Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.geol.msu.ru/deps/petro/Perchuk_publ/perchukTN05.pdf Äàòà èçìåíåíèÿ: Wed Feb 15 16:13:41 2012 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 01:52:33 2012 Êîäèðîâêà:

doi: 10.1111/j.1365-3121.2005.00647.x

Fluid-mediated modification of garnet interiors under ultrahigh-pressure conditions
Alexei L. Perchuk,
1

1,2

Michael Burchard, Walter V. Maresch and Hans-Peter Schertl
2

1

1

1

Institute of Mineralogy, Geology and Geophysics, Ruhr-University, D-44780 Bochum, Germany; IGEM RAS, Moscow 119017, Russia

AB STRA CT
We report the results of a novel experimental study on eclogitic garnets with abundant inclusions of clinozoisite, quartz and rutile subjected to temperatures (T) of 800­1100 °C and a pressure (P) of 4 GPa, representative of ultrahigh-pressure (UHP) metamorphic terranes such as the Kokchetav massif, Saxonian Erzgebirge, etc. The experiments reveal extremely rapid recrystallization and partial melting of garnet interiors controlled by fluids liberated from the breakdown of the hydrous mineral inclusions. The traditional assumption that inclusions of minerals and primary fluid inclusions should be representative of the peak or even earlier metamorphic history cannot be strictly applied in this case. We argue that inclusions in UHP garnets may mirror P­T conditions postdating growth of the host crystal or even P­T conditions never actually experienced by the rock itself. The above modification of garnet interiors produces a typical patchy microstructure that occurs in natural eclogitic garnet from the diamond-bearing UHP Kokchetav massif.

Terra Nova, 17, 545­553, 2005

Introduction
The study of mineral inclusions within strong refractory crystals like garnet, zircon or clinopyroxene (e.g. Chopin, 1984; Smith, 1984; Sobolev et al., 1994) has played a key role in identifying those metamorphic rocks that have experienced ultrahigh-pressure (UHP) conditions1 (e.g. Carswell and Compagnoni, 2003) and in delineating their pressure­temperature history. Combined data from both mineral inclusions and chemical zoning patterns in the host garnet can even lead to information on the P­T conditions of the subduction path preceding the metamorphic climax (e.g. Zhang et al., 1997). By contrast, minerals indicative of UHP conditions are rarely preserved in the rock matrix (e.g. Korsakov et al., 2004), due to later overprinting. So far, interest in such inclusions of UHP minerals in natural rocks has concentrated mainly on thermodynamic treatments of the relative P­ T stabilities of various mineral assemblages on the one hand, or structural
Correspondence: Dr Alexei Perchuk, IGEM, Russian Academy of Sciences, Staromonetny per. 35, Moscow 119017, Russia. Tel.: +7 95 230 8415; fax: +7 95 230 2179; e-mail: alp@igem.ru
1 The prefix òultraó is used for those highpressure metamorphic rocks that reached the P­T stability field of the SiO2 polymorph coesite.

polymorphs of simple systems such as SiO2 and carbon on the other (e.g. Chopin, 1984; Sobolev and Shatsky, 1990; Zhang et al., 1997), as well as elastic modelling of the relative P­V­ T properties of inclusions and host minerals (e.g. Gillet et al., 1984; van der Molen and van Roermund, 1986; Perrillat et al., 2003). Recently, polyphase inclusions in garnet from quartzofeldspathic UHP rocks of the Erzgebirge, Germany, have been interpreted as a supercritical dense fluid (Stockhert et al., 2001) or melt ¨ (Hwang et al., 2001) that has been entrapped in the host garnet. This paper presents the intriguing results of a novel experimental approach that aims to assess the behaviour of mineral inclusions in garnets at or close to the pressure and temperature conditions experienced by well-known UHP terranes (e.g. Kokchetav massif, Saxonian Erzgebirge, Norwegian and Greenland Caledonides). These experiments indicate that melt (and fluid) inclusions can be generated within an existing host garnet, i.e. formed in situ.

Modification of mineral inclusions in garnet during UHP experiments
Idiomorphic garnets from mafic eclogite of the Shubino area of the Maksyutov complex (Urals, Russia) (Lennykh and Valizer, 1999; locality III-1) were used as starting material. The mineral assemblage of the eclogite contains garnet, blue amphibole,

omphacite and epidote as major constituents, with minor amounts of phengite, quartz, rutile and chlorite. Garnets are idioblastic, up to 6 mm in size, and show concentric growth zoning characterized by a marked increase of MgO and a decrease of CaO from core to rim (Table 1). They contain numerous inclusions (Fig. 1a) of clinozoisite (Czo), quartz (Qtz), rutile (Rut), titanite (Ttn) and apatite (Ap), as well as minor chlorite (Chl), phengite (Phn) and amphibole (Am). The rock is from the HP­UHP melange zone, where peak metamorphic conditions scatter in the range of 600­ 680 °C and 1.5­3.2 GPa (Lennykh et al., 1995; Bostick et al., 2003). The lack of coesite or even radial cracks around the quartz inclusions in the garnet clearly indicates that this particular rock has not experienced UHP conditions. The garnets used for the experiments were selected on the basis of the following criteria: (1) sufficient number of hydrous mineral inclusions; (2) absence of initial zoning patterns related to the inclusions; (3) minimum number of cracks and thus maximum isolation of inclusions; (4) ease of extraction from the rock, i.e. without damage to the crystal; (5) garnet size less than that of the 3 mm inner diameter of the capsule, but large enough to allow recovery and handling after the experiment. The garnets were embedded in a dry, soft and inert matrix of SiO2 gel within Pd capsules. The experiments were carried out with
545

ñ 2005 Blackwell Publishing Ltd


.............................................................................................................................................................
Table 1 Representative microprobe analyses of host garnet, epidote inclusion and melt in the experiments.
Mineral Clino-zoisite Garnet initial Garnet initial Garnet new Garnet new Garnet new Garnet new Melt Melt Melt Melt Location/run condition Inclusion Core Rim 800 °C, 4 GPa, 96 h 900 °C, 4 GPa, 139 h 1000 °C, 4 GPa, 96 h 1100 °C, 4 GPa, 96 h 800 °C, 4 GPa, 96 h 900 °C, 4 GPa, 139 h 1000 °C, 4 GPa, 96 h 1100 °C, 4 GPa, 96 h SiO2 37.77 37.16 38.03 38.00 37.98 37.43 37.36 63.53 59.63 65.77 44.26 TiO
2

Fluid-mediated modification of garnet interiors · A. L. Perchuk et al.

Terra Nova, Vol 17, No. 6, 545­553

Al2O3 24.38 20.80 21.52 20.27 21.17 20.84 20.71 13.01 13.66 11.74 9.54

FeO

tot

MnO 0.15 0.91 0.49 0.53 0.37 0.54 0.66 0.03 0.05 0.01 0.15

MgO 0.07 3.13 6.35 0.88 1.76 1.82 2.67 0.29 0.35 0.16 0.78

CaO 23.28 9.67 6.90 19.02 18.42 16.50 12.46 3.38 6.89 6.66 12.68

Na2O 0.03 0.08 0.04 0.01 0.06 0.13 0.05 0.33 0.18 0.58 0.62

K2O 0.01 0.01 0.01 0.02 0.02 0.02 0.01 1.44 0.96 0.40 0.22

Total 98.08 99.56 100.27 100.73 99.92 99.76 100.58 85.29 86.71 89.25 90.02

0.14 0.18 0.01 0.29 0.45 0.49 0.94 0.63 1.42 1.55 7.17

12.26 27.55 26.92 21.71 19.68 21.90 25.72 2.64 3.58 2.31 14.60

Fig. 1 Pressure­temperature diagram of experimental conditions and geotherms of subducted slabs as suggested by Peacock (1990). Circles: P­T conditions of the experimental runs. Reactions: czo0.2 ¼ grs0.4)0.6 + ky + coe + v (see text for details of calculation); zo + ky + SiO2 + v ¼ melt after Boettcher (1970); other melting reactions with zoisite after Poli and Schmidt (2004). Dotted line: dehydration melting of inclusions in experiment. Phase transitions: graphite-diamond (Bundy, 1980); quartz-coesite (Bohlen and Boettcher, 1982). P­T conditions of eclogite facies metamorphic rocks from: G ­ Greenland (Gilotti and Ravna, 2002); K1 ­ Kokchetav massif, Kazakhstan (Shatsky et al., 1995); K2 ­ Kokchetav massif, Kazakhstan (Okamoto et al., 2002); E ­ Saxonian Erzgebirge, Germany (Massonne, 1999); N ­ Western Gneiss Region, Norway (Terry et al., 2000). Abbreviations: an, anorthite; coe, coesite; czo, clinozoisite; grs, grossular; ky, kyanite; v, H2O; zo, zoisite.
546

òNaCl-graphiteó pressure cells in endloaded 1/2¢¢ piston-cylinder apparatus, as described in detail by Fockenberg (1998), from 800 to 1100 °C at 4 GPa (Fig. 1). More detailed information on starting materials, as well as technical and analytical procedures are presented in the Supplementary material. Except for some interior regions in garnet in the 800 °C run (Fig. 2b), and for isolated inclusions of pure quartz, domains with partial melting were observed in the host garnets in the entire experimental range of 800­ 1100 °C at 4 GPa (Fig. 2c­e). The melt pockets can be both regular and irregular in shape (cf. Fig. 2d,f), quite often contain new euhedral kyanite, and are always associated with patches of Ca-rich garnet (Fig. 2, Table 1). Newly formed Ca-rich garnet is also developed around inclusions of clinozoisite in those areas where melting was not observed at 800 °C (Fig. 2b). It is significant that no such patches developed around isolated inclusions of SiO2. Radial cracks associated with the melt pockets (e.g. Fig. 2c,d) indicate that melting is accompanied by volume expansion. The composition and amount of melt vary from one melt pocket to the other in the garnets, depending on temperature, the types of mineral inclusions involved and their local modal proportions. As a general trend, however, the degree of melting and the compositional similarity in the melts produced increases with temperature, that is, the composition of the melt becomes more basic and less aluminous (Table 1, Figs 3a and 4). Our observations suggest that inclusions of clinozoisite served as a major
ñ 2005 Blackwell Publishing Ltd


Terra Nova, Vol 17, No. 6, 545­553

.............................................................................................................................................................
(a)

A. L. Perchuk et al. · Fluid-mediated modification of garnet interiors

(b)

Grtini

Qtz

Czo
Czo
Rut

Grtini Grtnew

(c)

(d)

mel

t

melt

Grtnew Grtini

Rut

Grtnew Czo

Grtini

(e)

(f)

Gr tnew

Gr tnew

melt
Grtini

melt

Fig. 2 Modification of garnet fabric during the experiment (backscattered electron images): (a) mineral inclusions in eclogitic garnet (Grtini) from Maksyutov complex, Urals, Russia before the experiment. Note the lack of zoning patterns associated with the inclusions; (b) rim of Ca-rich garnet (Grtnew) developed around clinozoisite inclusion (800 °C/4 GPa/96 h); (c) extensional fracture filled by new garnet (arrow) with small dark (fluid inclusions?) and light (Fe-rich phase) spots (800 °C/4 GPa/96 h). Vertical cracks in these and following images developed during decompression of the sample at low temperature ( <600 °C) conditions; (d) melt pockets associated with new garnet growth (dark grey) (900 °C/4 Gpa/139 h). Note the radial cracks around the melt pocket (arrowed); (e) melt pockets surrounded by the newly formed patchy garnet (1100 °C/4 GPa/96 h); (f) enlargement of Fig. 2(e) (rectangle) showing heterogeneity of the new garnet related to the melt pocket.

ñ 2005 Blackwell Publishing Ltd

547


.............................................................................................................................................................
(a)

Fluid-mediated modification of garnet interiors · A. L. Perchuk et al.

Terra Nova, Vol 17, No. 6, 545­553

Al2O3
800 oC 900 oC 1000 C 1100 oC
o

CaO
(b)

FeO+MgO Grs

800 C 900 C 1000 C 1100 C
o o o

o

New garnet

Core Rim

Initial garnet Alm Prp

Fig. 3 Compositions of melt (a) and garnet (b) at different experimental P­T conditions (for analyses see Table 1).

source of the H2O fluid within the garnet host, for which the following dehydration reaction appears to be the best candidate: 6Ca2 Al3 Si3 O12 (OH) ) 4Ca3 Al2 Si3 O
clinozoisite kyanite coesite fluid grossular 12

× 5Al2 SiO5 × SiO2 × 3H2 O :

Ï1÷

To test this, the corresponding isopleths of grossular content in garnet in equilibrium with a clinozoisite inclusion (Fig. 1) have been calculated using the internally consistent thermodynamic data set of Holland and
548

Powell (1998). The activity of grossular is based on the subregular quaternary solid solution model of Ganguly et al. (1998) for a fixed XFe/X 2 ¼ 4 Mg ratio and XMn ¼ 0.01, which corresponds to the newly formed garnet in the run at 800 °C, 4 GPa (see Table 1). The activity of clinozoisite for the composition of the inclusion presented in Table 1 was calculated for XFe ¼ Fe/(Fe + Ca + Mg + Mn), XMg ¼ Mg/(Fe + Ca + Mg + Mn), XMn ¼ Mn/(Fe + Ca + Mg + Mn), XCa ¼ Ca/(Fe + Ca + Mg + Mn).
2

the clinozoisite (Ca2Al2Si3O12(OH))­ epidote (Ca2FeAl2Si3O12(OH)) assemblage using the regular solution parameter of Gottschalk (2004). Our experimental data at 800 °C and 4 GPa (Table 1) are in a good agreement with the calculated grossular isopleth of 0.5 (Fig. 1). It is possible that garnet compositions developed in the high temperature runs (900­ 1100 °C) may have been less controlled by the reaction (1) because of the high degree of interior melting. It appears that the H2O fluid released from the clinozoisite (and probably other hydrous inclusions as well) triggers fluid-assisted growth of the Carich garnet and causes an abrupt decrease of the solidus temperature of the inclusion-garnet microsystem, resulting in partial melting. Phase relationships in an ACFM projection (Fig. 4) show that Ca-rich garnet ± Ky can grow during incongruent dehydration melting of primarily clinozoisite + Fe-rich garnet. However, other minor phases must also be involved, as witnessed by the presence of Ti and alkalis in the newly formed melts and the unusually high Ti content of the newly formed Ca-rich garnet patches. Thus, the development of Carich garnet is due to two different processes: the initial breakdown of clinozoisite and subsequent dehydration melting. More detailed studies are necessary to distinguish any compositional differences between the new garnet generations formed by these different mechanisms. In an analogous situation in nature, melts such as those observed here should crystallize to polyphase mineral inclusions plus an aqueous fluid upon subsequent cooling. This fluid could either be retained in the host mineral as primary or secondary fluid inclusions, or escape the host mineral to contribute to the òmatrixó metamorphic fluid. Depending on the P­T­fO2 evolution experienced by the garnet interiors, the newly formed polyphase inclusions could, but need not, consist of minerals analogous to the pre-melt assemblage.

`Patchy' garnet in UHP eclogite from Kokchetav massif
The occurrence of partial melting in diamondiferous gneisses of the UHP Kumdy-Kol area of the Kokchetav
ñ 2005 Blackwell Publishing Ltd


Terra Nova, Vol 17, No. 6, 545­553

.............................................................................................................................................................
A

A. L. Perchuk et al. · Fluid-mediated modification of garnet interiors

Ky
+Qtz

Czo
Grtini

Grtnew

C

FM

Fig. 4 ACFM (Al:Ca:Fe + Mg) diagram of phase relationships between the initial garnet (Grtini) and clinozoisite (Czo) and newly formed garnet and melt at different temperatures at 4 GPa (see also Fig. 3 and Table 1). Arrows indicate generalized compositional trends of the melt and garnet with increase of temperature. The tie-line connects the initial mineral compositions.

massif in Kazakhstan has been suggested on the basis of petrological and geochemical data (Shatsky et al., 1999; Hermann et al., 2001; Korsakov et al., 2004). We note that òpatchyó microstructures in garnet analogous to those produced in our experiments by interior dehydration melting (Fig. 2e) are also found in eclogite from the same area (Fig. 5a). Detailed study of the petrology and mineralogy of these mafic eclogites suggests peak metamorphic conditions along the òwarm slabó geotherm of Fig. 1 for P 4 GPa (Shatsky et al., 1995; Okamoto et al., 2002). Figure 4(a) shows the chemical heterogeneity of eclogitic garnet associated with a polymineralic inclusion surrounded by radial cracks. The inclusion has a negative crystal shape and consists of three compositionally discrete clinopyroxenes, clinozoisite, rutile and Al-rich titanite, which we have analysed (Table 2). We suggest that the inclusion could represent a recrystallized melt pocket (cf. Fig. 2e). If so, the original melt should have a composition close to the average composition of the clinopyroxenes (which dominate the polyphase inclusion volumetrically), but
ñ 2005 Blackwell Publishing Ltd

shifted slightly towards the clinozoisite and to a lesser extent towards rutile and/or titanite. These phases quite often appear as inclusions in (U)HP garnets. The comparatively small difference in composition between associated garnet patches and the host garnet is probably a result of the superposition of two processes: the development of new garnet by a dissolution/reprecipitation mechanism analogous to our experiments, and a subsequent diffusional modification of this heterogeneity. Although negligible on the time scale of our experiments, diffusional attenuation can be expected to have played an important part in modifying the garnet from the Kokchetav eclogite, which experienced temperatures (from T 950 °C at P > 4.3 GPa to T 600 °C at P 0.5 GPa) at which volume diffusion was very effective for durations up to 6 Myr (Hermann et al., 2001).

Discussion
Mineral inclusions protected from exterior fluid and deformation by enclosing garnet quite often remain

unreacted and persist metastably. Such inclusions are used in thermobarometry as monitors of the P­T conditions of the earlier metamorphic history (e.g. Krogh et al., 1994). Reequilibration is normally assumed to occur via processes such as Fe­Mg exchange reactions with the garnet host (e.g. Spear and Parrish, 1996), or phase transformation between polymorphs of SiO2 or carbon (e.g. Chopin, 1984). The first mechanism is particularly dominant in granulitic terranes (e.g. Perchuk et al., 1985), while polymorphic phase transformations of inclusions are mainly observed in UHP complexes (e.g. Schertl et al., 1991). In keeping with recent descriptions of trapped melt inclusions in metamorphic rocks (Hwang et al., 2001; Stockhert et al., ¨ 2001), our UHP experiments show that the above basic principles of inclusion thermobarometry cannot be strictly applied when inclusions of hydrous minerals break down. The liberated fluid triggers recrystallization of garnet interiors, obscuring the record of the earlier metamorphic history. The P­T conditions estimated from internal phase assemblages will mirror conditions postdating growth of the host crystal rather than earlier events. Indeed, these internal phase assemblages may even have formed at P­T conditions never actually experienced by the rock itself, due to internal overpressures caused by partial melting and/or recrystallization of mineral inclusion(s). Evidence for such overpressure is preserved in our run products as radial cracks around inclusions and/or melt pockets. It is important to note that the modal amounts of garnet in a subducted rock of basaltic composition will almost double at pressures between 3 and 4 GPa (Schmidt and Poli, 1998), implying that overgrowths of new garnet could significantly increase the ability of garnet to act as a confining vessel and to preserve fluids and/or local partial melts to greater depths. The extent of formation of the new Ca-rich garnet during our short-term experimental runs is incompatible with volume diffusion of divalent cations in garnet (e.g. Ganguly et al., 1998; Vielzeuf et al., 2005), so that rapid melt- and/or fluid-mediated growth and recrystallization in an essentially isochoric situation, must
549

melt


.............................................................................................................................................................
99.95 100.25 99.66 100.34 99.28 96.60 98.62 Garnet (light grey) Garnet (dark grey) Clinopyroxene 1 (inclusion) Clinopyroxene 2 (inclusion) Clinopyroxene 3 (inclusion) Clinozoisite (inclusion) Sphene (inclusion) 38.55 39.04 51.26 52.64 50.68 36.95 33.51 0.08 0.05 0.10 0.06 0.50 0.00 25.31 20.82 21.63 0.82 1.41 8.60 22.27 7.85 21.64 20.99 13.45 9.33 7.49 13.74 2.46 0.48 0.45 0.70 0.12 0.04 0.16 0.10 6.34 6.88 9.33 11.95 9.78 0.06 1.34 12.06 11.16 23.29 24.17 18.79 23.41 28.23 0.02 0.04 0.72 0.65 3.21 0.02 0.09 0.00 0.00 0.00 0.02 0.02 0.00 0.00 Total

Fluid-mediated modification of garnet interiors · A. L. Perchuk et al.

Terra Nova, Vol 17, No. 6, 545­553

Table 2 Representative microprobe analyses of host garnet and polymineralic inclusion of an eclogite from Kumdy-Kol, Kokchetav massif (see Fig. 5a,b)

Fig. 5 Backscattered electron images of patchy garnet with a polymineralic inclusion of three clinopyroxenes (Cpx1)3) + Al-rich titanite (Tit) + clinozoisite (Czo) + Rutile (Rut) surrounded by radial cracks. Mafic eclogite from the diamondiferous Kumdy Kol area, Kokchetav Massif, Kazakhstan. Note that the garnet fabric in (a) is very similar to the one produced by partial melting at 1100 °C and 4 GPa (Fig. 2f).

be called upon. Indeed, our experiments show that only a few days are necessary to develop such patchy fabrics in garnet at a temperature of 800 °C or more (Fig. 2b,e). This could make garnet an effective recorder of extremely short thermal pulses such as those evidenced by pseudotachylites in
550

subduction zones as a result of shear heating on faults at seismic strain rates (e.g. Austrheim and Andersen, 2004). In addition to the marked compositional shifts in CaO (and to a lesser extent in FeO and MgO), newly formed garnet contains appreciable

SiO2

TiO2

Al2O3

FeO

tot

MnO

MgO

CaO

Na2O

K2O

ñ 2005 Blackwell Publishing Ltd


Terra Nova, Vol 17, No. 6, 545­553

.............................................................................................................................................................
degrees with the host rock itself (Philippot et al., 1995). Metastably preserved inclusions of hydrous minerals in garnet may also have important rheological implications. Such inclusions may survive in garnets of mafic eclogite even if the host rock itself is almost dry. As suggested by Hacker et al. (2003), such rocks should contain ¸0.1 wt% of H2O at temperatures and pressures exceeding 800 °C and 3.3 GPa, respectively. Jin et al. (2001) have recently demonstrated that larger modal amounts of garnet will increase the strength of bimineralic (Grt-Omp) eclogite. On the basis of these data, one could expect a strengthening of the basaltic/gabbroic layer of subducted oceanic crust caused by intensive garnet growth in the pressure range of 3­4 GPa (Schmidt and Poli, 1998). The situation could be decidedly different, however, if the poikiloblastic interiors of the garnets are subjected to partial melting. Melt pockets in the garnet should make the strongest mineral in the eclogite assemblage weaker. This should obviously affect the strength of the whole rock and may have important consequences for the rheology of the subducted crust (e.g. Stockhert and ¨ Renner, 1998). The eclogitic garnets from the Maksyutov complex, Urals, used in this study contain mineral inclusions that are very common in HP terranes. It is evident, however, that other eclogitic garnets in HP terranes are likely to have different modal proportions of mineral inclusions and even inclusions of other types of hydrous phases such as lawsonite, amphibole, paragonite, chlorite, etc. Thus changes in the P­T conditions marking the onset of dehydration melting, and in the compositions of the melt and of the newly formed garnet are to be expected. Nevertheless, the general phenomenon should be the same. Garnet interiors should be modified considerably under extreme P­T conditions as inclusions of hydrous phase(s) break down. We conclude that the processes discussed above have important implications for mineral thermobarometry, fluid inclusion studies as well as rheologies of rocks from subduction/collision complexes, and suggest that such phenomena deserve further investigation in high-grade rocks of other tectonic environments.

A. L. Perchuk et al. · Fluid-mediated modification of garnet interiors

TiO2 (Table 1), suggesting that the garnet composition has been influenced by the coexistence of Ti-bearing melt. Such observations have important implications for conventional thermobarometry, because they indicate that garnet composition measured today in UHP rocks may indeed differ from the actual composition at peak metamorphic conditions. Patchy garnets in eclogites have previously been explained as relics incorporated in newly formed crystals (Perchuk and Philippot, 2000, their Fig. 3b) or as fluid-controlled growth of garnet coronas in metagabbros by the coalescence of small grains (Engvik et al., 2001, their Fig. 5d). Our experimental results suggest that internally controlled, fluid-mediated recrystallization and/ or partial melting can be a viable alternative for the origin of patchy fabrics. In particular, the presence of inclusions associated with the patchy garnet will be an important diagnostic criterion for internal partial melting and/or recrystallization. This is well documented in the present case by the formation of Ca-rich garnet patches from inclusions dominated by clinozoisite (Fig. 2b,c). Further systematic experimental studies will be needed to gauge the role of other factors such as variations in the nature and composition of the inclusions and their hosts, as well as the P­T conditions. The extension to the concepts of inclusion thermobarometry discussed above may also be important for fluid inclusion studies. Fluid inclusions are one of the major sources of information about fluids in subduction zones. Despite the fact that fluid inclusion isochores only rarely correlate with peak pressures and temperatures (Klemd et al., 1992), it is generally considered that such trapped fluid compositions are representative of the (U)HP metamorphic fluid (Scambelluri and Philippot, 2001). As garnet is considered to be one of the important containers for primary fluid inclusions (Touret, 2001), the possibility of additional sources for fluid inclusions deserves further consideration. Indeed, fluid released from mineral inclusions will most probably differ in composition from the matrix fluid, which originated from other sources and likely interacted to variable
ñ 2005 Blackwell Publishing Ltd

Acknowledgements
We thank L. Labrousse, J. Mosenfelder, L.L. Perchuk, P. Philippot and L.Ya. Aranovich for discussions, H.-J. Bernhard and R. Neuser for assistance during microprobe and SEM work, and T. Westphal for preparation of the samples. Constructive reviews by J. Hermann and an anonymous reviewer, as well as editorial comments by A. Kroner have led to significant improve¨ ment of the manuscript. A.L.P. was supported by an Alexander von Humboldt Foundation Research Fellowship and by RFBR grant 1645.2003.5. We also acknowledge support by the Deutsche Forschungsgemeinschaft through SFB 526.

References
Austrheim, H. and Andersen, T.B., 2004. Pseudotachylites from Corsica: fossil earthquakes from subduction complex. Terra Nova, 16, 193­197. Boettcher, A.L., 1970. The system CaO­ Al2O3­SiO2­H2O at high pressures and temperatures. J. Petrol., 11, 337­ 379. Bohlen, S.R. and Boettcher, A.L., 1982. The quartz­coesite transformation: a pressure determination and the effects of other components. J. Geophys. Res., 87, 7073­7078. Bostick, B.C., Jones, R.E., Ernst, W.G., Chen, C., Leech, M.L. and Beane, R.J., 2003. Low-temperature microdiamond aggregates in the Maksyutov Metamorphic Complex, South Ural Mountains, Russia. Am. Mineral., 88, 1709­1717. Bundy, F.P., 1980. The P­T phase and reaction diagram for elemental carbon. J. Geophys. Res., 85, 6930­6936. Carswell, D.A. and Compagnoni, R., 2003. Introduction with review of the definition, distribution and geotectonic significance of ultrahigh pressure metamorphism. EMU Notes Mineral., 5, 3­9. Chopin, C., 1984. Coesite and pure pyrope in high-grade blueschists of the Western Alps: first record and some consequences. Contrib. Mineral. Petrol., 86, 107­118. Engvik, A.K., Austrheim, H. and Erambert, M., 2001. Interaction between fluid flow, fracturing and mineral growth during eclogitization, an example from the Sunnfjord area, Western Gneiss region, Norway. Lithos, 57, 111­141. Fockenberg, T., 1998. An experimental investigation on the P­T stability of Mg­ staurolite in the system MgO­Al2O3­ SiO2­H2O. Contrib. Mineral. Petrol., 130, 187­198.

551


.............................................................................................................................................................
Ganguly, J., Cheng, W. and Chakraborty, S., 1998. Cation diffusion in aluminosilicate garnets: experimental determination in pyrope-almandine diffusion couples. Contrib. Mineral. Petrol., 131, 171­180. Gillet, P., Ingrin, J. and Chopin, C., 1984. Coesite in subducted continental crust: P­T history deduced from an elastic model. Earth Planet. Sci. Lett., 70, 426­436. Gilotti, J.A. and Ravna, E.J.K. 2002. First evidence of ultra-high pressure metamorphism in the North-East Greenland Caledonides. Geology, 30, 551­554. Gottschalk, M., 2004. Thermodynamic properties of zoisite, clinozoisite and epidote. Rev. Mineral. Geochem., 56, 83­124. Hacker, B.R., Abers, G. and Peacock, S.M., 2003. Subduction factory: theoretical mineralogy, densities, seismic wave speeds, and H2O contents. J. Geophys. Res., 108(B1), 2029 doi: 10.1029/ 2001JB001127. Hermann, J., Rubatto, D., Korsakov, A. and Shatsky, V.S., 2001. Multiple zircon growth during fast exhumation of diamondiferous, deeply subducted continental crust (Kokchetav Massif, Kazakhstan). Contrib. Mineral. Petrol., 141, 66­82. Holland, T.J.B. and Powell, R., 1998. An internally consistent thermodynamic data set for phases of petrological interest. J. Metamorph. Geol., 16, 309­343. Hwang, S.L., Shen, P.Y., Chu, H.T., Yui, T.F. and Lin, C.C., 2001. Genesis of microdiamonds from melt and associated multiphase inclusions in garnet of ultra high-pressure gneiss from Erzgebirge, Germany. Earth Planet. Sci. Lett., 188, 9­15. Jin, Z.-M., Zhang, J., Green, H.W., II and Jin, S., 2001. Eclogite rheology: implications for subducted lithosphere. Geology, 29, 667­670. Klemd, K., van den Kerhof, A.M. and Horn, E.E., 1992. High-density CO2­N2 inclusions in eclogite-facies metasediments of the Munchberg gneiss Complex, Germany. Contrib. Mineral. Petrol., 111, 409­419. Korsakov, A.V., Theunissen, K. and Smirnova, L.V., 2004. Intergranular diamonds derived from partial melting of crustal rocks at ultra-high pressure metamorphic conditions. Terra Nova, 16, 146­151. Krogh, E.G., Oh, C.W. and Liou, J.G., 1994. Polyphase and anticlockwise P­T evolution for the Franciscan eclogites and blueschists from Jenner, California, USA. J. Metamorph. Geol., 12, 121­ 134. Lennykh, V.L. and Valizer, P.M., 1999. High pressure metamorphic rocks of the Maksyutov complex (Southern Urals). Field Symposium Guide: 4th International Eclogite Field Symposium, Miass, 64 pp. Lennykh, V.I., Valizer, P.M., Beane, R., Leech, M. and Ernst, W.G., 1995. Petrotectonic evolution of the Maksyutov Complex, Southern Urals, Russia: implications for ultrahigh-pressure metamorphism. Int. Geol. Rev., 37, 584­600. Massonne, H.-J., 1999. A new occurrence of microdiamonds in quartzofeldspathic rocks of the Saxonian Erzgebirge, Germany, and their metamorphic evolution. In: Proceedings of VIIth International Kimberlite Conference, Cape Town (P.H. Nixon, ed.), 533­539. van der Molen, I., and van Roermund, H.L.M., 1986. The pressure path of solid inclusions in minerals: the retention of coesite inclusions during uplift. Lithos, 19, 317­324. Okamoto, K., Liou, J.G. and Ogasawara, Y., 2002. Petrology of diamond-grade eclogite from Kumdi-Kol. In: Diamondbearing Kokchetav massif, Kazakhstan (C.D. Parkinson, I. Katayama, J.G. Liou and S. Maruyama, eds), pp. 235­256. Universal Academic Press, Tokyo. Peacock, S.M., 1990. Fluid processes in subduction zones. Science, 248, 329­337. Perchuk, L.L., Aranovich, L.Ya., Podlesskii, K.K., Lavreut'eva, I.V., Gerasimov, V.Yu., Fed'kin, V.V., Kitsul, V.N. and Karsakov, L.P., 1985. Precambrian granulites of the Aldan shield, eastern Siberia, USSR. J. Metamorph. Geol. 3, 265­310. Perchuk, A.L. and Philippot, P., 2000. Nascent subduction: record in Yukon eclogites. Petrology, 8, 1­16. Perrillat, J.P., Daniel, I., Lardeaux, J.M. and Cardon, H., 2003. Kinetics of the coesite ­ quartz transition: application to the exhumation of ultrahigh-pressure rocks. J. Petrol., 44, 773­788. Philippot, P., Chevallier, P., Chopin, C. and Dubessy, J., 1995. Fluid composition and evolution in coesite-bearing rocks (Dora-Maira massif, Western Alps): implications for element recycling during subduction. Contrib. Mineral. Petrol., 121, 29­44. Poli, S. and Schmidt, M.W., 2004. Experimental subsolidus studies on spidote minerals. Rev. Mineral. Geochem., 56, 171­195. Scambelluri, M. and Philippot, P., 2001. Deep fluids in subduction zones. Lithos, 55, 213­227. Schertl, H.P., Schreyer, W. and Chopin, C., 1991. The pyrope-coesite rocks and their country rocks at Parigi, Dora Maira massif, Western Alps: detailed petrography, mineralogy and

Fluid-mediated modification of garnet interiors · A. L. Perchuk et al.

Terra Nova, Vol 17, No. 6, 545­553

PT-path. Contrib. Mineral. Petrol., 108, 1­21. Schmidt, M.W. and Poli, S., 1998. Experimentally based water budgets for dehydrating slabs and consequences for arc magma generation. Earth Planet. Sci. Lett., 163, 361­379. Shatsky, V.S., Jagoutz, E., Sobolev, N.V., Kozmenko, O.A., Parkhomenko, V.S. and Troesch, M., 1999. Geochemistry and age of ultrahigh pressure metamorphic rocks from the kokchetav massiv (northern Kazakhstan). Contrib. Minerol. Petrol., 137, 185­205. Shatsky, V.S., Sobolev, N.V. and Vavilov, M.A., 1995. Diamond-bearing metamorphic rocks of the Kokchetav massif (Northern Kazakhstan). In: Ultrahigh Pressure Metamorphism (R.G. Coleman and X. Wang, eds), pp. 427­455. Cambridge University Press, Cambridge. Smith, D.C., 1984. Coesite in clinopyroxene in the Caledonides and implications for geodynamics. Nature, 310, 641­644. Sobolev, N.V. and Shatsky, V.S., 1990. Diamond inclusions in garnets from metamorphic rocks: a new environment for diamond formation. Nature, 343, 742­746. Sobolev, N.V., Shatsky, V.S., Vavilov, M.A. and Goryainov, S.V., 1994. Zircon from ultrahigh pressure metamorphic rocks of folded regions as a unique container of inclusions of diamond, coesite, and coexisting minerals. Doklady Akad. Nauk, 334, 488­492. Spear, F.S. and Parrish, R., 1996. Petrology and petrologic cooling rates of the Valhalla Complex, British Columbia, Canada. J. Petrol., 37, 733­765. Stockhert, B. and Renner, J., 1998. ¨ Rheology of crustal rocks at ultrahigh pressure. In: When Continents Collide: Geodynamics and Geochemistry of Ultra-high Pressure Rocks (B.R. Hacker and J.G. Liou, eds), pp. 57­96. Kluwer Academic Publishers, Dordrecht. Stockhert, B., Duyster, J., Trepmann, C. ¨ and Massonne, H.-J., 2001. Microdiamond daughtercrystals precipitated from supercritical COH silicate fluids included in garnet, Erzgebirge, Germany. Geology, 29, 391­394. Terry, M.P., Robinson, P. and Ravna, E.J.K., 2000. Kyanite eclogite thermobarometry and evidence for thrusting of UHP over HP metamorphic rocks, NordÜyane, Western Gneiss Region, Norway. Am. Mineral., 85, 1637­1650. Touret, J.L.R., 2001. Fluids in metamorphic rocks. Lithos, 55, 1­4. Vielzeuf, D., Lupulescu, A., Perchuk, A.L., Laporte, D., Baronnet, A. and Addad,

552

ñ 2005 Blackwell Publishing Ltd


Terra Nova, Vol 17, No. 6, 545­553

.............................................................................................................................................................
A., 2005. Uphill diffusion and zero flux planes in garnets: an experimental and ATEM study. TMS Lett., in press. Zhang, R.Y., Liou, J.H., Ernst, W.G., Coleman, R.G.F., Sobolev, N.V. and Shatsky, V.S., 1997. Metamorphic evolution of diamond-bearing and associated rocks from the Kokchetav Massif, northern Kazakhstan. J. Metamorph. Geol., 15, 479­496. Received 3 November 2004; revised version accepted 5 July 2005

A. L. Perchuk et al. · Fluid-mediated modification of garnet interiors

Supplementary material
The authors have supplied an appendix, Appendix A1, containing the experimental procedures, as supplementary material which can be found on http://www.blackwell-synergy.com.

ñ 2005 Blackwell Publishing Ltd

553