fluid-assisted interaction of peraluminous metapelites

Precambrian Research 253 (2014) 114–145

Contents lists available at ScienceDirect

Precambrian Research

jo ur nal home p ag e: www.elsev ier .com/ locate /precamres

Fluid-assisted interaction of peraluminous metapelites withtrondhjemitic magma within the Petronella shear-zone,Limpopo Complex, South Africa

Oleg G. Safonova,b,c,∗, Daria S. Tatarinovab, Dirk D. van Reenenc, Maria A. Golunovaa,Vasily O. Yapaskurtb

a Institute of Experimental Mineralogy, Russian Academy of Sciences, Academician Ossipian str., 4, Chernogolovka, Moscow District, 142432, Russiab Department of Petrology, Moscow State University, Vorob’evy Gory, Moscow, 119899, Russiac Department of Geology, University of Johannesburg, Johannesburg, South Africa

a r t i c l e i n f o

Article history:Received 28 February 2014Received in revised form 30 May 2014Accepted 9 June 2014Available online 19 June 2014

Keywords:Granulite complexesTrondhjemitic meltsThermobarometryFluids

a b s t r a c t

The principal problem concerning the evolution of Precambrian granulite complexes located betweengranite-greenstone cratons is their interaction with the underthrusted greenstone belts during exhuma-tion from the lower crust. Besides evidence for pervasive and localized fluid fluxes arising fromdevolatilization of the greenschist and amphibolite-facies rocks in the course of their burial under-neath the granulites, this interaction might also be expressed in formation of diverse magmas ofgranitic, trondhjemitic and granodioritic composition. We present results of a petrological, fluid inclusionand thermobarometric study of interaction between fluidized trondhjemitic magma and peraluminousmetapelitic granulites associated with the regional high-grade Petronella shear zone located in theSouthern Marginal Zone (SMZ) of the Limpopo Granulite Complex (South Africa). The hot (T ∼ 1000 ◦C)trondhjemitic magma, which, presumably, originated from partial melting of a basaltic (amphibolite)material at the base of the granulite complex or at the top of the underthrusted greenstone blocks,intruded granulites at P ∼ 8.0–9.5 kbar (24–28 km depth) at 2.667 ± 0.9 Ga during the exhumation of theSMZ. The magma heated and assimilated orthopyroxene-cordierite metapelites and dragged them toa depth of 18–20 km (6.3–6.5 kbar). The magma was heterogeneously saturated with MgO, FeO, Al2O3

by the dissolved metapelites. This process provoked crystallization of several garnet generations fromthe trondhjemitic melt. Various mineral assemblages included in the different generations of garnetallowed application of TWQ method combined with PERPLE X pseudo sections to trace sub-isobariccooling of the magma from T ∼ 900–600 ◦C at 5.5–6.5 kbar. The isobaric cooling also affected the asso-ciated metapelites. Fluid inclusions trapped in garnet and quartz in the trondhjemite show that themagma transported carbonic fluid with densities corresponding to the late stages of magma cooling(600–650 ◦C and 5.5–6.5 kbar). Carbonic fluids coexisted with aqueous-salt fluids (preserved as inclusionswith salinity up to 20.58 wt.% NaCl eq.). These low water activity fluids (aH2O < 0.3) bearing Na, K andCa salts, being exsolved from the magma on cooling and solidification, provoked formation of complexNa-gedrite + biotite + sillimanite + quartz ± staurolite ± plagioclase-bearing assemblages after cordieritein metapelites at temperatures 630–570 ◦C and pressures 5.5–6.5 kbar. These data provide evidence thathot trondhjemitic melts played a critical role in the exhumation of granulites onto the adjacent granite-greenstone craton. The trondhjemite transferred heat from the lower to the middle crust and transportedlarge volumes of external aqueous-carbonic-salt fluids that participated in the rehydration of a significantportion of the SMZ.

© 2014 Elsevier B.V. All rights reserved.

∗ Corresponding author at: Institute of Experimental Mineralogy, RussianAcademy of Sciences, Academician Ossipian str., 4, Chernogolovka, Moscow District,142432, Russia. Tel.: +7 496 525 44 25; fax: +7 496 525 44 25.

E-mail address: [email protected] (O.G. Safonov).

1. Introduction

A much discussed issue regarding the evolution of Precambriangranulite complexes located between granite-greenstone cratonsis their interaction with the underthrusted greenstone belts dur-ing exhumation from the lower crust. Although the direct relations

http://dx.doi.org/10.1016/j.precamres.2014.06.0060301-9268/© 2014 Elsevier B.V. All rights reserved.

dx.doi.org/10.1016/j.precamres.2014.06.006

http://www.sciencedirect.com/science/journal/03019268

http://www.elsevier.com/locate/precamres

http://crossmark.crossref.org/dialog/?doi=10.1016/j.precamres.2014.06.006&domain=pdf

mailto:[email protected]

dx.doi.org/10.1016/j.precamres.2014.06.006

O.G. Safonov et al. / Precambrian Research 253 (2014) 114–145 115

Ab albiteAlm almandineAlOpx alumino-orthopyroxene end-member (AlAlO3)An anorthiteAnd andalusiteAnn anniteAnth anthophylliteBt biotiteCel celadoniteCor corundumCrd cordieriteEas eastoniteEn enstatitefCrd Fe-cordieriteFs ferrosilliteFsp ternary feldsparGed gedriteGrs grossularGrt garnetHc herciniteIlm ilmeniteKfs K-feldsparKy kyaniteMu muscoviteOpx orthopyroxeneOr orthoclasePhl phlogopitePl plagioclasePrp pyropeQtz quartzRu rutileSid siderophilliteSil sillimaniteSpl spinelSps spessartineSt stauroliteZo zoisiteZrc zircon

Compositional parametersXMg Mg/(Mg + Fe + Mn)XAl Al/(Si + Al + Ti + Mg + Fe + Mn) (for biotites and

orthopyroxenes)XCa Ca/(Ca + Na + K) (for plagioclase)XCa Ca/(Ca + Mg + Fe + Mn) (for garnet)XK K/(K + Na + Ca) (for alkali feldspar)

of the granulite complexes with the greenstone belts are rarelyexposed, detailed petrological studies of prime examples from theLimpopo Complex (South Africa), Lapland Granulite Belt (Russia-Finland), and Enisey Range (Russia) proved that the P–T–t andgeodynamic evolution of the greenstone cratons and juxtaposedgranulite complexes are closely interrelated (Perchuk et al., 1996,2000; Smit et al., 2000; Van Reenen et al., 2011). Exhumation ofthe granulite complexes from the lower crust always results inunderthrusting of the cratonic rocks followed by their subsequentjoint exhumation toward the surface (see Van Reenen et al., 2011and references therein). The tectonic juxtaposition of the granulitebelts with the cratonic complexes should not only be expressed bytheir joint P–T–t paths, but also by interrelated fluid and magmaticactivities. Volatile-rich cratonic rocks (greenschists, amphibolites,tonalite-trondhjemite gneisses) evidently served as strong sources

for various fluids and magmas that might have invaded the over-riding granulites from below (e.g. Huizenga et al., 2014).

Such relations are best studied for the Neoarchean LimpopoComplex (LC) where a large geological, structural, petrological, geo-chemical, geochronological, and isotopic database (e.g. Van Reenenet al., 1992, 2011 and references therein; Huizenga et al., 2014)demonstrated that the evolution of the granulite facies South-ern Marginal Zone (SMZ) of the Limpopo Complex was stronglyinfluenced by externally derived aqueous-carbonic and aqueous-salt (brine) fluids of reduced water activity. The source of thesefluids is assumed to be underthrusted greenstone rocks of thenorthern Kaapvaal Craton (NKVC). This assumption is sustainedby geophysical data showing that more than 60% of the SMZ atdepth is presently underlain by greenstone belts (e.g. De Beerand Stettler, 1992). Fluids being derived from devolatilization ofthe underthrusted greenstone material infiltrated overriding gran-ulites during the thrust-controlled exhumation onto the adjacentgranite-greenstone belts of the northern Kaapvaal Craton (NKVC)within the time period 2.69–2.62 Ga. At pressures about 6 kbarand temperatures 650–620 ◦C, these fluids established a so called“orthoamphibole isograd” in the hanging wall of the shallow north-dipping Hout River Shear Zone (HRSZ) (age 2.69 Ga) that bounds theSMZ in the south (Van Reenen et al., 2011; Huizenga et al., 2014).This isograd, defined in metapelites (Fig. 1a) by the disappear-ance of orthopyroxene (and the first appearance of anthophyllite)owing to the univariant reaction Opx + Qtz + H2O = Anth, subdividesthe SMZ into a northern granulite zone in which orthopyroxene isalways a stable phase in metapelites, and a southern zone of rehy-drated granulite in which orthopyroxene is completely absent frommetapelites. However, the late hydration of metapelites causedby infiltrating fluids is most clearly reflected by the divariantreaction Crd + H2O = Ged + Sil (or/and Ky) + Qtz. This reaction is notrestricted to the position of the isograd, but also affected cordieritein metapelites within the granulite zone. Orthopyroxene also dis-appears from quartz-feldspathic gneisses (Baviaanskloof gneisses)more or less at the position of the isograd, but is often still preservedas a relict phase in mafic, ultramafic gneisses and banded ironfor-mation south of the isograd. In addition to re-hydration, similarfluids also provoked extensive shear zone-hosted potassic meta-somatic alteration of the Baviaanskloof gneisses and formation oflode-gold mineralization in other rock types (e.g. Huizenga et al.,2014).

Fluids that resulted from devolatilization of the greenstonerocks are also considered to have triggered partial meltingeither at the top of the underthrusted greenstone pile or atthe base of the overthrusted granulites. Such partial melting ofamphibolite and tonalitic gneisses produced fluidized melts ofgranodioritic–trondhjemitic composition that were able to intrudeand interact with granulites, which had already been migmatizedat the peak of metamorphic conditions (e.g. Du Toit et al., 1983).Du Toit et al. (1983) first suggested, on the basis of intrusiverelationships with migmatized granulite, that the major pulse ofanatexis in the SMZ must have been linked to partial melting thatoccurred at depth below the granulite pile. In contrast, Stevens andvan Reenen (1992) and Stevens (1997) claimed that all anatecticevents, including the major pulse of anatexis could be explained bya mechanism of in situ dehydration melting of metapelitic rocksinvolving muscovite and biotite prior to the metamorphic peakat ∼2.72 Ga (Retief et al., 1990). However, dehydration melting asa viable mechanism to explain granite–granulite relations in theSMZ has been questioned by a number of authors. Kreissig et al.(2001) first showed that the timing ∼2.64 Ga of the major anate-ctic event exposed in the Bandelierkop quarry locality in the SMZ(Fig. 1a) was significantly later than the time of initial exhumation(∼2.69 Ga) and also than the time of peak metamorphism in theSMZ (∼2.72 Ga) (Retief et al., 1990). These authors also argued that

116 O.G. Safonov et al. / Precambrian Research 253 (2014) 114–145

Fig. 1. (a) Structural map of the South Marginal Zone (SMZ) and adjacent North-ern Kaapvaal Craton (NKVC). Major shear-zones: HRSZ, Hout River Shear Zone; 1,Matok shear-zone; 2, Petronella shear-zone; 3, Annaskraal shear-zone. Hatchedarea indicates the portion of the SMZ underlain by greenstone material of theNKVC. White circles show location of large trondhjemite and granodiorite intru-sions. White square shows a position of the studied locality at farm Petronella.White hexagon shows position of the locality at Bandelierkop quarry. (b) Generalview of the studied outcrop in the Sand River on farm Petronella. Light-coloredcoarse-grained trondhjemite body includes strongly attenuated lens-like blocks ofcordierite-orthopyroxene metapelite. Some blocks are continued into the trond-hjemite as long tails of garnet grains. 1–4, samples discussed in the paper (1,metapelite SAF12-2/2; 2, trondhjemite PET6; 3, trondhjemite PET10 and PET11; 4,garnet-rich reaction zone between metapelite and trondhjemite PET1 and PET2).

the main pulse of anatexis had been provoked by a mechanismof decompression melting during exhumation. The age ∼2.67 Garecently measured for the major trondhjemitic body that intrudedmetapelite at the Petronella locality (Fig. 1b) (Belyanin et al., 2014)also clearly postdates commencement of the exhumation stage(∼2.69 Ga) as has been dated from cordierite-bearing granulitesat Bandelierkop locality (Kreissig et al., 2001). Stable O-isotopedata recently obtained from metapelitic granulite and the associ-ated major leucocratic band at the Bandelierkop quarry also plainlyexcludes an isotopic link of the host metapelitic granulite with themajor pulse of anatexis (E.O. Dubinina and L.Ya. Aranovich, personalcommunications).

These new age and stable O-isotope data provide additionalinsight into the role of granitic magmatism during the tectonic,thermal and fluid evolution of the SMZ granulite terrain. In thepresent paper, we show first results of a study of the interactionbetween trondhjemitic intrusions and peraluminous metapelitewithin the high-grade Petronella shear-zone (Fig. 1a). A petro-logical study combined with fluid inclusion data, conventionalthermobarometry and pseudosection modeling allowed various

Table 1Bulk composition of garnet trondhjemites (PET6, PET10, PET11) and rocks from thereaction zone between trondhjemite and metapelite (PET1, PET2).

PET1 PET2 PET6 PET10 PET11

SiO2 61.85 56.70 69.75 75.40 75.65TiO2 0.74 – – – –Al2O3 12.82 19.22 17.60 13.31 14.01Cr2O3 0.07 – – – –Fe2O3 10.07 12.70 0.35 1.90 0.60MgO 5.78 4.05 – 0.54 0.19MnO 0.11 0.17 – – –CaO 2.17 1.71 2.33 1.89 2.10Na2O 2.96 3.59 6.49 4.45 4.92K2O 1.44 0.82 1.59 0.72 0.80BaO – 0.05 0.09 – –P2O5 0.12 0.05 0.08 – –SO3 1.41 – – – –Total 99.54 99.06 98.28 98.21 98.27A/(CNK) 1.95 3.14 1.69 1.89 1.79

Note: XRF (borate fusion), set of major elements without LOI. LLDs are about0.05 wt.%. Measurements below 0.05 wt.% are shown as “–”. LLD may be higher than0.05 wt.% for SO3. A/(CNK) = Al2O3/(CaO + Na2O + K2O).

reactions to be identified and calculated and sequences of min-eral assemblages preserved in the trondhjemite and metapelitemodeled. Finally, this integrated approach helped to specify thejoint P–T evolution and fluid regime of the trondhjemite intrusionand partially assimilated metapelite during exhumation.

2. Geological setting and sampling

Fig. 1a shows that, similar to other major trondhjemite-granodiorite bodies in the SMZ, the studied body is located inthe granulite subzone south of the Annaskraal shear zone, thatis at depth underlain by underthrusted greenstone material ofthe NKVC. The studied locality on farm Petronella (Google Earth:23◦20′42.43′′ S; 29◦36′42.72′′ E) includes a trondhjemitic body ofup to 1 km in width that intrudes metapelitic granulites in a dryriver bed (Fig. 1b). U/Pb age study of zircons extracted from thistrondhjemite body (samples PET10 and PET11 in Fig. 1b) by meansof LA-ICP-MS has indicated the presence of two age populations,2.667 ± 0.9 Ga and 2.726 ± 1.3 Ga (Belyanin et al., 2014). The olderzircon population represents inherited xenocrysts reflecting thepeak of metamorphism, whereas the younger age is considered torepresent the age of the emplacement of the trondhjemitic body. Itis interesting to note that the age of emplacement is very similar tothe age of the large Matok pluton, 2.671–2.664 Ga (Barton and VanReenen, 1992; Bohlender et al., 1992).

The studied outcrop exposed in the Sand River (Fig. 1b) showswater-smoothed exposures and large boulders of leucocratic rocksconsisting mainly of plagioclase and quartz (trondhjemites, seebelow) (samples PET6, PET10, PET11; Table 1). These rocks encom-pass blocks of dark-greenish-gray fine-grained metapelite andhighly attenuated fish-like lenses of the same rock type 20–50 cmwide and several meters in length. The largest metapelitic block inthe outcrop shown in Fig. 1b is about 1 m wide and 2 m long (sampleSAF12-2/2 is collected from this block). This block contains leucoc-ratic enclaves with numerous garnet grains of different sizes. Zonescrowded with garnet grains (samples PET1 and PET2 in Table 1)usually surrounds the metapelitic blocks, while long garnet tailsdefines the continuation of the attenuated metapelite blocks intothe trondhjemite body (Fig. 2a–d). Separate garnet grains are dis-persed in the trondhjemite way from the metapelite blocks (Fig. 2e).The studied samples were collected from the trondhjemite materialat different distance from the metapelite blocks, from centers of themetapelite blocks, from the garnet-rich trondhjemite-metapelitecontacts, as well as separate garnet grains that were extracted fromthe trondhjemite.


Fig. 2. Field relations of the rock types discussed in the paper. (a) Necked-down block of orthopyroxene-cordierite-biotite metapelite with large “tail” of garnet in quartz-plagioclase (trondhjemitic) matrix; separate garnet grains are dispersed in the trondhjemitic material; the block is characterized by elongated biotitic lenses of variable sizes.(b) Several meters-long garnet “tails” in the trondhjemite; dashed box shows the position of the detailed image in figure c. (c) The detailed view of the garnet cluster fromthe “tail” shown in the image b. It is clearly seen that biotite shells armor most garnet grains in the cluster, while separate grains in the trondhjemitic matrix are usuallydevoid of such shells. (d) Cluster of garnet grains in the trondhjemite. This image demonstrates that the garnet grains are crowded with diverse inclusions, while biotite shellssurround some grains. Greenish material between the garnet grains is a relic of orthopyroxene or cordierite from the original metapelite. The size of the largest garnet grainsis about 3 cm. (e) Garnet grains dispersed in the trondhjemite way from the metapelite blocks; the dashed insert illustrates large garnet grains with numerous inclusions. (f)Leucocratic orthopyroxene-garnet-K-feldspar-plagioclase-quartz enclave in metapelite from a locality just south of the locality shown in Fig. 1b.

The metapelite sample PET-5 was collected from a large out-crop of metapelitic granulite that is not closely associated withthe main trondhjemite body. This was done for comparison withmetapelite SAF12-2/2 that was trapped by trondhjemite (Fig. 1b).In addition, just south of the studied locality, we have collecteda sample (DR4-10) from a spectacular coarse-grained leucocraticenclave with large subhedral crystals of orthopyroxene and gar-net in quartz-feldspathic matrix (Fig. 2f) in order to compare thisenclave with garnet-bearing trondhjemite.

3. Analytical methods

Bulk rock analyses (Table 1) were performed using the XRFmethod at the central analytical facility (SPECTRAU) of the Univer-sity of Johannesburg (South Africa). After petrographic observation

using optical microscope, analyses of minerals were performedusing three analytical facilities.

(1) CamScan MV2300 (VEGA TS 5130MM) electron microscopeequipped with EDS INCA-Energy-350 and Tescan VEGA-II XMUmicroscope equipped with EDS INCA-Energy-450 and WDSOxford INCA Wave 700 at the Institute of Experimental Min-eralogy, Chernogolovka, Russia. Analyses were performed at20 kV accelerating voltage with a beam current of 10 nA anda beam diameter of 3 �m. Counting times was 100 s for all ele-ments. The ZAF matrix correction was applied. The followingstandards were used: SiO2 for Si and O, albite for Na, micro-cline for K, wollastonite for Ca, pure titanium for Ti, corundumfor Al, pure manganese for Mn, pure iron for Fe, periclase forMg, BaF2 for Ba.


(2) Jeol 6480 LV equipped with EDS detector INCA – Energy 350and WDS detector INCA Wave 500 (Oxford instruments) at theLaboratory of Local Methods of Analysis at the Department ofPetrology of the Moscow State University; analytical conditionsare 15 kV acceleration voltage, 15 nA beam current, countingtimes of 100 s.

(3) CAMECA SX100 equipped both with WDS and EDS detectorsat the central analytical facility (SPECTRAU) of the Universityof Johannesburg (South Africa). Analytical conditions are 15 kVacceleration voltage, 20 nA beam current, counting times of10–20 s for most elements, except 50 s for Na, and 1–5 �m beamspot size.

In cases where optical detection proved to be impossible, Ramanspectroscopy was applied for identification of kyanite and silli-manite in the products of retrograde reactions after cordierite inmetapelite. For this purpose, we used the JY Horiba XPloRa Jobinspectrometer (at Department of Petrology, Moscow State Univer-sity, Moscow, Russia) equipped with a polarized Olympus BX41microscope. Spectra were obtained using a 532-nm laser within therange 200–4000 cm−1 during 40 s (two times 20 s each). The spectrawere refined with LabSpec (version 5.78.24) software. Assignmentof the Raman bands of kyanite or sillimanite was carried out usingweb-based databank http://rruff.geo.arizona.edu/rruff/.

Fluid inclusions were investigated in double-polished sec-tions (∼150 to 200 �m thick) using the LINKAM THMSG 600heating–freezing stage at the Institute of Experimental Mineral-ogy. The stage works within the temperature range from −196 ◦C to600 ◦C with automatic heating/cooling with the rate 0.1–90 ◦C/min.Accuracy of the thermometric measurements is about ±0.1 ◦C. Sys-tematic calibration of the stage has been performed using natural(CO2, Camperio, Alps) and synthetic (H2O) inclusions in quartz.

4. Petrography and mineral assemblages of studied rocks

4.1. Garnet-bearing trondhjemite

Garnet-bearing trondhjemite is a leucocratic coarse-grainedrock consisting of plagioclase and quartz (Fig. 2a–e). Plagioclasein the studied samples varies within 50–70 vol.% and quartz within25–45 vol.%. However, some portions of the rock are totally com-posed of plagioclase with no or few quartz grains (Fig. 3a).K-feldspar is present (at least, in the studied samples) only asexsolution lamellae in plagioclase or as thin rims at the plagioclase-quartz contacts (Fig. 3a) and the K-feldspar content in the rock doesnot exceed 5 vol.%. Variations in bulk composition of this rock arerepresented by samples PET6, PET10 and PET11 (Table 1). The majorfeature of the rocks is low contents of MgO and Fe2O3 (and FeO),while the ratio A/(CNK) > 1 reflects the strong peraluminous char-acter. Following to the QAP diagram of the IUGS classification (LeMaitre et al., 2002), this rock can be classified as leucocratic tonalite.Taking into account the composition of plagioclase in the rock (seebelow) and its extreme leucocratic appearance, the rock will bedescribed as trondhjemite.

The trondhjemite shows subhedral to anhedral grains of plagio-clase and anhedral interstitial quartz (Fig. 3a). Large (up to severalcm) megacrystals of plagioclase are sporadically present in therock. Thin K-feldspar lamellae (Fig. 3a) concentrated in the centersof plagioclase crystals clearly represent exsolution from a ternaryhigh-temperature phase rather than from metasomatic replace-ment. Coarsening of antiperthite lamellae and change in theirorientation toward periphery of the plagioclase crystals (Fig. 3a)seems to reflect recrystallization (dissolution/precipitation) dur-ing cooling in presence of a fluid phase (e.g. Putnis, 2002). Locally,coarse graphic-like intergrowths of plagioclase and quartz are

also observed. Garnet is the major mafic mineral in the trond-hjemite and forms clusters in the vicinity of the metapeliteblocks that projects into the trondhjemite (Fig. 2a). These clus-ters build long “tails”, which trace the metapelite blocks into thetrondhjemitic body (Fig. 2a and b). Usually, garnet grains in the“tails” are armored by biotite shells or are locally surrounded byremnants of partially altered metapelitic cordierite or orthopy-roxene (Fig. 2c and d; Fig. 3b). These relations clearly indicatethat garnet in the trondhjemite is a product of interaction of themetapelite with the trondhjemite. Separate large garnet crystalsfar away from the metapelite blocks contain visible inclusions(Fig. 2e).

Two morphological types of garnets in the trondhjemite wererecognized during inspection of several thin sections. The com-mon morphological type (referred to as GRT-I) occurs either asseparate grains or as groups of two, three or several accreted crys-tals of various sizes (Fig. 3a, c and d). Although various types ofinclusion assemblages have been recognized in GRT-I, many grainsonly contain inclusions of quartz. The size of quartz inclusions insome garnet grains reaches 200–500 �m (Fig. 3d). Amoeboid mus-covite is also a common type of inclusions in GRT-I. Muscoviteis usually accompanied by inclusions of other minerals of whichquartz is most common, but rare inclusions of sillimanite werealso detected in association with muscovite. Biotite, which locallyaccompanies muscovite inclusions, usually is developed along con-tacts of muscovite with host garnet, indicating that this biotite isa product of later interaction of muscovite with garnet. Neverthe-less, separate euhedral inclusions of biotite are also present in GRT-I(Fig. 3b). Small euhedral inclusions of zoisite are found among mus-covite amoeboid inclusions in two garnet grains (Fig. 3d). Someinclusions of zoisite are in contact with muscovite. Plagioclaseinclusions (Fig. 3d) are very rare in garnet of the first morphologicaltype.

The second morphological type of garnets (GRT-II) appears aslarge, up to 1–2 cm in size, lace-like rounded or irregular grainswith numerous inclusions (Fig. 3h). Predominant inclusions insuch garnet grains are polycrystalline acicular or sheaf-like silli-manite aggregates often accompanied by corundum, spinel andbiotite (Fig. 3i and k). Spinel is usually attached to the con-tacts of sillimanite-corundum inclusions with host garnet and isheavily crowded with inclusions of both sillimanite and corundum(Fig. 3i–k). Spinel is not present at contacts of single sillima-nite inclusions in garnet. Biotite in the polyphase inclusions isclearly later, since it forms at contacts of sillimanite and corun-dum with host garnet and replaces spinel (Fig. 3i and k). Fig. 3hclearly shows that sillimanite-crowded garnet is overgrown byan outer zone. Inclusions of biotite are locally present in thiszone. In addition, polyphase inclusions consisting of biotite, mus-covite and sillimanite were also recognized (Fig. 3l). The outermostzones of the large garnet contain separate inclusions of mus-covite, which morphologically resemble the amoeboid inclusionsin GRT-I (Fig. 3e and g). In addition to large grains of gar-nets (Fig. 3h), irregular garnet aggregates were also identifiedin the trondhjemite. Usually, these aggregates grow around oron sillimanite at contacts with the plagioclase-quartz matrix(Fig. 3m and n). Such garnet aggregates are usually accompaniedor overgrown by garnet with biotite inclusions, similar to GRT-I(Fig. 3m).

Muscovite is very rare in the trondhjemite matrix, while biotiteforms shells around garnet grains. Such shells are abundant andthick around garnet in the vicinity of metapelite blocks (Fig. 2aand b), but only thin shells or separate flakes of biotite surroundgarnet located far away from the metapelite blocks (Fig. 2e). Inaddition, biotite forms small lenses in the trondhjemite (Fig. 2a),which, probably, originated from replacement of the metapelitefragments.

http://rruff.geo.arizona.edu/rruff/


Fig. 3. Textural features of the garnet trondhjemite. (a) Typical finely exsolved subhedral plagioclases and rounded garnets in the trondhjemite; the inset shows a detailedview of the exsolved plagioclases. (b) Relic of altered cordierite replaced by biotite and sillimanite. (c) Group of rounded garnet grains (the first morphological type ofgarnet) in the quartz-plagioclase matrix; garnet contains inclusions of quartz, and locally inclusions of muscovite, biotite and sillimanite (marked by the dashed box); matrixplagioclase is crowded with lamellae of K-feldspar. (d) Large garnet grains (the first morphological type of garnet) with dispersed inclusions of muscovite and zoisite andseparate inclusions of quartz and plagioclase; matrix plagioclase is crowded with lamellae of K-feldspar; K-feldspar also forms rare rims at contacts of quartz with plagioclase;dashed box indicates the position of image g. (e) Large subhedral inclusions of quartz in garnet of the first morphological type. (f) Small inclusions of biotite, muscovite andquartz at the periphery of the garnet grain. (g) Detailed view of amoeboid inclusions of muscovite accompanied by rare subhedral inclusions of zoisite in the garnet shown inimage d. (h) Composite garnet grain of the second morphological type with numerous inclusions of sillimanite and large polyphase inclusions; the wide core of garnet withsillimanite and polyphase inclusions is overgrown by outer zone with less abundant inclusions; dashed boxes indicate position of images i, j, k, and l; image l is partially outof view. (i) Detailed view of the polyphase inclusion; the inclusions is cored by sillimanite; spinel forming at contacts of sillimanite with the host garnet contains inclusionsof both sillimanite and corundum; later biotite is developed after sillimanite and spinel. (j) Prismatic inclusions of sillimanite and corundum in spinel from the polyphaseinclusion presented in the image i. (k) Polyphase inclusion consisting of sillimanite and corundum; spinel forms at contacts with the host garnet and contains corunduminclusions; the inclusions contains abundant late biotite. (l) Polyphase inclusion of biotite + muscovite + sillimanite and separate muscovite inclusions at the periphery ofthe garnet of the second morphological type. (m) Garnet aggregate around sheaf-like sillimanite; garnet captures inclusions of zircon; the right portion of the garnet doesnot contains sillimanite inclusions, but contains inclusions of biotite; late biotite also forms around garnet at its contacts with the plagioclase matrix. (n) Large aggregate offibrous sillimanite in the plagioclase matrix of the trondhjemite partially overgrown by garnet; garnet contains numerous inclusions of sillimanite; gray staff in the sillimaniteaggregates is late biotite.


Fig. 3. (Continued ).

4.2. Metapelites

Metapelites intruded by trondhjemite are typical gran-ulite facies rocks of the Bandelierkop Formation (e.g. VanReenen et al., 2011). Stevens and van Reenen (1992) dis-tinguished three major rock types among metapelitesof the Bandelierkop formation exposed within the Ban-delierkop quarry located about 25 km north-east of thePetronella locality (Fig. 1a): (1) garnet-orthopyroxene gneisses(+biotite + plagioclase + quartz), (2) garnet-cordierite gneisses(+orthopyroxene + biotite + plagioclase + quartz), and (3)cordierite-orthopyroxene gneisses (+biotite + plagioclase + quartz).Assemblages of these rock types reflect differences in bulk

composition as expressed by their Mg-number. A macroscopicstudy of the metapelitic blocks captured by the trondhjemite(Fig. 2a) at the Petronella locality showed that they usuallydo not contain garnet suggesting that they belong to the Mg-rich (cordierite-orthopyroxene) rock types of the Bandelierkopformation. Numerous garnets in the trondhjemite are in closecontact with the metapelitic blocks (Fig. 2a) or are attached toquartz-feldspathic (trondhjemitic) enclaves and lenses (Fig. 4a).These enclaves are usually located within local shears zones inthe metapelite that are either nearly concordant with the foliationor crosscut the foliation (Fig. 4a). These relations indicate theintrusive relations of the enclaves and lenses with the metapeliteand the formation of garnets via interaction of the trondhjemite


Fig. 4. Textural features of the cordierite-orthopyroxene-biotite-plagioclase-quartz metapelite. (a) Block of finely banded and foliated garnet-free metapelite with garnet-quartz-feldspathic enclaves attached to small shear-zones; it is clearly seen that the left portion of a larger enclave is nearly concordant with the foliation of the metapelite,whereas the right portion cross cuts the foliation. (b) BSE image of the metapelite SAF12-2/2; large orthopyroxene grains contain rounded inclusions of cordierite (marked withdashed frames) partially replaced by gedrite, biotite, sillimanite and locally staurolite, plagioclase and biotite; relics of cordierite are attached to boundaries of orthopyroxenegrains and sporadically form anhedral grains in quartz-plagioclase bands. (c) Cordierite grain replaced by the assemblage Ged + Sil + Bt in matrix of the metapelite SAF12-2/2.Plagioclase forms thin rims on cordierite at contacts with matrix quartz (indicated by black arrows). Close view of the plagioclase rim on cordierite (dashed box) is givenin figure d. (d) Close view of plagioclase rim with euhedral inclusions of staurolite and sillimanite on cordierite in contact with matrix quartz. (e) Large orthopyroxenegrain with inclusions of cordierite, biotite, and plagioclase. Cordierite inclusions are partially replaced by the assemblage gedrite + sillimanite + biotite. (f) Late assemblagegedrite + sillimanite (probably, kyanite) + biotite formed after cordierite (no relics of cordierite is found in this case) in the metapelite SAF12-2/2; staurolite, anthophyllite,and plagioclase are minor constituents of this assemblage; matrix plagioclase is locally intensively zoned in contact with the gedrite-sillimanite intergrowths with formationof Ca-rich zones (indicated by white arrow). (g) Isometric cluster consisting of intergrowths of orthopyroxene with coarse-grained cordierite in the quartz-plagioclase matrixof metapelite PET-5; the dashed box show the position of image h. (h) Late orthopyroxene-sillimanite, orthopyroxene-biotite, and biotite-sillimanite-quartz intergrowthsafter cordierite in the orthopyroxene-cordierite cluster in metapelite PET-5 (see image g).


Fig. 5. Textural features of the contact zone between trondhjemite and metapelite. (a) Resorbed garnet grain containing inclusions of quartz and biotite is partially surroundedby a biotite shell; dashed box shows the position of image d. (b) Growth of garnet after relict cordierite; cordierite is partially replaced by biotite-sillimanite-quartz aggregates;cores of matrix plagioclase are crowded by K-feldspar exsolution lamellae. (c) Relic of orthopyroxene contacting the resorbed garnet; both orthopyroxene and garnet areextensively resorbed by biotite-quartz intergrowths (see inset). (d) Later garnet-ilmenite-rutile micro assemblage after biotite; the inset shows the close view of the ilmenite-rutile intergrowths. (e) Growth of late idiomorphic garnet accompanied by sillimanite and quartz along the contacts of large garnet with plagioclase. (f) Myrmeckite-likeplagioclase-quartz textures intermixed with K-feldspar between garnet; growth of new Ca-rich garnet and locally Ti-poor biotite is visible along the grain boundaries of largegarnet grains; the inset (scale 100 �m) shows intergrowths of late garnet, biotite, quartz, plagioclase and K-feldspar.

material with the metapelite. This observation is fully consistentwith the relations of the smaller metapelite blocks that occurwithin the main trondhjemite body (Fig. 2a).

Cordierite-orthopyroxene metapelite are gray-greenish fine-grained foliated and banded rocks. The foliation in the metapeliteblocks is more distinct at the contact with the garnet-bearing trond-hjemite veins or with local shear-zones. The general characteristicof the metapelite varies because of different proportions of majorrock-forming minerals (orthopyroxene, cordierite, biotite, plagio-clase, quartz), various textural relationships, presence or absence ofminor and accessory minerals, and various retrograde assemblages.Two representative samples used for further thermobarometricstudy can shortly be described.

Metapelite SAF12-2/2 was collected from the block trappedin the trondhjemite (Fig. 1b). It is a clearly foliated rockwith orthopyroxene-rich and biotite-rich bands alternating with

plagioclase-quartz bands (Fig. 4b). The leucocratic bands show vari-able proportions of plagioclase and quartz, from about 10 vol.%of quartz up to 50 vol.% of quartz. Relics of cordierite, partiallyreplaced by intergrowths of gedrite, biotite, and aluminum sili-cate are usually present in the quartz-rich bands (Fig. 4b and c). Nosigns of kyanite in the intergrowths have been identified both opti-cally and by Raman analysis. Relic cordierite is locally surroundedby plagioclase rims, which contain minute crystals of sillima-nite and staurolite (Fig. 4d). Chains of attenuated orthopyroxenegrains and large biotite flakes are lined-up along the foliation(Fig. 4b). Large orthopyroxene grains contain rounded inclusionsof cordierite, plagioclase and biotite (Fig. 4e). All cordierite inclu-sions are partially or completely replaced by gedrite, sillimanite,and biotite aggregates. Relics of partially replaced cordierite arealso attached to boundaries of orthopyroxene grains (Fig. 4f).Gedrite-sillimanite-biotite intergrowths along the orthopyroxene


Fig. 6. Textural features of the leucocratic orthopyroxene and garnet-bearingenclave in metapelites.

boundaries imply the presence of former cordierite (Fig. 4f). In addi-tion to gedrite, sillimanite and biotite, the late assemblage includessubordinate anthophyllite, staurolite, and plagioclase. Zones of Ca-rich plagioclase form along contacts of matrix plagioclase with thegedrite-sillimanite textures (Fig. 4f).

In contrast to sample SAF12-2/2, metapelite PET-5 does notshow a clear foliation. Biotite flakes are chaotically oriented withinthe rock, which is characterized by two domains. The spectacu-lar textural feature of the first domain is near-isometric clustersconsisting of coarse worm-like orthopyroxene within a cordieritematrix (Fig. 4g). These clusters are usually surrounded by coarseorthopyroxene at the contact with quartz and sporadically containbiotite-replacing orthopyroxene. Texturally, these clusters are sim-ilar to spectacular cordierite-orthopyroxene coronas around garnetthat have been described from metapelite of the SMZ and inter-preted as products of the reaction Grt + Qtz = Opx + Crd during adecompression–cooling P–T path that had commenced at about2.69 Ga (Van Reenen, 1983; Stevens and van Reenen, 1992; Perchuket al., 1996, 2000; Smit et al., 2001; Van Reenen et al., 2011). No relicgarnet is present in the metapelite PET-5, that is characteristic forthe Mg-rich metapelites (e.g. Stevens and van Reenen, 1992). Butthe clusters imply that this mineral must have been present in theassemblage at the early stages of exhumation.

Orthopyroxene “worms” in the cordierite-orthopyroxene clus-ters in the metapelite PET-5 are usually rimmed with fineorthopyroxene-sillimanite intergrowths in contacts with cordierite(Fig. 4h). Quartz is usually absent in the intergrowths. It appearsjust locally in veinlets of acicular and fibrous sillimanite crossingcordierite, as well as in the symplectite-like textures consisting ofbiotite, quartz, sillimanite and orthopyroxene (Fig. 4h).

The second domain is composed of coarse-grained orthopyrox-ene, cordierite, biotite, plagioclase and quartz. This domain prob-ably represents strongly re-crystallized and coarsened (annealed)orthopyroxene-cordierite clusters.

4.3. The trondhjemite-metapelite reaction zone

The reaction zones between trondhjemite and metapelite blocksbear many features inherited from the trondhjemite and are usually

melanocratic because of the high content of biotite and garnet. Thebulk composition of these zones is highly variable with respect toall components (samples PET-1 and PET-2 in Table 1), because ofvariable proportions of biotite, garnet and quartz-feldspathic mate-rial. Cores of some plagioclase in the reaction zone are crowdedby lamellae of K-feldspar, whereas peripheral zones of plagioclasegrains are lamellae-free (Fig. 5a–c). Locally, peripheral zones ofplagioclases are represented by myrmeckite-like intergrowths ofplagioclase and quartz (Fig. 5c). Relics of orthopyroxene are abun-dant in the trondhjemite-metapelite reaction zone (Fig. 5c and f),while relict cordierite is very rare (Fig. 5b).

In contrast to garnet in the trondhjemite, garnet in the reac-tion zones at the contacts with the metapelite blocks are stronglyresorbed and completely or partially surrounded by thick biotiteshells (Fig. 5a). No inclusions of sillimanite, corundum, spinel,muscovite and zoisite are identified in garnet from the reac-tion zone. However, this garnet contains abundant inclusions ofbiotite and quartz (Fig. 5a–d). Matrix of the rock is characterizedby extensively developed biotite-quartz-plagioclase symplectiticaggregates, which are usually attached to plagioclase grain bound-aries. Biotite-quartz symplectitic intergrowths heavily replace bothgarnet and relic grains of cordierite and orthopyroxene (Fig. 5b, cand f).

Specific micro assemblages are developed along the contactsof garnet grains with surrounding minerals. Thin garnet zoneswith numerous inclusions of ilmenite and rutile locally appearsalong the contacts of garnet with biotite (Fig. 5d). At contactswith matrix plagioclase, idiomorphic garnet appears. They areusually accompanied by sillimanite-quartz veinlets crossing pla-gioclase (Fig. 5e). Idiomorphic garnet locally accompanied bybiotite was also found at contacts of large garnet grains with spec-tacular plagioclase-quartz myrmeckite-like textures developed inthe trondhjemite-metapelite reaction zone (Fig. 5f). Perthitic inter-growths of K-feldspar and plagioclase are locally attached to thesetextures (Fig. 5f).

4.4. Leucocratic orthopyroxene and garnet-bearing enclaves inmetapelites

Spectacular feature of the coarse-grained leucocratic enclavesin metapelites are large (locally, up to 3 cm) subhedral to euhe-dral orthopyroxene crystals (Fig. 2f). Smaller (0.5–2 mm) roundedgarnet grains are mostly attached to contacts of the orthopyroxenecrystals with the matrix consisting of plagioclase, K-feldspar, quartz(Fig. 6). K-feldspar (up to 50 vol.%) is a characteristic feature of theseenclaves, which discriminate them from the garnet trondhjemite.The monzonitic (subhedral plagioclase crystals in K-feldspar) tex-ture is locally developed in the enclave implying its formation withparticipation of a melt (Fig. 6). It is interesting to note that biotiteis totally absent in the enclaves, although contacting metapelitescontain abundant biotite.

5. Compositions of minerals

Garnet composition in the trondhjemite and the reaction zonevary within the narrow range: XMg = 0.29–0.37, XCa = 0.02–0.04, andXMn = 0.005–0.015 (Table 2 ). Differences between morphologicaltypes of garnet GRT-I and GRT-II in the trondhjemite and garnetfrom the reaction zone are mostly determined by different zoningpatterns.

Small grains of GRT-I (0.5–1 mm) in the trondhjemite usu-ally show very weak zoning (XMg = 0.34–0.35, XCa = 0.01–0.02, XMnbelow 0.01). Zoning of larger grains (up to 2.5 mm) containinginclusions of Qtz, Bt and Mu (Fig. 3c, e and f) is characterizedby a bowl-shaped symmetrical profiles of XMg and flat profiles of


Table 2Traverse through the garnet grain (GRT-I) from the trondhjemite (Fig. 6a).

Components 1 2 3 4 5 6 7 8 9 10 11

SiO2 38.24 38.46 38.44 38.23 38.19 38.35 38.40 38.50 38.40 38.47 38.72TiO2 0.01 0.01 0.00 0.01 0.00 0.03 0.02 0.02 0.01 0.02 0.01Al2O3 21.84 21.95 21.74 21.72 21.89 21.67 21.97 21.97 21.85 21.84 21.97Cr2O3 0.00 0.03 0.00 0.00 0.00 0.00 0.00 0.02 0.02 0.03 0.00FeO 28.42 29.66 29.19 29.10 29.44 29.40 29.49 29.53 29.19 28.08 27.51MnO 0.51 0.48 0.53 0.47 0.45 0.54 0.51 0.43 0.49 0.49 0.47MgO 9.09 8.60 8.53 8.58 8.61 8.59 8.83 8.80 8.66 9.30 9.71CaO 0.78 0.83 0.77 0.75 0.75 0.69 0.78 0.77 0.74 0.71 0.74Total 98.89 100.02 99.20 98.86 99.33 99.27 100.00 100.04 99.36 98.94 99.13Formula quantities normalized to 12 OSi 2.998 3.015 3.014 2.997 2.994 3.007 3.011 3.019 3.011 3.016 3.036Ti 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000Al 2.015 2.025 2.005 2.003 2.019 1.999 2.027 2.027 2.015 2.015 2.027Cr 0.000 0.002 0.000 0.000 0.000 0.000 0.000 0.001 0.001 0.002 0.000Fe 1.862 1.943 1.913 1.907 1.929 1.926 1.932 1.935 1.913 1.840 1.802Mn 0.034 0.032 0.035 0.031 0.030 0.036 0.034 0.028 0.032 0.032 0.031Mg 1.061 1.004 0.996 1.002 1.005 1.003 1.031 1.027 1.011 1.086 1.133Ca 0.066 0.070 0.065 0.063 0.063 0.058 0.066 0.065 0.062 0.060 0.062XMg 0.351 0.329 0.331 0.334 0.332 0.332 0.337 0.336 0.335 0.360 0.374XCa 0.022 0.023 0.022 0.021 0.021 0.019 0.021 0.021 0.021 0.020 0.021XMn 0.011 0.010 0.012 0.010 0.010 0.012 0.011 0.009 0.011 0.011 0.010

Note: 1, the left marginal point of the traverse; 11, the right marginal point of the traverse.

Traverse through the garnet grain (GRT-I) containing inclusions of muscovite and zoisite from the trondhjemite (Fig. 6b).

Components 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

SiO2 39.70 40.03 40.37 39.90 41.06 39.67 40.33 40.36 39.93 39.58 40.39 40.47 40.53 40.09 40.77 40.49TiO2 0.01 0.04 0.09 0.04 0.00 0.00 0.00 0.03 0.19 0.00 0.09 0.05 0.10 0.00 0.00 0.01Al2O3 22.27 22.61 23.51 22.82 23.06 22.45 21.69 22.37 22.28 21.87 22.26 22.04 22.96 22.51 22.00 23.24Cr2O3 0.14 0.09 0.04 0.03 0.00 0.05 0.10 0.00 0.04 0.00 0.07 0.00 0.06 0.00 0.09 0.00FeO 31.90 31.04 32.02 30.41 30.66 30.48 31.23 31.91 30.99 31.28 31.28 31.02 31.24 30.64 30.59 30.81MnO 0.63 0.74 0.42 0.38 0.52 0.61 0.43 0.39 0.50 0.34 0.37 0.43 0.38 0.70 0.54 0.70MgO 9.32 9.44 9.83 9.26 9.77 9.02 9.36 9.57 9.06 8.87 9.44 9.44 9.23 9.09 9.45 9.42CaO 0.68 0.80 0.79 0.66 0.78 0.71 0.74 0.57 0.73 0.79 1.24 1.22 1.33 1.28 1.10 0.63Total 101.65 100.79 100.07 100.50 100.85 100.99 100.88 100.20 100.72 100.73 100.14 100.67 100.83 100.31 100.54 100.30Formula quantities normalized to 12 OSi 2.961 2.971 2.935 2.981 2.987 2.983 3.019 2.981 2.987 2.996 2.987 3.002 2.973 2.987 3.025 2.973Ti 0.001 0.002 0.005 0.002 0.000 0.000 0.000 0.002 0.011 0.000 0.005 0.003 0.006 0.000 0.000 0.001Al 1.957 1.978 2.014 2.009 1.977 1.989 1.913 1.947 1.964 1.951 1.940 1.927 1.985 1.976 1.923 2.011Cr 0.008 0.005 0.002 0.002 0.000 0.003 0.006 0.000 0.002 0.000 0.004 0.000 0.003 0.000 0.005 0.000Fe 1.989 1.926 1.946 1.899 1.864 1.916 1.954 1.970 1.938 1.979 1.933 1.924 1.916 1.908 1.897 1.891Mn 0.040 0.047 0.026 0.024 0.032 0.039 0.027 0.024 0.032 0.022 0.023 0.027 0.024 0.044 0.034 0.044Mg 1.035 1.044 1.065 1.030 1.059 1.010 1.044 1.053 1.009 1.000 1.040 1.043 1.009 1.009 1.044 1.030Ca 0.054 0.064 0.062 0.053 0.061 0.057 0.059 0.045 0.058 0.064 0.098 0.097 0.104 0.102 0.087 0.050XMg 0.332 0.339 0.344 0.343 0.351 0.334 0.338 0.340 0.332 0.326 0.336 0.338 0.330 0.329 0.341 0.342XCa 0.017 0.021 0.020 0.018 0.020 0.019 0.019 0.015 0.019 0.021 0.032 0.031 0.034 0.033 0.029 0.016XMn 0.013 0.015 0.008 0.008 0.011 0.013 0.009 0.008 0.010 0.007 0.007 0.009 0.008 0.014 0.011 0.014


Traverse through the garnet grain (GRT-II) from the trondhjemite (Fig. 6c).

Components 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

SiO2 38.81 39.12 39.32 39.25 37.33 37.58 38.25 38.30 38.24 38.12 37.79 35.93 38.19 38.74 38.32 38.30 37.61TiO2 0.00 0.00 0.01 0.00 0.01 0.03 0.00 0.00 0.00 0.02 0.00 0.03 0.00 0.10 0.03 0.04 0.08Al2O3 21.25 22.46 21.69 22.40 21.09 20.77 21.46 21.44 20.99 21.60 21.57 28.04 21.82 22.39 21.80 21.59 21.77Cr2O3 0.00 0.00 0.00 0.00 0.00 0.13 0.02 0.00 0.00 0.00 0.00 0.00 0.13 0.00 0.00 0.03 0.20FeO 29.71 28.88 29.02 29.38 29.76 30.20 30.08 30.66 31.00 29.79 30.74 24.88 29.97 29.76 29.35 29.00 28.98MnO 0.34 0.34 0.50 0.85 0.51 0.29 0.55 0.32 0.20 0.51 0.41 0.26 0.35 0.40 0.55 0.35 0.46MgO 8.91 9.25 8.87 8.78 8.51 8.17 8.65 7.97 7.60 8.21 8.18 7.38 8.83 9.05 9.13 8.83 8.79CaO 0.67 0.65 0.70 0.71 0.92 0.71 0.70 0.70 0.89 0.76 0.75 0.62 0.75 0.79 0.53 0.70 0.79Total 99.69 100.70 100.11 101.37 98.13 97.91 99.78 99.54 99.00 99.02 99.83 97.85 100.13 101.24 99.74 99.08 98.73Formula quantities normalized to 12 OSi 3.018 2.993 3.032 2.996 2.968 2.999 2.986 3.001 3.019 2.993 2.962 2.788 2.969 2.966 2.978 2.994 2.961Ti 0.000 0.000 0.001 0.000 0.001 0.002 0.000 0.000 0.000 0.001 0.000 0.002 0.000 0.006 0.002 0.002 0.005Al 1.947 2.025 1.971 2.015 1.976 1.953 1.974 1.980 1.953 1.998 1.992 2.564 1.999 2.020 1.996 1.989 2.019Cr 0.000 0.000 0.000 0.000 0.000 0.008 0.001 0.000 0.000 0.000 0.000 0.000 0.008 0.000 0.000 0.002 0.012Fe 1.932 1.847 1.870 1.875 1.978 2.014 1.963 2.008 2.046 1.955 2.014 1.614 1.948 1.904 1.907 1.895 1.907Mn 0.022 0.022 0.033 0.055 0.034 0.020 0.036 0.021 0.013 0.034 0.027 0.017 0.023 0.026 0.036 0.023 0.031Mg 1.032 1.054 1.019 0.998 1.008 0.971 1.006 0.930 0.894 0.960 0.955 0.853 1.023 1.032 1.057 1.028 1.031Ca 0.056 0.053 0.058 0.058 0.078 0.061 0.059 0.059 0.075 0.064 0.063 0.052 0.062 0.065 0.044 0.059 0.067XMg 0.339 0.354 0.342 0.334 0.325 0.317 0.328 0.308 0.295 0.319 0.312 0.336 0.335 0.341 0.347 0.342 0.340XCa 0.018 0.018 0.019 0.019 0.025 0.020 0.019 0.019 0.025 0.021 0.021 0.020 0.020 0.021 0.014 0.020 0.022XMn 0.007 0.007 0.011 0.018 0.011 0.006 0.012 0.007 0.004 0.011 0.009 0.007 0.008 0.009 0.012 0.008 0.010



Table 2 (Continued)

Compositions of garnets from the trondhjemite-metapelite reaction zone.

Components 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

SiO2 36.45 37.56 37.58 37.89 37.54 38.53 37.40 38.08 38.28 37.74 38.00 37.58 37.55 38.82 37.71 37.94 37.70TiO2 0.00 0.02 0.00 0.00 0.01 0.00 0.00 0.06 0.02 0.01 0.03 0.25 0.26 0.09 0.00 0.00 0.05Al2O3 0.07 0.07 0.04 0.09 0.01 0.13 0.00 0.12 0.10 0.14 0.07 0.10 0.09 0.12 0.00 0.00 0.00Cr2O3 20.73 21.26 21.37 21.48 21.24 21.79 21.45 21.64 21.68 21.64 21.22 21.31 21.51 21.73 21.05 21.43 21.41FeO 29.65 28.97 28.45 28.62 29.23 30.02 29.62 29.91 29.59 29.36 29.59 31.34 31.29 31.10 28.36 28.08 28.17MnO 0.35 0.45 0.37 0.41 0.45 0.44 0.45 0.33 0.29 0.31 0.38 0.38 0.33 0.35 0.42 0.40 0.35MgO 7.93 8.61 8.68 8.64 8.46 8.26 8.20 7.84 8.45 8.54 8.04 7.07 6.96 7.07 8.34 8.42 8.59CaO 0.57 0.57 0.92 0.76 0.58 0.59 0.60 0.91 0.63 0.70 0.95 0.92 0.84 0.95 1.26 1.36 1.27Total 99.74 99.51 99.41 99.90 99.52 99.75 99.72 98.89 99.05 98.43 99.29 98.96 98.80 100.24 99.14 99.65 99.54Formula quantities normalized to 12 OSi 2.895 2.983 2.985 3.010 2.981 3.060 2.971 3.024 3.041 2.997 3.018 2.985 2.982 3.084 2.995 3.014 2.994Ti 0.000 0.001 0.000 0.000 0.001 0.000 0.000 0.004 0.001 0.001 0.002 0.015 0.016 0.005 0.000 0.000 0.003Al 1.937 1.986 1.997 2.008 1.985 2.036 2.004 2.022 2.026 2.022 1.983 1.991 2.010 2.031 1.967 2.003 2.001Cr 0.004 0.004 0.002 0.006 0.001 0.008 0.000 0.007 0.006 0.009 0.005 0.006 0.005 0.008 0.000 0.000 0.000Fe 1.968 1.923 1.888 1.899 1.940 1.992 1.966 1.985 1.964 1.948 1.964 2.080 2.077 2.064 1.883 1.864 1.870Mn 0.023 0.030 0.025 0.028 0.031 0.029 0.030 0.022 0.020 0.021 0.026 0.025 0.022 0.024 0.028 0.027 0.023Mg 0.937 1.018 1.026 1.022 1.000 0.977 0.970 0.927 0.999 1.010 0.951 0.837 0.822 0.836 0.987 0.996 1.015Ca 0.049 0.048 0.078 0.065 0.049 0.050 0.051 0.078 0.054 0.060 0.081 0.078 0.071 0.081 0.108 0.116 0.108XMg 0.315 0.337 0.340 0.339 0.331 0.320 0.322 0.308 0.329 0.332 0.315 0.277 0.275 0.278 0.328 0.332 0.337XCa 0.016 0.016 0.026 0.022 0.016 0.016 0.017 0.026 0.018 0.020 0.027 0.026 0.024 0.027 0.036 0.039 0.036XMn 0.008 0.010 0.008 0.009 0.010 0.010 0.010 0.007 0.006 0.007 0.009 0.008 0.007 0.008 0.009 0.009 0.008

Note: 1–14, traverse through garnet grain in Fig. 5a; 15–17, garnet associated with sillimanite forming euhedral facets at contact of garnet grains with plagioclase (Fig. 5e).

XCa and XMn (Fig. 7a). Since these garnets are plunged into thequartz-feldspathic matrix, the increase of XMg within the marginal400–450 �m (Fig. 7a) reflects the growth zoning rather than diffus-ional Fe–Mg exchange zoning. The zoning of GRT-I with inclusionsof muscovite and zoisite (Fig. 3d) with respect to XMg is unclear(Fig. 7b). Such near-sinusoidal profile can be related to late modi-fications of the garnet along the cracks or initial complexity of thegrains accreted from several smaller grains. Nevertheless, the trav-erse (Fig. 6b) shows clear zoning of the garnet with respect to XCa:the garnet zones containing inclusions of Mu and Zo show lowerCa content with respect to the inclusion-free garnet zones.

Large GRT-II grain (Fig. 3h) shows a complex zoning withthe general decrease of XMg down to 0.29 at the centralportions of the grain (Fig. 7c). The central portion includessheaf-like inclusions of sillimanite and polyphase inclusionssillimanite + corundum + spinel + late biotite. Composition of theperipheral zones with inclusions of biotite and muscovite is gener-ally similar to the composition of GRT-I.

Central zones of garnets from the trondhjemite-metapelite reac-tion zone show XMg = 0.34–0.35, that is slightly higher than XMgin the central portions of the GRT-I in the trondhjemite. How-ever, XCa and XMn in these garnets are very similar to those in thetrondhjemite. Large garnet grains corroded by the biotite show bell-shaped zoning pattern with decrease of XMg down to 0.31–0.29toward the contacts with the biotite shells. The lowest XMg (0.29)has been measured in garnet forming intergrowths with biotitecrowded with inclusions of ilmenite and rutile (Fig. 5d). The char-acteristic feature of the late garnet forming euhedral facets alongthe contacts of large garnet grains with plagioclase accompaniedby sillimanite veinlets (Fig. 5e) is higher XCa varying from 0.03 to0.045. Similar XCa is measured in garnets associated with K-feldsparand the Pl-Qtz textures (Fig. 5f).

Cordierites from both studied metapelites, SAF12-2/2 and PET-5,show relatively narrow range of XMg = 0.78–0.84 (Table 3). Nev-ertheless, this parameter regularly changes within this range.Inclusions of cordierite in large grains of orthopyroxene from themetapelite SAF12-2/2 (Fig. 4b) show the lowest XMg = 0.78–0.81(Table 3). This parameter increases in cordierite located atboundaries of orthopyroxene and reaches 0.83–0.84 for sep-arate cordierite grains within the quartz-plagioclase matrix(Fig. 4c). Mg-number of cordierite in the metapelite PET-5 slightlyincreases from centers of grains toward the contacts with the

Opx + Sil textures and Sil + Qtz veinlets crossing cordierite grains(Table 3).

Mg-numbers of the relict cordierite in the trondhjemite-metapelite reaction zone reaches 0.84–0.85. These values do notdepend on a degree of alteration of the relict cordierite aggre-gated with biotite + sillimanite and, thus, represent the initial XMgof cordierite involved in the interaction between trondhjemite andmetapelite. The XMg values are clearly higher than XMg of cordieritefrom metapelite suggesting re-equilibration of the cordierite withgarnet in the reaction zone.

Biotite in metapelite SAF12-2/2 and PET-5 can be clearly sub-divided with respect to XMg, XAl and the Ti content. Fig. 8a and bdemonstrate that biotite coexisting with orthopyroxene either aslarge flakes or as inclusions, i.e. can be assumed as primary in therock, are characterized by lower XMg and XAl, but the higher Ti con-tent in comparison to the biotite associated with late assemblagesGed + Sil + Qtz ± St (in SAF12-2/2) or Opx + Sil (in PET-5) (Table 4a). It is important to note that the XMg values of the biotite inclu-sions in orthopyroxene from the metapelite SAF12-2/2 show a widescatter, while XMg of large biotite flakes associated with orthopy-roxene is located within the narrow range about 0.66–0.67 (Fig. 8aand b) at similar XAl and the Ti content. This example clearly showsthat the inclusions were affected by strong Fe–Mg (and, possi-bly, slight Al) re-equilibration in contrast to the centers of largeflakes. Chlorine content of biotite from metapelite is almost belowdetection limit. Biotite from assemblages with gedrite is slightlyricher in F (0.29–0.33 wt.%) than large (primary) flakes of biotite(0.25–0.27 wt.%) (Table 4a).

In general, biotite from the garnet trondhjemite and thetrondhjemite-metapelite reaction zone show higher XMg in com-parison to biotite from the metapelite (Fig. 8c and d; Table 4b andc). There is practically no overlapping between compositions ofbiotite from the trondhjemite-metapelite reaction zone and com-positions of biotite from the metapelite. Thus, even if some relicsof metapelitic biotite were present in the trondhjemite-metapelitereaction zone, they were efficiently re-equilibrated with garnet. Atsimilar XAl and Ti contents, centers of large biotite flakes show muchnarrower scatter of XMg in comparison to biotite inclusions in gar-net and biotite contacting to garnet (Fig. 8c and d). Again, it is clearlyattributed to Fe–Mg re-equilibration of biotite in contact with gar-net in contrast to the centers of large flakes. Latest generations ofbiotite in the garnet trondhjemite and the trondhjemite-metapelite


Fig. 7. Compositional traverses through the garnet grains from the trondhjemite. (a) Variations of XMg, XCa and XMn in the grain of GRT-I; shaded area indicates the portion of thegarnet with small inclusions of biotite, muscovite and quartz (Fig. 3f), other portion of the grain is inclusion-free. (b) Center-to-rim variations of XMg, XCa and XMn in the grain ofGRT-I with inclusions of muscovite and zoisite (Fig. 3d); shaded area indicates the portion of the garnet crowded with inclusions, other portion of the grain is inclusion-free. (c)Variations of XMg, XCa and XMn in the large grain of GRT-II (Fig. 3h) containing inclusions of sillimanite and polyphase inclusions sillimanite + corundum + spinel + (late biotite)in the core portion and inclusions of biotite and muscovite in the peripheral portions; 1, zone with the large polyphase inclusion sillimanite + corundum + spinel + (late biotite)(Fig. 3i); 2, zones with sillimanite inclusions (+late biotite); 3, zone with another polyphase inclusion sillimanite + corundum + spinel + (late biotite) (Fig. 3k); 4, peripheralzone with muscovite inclusions (Fig. 3l); 5, peripheral zone with biotite inclusions.

reaction zone are characterized by low Ti contents, down to zeroin biotite replacing relict cordierite (Table 4b and c). In partic-ular, this observation implies that biotite developing within thesillimanite-corundum-spinel polyphase inclusions and at contactsof sillimanite inclusions in GRT-II (Fig. 3i and k–m) are late andrelated to the late interaction of host garnet with sillimanite, corun-dum and spinel.

Orthopyroxene from the metapelite SAF12-2/2 shows a rela-tively wide variations of XMg (0.55–0.61) and constant XAl at about

0.040–0.045 (Table 5, Fig. 9). The lowest XMg (0.55–0.57) is charac-teristic for the centers of large grains, whereas peripheral zones ofthe grains show increase of XMg up to 0.59–0.60. The highest XMgis detected in orthopyroxene contacting with gedrite-sillimanite-biotite-quartz reaction textures (Table 5). Orthopyroxene in themetapelite PET-5 are slightly more Mg-rich (XMg = 0.57–0.63) andAl-rich (XAl = 0.05–0.06). Orthopyroxenes forming intergrowthswith sillimanite after cordierite (Fig. 4h) show the highest XAl (up to0.08), whereas their XMg reaches 0.63 (Fig. 9). Relict orthopyroxene


Table 3Representative analyses of cordierites from metapelites SAF12-2/2 (1–8), PET-5 (9–13) and the relict cordierite from the trondhjemite-metapelite reaction zone (14–16).

Components 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

SiO2 48.78 49.06 49.68 49.18 49.24 49.06 49.51 49.14 53.14 45.79 49.96 49.34 48.83 48.29 47.80 48.12Al2O3 32.54 32.53 33.26 32.69 32.73 32.45 32.52 32.88 35.44 30.79 37.04 33.67 33.95 33.21 32.92 32.93FeO 4.35 4.65 5.38 4.27 4.19 3.64 3.95 3.70 4.54 4.62 3.54 3.76 3.59 3.57 4.00 3.97MnO 0.05 0.06 0.07 0.02 0.03 0.04 0.06 0.02 0.00 0.02 0.05 0.01 0.11 0.04 0.00 0.00MgO 10.46 10.44 10.58 10.51 10.58 10.89 10.76 10.87 11.14 9.63 10.32 10.95 11.41 11.35 10.82 10.99Na2O 0.08 0.09 0.11 0.07 0.08 0.07 0.07 0.07 0.00 0.03 0.04 0.19 0.04 0.13 0.11 0.09Total 96.26 96.83 99.08 96.74 96.85 96.15 96.87 96.68 104.26 90.88 100.95 97.92 97.93 96.58 95.65 96.10Formula quantities normalized to 18 OSi 5.034 5.041 5.008 5.046 5.045 5.051 5.065 5.032 5.041 5.014 4.890 4.994 4.937 4.958 4.963 4.972Al 3.957 3.938 3.951 3.952 3.951 3.937 3.920 3.968 3.962 3.973 4.272 4.016 4.045 4.017 4.027 4.009Fe 0.375 0.399 0.453 0.366 0.359 0.313 0.338 0.317 0.360 0.423 0.290 0.318 0.303 0.307 0.347 0.343Mn 0.004 0.005 0.006 0.002 0.003 0.003 0.005 0.002 0.000 0.002 0.004 0.001 0.009 0.003 0.000 0.000Mg 1.608 1.598 1.589 1.606 1.615 1.670 1.640 1.658 1.574 1.571 1.505 1.651 1.718 1.736 1.674 1.691Na 0.016 0.018 0.021 0.014 0.016 0.014 0.014 0.014 0.000 0.006 0.008 0.037 0.008 0.025 0.022 0.018XMg 0.809 0.798 0.776 0.814 0.817 0.841 0.827 0.839 0.814 0.787 0.837 0.838 0.846 0.849 0.828 0.831

Note: Metapelite SAF12-2/2. 1–3, inclusions in orthopyroxene partially replaced by the Ged + Sil(Ky) + Bt ± St aggregates (Fig. 4b); 4 and 5, grains initially contacting toorthopyroxene partially replaced by the Ged + Sil(Ky) + Bt ± St aggregates (Fig. 4b); 6 and 7, cordierite partially replaced by the Ged + Sil(Ky) + Bt ± St aggregates in the quartz-plagioclase matrix (Fig. 4c); 8, cordierite rimmed with plagioclase (with delicate inclusions of St and Sil; see Fig. 4d). Metapelite PET-5. 9 and 10, centers of cordierite grains inthe cordierite-orthopyroxene clusters (Fig. 4g); 11 and 12, cordierite contacting to the Opx + Sil textures (Fig. 4h); 13, cordierite at contact with the Opx + Bt + Qtz intergrowths(Fig. 4h).

from the trondhjemite-metapelite reaction zone differs from themetapelitic orthopyroxene by clearly higher XAl (0.060–0.075) atcomparable XMg (Table 5; Fig. 9). The lowest XAl and highest XMgare measured in orthopyroxene contacting with the biotite-quartzsymplectites (Table 5).

Cores of orthopyroxenes from the leucocratic orthopyroxeneand garnet-bearing enclaves from metapelites is characterized byelevated XAl (up to 0.08) at XMg = 0.58–0.60 (Fig. 9).

Muscovite included in garnet from the trondhjemite arecharacterized by an elevated Si content (Table 6), which

Fig. 8. Compositional variations of biotite in metapelite, garnet trondhjemite and the trondhjemite-metapelite reaction zone. (a) Variations of XMg and XAl in biotite frommetapelite SAF12-2/2 (1, centers of large biotite flakes associated with orthopyroxene; 2, biotite inclusions in orthopyroxene; 3, biotite associated with the assemblageGed + Sil(Ky) + Qtz ± St) and PET-5 (1, large biotite and biotite from the orthopyroxene-cordierite clusters; 2, late biotite associated with the assemblage Opx + Sil). (b)Variations of XMg and Ti content in biotite from metapelite SAF12-2/2 and PET-5 (data point markers see in figure a). (c) Variations of XMg and XAl in biotite from thetrondhjemite-metapelite reaction zone and the garnet trondhjemite. 1, centers of large biotite flakes building shells around garnet in the trondhjemite-metapelite reactionzone; 2, inclusions in garnet in the trondhjemite-metapelite reaction zone; 3, contacts with garnet in the trondhjemite-metapelite reaction zone; 4, biotite from symplectiteswith quartz after garnet and orthopyroxene in the trondhjemite-metapelite reaction zone; 5, late biotite replacing cordierite relics in the trondhjemite-metapelite reactionzone; 6, inclusions in GRT-I from the trondhjemite; 7, biotite associated with sillimanite-corundum-spinel polyphase inclusions in GRT-II.


Table 4(a) Representative analyses of biotites from metapelites PET-5 (1–3) and SAF12-2/2 (4–15).

Components 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

SiO2 37.53 37.20 37.29 37.88 38.21 37.87 37.49 39.00 39.28 39.50 38.65 35.86 36.16 36.64 35.73TiO2 5.66 5.56 4.59 3.42 3.40 3.59 4.01 1.57 0.59 1.03 2.57 3.31 3.29 3.60 3.55Al2O3 15.71 16.02 16.78 17.37 17.52 17.22 17.05 18.48 19.60 19.28 17.88 17.62 17.50 17.45 17.28FeO 13.47 14.29 12.46 13.03 13.34 13.26 13.18 12.40 11.91 11.72 10.93 12.42 11.70 14.48 13.81MnO 0.00 0.15 0.08 0.00 0.00 0.03 0.02 0.03 0.04 0.03 0.02 0.00 0.08 0.03 0.03MgO 13.69 13.36 14.97 14.75 14.84 14.71 14.20 16.31 16.96 16.63 16.75 15.53 15.73 13.96 13.89CaO 0.07 0.03 0.00 0.00 0.01 0.01 0.00 0.00 0.01 0.02 0.02 0.03 0.07 0.02 0.05Na2O 0.01 0.02 0.23 0.17 0.17 0.18 0.16 0.19 0.18 0.22 0.14 0.31 0.17 0.49 0.50K2O 10.28 10.25 10.47 8.88 8.37 8.47 9.12 8.39 7.18 7.62 8.45 9.41 9.51 9.16 9.03Cl n.d. n.d. 0.00 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.02 0.03 0.03 0.00 0.00F n.d. n.d. 0.00 0.25 0.27 0.25 0.26 0.33 0.30 0.29 0.33 n.d. n.d. n.d. n.d.Total 97.81 97.64 97.57 96.38 96.73 96.20 96.09 96.83 96.21 96.47 96.10 95.12 94.59 96.56 94.55Formula quantities normalized to 11 OSi 2.733 2.719 2.706 2.760 2.766 2.761 2.750 2.797 2.801 2.813 2.783 2.659 2.684 2.693 2.679Ti 0.310 0.305 0.250 0.187 0.185 0.197 0.221 0.085 0.032 0.055 0.139 0.184 0.184 0.199 0.200Al 1.348 1.380 1.435 1.491 1.495 1.479 1.474 1.562 1.647 1.618 1.517 1.540 1.531 1.511 1.527Fe 0.820 0.873 0.756 0.794 0.807 0.808 0.808 0.743 0.710 0.698 0.658 0.770 0.726 0.889 0.865Mn 0.000 0.009 0.005 0.000 0.000 0.002 0.001 0.002 0.002 0.002 0.001 0.000 0.005 0.002 0.002Mg 1.485 1.454 1.618 1.601 1.600 1.598 1.551 1.743 1.802 1.764 1.797 1.716 1.739 1.529 1.552Ca 0.005 0.002 0.000 0.000 0.001 0.001 0.000 0.000 0.001 0.002 0.002 0.002 0.006 0.002 0.004Na 0.001 0.003 0.032 0.024 0.024 0.025 0.023 0.026 0.025 0.030 0.020 0.044 0.025 0.069 0.072K 0.955 0.955 0.969 0.825 0.773 0.788 0.853 0.767 0.653 0.692 0.776 0.890 0.900 0.859 0.863Cl – – 0.000 0.001 0.001 0.001 0.000 0.000 0.000 0.000 0.002 0.003 0.004 0.000 0.000F – – 0.000 0.058 0.062 0.058 0.060 0.075 0.068 0.065 0.075 – – – –XMg 0.644 0.622 0.680 0.669 0.665 0.664 0.657 0.700 0.717 0.716 0.732 0.690 0.704 0.632 0.642XAl 0.201 0.205 0.212 0.218 0.218 0.216 0.217 0.225 0.235 0.233 0.220 0.224 0.223 0.222 0.224

Note: Metapelite PET-5. 1, large biotite in the recrystallized and coarsened portion of the rock; 2, biotite from the orthopyroxene-cordierite clusters; 3, biotiteintergrown with orthopyroxene and quartz at cordierite. Metapelite SAF12-2/2. 4–7, large biotite flakes at orthopyroxene; 8–11, biotite associated with gedrite,sillimanite (kyanite), staurolite and quartz; 12–15, inclusions in orthopyroxene.

(b) Representative analyses of biotites from the trondhjemite-metapelite reaction zone.

Components 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

SiO2 36.39 36.01 36.01 35.57 35.33 36.26 36.25 36.14 36.97 36.60 35.66 36.00 34.07 36.52 34.00 36.77TiO2 4.50 4.13 4.34 4.55 4.05 3.70 4.57 4.65 4.05 3.91 4.36 3.44 1.11 2.29 0.45 0.08Al2O3 17.19 17.16 17.31 16.92 16.83 18.11 16.89 16.58 17.69 17.38 16.89 17.05 15.90 17.48 16.82 19.65FeO 11.59 11.35 10.39 11.09 11.65 10.01 11.23 12.12 11.59 11.88 11.89 11.94 14.89 11.89 11.07 8.26MnO 0.00 0.00 0.05 0.04 0.03 0.05 0.02 0.00 0.04 0.00 0.01 0.00 0.13 0.00 0.00 0.00MgO 15.10 15.46 16.13 15.63 15.31 16.51 15.57 15.15 15.85 15.76 15.21 15.52 17.34 16.32 17.19 18.95CaO 0.04 0.00 0.03 0.00 0.02 0.06 0.01 0.05 0.01 0.07 0.00 0.06 0.09 0.03 0.09 0.05Na2O 0.33 0.27 0.32 0.22 0.19 0.57 0.34 0.16 0.20 0.24 0.19 0.24 0.00 0.16 0.00 0.14K2O 9.54 9.60 9.61 9.61 9.47 9.43 9.56 9.72 9.66 9.70 9.67 9.68 7.93 9.72 8.33 9.42Cl 0.00 0.00 0.00 0.00 0.02 0.00 0.00 0.00 0.01 0.00 0.00 0.01 0.00 0.00 0.00 0.02F n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 0.00 n.d. 0.00 n.d.Total 95.06 94.35 94.55 93.95 93.30 94.90 94.94 94.85 96.45 95.88 94.13 94.06 92.00 94.62 88.60 93.31Formula quantities normalized to 11 OSi 2.686 2.678 2.661 2.658 2.666 2.659 2.679 2.685 2.685 2.682 2.668 2.694 2.642 2.709 2.688 2.702Ti 0.250 0.231 0.241 0.255 0.230 0.204 0.254 0.260 0.221 0.216 0.245 0.194 0.065 0.128 0.027 0.004Al 1.495 1.503 1.507 1.490 1.496 1.565 1.471 1.452 1.514 1.500 1.490 1.503 1.453 1.528 1.567 1.702Fe 0.715 0.706 0.642 0.692 0.735 0.614 0.694 0.753 0.703 0.728 0.744 0.747 0.965 0.737 0.732 0.507Mn 0.000 0.000 0.003 0.003 0.002 0.003 0.001 0.000 0.003 0.000 0.000 0.000 0.009 0.000 0.000 0.000Mg 1.661 1.712 1.776 1.740 1.721 1.804 1.714 1.677 1.714 1.720 1.695 1.730 2.003 1.803 2.024 2.074Ca 0.003 0.000 0.003 0.000 0.002 0.004 0.001 0.004 0.001 0.005 0.000 0.005 0.007 0.002 0.008 0.004Na 0.047 0.040 0.046 0.032 0.028 0.082 0.048 0.023 0.028 0.034 0.028 0.034 0.000 0.023 0.000 0.020K 0.898 0.910 0.905 0.916 0.911 0.882 0.901 0.921 0.895 0.907 0.923 0.923 0.784 0.919 0.840 0.883Cl 0.000 0.000 0.000 0.000 0.002 0.000 0.000 0.000 0.001 0.000 0.000 0.002 0.000 0.000 0.000 0.002F 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000XMg 0.700 0.708 0.734 0.714 0.700 0.745 0.712 0.690 0.708 0.703 0.695 0.699 0.673 0.710 0.735 0.804XAl 0.220 0.220 0.221 0.218 0.218 0.229 0.216 0.213 0.221 0.219 0.218 0.219 0.204 0.221 0.223 0.243

Note: 1–3, centers of biotites forming shells around garnets; 4–6, inclusions in garnets; 7 and 8, contacts with garnets; 9 and 10, associated with ilmenite andrutile at contacts with garnets; 11 and 12, in symplectites with quartz after garnets; 13 and 14, in late symplectites with quartz after orthopyroxenes; 15, latebiotite associated with Ca-rich garnet and K-feldspar; 16, late biotite replacing cordierite relics.

(c) Representative analyses of biotites from the garnet trondhjemite.

Components 1 2 3 4 5 6 7 8 9 10

SiO2 39.14 38.58 37.85 36.01 36.21 36.96 37.15 36.76 36.16 36.28TiO2 4.18 4.17 2.81 3.35 1.96 1.46 0.64 1.55 1.22 0.40Al2O3 18.58 17.86 17.92 18.75 19.31 18.14 18.57 18.47 18.80 20.36FeO 9.75 8.81 9.38 10.82 10.53 9.84 9.67 10.78 10.82 9.68MnO 0.00 0.03 0.01 0.04 0.18 0.00 0.02 0.00 0.00 0.13MgO 17.56 16.95 17.33 16.08 17.33 17.22 17.44 17.31 15.94 18.80CaO 0.02 0.00 0.03 0.06 0.43 0.04 0.03 0.01 0.00 0.09Na2O 0.76 0.63 0.38 0.34 0.46 0.45 0.32 0.35 0.34 0.22


Table 4 (Continued )

(c) Representative analyses of biotites from the garnet trondhjemite.

Components 1 2 3 4 5 6 7 8 9 10

K2O 10.17 10.60 8.80 8.23 7.55 9.33 9.28 9.59 9.46 7.53Cl 0.00 0.07 0.03 0.01 0.00 0.00 0.11 0.12 0.00 0.06F 0.00 0.00 0.36 0.10 0.00 0.20 0.00 0.00 0.00 0.00Total 100.80 97.90 95.07 93.87 94.06 93.63 93.29 94.90 92.82 93.56Formula quantities normalized to 11 OSi 2.703 2.734 2.749 2.659 2.654 2.731 2.748 2.697 2.711 2.651Ti 0.217 0.222 0.153 0.186 0.108 0.081 0.036 0.085 0.069 0.022Al 1.512 1.491 1.534 1.631 1.667 1.579 1.619 1.597 1.661 1.753Fe 0.563 0.522 0.570 0.668 0.645 0.608 0.598 0.661 0.678 0.591Mn 0.000 0.002 0.001 0.003 0.011 0.000 0.001 0.000 0.000 0.008Mg 1.807 1.789 1.875 1.769 1.892 1.895 1.922 1.892 1.780 2.046Ca 0.001 0.000 0.002 0.005 0.034 0.003 0.002 0.001 0.000 0.007Na 0.102 0.087 0.053 0.049 0.065 0.064 0.046 0.050 0.049 0.031K 0.896 0.958 0.815 0.775 0.706 0.879 0.876 0.897 0.905 0.702Cl 0.000 0.008 0.004 0.001 0.000 0.000 0.000 0.000 0.000 0.000F 0.000 0.000 0.083 0.023 0.000 0.000 0.000 0.000 0.000 0.000XMg 0.762 0.774 0.767 0.725 0.742 0.757 0.762 0.741 0.724 0.773XAl 0.222 0.221 0.223 0.236 0.239 0.229 0.234 0.230 0.241 0.248

Note: 1–4, inclusions in GRT-I; 5, inclusion in the outer zone of GRT-II; 6–8, late biotite developed at the contacts of the sillimanite-corundum-spinel polyphase inclusionswith host garnet; 9 and 10, late biotite intergrown with sillimanite and muscovite.

Table 5Representative analyses of orthopyroxenes from the metapelite SAF12-2/2 (1–6), the metapelite PET-5 (7–12) and the trondhjemite-metapelite reaction zone (13–16).

Components 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

SiO2 50.07 50.75 50.58 50.49 50.09 50.06 51.50 51.04 52.63 50.54 50.84 50.12 48.70 48.27 48.10 49.53TiO2 0.07 0.08 0.07 0.06 0.09 0.07 0.14 0.19 0.25 0.06 0.20 0.01 0.13 0.08 0.05 0.08Al2O3 3.50 3.53 3.60 3.91 3.95 4.00 4.38 5.18 4.89 5.92 6.23 9.16 6.72 6.53 6.31 5.38FeO 26.06 25.73 26.30 24.60 25.03 24.64 22.81 23.64 25.11 23.30 23.44 21.42 23.58 23.58 24.05 22.99MnO 0.40 0.35 0.41 0.37 0.30 0.37 0.24 0.28 0.36 0.30 0.23 0.27 0.10 0.06 0.15 0.24MgO 18.92 18.97 18.68 20.41 19.26 19.75 21.67 19.73 20.98 20.99 20.77 19.83 20.14 20.38 19.63 20.83CaO 0.09 0.09 0.09 0.09 0.09 0.08 0.00 0.16 0.00 0.00 0.08 0.09 0.13 0.07 0.05 0.11Na2O 0.00 0.01 0.01 0.00 0.01 0.00 0.22 0.08 0.00 0.20 0.00 0.00 0.22 0.11 0.14 0.17Total 99.31 99.71 99.91 100.15 99.04 99.17 100.96 100.30 104.22 101.31 101.79 100.90 99.87 99.17 98.63 99.33Formula quantities normalized to 6 OSi 1.915 1.926 1.921 1.900 1.910 1.904 1.899 1.900 1.893 1.862 1.862 1.831 1.830 1.827 1.837 1.864Ti 0.002 0.002 0.002 0.002 0.003 0.002 0.004 0.005 0.007 0.002 0.006 0.000 0.004 0.002 0.001 0.002Al 0.158 0.158 0.161 0.173 0.178 0.179 0.190 0.227 0.207 0.257 0.269 0.394 0.298 0.291 0.284 0.239Fe 0.833 0.816 0.835 0.774 0.798 0.783 0.703 0.736 0.755 0.718 0.718 0.654 0.741 0.746 0.768 0.723Mn 0.013 0.011 0.013 0.012 0.010 0.012 0.007 0.009 0.011 0.009 0.007 0.008 0.003 0.002 0.005 0.008Mg 1.078 1.072 1.057 1.144 1.094 1.119 1.190 1.094 1.124 1.152 1.133 1.079 1.127 1.149 1.117 1.168Ca 0.004 0.004 0.004 0.004 0.004 0.003 0.000 0.006 0.000 0.000 0.003 0.004 0.005 0.003 0.002 0.004Na 0.000 0.001 0.001 0.000 0.001 0.000 0.016 0.006 0.000 0.014 0.000 0.000 0.016 0.011 0.008 0.011XMg 0.560 0.564 0.555 0.593 0.575 0.585 0.626 0.595 0.595 0.613 0.610 0.620 0.602 0.606 0.591 0.615XAl 0.039 0.040 0.040 0.043 0.044 0.045 0.048 0.057 0.052 0.064 0.067 0.099 0.074 0.073 0.071 0.060

Note: 1–3, centers of large orthopyroxene grains; 4–6, peripheral zones of large orthopyroxene grains; 7–9, orthopyroxenes from the cordierite-orthopyroxene clusters;10–12, orthopyroxenes from the orthopyroxene-sillimanite intergrowths; 13–15, centers of the relict grains; 16, orthopyroxene at the late biotite-quartz intergrowths.

Fig. 9. Compositional characteristics of orthopyroxene. 1, orthopyroxene fromthe metapelite SAF12-2/2; 2, orthopyroxene from the metapelite PET-5; 3, relictorthopyroxene in the trondhjemite-metapelite reaction zone; 4, orthopyroxenecoexisting with garnet in the Opx + Grt + Pl + Kfs + Qtz enclaves in metapelite.

shows a positive correlation with total Mg + Fe + Mn assum-ing presence of the celadonite component and the isomor-phism 2Al ↔ (Mg + Fe + Mn) + Si (Fig. 10a). The highest Si content(3.4–3.5 a.p.f.u.) is observed for muscovite coexisting with zoisite(Fig. 3d and g). Muscovite associated with quartz and biotite con-tains 3.2–3.3 Si a.p.f.u. Similar Si contents are characteristic formuscovite in the outer zones of GRT-II. The Si content of mus-covite shows a positive correlation with Mg-number (Fig. 10b).Muscovite included in the outer zones of GRT-II contains about0.026 a.p.f.u. Na (Table 6). Nevertheless, all muscovites show up to0.025–0.035 a.p.f.u. of Ca, while the total concentration K + Na + Cais below 1.0 a.p.f.u., suggesting the presence of the pyrophyllitecomponent in the solid solution. In general, compositional varia-tions of muscovite from the trondhjemite are within limits thatare typical for micas from peraluminous granites (Miller et al.,1981; Zane and Rizzo, 1999). Muscovite in the trondhjemite con-tains variable concentrations of F (0.26–0.81 wt.%) and negligibleconcentrations of Cl (Table 6).

Zoisite has been found in association with muscovite inclusionsin two garnet grains from trondhjemite (Fig. 3d and g). The FeO


Table 6Representative analyses of muscovites.

Components 1 2 3 4 5 6 7 8 9 10 11

SiO2 52.41 52.71 52.10 53.25 50.80 47.78 47.48 48.19 48.46 48.67 50.20TiO2 0.18 0.19 0.00 0.00 0.04 0.01 0.07 0.02 0.00 0.21 0.15Al2O3 28.94 28.77 29.97 27.63 30.66 29.82 30.12 29.50 29.28 32.67 32.07FeO 2.56 2.19 2.70 2.40 2.40 2.25 2.42 2.30 3.58 2.00 2.34MnO 0.00 0.00 0.05 0.04 0.03 0.11 0.00 0.00 0.00 0.21 0.09MgO 2.57 2.73 2.22 3.02 2.21 1.79 1.67 2.28 2.04 1.39 1.88CaO 0.40 0.42 0.27 0.39 0.37 0.47 0.32 0.20 0.31 0.38 0.14Na2O 0.00 0.13 0.00 0.18 0.03 0.05 0.10 0.20 0.19 0.20 0.21K2O 10.30 9.98 10.48 10.11 9.91 9.79 9.86 9.88 9.75 10.47 9.95Cl 0.04 0.04 0.03 0.00 0.00 0.00 0.06 0.00 0.00 0.06 0.02F 0.46 0.26 0.37 0.67 0.81 0.61 0.41 0.01 0.27 0.52 0.31Total 98.14 97.51 98.56 97.92 97.36 92.84 92.73 92.58 93.88 96.93 97.36Formula quantities normalized to 11 OSi 3.402 3.421 3.370 3.465 3.326 3.288 3.269 3.301 3.302 3.211 3.271Ti 0.009 0.009 0.000 0.000 0.002 0.001 0.004 0.001 0.000 0.010 0.007Al 2.214 2.200 2.284 2.118 2.365 2.418 2.443 2.381 2.351 2.540 2.462Fe 0.139 0.119 0.146 0.131 0.131 0.129 0.139 0.132 0.204 0.110 0.127Mn 0.000 0.000 0.003 0.002 0.002 0.006 0.000 0.000 0.000 0.012 0.005Mg 0.249 0.264 0.214 0.293 0.216 0.183 0.171 0.233 0.207 0.137 0.182Ca 0.028 0.029 0.019 0.027 0.026 0.035 0.024 0.015 0.023 0.027 0.010Na 0.000 0.016 0.000 0.023 0.004 0.007 0.013 0.027 0.025 0.026 0.027K 0.853 0.826 0.865 0.839 0.827 0.859 0.866 0.863 0.847 0.881 0.827Cl 0.004 0.004 0.003 0.000 0.000 0.000 0.007 0.000 0.000 0.007 0.002F 0.094 0.053 0.076 0.138 0.168 0.133 0.089 0.002 0.058 0.108 0.064

Note: 1–4, muscovite inclusions in garnet associated with zoisite; 5–7, muscovite inclusions in GRT-I; 8–11, muscovite inclusions in the outer zones of GRT-II.

Fig. 10. Compositional characteristics of muscovite included in garnets from thetrondhjemite. (a) Positive correlation of the Si content with the total Mg + Fe + Mn.(b) Positive correlation of the Si content with XMg. 1, muscovite inclusions in GRT-I; 2,muscovite inclusion in the outer zones of GRT-II; 3, muscovite inclusions associatedwith zoisite.

and MgO contents of this phase varies within 3.5–4.9 wt.% and2.6–2.9 wt.%, respectively.

Spinel involved in polyphase inclusions in GRT-II from the gar-net trondhjemite shows XMg = 0.32–0.39, while no clear regularitieswere found in variation of the spinel Mg-number. Following toformulas of spinel, the Fe3+/(Fe3+ + Al) ratio does not exceed 0.08.

Gedrite in the metapelite SAF12-2/2 contains 1.7–2.3 wt.% Na2Ocorresponding to 0.47–0.65 Na atoms per formula unit normalizedto 23 O (Table 7). Sodium positively correlates with Al suggest-ing Si ↔ Na + Al isomorphism along with the Tschermack-typesubstitution (Fig. 11a). Mg-number of gedrite shows a negativecorrelation with the Na and Al contents (Fig. 10b). Rare anthophyl-lite containing about 0.3 wt.% Na2O and 2.3 wt.% Al2O3 has beenanalyzed within the gedrite-sillimanite aggregates. It is associatedwith gedrite containing the lowest concentrations of Na2O (about1.7 wt.%) (Fig. 11b).

Mg-numbers of staurolite in the gedrite-sillimanite texturesafter cordierite in metapelite SAF12-2/2 vary within 0.27–0.32(Table 8), and are similar to Mg-numbers of staurolite associ-ated with the orthopyroxene-sillimanite-corundum intergrowthsin high-aluminous metapelite of SMZ (Belyanin et al., 2010). Themost Mg-rich staurolite is associated with rims of plagioclasearound cordierite (Fig. 4d).

Plagioclase in the metapelite SAF12-2/2 can be definitelysubdivided into two major groups, i.e. rock-forming matrixplagioclase and plagioclase related to the retrograde texturesGed + Sil + Qtz ± Bt ± St. The XCa of matrix plagioclase varies withina narrow interval 0.27–0.29 (Fig. 12; Table 9), while plagioclaseassociated with the retrograde textures show much higher XCa andclearly form two compositional types (Fig. 12). Plagioclase with XCavarying within 0.39–0.45 (Table 9; Fig. 12) usually forms rims onrelics of cordierite within the retrograde textures (see Fig. 4c and d).However, such plagioclase also forms as transitional zones betweencores of matrix plagioclase and plagioclase with extremely highXCa, 0.60–0.75 (Table 9; Fig. 12), which built narrow zones in con-tact with Na-gedrite-bearing reaction textures (example of such azone is pointed by a white arrow in Fig. 4f). Plagioclase from themetapelite PET-5 shows constant XCa at 0.25–0.27.

Plagioclase from garnet trondhjemite and the trondhjemite-metapelite reaction zone forms two textural types: (1) plagioclase


Table 7Representative analyses of gedrites (1–9) and one anthophyllite (10) in the metapelite SAF12-2/2.

Componets 1 2 3 4 5 6 7 8 9 10

SiO2 37.47 38.20 40.92 41.31 42.49 42.47 41.40 42.75 40.40 53.99TiO2 0.05 0.07 0.20 0.06 0.03 0.21 0.11 0.07 0.05 0.29Al2O3 23.56 22.82 19.13 18.23 17.97 17.64 19.26 17.16 20.61 2.28FeO 18.46 19.06 19.14 17.99 18.45 19.80 18.95 19.46 19.03 19.85MnO 0.33 0.26 0.36 0.29 0.24 0.36 0.30 0.31 0.30 0.24MgO 13.34 13.19 14.26 15.16 15.29 13.98 13.90 14.69 13.92 20.34CaO 0.21 0.14 0.44 0.23 0.31 0.38 0.37 0.38 0.28 0.02Na2O 2.25 2.30 1.79 1.72 1.67 1.88 2.06 1.71 2.21 0.27K2O 0.06 0.04 0.02 0.06 0.03 0.04 0.00 0.00 0.00 0.00Total 95.73 96.07 96.25 95.06 96.48 96.76 96.35 96.53 96.80 97.28Formula quantities normalized to 23 OSi 5.568 5.661 6.039 6.140 6.212 6.236 6.094 6.286 5.930 7.763Ti 0.006 0.007 0.022 0.007 0.003 0.023 0.012 0.008 0.006 0.031Al 4.126 3.985 3.327 3.193 3.097 3.052 3.341 2.973 3.564 0.387Fe 2.293 2.360 2.361 2.235 2.255 2.430 2.332 2.392 2.335 2.386Mn 0.042 0.033 0.045 0.036 0.030 0.045 0.037 0.039 0.037 0.030Mg 2.954 2.912 3.134 3.356 3.329 3.058 3.048 3.217 3.043 4.358Ca 0.034 0.023 0.069 0.036 0.049 0.060 0.058 0.060 0.044 0.003Na 0.649 0.659 0.511 0.496 0.472 0.535 0.588 0.487 0.629 0.077K 0.011 0.007 0.003 0.012 0.007 0.007 0.000 0.000 0.000 0.000XMg 0.56 0.55 0.57 0.60 0.59 0.55 0.56 0.57 0.56 0.64

with clear K-feldspar lamellae (Figs. 3a and 5b) and (2)lamellae-free plagioclase. However, lamellae-bearing plagioclaseis relatively rare in the reaction zone in comparison to thetrondhjemite. Lamellae-free plagioclase from the garnet trond-hjemite and the trondhjemite-metapelite reaction zone showsmuch lower XCa, 0.16–0.21 in comparison to plagioclase from

Fig. 11. Compositional variations of gedrite in the metapelite SAF12-2/2. (a) Positivecorrelation of Na and Al contents in formulas of gedrite normalized to 23 O. (b)Negative correlation of Na with XMg of gedrite. Dashed line indicates compositionsof coexisting gedrite and anthophyllite.

the metapelite (Table 9; Fig. 12). There is no overlapping ofcompositions of plagioclase from the trondhjemite-metapelitereaction zone and compositions of the metapelitic plagioclase(Fig. 12). The anorthite mole fraction of some plagioclase fromthe trondhjemite-metapelite reaction zone (0.24–0.27) approachesXCa of the metapelitic plagioclase. Usually, these compositions areobserved in plagioclase forming graphic-like intergrowths withquartz (Fig. 5c and f) or contacting with biotite-quartz sym-plectites in the matrix of the rock. The most sodic plagioclase(XCa = 0.13–0.18) is identified at contacts with Ca-rich garnet andsillimanite locally growing along the contacts of large garnet withplagioclase in the trondhjemite-metapelite reaction zone (Fig. 5e).

In comparison with plagioclase in metapelite, plagioclase fromthe trondhjemite and the trondhjemite-metapelite reaction zonecontains higher and widely varying concentrations of the ortho-clase component (Table 9; Fig. 12). The increase of the orthoclase

Table 8Representative analyses of staurolites in the metapelite SAF12-2/2.

Components 1 2 3 4 5

SiO2 28.29 26.61 27.26 27.26 27.17TiO2 0.11 0.24 0.51 0.39 0.12Al2O3 54.53 52.60 52.44 53.00 53.54FeO 10.87 12.84 12.65 11.24 12.47MnO 0.07 0.09 0.08 0.13 0.10MgO 2.92 2.65 2.81 2.68 2.83CaO 0.07 0.00 0.03 0.00 0.00Na2O 0.01 0.21 0.21 0.22 0.29K2O 0.00 0.01 0.03 0.01 0.04ZnO 0.73 0.85 1.09 0.98 1.10Cr2O3 0.00 0.29 0.21 0.11 0.04Total 97.60 96.40 97.33 96.02 97.72Formula quantities normalized to 22 OSi 3.721 3.603 3.654 3.671 3.622Ti 0.011 0.024 0.052 0.039 0.012Al 8.452 8.391 8.283 8.411 8.408Fe 1.195 1.453 1.418 1.265 1.390Mn 0.008 0.010 0.009 0.015 0.011Mg 0.572 0.535 0.562 0.538 0.562Ca 0.010 0.000 0.004 0.000 0.000Na 0.003 0.056 0.053 0.058 0.075K 0.000 0.002 0.005 0.002 0.006Zn 0.071 0.085 0.108 0.098 0.109Cr 0.000 0.031 0.022 0.012 0.005XMg 0.322 0.268 0.282 0.296 0.286

Note: 1–4, staurolites associated with gedrite-sillimanite (kyanite) textures; 5, del-icate staurolite crystals in the plagioclase rims around cordierite (Fig. 4d).


Fig. 12. Compositional variations of plagioclase in the studied rocks. 1, matrix pla-gioclase from the metapelite SAF12-2/2; 2, plagioclase related to the retrogradeassemblage Ged + Sil + Qtz + Bt + St after cordierite in the metapelite SAF12-2/2; 3,plagioclase from the trondhjemite-metapelite reaction zone; 4, plagioclase fromthe trondhjemite. Hatched markers denote re-integrated compositions of thelamellae-rich plagioclase grains. Re-integrated compositions of plagioclase from thetrondhjemite are shown in the inset.

component of plagioclase is reflected in the presence of abun-dant plagioclase with thin K-feldspar lamellae. Re-integratedcompositions of the lamellae-rich plagioclase indicate that XK ofthe initial plagioclase in the garnet trondhjemite varied within0.15–0.22 (Table 9). However, a significant number of the re-integrated analyses show XK up to 0.3 (Fig. 12; Table 9). There-integration of compositions of the lamellae-rich plagioclasefrom the trondhjemite-metapelite reaction zone shows lower XK,about 0.06 (Fig. 12; Table 9).

6. Fluid inclusions in the garnet trondhjemite

CO2 inclusions of diverse sizes and shapes, which seem to bedependent on the host minerals, are present both in quartz andgarnet (Fig. 13a–c). Small (5–15 �m) inclusions of CO2 are ubiqui-tous in quartz. They are rounded or show a negative crystal shapes.Large CO2 inclusions (up to 25 �m) of irregular shape are rare. Thesize of the CO2 inclusions in garnet varies from 5 up to 40 �m,and the shapes are extremely variable and range from irregularcurved inclusions with numerous branches to elongated inclusionswith zigzag-shaped tails with darkened boundaries (Fig. 13b). Thecarbonic inclusions texturally can be subdivided into primary andprimary-secondary varieties. The primary inclusions are single orform isolated groups, usually in the centers of grains and rarelyin their periphery. Primary inclusions are rare and only occur ingarnet (Fig. 13a). They are small (5–10 �m), usually with nega-tive crystal shape or oviform. Homogenization temperatures (Th)were measured from −25 ◦C to −17.7 ◦C, while the melting temper-ature is −57.8 ◦C. These values correspond to the range of densities1.054–1.02 g/cm3. Primary-secondary inclusions are detected inboth garnet and quartz. The homogenization temperature (density)of the primary-secondary inclusions in garnet varies from −32.1 ◦C(1.085 g/cm3) to 22 ◦C (0.75 g/cm3), while the melting tempera-ture (Tm) is −58.7 ◦C to −57.5 ◦C. Carbonic inclusions in quartz(Fig. 13c) record homogenization temperatures (densities) −32.7 ◦C(1.087 g/cm3) to 24.4 ◦C (0.719 g/cm3) (Tm = −58.1 ◦C to −57.7 ◦C).These variations reflect both conditions and time of entrapment,as well as the post-entrapment history. Slight decrease of Tm withrespect to pure carbon dioxide can be explained by the presence ofsmall amounts of N2 and/or CH4.

Aqueous-salt inclusions are only present in quartz and are rep-resented by primary-secondary varieties (Fig. 13d). They occuras slightly flattened inclusions of irregular shapes with sizes10–15 �m. Inclusions with sizes up to 25 �m are rare. Final melt-ing temperature varies from −7.5 ◦C to −17.2 ◦C, corresponding to11.11–20.58 wt.% NaCl equivalent. In all cases, brine inclusions areassociated with the primary-secondary carbonic inclusions, sug-gesting coexisting fluids during entrapment.

7. Thermobarometric methods

The most complete information about P–T conditions duringthe evolution of the studied rocks was obtained using a combi-nation of conventional thermobarometry and modeling of mineralassemblages using the pseudosection method for a given bulk com-position. Conventional thermobarometry accounts for the localequilibria preserved in a rock that passed through the complexP–T evolution recorded by a sequence of mineral assemblages andmineral zoning. In contrast, the pseudosection method allows thebulk compositional control on mineral equilibria to be analyzed. Forconventional point-to-point thermobarometry of the metapelite-trondhjemite rock association, we used the winTWQ (version 2.32)software (Berman, 2007) with end-member mineral according toBerman (1988). This method allows the determination of both Pand T conditions using all possible equilibria preserved in somespecific local mineral assemblage, and estimation of the degreeof equilibration of this assemblage. Solid solution models for gar-net, cordierite, spinel, orthopyroxene from Berman and Aranovich(1996) were applied in the calculations. Since the model for thecordierite solid solution implemented into the winTWQ does notinclude thermodynamic effects of H2O and CO2 in the cordierite,we used the following equations (Gerya and Perchuk, 1990) to cal-culate activities (ai) cordierite end-members (Mg-cordierite andFe-cordierite) in the fluid saturated solid solution to estimate aninfluence of volatiles in cordierite on the pressure estimates viacordierite-bearing equilibria:

RT ln aMgCrd = 2RT ln XCrdMg − 1333 + 0.617T − 0.336P

+ 1026(1 − XH2O) + 472(1 − XH2O)2 (a1)

RT ln aFeCrd = 2RT ln(1 − XCrdMg ) − 1333 + 0.617T − 0.336P

+ 1026(1 − XH2O) + 472(1 − XH2O)2, (a2)

where XH2O = H2O/(H2O + CO2) is a H2O mole fraction in the fluidequilibrated with cordierite. Unfortunately, there are no data onthe volatile content of cordierite in the SMZ metapelites. Followingto suggestion by Perchuk et al. (2000) and numerous studies offluid inclusions in the metapelites (Van Reenen, 1986; Van Reenenand Hollister, 1988; Baker et al., 1992; Van den Berg and Huizenga,2001; Huizenga et al., 2014), the value XH2O = 0.2 was taken intoaccount.

For modeling of the mineral assemblages in the garnettrondhjemite and metapelite, we applied the pseudosectionmethod using the PERPLE X software (Connolly, 2005) in ver-sion 6.6.8 for Windows with the standard properties databasehp02ver.dat and solution model database solute 09.dat (http://www.perplex.ethz.ch). The following models were applied forsolid solutions (see notations and references at http://www.perplex.ethz.ch/perplex solution model glossary.html): Gt(HP) forgarnet, TiBio(HP) or Bio(HP) for biotite, HCrd for hydrous cordierite,Mica(CHA1) for white mica solid solution, St(HP) for staurolite, andoAmph(DP) for orthoamphibole solid solution (including Na sub-stitution in gedrite). The mixing model of White et al. (2001) hasbeen used for the melt in the NCKFMASH system.

http://www.perplex.ethz.ch/

http://www.perplex.ethz.ch/

http://www.perplex.ethz.ch/perplex_solution_model_glossary.html

http://www.perplex.ethz.ch/perplex_solution_model_glossary.html


Table 9Representative analyses of plagioclases from the metapelite SAF12-2/2.

Components 1 2 3 4 5 6 7 8 9 10 11 12

SiO2 63.17 62.57 62.70 60.34 61.33 61.05 60.71 59.68 57.94 50.71 51.92 49.81Al2O3 25.23 24.78 24.63 24.70 24.83‘ 24.36 26.84 27.20 28.37 30.65 30.17 31.96CaO 6.41 5.94 6.07 6.23 5.88 5.75 8.55 8.82 9.35 14.02 12.15 14.91Na2O 8.01 8.16 8.36 8.30 8.36 8.43 6.94 6.66 6.14 3.69 4.38 3.10K2O 0.07 0.09 0.08 0.07 0.09 0.08 0.04 0.04 0.04 0.02 0.05 0.04Total 102.89 101.54 101.84 99.64 100.49 99.68 103.08 102.40 101.84 99.09 98.66 99.82Formula quantities normalized to 8 OSi 2.716 2.730 2.731 2.693 2.710 2.720 2.623 2.599 2.540 2.323 2.372 2.274Al 1.278 1.274 1.264 1.299 1.293 1.279 1.366 1.396 1.466 1.654 1.624 1.720Ca 0.295 0.278 0.283 0.298 0.278 0.274 0.396 0.411 0.439 0.688 0.595 0.729Na 0.667 0.690 0.706 0.718 0.716 0.728 0.581 0.562 0.522 0.328 0.388 0.274K 0.004 0.005 0.004 0.004 0.005 0.005 0.002 0.002 0.002 0.001 0.003 0.002XCa 0.305 0.285 0.285 0.292 0.278 0.272 0.404 0.422 0.456 0.677 0.604 0.725XK 0.004 0.005 0.004 0.004 0.005 0.005 0.002 0.002 0.002 0.001 0.003 0.002XNa 0.691 0.709 0.710 0.704 0.716 0.723 0.594 0.576 0.542 0.322 0.394 0.273

Note: 1, inclusion in matrix orthopyroxene (see Fig. 4e); 2–6, matrix plagioclases; 7–9, rims on cordierite (Fig. 4c and d); 10–12, plagioclase thin zones at contactswith gedrite (Fig. 4f).

Representative analyses of plagioclases from the trondhjemite-metapelite reaction zone.

Components 1 2 3 4 5 6 7 8 9 10 11

SiO2 63.11 63.33 61.61 60.91 62.22 62.97 63.61 63.96 63.80 63.88 63.80Al2O3 22.52 22.25 23.26 24.02 23.86 22.51 23.34 21.95 22.22 21.99 22.04CaO 3.51 3.25 4.36 5.19 5.04 3.37 3.87 3.39 3.41 3.40 3.25Na2O 9.40 9.53 8.72 8.42 8.68 9.60 9.58 8.95 9.01 8.88 9.21K2O 0.30 0.33 0.37 0.19 0.14 0.30 0.22 1.21 1.16 1.08 1.10BaO 0.16 0.03 0.14 0.10 0.00 0.11 0.00 0.22 0.27 0.25 0.15Total 99.01 98.72 98.46 98.83 99.93 98.86 100.63 99.46 99.60 99.23 99.39Formula quantities normalized to 8 OSi 2.816 2.830 2.772 2.734 2.756 2.814 2.793 2.844 2.833 2.844 2.840Al 1.184 1.172 1.233 1.270 1.245 1.186 1.207 1.150 1.163 1.154 1.156Ca 0.168 0.155 0.210 0.249 0.239 0.162 0.182 0.161 0.162 0.162 0.155Na 0.813 0.825 0.760 0.733 0.745 0.831 0.815 0.771 0.775 0.766 0.794K 0.017 0.019 0.022 0.011 0.008 0.017 0.013 0.069 0.066 0.061 0.062Ba 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.003 0.004 0.003 0.002XCa 0.168 0.155 0.212 0.251 0.241 0.160 0.180 0.161 0.162 0.164 0.153XK 0.017 0.019 0.022 0.011 0.008 0.017 0.012 0.069 0.065 0.062 0.062XNa 0.814 0.826 0.766 0.738 0.751 0.823 0.807 0.770 0.773 0.774 0.785

Note: 1 and 2, centers of the lamellae-free plagioclases; 3, graphic-like intergrowth with quartz (Fig. 5f); 4 and 5, at biotite-quartz symplectites in matrix; 6 and 7,at Ca-rich garnet and sillimanite (Fig. 5e); 8–11, re-integrated compositions of plagioclases with K-feldspar lamellae.

Representative analyses of plagioclases from the trondhjemite.

Components 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

SiO2 62.34 65.11 65.07 65.20 61.44 61.09 60.67 60.59 61.28 64.03 63.71 65.48 64.85 65.27 64.98Al2O3 23.55 23.91 23.25 23.38 21.87 21.66 21.61 21.62 20.28 22.36 22.68 22.94 23.09 22.69 23.06CaO 4.77 4.17 4.07 4.28 3.74 3.61 2.82 2.38 2.24 4.57 3.91 4.03 4.36 4.10 4.10Na2O 8.64 8.82 8.51 8.67 7.62 6.46 6.17 6.19 4.17 7.11 6.84 7.29 6.95 7.78 7.31K2O 0.27 0.39 0.43 0.34 2.51 4.20 5.08 5.16 8.74 3.93 3.69 3.62 3.52 2.59 3.00BaO n.d. n.d. 0.20 0.14 n.d. n.d. n.d. n.d. n.d. 0.00 0.09 0.05 0.00 0.00 0.08Total 99.66 102.47 102.68 102.34 97.24 97.14 96.46 96.06 96.83 102.58 101.72 104.36 103.22 102.85 103.85Formula quantities normalized to 8 OSi 2.770 2.803 2.822 2.815 2.816 2.819 2.825 2.830 2.876 2.810 2.814 2.819 2.813 2.829 2.814Al 1.233 1.213 1.188 1.190 1.181 1.178 1.185 1.190 1.121 1.156 1.180 1.164 1.180 1.159 1.177Ca 0.227 0.192 0.189 0.198 0.184 0.178 0.141 0.119 0.113 0.215 0.185 0.186 0.203 0.190 0.190Na 0.744 0.736 0.715 0.725 0.677 0.578 0.557 0.560 0.379 0.605 0.585 0.608 0.584 0.653 0.613K 0.015 0.021 0.024 0.019 0.147 0.247 0.302 0.307 0.523 0.220 0.208 0.199 0.195 0.143 0.166Ba 0.000 0.000 0.003 0.002 0.000 0.000 0.000 0.000 0.000 0.000 0.002 0.001 0.000 0.000 0.001XCa 0.230 0.202 0.204 0.210 0.182 0.178 0.141 0.121 0.111 0.207 0.189 0.187 0.206 0.193 0.196XK 0.016 0.023 0.026 0.020 0.146 0.246 0.302 0.312 0.515 0.212 0.212 0.200 0.198 0.145 0.171XNa 0.754 0.775 0.771 0.770 0.672 0.576 0.557 0.568 0.374 0.582 0.598 0.613 0.595 0.662 0.633

Note: 1–3, centers of the lamellae-free plagioclases; 4, inclusion in GRT-I; 5–15, re-integrated compositions of plagioclases with K-feldspar lamellae.

8. Mineral equilibria and P–T conditions of the studiedrocks

8.1. P–T conditions of the garnet trondhjemite andtrondhjemite-metapelite reaction zone

Two morphological types of garnet, GRT-I and GRT-II, contain-ing various inclusions clearly indicate the complex evolutionary

history of the garnet trondhjemite during cooling. Inclusions inGRT-I, i.e. biotite, muscovite, sillimanite, zoisite, and quartz, showthat this garnet represents the later generation with respect to GRT-II, which contains inclusions of sillimanite, corundum and spinel.This interpretation is proven by the presence of overgrowths inthe outer zones of GRT-II that contain biotite and muscovite inclu-sions (Fig. 3l and m). Abundance of quartz inclusions and scarcity ofplagioclase inclusions in GRT-I furthermore shows that this garnet


Fig. 13. Fluid inclusions in the garnet trondhjemite PET-A6. (a) Isolated primary inclusions of CO2 in garnet. (b) Trail of primary-secondary inclusions of CO2 along crack ingarnet. (c) Trail of primary-secondary inclusions of CO2 along crack in quartz. (d) Trail of primary-secondary aqueous-salt inclusions along crack in quartz.

formed after crystallization of a major portion of plagioclase, whichstrongly predominates over quartz in the rock.

Numerous inclusions of fibrous sillimanite aggregates andcorundum in the skeleton-like GRT-II (Fig. 3h, m and n) sug-gests the growth of these garnets as a result of sillimanite andcorundum reactions with the trondhjemite melt. Re-integrated pla-gioclase compositions plotted onto isotherms in the ternary systemAb–Or–An calculated from the model by Elkins and Grove (1990)are located between 850 and 1000 ◦C (Fig. 14) and these values areinterpreted to correspond to the temperature conditions at whichthe trondhjemite magma began interaction with the metapeliteblocks.

Coexistence of sillimanite, corundum, plagioclase and garnetwith the trondhjemite melt reflects the following equilibria:

Cor + SiO2 = Sil (1)

3An (in Pl) = Grs (in Grt) + 2Sil + SiO2 (2)

3An (in Pl) = Grs (in Grt) + 2Cor + 3SiO2 (3)

where SiO2 represents a component of the melt. The fact that quartzis absent from sillimanite inclusions in GRT-II (Fig. 3h, m and n) indi-cates that quartz was a later phase in the trondhjemite with respectto plagioclase. Thus, the SiO2 activity of the melt in the time ofthe GRT-II crystallization was below unity. Fig. 15a shows equilib-ria (1)–(3) calculated for compositions of garnet coexisting withsillimanite and corundum at pressures 7.5 and 6 kbar and varyingactivity of SiO2 (standardized to quartz). The average compositionof the re-integrated plagioclase in the trondhjemite was used in thecalculations, and the intersection of equilibria (1)–(3) defines thetemperature interval 970–880 ◦C at log(aSiO2

) varying from −0.047to −0.058 (Fig. 15a). These temperatures are in a perfect agreementwith the ternary feldspar thermometry (Fig. 14).

The spectacular feature of the polyphase sillimanite-corunduminclusions in GRT-II is the presence of spinel, which usually grows atthe contacts of the inclusions with host garnet (Fig. 3h–k). Absenceof spinel at the contacts of single sillimanite inclusions in garnet

Fig. 14. Compositions of re-integrated exsolved plagioclase from the garnet trond-hjemite plotted on isotherms in the ternary Ab–Or–An systems calculated from themodel by Elkins and Grove (1990) at pressure of 6.3 kbar; the inset shows a typicalexsolution texture used for re-integration of the plagioclase composition.

suggests that corundum is a necessary factor for the formation ofspinel inside the polyphase inclusions via the reaction

Grt + 5Cor = 3Spl + 3Sil. (4)

This reaction displaces to the right with decreasing pressureand increasing temperature. Combined with the garnet-spinelexchange reaction, (4) shows that spinel appeared in thepolyphase inclusions in garnet at 860–870 ◦C at 6.2–6.3 kbar(Fig. 15b). Accounting for the Fe3+ content of garnet and spinel asXMg = Mg/(Mg + Fe2+–Fe3+) increases the temperatures estimates byabout 20 ◦C (up to 890 ◦C). These temperatures are consistent withthe temperature range estimated from the re-integrated ternary


Fig. 15. Thermobarometry of the garnet trondhjemite. (a) Temperature and SiO2 activity of the trondhjemite melt coexisting with the assemblage gar-net + corundum + sillimanite + ternary feldspar at pressures 7.5 kbar (solid lines) and 6 kbar (dashed lines). (b and c) Two typical examples of the TWQ calculations forthe assemblage spinel + garnet + sillimanite + corundum + plagioclase indicating temperatures 860–870 ◦C and pressure 6.2–6.3 kbar; Fe3+ contents of garnet and spinel arenot taken into account in these calculations; thin dashed lines in the upper plot show some selected Qtz-bearing reactions as an illustration of the conclusion that quartz didnot participate in the assemblage. (d and e) Two typical examples of the TWQ calculations for the assemblage garnet + biotite + plagioclase + quartz indicating temperatures640–650 ◦C and pressure 6.0–6.5 kbar.

plagioclase and garnet–sillimanite–corundum-(melt) equilibria(Figs. 14 and 15a). Introduction of the plagioclase componentallows additional equilibria to support these P–T data (Fig. 15b andc). All equilibria involving quartz are beyond the equilibrium inter-section (Fig. 15b and c) suggesting that no quartz was present inthe system during formation of spinel.

The growth of GRT-II was subsequently followed by the for-mation of GRT-I that equilibrated with biotite, muscovite, quartz,

and locally sillimanite and zoisite. Rare K-feldspar-free plagio-clase inclusions in GRT-I (Fig. 3d) contain up to 0.54 wt.% K2O,corresponding to temperatures below 700 ◦C. According to thegarnet-biotite exchange equilibrium, compositions of biotite inclu-sions in GRT-I correspond to the temperature interval 600–650 ◦C.It is evident that biotite inclusions were trapped by GRT-I in thepresence of quartz and plagioclase. This provides an opportunity toestimate pressure for this assemblage based on equilibria involving


the anorthite component of plagioclase, the grossular componentof garnet, and the eastonite-siderophylite component of biotite.Fig. 15d and e shows the TWQ plots calculated for two exam-ples of micro assemblages involving biotite inclusions showed inFig. 3m and f. Fig. 14d represents garnet overgrowing GRT-II, whileFig. 15e represents a typical example of inclusions in GRT-I. Bothmicro assemblages give consistent P–T results at about 640 ◦C and6.0–6.5 kbar.

The principal assemblage of the trondhjemite-metapelite reac-tion zone is garnet + biotite + plagioclase + quartz (Fig. 5a–e). TheGrt-Bt Mg–Fe exchange equilibrium applied to cores of garnets andlarge biotite flakes (Fig. 5a) provides temperatures above 660 ◦C (atassumed pressure 6.3 kbar). Variability of the grossular content ingarnets and the anorthite content of plagioclase allows a wide pres-sure range to be calculated, from 5.7 to 7.5 kbar without any clearstatistical peak.

The specific feature of the trondhjemite-metapelite reactionzone is the presence of relicts of orthopyroxene and rare cordierite(Fig. 5b and c). Compositional differences between these relicminerals and minerals from the metapelite, i.e. slightly higher Mg-numbers and higher Al content of orthopyroxene (Fig. 9), implythat the relics were, probably, re-equilibrated with garnet formedduring interaction of the metapelite with the trondhjemite. Thisallows calculation of P–T conditions for the assemblage garnet-orthopyroxene-cordierite-plagioclase-quartz (Fig. 16a). All threeFe–Mg exchange equilibria in this system gave a consistent tem-perature of 620 ± 5 ◦C, while reaction

Grs + 2Prp + 3Qtz = 6En + 2An (5)

intersects them at pressure about 5.7 kbar. All reactions involv-ing AlOpx component in orthopyroxene (the equilibriumAlm = 3Fs + AlOpx is shown as an example in Fig. 15a) are farbeyond this intersection. According to the experimental studyby Aranovich and Berman (1997) on the orthopyroxene-garnetequilibrium in the FAS boundary system, the 6.0–6.7 wt.% ofAl2O3 preserved in the relic orthopyroxene coexisting with garnetwould correspond to temperatures above 800 ◦C. Aranovich andBerman (1997) specifically noted that thermometry based on theequilibrium

Alm = 3Fs + AlOpx (6)

usually gives higher temperatures in comparison to the Opx-GrtFe–Mg exchange equilibrium because of principally different clo-sure temperatures indicating a lack of a total equilibrium in theassemblage. In our case this difference is about 200 ◦C (Fig. 16a).Following the Aranovich and Berman (1997) strategy, we assumethat equilibrium (6) reflects the high temperature stage of the rockevolution. At high temperatures orthopyroxene and garnet wereequilibrated with the homogeneous ternary feldspar, a suggestionthat is fully proven by the presence of exsolved plagioclases in thereaction zone rock (Fig. 5b and c). With re-integrated compositionsof this plagioclase, the equilibria (5) and (6) intersect at tempera-tures about 820 ◦C and pressures 6.2–6.3 kbar (Fig. 16a).

Thus, the orthopyroxene-garnet-cordierite-plagioclase-quartzequilibria in the reaction zone between trondhjemite andmetapelite reflect a cooling path from above 800 ◦C down to 620 ◦Cwithin the pressure interval 6.3–5.7 kbar. This near-isobaric coolingpath is reflected in compositions of coexisting biotite and gar-net in the rock (Fig. 16b). The Grt-Bt Mg–Fe exchange equilibriumapplied to re-equilibrated contacting rims of the mineral and inclu-sions of biotite in garnet regularly provides lower temperatures(at assumed pressure 6.3 kbar) of 570–630 ◦C, in comparison totemperature estimated from the core compositions of these min-erals (Fig. 16b). The Mg–Fe exchange equilibrium of orthopyroxenewith biotite that forms intergrowths with quartz after orthopy-roxene shows temperatures consistent with the Grt-Bt equilibrium

Fig. 16. Thermobarometry of the trondhjemite-metapelite reaction zone. (a)Comparison of the Fe–Mg orthopyroxene-garnet, garnet-cordierite and cordierite-orthopyroxene equilibria with the equilibrium Alm = 3Fs + AlOpx for the composi-tion of the relict orthopyroxene; note that intersection of these equilibria with theequilibrium Grs + 2Prp + 3Qtz = 6En + 3An gives very close pressure values (5.7 and6.2 kbar, respectively). (b) The garnet-biotite Mg–Fe exchange equilibrium appliedto cores of garnet and large biotite flakes (dashed lines), and to re-equilibrated con-tacting rims and inclusions of biotite in garnet (solid lines); dotted line – the Mg–Feexchange equilibrium of orthopyroxene with biotite; pressure for the late stageof cooling is estimated from the equilibrium of biotite rims and late garnet zonescontaining inclusions of ilmenite intergrown with rutile (Fig. 5d). (c) An exampleof the TWQ calculation for the assemblage of the late Ca-rich idiomorphic garnetintergrown with Bt + Pl + Qtz + Kfs (Fig. 5f).

(Fig. 16b). The pressure value, about 6.2 kbar (Fig. 16c), for the endof this cooling path was estimated from the equilibrium of biotiterims and late garnet zones filled with inclusions of ilmenite inter-grown with rutile (Fig. 5d). Similar P–T parameters, 610–640 ◦Cat 5.5–6.5 kbar, were calculated using compositions of the lateCa-rich (1.0–1.4 wt.% CaO) idiomorphic garnet intergrown with Ti-poor biotite, plagioclase, quartz and K-feldspar (Fig. 5f). These P–T


parameters are very close to the haplogranite solidus in the pres-ence of water-rich fluids, namely H2O-KCl-NaCl (Aranovich et al.,2013) and H2O-CO2 (Ebadi and Johannes, 1991). This is also consis-tent with the presence of graphic textures in the rock, suggestingthe final crystallization of the near-eutectic melt (Fig. 5f).

Continuous cooling below the solidus presumably is reflectedby formation of Ca-enriched garnet, sillimanite and quartz afterplagioclase (Fig. 5e) via reaction

3An = Grs + 2Sil + Qtz. (7)

The formation of garnet-sillimanite-quartz intergrowths after pla-gioclase as a result of sub-isobaric cooling is consistent with thesuggestion by Perchuk et al. (1996, 2000).

8.2. Modeling of mineral assemblages of the garnet trondhjemiteusing the pseudosection method

P–T pseudosections were constructed for bulk compositionsof samples PET6 and PET11 (Table 1) (Fig. 17a–c). The samplesrepresent the garnet trondhjemite bulk compositions in differentportions of the intrusive body. Concentration of 1 wt.% H2O wasassumed for construction of the pseudosections.

For the bulk composition PET6, which is depleted in SiO2, MgO,FeO and enriched in Al2O3 (Table 1) addition of 1 wt.% of H2Oresulted in the trondhjemitic melt coexisting with a single ternaryfeldspar above 1000 ◦C (Fig. 17a and b). This reflects the presenceof the K-feldspar-rich cores of plagioclases in the rock (Fig. 3a).Quartz joins plagioclase during cooling (Fig. 17a and b), but doesnot exceed 5 vol.%. Absence of corundum and spinel in the high-temperature assemblage can be readily explained by the higherSiO2 content of the modeled samples (Table 1) in comparison to theplagioclase-rich domains in the trondhjemite containing corundumand spinel-bearing garnet grains (Fig. 3h, m and n). The pseudosec-tion constructed for 1 wt.% of H2O predicts that garnet crystallizesafter sillimanite at temperatures of about 860 ◦C (Fig. 17a). Thisresult is consistent with the abundance of sillimanite inclusions inGRT-II (Fig. 3h–n). However, an increase of the H2O content up to3 wt.% changes the relations between garnet and sillimanite, withgarnet crystallizing earlier than sillimanite (Fig. 17b). The watercontent also influences the exsolution of ternary feldspar into twofeldspars, reflecting the appearance of K-feldspar lamellae in pla-gioclase (Fig. 3a). At 1 wt.% H2O, ternary feldspar splits into twofeldspars at temperature about 700 ◦C (Fig. 17a). In contrast, twofeldspars appear on the solidus only if 3 wt.% H2O is added to thesystem (Fig. 17b). Muscovite formation at about 650 ◦C is accompa-nied by disappearance of sillimanite (Fig. 17a and b). Nevertheless, apseudosection constructed for 3 wt.% H2O (Fig. 17b) shows a narrowfield, where sillimanite coexists with muscovite. In fact, sillimaniteis extremely rare in the assemblage with muscovite inclusions inboth in GRT-I and GRT-II (Fig. 3c, d and f). The muscovite-sillimaniterelations in the trondhjemite were regulated by the reaction

Sil + Kfs + H2O = Mu + Qtz (8)

K-feldspar in this reaction is supplied by exsolution of the high-temperature plagioclase. It explains why the phase boundaries “Fspexsolution” and “Mu in” almost coincide in the pseudosection con-structed for 1 wt.% H2O (Fig. 17a).

Both pseudosections for the sample PET-6 (Fig. 17a and b)show that biotite appears immediately at the solidus (at about590–600 ◦C). This result agrees with formation of late Ti-poorbiotite around garnet in the trondhjemite, but disagrees with thepresence of biotite inclusions in GRT-I and GRT-II (Fig. 3f and m).This dilemma can be solved using the pseudosection for the bulkcomposition PET-11 (Fig. 17c), which is richer in SiO2 and onlyslightly richer in MgO and FeO. This pseudosection shows that this

change in the bulk composition significantly enlarges biotite sta-bility field up to about 740 ◦C. This is consistent with the strongpredominance of biotite inclusions in garnet from the reaction zonebetween the trondhjemite and metapelite (Fig. 5a and b), whichis strongly enriched in MgO and FeO (Table 1). According to thepseudosections, garnet crystallizes prior to sillimanite in PET-11.This result is consistent with extreme scarcity of sillimanite inclu-sions in GRT-I and its absence in garnet from the reaction zone. Twofeldspars also did not form in PET-11, which is definitely related toformation of biotite prior to exsolution of feldspar (compare Fig. 17aand b). It is evident that K2O released from plagioclase on coolingactively participated in the formation of biotite and, subsequently,also of muscovite.

Pseudosections predict a solidus at temperatures of about600 ◦C. This result is consistent with the conventional ther-mobarometry applied to the late assemblage of Ca-rich garnetand biotite associated with myrmeckite-like intergrowths of pla-gioclase, K-feldspar and quartz in the trondhjemite-metapelitereaction zone (Fig. 5f). These textures could represent products ofsolidification of the latest melt.

Thus, the pseudosections (Fig. 17a–c) closely reproduce mineralassemblages observed in the garnet trondhjemite as well as theirtemperature ranges compared to the results of conventional ther-mobarometry. Pseudosections also support the sub-isobaric P–Tpath for cooling of the trondhjemite, and show that after intru-sion, the P–T evolution of the metapelite was closely linked to theevolution of the trondhjemite.

8.3. P–T conditions of metapelites

The metapelite samples clearly show two principal assemblagesthat reflect different stages of the rock P–T evolution. Texturalrelations in the metapelite SAF12-2/2 imply that the assemblageOpx + Crd + Bt + Pl + Qtz was formed before or contemporarily withthe foliation (Fig. 4b) representing peak or near-peak metamor-phic conditions. Fig. 18a and b shows two typical results of theTWQ calculations for compositions of orthopyroxene, cordieriteand biotite in the sample SAF12-2/2. Since rare cordierite grainsin the matrix of the sample SAF12-2/2 are severely altered by latehydration products (Fig. 4a and c), the compositions of less alteredcordierite inclusions in orthopyroxene were chosen for calcula-tions. The diagrams shows a significant dispersion of temperaturevalues determined by three-major Fe–Mg exchange equilibria (Crd-Bt, Crd-Opx, Opx-Bt) for both cases. Taking into account muchlower enthalpy effect of the Crd-Bt exchange equilibrium and,thus, its stronger dependence on compositional errors, the tem-peratures calculated from the Crd-Opx and Opx-Bt equilibria aretaken as more reliable values. Nevertheless, these values are sig-nificantly different for two examples (730–750 ◦C and 650–660 ◦C,respectively). It is evident that this dispersion is related to strongFe–Mg re-equilibration of the cordierite inclusions during coolingand the formation of hydrous phases (gedrite, biotite, staurolite)along the contacts of the inclusions with host orthopyroxene.Despite the discrepancy of temperatures values, pressure valuescalculated from the more robust equilibria involving AlOpx com-ponent in orthopyroxene are relatively consistent, 7.5–8.0 kbarand 8.0–8.5 kbar, respectively. These pressure values correspond toequilibria of volatile-free cordierite. Taking into account the activ-ities of components in the volatile-bearing cordierite (a1) and (a2),the pressure range increases up to 9.5 kbar (Fig. 18a and b).

The presence of the cordierite-orthopyroxene clusters in themetapelite PET-5 (Fig. 4g) show that the evolution of the rockstarted with the formation of the assemblage, which includedformer garnet. However, garnet was decomposed to cordierite andorthopyroxene during the early stage of decompression and cool-ing (Van Reenen, 1983; Perchuk et al., 1996, 2000; Smit et al., 2001;


Fig. 17. P–T pseudosections calculated using the PERPLE X software for the garnet trondhjemite assemblages exemplified by two bulk compositions (Table 1). (a) For PET-6at 1 wt.% of H2O in the system, (b) for PET-6 at 3 wt.% of H2O in the system, and (c) for PET-11 at 1 wt.% of H2O in the system.

Van Reenen et al., 2011). Application of the Opx-Crd-Bt equilibria tothe compositions of minerals inside the clusters gives a wide scat-tering of temperatures from 680 up to 820 ◦C without any distinctstatistical peak. Again, it reflects a re-equilibration of the mineralcompositions after formation of the cordierite-orthopyroxene clus-ters. Pressure values calculated from the equilibria involving AlOpxcomponent in orthopyroxene vary within the range 7.9–9.1 kbarwith mean value about 8.5 kbar. Nevertheless, compositions ofcoexisting orthopyroxene, cordierite and biotite in the coarsenedportion of the metapelite PET-5 give a more narrow range of P andT that clusters around 770 ◦C and 8.5 kbar (Fig. 18c).

The early assemblage Opx + Crd + Bt + Pl + Qtz in the metapeliteSAF12-2/2 and PET-5 is followed by later assemblages. Thin zonesof orthopyroxene-sillimanite intergrowths along the contacts ofcordierite and orthopyroxene in the sample PET-5 indicate the reac-tion

Crd + AlOpx = 2Opx + 3Sil (9)

Formation of the intergrowths within the orthopyroxene-cordierite clusters (Fig. 4g and h), which are products of

decomposition of earlier garnet, clearly indicate the retrograde ori-gin of this assemblage after the major stage of decompression.Fig. 18d demonstrates the result of the TWQ calculation for thecompositions of orthopyroxene in the Opx + Sil intergrowths andcontacting cordierite (Fig. 4h). Orthopyroxene with the lowest XAlvalues was used for calculations in order to avoid an influence ofsuperposition from sillimanite during analyses of finely intergrownorthopyroxene and sillimanite. The intersection of the reaction(9) involving volatile-free cordierite with the exchange reactiondefines temperatures about 640 ◦C at pressure 4.5 kbar. For volatile-bearing cordierite (with activities calculated from the equations(a1) and (a2)), this pressure increases up to 5.7 kbar. The rangeof compositions of orthopyroxenes from the intergrowths andcontacting cordierites dives a range of temperatures 550–650 ◦Cand pressures (accounting for volatiles in cordierites) 5.3–5.7 kbar.

In the sample SAF12-2/2, the extensively developed late assem-blage Ged + Sil + Qtz ± Bt ± St originated after deformation andevidently was triggered by fluid influx along the foliation planes(Fig. 4e and f). Taking into account the significant sodium content of


Fig. 18. Thermobarometry of metapelite SAF12-2/2 and PET-5. (a and b) Two typical examples of the TWQ calculation for the assemblage orthopyrox-ene + cordierite + biotite + quartz in the metapelite SAF12-2/2; dashed lines show the equilibrium (4) with participation of volatile-bearing cordierite. (c) Example of theTWQ calculation for the assemblage orthopyroxene-cordierite clusters from metapelite PET-5 (Fig. 4g); dashed line shows the equilibrium (4) with participation of volatile-bearing cordierite. (d) Results of the TWQ calculations for late micro assemblages cordierite + orthopyroxene + sillimanite in the metapelite PET-5; dashed line shows theequilibrium (2) with participation of volatile-bearing cordierite.

gedrite (Table 7), its formation along with staurolite after cordieriteand plagioclase is described by reactions:

3Crd + Ab (in Pl) + H2O = Na-Ged + 5Sil + 7Qtz (10)

5Crd + Ab (in Pl) + 2H2O = Na-Ged + St + 14Qtz (11)

Crd + (1/2)H2O + (5/2)Sil = (1/2)St + (7/2)Qtz (12)

where Na-Ged is the NaMg6Al3Si6O22(OH)2 end-member. Con-sumption of the albite component of plagioclase for formation ofNa-gedrite is expressed in the formation of thin zones of extremelyanorthite-rich (up to An73) plagioclase in contact with the gedrite-sillimanite-quartz textures (Fig. 4f). In order to estimate thetemperature interval for the assemblage Na-Ged + Bt + Sil + St + Qtz,reactions (10)–(12) were modeled using the PERPLE X softwareassuming 1 wt.% H2O in the system at a pressure of 6.3 kbar.Fig. 19 shows the relative modal amounts of principal phases,i.e. cordierite, gedrite, staurolite, and biotite, in dependence ontemperature. The diagram predicts that Na-bearing gedrite is sta-ble within the wide temperature interval from above 850 downto 510 ◦C. This is consistent with the experimental and petro-logical facts that gedrite stability significantly increases with Nasubstitution (Fischer et al., 1999; Diener et al., 2008) and that Na-gedrite is stable at the UHT metamorphic conditions (cf. Kanazawaet al., 2009). The Na content of gedrite at temperatures 750–800 ◦Cis 0.63–0.66 a.p.f.u. (Fig. 19). Gedrite with such Na contents israrely present in the reaction textures in the metapelite SAF12-2/2 (Table 7), while the common Na concentration in gedrite in this

Fig. 19. Modal proportions of phases participating in reactions (10) and (11) (seetext) in dependence on temperature at 6.3 kbar and 1 wt.% H2O in the system calcu-lated using PERPLE X software. Gray solid line, modal proportions of cordierite. Blacksolid line, modal proportions of Na-bearing gedrite. Black long-dashed line, modalproportions of biotite. Black short-dashed line, modal proportions of staurolite. Val-ues along the gedrite line show Na content (a.p.f.u.) in gedrite at correspondingtemperature. Shaded box shows the possible temperature range for formation ofthe gedrite-sillimanite assemblage in the metapelite SAF12-2/2.


sample, 0.57–0.60 a.p.f.u. (Table 7), corresponds with the temper-ature interval 630–570 ◦C, where gedrite shows maximal modalcontent. Relics of cordierite can be present in this temperatureinterval, while staurolite begins to form. The calculated relationsare in a good agreement with the petrographic observations.Another for formation of the assemblage Ged + Bt + Sil + St + Qtz attemperatures below 600 ◦C is the presence of rare Al and Na-bearing anthophyllite in the gedrite-sillimanite aggregates (Fig. 11;Table 7). According to Spear (1980), the anthophyllite-gedritesolvus closes at about 600 ◦C. The temperature interval 630–570 ◦Cthus coincides with the temperatures of the possible solidus of thetrondhjemite magma.

8.4. P–T conditions of the orthopyroxene and garnet-bearingenclaves in metapelites

The Grt-Opx Fe–Mg equilibrium applied for the leucocra-tic orthopyroxene and garnet-bearing enclaves in metapelites(Fig. 2f) gives temperature values 810–820 ◦C. However, the morerobust equilibrium involving AlOpx component in orthopyroxene(Aranovich and Berman, 1997) defines temperatures up to 940 ◦C.The Grt-Opx-Pl-Qtz assemblage in these enclaves allows determi-nation of pressure at 7.1–7.5 kbar.

9. Discussion and model

9.1. Compilation of the P–T data, thermal effect of thetonalite-trondhjemite magmatism, and a geotectonic perspectivefor the SMZ

Table 10 and Fig. 20 compile the above P–T data and com-pare them with generalized P–T paths for the rocks of theSMZ (Van Reenen et al., 2011). The P–T estimates calculatedfrom the orthopyroxene-cordierite-biotite assemblages from themetapelites SAF12-2/2 and PET-5 (Fig. 18a–c) are shown as awide oval above the generalized decompression–cooling P–T pathfor the SMZ. As was shown above, this scatter of data resultedfrom strong Fe–Mg re-equilibration between minerals. We cannotexclude that temperatures for the orthopyroxene-cordierite-biotite assemblages from the metapelite were somewhat higher.Nevertheless, average pressures 8.5–9.0 kbar at temperatures750–800 ◦C (Fig. 19) are close enough to the generalizeddecompression–cooling P–T path. Evidence that the metapeliteunderwent decompression and cooling from higher P–T parame-ters is presented by cordierite-orthopyroxene clusters producedby decomposition of a precursor garnet (Fig. 4g).

The following group of P–T data for the metapelite includesestimates for orthopyroxene-sillimanite reaction textures pre-served in the metapelite PET-5 (Fig. 4h), which lie within550–650 ◦C at 5.3–5.7 kbar (Fig. 18d). The temperature inter-val 570–630 ◦C estimated for the gedrite-sillimanite-quartz-staurolite-biotite assemblage in the metapelite SAF12-2/2 (Fig. 4cand d) is very consistent with the temperatures calculated forthe orthopyroxene-sillimanite reaction textures in the metapelitePET-5. It shows that formation of these textures is a conse-quence of a single process. Simple connection between two setsof P–T data for metapelite defines a decompression–cooling pathsub-parallel to the generalized decompression–cooling P–T pathfor the SMZ (Fig. 20). However, such path cannot produce theorthopyroxene-sillimanite intergrowths since reaction (9) is dis-placed to the right either with increasing pressure or duringsub-isobaric cooling. The increase of pressure must be excludedsince the orthopyroxene-sillimanite textures are developed afterformation of the decompressional cordierite-orthopyroxene clus-ters. Thus, the assemblages Opx + Sil and Ged + Sil + Qtz + St + Bt

Fig. 20. Compilation of the P–T data for metapelite (empty ovals), garnet-bearingtrondhjemite (gray ovals), and the trondhjemite-metapelite reaction zone (blackboxes) compared to the generalized decompression cooling (black arrows) andsub-isobaric cooling (gray arrows) paths for the SMZ (Van Reenen et al., 2011).Gray dashed arrow shows a path for the Mg-Al-granulite DR19 (Belyanin et al.,2014). Empty rectangular box shows P–T parameters, which are calculated forOpx-Grt-Pl-Kfs-Qtz enclaves in the metapelite at farm Petronella. Small numbersshow ages of major stages of the SMZ evolution (2.72 Ga, peak of metamorphism;2.69 Ga, decompression–cooling; 2.68 Ga, average age of the Matok granite intru-sion; 2.62–2.60 Ga, sub-isobaric cooling). Large numbers show U/Pb zircon agesdetermined for the garnet trondhjemite at the Petronella locality (Fig. 1b) (Belyaninet al., 2014). Dashed P–T loop indicates the P–T evolution of underthrusted green-schist from the adjacent northern Kaapvaal Craton. The peak of metamorphism(∼2.68–2.69 Ga) coincides exactly with commencement of exhumation (∼2.69 Ga)in the SMZ (Kreissig et al., 2001).

after cordierite reflect the stage of sub-isobaric cooling of themetapelite at pressures 5.3–6.3 kbar. At this pressures and temper-atures below 600 ◦C, the sub-isobaric P–T path crosses the Sil-Kyphase boundary (Fig. 20). Although kyanite has not been found inthe present samples, several studies (e.g. Van Reenen, 1986; VanReenen et al., 2011) report kyanite in the late hydration productsof the SMZ metapelites.

The sub-isobaric cooling path at 5.3–6.3 kbar demands that themetapelite must be uplifted to about 18–20 km and simultaneouslyheated. Such path can be realized by dragging of the metapeliteby the hot exhuming trondhjemite magma, which reached a tem-perature of 1000 ◦C (Fig. 20). There is no record of this processpreserved in the studied metapelite. However, orthopyroxene andgarnet-bearing leucocratic enclaves (Fig. 2f) in metapelites that out-crops in the Sand River just south of the studied locality can bereadily interpreted as results of high-temperature influence of thehot trondhjemite magmas. The presence of abundant K-feldsparcombined with the results of thermobarometry implies that theenclaves could be products of local partial melting of biotite-richmetapelite at high temperatures, which, probably, had been trig-gered by the intruding trondhjemite en route during their jointexhumation with the metapelite. Our preliminary experimentaldata on dehydration partial melting of metapelite PET-5 at 8 kbarshowed that this rock produced K-rich granitic melt and new Al-rich orthopyroxene at 900 ◦C. This is additional proof that theenclaves are the product of the partial melting of metapelite causedby heating.

Exhumation at high temperatures is recorded in the polyphasesillimanite-corundum-spinel inclusions in GRT-II from the trond-hjemites, and specifically by reaction (4), which proceeds to theright with heating and decompression. This implies that heating


Table 10Compilation of the thermobarometric results for garnet trondhjemite and the reaction zone between the trondhjemite and metapelite.

Rock type Mineral assemblage Temperature (◦C) Pressure (kbar) P–T diagram

Garnet trondhjemite Ternary feldspar (Fsp) 850–1000 – Fig. 13Garnet trondhjemite GRT-II + Cor + Sil + Fsp + melt 970–880 7.5–6a Fig. 14aGarnet trondhjemite GRT-II + Cor + Sil + Spl + Fsp 860–890 6.2–6.3 Fig. 14b and cGarnet trondhjemite GRT-I (II) + Bt + Pl + Qtz 640–650 6.0–6.5 Fig. 14d and eReaction zone rock Grt + Opxb + Fsp + Qtz 820 ± 5 6.2–6.3 Fig. 15aReaction zone rock Grt + Opx + Crd + Pl + Qtz 620 ± 5 5.7–5.8 Fig. 15aReaction zone rock Grt + Bt (centers) + Pl + Qtz 660–680 5.7–7.5Reaction zone rock Opx + Bt + Qtz 660 –Reaction zone rock Grt + Bt (contacts, inclusions) 570–630 – Fig. 15bReaction zone rock Grt + Bt + Pl + Qtzc 610–640 5.5–6.5 Fig. 15cReaction zone rock Grt + Bt + Ilm + Ru + Qtz 610 6.2 Fig. 15b

a Assumed pressure.b The equilibrium Alm = 3Fs + AlOpx is used for temperature calculation, GRT-I and GRT-II – morphological types of garnets (see text).c Compositions of the late Ca-rich (1.0–1.4 wt.% CaO) idiomorphic garnet in contacts with the intergrowths Bt + Pl + Qtz + Kfs.

caused by intrusion of the exhuming hot magma resulted in for-mation of high-aluminous assemblages in the trapped metapelite.Although high-aluminous metapelitic assemblages are volumetri-cally insignificant (see reviews in Harley, 2008; Kelsey, 2008), suchassemblages are specific features of many Gondwanian granulitecomplexes of Southern India (cf. Bose et al., 2000; Santosh andSajeev, 2006; Rickers et al., 2001 and many references in thesepapers), Antarctica (Kelly and Harley, 2004 and references therein),and South Africa (Belyanin et al., 2010, 2012, 2014). Most authorsagree that these unique assemblages represent evidence for high toultra-high temperature metamorphism. However, metamorphismat temperatures above 900 ◦C definitely demands additional heatinput (Harley, 2008; Kelsey, 2008), which could be provided bymagmas intruding rocks under the granulite facies conditions. Therole of magmas in the evolution of the high-Al assemblages fromSouthern India has been stressed by Rickers et al. (2001) and Boseet al. (2000). Rickers et al. (2001, p. 563) concluded that the evo-lution of the Al-rich granulites of the Eastern Ghats Belt (SouthernIndia) “may be related to heating of the crust through magmaticaccretion. . ., fast uplift of the UHT granulites into mid-crustallevels. . ., emplacement of felsic magmas. . .resulting in reheatingof the crust to high-T conditions. . . and period of cooling to thestable geotherm. . .”. Bose et al. (2000) inferred a heating–coolingP–T trajectory at lower crustal depths for similar assemblages andargued that magmatic underplating provided the necessary heatinput. Our present results are consistent with these models andextend them through the conclusions that hot trondhjemite mag-mas served both as heat suppliers and as an additional driving forcefor exhumation of the heated granulites from the levels of 24–28 km(8–9 kbar) to the levels of about 16–19 km (5.3–6.5 kbar).

This geodynamic model is similar to the model accounting thevertical tectonic affected the entire Central Zone and northern por-tions of the SMZ of the Limpopo Complex (north of the Annaskraalshear zone, Fig. 1a). This process was accompanied by granitediapirism (Roering et al., 1992; Perchuk et al., 2008; Perchuk andGerya, 2011). Huizenga et al. (2011) described the P–T and fluid evo-lution of metapelite that interacted with granitic melts of the largeBulai pluton (2.612 ± 0.7 Ga) in the Central Zone. These authorsshowed that metapelitic enclaves in these granites also experi-enced sub-isothermal heating by the magma by about 50 ◦C atpressure about 5.5 kbar and accompanied by CO2 and brine fluids.The age of the granite diapirism in the Central Zone (∼2.69–2.63 Ga)is very close to the age of the trondhjemite bodies in the SMZ(2.64–2.667 Ga; Belyanin et al., 2014), which is, in turn, close tothe age of the large magmatic event represented by the Matokpolyphase pluton (Fig. 1a). It thus cannot be excluded that trond-hjemite intrusions might represent satellite bodies related to theMatok Complex. This implies that vertical tectonics and granitediapirism also affected the SMZ. The area south of the Annaskraal

shear-zone (Fig. 1a) mainly shows evidence for the isobaric-coolingstage of evolution (Perchuk et al., 1996, 2000), which has beeninterpreted to reflect evidence for major horizontal channeling ofthe rocks (thrusting at the middle crustal level) during the thrust-controlled emplacement of the SMZ onto the Kaapvaal craton.Nevertheless, the sub-isobaric cooling stage of evolution of the SMZmight also be related to the cooling effect of the reheated rocks.Table 10 summarizes the results of thermobarometry of the garnettrondhjemite and the trondhjemite-metapelite reaction zone andclearly indicates that the trondhjemite intrusion was affected bycooling from about 1000 ◦C to below 600 ◦C at nearly constant aver-age pressure about 5.7–6.5 kbar. This stage of evolution coincideswith sub-isobaric cooling as is recorded in metapelite (Fig. 20). Con-sistent with these data, we have recently shown that the isobariccooling evolution from a temperature above 900 ◦C at pressuresabout 6.5–7.0 kbar is also true for Mg-Al-assemblages preserved insample DR19 from the SMZ (Belyanin et al., 2014; Fig. 20).

9.2. Assimilation of metapelite by the trondhjemite magma

In addition to the thermal effect of the hot trondhjemite magma,its interaction with the country metapelite is clearly expressedby a process of assimilation. Assimilation can be described by theschematic reaction (Opx + Crd + Bt) + melt1 = Grt ± Sil ± Cor + melt2,where transition from melt1 to melt2 is expressed in saturationof the trondhjemite melt in the “metapelite components” such asMgO, FeO, Al2O3, TiO2. Pseudosections constructed for two differ-ent bulk compositions of the garnet trondhjemite (Fig. 17a–c) allowthe conclusion that variability of mineral assemblages in the rockwas primarily determined by variations in the bulk composition atsimilar cooling P–T paths. The variation of bulk composition reflectsheterogeneity of the trondhjemite melt dissolving metapelite,which evidently reflects unequal distribution of MgO, FeO, Al2O3,and TiO2 in this magma. The morphological garnet type GRT-II rep-resents portions of the melt that respectively was depleted in MgOand FeO and Al2O3-oversaturated (PET-6 in Table 1). The lowerMg-numbers of garnets recorded in cores of GRT-II (Fig. 7a) implyboth unequal spatial and temporal distribution of the Fe/Mg ratiowith the magmatic body. It is logical to assume that dissolutionof the metapelitic material might have started from stronger dif-fusion of Fe into the melt, whereas Mg has been concentrated inthe residuum. The slight increase of Mg-number of orthopyroxeneand cordierite preserved in the trondhjemite-metapelite reactionzone proves this suggestion. Further assimilation allowed more Mgto be introduced into the melt and, thus, crystallization of moreMg-rich garnets. GRT-I was forming in melts richer in SiO2, MgOand FeO (PET-10 and PET-11 in Table 1), whereas garnet-biotiteassemblages of the trondhjemite-metapelite reaction zone reflectoversaturation of the melt in MgO and FeO (PET-1 and PET-2 in


Table 1). Such process of dissolution of “foreign” (metapelitic) rocksin the granitic melts might be promoted by the sub-isothermaldecompression stage of evolution of the crust (e.g. García-Morenoet al., 2006).

9.3. Fluids in the trondhjemite melt and their effect on thetrapped metapelites

According to the experimental data by Johannes (1978),the H2O-saturated trondhjemite melts (normative plagioclaseAn15–20 + quartz) can exist at pressures 5–6 kbar within the tem-perature interval down to 680–690 ◦C. Addition of even minorMgO, FeO, K2O, as well as other volatiles (Cl, F), would expandthe stability of the melt to even lower temperatures. Pseu-dosections in Fig. 17a–c predict that the H2O-saturated meltexists down to 600–610 ◦C in systems corresponding to the bulkcompositions PET-6 and PET-11 (Table 1). Comparison of the pseu-dosections in Fig. 17a and b indicates that variations of the H2Ocontent strongly influenced mineral relations (garnet-sillimanite,muscovite-sillimanite, exsolution of feldspars) that crystallizedfrom the melt. The possible unequal distribution of fluids within thetrondhjemite melt during its cooling is recorded in the assemblagesincluded in GRT-I and GRT-II. Abundant sillimanite inclusions inGRT-II would imply dryer conditions. In contrast, the assemblageMu + Zo found among inclusions in GRT-I (Fig. 3d and g) implies ahigh degree of fluid saturation. Johannes (1984) showed that thereaction

4An (in Pl) + Or + H2O = Mu + 2Zo + 2Qtz (13)

intersects the H2O-saturated solidus of the granitic KCNASH sys-tem at pressure about 7 kbar and temperature 640 ◦C. However,this assemblage crystallized in the presence of garnet, which woulddisplace reaction (13) to lower pressures (Schliestedt and Johannes,1984) consistent with the results of our thermobarometric study.The crystallizing muscovite is characterized by high celadonite con-tent (Table 6; Fig. 10a and b). Zane and Rizzo (1999) paid attention tothe fact that muscovite with high celadonitic component content ischaracteristic for K-feldspar bearing granites with moderately highA/(CNK) index, rather than for highly peraluminous potassium-poor granites such as the studied trondhjemite. The only way toexplain this disagreement is to suggest that crystallization of theceladonite-rich muscovite occurred at the latest stages of the meltevolution after strong fractional differentiation, which led to strongaccumulation of SiO2 and K2O, as well as H2O and other volatilecomponents in the residual melt. The formation of the celadonitecomponent in muscovite in the garnet-bearing trondhjemite canthus be described by the following reaction

(1/9)Mu + (1/3)Prp + (8/3)Qtz + (4/9)K2O + (8/9)H2O = Cel

(14)

which indicates increase of K2O and H2O activities in the melt andin the coexisting fluids. It is evident that the increase of the alkaliactivity is related to accumulation of the salt components in the flu-ids. Increasing K2O activity is expressed in formation of K-feldsparrims at plagioclase-quartz boundaries (see Fig. 3d) (see Safonov andAranovich, 2014 for review).

Presence of salt components in the fluids coexisting with thetrondhjemite melt is supported by the presence of saline (up to20.58 wt.% NaCl equivalent) fluid inclusions in quartz (Fig. 13d).These inclusions coexist with CO2 inclusions, which are presentboth in quartz and GRT-I (Fig. 12a–c). This coexistence is anotherrepresentation of relatively high salinity of the aqueous fluids(Shmulovich and Graham, 2004; Aranovich et al., 2010). Themaximal densities of the primary and primary-secondary inclu-sions are consistent with the pressures calculated from mineral

Fig. 21. Isochores for primary CO2 inclusions (� = 1.054 g/cm3) and primary-secondary inclusions (� = 1.085 g/cm3) in garnet (solid lines), and for theprimary-secondary inclusions (� = 1.087 g/cm3) in quartz (dashed line), comparedto the major stages of the rock evolution (I, peak conditions for host metapelite; II,conditions of interaction with the tonalite magma; III, cooling and magma crystal-lization).

barometry (Fig. 21). It means that post-entrapment modificationsof the inclusions were insignificant. Nevertheless, isochores cor-responding to the densities of the CO2 inclusions are situatedbelow the P–T conditions calculated for the host orthopyroxene-cordierite-biotite metapelites, 750–800 ◦C and 8.5–9.5 kbar (Fig. 21,Stage I). They also define significantly higher pressures for tem-peratures 860 ± 5 ◦C in comparison to the pressures calculated forthe stage II (Fig. 21) of the metapelite interaction with the tonalitemagma (6.1 ± 0.4 kbar). However, the isochores perfectly fit theconditions about 600 ◦C and 5.7–6.5 kbar estimated for the stage IIIof final cooling and magma crystallization (Fig. 21). The above maxi-mal concentration of NaCl measured in fluid inclusions correspondsto just 7 mol.% of salt in the fluid. According to the mixing modelfor the H2O-NaCl fluid by Aranovich et al. (2010) this concentrationimposes a H2O activity about 0.83 at 6.3 kbar and 600 ◦C. How-ever, CO2 lowers H2O activity. Water activity during crystallizationof the trondhjemite melts was estimated using compositions ofcoexisting garnet, biotite and plagioclase. K-feldspar appears in thetrondhjemite only at the latest stages of the evolution forming exso-lution lamellae, rims at plagioclase-quartz contacts or rare grainsassociated with the plagioclase-quartz graphic-like intergrowths.However, plagioclases in the trondhjemite initially contained highconcentrations of the orthoclase component, now expressed asabundant exsolution lamellae (Table 9). Even lamellae-free plagio-clases in the trondhjemite contain up to 0.55 wt.% K2O (Table 9).This allows application of the orthoclase component in plagioclasefor calculation of water activity. In the present calculations, themodel by Elkins and Grove (1990) was used to compute the activ-ity of the orthoclase component in plagioclases. The results of suchcalculation vitally depend on the activity of the orthoclase com-ponent: the higher its activity, the lower is the water activity inthe fluid. For calculation of the water activity in the trondhjemite-metapelite reaction zone (Fig. 22), the re-integrated compositionsof lamellae-bearing plagioclases from this zone were used, sincethe average orthoclase content in these plagioclases, about 6 mol.%(Table 9), is very consistent with the garnet-biotite temperatures680–650 ◦C. For calculation of the water activity in the trondhjemite(Fig. 22), the compositions of lamellae-free plagioclases contain-ing 0.3–0.5 wt.% K2O (Table 9), were used. The calculated range ofthe water activity for the above temperature interval at 6.3 kbaris 0.31–0.17 (Fig. 22). These relatively low water activities implya high CO2 content of the fluid (probable XH2O = 0.3–0.4). Numer-ous studies (Van Reenen, 1986; Van Reenen and Hollister, 1988;Baker et al., 1992; Van den Berg and Huizenga, 2001; Huizengaet al., 2014) showed that fluids with exactly the same chemical


Fig. 22. Water activity in the H2O-CO2 fluid coexisting with the assemblage gar-net + biotite + quartz + plagioclase in the garnet trondhjemite (gray dot defined bythe intersection of dashed reaction curves), and in the trondhjemite-metapelitereaction zone (black dot defined by intersection of solid reaction curves) calculatedfor 6.3 kbar.

characteristics, i.e. immiscible CO2-rich fluids and aqueous brines,participated in the stabilization of the orthoamphibole isograd inthe SMZ at pressure about 6 kbar and temperatures below 650 ◦Cat the close of the Limpopo Orogeny (2.62 Ga). The mole fractionof H2O in the CO2-rich fluid phase was measured to vary within0.3–0.1. The fluids were, probably, generated by devolatilizationof underthrusted greenstones as they were heated by the over-thrusted granulites. The relatively low density of the CO2-rich fluidinclusions in the hydrated granulites indicates that the fluid pres-sure was less than the lithostatic pressure, indicating promotionof the rock permeability by hydration reactions (Huizenga et al.,2014).

Our present study evidently demonstrates that in addition tothe tectonically weakened zones (shear zones), another plausiblemedia for transferring fluids that participated in the re-hydrationof the SMZ metapelite is the trondhjemite magmas. The connec-tion between the trondhjemite intrusions and the re-hydration ofgranulites is substantiated by age data and comparable P–T con-ditions. P–T conditions measured for the trondhjemite-metapelitefluid–rock interaction at Petronella are the same as those sug-gested for the establishment of the isograd (Van Reenen, 1986; VanReenen and Hollister, 1988; Baker et al., 1992; Van den Berg andHuizenga, 2001; Huizenga et al., 2014). Being expelled from thetrondhjemite magma during cooling and crystallization, the fluidsinteracted with trapped and surrounding metapelite. The productsof this interaction reflect the composition of the fluids. Presence ofNa-rich gedrite, Ti-poor biotite, and sodic plagioclase show that inaddition to the “isochemical” reactions (10)–(12), the formation ofthese phases was manifested by reactions governed by activities ofalkalis, for example:

Crd + (7/15)Na2O + (1/3)H2O = (1/3)Na-Ged + (3/5)Ab + (6/5)Sil (15)

Crd + (1/6)Na2O + (1/3)H2O = (1/3)Na-Ged + (3/2)Qtz + (3/2)Sil (16)

Crd + (16/5)Sil + (7/10)Na2O + (1/2)H2O = (7/5)Ab + (1/2) St (17)

Crd + (1/5)K2O + (3/5)H2O = (1/5)St + (2/5)Phl + (11/5)Qtz (18)

Crd + (1/3)K2O + (2/3)H2O = (2/3)Phl + (4/3)Qtz + (10/6)Sil (19)

Reaction (17) represents formation of plagioclase rims oncordierite accompanied by delicate crystallites of staurolite(Fig. 4d), while reaction (15) describes formation of plagio-clase in the assemblage with Na-gedrite. The product plagioclasewith XCa = 0.39–0.45 is clearly different from the calcic plagio-clase (XCa = 0.6–0.75) produced by the “isochemical” reaction (10)(Table 9; Fig. 12). The composition of the plagioclase also indicatesparticipation of calcium in the reactions. There is no doubt that suchNa and Ca-rich fluids (presumably, NaCl and CaCl2) could easily be

produced by Na- and Ca-rich trondhjemite melts. Biotite formedin the retrograde reaction textures represents a typical productof cordierite interaction with K-bearing fluids (see experimentalexample in Safonov and Aranovich, 2014).

Despite high CO2 content in the fluids expelled by thetrondhjemite magmas, there is no evidence of carbonation orgraphitization in the studied metapelite. Absence of graphite inthe trondhjemite suggests that the oxygen fugacity during thetrondhjemite–metapelite interaction was above the CCO buffer.It is interesting to note that similar garnet-bearing leucocratictrondhjemite-granite rocks at the Bandelierkop locality (Fig. 1a)contain abundant graphite along with CO2 fluid inclusions, but veryrare aqueous inclusions. Relatively Fe-rich garnet-orthopyroxenemetapelite at this locality bears evidence for late carbonation(formation of late Fe-bearing dolomite). These data show thatessentially carbonic fluids accompanied mixing of the magmawith metapelite at the Bandelierkop locality, a process that isproven by oxygen isotope data (E.O. Dubinina and L.Ya. Ara-novich, personal communications). All available data thus indicatesthat trondhjemite magma that intruded the SMZ were, probably,heterogeneous with respect to the fluid budget. However, theseconclusions demand further study of trondhjemite bodies at otherlocalities within the SMZ.

10. Conclusions and perspectives

The present study indicates that at about 2.66–2.67 Ga, theSMZ was invaded by numerous fluidized magmas of trondhjemite,tonalite and granite composition. The age of these intrusions isslightly younger than the age of the main stage of the SMZ exhuma-tion (2.69 Ga) (Fig. 1a). This implies a close relationship of thesemagmas with commencement of interaction of overriding gran-ulite with underthrusted greenstone rocks of the NKVC. Althoughthis is beyond the scope of the present study, we can speculate thatthe buried rocks of the NKVC served as a source for these magmas.Their bulk composition, namely high contents of Al2O3, CaO andNa2O combined with the low K2O content, suggests a metabasalticor amphibolitic source for the trondhjemite. Experimental studieson dehydration and excess-water melting of amphibolitic rocks ofbasaltic composition (cf. Rapp et al., 1991; Winther, 1996) showthat this type of material is able to produce tonalite-trondhjemitemelts within the wide pressure interval (8–32 kbar) at tempera-tures 900–1000 ◦C. Our study shows that the trondhjemite meltsintruded the Petronella metapelite at similar temperature con-ditions. In addition, the dehydration melting of amphibolite (themechanism, which is most plausible at the granulite-greenstonecontact) produces 1–2 wt.% of aqueous fluids (Rapp et al., 1991),which readily can be transported upwards by the melts.

The present study demonstrated four major effects of the trond-hjemite melts that intruded granulites of the SMZ.

1. Geodynamic effect. The trondhjemite melts derived from the baseof the granulite pile or from the underthrusted greenstone mate-rial assisted in the exhumation of separate blocks of the SMZgranulites. Taking into account the notable volume of these meltsin the SMZ (Fig. 1a), the trondhjemite melts would provide acritical contribution to the exhumation of granulites onto theadjacent granite-greenstone craton. This suggestion might alsoinspire the numerical modeling of the joint uplift of the trond-hjemite intrusions and the metamorphic rocks, similar as hasbeen done for granite diapirism in the Central Zone (Perchuket al., 2008).

2. Thermal effect. The hot trondhjemite melts transferred additionalheat from the lower to the middle crust and served as an addi-tional source of heat for granulite metamorphism during the


retrograde stage. Heat produced by magmas can be an explana-tion for local partial melting of granulites and formation of UHTassemblages. This conclusion might inspire a study of the con-nection between granite-tonalite-trondhjemite intrusions andUHT phenomena in the SMZ (Tsunogae et al., 2004; Belyaninet al., 2010, 2012).

3. Assimilation effect. The hot trondhjemite melts actively assimi-lated country granulite rocks assisting in the mixing of granulitematerial with the cratonic material. This suggests a study of thegeochemical and isotopic (E.O. Dubinina and L.Ya. Aranovich,personal communications) signatures of such mixing and pro-vides a broad field for the experimental modeling of rock–meltinteraction at high-grade metamorphism.

4. Fluid effect. The trondhjemite melts transported or served as con-duits for large volumes of fluids produced by devolatilizationof the greenstone rocks. These fluids subsequently rehydrated asignificant portion of the SMZ. The present study thus for the firsttime provides clear evidence for a link between the retrogradehydration of SMZ granulites and the trondhjemite intrusions andcalls for further examination of this phenomenon with respectto compositional variability of the fluids.

Acknowledgments

We thank two reviewers who highly estimated our study, gavecomments, strong (and justified) criticism and useful suggestions.The authors thank all participants of the Limpopo field seminar in2013 for the fruitful discussions and comments on the preliminaryresults of the present study. The study is supported by the Rus-sian Scientific Foundation (project 14-17-00581 to OGS), partiallyby Russian Foundation for Basic Research (project 13-05-00353 toOGS), by grants from the National Science Foundation of SouthAfrica (GUN: 2053192 to DDvR) and the University of Johannesburgas part of the Russian-South African scientific collaboration. Wethank Konstantin Van, Dmitrii Varlamov, Alexei Nekrasov (groupof microprobe analysis and microscopy, IEM RAS) and GeorgeBelyanin (University of Johannesburg) for assisting with the micro-probe analyses. George Belyanin and Christian Reinke (SPECTRAU,University of Johannesburg) are thanked for bulk chemistry analy-ses of the samples. Pavel Pletchov (Department of Petrology, MSU)for providing facilities for the Raman analysis.

References

Aranovich, L.Y., Berman, R.G., 1997. A new garnet-orthopyroxene thermometerbased on reversed Al2O3 solubility in FeO-Al2O3-SiO2 orthopyroxene. Am. Min-eral. 82, 345–353.

Aranovich, L.Ya., Zakirov, I.V., Sretenskaya, N.G., Gerya, T.V., 2010. Ternary systemH2O–CO2–NaCl at high P–T parameters: an empirical mixing model. Geochem.Int. 48, 446–455.

Aranovich, L.Y., Newton, R.C., Manning, C.E., 2013. Brine-assisted anatexis: exper-imental melting in the system haplogranite–H2O–NaCl–KCl at deep-crustalconditions. Earth Planet. Sci. Lett. 374, 111–120.

Baker, J., Van Reenen, D.D., Van Schalkwyk, J.F., Newton, R.C., 1992. Constraints onthe composition of fluids involved in retrograde anthophyllite formation in theLimpopo Belt, South Africa. Precambrian Res. 55 (1), 327–336.

Barton Jr., J.M., Van Reenen, D.D., 1992. The significance of Rb–Sr ages of biotite andphlogopite for the thermal history of the Central and SMZs of the Limpopo Beltof southern Africa and the adjacent portions of the KC. Precambrian Res. 55,17–31.

Belyanin, G.A., Rajesh, H.M., Van Reenen, D.D., Mouri, H., 2010. Corun-dum + orthopyroxene ± spinel intergrowths in an ultrahigh-temperature Al–Mggranulite from the Southern Marginal Zone, Limpopo Belt, South Africa. Am.Mineral. 95 (1), 196–199.

Belyanin, G.A., Rajesh, H.M., Sajeev, K., Van Reenen, D.D., 2012. Ultrahigh-temperature metamorphism from an unusual corundum + orthopyroxeneintergrowth bearing Al–Mg granulite from the Southern Marginal Zone,Limpopo Complex, South Africa. Contrib. Mineral. Petrol. 164, 457–475.

Belyanin, G., Van Reenen, D.D., Safonov, O.G., 2014. Response to comments by Nicoliet al. on paper by Belyanin et al. (2012). Contrib. Mineral. Petrol. 167, 1–5.

Berman, R.G., 1988. Internally-consistent thermodynamic data for stoichiometricminerals in the system Na2O-K2O-CaO-MgO-FeO-Fe2O3-Al2O3-SiO2-TiO2-H2O-CO2. J. Petrol. 29, 445–522.

Berman, R.G., 2007. WinTWQ (version 2.3): a software package for perform-ing internally-consistent thermobarometric calculations. Geological Survey ofCanada Open File 5462.

Berman, R.G., Aranovich, L.Ya., 1996. Optimized standard state and solution prop-erties of minerals. I. Model calibration for olivine, orthopyroxene, cordierite,garnet, and ilmenite in the system FeO-MgO-CaO-Al2O3-TiO2-SiO2. Contrib.Mineral. Petrol. 126, 1–24.

Bohlender, F., Van Reenen, D.D., Barton Jr., J.M., 1992. Evidence for metamorphicand igneous charnockites in the Southern Marginal Zone of the Limpopo Belt.Precambrian Res. 55, 429–449.

Bose, S., Fukuoka, M., Sengupta, P., Dasgupta, S., 2000. Evolution of high-Mg-Al gran-ulites from Sunkarametta, Eastern Ghats, India: evidence for a lower crustalheating–cooling trajectory. J. Metamorph. Geol. 18 (3), 223–240.

Connolly, J.A.D., 2005. Computation of phase equilibria by linear programming: atool for geodynamic modeling and its application to subduction zone decarbon-ation. Earth Planet. Sci. Lett. 236, 524–541.

De Beer, J.H., Stettler, E.H., 1992. The deep structure of the Limpopo Belt. PrecambrianRes. 55, 173–186.

Diener, J.F.A., Powell, R., White, R.W., 2008. Quantitative phase petrology ofcordierite–orthoamphibole gneisses and related rocks. J. Metamorph. Geol. 26(8), 795–814.

Du Toit, M.C., Van Reenen, D.D., Roering, C., 1983. Some aspects of the geology,structure and metamorphism of the Southern Marginal Zone of the LimpopoMetamorphic Complex. In: Van Biljon, W.J., Leg, J.H. (Eds.), The Limpopo Belt.Geological Society of South Africa Special Publication 8, pp. 89–102.

Ebadi, A., Johannes, W., 1991. Beginning of melting and composition of first melt inthe system Qz-Ab-Or-H2O-CO2. Contrib. Mineral. Petrol. 106, 286–329.

Elkins, L.T., Grove, T.L., 1990. Ternary feldspar experiments and thermodynamicmodels. Am. Mineral. 75, 544–559.

Fischer, H., Schreyer, W., Maresch, W.V., 1999. Synthetic gedrite: a stable phase in thesystem MgO-Al2O3-SiO2-H2O (MASH) at 800 ◦C and 10 kbar water pressure, andthe influence of FeNaCa impurities. Contrib. Mineral. Petrol. 136 (1-2), 184–191.

García-Moreno, O., Castro, A., Corretgé, L.G., El-Hmidi, H., 2006. Dissolution oftonalitic enclaves in ascending hydrous granitic magmas: an experimentalstudy. Lithos 89 (3), 245–258.

Gerya, T.V., Perchuk, L.L., 1990. GEOPATH – A Thermodynamic Database for Geother-mobarometry and Related Calculations with the IBM PC computer. Program andAbstract. Calgary University Press, Calgary, Canada, pp. 59–61.

Harley, S.L., 2008. Refining the P–T records of UHT crustal metamorphism. J. Meta-morph. Geol. 26 (2), 125–154.

Huizenga, J.M., Perchuk, L.L., van Reenen, D.D., Flattery, Y., Varlamov, D.A., Smit, C.A.,Gerya, T.V., 2011. Granite emplacement and the retrograde PT-fluid evolutionof Neoarchean granulites from the Central Zone of the Limpopo Complex. Geol.Soc. Am. Mem. 207, 125–142.

Huizenga, J.M., Van Reenen, D.D., Touret, J.L.R., 2014. Fluid-rock interaction inretrograde granulites of the Southern Marginal Zone, Limpopo high-grade ter-rain, South Africa. Geosci. Front., http://dx.doi.org/10.1016/j.gsf.2014.01.004 (inpress).

Johannes, W., 1978. Melting of plagioclase in the system Ab–An–H2O andQz–Ab–An–H2O at PH2O = 5 kbars, an equilibrium problem. Contrib. Mineral.Petrol. 66 (3), 295–303.

Johannes, W., 1984. Beginning of melting in the granite system Qz-Or-Ab-An-H2O.Contrib. Mineral. Petrol. 86 (3), 264–273.

Kanazawa, T., Tsunogae, T., Sato, K., Santosh, M., 2009. The stability and origin ofsodicgedrite in ultrahigh-temperature Mg–Al granulites: a case study from theGondwana suture in southern India. Contrib. Mineral. Petrol. 157 (1), 95–110.

Kelly, N.M., Harley, S.L., 2004. Orthopyroxene–corundum in Mg–Al-rich granulitesfrom the Oygarden Islands, East Antarctica. J. Petrol. 45 (7), 1481–1512.

Kelsey, D.E., 2008. On ultrahigh-temperature crustal metamorphism. GondwanaRes. 13 (1), 1–29.

Kreissig, K., Holzer, L., Frei, I.M., Villa, J.D., Kramers, J.D., Kröner, A., Smit, C.A., VanReenen, D.D., 2001. Geochronology of the Hour River Shear Zone and the meta-morphism in the Southern Marginal Zone of the Limpopo belt, Southern Africa.Precambrian Res. 109, 145–173.

Le Maitre, R.W., Streckeisen, A., Zanettin, B., Le Bas, M.J., Bonin, B., Bateman, P., Bel-lieni, G., Dudek, A., Efremova, S., Keller, J., Lameyre, J., Sabine, P.A., Schmid, R.,Sörensen, H., Wooley, A.R., 2002. Igneous Rocks – A Classification and Glossaryof Terms. Recommendations of the IUGS Subcommission on the Systematics ofIgneous Rocks, 2nd ed. Cambridge University Press, Cambridge.

Miller, C.F., Stoppard, E.F., Bradfish, L.J., Dollase, W.A., 1981. Composition of plutonicmuscovite: genetic implications. Can. Mineral. 19, 25–34.

Perchuk, L.L., Gerya, T.V., 2011. Formation and evolution of Precambrian gran-ulite terranes: a gravitational redistribution model. Geol. Soc. Am. Mem. 207,289–310.

Perchuk, L.L., Gerya, T.V., Van Reenen, D.D., Safonov, O.G., Smit, C.A., 1996. TheLimpopo Metamorphic Belt, South Africa: 2. Decompression and cooling regimesof granulites and adjacent rocks of the Kaapvaal Craton. Petrology 4, 571–599.

Perchuk, L.L., Gerya, T.V., Van Reenen, D.D., Smit, C.A., Krotov, A.V., Safonov, O.G.,2000. Comparative petrology and metamorphic evolution of the Limpopo (SouthAfrica) and Lapland (Fennoscandia) high grade terrains. Mineral. Petrol. 69,69–107.

Perchuk, L.L., Van Reenen, D.D., Smit, C.A., Boshoff, R., Belyanin, G.A., Yapaskurt, V.O.,2008. Role of granite intrusions for the formation of ring structures in granulite

http://refhub.elsevier.com/S0301-9268(14)00197-1/sbref0005









































































































































































































































































































































































































































































































































































































































dx.doi.org/10.1016/j.gsf.2014.01.004














































































































































































































































































complexes: examples from the Limpopo belt, South Africa. Petrology 16 (7),652–678.

Putnis, A., 2002. Mineral replacement reactions: from macroscopic observations tomicroscopic mechanisms. Mineral. Mag. 66 (5), 689–708.

Rapp, R.P., Watson, E.B., Miller, C.F., 1991. Partial melting of amphibolite/eclogiteand the origin of Archean trondhjemites and tonalites. Precambrian Res. 51 (1),1–25.

Retief, E.A., Compston, W., Armstrong, R.A., Williams, I.S., 1990. Characteristicsand preliminary U–Pb ages of zircon from Limpopo Belt lithologies. ExtendedAbstracts, the Limpopo Belt. In: A Field Workshop on Granulites and Deep CrustalTectonics, June 21–July 1, 1990, Rand Afrikaans University (now University ofJohannesburg).

Rickers, K., Raith, M., Dasgupta, S., 2001. Multistage reaction textures in xenolithichigh-MgAl granulites at Anakapalle, Eastern Ghats Belt, India: examples of con-tact polymetamorphism and infiltration-driven metasomatism. J. Metamorph.Geol. 19 (5), 563–582.

Roering, C., van Reenen, D.D., Smit, C.A., Barton Jr., J.M., de Beer, J.H., de Wit, M.J.,Stettler, E.H., van Schalkwyk, J.F., Stevens, G., Pretorius, S., 1992. Tectonic modelfor the evolution of the Limpopo Belt. Precambrian Res. 55 (1–4), 539–552.

Safonov, O.G., Aranovich, L.Y., 2014. Alkali control of high-grade metamorphism.Geosci. Front., http://dx.doi.org/10.1016/j.gsf.2014.03.010 (in press).

Santosh, M., Sajeev, K., 2006. Anticlockwise evolution of ultrahigh-temperaturegranulites within continental collision zone in southern India. Lithos 92 (3),447–464.

Schliestedt, M., Johannes, W., 1984. Melting and subsolidus reactions in the sys-tem K2O-CaO-Al2O3-SiO2-H2O: corrections and additional experimental data.Contrib. Mineral. Petrol. 88 (4), 403–405.

Shmulovich, K.I., Graham, C.M., 2004. An experimental study of phase equilibriain the systems H2O-CO2-CaCl2 and H2O-CO2-NaCl at high pressures and tem-peratures (500–800 ◦C, 0.5–0.9 GPa): geological and geophysical applications.Contrib. Mineral. Petrol. 146, 450–462.

Smit, C.A., Van Reenen, D.D., Gerya, T.V., Varlamov, D.A., Fed’kin, A.V., 2000.Structural-metamorphic evolution of the Southern Yenisey Range of EasternSiberia: implications for the emplacement of the Kanskiy granulite Complex.Mineral. Petrol. 69 (1–2), 35–67.

Smit, C.A., Van Reenen, D.D., Gerya, T.V., Perchuk, L.L., 2001. P–T conditions ofdecompression of the Limpopo high-grade terrain: record from shear zones.J. Metamorph. Geol. 19, 249–268.

Spear, F.S., 1980. The gedrite-anthophyllite solvus and the composition limits oforthoamphibole from the Post Pond Volcanics, Vermont. Am. Mineral. 65,1103–1118.

Stevens, G., 1997. Melting carbonic fluids and water recycling in the deep crust, anexample from the Limpopo Belt, South Africa. J. Metamorph. Geol. 15, 141–154.

Stevens, G., van Reenen, D.D., 1992. Partial melting and the origin of metapeliticgranulites in the Southern Marginal Zone of the Limpopo Belt, South Africa.Precambrian Res. 55, 303–319.

Tsunogae, T., Miyano, T., Van Reenen, D.D., Smit, C.A., 2004. Ultrahigh-temperaturemetamorphism of the southern marginal zone of the Archean Limpopo Belt,South Africa. J. Mineral. Petrol. Sci. 99, 213–224.

Van den Berg, R., Huizenga, J., 2001. Fluids in granulites of the Southern MarginalZone of the Limpopo belt, South Africa. Contrib. Mineral. Petrol. 141 (5), 529–545.

Van Reenen, D.D., 1983. Cordierite + garnet + hypersthene + biotite-bearing assem-blages as a function of changing metamorphic conditions of the southernmarginal zone of the Limpopo metamorphic complex, South Africa. Geol. Soc. S.Africa Spec. Publ. 8, 143–167.

Van Reenen, D.D., 1986. Hydration of cordierite and hypersthenes and a descriptionof the retrograde orthoamphibole isograd in the Limpopo Belt, South Africa. Am.Mineral. 71, 900–915.

Van Reenen, D.D., Hollister, L.S., 1988. Fluid inclusions in hydrated granulite faciesrocks, southern marginal zone of the Limpopo Belt, South Africa. Geochim. Cos-mochim. Acta 52 (5), 1057–1064.

Van Reenen, D.D., Roering, C., Ashwal, L.D., de Wit, M.J. (Eds.), 1992. The ArchaeanLimpopo Granulite Belt: Tectonics and Deep Crustal Processes. PrecambrianResearch, 55, no. 1–4. , p. 587 pp.

Van Reenen, D.D., Smit, C.A., Perchuk, L.L., Roering, C., Boshoff, R., 2011. Thrustexhumation of the Neoarchean ultrahigh-temperature Southern Marginal Zone,Limpopo Complex: convergence of decompression–cooling paths in the hangingwall and prograde P–T paths in the footwall. Geol. Soc. Am. Mem. 207, 189–212.

White, R.W., Powell, R., Holland, T.J.B., 2001. Calculation of partial melting equilibriain the system Na2O-CaO-K2O-FeO-MgO-Al2O3-SiO2-H2O (NCKFMASH). J. Meta-morph. Geol. 19, 139–153.

Winther, K.T., 1996. An experimentally based model for the origin of tonalitic andtrondhjemitic melts. Chem. Geol. 127 (1), 43–59.

Zane, A., Rizzo, G., 1999. The compositional space of muscovite in granitic rocks. Can.Mineral. 37, 1229–1238.










































































































































dx.doi.org/10.1016/j.gsf.2014.03.010

























































































































































































































































































































































































































































































fluid-assisted interaction of peraluminous metapelites

Documents