Granulites and charnockites of the Gruf Complex: Evidence for …€¦ · Granulites and...

Lithos 124 (2011) 17–45

Contents lists available at ScienceDirect

Lithos

j ourna l homepage: www.e lsev ie r.com/ locate / l i thos

Granulites and charnockites of the Gruf Complex: Evidence for Permian ultra-hightemperature metamorphism in the Central Alps

A. Galli ⁎, B. Le Bayon 1, M.W. Schmidt, J.-P. Burg, M.J. Caddick, E. ReusserDepartment of Earth Sciences, ETH Zurich, Sonnegstrasse 5, CH-8092 Zurich, Switzerland

⁎ Corresponding author.E-mail address: [email protected] (A. Galli)

1 Present address: BRGM, 3 avenue Claude-GuillemiCedex 2 France.

0024-4937/$ – see front matter © 2010 Elsevier B.V. Aldoi:10.1016/j.lithos.2010.08.003

a b s t r a c t
a r t i c l e i n f o
Article history:Received 13 January 2010Accepted 3 August 2010Available online 10 August 2010

Keywords:Gruf ComplexGranulitesCharnockitesUHT metamorphismCentral Alps

We present a detailed field and petrological study of charnockites and ultra-high temperature (UHT)granulites from the Gruf Complex, eastern Central Alps. Charnockites occur as up to 0.5 km wide and 8 kmlong, internally boudinaged, opx-bearing sheet-like bodies within the regionally dominant migmatiticbiotite-orthogneisses. Granulites occur as garnet–orthopyroxene–biotite–alkali feldspar-bearing schlieren(±sapphirine, sillimanite, cordierite, corundum, spinel, plagioclase, and quartz) within charnockites and asresidual enclaves both in the charnockites and the migmatitic orthogneisses. Thermobarometric calculations,P–T pseudosections and orthopyroxene Al content, show that both charnockites and granulites equilibratedat metamorphic peak conditions of T=920–940 °C and P=8.5–9.5 kbar. Peak assemblages weresubsequently overprinted by intergrowth, symplectite and corona textures involving orthopyroxene,sapphirine, cordierite and spinel at T=720–740 °C and P=7–7.5 kbar. We suggest that granulites andcharnockites are lower crustal relicts preserved in the migmatitic orthogneisses. Garnet diffusion modellingshows that metamorphic garnet–opx±sapphirine±sillimanite peak assemblages and post-peak reactiontextures always involving cordierite developed during two separate metamorphic cycles. Peak assemblagesreflect UHT metamorphism related to post-Varican Permian extension, but post-peak coronae andsymplectites formed during the mid-Tertiary, upper amphibolite facies, Alpine regional metamorphism.Fluid-absent partial melting of pelitic and psammitic sediments during the Permian UHT event lead to theformation of charnockitic magmas and granulitic residues. Intense melt loss and thorough dehydration of thegranulites (although retaining biotite) favoured the partial preservation of peak mineral assemblages duringAlpine metamorphism.

.n, BP 36009, F-45060 Orléans

l rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

Understanding ultra-high temperature (UHT) metamorphism atcrustal levels, defined by temperatures in excess of 900 °C and typicalpressures of 7–13 kbar (Harley, 1998a), is one of the most challengingtasks inmetamorphic petrology (Harley, 2004, 2008; Kelsey, 2008). Mostcommonly, very high-grade mineral assemblages including sapphirine,aluminous orthopyroxene, garnet, sillimanite, spinel, cordierite orcorundum, and post-peak reaction-textures such as moats and symplec-tites are preserved in Mg–Al-rich metapelites (Harley et al., 1990; Mouriet al., 1996; Raith et al., 1997; Moraes et al., 2002; Kelsey et al., 2003;Sajeev and Osanai, 2004; Osanai et al., 2006; Barbosa et al., 2006; Santoshand Sajeev, 2006; Tong andWilson, 2006; Brandt et al., 2007; Leite et al.,2009). Thus, despite being volumetrically rare, such rock types are themost appropriate for investigating UHT conditions, and are fundamental

for reconstructing the tectono-metamorphic history of high-gradeterrains because of their multitude of mineral possible reactions.

During the last two decades, our ability to better determine themetamorphic conditions of high-grade metamorphic rocks hasimproved significantly. Advances have been made in the accuracy of“conventional thermometers” (for review see Harley, 2008), the useof Al-solubility in orthopyroxene (Harley and Green, 1982; Aranovichand Berman, 1997; Harley, 1998b; Harley and Motoyoshi, 2000),retrieval methods in order to overcome the diffusive effect of resettingof mineral composition during cooling and/or retrograde metamor-phism (Fitzsimons and Harley, 1994; Pattison and Begin, 1994;Pattison et al., 2003), and the development of internally-consistentthermodynamic datasets (e.g. Powell and Holland, 1988; Holland andPowell, 1990, 1998; Berman and Aranovich, 1996). As a result of theseadvances, more than 40 UHT localities have been identifiedworldwide (Brown, 2006, 2007, review in Kelsey, 2008).

In this study, we focus on the granulites and charnockites of theGruf Complex (Central Alps), representing a long-standing enigma inAlpine geology. In the Gruf Complex, rare blocks of dark Mg–Al-richsapphirine-bearing granulites were discovered by Cornelius (1916)and Wenk et al. (1974) in two talus slopes in Val Codera (locations

http://dx.doi.org/10.1016/j.lithos.2010.08.003

mailto:[email protected]

http://dx.doi.org/10.1016/j.lithos.2010.08.003

http://www.sciencedirect.com/science/journal/00244937

18 A. Galli et al. / Lithos 124 (2011) 17–45

indicated in Wenk and Cornelius, 1977), but have never been foundin-situ, in part because of the hostile topography of the Gruf area.Estimates of metamorphic peak-conditions suggested TN800 °C andP=10±2 kbar, conditions pre-dating the formation of symplectitictextures and coronae at T=750 °C and P=5 kbar (Barker, 1964;Ackerman and Seifert, 1969; Droop and Bucher-Nurminen, 1984).Droop and Bucher-Nurminen (1984) suggested that the Grufgranulites formed during a single Alpine metamorphic cycle charac-terised by nearly isothermal decompression. Zircons from theGruf granulites (Liati and Gebauer, 2003) yielded SHRIMP ages of272.0±4.1 Ma for oscillatory cores, interpreted as a magmaticcrystallisation age, and 32.7±0.5 Ma for the homogeneous rims,inferred to represent the age of the granulite facies metamorphism.Similar ages of 33±4.4 Ma (Schmitz et al., 2009) have been reportedfrom monazite grains included in or intergrown with high-temper-ature minerals (e.g. sapphirine, orthopyroxene).

However, metamorphic conditions of TN800 °C and P=10±2 kbardo not fit any knownmetamorphic event of regional significance in theCentral Alps, questioning a close correlation of the Gruf Complex withother Alpine tectonic units. Furthermore, the strong refractory charac-ter, the occurrence of sapphirine–quartz inclusions ingarnet (DroopandBucher-Nurminen, 1984) and the highly aluminous character oforthopyroxene porphyroblasts (Al2O3

opx≈9 wt.%), strongly suggestthat the Gruf granulites might have formed at higher temperaturethan previous estimates, namely at UHT conditions, making theirregional metamorphic significance even more puzzling.

The aims of this study are: (i) to place the Gruf granulites in theirgeological context and characterise the different granulite types; (ii) tounderstand the regional extent, abundance and relations to the countryrocks of these granulites; (iii) to refine the estimates of peak P–Tconditions and the post-peak history with modern thermodynamic tools(P–Tpseudosections, “conventional thermobarometry”accounting for Fe–Mg back-diffusion, Al-solubility in orthopyroxene) and (iv) to integratethese granulites within the tectonic and metamorphic evolution of theCentral Alps.

2. Geological setting

The Central Alps are characterised by a Barrovian-type metamorphicbelt (Niggli, 1970;Freyetal., 1974;Wenk,1975;ToddandEngi, 1997; Freyand Ferreiro Mählmann, 1999; Engi et al., 2004). Mineral isograds,isotherms and isobars have a concentric shape and define the LepontineDome (Wenk, 1955; Trommsdorff, 1966; Todd and Engi, 1997). Thispattern crosscuts tectonic nappe boundaries, indicating that Barrovian

Fig. 1. Sketch map of the eastern Central Alps, on the Switzerland–Italy border (in

peak-conditions were achieved after nappe stacking (Niggli and Niggli,1965; Wenk, 1970; Trommsdorff, 1966; Bernotat and Bambauer, 1982;Todd and Engi, 1997).Metamorphic conditions increase southwards fromupper greenschist to upper amphibolite facies, with migmatizationoccurring in the southern Lepontine Dome (Berger et al., 2005; Burriet al., 2005). Anatexis occurred mostly through water-assisted melting(Berger et al., 2008) at about 700 °C and 6–8 kbar (Burri et al., 2005)between 32 and 22Ma (Hänny et al., 1975; Köppel et al., 1981; Gebauer,1996; Berger et al., 2009; Rubatto et al., 2009). Further south, theLepontineDome is truncated by the Insubric Line (Fig. 1), amajor tectonicboundary separating the high-grade Lepontine from the Southern Alpswhich experienced almost noAlpinemetamorphism (Schmid et al., 1989;Steck and Hunziker, 1994).

The Gruf Complex, which is approximately 20×10 km in size, islocated at the south-easternmargin of the Lepontine Dome. To the north,it is borderedby theMesozoicophioliticChiavennaUnit, to theeast it is cutby the 30–32Ma old calc-alkaline Bergell Pluton, while to the south andwest it is confined by the Novate S-type Granite and the Adula Nappe(Fig. 1). TheGruf Complex predominantly consists ofmigmatitic quartzo–feldspathic orthogneisses, paragneisses and biotite–sillimanite–garnet–(±cordierite)-bearing metapelitic rocks (Fig. 2). The latter show mid-Tertiary metamorphic conditions of upper amphibolites facies (Bucher-Nurminen and Droop, 1983). Migmatization mostly occurred contempo-raneously with (or slightly post-dating) emplacement of the BergellPluton (Berger et al., 1996;Davidsonet al., 1996).Numerousm-size lensesand cm-size nodules of mafic, ultramafic, metapelitic and calcareouscomposition occur throughout the migmatitic gneiss sequence (Artus,1959; Moticska, 1970; Wenk, 1973; 1982; 1986; 1992; Diethelm, 1989),and are particularly concentrated close to the contact with the BergellPluton (Moticska, 1970; Wenk and Cornelius, 1977). These rocks havebeen regardedas relicts of a formernappeboundary (Diethelm,1989) andwere interpreted as the equivalent of the Chiavenna Unit and theBellinzona–Dascio Zone (Diethelm, 1989; Davidson et al., 1996; Schmidet al., 1996). Rare blocks of sapphirine-bearing granulites reported fromthe Gruf Complex represent the only crustal rocks formed at TN800 °Cwithin the entire Central Alps.

3. Field relations

3.1. Rock types

Detailed mapping (1:25,000) of the Gruf Complex has revealed theoccurrence of charnockite sheets and six different types of granulite(Fig. 2): dark, massive sapphirine–orthopyroxene–cordierite–garnet

set). Frame: study area. Swiss co-ordinates are given with units in kilometres.

Fig. 2. Simplified geological map of the Gruf unit and locations of charnockites and different types of granulites in simplified geological map of the Gruf Complex. Location of theinvestigated samples is also indicated.

19A. Galli et al. / Lithos 124 (2011) 17–45

granulite (type A), dark, massive sapphirine-bearing orthopyroxene–sillimanite–cordierite–garnet granulite (type B), dark, massive orthopyr-oxene–cordierite–garnet granulite (type C), dark, foliated sapphirine–orthopyroxene–garnet granulite (type D), felsic, foliated sapphirine-bearing orthopyroxene–garnet granulite (type E) and felsic, foliatedsapphirine–orthopyroxene granulite (type F). Type A and B granulitescorrespond to the sapphirine-bearing samples described in boulders fromthe debris scree to the east of Bresciadega in Val Codera (Cornelius, 1916),and investigated by Droop and Bucher-Nurminen (1984). Type C to Fgranulites and charnockites have not been previously identified.

3.2. Occurrence

The different types of granulite are widespread throughout the entireGruf Complex (Fig. 2). TypesAandBhavebeen identifiedas cm- tometer-size enclaves (Fig. 3a), both within migmatitic biotite-rich orthogneisses(Valle dei Vanni, Swiss coordinates 763′000/130′200) and withincharnockites (Val Piana, 761′050/127′070). However, no type A or Bgranulite has been found in outcrops at the historical locality nearBresciadega. Type C granulites have been found exclusively as meter-sizeboulders in the main rivers of Val Codera (758′650/125′400) and ValSchiesone (752′180/129′340). Type D, E and F granulites form up to 1 mlong and40 cm thick “schlieren”within charnockites (Fig. 3b) inVal Piana(761′000/127′100), Valle del Conco (760′160/128′540), and Val Aurosina(758′030/128′140), also occurring as boulders in the main river of ValCodera. Large (b0.5 km wide and b8 km long), east–west strikingcharnockites constitute internally boudinaged, sheet-like bodies withinthe regionally dominant migmatitic biotite-rich orthogneisses (Fig. 3c).Two major bodies have been mapped in the core of the Gruf Complex,near the crest between Val Codera and Val Chiavenna (Fig. 2). Severalsmaller lenses (b500 m long) occur in Val Piana, Valle del Conco and ValAurosina. In general, the charnockitic boudins and lenses are internallyundeformed or preserve magmatic flow structures. Up to 50 cm widemylonites commonly separate charnockites from the adjacent rocks.

4. Whole rock composition and chemographic analysis

Based on their peak mineral assemblages (Table 1), the six types ofgranulite and charnockites can be classified into fourmajor groups. TypesAandDare sapphirine-bearing, sillimanite- andquartz-absent; typeBandthe coarse-grained domains of type C are sillimanite-bearing, quartz- andsapphirine-absent; type E and F are quartz- and sapphirine-bearing,sillimanite-absent; fine-grained domains of type C and the charnockitesare quartz-bearing, sillimanite- and sapphirine-absent. All rock types aregarnet-, orthopyroxene- and biotite-bearing.

Bulk compositionsof homogeneous charnockites and typeA, B, C andF granulites were obtained using a Panalytical Axios wave-lengthdispersive XRF spectrometer (WDXRF, 2.4 kV) at ETH Zürich. Granulitetypes D and E, occurring as irregular schlieren in charnockites, were notanalysed because they could not be separated from the host rock.Representative rock compositions are given in Table 2 and discussed interms of XMg, A/AFM [A=Al2O3–(K2O+Na2O+CaO)] and S/SFM[S=SiO2–Al2O3–CaO–5 (K2O+Na2O)+FeO+MgO], following the ap-proach of Harley (2008). The relationships between rock compositionsand peak mineral assemblages are displayed via MFAS ((Mg+Fe)O–SiO2–Al2O3) diagrams projected from the three feldspar components(Fig. 4a and b). Relations in quartz-saturated rock types are alsorepresented by an AFM (Al2O3–FeO–MgO) diagram projected fromquartz and the three feldspar components (Fig. 4c). We note here thatslight discrepancies between bulk compositions and the trianglesdefined by mineral tie-lines are most likely related to the oxidationstate of Fe: all Fe is calculated as Fe2+, thus overestimating the (Mg,Fe)O-component and underestimating the Al2O3 component, to whichferric iron would be added.

Granulite typeA is quartzundersaturated(40–42 wt.%SiO2, S/SFM:59–60), rich in alumina (24–25 wt.% Al2O3, A/AFM: 25–27) and highlymagnesian (XMg≈0.74). In a MFAS diagram, (Fig. 4a), type A composi-tions plot on the silica-poor side of the sillimanite–orthopyroxene tie-line,within the sapphirine–garnet–orthopyroxene field (field 1 in Fig. 4a).

image of Fig.�2

Fig. 3. Field aspects of granulites and charnockites. (a) Retrogressed type A granulite enclave within migmatitic biotite-orthogneiss (Valle dei Vanni, 763′024/130′193); (b) type Egranulite schlieren (sample PiaGra1) within charnockite (sample PiaMig1, Val Piana, 761′118/127′060); (c) internally boudinaged, sheet-like body of charnockite within migmatiticbiotite-rich orthogneiss (upper Val Casnaggina, 762′644/130′526). Few cm-thin mylonitic zones occur at the contacts between charnockite andmigmatitic biotite-orthogneiss (bolddashed line) and between charnockitic boudins (thin dashed lines).


Type B and C-coarse granulites are slightly richer in silica than typeA (43–46 wt.% SiO2, S/SFM: 62–63), again containing high aluminacontents (23–26 wt.% Al2O3, A/AFM: 25–29). Type B granulites aremore magnesian (XMg: 0.72–0.76) than type C-coarse (XMg≈0.58). InMFAS space, they plot within the sillimanite–garnet–orthopyroxenefield (field 2 in Fig. 4a), in accordance with the observed sillimanite-bearing, sapphirine-absent peak assemblage.

Type F granulite is considerably richer in silica (~71 wt.% SiO2, S/SFM:92) and poorer in alumina (~15 wt.% Al2O3, A/AFM: 32) than types A, Band C-coarse, with XMg≈0.63. In a MFAS diagram, this plots on thealumina-rich side of the garnet–quartz tie-line, within the sapphirine–garnet–quartz triangle (field 3 in Fig. 4b). It falls in the sapphirine–orthopyroxene–garnet triangle of an AFM diagram (field 5 in Fig. 4c).

Type C-fine granulites and charnockite are quartz saturated.However, type C-fine granulites are poorer in silica (~49 wt.% SiO2,S/SFM: 65) than charnockite (61–75 wt.% SiO2, S/SFM: 78–95), andare considerably poorer in silica and richer in both iron andmagnesium than the type F granulite. Alumina content (~16 wt.%Al2O3) and XMg (~0.60) of type C-fine granulites are similar to type F,but the A/AFM value is lower (A/AFM: 16). Charnockites displayvariable alumina content (13–19 wt.% Al2O3, A/AFM: 4–34) and areless magnesian than all granulites (XMg: 0.30–0.44). In a FMASdiagram, the compositions of both C-fine granulite and charnockite lieon the alumina-poor side of the garnet–quartz tie-line, within thequartz–garnet–orthopyroxene field (labelled 4 in Fig. 4b). Type C-finegranulite plots close to the garnet–orthopyroxene tie-line, inaccordance with the high proportion of garnet and orthopyroxene

in the rock, while charnockites plot much closer to quartz, inagreement with the high amount of quartz observed. In an AFMdiagram, type C-fine plots within the orthopyroxene–garnet–biotitefield (labelled 6 in Fig. 4c), while charnockites plot within theorthopyroxene–garnet–ilmenite triangle (7 in Fig. 4c), again inaccordance with the observed peak mineral assemblages.

5. Mineralogical description of granulite types and charnockites

Mineral assemblages and textural relations for the differentgranulites and charnockite are summarized in Table 1. The averagerelative abundance of the main mineral phases present in eachinvestigated sample is given in the text as vol.%.

5.1. Dark, massive sapphirine–orthopyroxene–cordierite–garnetgranulite (Type A, sample BRE7)

The rock is melanocratic, coarse-grained and massive. The peakmetamorphic assemblage contains garnet (20%), brownish, prismaticorthopyroxene (12%), darkblue, prismatic sapphirine (20%), biotite (17%),and alkali feldspar (8%). Garnet occurs as large (b2.5 cm), roundedporphyroblasts with irregular lobate shapes (Fig. 5a). Rims are generallyreplacedby late biotite or lamellar orthopyroxene–cordierite symplectites(Fig. 5b). Orthopyroxene (opx1) forms idiomorphic porphyroblasts up to4 mm across, with lobate and corroded grain boundaries (Fig. 5a). Opx1may be intergrown with sapphirine and partially replaced by late biotite.Prismatic sapphirine occurs as idiomorphic porphyroblasts up to 3mm

image of Fig.�3

Table 1Main mineral assemblages and textural features in Gruf granulites and charnockite investigated in this study.

Rock type Type A Type B Type C (coarse/fine) Type D Type E Type F Charnockite

Sample name BRE7 BRE10 GraG/GraF CodGra5 PiaGra1 Conc13 PiaMig1

Location Bresciadega Bresciadega Codera river Codera River Val Piana Valle del Conco Val Piana

MineralsGarnetMatrix porphyroblast P P P/P P P P PIncluded in opx opx opx/– – – – –

OrthopyroxeneMatrix porphyroblast P P P/P P P P PSym around Grt pP pP pP/pP pP – – –

Sym around Opxpor pP pP –/– pP – – –

Inclusion grt, bt grt, bt grt/– grt grt – –

SapphirineMatrix porphyroblast P –* –/– P – – –

Sym around Grt – pP –/– pP – – –

Sym around Sil – pP pP/– – P P –

Sym around Crn – – –/– – P – –

Inclusion grt, opx, bt grt, opx grt/grt grt grt – –

SillimaniteMatrix porphyroblast – P P/– – prP prP –

Inclusion grt, spr, ksp grt, ksp, spr grt, opx/grt grt grt – grt

BiotiteMatrix P P P/P P P P PCoronae around opx pP pP pP/pP pP pP pP pPInclusion grt, opx grt, opx, sil grt, opx/grt, opx grt, opx grt grt, opx grt, opx

CordieriteCoronae pP pP pP/pP pP – – –

CorundumMatrix – – –/– – prP – –

Inclusion grt – grt/– – – – –

SpinelSym around Grt – pP –/– pP – – –

Sym around Sil – pP pP/– – pP pP pPSym around Crn prP – –/– – pP – –

Inclusion grt, opx, spr grt, spr grt, opx/grt opx, spr grt grt –

Alkali feldsparMatrix P P –/P P P P PInclusion grt grt, opx, sil –/– grt, opx grt – grt

PlagioclaseMatrix – P –/P P P P PSym around opx pPInclusion grt grt, opx grt/grt grt, spr grt – grt, opx

QuartzMatrix – – –/P – P P PInclusion grt grt, opx grt/grt, opx – grt – grt, opx

P: part of the metamorphic peak assemblage; prP: part of the metamorphic pre-peak assemblage; pP: part of the metamorphic post-peak assemblage. List of abbreviations forinclusions: grt — in garnet; opx — in orthopyroxene; spr — in sapphirine; sil— in sillimanite; ksp — in alkalifeldspar; bt — in biotite. *Prismatic sapphirine occurs in type B granuliteexclusively in micro-domains of different bulk-rock composition and are therefore not considered part of the peak assemblage representative for type B granulite (see text).


long (Fig. 5a), with inclusions of sillimanite and spinel (Fig. 5c). Generally,sapphirine is separated from the matrix by a thin moat of cordierite or isembayed by orthopyroxene porphyroblasts (Fig. 5c). Droop and Bucher-Nurminen (1984) reported rare sapphirine–quartz inclusions in the rimsof garnet porphyroblasts. Biotiteoccurs asmillimetric, unorientedflakes inthematrix as part of the peak assemblage or as fine-grained inclusions ingarnet and opx1 porphyroblasts. Alkali feldspar forms microperthitic,xenomorphic porphyroblasts up to 5 mm in size.

Lamellar orthopyroxenes (3%), cordierite (10%), plagioclase (2%) and alate biotite generation (8%) characterise the post-peak assemblage.Lamellar orthopyroxene (opx2), together with cordierite, forms symplec-titic coronae around garnet porphyroblasts (Fig. 5b). Opx3 with thin

interstitial plagioclase patches forms lamellar symplectites at the opx1–alkali feldspar interface (Fig. 5d). Cordieritemantles prismatic sapphirineor, together with opx2, forms symplectites around garnet (Fig. 5b).Plagioclase occurs as interstitial patches between opx1 and alkali feldspar,rimming opx3 lamellar symplectites. New biotite flakes grow partlyreplacing the rims of opx and garnet porphyroblasts.

Spinel and sillimanite are exclusively included in the cores of bothprismatic sapphirine and garnet porphyroblasts, and are absent from thematrix. Spinel inclusions, in contact with quartz, as well as spinel–sapphirine–corundum composite inclusions, occur in the rim of garnetporphyroblasts. Corundum, quartz, staurolite, sphene and rutile occur onlyas inclusions in garnet porphyroblasts.

Table 2Major element bulk composition (wt.%) for different types of granulites and charnockite.

Rock type Type A Type B Type C (coarse) Type C (fine) Type F Charno “Calculated residium”**

Sample BRE7 BRE10 GraG GraF Conc13 PiaMig1

SiO2 40.224 46.710 43.340 48.480 71.613 70.170 41.21TiO2 0.560 0.583 0.542 2.037 0.268 0.417 2.77Al2O3 25.013 23.614 23.885 15.792 15.440 14.804 16.54FeO* 11.314 9.137 15.516 15.380 1.502 3.167 19.53MnO 0.125 0.096 0.122 0.121 0.006 0.070 0.20MgO 18.081 15.590 12.151 13.075 1.456 0.884 15.76CaO 0.559 0.329 0.760 1.358 2.946 2.446 1.52Na2O 0.334 0.359 0.421 0.232 1.861 3.876 0.02K2O 2.550 2.636 1.475 2.434 3.970 2.912 1.87P2O5 0.069 0.070 0.085 0.164 0.103 0.149 –

Cr2O3 0.001 0.001 0.001 0.046 0.002 0.001 –

NiO 0.001 0.001 0.005 0.023 0.001 0.001 –

H2O 0.894 0.806 1.631 0.213 0.487 0.369 0.58Tot 99.724 99.933 99.933 99.353 99.655 99.265 100XMg 74 75 58 60 63 33 45A/AFM 25 27 28 16 32 11S/SFM 59 63 63 65 92 90

*Total iron as FeO; H2O content estimated from Loss of Ignition (LOI); XMg=Mg/(Mg+Fe2+); A: Al2O3–(K2O+Na2O+CaO); S: SiO2–Al2O3–CaO–5*(K2O+NaO)+FeO+MgO;**Calculated bulk-rock composition following progressive melt extraction from a composite of rocks GraF and PiaMig1 during P–T path a–b–e–c–d (Fig. 16a), explained inSection 9.3.1.


5.2. Dark, massive sapphirine-bearing orthopyroxene–sillimanite–cordierite–garnet granulite (Type B, sample BRE10)

The rock is melanocratic, coarse-grained and massive. The metamor-phic peak assemblage is comprised of garnet porphyroblasts (20%),brownish, prismatic orthopyroxene (12%), light greyish, prismaticsillimanite (15%), biotite (20%), alkali feldspar (10%) and plagioclase(2%). Compared to type A granulites, the peak assemblage includessillimanite but is sapphirine-absent.

Garnet occurs as large (b2.0 cm), xenomorphic, strongly corrodedporphyroblasts with irregular shapes. Frequently, garnet is almostcompletely replaced by opx porphyroblasts or is preserved asdisrupted and relict inclusions. Garnet is commonly surrounded bypolymineralic post-peak symplectitic coronae of orthopyroxene andcordierite±sapphirine and/or spinel. Prismatic orthopyroxene (opx1)forms large (≤3 mm), hypidiomorphic porphyroblasts, in partreplaced by late biotite. Common inclusions in prismatic opx1 arebiotite, sillimanite, relicts of garnet porphyroblasts, as well as perthiticalkali feldspar rimmed by a thin film of plagioclase. Opx1 frequentlysurrounds garnet porphyroblasts. Coarse prismatic sillimanite porphyro-blasts (up to 2–3 mm) are abundant in the matrix. Sillimanite grains arecorroded and exhibit an inner symplectitic corona of sapphirine±spinel,and an outer moat of cordierite (Fig. 5e). Less frequently, relict sillimanitegrains are preserved as inclusions in opx porphyroblasts (Fig. 5f). Biotite ispart of the matrix and occurs as millimetric flakes with a only weakpreferred orientation. Alkali feldspar and plagioclase occur as isolated,xenomorphic grains within the matrix. Locally, in silica-poorer micro-domains similar to type A granulite, rare prismatic sapphirine occursinstead of sillimanite as part of the peak assemblage. Such domains aretreated as type A granulite and are not representative of type B granulite.

Lamellar sapphirine (4%), spinel (1%), orthopyroxene (3%), cordierite(13%) and late biotite mark the post-peak assemblage. Sapphirine occursas symplectitic sapphirine–cordierite (±spinel) corona surroundingprismatic sillimanite (Fig. 5e). Sapphirine2, together with opx2, spineland cordierite, formspart of complex symplectites partly replacing garnet.Rare, lamellar spinel grains compose symplectitic coronae around garnetand sillimanite porphyroblasts. Lamellar orthopyroxene shows the sametextural pattern as opx2 and opx3 in type A granulites. Cordierite occurswithin symplectitic corona around garnet or as corona surroundingsillimanite–sapphirine (±spinel) aggregates (Fig. 5e). Late biotite showsthe same textural relations as in type A granulite.

Quartz, sphene and rutile are accessories and occur only asinclusion in garnet porphyroblasts.

5.3. Dark, massive orthopyroxene–cordierite–garnet granulite (Type C,samples GraG and GraF)

The rock is melanocratic, massive and characterised by domains ofdifferent grain-size and mineral assemblage (Fig. 6a). In this study, thecoarser-grained rock domains are referred as type C-coarse granulite(sample GraG), while the finer-grained domains are referred as type C-fine granulite (sample GraF).

The metamorphic peak assemblage of type C-coarse granuliteconsists of garnet (30%), brownish, prismatic orthopyroxene (28%),prismatic sillimanite (8%) and biotite (9%).

Garnet (b2.0 cm) occurs as rounded grains with lobate shapes and iscommonlypartly replacedbyorthopyroxene, cordierite and latebiotite, oroccurs as inclusions in opx (Fig. 6b). Garnet rarely has symplectitic opx–cordierite coronae. Prismatic orthopyroxene (opx1) forms large (b 1.5 cm),hypidiomorphic porphyroblasts overgrowing sillimanite and garnet(Fig. 6b). Prismatic sillimanite is common in the matrix or as needle-likeinclusions in opx1 and garnet. The grains generally exhibit an innersymplectitic corona of spinel–cordierite±rare sapphirine and an outercorona of cordierite (Fig. 6c). Biotite occurs in the matrix as unorientedflakes.

Lamellar orthopyroxene (2%), spinel (2%), rare sapphirine (1%),cordierite (14%) and late biotite (6%) form the post-peak mineralassemblage. Lamellar orthopyroxene (opx2), together with cordierite,occurs in symplectites around garnet. Lamellar spinel and raresapphirine form polymineralic spinel–sapphirine–cordierite coronae,partially or completely replacing prismatic sillimanite, or occurring astiny inclusions in garnet. Cordierite surrounds garnet porphyroblastsand prismatic sillimanite, while late biotite partly replaces the rims ofgarnet and orthopyroxene porphyroblasts. Corundum, staurolite, rutile,sphene, apatite, quartz and plagioclase are exclusively observed asinclusions in garnet porphyroblasts. Rarely, sapphirine and quartzincluded in garnet are in mutual contact (Fig. 6d).

The peak mineral assemblage of type C-fine granulite comprisesgarnet (35%), orthopyroxene (18%), biotite (10%), quartz (10%), plagio-clase (10%) and alkali feldspar (3%). Garnet generally occurs as small(b1 mm), rounded porphyroblasts showing lamellar symplectites oforthopyroxene and cordierite (Fig. 6e and f). Scarce larger garnet grains(b4 mm)are also present.Orthopyroxene formshypidiomorphic porphyr-oblasts, partially replaced by late biotite. Commonly, opx mantles quartzgrains, separating quartz from biotite (Fig. 6e). Biotite forms unorientedflakes. Unlike the type C-coarse granulite, fine-grained domains containquartz, plagioclase, and alkali feldspar, as part of the matrix.


Lamellar orthopyroxene (4%), cordierite (5%) and late biotite (5%)form the post-peak mineral assemblage. Lamellar orthopyroxene(opx2), together with cordierite, develops as symplectite around

Fig. 4. a–b) MFAS (MgO+FeO–Al2O3–SiO2) diagram projected from the three feldsparcomponents representing bulk compositions and analysed peak mineral assemblages.Total iron as FeO. This ternary diagram does not display Fe–Mg relations betweenphases. a) for type A, B granulites and coarse-grained domains of type C; b) for fine-grained domains of type C and F granulites and charnockites; c) AFM (Al2O3–FeO–MgO)diagram projected from quartz and the three feldspar components representing bulkcompositions and analysed peak mineral assemblages of type C-fine and F granulitesand charnockites.

garnet (Fig. 6f). Cordierite displays the same textural features as in thetype C-coarse granulites. Sapphirine, spinel, sillimanite, apatite, spheneand rutile occur exclusively as inclusions in garnet porphyroblasts.

5.4. Dark, foliated sapphirine–orthopyroxene–garnet granulite (Type D,sample CodGra5)

The rock ismelanocratic,fine- tomedium-grainedandweakly foliated.Garnet (15%), prismatic orthopyroxene (10%), sapphirine (4%), biotite(36%), rare alkali feldspar (2%) and antipertithic plagioclase (9%) form thepeak mineral assemblage. Garnet occurs as rounded, xenomorphicporphyroblasts (3–4 mm in size). These exhibit irregular, lobate-shapedrims and are strongly replaced by symplectitic aggregates of orthopyrox-ene, cordierite and plagioclase±sapphirine, spinel (Fig. 7a). Prismaticorthopyroxene is present as up to 4 mm long and 3 mm wide,hypidiomorphic porphyroblasts, generally embayed by late biotite.Sapphirine (spr1) occurs in thematrix as prismatic grain strongly replacedby symplectitic aggregates of spinel and plagioclase surrounded by anouter corona of cordierite. Biotite flakes are part of the matrix. Biotiteshows a well-developed preferred orientation and defines the main rockfoliation. Alkali feldspar and plagioclase compose the matrix.

The post-peak mineral assemblage is marked by lamellar orthopyr-oxene (5%), lamellar sapphirine (6%), spinel (4%), cordierite (5%) and latebiotite (4%). Lamellar orthopyroxene (opx2) forms part of lobate opx2–cordierite–plagioclase±sapphirine2±spinel symplectites, mantling andpartly replacing garnet porphyroblasts. Opx2 is often surroundedby a thinfilm of plagioclase. Intergrowths of opx2–plagioclase–cordierite–sapphirine2 can completely replace garnet porphyroblasts (Fig. 7b).Opx3 forms lamellar aggregates between porphyroblastic opx1 andplagioclase. Lamellar sapphirine2 is part of opx2–cordierite–plagioclase–sapphirine2 ± spinel symplectites around garnet porphyroblasts. Spinel,togetherwith sapphirine2, plagioclase, opx2 and cordierite, forms lamellarsymplectites surrounding garnet porphyroblasts and, together withplagioclase and cordierite partly replaces spr1.Cordieriteoccurs in coronaearound garnet porphyroblasts and sapphirine1–plagioclase–spinel sym-plectites. Late biotite partly replaces garnet and orthopyroxene porphyr-oblasts. Apatite, sillimanite and rutile are accessories and occur asinclusions in garnet porphyroblasts.

5.5. Felsic, foliated sapphirine-bearing orthopyroxene–garnet granulite(Type E, sample PiaGra1)

The rock is more leucocratic than other granulite types and isfoliated. Garnet porphyroblasts (10%), prismatic orthopyroxene(10%), sapphirine (1%), biotite (32%), quartz (15%), plagioclase(15%) and perthitic alkali feldspar (10%) compose the metamorphicpeak mineral assemblage. Garnet occurs in two generations. Garnet1consists in coarse-grained, up to 1.5 cm across, atoll-like, hypidio-morphic porphyroblasts wrapped by the foliation. The grains arerounded and show irregular lobate shapes. Late biotite and chloritecoronae commonly surround the porphyroblasts. Garnet2 occurs asb0.2 mm, xenomorphic grains overgrowing the main foliation.Orthopyroxene forms hypidiomorphic porphyroblasts, commonlysurrounded by late biotite coronae or partially replaced by lamellarsymplectites of biotite–quartz. The grains are weakly orientedsubparallel to the main foliation. Sapphirine occur in lobate sapphi-rine–plagioclase symplectites, mantled by equant spinel grains. Theseaggregates are elongated within the rock foliation and partly or

image of Fig.�4

Fig. 5. Photomicrographs of representative mineral assemblages and textural relationships in type A and B granulites: (a) xenomorphic garnet porphyroblast, bluish, prismatic sapphirine andpinkish orthopyroxene porphyroblasts (opx1) and biotite (sample BRE7); (b) opx2–cordierite symplectites replacing mostly garnet porphyroblasts (BRE7); (c) sapphirine grains with spinelinclusions embayedby large opxporphyroblasts (BRE7); (d) orthopyroxene–plagioclase symplectites (opx3) aroundperthitic alkali feldspar andorthopyroxeneporphyroblasts and thinmoat ofcordieritemantling sapphirine porphyroblast (BRE7); (e) double corona of sapphirine–cordierite symplectites (inner) and cordierite (outer) partially replacing prismatic sillimanite formed viareaction 15 (sample BRE10); (f) relict sillimanite overgrown by orthopyroxene porphyroblasts (BRE10).


completely replace former relict sillimanite or corundum (Fig. 7c).Biotite is abundant in the rock matrix occurring as fine-grained flakes(b0.2 mm) that delineate a pronounced preferred orientation definingthe main rock foliation. Quartz, plagioclase and perthitic alkali feldspar arepresent in the rock matrix as fine-grained (b0.2 mm) interstitialaggregates with a slightly preferred shape orientation.

Spinel (2%), late biotite (4%) and chlorite (2%) form the post-peakmineral assemblage. Spinel occurs in sapphirine–plagioclase–spinelsymplectites, mantling sapphirine grains. Late biotite and chloritedevelop at the garnet and orthopyroxene rims.

Relict corundum and sillimanite are preserved in minor amount asstrongly replaced pre-peak mineral phases displaying sapphirine–plagioclase–spinel symplectites (Fig. 7c). Sphene, zircon and monazite areaccessories.

5.6. Felsic, foliated sapphirine–orthopyroxene granulite (Type F, sampleConc13)

These granulites are leucocratic, medium- to coarse-grained andfoliated. The metamorphic peak mineral assemblage comprises garnet

(3%), orthopyroxene (3%), sapphirine (5%), biotite (4%), quartz (36%),plagioclase (31%) and alkali feldspar (10%). Garnet and orthopyroxeneoccur as 1–2mm large, strongly corroded, xenomorphic porphyroblastsdisplaying irregular lobate shapes (Fig. 7d). Sapphirine forms up to 1 mmlong lamellar aggregates which partially or completely substituteprismatic sillimanite (Fig. 7d). The aggregates are elongated within themain foliation and commonly display symplectitic coronae of equantspinel and plagioclase. Biotite forms rare brownish, fine-grained(b0.1 mm) patches in the matrix showing a weak preferred shapeorientation. Elongated and oriented quartz grains up to 3 mm in lengthand 1 mm in width, granoblastic aggregates of plagioclase and subordi-nate perthitic alkali feldspar compose the rock matrix.

Spinel (2%), late biotite (1%) and a second plagioclase generation(4%) form the post-peak mineral assemblage. Lamellar spinel and lateplagioclase form symplectitic aggregates around sapphirine. Latebiotite partly replaces orthopyroxene and garnet porphyroblasts.

Prismatic sillimanite (1%) occasionally preserved in the cores ofsapphirine and rare, fine-grained spinel inclusions in garnet areinterpreted as relics of the pre-peakmineral assemblage. Ilmenite, apatite,monazite and zircon are accessories.

image of Fig.�5

Fig. 6. Field aspect and photomicrographs of representative mineral assemblages and textural relationships in type C granulites: (a) boulder of type C granulite showing coarse-grained (type C-coarse, sample GraG, to the right) and finer-grained (type C-fine, sample GraF) domains; (b) orthopyroxene porphyroblast overgrowing garnet and sillimanitegrains and spinel–cordierite corona around sillimanite (GraG); (c) sapphirine–spinel–biotite–cordierite corona mantling prismatic sillimanite formed via reaction 15 (GraG);(d) back-scatter image of quartz–sapphirine inclusion in garnet porphyroblast (GraG); (e) opx1 porphyroblasts partly replaced by biotite, symplectitic opx2–cordierite mantlinggarnet formed via reaction 11 and opx3 coronae around quartz grain formed probably by fluid-absent biotite melting via reaction 2 (GraF); (f) back-scatter image of opx2–cordieritesymplectites around garnet formed via reaction 11 (GraF).


5.7. Charnockite

Charnockites are leucocratic, medium- to coarse-grained andweakly foliated. These rocks are composed of quartz (40%), plagio-clase (30%), perthitic alkali feldspar (15%), rounded garnet (5%),brownish orthopyroxene (5%), rare biotite (3%) and ilmenite (1%).Amphibole, rutile, apatite and zircon are accessories.

Garnet is present as tiny (b0.5 mm), subrounded, hypidiomorphicgrains showing irregular edges. Garnet is commonly partly substituted bylate biotite. Orthopyroxene occurs as small (up to 0.5 mm), hypidio-morphic porphyroblasts. The grains are frequently corroded or partiallyreplaced by polymineralic coronae of late biotite, amphibole and lateilmenite (Fig. 7e and f). Biotite forms part of the metamorphic peakmineral assemblage and occurs as rare, slightly oriented, small flakes (upto 0.5 mm). Ilmenite occurs as xenomorphic grains in the matrix. Latebiotite2 grows at the expense of garnet or, together with amphibole andilmenite2, surrounding opx porphyroblasts. Amphibole occurs exclusivelyas post-peak phase partly replacing opx. Quartz, plagioclase and perthiticalkali feldspar formtherockmatrix.Generally, grains are slightlyelongatedand, together with rare biotite1, define the main rock foliation.

6. Mineral chemistry

Mineral chemical compositions were obtained using the electronmicroprobe JEOL-8200 at ETH Zurich. Operating conditions for spotanalyses were 15 kV accelerating voltage, 20 nA sample current, 40 scounting time and beam size of b5 μm. Natural and synthetic standardswere used for calibration. Minerals were analysed for SiO2, TiO2, Cr2O3,Al2O3, FeO, MnO, MgO, CaO, Na2O, and K2O. Fe3+ was calculated fromcharge balance. Ferric iron in sapphirine was estimated using theapproach of Higgins et al. (1979): VIFe3+=IVAl3+−(VIAl+Crtot).Representative mineral data are presented in Tables 3–6.

6.1. Garnet

Garnets in granulite types A, B, C-coarse, D and E (Fig. 8a andTable 3) are almost pure pyrope–almandine solid solutions with smallgrossular and spessartine components (Xgrsb0.05, Xspsb0.03). Gar-nets in types A, B, C-coarse and D granulites are relatively rich inmagnesium (Xpyr: 0.45–0.57) than garnets of type E (Xpyr: 0.38–0.41).Garnets in type C-fine and F granulites are generally richer in iron

image of Fig.�6

Fig. 7. Photomicrographs of representative mineral assemblages and textural relationships in type D, E and F granulites and in charnockites: (a) opx–cordierite–plagioclase–sapphirine symplectite replacing garnet porphyroblast via reaction 12 (sample CodGra5); (b) orthopyroxene–sapphirine intergrowth pseudomorph after garnet formed viareaction 12 (CodGra5); (c) spinel–sapphirine–corundum symplectites formed via reaction 9 (sample PiaGra1); (d) sillimanite replaced by sapphirine–plagioclase symplectites andrelictic opx (Conc13); (e) orthopyroxene porphyroblast displaying biotite–amphibole coronae (sample PiaMig1); (f) back-scatter image of biotite–ilmenite–amphibole coronaesurrounding orthopyroxene (PiaMig1).


(Xpyr: 0.35–0.48) than granulites type A, B and C-coarse and showvariable grossular content (Xgrs: 0.03–0.13). Garnets of charnockitesare considerably richer in iron (Xpyr: 0.21–0.27) than garnets ingranulites and display low grossular (Xgrs: 0.04–0.05) and spessartine(Xsps: 0.05–0.07) contents.

Different zoning patterns exist. Most garnet porphyroblasts arecharacterised by decreasing Mg content and slightly increasing Fe2+,Ca and Mn towards the rim. Some garnet porphyroblasts in granulitesof types A, B and C-fine exhibit more complex 4-phase zoningfeatures. Twomajor patterns are recognized in type A and B granulites(see also Droop and Bucher-Nurminen, 1984; Schefer, 2005): i) a largehomogeneous core, an inner zone of rimwards decreasing Mg andincreasing Ca, an outer zone of increasing Mg and decreasing Ca andFe2+, and a rim with decreasing Mg, and slightly increasing Ca andMn; ii) a homogeneous core, an inner zone of rimwards sharplyincreasing Mg and decreasing Fe2+, Ca and Mn, an outer zone ofgently decreasing Mg and increasing Fe2+, and a rim of abruptdecrease of Mg and increase of Fe2+. Garnets in type C-fine granulitesdisplay a homogeneous core, an inner zone of rimwards increasingMgand decreasing Fe2+, an outer zone of homogeneous Mg and Fe2+ but

gently decreasing Ca, and a rim of abruptly increasing Fe2+ andstrongly decreasing Mg and Ca (Fig. 8b).

These complex zoning patterns suggest that the rocks underwentmultiple stages of equilibrium. Zoning pattern i) in type A and Bgranulites probably formed as a consequence of (re)-heating after aperiod of homogenisation of the garnet core. Instead, zoning patternii) in type A and B granulites and zoning in type C-fine granulite mayreflect prograde garnet growth during heating and progressiveextraction of melt, leading to the observed increase of Xpyr from thecore rimwards (Fig. 8b).

6.2. Orthopyroxene

The composition of orthopyroxene varies with its texture (Fig. 9aand Table 4). In type A, B and C-coarse granulites, the highest Al2O3

contents are preserved in the core of large opx1 porphyroblasts (7.6–9.8 wt.% Al2O3 in types A and B; 8.7–9.3 wt.% Al2O3 in type C-coarse),steeply decreasing rimwards probably due to decreasing metamor-phic conditions (5.7–6.7 wt.% Al2O3 in types A; 4.7–5.5 wt.% Al2O3 intype B; 5.8–6.4 wt.% Al2O3 in type C-coarse). In type C-fine and D

image of Fig.�7

Table 3Representative garnet analyses.

Sample BRE7 BRE10 GraG GraF CodGra5 PiaGra1 Conc13 PiaMig1

Texture Co R Co R Co R Co R Co R Co R Co2 Co Co

SiO2 40.570 40.100 40.060 39.970 39.240 38.340 39.880 39.670 39.960 39.500 38.100 38.500 38.82 39.590 38.140TiO2 0.012 0.023 0.002 0.000 0.010 0.011 0.096 0.056 0.028 0.027 0.022 0.010 0.024 0.061 0.022Cr2O3 0.000 0.005 0.021 0.000 0.028 0.000 0.066 0.064 0.000 0.011 0.008 0.000 0.003 0.008 0.000Al2O3 23.200 23.040 22.970 22.700 22.180 22.010 22.180 22.320 22.660 22.710 21.740 22.260 22.205 22.260 21.560FeO 21.400 23.280 22.430 23.770 24.330 26.710 23.040 25.210 23.030 23.520 27.820 28.440 28.23 22.550 31.320Mno 0.256 0.701 0.617 0.574 0.322 0.461 0.599 0.459 0.726 0.892 1.228 0.902 1.58 0.442 2.240MgO 15.130 13.140 14.270 12.620 13.130 11.090 10.930 10.530 12.890 12.420 10.060 9.830 9.465 11.600 6.250CaO 0.753 1.400 0.907 1.064 1.026 1.271 4.180 2.650 1.780 1.710 0.732 0.775 0.753 3.290 1.650Na2O 0.018 0.011 0.035 0.012 0.018 0.017 0.000 0.009 0.012 0.018 0.000 0.001 0.037 0.003 0.000K2O 0.000 0.006 0.001 0.006 0.000 0.018 0.007 0.003 0.000 0.007 0.019 0.000 0.01 0.001 0.000Tot 101.338 101.706 101.313 100.716 100.285 99.928 101.978 100.971 101.087 100.744 99.728 100.718 101.13 99.804 101.182O 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12Si 2.966 2.958 2.947 2.986 2.941 2.922 2.990 2.991 2.968 2.950 2.935 2.940 2.959 2.990 2.970Ti 0.001 0.001 0.000 0.000 0.001 0.001 0.005 0.003 0.002 0.002 0.001 0.001 0.001 0.003 0.001Cr 0.000 0.000 0.001 0.000 0.002 0.000 0.004 0.004 0.000 0.001 0.000 0.000 0 0.000 0.000Al 1.999 2.003 1.992 1.999 1.960 1.977 1.960 1.984 1.984 1.999 1.974 2.003 1.995 1.982 1.979Fe2+ 1.238 1.356 1.263 1.455 1.367 1.521 1.399 1.565 1.352 1.369 1.637 1.699 1.709 1.392 1.960Fe3+ 0.070 0.080 0.117 0.030 0.158 0.181 0.046 0.025 0.079 0.100 0.155 0.116 0.09 0.032 0.079Mn 0.016 0.044 0.038 0.036 0.020 0.030 0.038 0.029 0.046 0.056 0.080 0.058 0.102 0.028 0.148Mg 1.649 1.445 1.565 1.405 1.467 1.260 1.221 1.183 1.427 1.383 1.155 1.119 1.075 1.306 0.725Ca 0.059 0.111 0.072 0.085 0.082 0.104 0.336 0.214 0.142 0.137 0.060 0.063 0.061 0.266 0.138Na 0.003 0.002 0.005 0.002 0.003 0.002 0.000 0.001 0.002 0.003 0.000 0.000 0.005 0.000 0.000K 0.000 0.001 0.000 0.001 0.000 0.002 0.001 0.000 0.000 0.001 0.002 0.000 0.001 0.000 0.000Sum 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8XMg 0.558 0.502 0.531 0.486 0.490 0.425 0.458 0.427 0.501 0.515 0.392 0.381 0.374 0.522 0.262XMg* 0.571 0.516 0.553 0.491 0.518 0.453 0.466 0.431 0.486 0.498 0.414 0.397 0.386 0.516 0.270Xpyr 0.557 0.489 0.533 0.471 0.499 0.432 0.408 0.396 0.481 0.469 0.394 0.381 0.365 0.436 0.244Xalm 0.418 0.459 0.430 0.488 0.465 0.522 0.467 0.523 0.456 0.465 0.558 0.578 0.58 0.465 0.660Xgrs 0.020 0.037 0.024 0.029 0.028 0.036 0.112 0.072 0.048 0.046 0.021 0.022 0.035 0.089 0.046Xsps 0.005 0.015 0.013 0.012 0.007 0.010 0.013 0.010 0.015 0.019 0.027 0.020 0.021 0.009 0.050

Co— core of porphyroblast; R— rim of porphyroblast. Fe3+ calculated after charge balance based on 8 cations. XMg=Mg/(Mg+Fetot); XMg*=Mg/(Mg+Fe2+); Xpyr=Mg/(Mg+Fe2++Ca+Mn); Xalm=Fe2+/(Fe2+ + Mg+Ca+Mn); Xgrs=Ca/(Ca+Fe2++Mg+Mn); Xsps=Mn/(Mn+Fe2++Mg+Ca).


granulites, opx1 porphyroblasts show an increase in Al2O3 from abroad core relatively poor in alumina (3.5–4.8 wt.% Al2O3 in type C-fine; 4.7–5.5 wt.% Al2O3 in type D) to narrow rims rich in alumina(8.0–8.8 wt.% Al2O3 in type C-fine; 7.0–7.9 wt.% Al2O3 in type D),while XMg decreases more regularly rimwards (Fig. 9b). This zoningprobably reflects heating and prograde orthopyroxene growthoccurred during progressive melt extraction. Small opx blasts fromtype E granulites are unzoned and show a low Al2O3-content of 5.7–6.6 wt.% and XMg values of about 0.64. Opx porphyroblasts of type Fgranulites has high Al2O3-content of 8.13–10.15 wt.% but low XMg

values of 0.58–0.60. Opx in charnockites exhibits low Al2O3-content of4.3–4.7 wt.% and XMg of 0.51–0.53.

Symplectitic opx2 and opx3 in types A and B granulites displaysimilar XMg and slightly lower Al2O3 contents (4.9–6.7 wt.% Al2O3 intype A; 4.9–5.8 wt.% Al2O3 in type B) than porphyroblast rims. Opx2symplectites in type C-fine granulites have an intermediate compo-sition between porphyroblast core and rim (4.4–5.9 wt.% Al2O3),whilst opx2 of type D granulites are richer in alumina than opx1 rims(7.9–8.6 wt.% Al2O3). In general, Fe3+ represents less than 13% of Fetot,except for type C-coarse and type D granulites which have Fe3+≈17–20%. No systematic zoning of Fe3+ has been noticed.

6.3. Sapphirine

Prismatic sapphirine porphyroblasts in type A granulite display anintermediate (Fetot+Mg):Al2:Si composition between 2:2:1 and7:9:3 (Fig. 10 and Table 5) and an XMg value gently decreasingtowards the rim from 0.81–0.86 in the core to 0.76–0.81 at the rim.Symplectitic sapphirine shows a wider compositional range. Lamellarsapphirine symplectites in type B and E granulites are relatively rich inalumina and poor in silica respect to sapphirine porphyroblasts andhave a XMg≈0.73–0.82 and XMg≈0.67–0.68, respectively. Symplec-

titic sapphirine grains in type C-coarse granulites display a similar Si-content as symplectitic sapphirine in type B and E granulites but arepoorer in alumina, with an XMg≈0.70–0.75. Symplectitic sapphirinesin type D and F granulites exhibit a strong compositional variation andhave XMg≈0.71–0.73 and 0.67–0.69 respectively. All textural types ofanalysed sapphirine lie off the ideal line of Tschermak's substitution(Al2Si−1 Mg−1), suggesting a considerable incorporation of Fe3+ inthe structure. Similarly to opx, sapphirine in type C-coarse and type Dgranulites show the highest Fe3+ content. Estimates of ferric ironcontent after Higgins et al. (1979) yield for all analysed samples Fe3+/Fetot values of ~0.15–0.30.

6.4. Biotite

Biotite from all different types of granulites is highly magnesianwith XMg=0.69–0.82 and is rich in titanium (0.07–0.25 Ti p.f.u.calculated for 11 oxygens). Biotite in type A, B and D granulitesdisplays the highest XMg values of 0.76–0.82, reflecting the highestbulk-rock XMg, whereas biotite in charnockites is considerably lessmagnesian with XMg≈0.45–0.64 (Table 6). In general, biotite ingranulites, particularly in types C-fine, D and E, shows a negativecorrelation between XMg and Ti (Fig. 11).

6.5. Cordierite

Cordierite has a homogeneous, highly magnesian composition in allanalysed samples (Table 6)with XMg values comprised between 0.81 and0.90. No relevant compositional variations between different texturaltypes have been observed. The total sum of oxides is much lower than100%, suggesting the presence of H2O+CO2 in the structure.

Table 4Representative orthopyroxene analyses.

Sample BRE7 BRE10 GraG GraF CodGra5 PiaGra1 Conc13 PiaMig1

Texture Co R Sym1 Sym2 Co R Sym Co R Co R Sym Co R Sym Co Co Co

SiO2 50.650 50.530 52.520 51.090 50.980 52.040 51.860 49.010 49.600 52.090 49.300 51.010 50.710 49.130 48.590 49.240 46.580 48.540TiO2 0.001 0.068 0.015 0.035 0.092 0.037 0.040 0.139 0.088 0.117 0.153 0.074 0.049 0.026 0.060 0.049 0.178 0.090Cr2O3 0.000 0.012 0.000 0.029 0.000 0.033 0.000 0.000 0.000 0.000 0.050 0.140 0.000 0.000 0.000 0.001 0.000 0.000Al2O3 8.720 6.490 5.350 5.870 8.820 5.170 5.350 9.190 5.800 4.150 8.110 5.040 4.910 7.220 8.250 6.570 9.830 4.530FeO 12.070 17.640 16.390 17.930 10.990 17.220 17.270 16.130 20.100 19.400 19.650 19.970 17.890 19.190 19.330 21.650 23.460 28.620Mno 0.118 0.222 0.145 0.186 0.068 0.181 0.150 0.076 0.099 0.097 0.107 0.119 0.286 0.351 0.295 0.274 0.255 0.922MgO 27.860 25.130 26.040 24.840 28.730 25.850 25.520 25.070 23.630 24.220 22.610 23.270 25.420 23.670 23.300 21.630 19.550 17.120CaO 0.074 0.053 0.054 0.082 0.135 0.027 0.059 0.119 0.067 0.133 0.073 0.076 0.124 0.082 0.101 0.059 0.066 0.153Na2O 0.000 0.000 0.012 0.009 0.005 0.008 0.020 0.000 0.000 0.000 0.000 0.001 0.017 0.010 0.009 0.131 0.000 0.000K2O 0.014 0.033 0.006 0.005 0.009 0.001 0.000 0.000 0.004 0.003 0.012 0.000 0.000 0.016 0.009 0.269 0.002 0.015Tot 99.507 100.177 100.532 100.075 99.828 100.568 100.269 99.734 99.388 100.210 100.065 99.969 99.406 99.693 99.944 99.873 99.921 99.990O 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6Si 1.797 1.825 1.883 1.852 1.794 1.870 1.870 1.766 1.825 1.901 1.804 1.877 1.848 1.796 1.773 1.819 1.736 1.860Ti 0.000 0.002 0.000 0.001 0.002 0.001 0.001 0.004 0.002 0.003 0.004 0.002 0.001 0.001 0.002 0.001 0.005 0.003Cr 0.000 0.000 0.000 0.001 0.000 0.001 0.000 0.000 0.000 0.000 0.001 0.004 0.000 0.000 0.000 0.000 0.000 0.000Al(T) 0.203 0.175 0.117 0.148 0.206 0.130 0.130 0.234 0.175 0.099 0.196 0.123 0.152 0.204 0.227 0.181 0.264 0.140Al(M) 0.162 0.102 0.109 0.103 0.160 0.089 0.098 0.157 0.077 0.079 0.154 0.096 0.058 0.107 0.128 0.106 0.167 0.065Fe2+ 0.316 0.462 0.483 0.500 0.281 0.479 0.490 0.417 0.526 0.578 0.568 0.596 0.452 0.490 0.492 0.574 0.644 0.847Fe3+ 0.042 0.071 0.008 0.043 0.042 0.039 0.031 0.070 0.093 0.014 0.033 0.018 0.093 0.096 0.097 0.094 0.087 0.070Mn 0.004 0.007 0.004 0.006 0.002 0.006 0.005 0.002 0.003 0.003 0.003 0.004 0.009 0.011 0.009 0.009 0.008 0.030Mg 1.473 1.353 1.392 1.342 1.507 1.384 1.372 1.347 1.296 1.317 1.233 1.276 1.380 1.290 1.267 1.191 1.086 0.978Ca 0.003 0.002 0.002 0.003 0.005 0.001 0.002 0.005 0.003 0.005 0.003 0.003 0.005 0.003 0.004 0.002 0.003 0.006Na 0.000 0.000 0.001 0.001 0.000 0.001 0.001 0.000 0.000 0.000 0.000 0.000 0.001 0.001 0.001 0.009 0.000 0.000K 0.001 0.002 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.000 0.000 0.001 0.000 0.013 0.000 0.001Sum 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4XMg 0.804 0.717 0.739 0.712 0.823 0.728 0.725 0.735 0.677 0.690 0.672 0.675 0.717 0.687 0.682 0.640 0.598 0.516XMg* 0.823 0.745 0.742 0.728 0.843 0.743 0.737 0.764 0.712 0.695 0.685 0.682 0.753 0.725 0.720 0.675 0.628 0.536y(Al) 0.182 0.138 0.113 0.125 0.183 0.109 0.114 0.195 0.126 0.089 0.175 0.109 0.105 0.156 0.177 0.143 0.216 0.102A/AFM 0.166 0.128 0.107 0.117 0.167 0.103 0.107 0.176 0.116 0.086 0.160 0.104 0.103 0.149 0.168 0.133 0.192 0.097Fe3+/Fetot 0.142 0.132 0.017 0.079 0.131 0.075 0.059 0.143 0.150 0.023 0.056 0.030 0.170 0.164 0.165 0.141 0.119 0.076

Co — core of porphyroblast; R — rim of porphyroblast; Sym — symplectites. Fe3+ calculated after charge balance based on 4 cations. XMg=Mg/(Mg+Fetot); XMg*=Xen=Mg/(Mg+Fe2+); y(Al)=(Altot/2).

28A.G

alliet

al./Lithos

124(2011)

17–45

Table 5Representative sapphirine and spinel analyses.

Sapphirine Spinel

Sample BRE7 BRE10 GraG CodGra5 PiaGra1 Conc13 BRE7 BRE10 GraG CodGra5 PiaGra1 Conc13

Texture Co R Sym Sym Sym Sym Sym In In Sym In Sym Sym Sym Sym

SiO2 14.020 13.810 13.100 12.570 13.260 12.750 13.690 0.033 0.170 0.195 0.057 0.137 0.020 0.047 0.000TiO2 0.019 0.050 0.062 0.033 0.020 0.006 0.014 0.000 0.032 0.018 0.019 0.020 0.011 0.012 0.046Cr2O3 0.025 0.000 0.000 0.029 0.012 0.008 0.062 0.002 0.009 0.005 0.000 0.038 0.163 0.047 0.051Al2O3 61.490 60.840 61.310 61.830 59.420 60.080 58.010 62.700 60.690 61.400 58.400 60.840 59.310 60.010 58.450FeO 6.140 8.230 8.360 10.020 10.810 11.740 12.640 23.340 26.670 26.580 32.710 26.250 28.860 32.210 31.610Mno 0.041 0.050 0.038 0.021 0.130 0.133 0.031 0.078 0.074 0.090 0.084 0.034 0.122 0.135 0.022MgO 18.120 17.120 16.600 15.150 15.810 14.210 14.850 13.930 12.410 11.810 8.410 11.710 10.240 7.440 9.190CaO 0.018 0.007 0.008 0.013 0.048 0.057 0.021 0.000 0.000 0.004 0.017 0.000 0.000 0.003 0.095Na2O 0.002 0.016 0.000 0.011 0.000 0.517 0.010 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000K2O 0.003 0.000 0.009 0.000 0.005 0.167 0.001 0.005 0.001 0.000 0.000 0.000 0.000 0.016 0.014Tot 99.877 100.124 99.487 99.677 99.514 99.668 99.328 100.088 100.055 100.102 99.698 99.028 100.278 99.920 99.478O 10 10 10 10 10 10 10 4 4 4 4 4 4 4 4Si 0.827 0.819 0.782 0.755 0.800 0.770 0.834 0.001 0.005 0.005 0.002 0.004 0.001 0.001 0.000Ti 0.001 0.002 0.003 0.001 0.001 0.000 0.001 0.000 0.001 0.000 0.000 0.000 0.000 0.000 0.001Cr 0.001 0.000 0.000 0.001 0.001 0.000 0.003 0.000 0.000 0.000 0.000 0.001 0.004 0.001 0.001Al 4.273 4.253 4.317 4.379 4.223 4.275 2.166 1.940 1.906 1.930 1.897 5.996 1.914 1.947 1.892Fe2+ 0.230 0.299 0.299 0.394 0.368 0.407 2.000 0.454 0.511 0.534 0.654 0.533 0.580 0.692 0.620Fe3+ 0.073 0.109 0.119 0.110 0.177 0.185 0.481 0.058 0.084 0.059 0.099 0.059 0.081 0.049 0.105Mn 0.002 0.003 0.002 0.001 0.007 0.007 0.163 0.002 0.002 0.002 0.002 0.001 0.003 0.003 0.001Mg 1.592 1.513 1.478 1.357 1.421 1.278 0.002 0.545 0.493 0.469 0.345 0.470 0.418 0.305 0.376Ca 0.001 0.000 0.001 0.001 0.003 0.004 1.348 0.000 0.000 0.000 0.001 0.000 0.000 0.000 0.003Na 0.000 0.002 0.000 0.001 0.000 0.060 0.001 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000K 0.000 0.000 0.001 0.000 0.000 0.013 0.001 0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.000Sum 7 7 7 7 7 7 7 3 3 3 3 3 3 3 3XMg 0.840 0.788 0.780 0.729 0.723 0.683 0.677 0.515 0.453 0.442 0.314 0.443 0.387 0.292 0.341XMg* 0.874 0.835 0.832 0.775 0.794 0.758 0.737 0.546 0.491 0.468 0.346 0.469 0.419 0.306 0.377Fe3+/Fetot 0.240 0.268 0.284 0.218 0.325 0.313 0.253 0.114 0.141 0.099 0.132 0.099 0.122 0.067 0.132A/AFM 0.701 0.701 0.708 0.714 0.702 0.717 0.695

Co— cores; R— rims; Sym— symplectites; In— inclusions in garnet porphyroblasts; XMg=Mg/(Mg+Fetot); XMg*=Mg/(Mg+Fe2+); A/AFM=Al/(Al+Fe2++Mg); Fe3+ calculatedin sapphirine after charge balance based on 7 cations following the approach of Higgins et al. (1979) and in spinel after charge balance based on 3 cations.

Table 6Representative biotite, cordierite and plagioclase analyses.

Biotite Cordierite Plagioclase

Sample BRE7 BRE10 GraG GraF CodGra5

PiaMig1

Conc13 PiaMig1 BRE7 BRE10 GraG GraF CodGra5

GraF CodGra5

PiaMig1

Conc13 PiaMig1

Texture M M M M M M M M C C C C C M M M M M

SiO2 38.060 38.290 36.600 37.770 38.090 38.17 38.040 36.140 49.473 49.561 48.840 49.730 49.013 51.410 57.618 63.024 56.440 60.943TiO2 3.430 3.450 3.290 4.460 2.140 2.57 2.260 4.140 0.001 0.004 0.006 0.016 0.009 0.015 0.009 0.012 0.033 0.015Cr2O3 0.000 0.000 0.000 0.048 0.052 0.0056 0.000 0.017 0.008 0.004 0.010 0.003 0.018 0.004 0.008 0.002 0.000 0.003Al2O3 17.150 16.530 17.140 15.750 16.550 16.15 18.420 15.100 33.870 33.360 32.334 33.300 33.199 30.129 25.918 22.522 27.340 23.877FeO 9.290 9.860 11.860 11.020 9.980 10.75 11.470 18.880 2.889 2.934 3.693 3.298 3.147 0.069 0.120 0.032 0.062 0.075Mno 0.013 0.008 0.005 0.013 0.050 0.0593 0.004 0.211 0.029 0.034 0.018 0.022 0.061 0.005 0.007 0.005 0.000 0.008MgO 18.330 18.150 16.490 16.890 18.710 17.79 15.980 11.850 11.968 11.862 11.515 11.686 11.817 0.002 0.004 0.004 0.000 0.002CaO 0.006 0.000 0.000 0.028 0.000 0.0115 0.007 0.003 0.013 0.018 0.013 0.022 0.020 13.429 8.525 4.018 9.500 5.541Na2O 0.336 0.221 0.266 0.2293 0.135 0.0985 0.170 0.0949 0.099 0.085 0.085 0.072 0.060 3.762 6.304 8.508 5.950 7.617K2O 9.190 9.39 9.39 9.25 9.540 9.59 9.500 9.32 0.007 0.009 0.007 0.016 0.003 0.082 0.196 0.330 0.155 0.288Tot 95.806 95.899 95.313 95.458 95.247 95.195 95.581 95.756 98.358 97.870 96.520 98.165 97.347 98.908 98.707 98.457 99.480 98.369O 11 11 11 11 11 11 11 11 18 18 18 18 18 8 8 8 8 8Si 2.489 2.511 2.453 2.506 2.519 2.535 2.508 2.494 4.968 5.001 5.017 5.010 4.980 2.363 2.621 2.843 2.550 2.764Ti 0.169 0.170 0.166 0.223 0.106 0.128 0.112 0.215 0.000 0.000 0.000 0.001 0.001 0.001 0.000 0.000 0.001 0.001Cr 0.000 0.000 0.000 0.003 0.003 0.000 0.000 0.001 0.001 0.000 0.001 0.000 0.001 0.000 0.000 0.000 0.000 0.000Al 1.322 1.278 1.354 1.232 1.290 1.265 1.431 1.228 4.009 3.968 3.915 3.955 3.977 1.632 1.390 1.198 1.456 1.276Fetot 0.508 0.000 0.665 0.000 0.552 0.597 0.632 1.090 0.243 0.248 0.317 0.278 0.267 0.003 0.005 0.001 0.002 0.003Mn 0.001 0.541 0.000 0.611 0.003 0.003 0.000 0.012 0.003 0.003 0.002 0.002 0.005 0.000 0.000 0.000 0.000 0.000Mg 1.787 0.000 1.647 0.001 1.844 1.761 1.570 1.219 1.791 1.784 1.763 1.755 1.790 0.000 0.000 0.000 0.000 0.000Ca 0.000 1.774 0.000 1.670 0.000 0.001 0.001 0.000 0.001 0.002 0.001 0.002 0.002 0.661 0.416 0.194 0.460 0.269Na 0.043 0.000 0.035 0.002 0.017 0.013 0.022 0.013 0.019 0.017 0.017 0.014 0.012 0.335 0.556 0.744 0.521 0.670K 0.767 0.028 0.803 0.029 0.805 0.813 0.799 0.821 0.001 0.001 0.001 0.002 0.000 0.005 0.011 0.019 0.009 0.017Sum 7.086 0.785 7.123 0.783 7.139 7.116 7.075 7.093 11.037 11.024 11.034 11.019 11.036 5 5 5 5 5XMg 0.779 0.766 0.713 0.732 0.770 0.747 0.713 0.528 0.881 0.878 0.847 0.863 0.870XAb 0.335 0.566 0.778 0.526 0.701XAn 0.660 0.423 0.202 0.465 0.282XOr 0.005 0.012 0.020 0.009 0.017

M — matrix; C — corona; XMg=Mg/(Mg+Fetot); XAb=Na/(Na+Ca+K); XAn=Ca/(Ca+Na+K); XOr=K/(K+Na+Ca).


Fig. 8. (a) Compositions of garnet porphyroblast cores; (b) chemical zoning pattern of garnet porphyroblast in sample GraF (type C-fine granulite).


6.6. Spinel

Spinel is a hercynite–spinel solid solution (Table 5) with very lowCr2O3 contents (b0.10 wt.%). Spinel inclusions in garnet and prismaticsapphirine in type A granulites yield high XMg values of 0.45–0.53.Symplectitic spinel mantling sillimanite or spinels forming lobatesymplectites are generally less magnesian with XMg values from 0.26to 0.45. Spinel inclusions are richer in Fe3+ (Fe3+/Fetot≈11–25%)than symplectitic spinels (Fe3+/Fetot≈5–14%).

6.7. Plagioclase

Plagioclase compositions vary strongly with granulite type(Table 6). Matrix plagioclase of type C-fine granulites has the highestXAn of 0.57–0.76. Plagioclase in the matrix of type F granulites hasXAn=0.44–0.57, matrix plagioclase of type D granulites have

XAn=0.38–0.44, whereas matrix plagioclase of type E granulitesand charnockites have XAn=0.12–0.30. Within individual samples,plagioclase grains in symplectites with sapphirine±spinel in type Dand E granulites are slightly more anorthitic than matrix plagioclase,displaying a XAn=0.41–0.55 and XAn=0.25–0.34, respectively.

7. Textural evolution

Mg–Al-rich granulites preserve reaction textures such as coronaeand symplectites that allow deciphering their metamorphic history.Textures achieved during the prograde evolution are rare and usuallyoverprinted by metamorphic peak or post-peak assemblages. Morefrequently, high-grade rocks display reaction textures developed atlower grade conditions. This may help to constrain part of theirmetamorphic evolution, for example allowing post-peak isothermaldecompression to be distinguished from isobaric cooling (Harley,

image of Fig.�8

Fig. 9. (a) Composition of different textural types of orthopyroxene in Al2O3 vs. XMg space; (b) XMg zoning in orthopyroxene porphyroblast in sample GraF (type C-fine granulite);c) Al2O3 zoning in orthopyroxene porphyroblast in sample GraF (type C-fine granulite).


1989), so that a high-grade terrain can be assigned to a specifictectonic framework (Ellis, 1987; Bohlen, 1991; Platt et al., 1998).

7.1. Prograde evolution

7.1.1. Earliest recognizable mineral assemblageInformation on the early metamorphic history of the Gruf

granulites is essentially provided only by inclusions in the cores ofporphyroblasts in type A and C-coarse granulites. Staurolite inclusionsin garnet cores of both types of granulite, together with quartz,

plagioclase, biotite and alumosilicate inclusions, represent parts of theearliest preserved mineral assemblage. The reaction

St = Grt + Spl + Ky= Sil + H2O ð1Þ

was experimentally calibrated by Richardson (1968) andmay explain thewidespread occurrence of spinel and composite staurolite–spinel inclu-sions in garnet porphyroblasts, probably formed during granulitemetamorphism of earlier amphibolite facies staurolite–quartzassemblages.

image of Fig.�9

Fig. 10. Composition of different textural types of sapphirine in the Al vs. Si diagram (a.p.f.u. for 10 oxygens). 7:9:3 and 2:2:1 in the figure imply representative compositions ofMg1.75Al0.45Si0.75O10 and Mg2Al4SiO10, respectively.


Prismatic sillimanite in type B and type C-coarse granulites (Figs. 5eand 6c) resembles in both shape and internal structure the sillimanitepseudomorphs after kyanite described by Lal et al. (1984), Raith et al.(1997), and Tong and Wilson (2006). We thus propose that kyanite wasthe stable Al2SiO5 polymorph during the earlier metamorphic history.

7.1.2. Heating, partial melting and peak assemblagesWe suggest that during prograde heating the early mineral

assemblage was involved in a series of fluid-absent biotite meltingreactions affecting all investigated types of granulite. Melting most

Fig. 11. Composition of bio

likely persisted until the thermal peak was attained. Melting is heldresponsible for the partial or complete consumption of quartz and/orbiotite from the rock matrix and for the production of garnet andorthopyroxene porphyroblasts. The macroscopically observable segre-gation of melt led to the development of orthopyroxene-bearingleucosome (charnockite, although we note that this almost certainlydoes not record any single pristine extracted melt composition) andresidual melanosome (granulite) domains. Experimentally calibratedmelting reactions in the meta-aluminous system are:

Qtz + Bt = Opx + Kfs + melt ð2Þ

tite in Ti vs. XMg space.

image of Fig.�10

image of Fig.�11


Qtz + Bt = Grt + Kfs + melt ð3Þ

Qtz + Bt = Grt + Opx + Kfs + melt ð4Þ

Qtz + Bt + Pl = Grt + Opx + Kfs + melt ð5Þ

(Vielzeuf and Holloway, 1988; Vielzeuf and Montel, 1994; Clemenset al., 1997), whereas possible melting reactions in more aluminoussystems are:

Bt + Sil + Qtz = Spr + Ksp + melt ð6Þ

and

Bt + Sil + Spl = Opx + Spr + Ksp + melt ð7Þ

(Hensen and Harley, 1990). Reactions (2) to (5) explain the commonoccurrence of biotite and quartz inclusions in orthopyroxene andgarnet in type C-fine, E, and F granulites and charnockite, as well asthe orthopyroxene coronae separating quartz from biotite in type C-fine granulites (Fig. 6e). Reactions (6) and (7) led to the formation ofsapphirine porphyroblasts in the rock matrix of type A, D and Fgranulites, and explain the occurrence of spinel (in type A granulite)and rare sillimanite (in type A and F granulites) inclusions in prismaticsapphirine porphyroblasts (Fig. 5c). Reaction (6) would also accountfor the rare occurrence of composite sapphirine–quartz inclusions inthe rims of garnet porphyroblasts in type A and C-coarse granulites.During the melting event, early spinel was consumed from the matrixof type A granulites and is preserved exclusively as inclusions inprismatic sapphirine (Fig. 5c) and in garnet porphyroblasts. Compos-ite inclusions of spinel–corundum and spinel–sapphirine–corundumin garnets from type A granulites result from the completed reaction

Spl + Sill = Grt + Crn ð8Þ

(which exhausted sillimanite), and from the equilibrium

Grt + Spl + Crn = Spr: ð9Þ

Reaction (9) also explains the prograde consumption of corundumfrom the matrix of type E granulites, where resorption of corundumand garnet by sapphirine, which is typically attributed to heating(Kelsey et al., 2005), led to the formation of the observed corundum–

sapphirine–spinel symplectites (Fig. 7c).

7.2. Post-peak evolution

The post-peak evolution is characterised by the development ofcomplex symplectitic textures. As proposed by Fyfe (1973), Powell(1983), Waters (1988) andWhite and Powell (2002), we suggest thatsubstantial melt and H2O loss may have precluded the completion ofretrograde reactions, preserving parts of the peak mineralassemblages.

7.2.1. Textures involving garnetIn type A–D granulites, garnet porphyroblasts generally display

polymineralic symplectitic coronae (Figs. 5a,b, 6e,f and 7a). Asproposed by Droop and Bucher-Nurminen (1984), the reaction

Grt + Kfs + H2O = Opx + Crd + Bt ð10Þ

accounts for both the observed opx–cordierite symplectites betweengarnet and orthopyroxene (e.g. Fig. 5b) and the biotite coronaearound garnet. In the quartz-bearing type C-fine granulites (Fig. 6eand f) the same texture could also be formed by the reaction

Grt + Qtz = Opxsym + Crd ð11Þ

(experimentally calibrated by Hensen and Green, 1970 and Schreyerand Abraham, 1978). However, it is unlikely that reaction (11)accounts for the formation of opx–cordierite symplectites in type A, Band D granulites, where quartz was probably exhausted during partialmelting.

We interpret the formation of composite symplectites of ortho-pyroxene, cordierite, sapphirine and spinel in type B and D granulites(Fig. 7a) as the result of reactions

Grt = Opx + Spr + Crd � Plð Þ ð12Þ

and

Grt = Opx + Spl + Crd � Plð Þ ð13Þ

(Hensen, 1971). Reaction (12) also explains the opx–sapphirineintergrowth pseudomorph after garnet, as observed in type Dgranulites (Fig. 7b). The same intergrowth is commonly substitutedby spinel and cordierite, probably via the reaction

Spr + Opx = Crd + Spl ð14Þ

which is rendered possible through the involvement of significantFe3+.

7.2.2. Textures involving sillimaniteIn type B and C-coarse granulites, prismatic sillimanite is typically

mantled by composite symplectitic coronae and moats of sapphirine–cordierite±spinel±biotite (Figs. 5e, 6c). The reactions

Sil + Opx = Spr + Crd ð15Þ

(experimentally investigated by Schreyer, 1970) and

Sil + Grt = Spr + Crd ð16Þ

(Hensen, 1971) account for the formation of sapphirine–cordieritecoronae, whereas the reaction

Spr = Spl + Crd ð17Þ

explains the growth of fine-grained spinel at the expenses ofsymplectitic sapphirine (Fig. 6c).

8. P–T estimates

8.1. Prograde conditions

The quantitative reconstruction of the prograde segment of a P–Tpath for granulite facies rocks is generally problematic because of i)overprinting of the earliest mineral assemblages by peak and/or post-peakmineral assemblages; ii) resetting of original phase compositionsduring the metamorphic evolution; and iii) modification of theoriginal bulk composition through melt loss.

Nevertheless, we reconstruct the early metamorphic history of theGruf granulites and charnockites by combining available experimen-tal data for pelitic and semi-pelitic compositions (Fig. 12). Metamor-phic evolution probably began in the stability field of kyanite, withincreasing temperature leading to the formation of sillimanitepseudomorphs after kyanite (Fig. 5e). At approximately 840 °C and9–10 kbar, reactions (8) and (9) occurred in granulite types A and C-coarse, producing corundum and sapphirine grains now preserved ascorundum and sapphirine–corundum–spinel inclusions in garnetporphyroblasts. At approximately 875 °C, the upper stability limit ofstaurolite was overstepped, consuming staurolite from the matrix viareaction (1) (Richardson, 1968). However, rare staurolite grains havebeen trapped as “relict” inclusions in garnet porphyroblasts of type A

Fig. 12. P–T diagram showing the locations of experimentally determined reactions occurred during prograde metamorphism. References for reactions in the text. Upper stability ofTi-bearing biotite after Stevens et al. (1997).


and C-coarse granulites. Fluid-absent biotite melting experiments onmagnesian pelites and metagreywackes topologically analysed in thesystem KFMASH (Vielzeuf and Holloway, 1988; Hensen and Harley,1990; Vielzeuf and Montel, 1994; Clemens et al., 1997) show thatfurther heating would progressively lead to reactions (2), (4) and (7)in all granulite types at temperatures to 900 °C, producing coarse-grained opx, garnet and sapphirine porphyroblasts. However, experi-ments in the system KFMASNaCaTi (Stevens et al., 1997) indicate thatTi will stabilize biotite to approximately 40–50 °C higher temperature.In accordance with the high-Ti character of the biotite observed in allof our granulite types (Fig. 11), reactions (2), (4) and (7) may thushave been completed to temperatures of 930–940 °C.

8.2. Metamorphic peak conditions

Three approaches have been applied and compared in order tocalculate metamorphic peak conditions: “conventional geothermo-barometry”, P–T pseudosections and contouring of P–T space for Al-in-orthopyroxene contents.

8.2.1. Conventional thermobarometryThe most robust available mineral composition thermobarometers

consider the Al content of orthopyroxene coexisting with garnet,because of the relative immobility of Al with respect to Fe–Mg (Harleyand Green, 1982; Aranovich and Berman, 1997). Resetting of mineralchemistry during cooling and/or retrograde metamorphism is a majorsource of uncertainty when applying conventional thermobarometryparticularly for those based on Fe–Mg exchange to high-grade rocks(Frost and Chacko, 1989). This has partly been overcome by employ-ing retrieval techniques (Fitzsimons and Harley, 1994; Pattison andBegin, 1994; Pattison et al., 2003) in which measured Fe–Mg mineralcompositions are corrected such that calculated temperaturesconverge in P–T space with those derived from the Al content.

We estimated metamorphic peak-conditions for granulites andcharnockites using the orthopyroxene–garnet thermobarometers ofboth Harley and Green (1982) and Aranovich and Berman (1997), andwe performed correction for Fe–Mg diffusion with the program RCLC–P (Pattison et al., 2003). For granulite types A, B, C-coarse and E, andfor charnockites, average core compositions of garnet and orthopyr-

oxene porphyroblasts were chosen for the calculation. Based onincreasing alumina contents towards the rim of orthopyroxeneporphyroblasts of type D granulites, we paired average garnet corecomposition with average orthopyroxene rim composition for thisrock type, inferred to represent equilibrated mineral compositions atthe peak of metamorphism. Due to the complex zoning displayed bygarnet and opx porphyroblasts in type C-fine granulite (Figs. 8b, 9band c) and because of the wide compositional range displayed byporphyroblasts in type F granulite (Figs. 8a and 9a), we did notinvestigate these rock types, for which it was impossible to pairmineral compositions representative of equilibrium at the peak ofmetamorphism. The content of Al on theM1 site of opxwas calculatedas y(Al)=AlM1=Altotal/2. In this estimate we did not consider ferriciron, following the approach of Harley and Green (1982), Fitzsimonsand Harley (1994) and Aranovich and Berman (1997). This approachprovides a maximum estimate for y(Al) and may lead to over-estimated peak temperature (Harley, 1989; Nandakumar and Harley,2000; Pattison et al., 2003). Typically, calculated temperatures willdecrease by approximately 20–50 °C when the effect of ferric iron onthe M1 site in orthopyroxene is taken into account (Harley, 2008).Peak-temperatures were calculated for pressures of 8, 9 and 10 kbar.This pressure range has been inferred from peak assemblagestabilities as deduced from pseudosections (Section 8.2.2). Mineralcompositions and results are presented in Table 7 and Fig. 13.

The investigated samples of type A, B, D and E granulites and thecharnockite yield similar temperatures of 900–920 °C at 8 kbar, 930–950 °C at 9 kbar and 960–980 °C at 10 kbar, using the Al-in-opxthermobarometer of Harley and Green (1982). Estimated temperaturesfor coarse-grained type C are approximately 30 °C higher at allconsidered pressures. The Aranovich and Berman (1997) calibrationyielded considerably lower temperatures than the Harley and Green(1982) calibration for all the investigated samples (Fig. 13). The formercalibration is strongly susceptible to Fe–Mg diffusion on cooling,generally yielding lower temperatures than the Aranovich and Bermanexpression (see discussion inHarley, 2008). Correction of the Aranovichand Berman (1997) thermometer with RCLC–P (Pattison et al., 2003)yielded results similar to those from the Harley and Green (1982)thermobarometer. Estimates for type A, B, D and E granulites correctedfor Fe–Mg diffusion consistently yielded temperatures of 910–925 °C at

image of Fig.�12

Table 7Tpeak estimates for granulites Types A–E and charnockites at P=8, 9, 10 kbar.

Sample Garnet Orthopyroxene HaG_(1982) AaB_(1997) RCLC-P

Xpyr Xgrs Xen y(Al) 8/9/10 kbar 8/9/10 kbar 8/9/10 kbar

BRE7 (type A) 0.544 0.030 0.807 0.183 917/950/982 °C 813/860/906 °C 908/936/963 °CBRE10 (type B) 0.521 0.024 0.788 0.183 920/952/984 °C 819/866/912 °C 924/951/963 °CGraG (type Cc) 0.448 0.033 0.750 0.195 954/987/1020 °C 818/864/911 °C 976/1002/1028 °CCodGra5 (type D) 0.526 0.049 0.735 0.153 916/947/978 °C 899/946/992 °C 918/944/970 °CPiaGra1 (type E) 0.405 0.021 0.676 0.136 898/927/956 °C 858/902/947 °C 910/935/959 °CPiaMig1 (charno) 0.305 0.047 0.544 0.103 906/933/961 °C 862/904/946 °C 951/972/993 °C

Abbreviations: HaG_(1982) — Harley and Green (1982); AaB_(1997) — Aranovich and Berman (1997). Uncertainties are±40 °C. Xpyr, Xgrs, Xen; y(Al) as in Tables 3 and 4.


8 kbar, 935–950 °C at 9 kbar and 960–970 °C at 10 kbar. With thediffusion correction, the calculated temperatures for charnockite andcoarse-grained type C granulites are about 20–30 °C and 50–60 °Chigher than for the other granulites, respectively (Fig. 13).

8.2.2. P–T pseudosections and Al-in-orthopyroxene contoursPseudosections were constructed to represent the stability fields of

coexisting minerals in P–T space for a specific bulk-rock compositionemploying an internally-consistent thermodynamic dataset (here the2004 update of Holland and Powell, 1998), and mineral solutionmodels as listed in the caption to Fig. 14. Key advantages of thisapproach include the ability to represent more complex chemicalsystems than FMAS (Hensen, 1971; Hensen and Green, 1973; Hensen,1987; Kelly and Harley, 2004) or KFMASH (Hensen and Harley, 1990;Srogi et al., 1993; Sengupta et al., 1999; McDade and Harley, 2001),not having to rely directly on measured mineral compositions thatmay have re-equilibrated during cooling, and the ability to contour formineral proportions and mineral chemistry. In particular, combiningcalculated contours of Al-in-opx with the observed mineral compo-sition and/or zoning trends can allow the metamorphic history of aspecific rock to be unravelled in more detail (Harley, 1998a; Kelsey etal., 2003; Baldwin et al., 2005).

Nevertheless, several important limitations for the use of pseudo-sections for high-grade rocks have to be considered. Major problemsarise from uncertainties in the thermodynamic properties of sapphi-rine (review in Harley, 2008), and particularly the lack of data

Fig. 13. Summary of conventional thermobarometric calculation at 8, 9 and 10 kbar wthermobarometer and correction for Fe–Mg back exchange with the program RCLC–P (Pattsample BRE10; C: type C-coarse, sample GraG; D: type D, sample CodGra5; E: type E, sampcalculated Harley and Green (1982) and RCLC–P (Pattison et al., 2003) data were plotted a

available to quantify the role of ferric iron on the stability of crucialhigh-temperature phases such as sapphirine, spinel and orthopyrox-ene (Annersten and Seifert, 1981; Hensen, 1986; Sandiford et al.,1987; Powell and Sandiford, 1988; Das et al., 2001;White et al., 2002).A further problem is the choice of an appropriate “reactive” bulk-composition for the problem to be treated (Stüwe, 1997). Despitethese limitations, pseudosections offer a complementary approach totraditional thermobarometry when investigating metamorphic pro-cesses in high-grade rocks.

For the bulk compositions of granulite types A (sample BRE7), B(BRE10), C-fine (GraF) and F (Conc13), we calculated P–T pseudosec-tions in Na2O–CaO–K2O–FeO–MgO–Al2O3–SiO2–H2O–TiO2–MnO(NCKFMASHTiMn), employing the 2007 version of Perplex (Connolly,2005). The inclusion of MnO was not strictly required for thesecalculations and can sometimes be detrimental (see discussion ofWhite et al., 2007), but was justified by the garnet-zoning modellingthat follows (Section 9.3) and probably has limited negativeconsequences for samples as garnet-rich as those examined here.Pseudosections were contoured for Al-content in orthopyroxene(using the Powell and Holland, 1999, pyroxene model, contoured inwt.% Al2O3 in Fig. 14a–d). We note here that because of the lack ofappropriate solution models accounting for Fe3+, we do not includeferric iron in these calculations. However, Fe3+ will enhance thestability of all phases that readily incorporate it (i.e. sapphirine,orthopyroxene and spinel), moving the Al-in-opx isopleths. Therefore,these results are likely to slightly overestimate peak P–T conditions.

ith the Harley and Green (1982) thermobarometer, Aranovich and Berman (1997)ison et al., 2003). Letters refer to type of granulite (A: type A, sample BRE7; B: type B,le PiaGra1; F: type F, sample Conc13; H: charnockite, sample PiaMig1). For clarity, thet slightly lower and higher pressures, respectively.

image of Fig.�13

Fig. 14. (a–d) P–T pseudosections for type A (BRE7), B (BRE10), C-fine (GraF) and F (Conc13) granulites calculated in the system NaCaMnTiKFMASH and contoured for wt.% Al2O3 inorthopyroxene. Solution-phase models used in the calculations are: garnet–Gt(WPH): White et al. (2007); biotite–TiBio(WPH): White et al. (2007); orthopyroxene–Opx(HP):Powell and Holland (1999); sapphirine–Sapp(KWP): Kelsey et al. (2004); cordierite–hCrd: Holland and Powell (1998); feldspars–feldspar: Furman and Lindsley (1988); spinel–Sp(HP): Holland and Powell (1998); osumilite–Osm(HP): Holland et al. (1996); ilmenite–IlGkPy: Andersen and Lindsley (1988); melt–melt(HP):White et al. (2007), names as definedin solute.dat of Perplex; stars represent conventional thermobarometric estimates corrected for Fe–Mg back exchange at 9 kbar (Pattison et al., 2003).


In the pseudosection calculated for sample BRE7 (type A granulites,Fig. 14a), the Bt–Melt–Spr–Ksp–Opx–Grt field corresponds to theobserved peak assemblage. The appearance of sapphirine and thesimultaneous disappearance of sillimanite are predicted to occur atTN915 °C, while the disappearance of orthopyroxene with increasingpressure limits the peak pressure to 9.5–12 kbar depending on T. Theobserved peak assemblage is predicted for a wide P–T range.Contouring the pseudosection for Al in orthopyroxene (measuredaverage: 8.8 wt.% Al2O3) allows us to restrict our P–T estimate to the

lower-T domain of this field (about 915–950 °C and 8–9.5 kbar). Theseconditions are in good agreement with conventional thermobaro-metric estimates corrected for Fe–Mg back-exchange, which yield≈936 °C at ≈9 kbar (Table 7).

The high sensitivity of silica undersaturated rocks to bulk silicacontent (Kelsey et al., 2005), is illustrated by comparing pseudosec-tions of type A and B granulites. A pseudosection for the SiO2-richerand FeO–MgO-poorer type B granulite (sample BRE10) predicts asubstantial increase in quartz and sillimanite stabilities, and a retreat

image of Fig.�14


of the sapphirine field towards higher temperatures (Fig. 14b). Thestability field corresponding to the observed peak-assemblage (Bt–Melt–Ksp–Pl–Sil–Opx–Grt) is limited to a narrow P–T range between900–945 °C and 8–10 kbar. This is supported by the Al content in opx(measured average 8.7 wt.% Al2O3) and is in broad agreement withconventional thermobarometric estimates (Fe3+-absent calculationsgiving 951 °C at 9 kbar, Table 7).

A pseudosection calculated for sample GraF (type C-fine granulite,Fig. 14c) predicts that the observed peak assemblage (Grt–Opx–Bt–Ksp–Pl–Qtz) is stable at 900–930 °C (depending on pressure). Athigher temperatures, quartz and subsequently biotite would becompletely consumed by fluid-absent biotite melting reactions (seereactions (2) to (5) in Section 7.1.2). The Al-in-opx isoplethcorresponding to the measured peak composition (~ 8.4 wt.% Al2O3)is located at slightly higher temperature than the stability of quartzand biotite, predicting peak conditions of 930–1020 °C at 8–10 kbar.This discrepancy could be resolved by expansion of the biotitestability field with addition of components such as F, Cl or Fe3+, whichare not considered here (Peterson et al., 1991; Hensen and Osanai,1994; Mouri et al., 1996), or with use of a newer biotite model whichpermits M2 site Ti and is shown to enhance high-T biotite stability(Tajčmanová et al., 2009). Conversely, the Al-in-opx isopleth is likelyto move towards slightly lower T and P if Fe3+ is included. Due to thelow Fe3+ content estimated in opx (~5% of Fetotal), however, wefavour the former explanation and suggest that type C-fine granuliteformed at approximately 950 °C and 8–9 kbar. This is in accordancewith conventional thermobarometry estimates for the coarser-grained type C-coarse granulites, which yield TN950 °C (Table 7).

For type-F granulites, the occurrence of the peak assemblage(Melt–Spr–Ksp–Pl–Grt–Opx–Qtz–Ru) in the appropriate pseudosec-tion (Fig. 14d) limits the minimum peak temperature to 915 °C atpressures between 7 and 10 kbar. Again, Al-in-opx isopleths (mea-sured average 9.0 wt.% Al2O3) help to refine the estimated conditionsto T≈915–940 °C and P≈8–9 kbar.

8.3. Post-peak conditions

During the post-peak history, the Gruf granulites developed aseries of symplectitic assemblages (Section 7.2). Decreased cationdiffusion rates at lower temperatures and/or loss of melt (White andPowell, 2002) lead to the development of micro-scale volumes ofequilibration in these rocks. Estimating the effective bulk compositioninvolved in any specific reaction corona/symplectite is problematicand in this case, identification of a single reaction using thepseudosection approach may be seriously misleading (Stüwe, 1997).

Effectively, pseudosections for type A, B, C-fine and F compositionspredict that the post-peak reactions forming cordierite at the expense

Table 8Calculation of metamorphic post-peak conditions for samples BRE10 (type B granulite), GraGusing the P–T average approach of Powell and Holland (1994).

Sample Post-peakassemblage

Garnet Opx Spr Spl Cr

x(g) z(g) x(opx) y(opx) Q(opx) x(sa) y(sa) x(sp) x(

BRE10 Grt, Opx, Spr,Spl, Crd, Bt, Sil

0.50 0.03 0.27 0.12 0.15 0.17 0.22 0.53 0.

GraG Grt, Opx, Spr,Spl, Crd, Bt, Sil

0.55 0.04 0.30 0.13 0.15 0.20 0.22 0.55 0.

GraF Grt, Opx, Crd,Bt, Qtz

0.57 0.07 0.32 0.11 0.15 – – – 0.

CodGra5 Grt, Opx, Spr,Spl, Crd, Bt, Pl

0.50 0.05 0.28 0.17 0.15 0.20 0.22 0.60 0.

Garnet: x(g)=Fe/(Fe+Mg); z(g)=Ca/(Ca+Fe+Mg); orthopyroxene: x(opx)=Fe/(Fe+Mg); yy(sa)=x(Al,M1); spinel: x(sp)=Fe/(Fe+Mg); cordierite: x(cd)=Fe/(Fe+Mg); h(cd)=H2O oplagioclase: ca(pl)=Ca/(Ca+Na+K); k(pl)=K/(K+Ca+Na). Used activity models: garnet andafter Kelsey et al. (2004); spinel and cordierite after Holland and Powell (1998); plagioclase after

of sillimanite and garnet should occur at≈900 °C and≈7–8 kbar.However, this is inconsistent with the measured Al-content oforthopyroxene porphyroblast rims (5–6 wt.%) and symplectitic opx(5–5.5 wt.%) in all of the considered samples (Table 4). Al-in-opxcontours in pseudosections (Fig. 14) suggest that the observed Al2O3

decrease (3–3.5wt.% relative to porphyroblast cores) corresponds to atemperature decrease of approximately 150 °C, compatible withisopleth spacing estimates of~40–50 °C per 1 wt.% Al2O3 (Aranovichand Berman, 1996; Harley and Motoyoshi, 2000).

As an alternative to the pseudosection approach, we calculated theP–T conditions of the most relevant reactions leading to formation ofcordierite-bearing post-peak coronae and symplectites in samplesBRE10 (type B granulite), GraG (type C-coarse), GraF (type C-fine)and CodGra5 (type D), using the measured compositions of theparticipating phases. Post-peak mineral assemblages occurring in eachsample and phase activities used in the calculations are given in Table 8.

Sample BRE10 displays the widest range of post-peak reactiontextures, involving the destabilisation of both sillimanite and garnet.For this sample, the intersection of the slightly T-sensitive cordieriteproducing reactions (12, 13, 15, 16) with the Grt–Opx thermo-barometer corrected for Fe–Mg back diffusion (calculated by pairingOpxsym with garnet rim compositions, using RCLC–P, Pattison et al.,2003) yields 720–750 °C and 6–7.5 kbar for symplectite formation(Fig. 15). The location of the spinel–cordierite forming reaction (17),although having a large uncertainty, is in agreement with this result.The location of reactions (11), (12) and (15), calculated respectivelyfor samples GraF, CodGra5 and GraG, also agree well with these post-peak conditions (Fig. 15).

P–T conditions for texturally-equilibrated post-peak assemblagesof samples BRE10, GraG and GraF were also calculated with the P–Taverage approach of Powell and Holland (1994). Results coincide withthe estimates from single reactions, yielding consistent conditions of720–740 °C and 6.5–7.5 kbar (Table 8). A P–T average calculation onsample CodGra5, although yielding lower temperatures of 684±68 °Cand higher pressures of 8.1±1.1 kbar (Table 8), is in error of thisrange. We note that decreased Al-contents in opx rims andsymplectites (Table 4) fit well with Al-in-opx compositions predictedby the pseudosections at these inferred conditions.

9. Discussion

9.1. UHT: age and significance

Our thermobarometry on charnockites and six granulite typesdistributed over the entire Gruf Complex shows that these rocksformed at temperatures in excess of 900 °C and pressures of 8.5–9.5 kbar. The high temperature attained by the Gruf charnockites and

(type C-coarse granulite), GraF (type C-fine granulite) and CodGra5 (type D granulite)

d Bt Pl P–T-estimate

cd) h(cd) x(bi) y(bi) Q(bi) ca(pl) k(pl) T (°C) P (kbar) Cor Fit

13 0.10 0.25 0.52 0.15 – – 721±59 7.3±0.5 0.61 3.48

15 0.10 0.28 0.55 0.15 – – 731±59 7.1±0.5 0.63 3.55

14 0.10 0.27 0.51 0.15 – – 745±51 7.7±0.5 −0.16 2.33

13 0.10 0.24 0.47 0.15 0.50 0.01 684±68 8.1±1.1 0.19 5.27

(opx)=x(Al,M1)=Al/2;Q(opx)=2 (x(Fe,M2)–x(opx)); sapphirine: x(sa)=Fe/(Fe+Mg);n hydroxil site; biotite: x(bi)=Fe/(Fe+Mg); y(bi)=x(Al,M1); Q(bi)=3 (x−x(Fe,M2));biotite afterWhite et al. (2007); orthopyroxene after Powell and Holland (1999); sapphirine(Holland and Powell, 2003).

Fig. 15. P–Tdiagramshowing the calculated locationsof observed reactions in coronae and symplectites of Gruf granulites (black lines—BRE10; dotted line—CodGra5; dotted–dashed line—GraG; dashed line—GraF). The star and the dashed ellipse represent P–T average estimate for sample BRE10 and corresponding uncertainty. The grey ellipse corresponds to the best estimatepost-peak conditions consistent with the central Alpine Lepontine metamorphism (dashed box, after Heitzmann, 1975; Schmidt, 1989; Engi et al., 1995; Stucki, 2001; Nagel et al., 2002 andBurri et al., 2005). Reactions are written with the high-T assemblage to the right of the “=” sign. Numbers refer to reaction numbers in the text.


granulites is unique evidence for UHT metamorphism in the GrufComplex and, more generally, in the Central Alps. Such high-gradeconditions within a growing orogen are uncommon and very difficultto integrate within the framework of the Alpine metamorphism aspreviously inferred by Droop and Bucher-Nurminen (1984), Liati andGebauer (2003) and Schmitz et al. (2009). It is likely that suchconditions can only be attained if mantle diapirism (coeval with orindependent of slab breakoff) advects the necessary heat, but such anevent should influence a region far larger than just the Gruf Complex(e.g. Schubert et al., 2001). An alternative explanation sees the Grufcharnockites and granulites as lower crustal relics of a pre-Alpinehigh-grade metamorphism. In this case, a UHT Alpine event isunnecessary and themost likely candidate for such high temperaturesis the post-Variscan Permian event responsible for both thewidespread formation of granulites and charnockites in the Europeancrust (Lorenz and Nicholls, 1976; Lorenz and Nicholls, 1984; Schusterand Stüwe, 2008) and the vast southern European rhyolitic province(Lorenz and Nicholls, 1984). Granulites related to the extensive riftingthat occurred in the Permian are exposed in several parts of the Alpsand record similar P–T conditions to the Gruf granulites. Metapeliticgranulites formed at≈850 °C and≈8 kbar are present in theMalencoUnit, 15 km east of the Gruf Complex (Müntener et al., 2000).Granulites of pelitic composition that occur in the Sondalo Complex(Braga et al., 2001), 50 km to the east-south-east of the Gruf Complex,yield a temperature of approximately 900 °C and pressure of 8 kbar(Braga et al., 2003). Permian granulites recording TN850 °C are alsoexposed in the Ivrea and Sesia Zones (e.g. Barboza and Bergantz, 2000;Lardeaux and Spalla, 1991; Rebay and Spalla, 2001), approximately 60and 100 kmwest-south-west of the Gruf, respectively (although thesetwo distances are reduced by 30 km after reintegration of post-Oligocene dextral displacement along the Periadriatic Fault System;Müller et al., 2001).

Several arguments indicate a pre-Alpine age for the granulitefacies metamorphism studied here: i) ultramafic, mafic and carbonaterocks occurring in the Gruf Complex and regarded as equivalent to theMesozoic Chiavenna Unit (Diethelm, 1989; Davidson et al., 1996;Schmid et al., 1996) show no sign of UHT-metamorphism; ii) intensefluid-absent biotite melting as observed for these granulites is absent

from the mid-Tertiary upper amphibolite-facies migmatites whichvolumetrically dominate the Gruf Complex (Fig. 2). These ratherdeveloped by fluid-saturated and also fluid-absent muscovite meltingreactions which formed widespread migmatites throughout theCentral Alps (Burri et al., 2005); iii) the 32–30 Ma granodioritic–tonalitic Bergell Intrusion (Von Blanckenburg, 1992) induced a localincrease in partial melting of proximal metapelitic rocks, which isinconsistent with host-rocks that were already at UHT conditions; iv)SHRIMP ages of zircons from the charnockites consistently yield aPermian intrusion age of 266–280 Ma (Galli et al., in review). Texturalevidence and inclusions within these crystals strongly imply granu-lite-facies mineralogy at this time.

We propose that a single, well defined population of 272.0±4.1 Ma ages from zircon cores from the type A and B Gruf granulites(Liati and Gebauer, 2003) corresponds to the age of the UHT event.This disagrees with the interpretation of Liati and Gebauer (2003),who attribute the 272 Ma age to crystallisation of a supposed graniteprecursor, the magmatic character being inferred from oscillatoryzoning of the inner portions of the zircons. However, the high Mg–Alcharacter of the granulites, the widespread occurrence of prismaticsillimanite (type B), and staurolite inclusions in garnet (types A and C-coarse) indicate a Mg–Al-rich metapelitic precursor protolith. Thisinterpretation is supported bymass balance calculations showing thatthe required amount of granitic melt extracted fromMg-rich pelites inorder to reach a residual composition corresponding to type A or Bgranulites would be 40–50%, in accordance with the amount of meltformed by first fluid-absent muscovite and then fluid-absent biotitemelting of Mg-rich metapelites at 920–940 °C and 9 kbar (Stevens etal., 1997). In the case of a granitic protolith, extraction ofN90% meltwould be required to generate a somewhat similar but stronglyperaluminous residual composition. If this melting was related to postnappe stacking Lepontine metamorphism, then it should have beenwidespread throughout the Gruf (and possibly other units), which isnot the case.

In our interpretation, Permian oscillatory zoning in zircon formedduring extensive partialmelting related to peak granulitemetamorphism,and Oligocene homogeneous zircon rims recrystallised during theAlpine, upper-amphibolite-facies Lepontine metamorphism. This

image of Fig.�15


interpretation is also in agreement with granulites constitutingschlieren and enclaves within the charnockites (Fig. 3b), suggestingthat pelitic rocks underwent intense partial melting (during thePermian UHT event), leading to the formation of the orthopyroxene-bearing charnockitic melts and granulitic residuals.

9.2. Post-peak evolution

Commonly, reaction textures such as the observed symplectitesand coronaewhich partly or completely replace garnet and sillimanitehave been used to infer rapid isothermal decompression (Hensen,1987; Harley, 1989; 1998a). In fact, the Gruf Complex has beenconsidered a classical high-grade terrain that experienced a strong(Alpine) isothermal decompression (ITD) along a steep retrograde P–T path (Droop and Bucher-Nurminen, 1984; Harley, 1989). Thisinterpretation arose due to estimated P–T conditions of 830 °C,10 kbar for peak conditions and 750 °C, 5 kbar for symplectiteformation (Droop and Bucher-Nurminen, 1984), implying a decom-pressional cooling path with a dP/dT gradient of 60 bar/°C (Fig. 16).

Our thermobarometric calculations, however, suggest that theserocks may have formed at temperatures in excess of 900 °C andpressures of 8.5–9.5 kbar (approximately 100 °C hotter than previ-ously thought). Calculations also suggest that the peak assemblagesubsequently partly re-equilibrated at 720–740 °C and 7–7.5 kbar(deeper than previous estimates). Thus, the inferred post-peak dP/dTgradient is~10 bar/°C, much less than previously suggested. Thisconsiderably lower dP/dT forces a reinterpretation of the tectono-metamorphic evolution and significance of the Gruf granulites in anAlpine context. Whereas granulite and charnockite peak-conditionsare incompatible with any other metamorphic conditions determinedfor Lepontinemetamorphism, post-peak conditions correspond to thegenerally accepted mid-Tertiary regional metamorphic conditions forthis part of the Lepontine Dome (Bucher-Nurminen and Droop, 1983;Burri et al., 2005) and to intrusion pressures of the adjacent Bergelltonalite (Reusser, 1987). This suggests three possible tectonicscenarios: i) granulites and charnockites formed at ultra-high

Fig. 16. Comparison between single metamorphic cycle model (Droop and Bucher-Nurmiellipses: estimates of peak and post-peak conditions after Droop and Bucher-Nurminen (19Alpine metamorphic cycle (Droop and Bucher-Nurminen, 1984); black bold ellipses: estimevolution for the Gruf granulites at 272 Ma, B2 — estimated metamorphic path for the Alpiestimates for Malenco, Sondalo, Ivrea and Sesia granulites (Müntener et al., 2000; Braga et al.dashed square: estimates of the central Alpine metamorphic conditions (Heitzmann, 1975;

temperature conditions during the Permian event and cooled slowlyover the following 250 Ma to 720–740 °C (b1 °C Myr−1); ii) granu-lites and charnockites were formed by Permian UHT metamorphism,were rapidly exhumed and cooled, and were not affected by thesubsequent Lepontine metamorphism; iii) granulites and charnock-ites experienced UHT metamorphic conditions during the Permianand were subsequently overprinted by the Alpine Lepontine event.Thus, peak- and post-peak conditions may not represent progressivestages of the same Alpine metamorphic cycle, reflecting instead twodifferent and tectonically unrelated metamorphic events (Fig. 16).

9.3. Duration of UHT metamorphism and subsequent cooling: insightfrom garnet zoning

In order to approximately constrain the duration of the thermalevent and test whether any of the three scenarios outlined above canbe excluded, we investigated major element zoning in garnetporphyroblasts, using techniques similar to several recent works(e.g. Ague and Baxter, 2007; Dachs and Proyer, 2002; Lasaga and Jiang,1995; Storm and Spear, 2005).

9.3.1. Calculation methodsGarnet crystals from the Gruf Complex typically record “prograde”

zoning patterns (Droop and Bucher-Nurminen, 1984) but do notcontain the distinct steps in zoning patterns which have been used toinfer pre-diffusion compositional zoning (e.g. Ague and Baxter, 2007).Because this information is required to forward model the intra-crystalline diffusive history, we use an alternative technique,described in more detail in Caddick et al. (in press) and Caddick andThompson (2008). Briefly, this approach follows suggestions (e.g.Konrad-Schmolke et al., 2005; Konrad-Schmolke et al., 2007; Spear,1988; Spear et al., 1991) that equilibrium thermodynamic calculationscan provide an initial growth zoning for any given P–T path, whichthen can be used to assess the extent of diffusive resettingexperienced.

nen, 1984) and two metamorphic cycles model for the Gruf granulites. Light greyish84); light greyish dotted line (A): isothermal decompression (ITD) path during a singleates of peak and post-peak conditions from this study; B1 — Permian metamorphic

ne regional metamorphism at 32 Ma; black ellipses: Permian metamorphic conditions, 2003; Barboza and Bergantz, 2000; Lardeaux and Spalla, 1991; Rebay and Spalla, 2001);Schmidt, 1989; Engi et al., 1995; Stucki, 2001; Nagel et al., 2002 and Burri et al., 2005).

image of Fig.�16


Preserved garnet zoning (e.g. Fig. 8) and the inferred prograde P–Thistory of Gruf Complex samples (Figs. 12 and 14) suggest that garnetgrowth began significantly before peak temperature, requiring bothassumption of the early prograde P–T path and sensitive treatment ofbulk-rock composition. We assume here a simple path involvingheating during burial to 11 kbar, then decompression to 8.5 kbarduring further heating to 950 °C (path a–b, Fig. 17a). Following peak-T(point b of Fig. 17a), two types of path continuation were tested;either cooling directly through 720 °C to 400 °C (dashed black curve

Fig. 17. Modelled garnet zoning profiles formed during various P–T–t histories. (a) P–T pathstage cooling path (dashed black curve b–c–d) or cooling followed by later reheating then ediameter spherical crystal along P–T path a–b–c–d, without considering intra-crystalline dconsidering diffusion and with 250 Myr cooling from UHT to “Lepontine conditions” and 20lines represent microprobe traverses through natural crystals (see also Fig. 8) (d–g) Profiles240 Myr at either 450 °C, 550 °C, 600 °C or 650 °C, and then a 30 Myr “Lepontine event”. Bonatural crystals.

b–c–d, Fig. 17a) or cooling first to some lower temperature(represented by e on Fig. 17a although various temperatures weretested) then heating again to inferred Alpine conditions (solid bluecurve b–e–c–d, Fig. 17a). Garnet composition and modal proportionwere calculated at approximately 200 points along each P–T path withPerple_X (dataset and mineral solution models as described inSection 8.2.2). Calculations used a starting bulk-rock compositioncalculated by mixing the compositions of samples GraF and PiaMig1(Table 2) in the ratio 60% to 40%, respectively, to simulate an

types modelled, with a common prograde history (black curve a–b) then either a onexhumation (blue curve b–e–c–d). (b) Core to rim zoning profile established in a 3 mmiffusion. (c) Profile established by a one-stage heating-cooling path (as panel a), butMyr cooling from “Lepontine conditions”. Bold curves are model outputs, fine dashedestablished by two-phase P–T paths, with rapid early UHT metamorphism followed byld curves are model outputs, fine dashed lines represent microprobe traverses through

image of Fig.�17


estimated pre-melt-loss composition. The composition of any meltformed along the P–T path was progressively removed from the“residual bulk-rock composition”, which provides an additional checkof calculation consistency; the calculated “residual bulk-rock compo-sition” tending to that of sample GraF as the calculation progressesand melt is extracted. The final model composition agrees relativelywell with that of sample GraF, albeit with higher MgO and FeOcontents (final column, Table 2). All of the melt produced at each P–Tstep was removed from the residual composition, so the final restitecomposition is almost completely devoid of Na2O.

Garnet modal proportion along the P–T path was converted tocrystal radius assuming a spherical geometry. A final crystal diameterof 3 mmwas prescribed and it was assumed that crystals nucleated atthe garnet-in reaction (i.e. without reaction overstepping). Garnetcompositions were recorded along retrograde paths (to indicate theorientation of late-stage resetting), but crystal resorption was notpermitted because detailed simulation of symplectite formation isoutside the scope of the simple 1-dimensional approach applied here.This undoubtedly results in uncertainty in modelled crystal rimcompositions, but as shown below permits extraction of theapproximate timescale of the UHT event. A model time was assignedto each of the P–T points at which phase abundance and compositionwere calculated, yielding a series of time-steps. In each time-step themodel crystal was allowed to instantaneously grow or change its rimcomposition (as predicted by the Perple_X calculations), and diffusivere-equilibration of the zoning profile established by previous stepswas calculated. Composition-dependent Fe, Mg, Ca and Mn diffusiv-ities of Carlson (2006) were used with the Lasaga (1979) formulationfor multi-component diffusion in ionic lattices. Diffusivity wascalculated at each time-step (i.e. for each P–T condition) and at eachof the 100 recorded model positions (i.e. at each crystal composition).

9.3.2. ResultsIn the absence of intra-crystalline diffusion, P–T path a–b–c–d

(Fig. 17a) results in a garnet zoning profile showing decreasing XMn

and XCa, and increasing XMg from core to rim (Fig. 17b). XFe initiallyincreases rimwards, then decreases and finally flattens towards thecrystal rim. Permitting intra-crystalline diffusion throughout the sameP–T history allows us to model the effects of a single long period ofcooling fromUHT to “Lepontine conditions” (e.g. Fig. 17c). Herewe re-model path a–b–c–d, setting the duration of b–c to 250 Myr (an initialcooling rate of ca. 1 °C Myr−1) and the duration of c–d to 20 Myr(consistent with Alpine exhumation rates of Rubatto et al., 2009). Theduration of prograde phase a–b is not known, but multiple testedpossibilities yield effectively the same result because the subsequentlong period of cooling from UHT conditions removes all evidence ofearlier zoning (e.g. Fig. 17c). In particular, increases in XMg anddecreases in XFe and XMn measured from core to near-rim in naturalcrystals (Fig. 8) are poorly represented (and indeed reversed) by themodel result (Fig. 17c). The model result also has very high XMg andlow XCa in the crystal core because of the long duration spent at high-T, where equilibrium (i.e. crystal rim) compositions differ significantlyfrom those experienced earlier in the growth history and intra-crystalline diffusivities are sufficiently high to modify the crystal corecomposition. We emphasise here that sluggish post high-T interactionbetween the crystal of interest and the matrix (due to slow inter-granular diffusivity or physical shielding by growth of a new rim) willreduce this significant modification of modelled crystal bulk compo-sition. For example, complete closed-system behaviour followingpeak-T would result in fixed total Fe, Mg, Ca and Mn contents in eachcrystal, but additional calculations (not shown) suggest that a250 Myr cooling phase (b–c) would simply yield entirely unzonedcrystals, preserving none of the prograde zoning seen in naturalcrystals (Fig. 8 and Droop and Bucher-Nurminen, 1984).

An alternative P–T history involves rapid cooling from peak-T,followed by Alpine reheating and subsequent exhumation (path a–b–

e–c–d, Fig. 17a). Intra-crystalline diffusion was modelled along thispath, for post-UHT residence temperatures of 450 °C, 550 °C, 600 °Cand 650 °C (Fig. 17d–g). In all cases, the fit to measured crystal zoningprofiles is relatively good if the UHT event is sufficiently fast. It isbeyond the scope of this modelling to more accurately estimate theduration of this event, but total timescales of less than 20 Myr for pathsegment a–b–e are required to maintain prograde zoning if theheating and cooling rates are constant throughout the model. Thisduration can be extended if the proportion of the total path time thatis spent near peak-T is reduced.

For the purpose of illustration, Fig. 17d–g represents a very rapidUHT event (1 Myr) followed by 240 Myr residence at either 450 °C,550 °C, 600 °C or 650 °C, then reheating to inferred “peak Lepontineconditions” and subsequent exhumation (using the cooling curves ofRubatto et al., 2009). Results for residence at each temperature aresimilar, highlighting the huge difference in diffusivities between UHT(where 1 Myr of diffusion significantly modifies zoning profiles) andlow-grade conditions (where 240 Myr of diffusion has relatively littleeffect on the preserved profile of 3 mm diameter crystals). Permian toAlpine residence temperatures between 550 and 600 °C fit observa-tions best (Fig. 17e–f), with hotter models flattening profiles in allcomponents (particularly Mg and Fe). We emphasise, however, that itis not possible to particularly accurately resolve the lower bound onthe residence temperature because results from all models are similarand because of the likely uncertainties arising from (i) the modelassumptions described above, (ii) the activity models used to predictgrowth compositions, and (iii) the diffusivity data employed. Despitethis, it is clear that all models involving rapid post-UHT cooling andsubsequent heating (Fig. 17d–g) preserve garnet zoning trendsbroadly consistent with natural crystals, unlike the 1-stage model(Fig. 17c).

All models poorly fit the outermost parts of the observed garnetprofile. This is presumably partly due to a combination of theassumptions listed above, although longer durations spent just beforepeak-T and during parts of the exhumation path would result inslightly better fits.

9.4. Alpine overprint

It is not possible with the current garnet zoning model to easilydistinguish whether the post-peak symplectite assemblages grewduring cooling from the UHT event, or re-heating during theOligocene. However, various lines of evidence lead us to attributethe formation of symplectitic, mostly cordierite-bearing, post-peakassemblages to the Central Alpine Lepontine metamorphism. Theintense loss of melt during the pre-Alpine granulitic event lead to asubsequent absence or scarcity of melt or intergranular fluid, suchthat reaction kinetics may have been too slow to completelyrecrystallize the peak-assemblages, leading to symplectite andcoronae formation (Fyfe, 1973; Powell, 1983; Waters, 1988; Powelland Downes, 1990; White and Powell, 2002). Hence, we propose thatowing to their extremely refractory bulk-rock composition, the Grufgranulites only partially re-equilibrated during the Central Alpineupper-amphibolite facies metamorphism (at 720–740 °C and 7–7.5 kbar). This event was responsible for the development of 32 Mametamorphic rims on zircons separated from the Gruf granulites (Liatiand Gebauer, 2003). Monazite ages of 33 Ma (Schmitz et al., 2009) arealso consistent with this, although we acknowledge Schmitz et al.'sinterpretation that this records granulite-facies rather than amphib-olite-grade metamorphism. The possibility that this may record Pbloss due to diffusion in pre-existing (Permian) grains is controversial(see for example Cherniak et al., 2004; Smith and Giletti, 1997; Spearand Parrish, 1996) as monazite resetting temperatures are in thevicinity of the Lepontine metamorphic temperatures. Without furthertextural study, it is difficult to assess whether partial re-crystallisationof earlier grains may have occurred.


10. Conclusions

1. Previous to this study, Gruf granulites were represented only byblocks in a scree fan mostly situated below a steep mountain cliff,the top of which comprises the Bergell Intrusion. The discovery ofin-situ granulites and charnockites over much of the Gruf areaexcludes any genetic link between the high-temperature granulitesand the Bergell Intrusion (e.g. blocks of metapelites fallen into atonalitic or more basic magma).

2. Granulite metamorphism in this region is widespread in a beltranging from the Malenco to the Sesia Zone. During the Permian,the Gruf Complex was most likely part of this granulite belt. Theoccurrence of granulite facies rocks is in clear contrast to theadjacent gneissic Tambo Nappe and Bellinzona–Dascio Zone whichexperienced amphibolite-grade Alpine metamorphism, with noevidence of inherited UHT rocks (Heitzmann, 1975; Baudin andMarquer, 1993; Stucki et al., 2003). The Gruf also clearly contrastswith the Adula Nappe, which preserves evidence of eclogite faciesconditions (recording increasing depth southward from P≈10–17 kbar and T≈450–640 °C to PN25 kbar and T≈750–800 °C;Heinrich, 1986; Meyre et al., 1997; Nimis and Trommsdorff, 2001;Dale and Holland, 2003) followed by Alpine amphibolite-faciesmetamorphism (Nagel et al., 2002). Such eclogite relics have neverbeen found in the Gruf Complex.

3. Metamorphic conditions of 720–740 °C and 7–7.5 kbar, recorded incoronae and symplectites of the granulites and charnockites, fitwell with intrusion pressures of the Bergell tonalite (P~6.5 kbar,determined 1–3 km to the east; Reusser, 1987). They also agreewith amphibolite-grade P–T conditions established for Lepontinemetamorphism in the Adula Nappe, the Bellinzona–Dascio Zone20 km to the west (TN700 °C and P≈6–7 kbar; Heitzmann, 1975;Schmidt, 1989; Engi et al., 1995; Stucki, 2001; Nagel et al., 2002;Burri et al., 2005), and from migmatites found within the GrufComplex itself.

4. Models of garnet zoning patterns do not yield very accurate P–T–thistories for the complex granulite samples, but strongly suggestthat UHT conditions were brief, and that granulitic rocks resided atsignificantly lower temperatures (b 550–600 °C) for much of theirhistory, before experiencing Oligocene Alpine metamorphism.

5. A Permian age for Gruf Complex granulites and charnockitesrenders complex reinterpretations of the metamorphic history ofthe Central Alps, requiring an Oligocene UHT event, unnecessary.Identification of a two-stage history, which does not require acommon thermo-tectonic setting for both UHT rocks (whosethermal peak was seemingly short-lived) and “classical Alpine”amphibolite-grade migmatites makes their geodynamic interpre-tation far simpler.

Acknowledgments

We are thankful to P. Nievergelt for discussions improving themanuscript and to J.A.D. Connolly for help with Perplex. Our heartfeltthanks go to S. Ghizzoni, M. Cotola, G. Rochin and the family Biavaschifor their help during the field work. Thanks also to F. Pirovino toprovide excellent quality thin sections.

References

Ackermand, D., Seifert, F., 1969. Druck- und Temperaturbedingungen bei der Bildungder sapphirinführenden Gesteine von Val Codera. Fortschritte der Mineralogie 47.

Ague, J.J., Baxter, E.F., 2007. Brief thermal pulses during mountain building recorded bySr diffusion in apatite and multicomponent diffusion in garnet. Earth and PlanetaryScience Letters 261, 500–516.

Andersen, D.J., Lindsley, D.H., 1988. Internally consistent solution models for Fe–Mg–Mn–Ti oxides — Fe–Ti oxides. American Mineralogist 73, 714–726.

Annersten, H., Seifert, F., 1981. Stability of the assemblage orthopyroxene–sillimanite–quartz in the system MgO–FeO–Fe2O3–Al2O3–SiO2–H2O. Contributions to Miner-alogy and Petrology 77, 158–165.

Aranovich, L.Y., Berman, R.G., 1996. Optimized standard state and solution properties ofminerals II. Comparisons, predictions, and applications. Contributions to Mineral-ogy and Petrology 126, 25–37.

Aranovich, L.Y., Berman, R.G., 1997. A new garnet–orthopyroxene thermometer basedon reversed Al2O3 solubility in FeO–Al2O3–SiO2 orthopyroxene. American Miner-alogist 82, 345–353.

Artus, F., 1959. Über metamorphe Breccien basischer Gesteine im Granit des BergellerMassivs. Hamburger Beiträge zur Angewandten Mineralogie, Kristallphysik undPetrogenese 2, 1–39.

Baldwin, J.A., Powell, R., Brown, M., Moraes, R., Fuck, R.A., 2005. Mineral equilibriamodeling of ultrahigh-temperature metamorphism: an example from the Anápolis–Itauçu Complex, central Brazil. Journal of Metamorphic Geology 23, 511–531.

Barbosa, J., Nicollet, C., Leite, C., Kienast, J.-R., Fuck, R.A., Macedo, E.P., 2006. Hercynite–quartz-bearing granulites from Brejões Dome area, Jequié Block, Bahia, Brazil:influence of charnockite intrusion on granulite facies metamorphism. Lithos 92,537–556.

Barboza, S.A., Bergantz, G.W., 2000. Metamorphism and anatexis in the mafic complexcontact Aureole, Ivrea Zone, northern Italy. Journal of Petrology 41, 1307–1327.

Barker, F., 1964. Sapphirine-bearing rock, Val Codera, Italy. American Mineralogist 49,146–152.

Baudin, T., Marquer, D., 1993. Metamorphism and deformation in the Tambo nappe(Swiss Central Alps): evolution of the phengite substitution during Alpinedeformation. Schweizerische Mineralogische und Petrographische Mitteilungen73, 285–299.

Berger, A., Rosenberg, C., Schmid, S.M., 1996. Ascent, emplacement and exhumation ofthe Bergell pluton within the Southern Steep Belt of the Central Alps.Schweizerische Mineralogische und Petrographische Mitteilungen 76, 357–382.

Berger, A., Mercolli, I., Engi, M., 2005. The central Lepontine Alps: explanatory notesaccompanying the tectonic-geological map sheet Sopra Ceneri (1:100,000).Schweizerische Mineralogische und Petrographische Mitteilungen 85, 109–146.

Berger, A., Burri, T., Alt-Epping, P., Engi, M., 2008. Tectonically controlled fluid flow andwater-assisted melting in the middle crust: an example from the Central Alps.Lithos 102, 598–615.

Berger, A., Rosenberg, C., Schaltegger, U., 2009. Stability and isotopic dating of monaziteand allanite in partially molten rocks: examples from the Central Alps. SwissJournal of Geosciences 102, 15–29.

Berman, R.G., Aranovich, L.Y., 1996. Optimized standard state and solution properties ofminerals I. Model calibration for olivine, orthopyroxene, cordierite, garnet, andilmenite in the system FeO–MgO–CaO–Al2O3–TiO2–SiO2. Contributions to Miner-alogy and Petrology 126, 1–24.

Bernotat, W.H., Bambauer, H.U., 1982. The microcline/sanidine transformation isograd inmetamorphic regions. II. The region of Lepontine metamorphism, central Swiss Alps.Schweizerische Mineralogische und Petrographische Mitteilungen 62, 231–244.

Bohlen, S.R., 1991. On the formation of granulites. Journal of Metamorphic Geology 9,223–229.

Braga, R., Giacomini, F., Messiga, B., Tribuzio, R., 2001. The Sondalo gabbroic complex(Central Alps, orthern Italy): evidence for emplacement of mantle-derived meltsinto amphibolite-facies metapelites. Physics and Chemistry of the Earth. Part A:Solid Earth and Geodesy 26, 333–342.

Braga, R., Callegari, A., Messiga, B., Ottolini, L., Renna, M.R., Tribuzio, R., 2003. Origin ofprismatine from the Sondalo granulites (Central Alps, northern Italy). EuropeanJournal of Mineralogy 15, 393–400.

Brandt, S., Will, T.M., Klemd, R., 2007. Magmatic loading in the Proterozoic EpupaComplex, NW Namibia, as evidenced by ultrahigh-temperature sapphirine-bearingorthopyroxene–sillimanite–quartz granulites. Precambrian Research 153,143–178.

Brown, M., 2006. Duality of thermal regimes is the distinctive characteristic of platetectonics since the Neoarchaean. Geology 34, 961–964.

Brown, M., 2007. Metamorphic conditions in orogenic belts: a record of secular change.International Geology Review 49, 193–234.

Bucher-Nurminen, K., Droop, G., 1983. The metamorphic evolution of garnet–cordierite–sillimanite–gneisses of the Gruf–Complex, Eastern Pennine Alps.Contributions to Mineralogy and Petrology 84, 215–227.

Burri, T., Berger, A., Engi, M., 2005. Tertiary migmatites in the Central Alps: regionaldistribution, field relations, conditions of formation, and tectonic implications.Schweizerische Mineralogische und Petrographische Mitteilungen 85, 215–232.

Caddick, M.J., Konopásek, J., Thompson, A.B., 2010. Preservation of garnet growthzoning and the duration of prograde metamorphism. Journal of Petrology 51 (11),2327–2347.

Caddick, M.J., Thompson, A.B., 2008. Zoned garnets and the duration of metamorphicevents. Geochimica et Cosmochimica Acta 72 (12), 127.

Carlson, W.D., 2006. Rates of Fe, Mg, Mn and Ca diffusion in garnet. AmericanMineralogist 91, 1–11.

Cherniak, D.J., Watson, E.B., Grove, M., Harrison, T.M., 2004. Pb diffusion in monazite: acombined RBS/SIMS study. Geochimica et Cosmochimica Acta 68 (4), 829–840.

Clemens, J.D., Droop, G.T.R., Stevens, G., 1997. High-Grade Metamorphism, Dehydrationand Crustal Melting: A Reinvestigation Based on New Experiments in the Silica-Saturated Portion of the System KAlO2–MgO–SiO2–H2O–CO2 at P≤1.5 GPa.

Connolly, J.A.D., 2005. Computation of phase equilibria by linear programming: a toolfor geodynamic modelling and its application to subduction zone decarbonation.Earth and Planetary Sciences Letters 236, 524–541.

Cornelius, H.P., 1916. Ein alpines Vorkommen von Sapphirin. Centralblatt für Mineralogie,Geologie und Paläontologie, 265–269.

Dachs, E., Proyer, A., 2002. Constraints on the duration of high-pressure metamorphismin the Tauern Window from diffusion modelling of discontinuous growth zones ineclogite garnet. Journal of Metamorphic Geology 20, 769–780.


Dale, J., Holland, T.J.B., 2003. Geothermobarometry, P–T paths and metamorphic fieldgradients of high-pressure rocks from the Adula Nappe, Central Alps. Journal ofMetamorphic Geology 21, 813–829.

Das, K., Dasgupta, S., Miura, H., 2001. Stability of osumilite coexisting with spinel solidsolution in metapelitic granulites at high oxygen fugacity. American Mineralogist86, 1423–1434.

Davidson, C., Rosenberg, C., Schmid, S.M., 1996. Synmagmatic folding of the base of theBergell pluton, Central Alps. Tectonophysics 265, 213–238.

Diethelm, K.H., 1989. Petrographische und geochemische Untersuchungen anbasischen Gesteinen der Bergeller Intrusion. Ph.D. Thesis, ETH Zurich, Nr. 8855.

Droop, G., Bucher-Nurminen, K., 1984. Reaction textures and metamorphic evolution ofsapphirine-bearing granulites from the Gruf Complex, Italian Central Alps. Journalof Petrology 25, 766–803.

Ellis, D.J., 1987. Origin and evolution of granulites in normal and thickened crust.Geology 15, 167–170.

Engi, M., Todd, C.S., Schmatz, D.R., 1995. Tertiary metamorphic conditions in the easternLepontine Alps. Schweizerische Mineralogische und Petrographische Mitteilungen75, 347–369.

Engi, M., Bousquet, R., Berger, A., 2004. Explanatory notes to the map: metamorphicstructure of the Alps. Central Alps. Mitteilungen Österreiche MineralogischeGesellschaft 149, 53–69.

Fitzsimons, I.C.W., Harley, S.L., 1994. The influence of retrograde cation-exchange ongranulite P–T estimates and a convergence technique for the recovery of peakmetamorphic conditions. Journal of Petrology 35, 543–576.

Frey, M., Hunziker, J.C., Frank, W., Bocquet, J., Dal Piaz, G.V., Jäger, E., Niggli, E., 1974.Alpine metamorphism of the Alps. A review. Schweizerische Mineralogische undPetrographische Mitteilungen 54, 247–290.

Frey, M., Ferreiro Mählmann, R., 1999. Alpine metamorphism of the Central Alps.Schweizerische Mineralogische und Petrographische Mitteilungen 79, 135–154.

Frost, B.R., Chacko, T., 1989. The granulite uncertainty principle: limitations onthermometry in granulites. Journal of Geology 18, 441–454.

Fuhrman, M.L., Lindsley, D.H., 1988. Ternary-feldspar modelling and thermometry.American Mineralogist 73, 201–215.

Fyfe, W.S., 1973. The granulite facies, partial melting and the Archaean crust.Philosophical Transactions of the Royal Society of London, Series A. Mathematicaland Physical Sciences 247, 457–461.

Galli, A., Le Bayon, B., Schmidt, M.W., Burg, J.P., Larionov, A., Sergejev, S., in review.Geochronological constraints in the Gruf Complex: evidence for charnockiticmagmatism in Permian time. Contributions to Mineralogy and Petrology.

Gebauer, D., 1996. A P–T–t path for an (ultra?-) high-pressure ultramafic/mafic rock-association and its felsic country-rocks based on SHRIMP-dating of magmatic andmetamorphic zircon domains. Example: Alpe Arami (Central Swiss Alps). In: Basu,A., Hart, S.R. (Eds.), Earth Processes: Reading the Isotopic Code. : GeophysicsMonograph Serie, 95. American Geophysics Union, Washington, pp. 269–276.

Harley, S.L., 1989. The origins of granulites: a metamorphic perspective. GeologicalMagazine 126, 215–247.

Harley, S.L., 1998a. On the occurrence and characterisation of ultrahigh-temperaturecrustal metamorphism. In: Treloar, P.J., O'Brien, P.J. (Eds.), What Drives Metamor-phism and Metamorphic Reactions? : Geological Society of London SpecialPublications, London, 138, pp. 81–107.

Harley, S.L., 1998b. An appraisal of peak temperatures and thermal histories inultrahigh-temperature (UHT) crustal metamorphism: the significance of alumi-nous orthopyroxene. In: Motoyoshi, Y., Shiraishi, K. (Eds.), Origin and Evolution ofContinents, 53. Memoir National Institute of Polar Research, Tokyo, pp. 49–73.Special Issue.

Harley, S.L., 2004. Extending our understanding of ultrahigh temperature crustalmetamorphism. Journal of Mineralogical and Petrological Sciences 99, 140–158.

Harley, S.L., 2008. Refining the P–T records of UHT crustal metamorphism. Journal ofMetamorphic Geology 26, 125–154.

Harley, S.L., Green, D.H., 1982. Garnet-orthopyroxene barometry for granulites andperidotites. Nature 300, 697–701.

Harley, S.L., Motoyoshi, Y., 2000. Al zoning in orthopyroxene in a sapphirine quartzite:evidence forN1120 °C UHT metamorphism in the Napier Complex, Antarctica, andimplications for the entropy of sapphirine. Contributions to Mineralogy andPetrology 138, 293–307.

Harley, S.L., Hensen, B.J., Sheraton, J.W., 1990. Two-stage decompression in orthopyr-oxene–sillimanite granulites from Forefinger Point, Enderby Land, Antarctica:implications for the evolution of the Archean Napier Complex. Journal ofMetamorphic Geology 8, 591–613.

Heinrich, C., 1986. Eclogite facies regional metamorphism of hydrous mafic rocks in theCentral Alpine Adula nappe. Journal of Petrology 27, 123–154.

Heitzmann, P., 1975. Zur Metamorphose und Tektonik im südostlichen Teil derLepontinischen Alpen (Provincia di Como, Italia). Schweizerische Mineralogischeund Petrographische Mitteilungen 55, 467–522.

Hensen, B.J., 1971. Theoretical phase relations involving cordierite and garnet in thesystem MgO–FeO–Al2O3–SiO2. Contributions to Mineralogy and Petrology 33,191–214.

Hensen, B.J., 1986. Theoretical phase relations involving cordierite and garnet revisited:the influence of oxygen fugacity on the stability of sapphirine and spinel in thesystem Mg–Fe–Al–Si–O. Contributions to Mineralogy and Petrology 92, 362–367.

Hensen, B.J., 1987. P–T grids for silica-undersaturatedgranulites in the systemMAS (n+4)and FMAS (n+3)— tools for thederivations of P–Tpaths ofmetamorphism. Journal ofMetamorphic Geology 5, 255–271.

Hensen, B.J., Green, D.H., 1970. Experimental data on coexisting cordierite and garnetunder high grade metamorphic conditions. Physics of the Earth and PlanetaryInteriors 3, 431–440.

Hensen, B.J., Green, D.H., 1973. Experimental study of the stability of cordierite andgarnet in pelitic compositions at high pressures and temperatures. III. Synthesis ofexperimental data and geological applications. Contributions to Mineralogy andPetrology 38, 151–166.

Hensen, B.J., Harley, S.L., 1990. Graphical analysis of P–T–X relations in granulite faciesmetapelites. In: Ashworth, J.R., Brown, M. (Eds.), High-temperature Metamorphismand Crustal Anatexis, Series, 2. The Mineralogical Society of Great Britain, Unwin-Hyman, London, pp. 19–56.

Hensen, B.J., Osanai, Y., 1994. Experimental study of dehydration melting of F-bearingbiotite in model pelitic compositions. Mineralogical Magazine 58A, 410–411.

Higgins, J.B., Ribbe, P.H., Herd, R.K., 1979. Sapphirine I. Crystal chemical contributions.Contributions to Mineralogy and Petrology 68, 349–356.

Holland, T.J.B., Powell, R., 1990. An enlarged and updated internally consistent data-setwith uncertainties and correlations: the system K2O–Na2O–CaO–MgO–MnO–FeO–Fe2O3–Al2O3–TiO2–SiO2–C–H2–O2. Journal of Metamorphic Geology 8, 89–124.

Holland, T.J.B., Powell, R., 1998. An internally consistent thermodynamic dataset forphases of petrological interest. Journal of Metamorphic Geology 16, 309–343.

Holland, T.J.B., Powell, R., 2003. Activity-composition relations for phases inpetrological calculations: an asymmetric multicomponent formulation. Contribu-tions to Mineralogy and Petrology 145, 492–501.

Holland, T.J.B., Babu, E.V.S.S.K., Waters, D.J., 1996. Phase relations of osumilite anddehydration melting in pelitic rocks: a simple thermodynamic model for theKFMASH system. Contributions to Mineralogy and Petrology 124, 383–394.

Hänny, R., Grauert, B., Soptrajanova, G., 1975. Paleozoic migmatites affected by high-grade Tertiary metamorphism in the Central Alps (Valle Bodengo, Italy). Ageochronological study. Contributions to Mineralogy and Petrology 51, 173–196.

Kelly, N.M., Harley, S.L., 2004. Orthopyroxene-corundum in Mg–Al-rich granulites fromthe Oygarden Islands, east Antarctica. Journal of Petrology 45, 1481–1512.

Kelsey, D.E., 2008. On ultrahigh-temperature metamorphism. Gondwana Research 13,1–29.

Kelsey, D.E., White, R.W., Powell, R., 2003. Orthopyroxene–sillimanite–quartz assem-blages: distribution, petrology, quantitative P–T–X constraints and P–T paths.Journal of Metamorphic Geology 21, 439–453.

Kelsey, D.E., White, R.W., Holland, T.J.B., Powell, R., 2004. Calculated phase equilibria inK2O–FeO–MgO–Al2O3–SiO2–H2O for sapphirine–quartz-bearing mineral assem-blages. Journal of Metamorphic Geology 22, 559–578.

Kelsey, D.E., White, R.W., Powell, R., 2005. Calculated phase equilibria in K2O–FeO–MgO–Al2O3–SiO2–H2O for silica-undersaturated sapphirine-bearing mineralassemblages. Journal of Metamorphic Geology 23, 217–239.

Konrad-Schmolke, M., Handy, M.R., Babist, J., O'Brien, P.J., 2005. Thermodynamicmodelling of diffusion-controlled garnet growth. Contributions to Mineralogy andPetrology 149 (2), 181–195.

Konrad-Schmolke, M., O'Brien, P.J., De Capitani, C., Carswell, D.A., 2007. Garnet growthat high- and ultra-high pressure conditions and the effect of element fractionationon mineral modes and composition. Lithos 103, 309–332.

Köppel, V., Günthert, A., Grünenfelder, M., 1981. Patterns of U–Pb zircon and monaziteages in polymetamorphic units of the Swiss Alps. Schweizerische Mineralogischeund Petrographische Mitteilungen 61, 97–119.

Lal, R.K., Ackermand, D., Raith, M., Raase, P., Seifert, F., 1984. Sapphirine-bearingassemblages from Kiranur, Southern India: a study of chemographic relationshipsin the Na2O–FeO–MgO–Al2O3–SiO2–H2O system. Neues Jahrbuch für Mineralo-gische Abhandlungen 150, 121–152.

Lardeaux, J.M., Spalla, M.I., 1991. From granulites to eclogites in the Sesia zone (ItalianWestern Alps): a record of the opening and closure of the Piedmont ocean. Journalof Metamorphic Geology 9, 35–59.

Lasaga, A.C., 1979. Multicomponent exchange and diffusion in silicates. Geochimica etCosmochimica Acta 43, 455–469.

Lasaga, A.C., Jiang, J., 1995. Thermal history of rocks: P–T–t paths from geospeedometry,petrologic data, and inverse theory techniques. American Journal of Science 295,697–741.

Leite, C.de.M.M., Barbosa, J.S.F., Goncalves, P., Nicollet, C., Sabaté, P., 2009. Petrologicalevolution of silica-undersaturated sapphirine-bearing granulite in the Paleoproter-ozoic Salvador-Curaçá Belt, Bahia, Brazil. Gondwana Research 15, 49–70.

Liati, A., Gebauer, D., 2003. Geochronological constraints for the time ofmetamorphism in theGruf Complex (Central Alps) and implications for the Adula-Cima Lunga nappe system.Schweizerische Mineralogische und Petrographische Mitteilungen 83, 159–172.

Lorenz, V., Nicholls, I.A., 1976. The Permocarboniferous Basin and Range Province of Europe.An application of plate tectonics. In: Falke, H. (Ed.), The Continental Permian in Central,West, andSouthEurope.D.Reidel PublishingCompany,Dordrecht,Holland,pp. 313–342.

Lorenz, V., Nicholls, I.A., 1984. Plate and intraplate processes of Hercynian Europeduring the Late Paleozoic. Tectonophysics 107, 25–56.

McDade, P., Harley, S.L., 2001. A petrogenetic grid for aluminous granulite faciesmetapelites in the KFMASH system. Journal of Metamorphic Geology 19, 45–59.

Meyre, C., De Capitani, C., Partzsch, J., 1997. A ternary solid solution model foromphacite and its application to geothermobarometry of eclogites from the MiddleAdula nappe (Central Alps, Switzerland). Journal of Metamorphic Geology 15,687–700.

Moraes, R., Brown, M., Fuck, R.A., Camargo, M.A., Lima, T.M., 2002. Characterization and P–Tevolution of melt-bearing ultrahigh-temperature granulites: an example from theAnápolis–Itauçu Complex of the Brasília Fold Belt, Brazil. Journal of Petrology 43,1673–1705.

Moticska, P., 1970. Petrographie und Strukturanalyse des westlichen Bergeller Massivsund seines Rahmens. Schweizerische Mineralogische und Petrographische Mittei-lungen 50, 355–443.

Mouri, H., Guiraud, M., Hensen, B.J., 1996. Petrology of phlogopite–sapphirine-bearingAl–Mg granulites from Ihouhaouene, In Ouzzal, Hoggar, Algeria: an example of


phlogopite stability at high temperature. Journal of Metamorphic Geology 14,725–738.

Müller, W., Prosser, G., Mancktelow, N.S., Villa, I.M., Kelley, S.P., Viola, G., Oberli, F., 2001.Geochronological constraints on the evolution of the Periadriatic Fault System(Alps). International Journal of Earth Sciences 90, 623–653.

Müntener, O., Hermann, J., Trommsdorff, V., 2000. Cooling history and exhumation oflower crustal granulite and Upper Mantle (Malenco, Eastern Central Alps). Journalof Petrology 41, 175–200.

Nagel, T., De Capitani, C., Frey, M., 2002. Isograds and P–T evolution in the easternLepontine Alps (Graubünden, Switzerland). Journal of Metamorphic Geology 20,309–324.

Nandakumar, V., Harley, S.L., 2000. A reappraisal of the pressure–temperature path ofgranulites from the Kerala Khondalite Belt, southern India. Journal of Geology 108,687–703.

Niggli, E., 1970. Alpine Metamorphose und alpine Gebirgsbildung. Fortschritte derMineralogie 47, 16–26.

Niggli, E., Niggli, C., 1965. Karten der Verbreitung einiger Mineralien der alpidischenMetamorphose in den Schweizer Alpen (Stilpnomelan, Alkali-Amphibol, Chlor-itoid, Staurolith, Disthen, Sillimanit). Eclogae Geologicae Helvetiae 58, 335–368.

Nimis, P., Trommsdorff, V., 2001. Revised thermobarometry of Alpe Arami and othergarnet peridotites from the Central Alps. Journal of Petrology 42, 103–115.

Osanai, Y., Sajeev, K., Owada, M., Kehelpannala, K.V.W., Prame, W.K.B., Nakano, N.,Jayatileke, S., 2006. Metamorphic evolution of high-pressure and ultrahigh-temperature granulites from the Highland Complex, Sri Lanka. Journal of AsianEarth Sciences 28, 20–37.

Pattison, D.R.M., Begin, N.J., 1994. Zoning patterns in orthopyroxene and garnet ingranulites: implications for geothermometry. Journal of Metamorphic Geology 12,387–410.

Pattison, D.R.M., Chacko, T., Farquhar, J., McFarlane, C.R.M., 2003. Temperatures of granulite-facies metamorphism: constraints from experimental phase equilibria and thermo-barometry corrected for retrograde exchange. Journal of Petrology 44, 867–900.

Peterson, J.W., Chacko, T., Kühner, S.M., 1991. The effect of fluorine on the vapour-absent melting of phlogopite+quartz: implication for deep-crustal processes.American Mineralogist 76, 470–476.

Platt, J.P., Soto, J.-I., Whitehouse, M.J., Hurford, A.J., Kelly, S.P., 1998. Thermal evolution,rate of exhumation, and tectonic significance of metamorphic rocks from the floorof the Alboran extensional basin, western Mediterranean. Tectonics 17, 671–689.

Powell, R., 1983. Processes in granulite-facies metamorphism. In: Atherton, M.P.,Gribble, C.D. (Eds.), Migmatites, Melting and Metamorphism. Shiva Publishing,Nantwich, Cheshire, pp. 127–139.

Powell, R., Holland, T.J.B., 1988. An internally consistent thermodynamic dataset withuncertainties and correlations. 3. Applications, methods, worked examples and acomputer program. Journal of Metamorphic Geology 6, 173–204.

Powell, R., Sandiford, M.J., 1988. Sapphirine and spinel phase relationships in thesystem FeO–MgO–Al2O3–TiO2–O2 in the presence of quartz and hypersthene.Contributions to Mineralogy and Petrology 98, 64–71.

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 CrustalAnatexis. The Mineralogical Society of Great Britain, Series, 2. Unwin-Hyman,London, pp. 105–123.

Powell, R., Holland, T.J.B., 1994. Optimal geothermometry and geobarometry. AmericanMineralogist 79, 120–133.

Powell, R., Holland, T.J.B., 1999. Relating formulations of the thermodynamics ofmineral solid solutions: activity modelling of pyroxenes, amphiboles and micas.American Mineralogist 84, 1–14.

Raith, M., Karmakar, S., Brown, M., 1997. Ultra-high temperature metamorphism andmultistage decompressional evolution of sapphirine granulites from the Palni HillRanges, southern India. Journal of Metamorphic Geology 15, 379–399.

Rebay, G., Spalla, M.I., 2001. Emplacement at granulite facies conditions of the Sesia–Lanzo metagabbros: an early record of Permian rifting? Lithos 58, 85–104.

Reusser, E., 1987. Phasenbeziehungen im Tonalit der Bergeller Intrusion. Bericht überdie 62. Hauptversammlung in Luzern. Schweizerische Mineralogische undPetrographische Mitteilungen 67, 377–378.

Richardson, S.W., 1968. Staurolite stability in a part of the system Fe–Al–Si–O–H.Petrology 9, 757–768.

Rubatto, D., Hermann, J., Berger, A., Engi, M., 2009. Protracted fluid-present meltingduring Barrovian metamorphism in the Central Alps. Contributions to Mineralogyand Petrology. doi:10.1007/s00410-009-0406-5.

Sajeev, K., Osanai, Y., 2004. Ultrahigh-temperature metamorphism (1150 °C, 12 kbar)and multistage evolution of Mg-, Al-rich granulites from the central HighlandComplex, Sri Lanka. Journal of Petrology 45, 1821–1844.

Sandiford, M., Neale, F., Powell, R., 1987. Metamorphic evolution of aluminousgranulites from Labwor Hills, Uganda. Contributions to Mineralogy and Petrology95, 217–225.

Santosh, M., Sajeev, K., 2006. Anticlockwise evolution of ultrahigh-temperaturegranulites within continental collision zone in southern India. Lithos 92, 447–464.

Schefer, S., 2005. Deformation und Metamorphose im Gruf-Komplex (1), Val Codera,Italien. Unpublished Diploma Thesis, University of Basel.

Schmid, S.M., Aebli, H.R., Heller, F., Zingg, A., 1989. The role of the Periadriatic linein the tectonic evolution of the Alps. In: Dietrich, D., Coward, M.D. (Eds.),Alpine Tectonics: Geological Society of London Special Pubblications, London, 45,pp. 153–171.

Schmid, S.M., Pfiffner, O.A., Froitzheim, N., Schönborn, G., Kissling, E., 1996.Geophysical–geological transect and tectonic evolution of the Swiss–Italian Alps.Tectonics 15, 1036–1064.

Schmidt, M.W., 1989. Petrography and structural evolution of ophiolitic remnants inthe Bellinzona Zone, Southern Steep Belt, Central Alps (CH, I). SchweizerischeMineralogische und Petrographische Mitteilungen 69, 393–405.

Schmitz, S., Möller, A., Wilke, M., Malzer, W., Kannigiesser, B., Bousquet, R., Berger, A.,Schefer, S., 2009. Chemical U–Th–Pb dating of monazite by 3D-Micro X-rayfluorescence analysis with synchrotron radiation. European Journal of Mineralogy21, 927–945.

Schreyer, W., 1970. Metamorphose pelitischer Gesteine im Modellsystem MgO–Al2O3–

SiO2–H2O. Fortschritte der Mineralogie 47, 124–165.Schreyer, W., Abraham, K., 1978. Symplectitic cordierite–orthopyroxene–garnet

assemblages as products of contact metamorphism of pre-existing basementgranulites in the Vredefort structure, South Africa and their relation topseudotachylite. Contributions to Mineralogy and Petrology 68, 53–62.

Schubert, G., Turcotte, D.L., Olson, P., 2001. Mantle Convection in the Earth and Planets.Cambridge University Press, Cambridge. 941 pp.

Schuster, R., Stüwe, K., 2008. Permian metamorphic event in the Alps. Geology 36,603–606.

Sengupta, P., Sen, J., Dasgupta, S., Raith, M., Bhui, U.K., Ehl, J., 1999. Ultra-hightemperature metamorphism of metapelitic granulites from Kondapalle, EasternGhats Belt: implications for the Indo–Antarctic correlation. Journal of Petrology 40,1065–1087.

Smith, H.A., Giletti, B.J., 1997. Lead diffusion in monazite. Geochimica et CosmochimicaActa 61, 1047–1055.

Spear, F.S., 1988. Metamorphic fractional crystallization and internal metasomatism bydiffusional homogenization of zoned garnets. Contributions to Mineralogy andPetrology 99, 507–517.

Spear, F.S., Kohn, M.J., Florence, F.P., Menard, T., 1991. A model for garnet andplagioclase growth in pelitic schists: implications for thermobarometry and P–Tpath determinations. Journal of Metamorphic Geology 8, 683–696.

Spear, F.S., Parrish, R.R., 1996. Petrology and cooling rates of the Valhalla Complex,British Columbia, Canada. Journal of Petrology 37, 733–765.

Srogi, L., Wagner, M.E., Lutz, T.M., 1993. Dehydration partial melting and disequilibriumin the granulite-facies Wilmington Complex, Pennsylvania–Delaware Piedmont.American Journal of Science 293, 405–462.

Steck, A., Hunziker, J., 1994. The Tertiary structural and thermal evolution of the CentralAlps — compressional and extensional structures in an orogenic belt. Tectono-physics 238, 229–254.

Stevens, G., Clemens, J.D., Droop, G.T.R., 1997. Melt production during granulite-faciesanatexis: experimental data from “primitive” metasedimentary protoliths. Con-tributions to Mineralogy and Petrology 128, 352–370.

Storm, L.C., Spear, F.S., 2005. Pressure, temperature and cooling rates of granulite faciesmigmatitic pelites from the southern Adirondack Highlands, New York. Journal ofMetamorphic Geology 23, 107–130.

Stucki, A., 2001. High grade Mesozoic ophiolites of the Southern Steep Belt, CentralAlps. Unpubl. PhD thesis ETH Zurich, No. 14206.

Stucki, A., Rubatto, D., Trommsdorff, V., 2003. Mesozoic ophiolite relics in the SouthernSteep Belt of the Central Alps. Schweizerische Mineralogische und PetrographischeMitteilungen 83, 285–299.

Stüwe, K., 1997. Effective bulk composition changes due to cooling: a model predictingcomplexities in retrograde reaction textures. Contributions to Mineralogy andPetrology 129, 43–52.

Tajčmanová, L., Connolly, J.A.D., Cesare, B., 2009. A thermodynamic model for titaniumand ferric iron solution in biotite. Journal of Metamorphic Geology 27, 153–165.

Todd, C.S., Engi, M., 1997. Metamorphic field gradients in the Central Alps. Journal ofMetamorphic Geology 15, 513–530.

Tong, L., Wilson, C.J.L., 2006. Tectonothermal evolution of the ultrahigh temperaturemetapelites in the Rauer Group, east Antarctica. Precambrian Research 149, 1–20.

Trommsdorff, V., 1966. Progressive Metamorphose kieseliger Karbonatgesteine in denZentralalpen zwischen Bernina und Simplon. Schweizerische Mineralogische undPetrographische Mitteilungen 46, 431–460.

Vielzeuf, D., Holloway, J.R., 1988. Experimental determination of the fluid-absentmelting relations in the pelitic system. Consequences for crustal differentiation.Contributions to Mineralogy and Petrology 98, 257–276.

Vielzeuf, D., Montel, J.M., 1994. Partial melting of metagreywackes. Part I. Fluid-absentexperiments and phase relationships. Contributions to Mineralogy and Petrology117, 375–393.

Von Blanckenburg, F., 1992. Combined high-precision chronometry and geochemicaltracing using accessory minerals: applied to the Central-Alpine Bergell intrusion(Central Europe). Chemical Geology 100, 19–40.

Waters, D.J., 1988. Partial melting and the formation of granulite facies assemblages inNamaqualand, South Africa. Journal of Metamorphic Geology 6, 387–404.

Wenk, E., 1955. Eine Strukturkarte der Tessiner Alpen. Schweizerische Mineralogischeund Petrographische Mitteilungen 35, 311–319.

Wenk, E., 1975. Zur alpinen Metamorphose. Schweizerische Mineralogische undPetrographische Mitteilungen 55, 116–125.

Wenk, H.-R., 1970. Zur Regionalmetamorphose und Ultrametamorphose der Zentra-lalpen. Fortschritte der Mineralogie 47, 34–51.

Wenk, H.-R., 1973. The structure of the Bergeller Alps. Eclogae Geologicae Helvetiae 66,255–291.

Wenk, H.-R., 1982. A geological history of the Bergell granite and related rocks.Transformists Petrology. Theophrastus publications, Athens, pp. 113–148.

Wenk, H.-R., 1986. Introduction to the geology of the Bergell Alps with guide forexcursions. Jahresbericht der Naturforschenden Gesellschaft Graubünden 103,29–90.

Wenk, H.-R., 1992. Note esplicative. Foglio 70 Sciora. Servizio idrologico e geologiconazionale. 61 pp.


Wenk, H.R., Wenk, E., Wallace, J.H., 1974. Metamorphic mineral assemblages in peliticrocks of the Bergell Alps. Schweizerische Mineralogische und PetrographischeMitteilungen 54, 507–554.

Wenk, H.R., Cornelius, S.B., 1977. Geologischer Atlas der Schweiz, 1:25000. Publikatio-nen Schweizerische Geologische Kommission, Blatt Sciora.

White, R.W., Powell, R., 2002. Melt loss and the preservation of granulitie facies mineralassemblages. Journal of Metamorphic Geology 20, 621–632.

White, R.W., Powell, R., Clarke, G.L., 2002. The interpretation of reaction textures in Fe-rich metapelitic granulites of the Musgrave Block, central Australia: constraintsfrom mineral equilibria calculations in the system K2O–FeO–MgO–Al2O3–SiO2–

H2O–TiO2–Fe2O3. Journal of Metamorphic Geology 20, 41–55.White, R.W., Powell, R., Holland, T.J.B., 2007. Progress relating to calculation of partial

melting equilibria for metapelites. Journal of Metamorphic Geology 25, 511–527.

Granulites and charnockites of the Gruf Complex: Evidence for …€¦ · Granulites and...

Documents

Transcript of Granulites and charnockites of the Gruf Complex: Evidence for …€¦ · Granulites and...