga (Cox et al., 1998; Indares et al., 1998). Maximum P–T conditions of 1600–1800 MPa and 800–900°C were calculated from the formerly deepest crustal levels (Lelukuau terrane and Lac Espadon Suite of the Tshenukutish terrane; Fig. 1c; Indares, 1997; Cox & Indares, 1999a). Mineral assemblages in metagabbro from middle-crustal levels (Baie du Nord segment of the Tshenukutish terrane; Fig. 1c) are variably retrogressed and record P–T conditions in the range of 1200–1400 MPa and 800°C (Cox & Indares, 1999b). High-P–T metamorphism in the MIZ is attributed to tectonic burial during a Grenvillian crustal shortening episode, followed shortly after by a thermal perturbation related to emplacement of within-plate gabbro at 1·04 Ga (Indares et al., 1998). A last stage of Grenvillian magmatic activity is expressed by granite intrusions at the highest levels of the MIZ (Boundary zone; Tshenukutish terrane, Fig. 1c) at 1·017 Ga (Indares et al., 1998). Exhumation of the MIZ is consistent with extrusion from the lower levels of the thickened crust, controlled by thermal weakening during the high-T metamorphic event and by the presence of a crustal-scale ramp (Archaean basement in the underlying Gagnon terrane) over which tectonic transport was preferentially channelled (Indares et al., 2000).
|
Setting and general characteristics of the metapelites
Metapelites were recognized in two distinct structural levels of the Tshenukutish terrane (Fig. 1c): in the SW Baie du Nord segment (locations BNS1 to BNS3) as deca- to hectometric size rafts in 1450 Ma diorite, and in the overlying Boundary zone (locations BZ1 and BZ2), as rafts and pods in the 1017 Ma Hart Jaune granite. In all locations they are associated with minor quartzite and calcsilicate rock and in the Boundary zone they alternate with metagreywacke. The metapelites comprise aluminous layers consisting of Qtz–Pl–Kfs–Ky–Grt–Ru ± Bt [mineral abbreviations are after Kretz (1983)
] in a variety of textural relationships, intermixed with leucosome. In location BNS1 they contain abundant biotite and up to 20% diffuse leucosomatic pods. In location BNS2 biotite is rare and the leucosome (up to 30% of the rock) occurs as diffuse anastomosing veins. In both locations leucosome contains garnet and kyanite. Metapelites from location BNS3 consist of Grt–Ky-rich aggregates with interstitial biotite in a Kfs-rich quartzo-feldspathic matrix. In the highest levels of the Boundary zone (location BZ1) aluminous layers are Bt rich and contain 5% Ky-bearing leucosomatic pods, whereas below the Lac Espadon suite (Fig. 1; location BZ2) Grt–Ky–Kfs layers and pods occur in granite. In both locations metagreywacke contains the assemblage Grt–Bt–Pl–Qtz and is leucosome free. |
GEOCHRONOLOGY |
|---|
Published metamorphic ages in the MIZ were determined from meta-igneous rocks from all the structural levels, except for the top of the Boundary zone, and are consistent with the culmination of the high-P–T metamorphism at
1050–1040 Ma (U/Pb on zircon; Cox et al., 1998; Indares et al., 1998). In this section, we present new U–Pb data on accessory minerals from the metapelites of the Tshenukutish terrane, to further constrain the Grenvillian thermal history and explore the possibility of additional earlier metamorphic events.
Leucosome-rich samples of 10–20 kg from each location were processed using standard procedures at Memorial University. The U and Pb isotopic ratios were measured by thermal ionization mass spectrometry (TIMS) using a Finnigan MAT 262 mass spectrometer equipped with an ion-counting secondary electron multiplier [for details, see Dubé et al. (1996)]. All samples contained monazite and samples from BNS2, BZ1 and BZ2 also contained rutile (Table 1). For each sample, two monazite and rutile fractions were selected under the microscope according to criteria of mineral morphology and clarity. Grains with alteration, fluid or solid inclusions were avoided. Monazite fractions were all composed of large (100–200 µm) clear yellow grains and all were abraded (see Krogh, 1982) to remove outer surfaces, which may have undergone some alteration or leaching by grain boundary fluids (Table 1 and Fig. 2). All fractions gave duplicated concordant results (Table 1 and Fig. 2), demonstrating the high quality of these grains. Ages for monazite reported from the Baie du Nord segment are 1040 ± 2 Ma for BNS1 and 1033 ± 2 Ma for BNS2 and BNS3. They are based on assessment of all the ages for both fractions in each case, as they all collectively overlap within combined uncertainties. The age of the sample from BZ1 of the Boundary zone (1017 ± 2 Ma) is based on the 207Pb/206Pb age of the older concordant fraction, as the lower analysis, M1, might have undergone minor Pb loss (Fig. 2). Finally, the two monazite fractions from sample BZ2 yield an age of 1019 ± 2 Ma (Table 1 and Fig. 2). Large rutile fractions were prepared of clearest crack-free dark red prisms and two fractions of each were analysed. The age of rutile from BNS2 (929 ± 2 Ma) is based on the duplicated 206Pb/238U age, as the other ages have higher uncertainties as a result of the significant correction for common lead in these grains. Rutile ages from the Boundary zone (934 ± 4 Ma for BZ1 and 930 ± 3 Ma for BZ2) are based on the duplicated 207Pb/206Pb ages in these fractions, which have significantly less common lead.
|
|
The reported monazite ages overlap with the metamorphic age range obtained by zircons from the mafic rocks of the MIZ (1050–1020 Ma; Indares et al. 1998) and they provide no evidence for potentially earlier metamorphic events in the area. In the Baie du Nord segment, monazite ages between 1040 and 1033 Ma overlap with zircon metamorphic ages from metagabbros (1046–1030 Ma; Cox et al., 1998), indicating fast cooling following peak T conditions. In the Boundary zone, monazite ages are several million years younger (1019–1017 Ma) than in the Baie du Nord segment. Owing to the lack of metamorphic zircon data in this area, the interpretation of these ages is less straightforward; they may indicate either younging of the metamorphic peak towards the top of the structural pile, or delayed cooling. The Hart Jaune granite that encloses the supracrustal rafts in location BZ1 has a crystallization age of 1017 ± 2 Ma, the same as the monazite age in this location, suggesting high heat flow at the time of closure of monazite in the southern Boundary zone. Rutile ages at 929–934 Ma are typical for the entire Manicouagan Imbricate zone (Cox et al., 1998) and indicate that during late stages of cooling (i.e. below the closure temperature of rutile) the whole zone was behaving as a coherent unit.
|
HIGH-P–T PARTIAL MELTING OF METAPELITES: THEORETICAL BACKGROUND |
|---|
P–T framework of partial melting reactions for Ky-bearing pelitic rocks
The general characteristics of the metapelites from the Tshenukutish terrane are consistent with anatexis by dehydration melting of micas in the stability field of kyanite. Extrapolation of experimental data in the kyanite P–T field (Le Breton & Thompson, 1988
; Vielzeuf & Holloway, 1988) shows that with rising temperature and under fluid-absent conditions (expected to prevail in the lower crust), metapelites undergo dehydration melting first of muscovite [reaction (R1): Ms ± Pl + Qtz = Kfs + Ky + L] and then of biotite [reaction (R2): Bt + Ky ± Pl + Qtz = Grt + Kfs + L]. In the KFMASH system (R1) is univariant whereas (R2) defines a divariant P–T field whose width depends on the XMg of the rock. For common pelites with XMg = 0·40–0·45, biotite is eliminated by (R2) at T conditions below those for the next reaction, Bt + Grt ± Pl + Qtz = Opx + Ky + Kfs + L (RI), which is univariant (Carrington & Harley, 1995; Fig. 3a). Addition of albite (NaKFMASH system) lowers the temperature of all reactions by 50°C (Fig. 3a) owing to preferential incorporation of Na in the melt (Thompson & Tracy, 1979). In contrast, addition of anorthite (CaNaKFMASH) increases the variance and shifts all reactions to higher temperatures relative to the NaKFMASH system. Because common high-grade pelitic rocks have An-poor plagioclase (up to An25), partial melting reactions can be adequately represented in the NaKFMASH system (Fig. 3a). Reaction (R2) has a steeper P–T slope relative to (R1); at high pressures the two reactions intersect with the field of (R2) ending at (R1) and the continuous Bt + Ky ± Pl + Qtz = Grt + Ms reaction controlling phase equilibria on the low-T side of (R1) (Spear et al., 1999; Fig. 3a).
|
This P–T framework does not take into account Fe–Mg substitutions in white mica. However, with increasing pressure white mica becomes increasingly phengitic. The dehydration melting of phengite may be represented by the reaction (R1a): Phe ± Pl + Qtz = Ky + Kfs + Bt + L, which differs from (R1) in that it also produces biotite (Thompson, 1982
; Vielzeuf & Holloway, 1988), and is divariant in the NaKFMASH system. Reaction (R1a) is considered to cover a narrow T interval, slightly on the high-T side of (R1). Nevertheless, its divariant character has implications on phase relationships at high pressures. The intersection of two divariant reactions in any system defines a univariant reaction whose length depends upon the bulk composition (see Hensen, 1971; Powell & Downes, 1990; Carrington & Harley, 1995). In the NaKFMASH system with Grt–Phe–Qtz–Pl–Bt–Ky–Kfs–L as possible phases, the intersection of (R1a) and (R2) in P–T space defines the reaction Grt + Phe + Pl + Qtz = Bt + Ky + Kfs + L (RII), which involves all the phases, with Phe and Grt on the other side of (R1a) and (R2) because Grt is absent from (R1a) and Phe is absent from (R2) (Fig. 3b). Following the phase rule, we can also write six additional divariant reactions originating at (RII), each one corresponding to a missing phase and whose relative locations can be established by using Schreinemaker’s criteria. Among them, the Bt + Phe + Pl + Qtz = Grt + Kfs + L [(R3); Ky absent] and the Phe + Pl + Qtz = Grt + Ky + Kfs + L [(R4); Bt absent] reactions correspond to the high-P equivalents of (R1a) and (R2). According to this phase diagram, at high pressures phengite and biotite would melt concurrently by reaction (R3), with any remaining white mica melting subsequently by reaction (R4). This reversal in the melting order of micas has been discussed by Le Breton & Thompson (1988) and Vielzeuf & Holloway (1988), and reactions (R1a), (R2), (R3) and (R4) have been portrayed in the KFASH system by Thompson (1982) as a set of univariant reactions. Reaction (RII) is for the first time defined here. The location of reactions (RII), (R3) and (R4) may be constrained by high-P partial melting experiments involving phengite that are at present lacking, therefore the phase diagram of Fig. 3b is purely qualitative. We note that the importance of these reactions for natural rocks depends on the amount of Fe–Mg substitution in phengite; if this amount is low only minor garnet and biotite would be involved in (R1a), (R3) and (RII), and the dominant white-mica-out reaction would essentially be (R1). Nevertheless, the P–T grid involving phengite (Fig. 3b) is theoretically valid and together with the grid involving muscovite (Fig. 3a), provides the basis for interpreting textures related to partial melting in high-P metapelites. However, complications are likely to arise by retrograde growth of micas as a result of fluid release during crystallization of the melt and by loss of original melting-related textures by recrystallization.
Garnet zoning
Partial melting also affects the composition of solid solutions such as garnet and plagioclase. Incorporation of Na in the melt leads to enrichment of the residual plagioclase in Ca, provided it can react. Consequently, as Grs and An are related by the GASP reaction, the onset of melting will result in a step increase of the Grs content of garnet (Spear & Kohn, 1996), notably at high pressures. Although growth zoning of garnet in terms of Fe, Mg and Mn is commonly altered by diffusion at high temperatures, Ca zoning has the potential to be preserved, owing to its slower diffusion rates. Additional constraints may be placed by examining the zoning pattern of trace elements in garnet, in conjunction with the type of accessory phases present in its immediate environment, because they diffuse at slow rates. In particular, it has been suggested that zoning of elements such as Y, P and Cr may provide additional means of identifying garnet that grew during partial melting (Hiroi & Ellis, 1994; Spear & Kohn, 1996; Pyle & Spear, 1999) and detecting episodes of garnet consumption (Pyle & Spear, 1998, 1999).
In the following sections we attempt to identify partial melting reactions in the metapelites from the Tshenukutish terrane based upon sample-scale textural relationships. Potential links between garnet zoning and partial melting are investigated by means of X-ray maps and zoning profiles of major elements, Y, P and Cr (see the Appendix for analytical techniques). The P–T evolution is tentatively constrained by phase diagrams and, where applicable, thermobarometry.
|
TEXTURES AND MINERAL CHEMISTRY |
|---|
Baie du Nord Segment: samples from locations BNS1 and BNS2
Textures and compositions of feldspars and biotite
The selected samples have typical pelite composition (Table 2) with XMg = 0·52 (BNS1) to XMg = 0·47 (BNS2), and are texturally heterogeneous. They comprise coarse-grained areas of variably recrystallized quartz ribbons with subordinate plagioclase, and pods with garnet, kyanite and perthitic K-feldspar porphyroblasts in a finer-grained Kfs–Pl–Qtz matrix (Fig. 4a). Accessory phases are graphite, monazite and xenotime. Garnet is up to 4 mm in diameter, and contains inclusions of quartz. Its contacts with the Kfs–Pl–Qtz matrix are commonly overgrown by Bt ± Pl aggregates. Kyanite is separated from the coarse-grained areas by a rim of K-feldspar (Fig. 4b) and in the Kfs–Pl–Qtz matrix it is locally rimmed by garnet in turn overgrown by biotite. Biotite also occurs throughout the matrix (BNS1), or as rare interstitial laths (BNS2). Small rounded grains (
50 µm in diameter) of monazite and xenotime are included in the Bt ± Pl aggregates and larger monazite occurs in the Kfs–Pl–Qtz matrix.
|
|
The composition of plagioclase depends upon the textural context. In the coarse-grained areas, large plagioclase away from garnet is the most calcic (cores, An30; rims, An21; BNS1), whereas adjacent to garnet it is homogeneous (An22–24 in BNS1; An12–15 in BNS2). Plagioclase in the matrix is Ab rich (An05–10), and in the Bt ± Pl aggregates it displays a marked increase in An towards garnet, to An20–24 (BNS1) and An18 (BNS2). The perthitic K-feldspar has a bulk composition of
Ab15An2Or82. In samples from BNS1 XFe of biotite overgrowing garnet increases from 0·24 to 0·32–0·35 away from garnet, and TiO2 is in the range of 2–4%. Matrix biotite has intermediate XFe (0·28–0·30) and higher TiO2 (3·5–5%) than in the reaction zones. In contrast, biotite in samples from BNS2 is chemically homogeneous with XFe = 0·34–0·38 and TiO2 = 3·5–5%.
Garnet composition and zoning
In samples from both locations garnet is Prp rich. It has chemically homogeneous cores (BNS1, Alm46–48Prp42–44Grs9Sps1; BNS2, Alm52–55Prp39–41Grs6–7Sps1) whereas rims adjacent to the Ab–Kfs–Qtz matrix are zoned (Figs 5a and 6a). In these rims, Grs increases up to 12% (smoothly in BNS1 and sharply in BNS2) and then decreases to 5–8%. This decrease is sharpest in the outer 100–150 µm next to Bt ± Pl aggregates, where it is accompanied by an abrupt decrease in Prp and an increase in Alm. In addition, the outer 200 µm of the least overgrown garnets in samples from BNS2 are characterized by a concentric trough in the Prp profile (Fig. 6a), which, although variable in intensity, is independent of the Grs zoning. Garnet rims around kyanite are quasi-homogeneous with Alm59–62Prp35–37Grs4Sps1, the same as the rims of the porphyroblasts. Garnet porphyroblasts from both locations can be divided into two domains based on trace element zoning. The core domain, bounded by the ring of maximum Grs content, is characterized by a strong bell-shaped decrease in Y (from 2500 to 250 ppm), a smooth increase in P (to 300 ppm) and flat Cr (Figs 5a and 6a). In contrast, the Grs-enriched rims display only a slight decrease in Y, increasing Cr (to 120 ppm) and a sharp decrease in P restricted to the outer 200 µm. These trends suggest a strong correlation between Ca and Y and Cr zoning, as also reported by Spear & Kohn (1996).
|
|
Interpretation
Textural heterogeneity of samples from BNS1 and BNS2 is probably a result of partial melting, with the Qtz Ca–Pl areas and the Kfs–Ab–Qtz pods representing solid leftover and melt domains, respectively. The presence of kyanite in the pods is consistent with dehydration melting of white mica [reaction (R1) or (R1a)]. Within the coarse-grained domains, the presence of K-feldspar rims separating kyanite and quartz is consistent with subsequent action of (R2) in which kyanite, quartz, and biotite are reactants. These rims together with the large perthitic grains floating on the matrix suggest that (R2) produced a notable amount of peritectic K-feldspar. The rimward decrease of An in plagioclase from the restitic Qtz-rich areas may be linked with Grs production during garnet growth. In contrast, low and uniform An content in plagioclase grains from the pods is consistent with crystallization from melt. Finally, in both locations, the Bt ± Pl aggregates are interpreted to have formed at the expense of garnet during melt crystallization [reaction (R2) in the opposite sense]. This is consistent with An increase in plagioclase rims touching garnet in these textural domains.
Retrograde biotite produced by net transfer reactions is characterized by higher XFe than peak biotite (if any of it remains), although it can subsequently shift to lower values, especially in the vicinity of garnet, by continuing Fe–Mg exchange between the two phases at late stages of cooling (Spear & Florence, 1992; Kohn et al., 1997). Therefore, the observed trend in biotite composition from sample BNS1 (i.e. high XFe in biotite away from garnet in the reaction zones and intermediate values in biotite from the matrix) may be a relict of an original contrast between peak biotite in the matrix that survived melting, and retrograde biotite that formed during melt crystallization. Nevertheless, owing to fast diffusion in biotite, the chances that matrix biotite retained its peak composition are rather slim (see also Spear & Parrish, 1996).
Operation of reaction (R2) implies that garnet, at least in part, was produced together with melt. Homogeneous garnet cores surrounded by Grs-enriched rims are consistent with growth in at least two stages, first by subsolidus reactions and second by (R2) (see also Spear & Kohn, 1996; Kohn et al., 1997). We note that rims adjacent to the coarse grain domains are lacking Grs enrichment (Fig. 6a), indicating that growth by (R2) was irregular and mainly occurred next to melt pods. The smooth Grs decrease in most of the rim segments outwards from the Grs peak is consistent with isothermal garnet growth during (R2), as the Grs isopleths have a positive slope in P–T space (Spear et al., 1999). However, the steeper decrease in the outer few tens of micrometres, together with the retrograde Fe–Mg zoning, is probably related to garnet resorption by reaction (R2) in reverse sense. An anomalous feature of some garnets from BNS2 is the concentric Prp trough (Fig. 6a). A similar trend is reported by Gardien & Thompson (1995) from garnet experimentally produced at temperatures above the dehydration melting of biotite, suggesting a possible link with the elimination of biotite from the reservoir. An attractive but speculative explanation for this feature involves: (1) equilibrium (and further growth?) of the outer rims of garnet in the presence of an Fe-rich opaque, following elimination of biotite, leading to a drop in Prp; (2) subsequent increase of Prp at the rims as a result of Fe–Mg exchange between biotite and garnet upon retrograde development of the former. In this case, the Prp trough observed would be the net result of diffusional adjustments of the rim.
The reaction history of garnet can be investigated further by the zoning trends of Y, Cr and P (Fig. 6b). The Y pattern is typical of subsolidus garnet that started growing in the presence of an Y-saturating phase (e.g. xenotime; Pyle & Spear, 1999, 2000). In samples from BNS1 and BNS2, no xenotime was found as inclusions in garnet, but its presence in the retrograde biotite overgrowing garnet attests to a high Y concentration in the garnet domains. We also note that the change in the gradient of Y decrease in the Ca-enriched rims coincides with a drastic change in the Grt-forming reaction, from a subsolidus reaction to (R2). Increasing Cr at the rim segment is consistent with garnet growing at the expense of biotite, i.e. by (R2) (Spear & Kohn, 1996). The overall P pattern seems to be insensitive to the changes in the garnet-forming reactions because it does not show any discontinuity at the boundary of the Ca-enriched zones. As for the sharp decrease of P in the outer rims of these zones, it may be linked to formation of monazite in the sites of garnet consumption.
Baie du Nord segment: location BNS3
Textures, and compositions of feldspars and biotite
Selected samples from this location contain a Qtz–Kfs–Pl-bearing matrix with garnet (6–9 mm in diameter), kyanite and K-feldspar porphyroblasts, and minor biotite. Accessory phases are apatite and monazite. These samples have a composition of typical pelite with XMg = 0·40 (Table 2). Plagioclase is Ab rich (An11–15) with highest An contents in grains touching garnet. The K-feldspar porphyroblasts are perthitic (with bulk composition Ab31–34An4Or62–66), with lamellae of the same composition as matrix plagioclase. Biotite forms clusters invading garnet or laths dispersed in the matrix, and is homogeneous with XFe = 0·40–0·44, and TiO2 = 4–5%. Kyanite is locally separated from K-feldspar and garnet by a rim of quartz.
Garnet: textures, composition and zoning
Samples from location BNS3 display two textural types of garnet, of similar size range. Garnet I comprises three concentric textural domains (Fig. 4c): a core with abundant micrometre-size quartz and a few apatite inclusions surrounded by two clear zones (middle and rim domains) separated by a narrow ring of coarse kyanite, quartz and apatite inclusions. Garnet II has a wide core domain with abundant inclusions, surrounded by a clear rim of irregular width (Fig. 4d). Inclusions are composed of quartz, rutile, kyanite, apatite, and, rarely, monazite. In addition, coarse inclusions of quartz together with intergrowths of fine-grained material, are widespread (Fig. 4d). The intergrowths consist of K-feldspar (and locally albite) intermixed with chemically heterogeneous material of average composition 33–40% Al2O3, 47–54% SiO2 (corresponding to a cationic ratio Al:Si 1), subordinate K2O (0·5–2%), Na2O (0–1·5%), MgO (2%), and FeO (0·5–1·5%) with analyses closing at 90%. This composition probably represents a mixture of phases dominated by kaolinite, as suggested by the Al and Si contents and the low totals. In addition, rims between adjacent garnet grains are the locus of symplectitic pods of skeletal kyanite and quartz separated by fine intergrowths of the same material as in the garnet inclusions (Fig. 4d). In both cases, the presence of quartz, K-feldspar and, possibly, kaolinite suggests that they may be melt pods trapped in garnet. These pods may have crystallized quartz and white mica in isolation from the matrix, with white mica subsequently breaking down to kaolinite and K-feldspar owing to excess H2O that remained trapped.
Garnet I exhibits spectacular zoning (Figs 5b and 6b). The core and the middle domains are characterized by a sharp outward decrease in Grs (from 15 to 5%), compensated by an increase in both Prp (from 22 to 31%), and Alm (from 58 to 64%), whereas XFe decreases slightly (from 0·71 to 0·68). In the rim domain, Grs first increases to 8% and finally drops to 3%, Prp drops to 24%, and Alm increases to 74% concurrently with increase in XFe (to 0·75). Sps displays a smooth but perceptible outward decrease (from 3 to 1%). This, together with Grs zoning at least out to the second Ca peak, and the compensatory Alm and Prp patterns, provides evidence for preservation of growth zoning although undoubtedly modified by diffusion. In contrast, the outer rim zoning is typical of diffusion-controlled retrograde zoning. The cores of garnet II (Figs 5c and 6c) are chemically homogeneous with Alm62Prp29Grs7–9Sps1 whereas in the outer 1000 µm Grs drops to 3%, and Alm increases to 70%. We note that the composition of the cores of garnet II is similar to that of the inner part of the rim domains of garnet I. Chemical trends at the outermost rims are consistent with retrograde resetting.
In the core and middle domains of garnet I Y decreases outwards (from 720 to 250 ppm) in two steps (Fig. 6). In the same domains, P decreases (from 200 to 100 ppm) whereas Cr increases slightly (Figs 5b and 6b). In contrast, the rim domain is characterized by flat Y, increasing Cr (to 200 ppm) and higher P (150–200 ppm) than in the middle domain. The most striking feature, however, is a marked concentric peak in P (to 650 ppm), accompanied by much weaker humps in Cr and Y, just outside the boundary between the middle and the rim domains. Garnet II displays relatively constant Y (250 ppm), P (150 ppm) and Cr (180 ppm) with the exception of: (1) a decrease in Y (from 300 to 250 ppm) away from the monazite-bearing central region; (2) higher P (200 ppm) and a trough in Cr and Y in parts of the rim domain (Figs 5c and 6c). These latter trends are asymmetric and apparently unrelated to any specific textural setting. We note that Y, P and Cr contents in garnet II are similar to those of the rim domain in garnet I and that Cr contents in both garnet types are higher than in garnets from BNS1 and -2 (Fig. 6).
Interpretation
The lack of recognizable restitic areas in samples from location BNS3 and chemical homogeneity of the matrix phases may be attributed to more extensive partial melting and/or subsequent recrystallization than in the previous locations. However, large garnets have the potential to preserve some record of the partial melting history.
In garnet I, the following features are particularly relevant. (1) The relatively coarse-grained Qtz–Ky inclusions bounding the rim from the middle domain attest to a break in the growth history of garnet. In this setting, the narrow concentric peak in P (and to a lesser extent in Cr and Y) immediately outside that boundary attests to partial resorption before the growth of the rim domain. Release of these elements from the consumed garnet would lead to a local increase in their concentration in the immediate matrix. If they are not redistributed among matrix phases, they would subsequently reincorporate in garnet upon renewed growth, creating spikes (Pyle & Spear, 1998, 1999). (2) Grs increase in the rim domain reflects a renewed increase in the availability of Ca, as expected from the continuing dissolution of the Ab component of plagioclase in melt while garnet had stopped growing. These features imply at least two episodes of garnet growth (first of the core and middle domains and second of the rim domain) separated by an event of garnet consumption. Enrichment in Grs and increasing Cr are consistent with growth of the rim domain by reaction (R2), whose operation is also implied by the overall mineral assemblage. In this context, a reaction operating before (R2) has to be envisaged to account for the evidence of garnet consumption discussed above. During prograde metamorphism, and in a P–T framework of reactions involving muscovite (Fig. 3a) garnet is expected to grow by the continuous reaction Bt + Ky ± Pl + Qtz = Grt + Ms up to crossing (R1), which does not involve garnet. Therefore, this framework cannot explain the resorption texture in garnet. However, if we consider reactions involving phengite, then garnet may be partially consumed by the discontinuous reaction (RII) just before entering the P–T field of (R2) (Fig. 3b), in accordance with the textures and zoning of the rim domain of garnet I.
In the core and middle domains, the sharp outward decrease of Grs from values as high as Grs15 (Fig. 6c) may attest to isothermal garnet growth at high-P conditions in the presence of An-rich plagioclase. As the samples are not particularly Ca rich (Table 2), high An content in plagioclase during early stages of garnet growth may be the result of a melting reaction, such as reaction (R3). This scenario, although consistent with high Cr contents relative to samples from the other locations, is, however, too speculative; the Y decrease in steps attests to a more complex reaction history with no record preserved in the zoning of the other elements. In addition, the absence of Y-bearing inclusions in garnet (xenotime or monazite) and in its immediate environment precludes any tentative interpretation of the Y zoning. Finally, in contrast to the apatite-free samples from BNS1 and -2, P zoning in the core and middle domain of garnet I mimics the Ca zoning, suggesting that in the presence of apatite Ca in garnet controls the P content.
In general, garnet II displays relatively homogeneous cores of composition similar to the rim domain of garnet I. As the two types of garnet have similar sizes, the differences in their zoning patterns are probably due to different reaction histories. Indeed, the composition of garnet II, together with the lack of textural discontinuities, the presence of the same type of inclusions as in the ring between the middle and rim domains of garnet I, and the lack of spikes on the trace element profiles suggest that garnet II was produced by the same reaction as the rims of garnet I, i.e. (R2). Growth of this garnet by reaction (R2) is further supported by the presence of inclusions that may represent trapped melt (see above). Relatively high Y contents in some garnet cores are linked with the presence of monazite inclusions. P plateaux in some garnet rims in conjunction with the lack of apatite inclusions in the rim domains is consistent with dissolution of apatite in the melt leading to a change in the partition coefficient of P between garnet and the matrix. In addition, troughs in the Cr profiles at the same rims may indicate growth above the elimination of biotite from the system, by consumption of some Fe-oxide. This is consistent with a generalized Alm increase towards these rims, although the same trend may also be interpreted as retrograde. Another characteristic of the garnet II core domains is abundance of apatite inclusions. If these porphyroblasts grew entirely by reaction (R2), there may be a link between concentration of apatite and garnet nucleation.
Aluminous gneisses from the Boundary zone (locations BZ1 and BZ2)
The aluminous layers of the highest levels of the Boundary zone (location BZ1) and the aluminous pods of the lowest levels (location BZ2) are characterized by low XMg values (0·23–0·27; Table 2), suggesting the possibility of a common Fe-rich protolith. Apart from this, the rocks from the two locations have contrasting characteristics. The aluminous layers are pelitic in composition (Table 2) and consist of the assemblage Qtz–Pl–Ksp–Grt–Ky–Bt–Ru, whereas the aluminous pods have an Al-rich, Si-poor chemistry (Table 2) and are mainly composed of Ky–Grt–Kfs. Both rock types contain apatite and monazite as accessory phases.
Textures and mineral chemistry
In the aluminous layers (location BZ1) garnet occurs as porphyroblasts up to 4 mm in diameter with quartz, plagioclase, rutile, ilmenite and apatite inclusions in the core and clear rims (Fig. 4e). Ilmenite was not found in the matrix, hence it is not considered as part of the peak assemblage. Biotite and kyanite are evenly distributed throughout the matrix. Apatite and monazite mainly occur as inclusions in biotite. K-feldspar is mostly microcline but some large crystals are perthitic. Plagioclase (An19–25Or0–2) is homogeneous and adjacent to garnet is most depleted in anorthite (An19–21). Exsolution lamellae of plagioclase in K-feldspar have the same composition as the matrix grains, and bulk composition of individual K-feldspar grains falls in the range Ab12–16An2–3Or82–85. Biotite is chemically homogeneous at the sample scale with XFe = 0·57–0·61 and TiO2 = 4·2–4·6%. Garnet is Alm rich in accordance with the high-Fe bulk content of the samples (Table 1). It is unzoned (Alm75–78Prp16–18Grs5–7Sps1), with the exception of rims adjacent to biotite (outer 250 µm), where Alm increases to 80% and Prp drops to 12%. The only detectable zoning in trace elements is an outward increase in P whereas Y and Cr contents appear to be uniform. The Y content, in particular, is below 250 ppm, i.e. lower than in garnets from the previously described locations.
The aluminous pods (location BZ2) consist of garnet and kyanite porphyroblasts (up to 2 mm in diameter), in a matrix of K-feldspar (Fig. 4f), with rutile and apatite as minor phases. Apatite occurs both in the matrix and as inclusions in garnet and kyanite. Garnet porphyroblasts result from the amalgamation of small subidiomorphic grains and are commonly rimmed by Ky–Qtz–Bt or Ky–Qtz–Ab symplectite. Locally, garnet rims kyanite. Kyanite contains inclusions of K-feldspar, apatite and quartz. K-feldspar (An14–20Or78–84) shows very fine perthitic lamellae that were too closely spaced to allow separate analysis of the exsolutions and the host grain. Garnet (Alm68–70Prp19–22Grs8–12Sps1) is homogeneous except for a slight decrease of Grs in the outer 100–200 µm of individual subgrains. Y and Cr are uniform whereas P mimics the Grs zoning. As plagioclase is absent, Grs decrease at the rims is attributed to depletion of Ca in the matrix during the last stages of garnet growth. Compared with the previous case, garnet from BZ2 is more Prp and Grs rich (for similar bulk Ca and XMg; Table 2).
Interpretation
In samples from location BZ1 the assemblage Qtz–Pl–Kfs–Grt–Ky–Bt and the presence of kyanite in the leucosome is consistent with elimination of white mica by reaction (R1) [or (R1a)]. Apparently low amounts of leucosome (5%) and the lack of any textural evidence of reaction (R2) suggest that dehydration melting of biotite did not occur and that garnet is the product of a subsolidus reaction. In contrast, large amounts of K-feldspar in this rock are consistent with extensive partial melting, and uniformly low Y contents in garnet, despite the widespread presence of monazite in the sample, are consistent with garnet growing in the presence of melt. It is therefore possible that reaction (R2) occurred, but during subsequent melt crystallization, the rock experienced pervasive textural re-equilibration erasing the record of previous metamorphic history. Overall chemical homogeneity in terms of the major constituents of garnet, biotite and plagioclase suggests that diffusion was efficient at the sample scale during the re-equilibration stage. Garnet zoning in terms of Prp and Alm in rims adjacent to biotite is consistent with subsequent retrograde Fe–Mg exchange between the two phases. Any evidence of this in biotite is lacking, probably as a result of faster diffusion rates. In addition, the low magnitude of retrograde zoning indicates fast cooling.
The Al-rich and Si-poor bulk chemistry and the mineralogy of the aluminous pods (location BZ2) is typical of peritectic aggregates after complete dehydration melting of both micas. In this location, the presence of K-feldspar inclusions in kyanite suggests that the latter grew in the stability field of the former, whereas local mantling of kyanite by garnet suggests subsequent operation of reaction (R2) (Fig. 3). Lack of major and trace element zoning in garnet is consistent with a simple growth history by reaction (R2). We note that garnet composition is broadly similar to that of garnet from location BZ1. Finally, Ky–Qtz–Bt symplectite around garnet is attributed to limited back-reaction (R2) during cooling.
|
P–T ESTIMATES |
|---|
Constraints from the P–T grid
First-order constraints on the P–T field of the metamorphic peak can be placed by using the P–T grid for the NaKFMASH system after Spear et al. (1999
; Figs 3a and 7). For all the studied rocks, lower P–T limits are defined by the Sil–Ky transition and reaction (R1), which marks the onset of dehydration melting. This is also valid for the samples from BNS3, which may have experienced reaction (RII) because the later should lie close to the extension of (R1) [or (R1a)]. Evidence for elimination of prograde biotite, at least in samples from BNS2 and BNS3, implies peak T conditions in excess of 850–875°C, the latter corresponding to culmination of (R2) in metapelites with intermediate XMg at 1000 MPa (LeBreton & Thompson, 1988; Vielzeuf & Holloway, 1988). This may not be valid for the aluminous pods from location BZ2, which are expected to lose biotite at lower temperatures owing to their Fe-rich character (Carrington & Harley, 1995). However, an upper-T limit of 900–950°C is set for samples from BZ by reaction (RI) because no orthopyroxene was observed in adjacent metagreywackes.
|
Figure 7a also shows semi-quantitative contours of XFe in garnet after Spear et al. (1999). If garnet preserves growth zoning these contours may be used to roughly constrain: (1) the P–T conditions at which (R1) is crossed [by the intersection of (R1) with the contour of the XFe in the first garnet growing with melt]; (2) a thermal maximum, set by the contour of the lowest XFe in garnet; (3) the T conditions of melt crystallization [by (R2) in reverse sense] by using XFe of garnet adjacent to the reaction zones (but not in direct contact with biotite, to minimize effects of subsequent Fe–Mg exchange).
For instance, XFe 0·56–0·58 in high-Grs rings in garnet from BNS1 implies that (R1) was crossed at 800°C and 1550–1650 MPa (Fig. 7a). Minimum XFe 0·52 close to the rims of the same garnets suggests T conditions of 850°C at advanced stages of (R2). Finally, XFe 0·65 at garnet rims in reaction zones, not directly touching biotite, are consistent with T conditions of 750°C at the end of melt crystallization. In the last case, P conditions are bracketed by the intersections of the XFe contour with reaction (R1) and with the Ky–Sil transition line, between 1200 and 900 MPa (Fig. 7a). In the case of samples from BNS1 and BNS2 this approach yields a realistic picture of the overall P–T paths in the field of reaction (R2), with rising T while (R2) is in progress, followed by cooling and decompression during melt crystallization and formation of Bt ± Pl aggregates after garnet (Fig. 7a). However, it should be kept in mind that these P–T paths are schematic and their exact shape and location should be viewed with caution for various reasons: (1) XFe contours are semi-quantitative (see Spear et al., 1999), P–T ranges are slightly underestimated because Ca is not taken into account, and T conditions of white mica dehydration melting are minimum if reaction (R1a) [or (RII)] occurred instead of (R1); (2) partial resetting by diffusion at the thermal peak will tend to decrease the XFe of earlier grown garnet, therefore the XFe measured in high-Grs garnet rings should be viewed as minimum (and estimated pressures as maximum); (3) no distinction can be made between heating under constant or decreasing pressure along the prograde segment of the paths; (4) minimum XFe provides only a minimum estimate of the thermal peak because the isopleths are not valid at temperatures above the elimination of biotite and also because retrograde diffusional resetting would tend to increase the XFe.
XFe in garnets from BNS3 is higher than in BNS1 and BNS2 and displays little variation, resulting in a short apparent P–T path on the low-P corner of the field of reaction (R2) with kyanite. This path is inconsistent with evidence of extensive partial melting in the kyanite field [and of likely higher pressures if reaction (RII) occurred]. Garnets from BZ do not preserve growth zoning and display the lowest XFe (0·8), with contours falling below the stability field of kyanite coexisting with melt. In the case of garnet from BZ1, this supports evidence of extensive re-equilibration after crystallization of the melt discussed above; nevertheless, the absence of muscovite or sillimanite is inconsistent with the measured XFe. Garnet from the restitic aluminous pods (BZ2) clearly was not in equilibrium with melt and biotite; nevertheless, XFe should indicate a temperature of advanced stages of (R2) and this is not the case. It should be noted that the studied samples display a marked correlation between bulk XFe and XFe of garnet and that only samples with intermediate bulk XFe (XMg = 0·47–0·52; BNS1–2; Table 2) have garnet with XFe consistent with extensive partial melting in the kyanite field of (R2) (Fig. 7a). It is therefore suggested that in Fe-rich samples, some alternative equilibria between garnet ± biotite and an Fe-rich phase controls the XFe in garnet shifting the contours proposed by Spear et al. (1999) to higher temperatures.
Thermobarometry
Despite the presence of garnet, biotite, kyanite, plagioclase and quartz in all samples but BZ2, application of thermobarometry suffers major limitations as a result of the lack of suitable mineral assemblages representative of the thermal peak. For instance, in locations BNS1 and BNS2 kyanite is not in contact with quartz and plagioclase, therefore the GASP barometer cannot be used. Nevertheless, the pressure at which (R1) was crossed can be estimated in sample BNS2 by using garnet rims devoid of Ca enrichment (see Table 3) and rims of adjacent large restitic plagioclase, based on the assumption that these phases were in equilibrium with kyanite and quartz at that stage of the metamorphic evolution and that they did not reset subsequently. Calculated with the GASP reaction, this pressure is 1650 MPa at 810°C (Fig. 7b; for method see Table 3 footnotes), i.e. slightly above the P range estimated with the XFe contours. BNS1 garnets are entirely surrounded by Grs-enriched rims. However, pressures at (R1) may be constrained to some extent by using the outer core of garnet (at the boundary with the Ca-enriched zone), and rims of the large restitic plagioclase in the matrix. Calculated GASP isopleths cross (R1) at 1400 MPa, i.e. at the lower-P side of the range obtained with the XFe contours, and may be viewed as minimum in case the rims of the restitic plagioclase were enriched in An during partial melting. In samples from BNS3 no subsolidus plagioclase is preserved. However, GASP isopleths for garnet with maximum Grs content (in the Grs-enriched rims of garnet I and at the core of garnet II) and matrix plagioclase, should give a rough upper-P limit for the initiation of (R2) (upper because plagioclase in equilibrium with that garnet was more An rich than the final matrix plagioclase). These isopleths cross reaction (R1) at 1650 MPa.
|
In the Baie du Nord segment, Grt–Bt thermometry can be used to constrain the thermal peak in samples from BNS1 that may contain peak biotite, and that also display graphite implying minimum Fe3+ content in biotite (Spear & Florence, 1992). Nevertheless, calculated T conditions using biotite away from the reaction zones with maximum-Prp garnet from the Grs-enriched rim are of 900°C (see Table 3), above the upper-T limit of (R2) for pelites of intermediate XMg. Such high-T conditions suggest that if biotite survived the thermal peak, sample-scale Fe–Mg diffusion during subsequent melt crystallization has shifted its XFe ratio to higher values (see also Spear & Parrish, 1996). Finally, application of the two-feldspar thermometer (Kroll et al., 1993) is dubious because the compositions of equilibrium feldspar pairs cannot be established with confidence.
In all samples from BNS, P–T conditions of local equilibrium between garnet and retrograde Bt ± Pl aggregates can be calculated by using touching garnet and plagioclase rims and biotite not in contact with garnet (to minimize effects of Fe–Mg exchange subsequent to closure of the net transfer reactions). Grt–Bt thermometry gave scattered results; however, intersection of the GASP isopleths with the XFe contours yield consistently 1100 MPa and 750°C (BNS1 and BNS2) and 725°C and 900 MPa (BNS3), near the lower P–T corner of the kyanite field of (R2) (Fig. 7b).
Samples from location BZ1 are the only ones to contain Grt–Pl–Bt–Ky–Qtz in apparent textural and chemical equilibrium. P–T conditions were calculated with the GASP and Grt–Bt reactions, using the compositions of the garnet cores, matrix biotite, and plagioclase adjacent to garnet with the lowest An contents. These compositions should be representative of the conditions at which sample-scale equilibrium was achieved; this is probably distinct from the thermal peak, because, as discussed above, this sample provides indirect evidence of late re-equilibration. However, these samples do not contain graphite or matrix ilmenite, therefore the Fe3+ content of biotite is unconstrained (Williams & Grambling, 1990). Calculated P–T conditions fall in the range 1100–1200 MPa and 830–860°C (Fig. 7), within the stability field of kyanite, but 100°C above (R1). However, taking into account that for Fe-rich pelites the divariant field of reaction (R2) is displaced to lower temperatures relative to intermediate XMg pelites, biotite should not be stable at these conditions. In this context, we consider that calculated temperatures are probably overestimated owing to the use of total Fe as Fe2+ in biotite.
|
DISCUSSION AND CONCLUSIONS |
|---|
Metamorphic context, ages, and implications
Metapelites from the Tshenukutish terrane provide strong evidence of dehydration melting of micas in the kyanite field. This is consistent with P–T conditions of 1200–1600 MPa and 800–900°C recorded by high-P textures of metagabbros from the same terrane (Cox & Indares, 1999a
, 1999b). High-P–T metamorphism is related to a Grenvillian crustal shortening event. In the lower levels of the Tshenukutish terrane (Baie du Nord segment) its peak is constrained to 1040–1030 Ma (U–Pb on zircon, Cox et al., 1998). This contribution provided new age constraints from U–Pb data on monazite. In the Baie du Nord segment monazite ages range between 1040 and 1033 Ma and overlap with ages of metamorphic zircons, indicating fast cooling following peak conditions. In the higher structural levels (Boundary zone), monazite displays ages of 1019 and 1017 Ma. This suggests younging of the metamorphic peak, or delayed cooling, toward the higher structural levels and seems to be related to high heat flow in parts of the Boundary zone at that time, as attested to by the presence of the 1017 Ma Hart Jaune granite (Fig. 1c).
Reaction sequences and P–T evolution
Partial melting reactions and prograde P–T evolution
The migmatitic aluminous gneisses from Tshenukutish terrane retain variable record of their partial melting history. Samples from BNS1 and BNS2 that best preserve textural differences between former melt and subsolidus domains provide evidence for dehydration melting first of white mica [muscovite with possibly a phengitic component; reaction (R1) or (R1a), Fig. 7] and subsequently of biotite, the latter being associated with growth of K-feldspar and garnet rimming kyanite and subsolidus garnet [reaction (R2), Fig. 7]. Garnet formed by (R2) is distinctively richer in Ca and Cr and poorer in Y than subsolidus garnet. In samples from BNS3, most garnet formed entirely by reaction (R2) with the exception of one sample where growth by (R2) was preceded by partial resorption of an earlier garnet. This provides evidence that a discontinuous NaKFMASH reaction involving phengite, Phe + Grt + Pl + Qtz = Bt + Kfs + Ky + L (Fig. 7b), defined for the first time here on theoretical grounds, may have taken place.
Dehydration melting of white mica in samples from BNS1 and BNS2 was constrained to 1400–1600 MPa and 800°C, by using contours of XFe in garnet in the NAKMASH system and GASP isopleths, with subsequent elimination of biotite at temperatures in excess of 850–875°C. In addition to more extensive textural re-equilibration following (R2), which disallows use of GASP isopleths, garnet XFe contours in samples from BNS3 and BZ1, which are increasingly Fe rich, gave suspiciously low-P results that in fact suggest an additional strong correlation between XFe in garnet and bulk Fe. Therefore, P–T conditions of (RII) (if it occurred) in BNS3 could not be constrained. In Fig. 3b, reaction (RII) appears on the high-P side of (R1a) [and, obviously, (R1)]. However, this is valid only for a specific bulk composition; as it is a general trend in the KFMASH system, reactions tend to be displaced to lower P–T conditions with increasing bulk Fe. In addition, the extent to which reaction (RII) will occur also depends on the amount of celadonite substitution of the white mica entering the melt domain, which also may be a function of the bulk composition. Therefore evidence for reaction (RII) in BNS3 may be due to the higher XFe of this sample rather to a difference in pressure between BNS3 and BNS1–2.
We note that all samples contain notable amounts of peritectic K-feldspar. Partial melting experiments in the sillimanite field produced small amounts of that phase (Carrington & Harley, 1995) but its abundance is predicted to increase with pressure owing to decreasing melt proportion (Castro et al., 1999), which is linked with a pressure-related increase in the solubility of H2O (Holtz & Johannes, 1994).
Retrograde reactions
In all samples melt crystallization was accompanied by development of biotite at the expense of garnet and, owing to the absence of white mica, it was probably completed at temperatures within the range of reaction (RII). This is consistent with P–T conditions of 750°C and 1100 MPa estimated at the intersection of relevant XFe contours and GASP isopleths in samples from BNS1 and BNS2 (Fig. 7b). Although the absence of retrograde muscovite may be attributed to melt escape (Kohn et al., 1997; Spear et al., 1999), in the present case we favour an alternative explanation. As partial melting started at higher pressures relative to those of melt crystallization and because of the positive slope of reaction (R1) (see Fig. 7b), the rocks crossed a wider (R2) field upon cooling, allowing thus complete melt crystallization within that field. It is therefore suggested that in addition to melt escape, the shape of the P–T path can control the development of retrograde muscovite.
|
APPENDIX: ANALYTICAL TECHNIQUES |
|---|
 TOPABSTRACTINTRODUCTIONREGIONAL CONTEXTGEOCHRONOLOGYHIGH-P-T PARTIAL MELTING OF...TEXTURES AND MINERAL CHEMISTRYP-T ESTIMATESDISCUSSION AND CONCLUSIONS APPENDIX: ANALYTICAL TECHNIQUES REFERENCES |
|---|
Mineral analyses were performed on a CAMECA SX50 electron microprobe at Memorial University. Point analyses of major elements in garnet, biotite, plagioclase and K-feldspar were performed in energy dispersive (ED) mode with a 20 nA specimen current, 15 kV accelerating voltage and 75–100 s counting time, except for feldspars, where a 10 nA specimen current was used to avoid Na loss. Point analyses in garnet including Y, P and Cr were performed in ED mode for major elements and wavelength dispersive (WD) mode for traces, with 15 kV accelerating voltage, 50 nA specimen current and 50 s counting time for the traces. Compositional maps of Ca, Fe, Mg, Y, P and Cr in garnet were produced in WD mode using a grid of 256 x 256 pixels, 20 kV accelerating voltage, and 50 nA specimen current for Ca, Fe and Mg, and 300 nA for the trace elements. In addition, closely spaced quantitative analyses for major elements and qualitative analyses of Y, P and Cr were performed along specific transects to obtain zoning profiles. Y, P and Cr profiles were obtained in WD mode, with 15 kV accelerating voltage, 300 nA specimen current and 30 s counting time.
|
ACKNOWLEDGEMENTS |
|---|
We thank F. Spear for a very challenging review of the manuscript, and the second reviewer, P. J. O’Brien. M. Piranian provided support on the microprobe, and R. Hicks and R. Cox provided technical assistance in the U–Pb laboratory. This project was supported by NSERC and LITHOPROBE grants to the authors. This is LITHOPROBE Contribution 1210.
|
FOOTNOTES |
|---|
*Corresponding author. Telephone: (709)-737-2456. Fax: (709)-737-2589. E-mail: afin{at}sparky2.esd.mun.ca
|
REFERENCES |
|---|
Barbey, P., Macaudiere, J. & Nzenti, J. P. (1990). High-pressure dehydration melting of meta-pelites: evidence from the migmatites of Yaoundé (Cameroon). Journal of Petrology 31, 401–427.
Berman, R. G. (1991). Thermobarometry using multi-equilibrium calculations: a new technique, with petrological implications. Canadian Mineralogist 29, 833–855.[Web of Science]
Berman, R. & Aranovich, L. Y. (1996). Optimized standard state and solution properties of minerals. I. Model calibration for olivine, orthopyroxene, cordierite, garnet and ilmenite in the system FeO–MgO–CaO–Al2O3–TiO2–SiO2. Contributions to Mineralogy and Petrology 126, 1–24.
Carrington, D. P. & Harley, S. L. (1995). Partial melting and phase relations in high grade metapelites: an experimental petrogenetic grid in the KFMASH system. Contributions to Mineralogy and Petrology 120, 270–291.
Carswell, D. A. & O’Brien, P. J. (1993). Thermobarometry and geotectonic significance of high-P granulites: examples from the Moldanubian zone of the Bohemian massif in lower Austria. Journal of Petrology 34, 427–459.
Castro, A., Patiño-Douce, A. E., Guillermo Corretgé, L., de la Rosa, J. D., El-Biad, M. & El-Hmidi, H. (1999). Origin of peraluminous granites and granodiorites, Iberian massif, Spain: an experimental test of granite petrogenesis. Contributions to Mineralogy and Petrology 135, 255–276.
Cox, R. & Indares, A. (1999a). Transformation of Fe–Ti gabbro to coronite, eclogite and amphibolite in the Baie du Nord segment, eastern Grenville Province. Journal of Metamorphic Geology 17, 537–555.
Cox, R. & Indares, A. (1999b). High pressure and temperature metamorphism of the mafic and ultramafic Lac Espadon suite, Manicouagan Imbricate Zone, eastern Grenville Province. Canadian Mineralogist 37, 335–357.[Web of Science]
Cox, R., Indares, A. & Dunning, G. (1998). Petrology and U–Pb geochronology of mafic high-pressure metamorphic coronites from the Tshenukutish Domain, eastern Grenville Province. Precambrian Research 90, 59–83.
Dubé, B., Dunning, G., Roddick, C. & Lauziere, K. (1996). New insights into the Appalachian orogen from geology and geochronology along the Cape Ray Fault, SW Newfoundland. Geological Society of America Bulletin 108, 101–116.
Fuhrman, M. L. & Lindsley, D. H. (1988). Ternary feldspar modeling and thermometry. American Mineralogist 73, 201–215.[Abstract]
Gardien, V., Thompson, A. B., Grujic, D. & Ulmer, P. (1995). Experimental melting of biotite + plagioclase + quartz + muscovite assemblages and implications for crustal melting. Journal of Geophysical Research 100(B8), 15581–15591.[Web of Science]
Godard, G., Martin, S., Prosser, G., Kienast, J. R. & Morten, L. (1996). Variscan migmatites, eclogites and garnet peridotites of the Ulten zone, eastern Austroalpine system. Tectonophysics 259, 313–341.
Hensen, B. J. (1971). Theoretical phase relations involving cordierite and garnet in the system MgO–FeO–Al2O3–SiO2. Contributions to Mineralogy and Petrology 33, 191–214.
Holz, T. & Johannes, W. (1994). Maximum and minimum water contents of granitic melts: implications for chemical and physical properties of ascending magmas. Lithos 32, 149–159.
Hiroi, Y. & Ellis, C. (1994). The use of garnet porphyroblasts in highly deformed pelitic rocks to infer presence of partial melting. EOS Transactions, American Geophysical Union 75, 364.
Indares, A. (1995). Metamorphic interpretation of high-P–T metapelites with preserved growth zoning in garnet, eastern Grenville Province, Canadian Shield. Journal of Metamorphic Petrology 13, 475–486.
Indares, A. (1997). Grt–Ky clinopyroxenites and Grt–Ky restites from the Manicouagan Imbricate Zone: an unusual case of high P–T metamorphism in the Grenville Province. Canadian Mineralogist 35, 1161–1171.[Web of Science]
Indares, A., Dunning, D., Cox, R., Gale, D. & Connelly, J. (1998). High-P–T-rocks from the base of thick continental crust: geology and age constraints from the Manicouagan Imbricate Zone, eastern Grenville Province. Tectonics 17, 426–440.[Web of Science]
Indares, A., Dunning, D. & Cox, R. (2000). Tectono-thermal evolution of deep crust in a Mesoproterozoic continental collision setting: the Manicouagan example Canadian Journal of Earth Sciences 37, 325–340.[Web of Science]
Kohn, M. J., Spear, F. S. & Valley, J. W. (1997). Dehydration-melting and fluid recycling during metamorphism: Rangeley Formation, New Hampshire, USA. Journal of Petrology, 38, 1255–1277.
Kretz, R. (1983). Symbols for rock-forming minerals. American Mineralogist 68, 277–279.[Abstract]
Krogh, T. (1982). Improved accuracy of U–Pb zircon ages by the creation of more concordant systems using an air abrasion technique. Geochimica et Cosmochimica Acta 37, 637–649.
Kroll, H., Evangelakakis, C. & Voll, G. (1993). Two-feldspar geothermometry: a review and revision for slowly cooled rocks. Contributions to Mineralogy and Petrology 114, 510–518.
Le Breton, N. & Thompson, A. B. (1988). Fluid-absent (dehydration) melting of biotite in metapelites in the early stages of crustal anatexis. Contributions to Mineralogy and Petrology 99, 226–237.
Liu, Y. & Zhong, Y. (1997). Petrology of high-pressure granulites from the eastern Himalayan syntaxis. Journal of Metamorphic Geology 15, 451–466.
O’Brien, P. J., Kroner, A., Jaeckel, P., Hegner, E., Zelazniewicz, A. & Kryza, R. (1997). Petrological and isotopic studies on Palaeozoic high-P granulites, Gory Sowie Mts, Polish Sudetes. Journal of Petrology 36, 433–456.
Powell, R. & Downes, J. (1990) Garnet porphyroblast-bearing leucosomes in metapelites: mechanisms, phase diagrams, and an example from Broken Hill, Australia. In: Ashworth, J. R. & Brown, M. (eds) High Temperature Metamorphism and Crustal Anatexis, Mineralogical Society Series 2. London: Unwin Hyman, pp. 105–123.
Pyle, J. M. & Spear, F. (1998). Trace element distribution in garnet: a road map for following pelite reaction histories. Geological Society of America Annual Meeting, Toronto, Abstracts with Programs 30(7), A-231.
Pyle, J. M. & Spear, F. (1999). Yttrium zoning in garnet: coupling of major and accessory phases during metamorphic reactions. Geological Materials Research 1(6), 1–49.
Pyle, J. M. & Spear, F. (2000). An empirical garnet (YAG)–xenotime thermometer. Contributions to Mineralogy and Petrology 200, 51–58.
Spear, F. & Florence, F. P. (1992). Thermobarometry in granulites: pitfalls and new approaches. Precambrian Research 55, 209–241.
Spear, F. S. & Kohn, M. J. (1996). Trace element zoning in garnet as a monitor of crustal melting. Geology 24, 1099–1102.
Spear, F. S. & Parrish, R. R. (1996). Petrology and cooling rates of the Valhalla Complex, British Columbia, Canada. Journal of Petrology 37, 733–765.
Spear, F. S., Kohn, M. J. & Cheney, J. T. (1999). PT paths from anatectic pelites. Contributions to Mineralogy and Petrology 134, 17–32.
Thompson, A. B. (1982). Dehydration melting of pelitic rocks and the generation of H2O undersaturated granitic liquids. American Journal of Science 282, 1567–1595.
Thompson, A. B. & Tracy, R. J. (1979). Model systems for anatexis of pelitic rocks. II. Facies series melting and reactions in the system CaO–KaAlO2–NaAlO2–Al2O3–SiO2–H2O. Contributions to Mineralogy and Petrology 70, 429–438.
Vielzeuf, D. & Holloway, J. R. (1988). Experimental determination of the fluid absent melting relations in the pelitic system. Contributions to Mineralogy and Petrology 98, 257–276.[Web of Science]
Vielzeuf, D. & Pin, C. (1989) Geodynamic implications of granulitic rocks in the Hercynian belt. In Daly, J. S., Cliff, R. A. & Yardley, B. W. D. (eds) Evolution of Metamorphic Belts. Geological Society, London, Special Publications 43, 343–348.
Williams, M. L. & Grambling, J. A. (1990). Manganese, ferric iron and the equilibrium between garnet and biotite. American Mineralogist 75, 886–908.[Abstract]
Willner, A. P., Rötzler, K. & Maresch, W. V. (1997). Pressure–temperature and fluid evolution of quartzofeldspathic metamorphic rocks with a relic high-pressure granulite-facies history from the central Erzgebirge (Saxony, Germany). Journal of Petrology 383, 307–336.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Ferrando, F. Rolfo, and B. Lombardo Fluid evolution from metamorphic peak to exhumation in Himalayan granulitised eclogites, Ama Drime range, southern Tibet European Journal of Mineralogy, July 1, 2007; 19(4): 439 - 461. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||