| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology | Volume 43 | Number 10 | Pages 1947-1974 | 2002
© Oxford University Press 2002
Equilibrium and Disequilibrium Trace Element Partitioning in Hydrous Eclogites (Trescolmen, Central Alps)
1MINERALOGISCHES INSTITUT, UNIVERSITÄT HEIDELBERG, IM NEUENHEIMER FELD 236, 69120 HEIDELBERG, GERMANY
2INSTITUT FÜR GEOLOGISCHE WISSENSCHAFTEN, UNIVERSITÄT GREIFSWALD, JAHNSTR. 17A, 17487 GREIFSWALD, GERMANY
3DEPARTMENT OF EARTH SCIENCES, MEMORIAL UNIVERSITY OF NEWFOUNDLAND, ST. JOHN’S, NEWFOUNDLAND, A1B 3X5, CANADA
Received January 5, 2000; Revised typescript accepted March 28, 2002
 |
ABSTRACT |
|---|
TOP
ABSTRACT
INTRODUCTIONDespite the widespread presence of hydrous phases in subduction- related systems, experimental DMin/Fluid trace element values for many hydrous phases are lacking. To fill this gap, we present a set of DMin/Clinopyroxene values (where Min indicates amphibole, zoisite, phengite, paragonite or apatite) derived from equilibrium parageneses of eclogites from Trescolmen (Central Alps, Switzerland). These data can be combined with experimental data for DClinopyroxene/Fluid, to estimate DMin/Fluid values for the hydrous phases, thus circumventing experimental problems with the direct determination of such values. We analysed Li, Be, B, Sr, Y, Zr, Nb, Ba, Ce, Nd, Sm, Pb, Th and U in coexisting phases by laser ablation microprobe inductively coupled plasma mass spectrometry. Many of the values are extremely low; for example, Nb, Ba, Ce, Th and U are in the lower ppb range in clinopyroxene. Attainment of equilibrium was evaluated by textural, and major and trace element characteristics. Non-equilibrated assemblages are common in most eclogite localities, including Trescolmen, and using such samples would lead to the derivation of erroneous values for equilibrium partitioning. However, four of the 10 studied eclogites from Trescolmen having homogeneous clinopyroxene compositions and preferred orientation of high-pressure phases yielded consistent DMin/Clinopyroxene values in all four samples (where Min indicates amphibole, phengite, paragonite, apatite), and hence were studied in detail. The low abundances in some phases result from strong preferential incorporation of trace elements into other minor phases. From the investigated hydrous phases (amphibole, zoisite, clinozoisite, phengite, paragonite, apatite and talc), zoisite was found to be the most important carrier of Sr, light rare earth elements, Pb, Th and U, whereas phengite hosts Ba and is, along with clinopyroxene and paragonite, an important phase for B. However, because of their low modal abundance in eclogite-facies rocks, phengite and paragonite do not control the B whole-rock budget. We infer that estimated DMin/Clinopyroxene values from equilibrium assemblages can be used as a good approximation for partition coefficients under the given PT conditions (
650°C and 2·0 GPa). Deformation-induced dynamic recrystallization seems to be the driving process necessary for a close approach towards trace element equilibrium. This process should work efficiently during subduction close to the slab–mantle interface, affecting sediments and low-T altered mid-ocean ridge basalt, but is of less importance inside the slab. Knowledge of D values from equilibrium assemblages determined in this study is useful not only for large-scale subduction-zone modelling, but also as an aid in the selection of equilibrium subassemblages in imperfectly equilibrated samples, so that a crystallization sequence can be derived. KEY WORDS: Central Swiss Alps; eclogites; partition coefficients; phengite; zoisite
|
INTRODUCTION |
|---|
The trace element composition of fluids emanating from subducting oceanic crust undergoing prograde dehydration reactions can be characterized experimentally by determining partition coefficients between mineral and fluid (DMin/Fluid). Data for DMin/Fluid exist for a few trace elements in relevant high P/low T phases such as clinopyroxene, garnet, amphibole, phengite, apatite and rutile (Beswick, 1973; Volfinger, 1976; Ayers & Watson, 1993; Brenan et al., 1994, 1995a, 1995b, 1998; Ayers et al., 1997; Stalder et al., 1998; Melzer et al., 1998; Najorka et al., 1999; Melzer & Wunder, 2000). The phase relations of metabasaltic assemblages are complex. Schmidt & Poli (1998) concluded that, in addition to the above-mentioned phases, chlorite, lawsonite, zoisite, paragonite, chloritoid and talc can be important phases in mafic rocks undergoing subduction. However, there are no fluid partitioning data for these hydrous minerals.
One can augment experimental mineral–fluid values by combining experimentally determined DMin1/Fluid with measured DMin1/Min2 in equilibrium blueschist- and eclogite-facies assemblages (Zack & Foley, 1997; Brenan et al., 1998). This is not a straightforward approach, as convincing examples of disequilibrium in metamorphic rocks were found in trace element zoning studies of garnet by Hickmott & Shimizu (1990). Those workers inferred that trace-element-enriched surface layers were trapped into the crystal structure in certain environments, leading to a disequilibrium pattern (see also Watson, 1996). Getty & Selverstone (1994) used trace element partition coefficients calculated from coexisting phases to argue against the attainment of equilibrium in Tauern eclogites. Furthermore, several studies in high-pressure rocks have pointed out the difficulty in obtaining meaningful isotopic ages from mineral separates as a result of disequilibrium on a hand-specimen scale (e.g. Thöni & Jagoutz, 1992; but see Zheng et al., 2002). However, an increasing number of in situ studies have demonstrated that a close approach to equilibrium is reached between adjacent minerals for a range of trace elements in regional and subduction-related environments (Messiga et al., 1995; Kretz et al., 1999; Yang et al., 1999).
This paper reports data for a suite of eclogite-facies rocks from Trescolmen (Adula nappe, Swiss Alps) where eclogites with pronounced foliation show strong evidence for a close approach to equilibrium trace element distributions. The eclogite locality at Trescolmen has been extensively studied (Heinrich, 1982, 1986; Santini, 1992; Meyre & Puschnig, 1993; Meyre et al., 1997, 1999; Meyre & Frey, 1998; Becker et al., 1999, 2000; Wiesli et al., 2001; Zack et al., 2001, 2002) and is characterized by widespread, well-crystallized hydrous high-pressure phases such as barroisitic amphibole, phengite, paragonite, zoisite, talc and apatite in addition to garnet and clinopyroxene. We present Min/clinopyroxene D values (where Min indicates amphibole, phengite, zoisite, paragonite and apatite) for Li, Be, B, Sr, Y, Zr, Nb, Ba, Ce, Nd, Sm, Pb, Th and U because D values are best studied for clinopyroxene (Brenan et al., 1995a, 1995b, 1998; Ayers et al., 1997; Stalder et al. 1998). A discussion of other subduction-relevant trace elements from the same sample suite has been presented in companion papers concerning phengite (Cs, Rb; Zack et al., 2001) and rutile systematics (Sb, W, Sn, Mo; Zack et al., 2002).
The data include petrographic observation, high-resolution backscattered imaging, thin-section mapping, whole-rock X-ray fluorescence (XRF) and solution inductively coupled plasma mass spectrometry (ICP-MS) analysis as well as quantitative electron microprobe and laser ablation microprobe-ICP-MS (LAM-ICP-MS) analysis of major and trace element compositions. This integrated approach delineates textural and chemical criteria for the recognition of equilibrium in metamorphic samples. We will also show that partitioning data can yield information about the crystallization history of assemblages that have not attained trace element equilibration.
|
GEOLOGICAL SETTING |
|---|
The Adula nappe (Central Alps, Switzerland) displays a classic example of the progressive increase in metamorphic grade of eclogite-facies rocks (from
500°C and 1·2 GPa to 800°C and >2·0 GPa; Heinrich, 1986), which occurs over a distance of 50 km from north to south. High-pressure conditions were reached during Alpine orogeny, when the European passive margin was subducted underneath the Apulian plate (Froitzheim et al., 1996). Near Lagh de Trescolmen, numerous eclogite bodies are embedded in garnet mica schists over an area of 1 km2 (Fig. 1). Metamorphism at 650°C and 2·4 GPa was followed by extensive recrystallization at 650°C and 2·0 GPa [the Trescolmen phase of Meyre & Puschnig (1993)]. During the Trescolmen phase, high-pressure hydrous phases such as amphibole and paragonite overgrew the peak pressure foliation mostly defined by clinopyroxene. This evidence for fluid influx into the Trescolmen eclogites during the early stages of the uplift path (Zack et al., 2001) suggests that fluid influx promoted thorough recrystallization of all eclogite-facies phases, although we note that not all eclogites at Trescolmen experienced this fluid influx (Becker et al., 1999). The Trescolmen eclogites are extremely fresh, and amphibolite-facies overprinting is confined to the centimetre- to decimetre-wide outermost margins of eclogite bodies (Heinrich, 1982), millimetre-wide fissures and micrometre-wide amphibole–plagioclase intergrowths at some grain boundaries inside the bodies.
|
, sample localities.The samples were collected from the Lagh de Trescolmen area (Table 1). Samples Ad25, Z6-59-1 and Z8-59-5 were obtained from outcrops, and the others were obtained from angular blocks, the largest of which are 3 m wide (Z6-52-1, Z8-77-5). Samples Z6-59-1 and Z8-59-5 are from the same outcrop, in which talc inclusions in amphibole are visible in some parts (Z8-59-5) of a 2 m boudin (Zack et al., 2001). A sample from a talc-free area (Z6-59-1) was analysed by LAM. CHM30a and CHM30b represent phengite-rich and zoisite-rich domains, respectively, of sample CHM30. These samples are separated by <40 cm. Ad25 and CHM30a are subsamples of Ad25-9-3 (Heinrich, 1986) and CHM30 (Meyre et al., 1997), respectively.
|
|
ANALYTICAL METHODS |
|---|
Major and minor element compositions of minerals were measured with a Cameca SX51 electron microprobe in Heidelberg and a JEOL JXA 8900 RL electron microprobe in Göttingen, both equipped with five wavelength-dispersive spectrometry (WDS) systems and synthetic mineral standards. Operating conditions were 15 kV acceleration voltage, 12 nA beam current on the Faraday cup, 1–5 µm beam diameter and 15–30 s counting time on the peak of all measured elements. A PAP correction for the Heidelberg microprobe data and a
z correction for the Göttingen microprobe data were applied. Clinopyroxene and garnet were analysed under operating conditions of 20 kV, 80 nA, 10 µm beam diameter and counting times on the peak of 15 s for Na, Ca, K and Fe (P in apatite); 30 s for Al, Si, Mg and Mn (Cl and F in apatite); and 60 s for Cr and Ti. Modal abundances of primary phases in Table 1 were obtained by element mapping with the Göttingen electron microprobe as outlined by Zack et al. (2002). Trace element analyses were conducted with the laser ablation system at Memorial University of Newfoundland (MUN). The regular setting of laser source, mass spectrometer and standardizing procedures have been described in detail by Horn et al. (1997) and references therein. Although the precision of analyses of natural minerals in thin sections is decreased with respect to analyses of glass standards as a result of variable ablation pit sizes and mineral matrices, a comparison of electron microprobe and laser ablation microprobe analyses of Ba in phengite (Zack et al., 2001) and selected trace elements in rutile (Zack et al., 2002) indicates an accuracy of laser ablation measurements in the range of 15–20%. The laser was focused 100–500 µm above the mineral surface to reduce fractionation effects. This was found to be most important for Pb measurements. Pb fractionation (compared with the Ca signal) during a 1 min ablation reduces from 50% for a beam focused at the sample surface to 20% for a beam focused 500 µm above the surface (2 mJ pulse energy in both cases). Fractionation effects of up to 20% are not a serious problem because fractionation during ablation is similar for glass standard and minerals.
We used a rigorous cleaning procedure to remove surface contamination. Elements such as B, Zr, Ce, Ba and Pb appear to be anomalously enriched on grease-contaminated surfaces (e.g. apparent Ba concentration in clinopyroxene can vary from 5 ppm down to 0·05 ppm depending on cleaning procedure). Thin sections were therefore polished with 0·3 µm aluminium oxide powder, washed in an ultrasonic bath for 2 min and then cleaned with pure alcohol before each LAM session. Thin sections were inspected with a high-magnification reflected-light microscope. If grease droplets 
1 µm diameter were detected, cleaning with alcohol was repeated until grease droplets completely disappeared.
Solid and fluid inclusions as well as alteration along microcracks can change trace element compositions by several orders of magnitude. Inclusion-free areas were selected with the petrographic microscope, which in the MUN system is an integral part of the LAM instrumental system (Jackson et al., 1992). Submicroscopic inclusions were detected by inspecting each block of data acquisition in time-resolved mode. ‘Spiky’ signals were excluded from data reduction under the assumption that they are due to inclusions. Under perfect conditions, regular ablation behaviour can be shown for element concentrations greater than 100 ppb. Below this value, signals arrive too irregularly to demonstrate homogeneous distribution in the ablated mineral. Values <100 ppb are therefore maxima.
To reach the lower ppb range, element menus (usually all trace elements of interest were analysed simultaneously) were split and the largest practically possible laser pulse energies, depending on the size of optically inclusion- and alteration-free areas, were applied (for clinopyroxene between 1·5 and 2·7 mJ at 10 Hz with a beam focused 300–500 µm above the sample surface). For these reduced element menus (as low as six elements), the ICP-MS system was tuned to have the highest sensitivity in the chosen mass range. Helium was used as a carrier gas for the ablated material when light elements were analysed. This increases sensitivity for Li, Be and B by a factor of >20 when compared with Ar as a carrier gas. With He as a carrier gas, calculated detection limits (2
above background signals) on ablation pits of 60 µm width are 0·04, 0·3 and 0·7 ppm for Li, Be and B, respectively. Optimized sensitivity for the heavier elements in combination with a reduced element menu and high laser pulse energies leads to detection limits for U in clinopyroxene of between 0·6 and 10 ppb in different samples, depending on crystal size and inclusion density. For most trace elements, Ar was used as the carrier gas, as it was found that sensitivity is similar for both carrier gases so that the cheaper option can be used. Garnet contents are below 100 ppb for most trace elements of interest, so that the beam diameter was always chosen to be at least 60 µm. Because of the small-scale features (
20 µm; see below), measured garnet concentrations are always mixed analyses of different garnet generations.
For major and trace element whole-rock determination, fist-size samples were crushed to centimetre-size splits, which were washed in an ultrasonic bath and hand-picked to avoid any trace of visible alteration. These splits were used for standard analyses with XRF spectrometry (Universität Göttingen and Memorial University) as well as with single-collector ICP-MS (Memorial University).
|
WHOLE-ROCK GEOCHEMISTRY AND GEOTECTONIC POSITION OF THE ADULA NAPPE |
|---|
The eclogites of Trescolmen range widely in composition. Their protoliths can be subdivided into accumulative mafic and basaltic types on the basis of their Al2O3 and TiO2 contents [Fig. 2; compare Miller et al. (1988) and Miller & Thöni (1995)]. Samples CHM30a, CHMb, Z6-59-1 and Z6-55-3 appear to be cumulates whereas Z5-52-1 is a FeTi basalt and the remaining samples are basalts. Coarse-grained, unfoliated eclogite samples all belong to the cumulate group, whereas three out of the four foliated, medium- to fine-grained eclogites are basaltic (Tables 1 and 2). Using eclogite grain size as an indicator of protolith coarseness, in the following discussion we classify samples with a cumulate chemistry and/or coarse grain size as metagabbroic. In addition to low TiO2 and variable Al2O3 contents, metagabbroic samples also display the highest MgO (9·7–12·1%) and Cr contents (810–1020 ppm) in the suite. As no whole-rock data exist for Ad25, we classify this sample as cumulates on the basis of high mg-number in garnet (Table 3) and clinopyroxene (Table 4) and on the low modal abundance of rutile (Table 1). This sample is special as it is the only fine-grained cumulate and exhibits a pronounced foliation, which we tentatively interpret as a deformed metagabbroic sample.
|
, basalts; 
, FeTi basalt. Dashed line distinguishes kyanite-free eclogites (Z6-50-2, Z6-52-1, Z6-55-3, Z6-55-4) from kyanite-bearing eclogites (
|
|
|
Trescolmen metabasaltic eclogites are identical in their major element composition to mid-ocean ridge basalts (MORB). This is illustrated by comparing fractionation sequences of MORB with the Trescolmen eclogites. Using TiO2 as a fractionation index, the Trescolmen eclogites span the whole range of MORB compositions and depletion or enrichment trends for Cr, MgO, SiO2 and FeO (Fig. 3). Only CaO and especially Na2O contents seem to be disturbed (Fig. 3c and f), which can be best explained by spilitization. Generally immobile elements such as Ti, Zr, Y and Sr are used to classify basaltic rocks and their metamorphic equivalents (Pearce & Cann, 1973). Pearce & Cann stated that only basalts with 12% < MgO + CaO < 20% should be used, and samples Z8-77-5 and Z6-55-4 have slightly higher values. However, plotting the Trescolmen eclogites on the Ti–Zr–Y and Ti–Zr–Sr discrimination diagrams of Pearce & Cann (1973) illustrates that all samples fall into the field defined for MORB-type basalts.
|
Further evidence for spilitization is provided by oxygen isotope measurements of Trescolmen eclogites. Oxygen isotope data for reconstructed whole rocks (measurements were performed on fresh clinopyroxene and garnet separates) give values between 5·5 and 7·5
(Kohn & Valley, 1998; Wiesli et al., 2001). On the basis of these heavy oxygen isotopes, Wiesli et al. (2001) suggested a low-T hydrothermally altered oceanic crust as the protolith of their investigated Trescolmen eclogites, although they considered the possibility that minor oxygen isotope modifications occurred by interaction with fluids from surrounding garnet mica schist. These observations are consistent with our earlier findings based on Cs–Rb–Ba systematics, where both a low-T hydrothermal and an eclogite-facies fluid infiltration signature can be traced in amphibole and phengite in the eclogites (Zack et al., 2001). Here, eclogite–metapelite interaction was very minor [discussed by Zack et al. (2001)], in contrast to massive phengite formation in rinds of high-grade blocks embedded in mélange terranes (Sorensen et al., 1997). In any case, hydrothermal alteration of the eclogite protoliths seems to be a multi-stage process, as high sodium contents are decoupled from high potassium contents, which is not uncommon in altered oceanic crust (e.g. Stroncik & Schmincke, 2001). Furthermore, we note a positive correlation of Li, light rare earth elements (LREE), U and Th contents with sodium in the Trescolmen eclogites (Table 2), a feature that has so far no direct analogue in modern ocean crust. This compares well with the finding of Becker et al. (1999, 2000), who correlated enrichment in U and in heavy Sr and Pb isotopes in Trescolmen eclogites with metasomatism before high-pressure metamorphism.
The operation of several processes in the formation of the eclogite protoliths is demonstrated by the variety of mineral parageneses at Trescolmen. Three out of four paragonite-bearing eclogites are the most Na2O-enriched samples (3·7 wt % Na2O, Fig. 3f) and zoisite occurs in eclogites with 12 wt % CaO (Fig. 3c). The modal abundance of phengite is directly linked to whole-rock potassium enrichment, as it is the dominant K-bearing phase (Zack et al., 2001). Major element variation in the protoliths has a direct control on the occurrences of talc, chromite and kyanite: talc occurs only in two of the three MgO-rich eclogites (Fig. 3b), the only chromite relics were found in the most Cr-rich sample (Fig. 3a) and kyanite is restricted to samples with 15 wt % Al2O3 (Fig. 2).
It has to be emphasized that the major and immobile trace element composition of the eclogites does not necessarily imply an origin for their protoliths as MORB at spreading ridges. In continental rift tectonic settings, MORB-type magmas may be injected into and/or extruded onto thinned continental crust. The origin of the metabasalts of the Adula nappe in a transitional continental–oceanic rifting environment avoids the necessity to emplace them later into the metapelite host rocks. If the pelites and basalts were juxtaposed before collision, this explains why the Trescolmen metapelites record the same high-pressure conditions as the enclosed eclogites (Meyre et al., 1999). This rift basin setting also explains the almost complete lack of ultramafic rocks in the Northern and Middle Adula nappe (Heinrich, 1986), and the local presence of FeTi metabasalts (e.g. sample Z6-52-1). FeTi basalt is a highly fractionated constituent of the oceanic crust. Single, non-replenished magma batches injected into cool surroundings, which can lead to such fractionation, are the exception in normal oceanic crust (occurring at propagating spreading segments; Clague et al., 1981), but are common, besides also in flood basalt provinces, in continental rift systems. The Adula nappe is interpreted to be the former outermost passive margin of the European continent before the Alpine orogeny (Froitzheim et al., 1996) and magmatism accompanied the rifting stage (Kurz et al., 1998b). Despite the likely tectonic setting of the Trescolmen eclogites in a rifting environment, the samples chemically resemble MORB. In this respect, Trescolmen eclogites may be used as chemical analogues for subducting oceanic crust.
|
MINERAL CHEMISTRY |
|---|
Minerals were studied by high-resolution backscattered electron imaging and analysed with the electron microprobe (selected electron microprobe data are given in Tables 3–5) to define areas in each phase that were stable during the Trescolmen stage. Here, during deformation accompanied by fluid influx (see above), clinopyroxene partly recrystallized together with hydrous phases at
650°C and 2 GPa, so that an approach to equilibrium is most likely under these conditions. The selected areas were then analysed by LAM for their trace element contents. The minerals are described below, discussing especially features that may be indicative of equilibrium or disequilibrium and fluid infiltration. Additional data for amphibole and phengite have been presented by Zack et al. (2001).
|
Clinopyroxene
Clinopyroxene ranges from 27 to 57 mol % jadeite with mg-number from 76 to 95 (Fig. 4a and b). Clinopyroxene analyses from samples Z6-50-2, Z6-50-13, Z6-55-4, CHM30a and CHM30b define distinct clusters in a plot of mg-number vs Jd (Fig. 4a). Large heterogeneities in clinopyroxene composition are confined to samples Z6-50-14, Z6-55-3, Z6-59-1 and Ad25 (Fig. 4b) and reflect textural features and/or whole-rock geochemistry. For instance, samples Ad25, Z6-55-3 and Z6-59-1 are metagabbroic and, except Ad25, lack a penetrative foliation. Sample Z6-55-3 provides a spectacular example of disequilibrium at a thin-section scale: in addition to a pronounced variation in Jd content and mg-number, this sample contains millimetre-wide mineral aggregates of a striking green colour in thin section, which consist of phases rich in Cr (e.g. clinopyroxene and phengite with up to 2 wt % Cr2O3, compared with 0·2 wt % Cr2O3 outside the Cr-rich clusters) and minute inclusions of chromite in clinopyroxene. Chromite inclusions in such Cr-rich clinopyroxene grains have been interpreted to be relict magmatic grains (see Messiga et al., 1999). Cr mobility during eclogite-facies metamorphism therefore appears to have been limited in this sample to the size of the Cr-rich clusters (1 mm).
|
Sample Z6-52-1 presents a case where homogeneous and heterogeneous clinopyroxene can be found together. Fine-grained clinopyroxene shows a core to rim decrease from Jd54 to Jd47, with a corresponding decrease of mg-number from 83 to 79 (Figs 4a and 5a). Large, idiomorphic and inclusion-poor clinopyroxene grains (commonly aligned in the foliation) in this sample are unzoned, with Jd contents and mg-number identical to the rim compositions of smaller grains. As discussed below, such large, homogeneous clinopyroxene grains are considered to have been formed during the Trescolmen stage, which closely approached equilibrium conditions.
|
|
Garnet
In general, all samples show garnet zonation typical for growth under prograde metamorphic conditions, characterized by an increase in mg-number and decrease in MnO content from grain cores to rims (Tracy, 1982). CaO content also decreases from core to rim (Fig. 6). Garnet rims are clear and contain sparse inclusions of quartz and omphacite (observed only in Z8-59-5). Backscattered electron images show patchy zoning of garnet cores, in which old core areas (GRT1 in Fig. 5b) are replaced along irregular annealed cracks and fractures by recrystallized garnets (GRT2a). These cracks contain garnet with rim-like compositions (GRT2b). Erambert & Austrheim (1993) described similar features in garnet as annealed cracks, formed under eclogite-facies conditions, and interpreted them as fluid pathways. As garnet rims are in equilibrium with omphacite and cracks giving evidence for fluid presence, we correlate GRT2a and GRT2b with the high-pressure fluid infiltration of the Trescolmen stage. A second type of annealed straight and anastomosing bright fractures (GRT3) is visible in the most altered samples (Z6-55-3, Z8-77-5) and postdates eclogite-facies garnet rims. The late annealed fractures in garnet grains are displaced from typical garnet zoning (Ca–Mg exchange, Fig. 6) towards greater Fe and Mn contents at a given Ca content. This trend is also seen at the outermost rims of some garnet grains from Trescolmen [e.g. sample Z6-50-12 of Meyre et al. (1999)]. Late annealed fractures and the outermost rim zoning probably reflect amphibolite-facies overprinting, and resemble annealed garnet fractures described by Kurz et al. (1998a).
|
, compositions of late annealed cracks in garnets from sample Z6-55-3 (GRT3 in Fig.
Zoisite, clinozoisite and allanite
Zoisite was found in samples Ad25, CHM30b and Z8-77-5. It coexists with clinozoisite in sample Z8-77-5, in which distinct crystals of zoisite and clinozoisite, being in mutual grain contact with each other, show limited variations in composition (XFe from 0·037 to 0·058 in zoisite and from 0·098 to 0·132 in clinozoisite), consistent with an inferred miscibility gap between zoisite and clinozoisite (Franz & Selverstone, 1992). Paragonite eclogites from Dabie Shan, which yield similar PT estimates (1·9 GPa and 700°C; Okay, 1995) display apparently coexisting zoisite and clinozoisite of nearly identical compositions (XFe is 0·049 for zoisite and 0·135 for clinozoisite; Okay, 1995). Zoisites varies in XFe from 0·021 to 0·042 in Ad25 and from 0·037 to 0·055 in CHM30b. These compositions might reflect a lower Fe3+ concentration in the whole rocks in Ad25 and CHM30b compared with Z8-77-5, as Fe3+-rich clinozoisite is not found in the former samples.
The zoisite grains in samples Ad25 and CHM30b show pronounced zoning in backscattered electron images (Fig. 5c and d), which correlates with Sr contents. For instance, bright areas of zoisite grains in CHM30b (Fig. 5c) contain up to 21 000 ppm Sr (ZOI3), whereas darker areas contain 3500 ppm Sr (ZOI2). In this sample, zoisite shows evidence for growth zoning and later resorption. Zoisite cores, rich in Sr (ZOI1) formed under eclogite-facies conditions, as indicated by numerous omphacite inclusions. Sr is enriched in zoisite cores and depleted in the surrounding matrix. Later zoisite (ZOI2) contains less Sr, similar to more common Mn growth zoning patterns of garnet. Growth continued under eclogite-facies conditions, documented by rare omphacite inclusions. Extreme enrichments of Sr of up to 21 000 ppm occur only in the outermost rim of zoisites (ZOI3). These zones are unevenly distributed around zoisites and mostly smaller than 30 µm. These features point to zoisite resorption that released Sr, which was redistributed in the surface layers of the remaining zoisite. As amphibolite overprinting is present on grain boundaries, the high-Sr rims formed as zoisite began to break down at amphibolite-facies conditions.
The zoning of zoisite grains in sample Ad25 (Fig. 5d) is very complex. Cores (ZOI1) contain 2800 ppm Sr and show small-scale heterogeneities in backscattered images. These are surrounded by narrow Sr-rich zones (up to 10 000 ppm Sr; ZOI2). At grain rims, Sr drops to levels of 1000 ppm (ZOI3). The origin of two or more zones rich in Sr might point to episodic fluid infiltration into the eclogite bodies, with Sr transported by the fluid. Alternatively, these rims might also be explained by resorption, as for CHM30b.
Allanite occurs as an accessory phase in samples Z6-50-2, Z6-50-13, Z6-52-1 and Z6-55-3, typically as irregular grains <40 µm long. An exception is sample Z6-50-13, in which several elongated grains up to 400 µm in size show a preferred orientation identical to the eclogite-facies foliation and contain omphacite inclusions. At least in this sample, allanite grew under eclogite-facies conditions.
Amphibole, phengite, paragonite, talc and apatite
Amphibole is present in every sample. According to the classification of Leake et al. (1997), amphibole from most samples is barroisite. A strong correlation between Na on the M4 site in amphibole cores and XJd of associated omphacite indicates that amphibole was stable at eclogite-facies conditions (Heinrich, 1986). Na(M4) increases with Na+K(A-site) in amphibole cores, but core to rim zoning shows a decrease of Na(M4) and increase of Na+K(A-site) towards the rim. Amphibole rim compositions are not correlated with those of associated omphacite: the amphibole rims must have grown after the Trescolmen stage, either during the uplift still under eclogite-facies conditions, or during the amphibolite-facies overprint.
Phengites contain 37–47 mol % of celadonite, and display Na/(Na + K) of 0·05–0·20, with a negative correlation between these two parameters. Paragonite and celadonite zoning of phengite rims reflects late diffusion along grain boundaries and cleavage planes (Zack et al., 2001). Laser ablation measurements have therefore been limited to the cores of fresh phengites with no observable cleavage planes.
Other hydrous phases lack pronounced chemical variations. Paragonite shows limited celadonite and muscovite substitution (up to 1·4 wt % K2O; Table 5). Talc is almost a pure end-member, and contains only 0·45 wt % Al2O3 and 2·7 wt % FeO as minor components (Table 5). Apatite contains 4·3–5·4 wt % F and low Cl concentrations (200–500 ppm Cl; Table 5).
|
MASS BALANCE AND DOMINANT TRACE ELEMENT CARRIERS |
|---|
Bulk-rock data and modal abundances of the high-pressure phases have been combined with the results from the LAM data of in situ mineral analyses to perform mass balance calculations. Differences between measured and calculated whole-rock compositions can point to the presence of important, unanalysed accessory phases, and the combination of the trace element content and modal abundance of each phase indicates the dominant carriers of each trace element in eclogites. Calculations are reported for two samples that represent phengite-rich eclogite (Z6-50-13) and zoisite-rich eclogite (CHM30b).
The trace element budget of sample Z6-50-13 has been successfully modelled for a range of trace elements (Zack et al., 2001, 2002). Phengite dominates the Ba budget of the whole rock, rutile dominates Nb and zircon dominates Zr, and so Ba, Nb and Zr are not plotted in Fig. 7a and b. For the remaining elements, the differences between calculated and measured whole-rock concentrations of Li, Sr, Pb and Y are generally <20% (slightly >20% in the case of Li). This is treated as a balanced calculation (Sorensen & Grossman, 1989). For Be and B, no whole-rock data exist and the calculated whole rock is normalized to 100%. The difference from the 100% level of the calculated whole rock in Fig. 7a is due to not plotting amphibole, which makes up 3% of the Be and 6% of the B budget. For all other elements, the contribution of amphibole is <1% in eclogite sample Z6-50-13. Calculated values are much less than whole-rock compositions for Ce, Nd, Sm, Th and U. The deficits for REE are probably explained by the presence of non-analysed allanite in Z6-50-13 (confirmed by qualitative energy-dispersive spectrometry analyses). Allanite is extremely rich in LREE (here Ce) and poorer in HREE (heavy REE; here Y is treated as an HREE). Measured and calculated whole-rock REE concentrations balance using an average composition of allanites from garnet amphibolites (38 000 ppm Ce, 17 000 ppm Nd; Sorensen & Grossman, 1989) and assuming a modal allanite abundance of only 0·04%. Allanite with 1400 ppm Th and 600 ppm U would balance the whole-rock deficit of these trace elements. Th and U are also found in zircon. Using a modal abundance of zircon calculated with the Zr whole-rock content (0·018%; Zack et al., 2002), 4300 ppm Th and 2000 ppm U in zircon would eliminate the deficits in the Th and U budget.
|
Mass balance for sample CHM30b can be achieved for all trace elements except Y and Th (Fig. 7b). The deficit in Y can be explained by a lack of analyses of garnet cores (as a result of a high inclusion density in the cores). Trescolmen garnets are probably zoned in Y with highest concentration in the core, as commonly observed in prograde growth zoning of garnets (e.g. Hickmott & Shimizu, 1990; Yang & Rivers, 2001). The deficit in Th is less satisfactorily explained. Eclogite-facies zircons do not show extreme enrichments in Th over U (Creaser et al., 1997). Given the large compositional differences in a single zoisite grain (Fig. 5c), which produce Th/U ratios between 0·4 and 5·1, the data from one grain may fail to represent the average composition of zoisite in the rock.
Li, Be and B
Samples Z6-50-13 and CHM30b both show the dominant role of clinopyroxene for the Li and Be budget. Other phases incorporate <10% of the whole-rock value. B is more evenly distributed among different phases: in sample Z6-50-13, paragonite, phengite, garnet and clinopyroxene each contribute >10% of B to the whole-rock budget. Although phengite is an important mineral for B, it is neither abundant nor does it contain the highest concentration of B: paragonite contains four times as much B as does phengite. Domanik et al. (1993) reported high B concentrations in phengite (60–100 ppm) and significant amounts in omphacite ( 5 ppm) from eclogites. Be contents of phengite are similar to those of clinopyroxene in the Trescolmen eclogites. However, the mass balance calculations for Trescolmen do not confirm the conclusion of Domanik et al. that phengite is the primary host of B and Be in samples that contain clinopyroxene as a major phase. Guidotti & Sassi (1998) showed that Li is commonly <100 ppm in micas from regional metamorphic metapelitic rocks, a conclusion that holds for Trescolmen eclogite-facies phengite.
Sr and Pb
Sr and Pb are similarly partitioned in both samples. In CHM30b, zoisite contains >80% of the whole-rock budget of both elements. Clinopyroxene is the only other significant carrier of Sr in CHM30b. In sample Z6-50-13, which lacks zoisite, Sr and Pb are accommodated in apatite, paragonite, phengite and clinopyroxene. In detail, clinopyroxene carries most of the Sr (40%; 20% for Pb) whereas phengite is most important for the Pb budget (50%; 20% for Sr). Interestingly, allanite does not account for much Sr or Pb, given the almost perfect agreement between calculated and measured whole-rock concentrations. However, the low calculated modal abundances permit 5000 ppm Sr and 200 ppm Pb in allanite, which would then account for <3% of the whole-rock composition. The importance of zoisite (or epidote-group minerals) as a carrier for Sr and Pb has been reported both as absolute concentrations (Hickmott et al., 1992; Domanik et al., 1993) and dominance in mass balance calculations (Sorensen & Grossman, 1989; Nagasaki & Enami, 1998). Although dehydration of zoisite could release large quantities of Sr and Pb into the fluid (Hickmott et al., 1992; Nagasaki & Enami, 1998), in the absence of zoisite, significant quantities of Sr and Pb can also be incorporated in clinopyroxene, which is stable beyond zoisite stability (Schmidt & Poli, 1998). A strong zoisite signature in the released fluid will therefore not inevitably develop upon zoisite breakdown, but instead depends on a more complex interplay of partition coefficients between zoisite, clinopyroxene and fluid and/or the approach towards equilibrium conditions.
REE and Y
Even though the REE concentrations and modal abundances of allanite (in Z6-50-13) and zoisite (in CHM30b) are completely different, both phases are important for the REE inventory of the rock. More than 90% of the LREE (Ce) are accommodated by allanite and zoisite, but the dominance of these two minerals in whole-rock budgets decreases steadily towards the HREE, with the relative proportions in allanite and zoisite Ce > Nd > Sm > Y. Instead of zoisite and allanite, garnet dominates the HREE (proxied by Y) budget (in Z6-50-13 >80%). Although apatite may contain >180 ppm of Ce, it contains less REE than zoisite or allanite, and is less abundant than zoisite in zoisite-bearing eclogites. In eclogites from Liguria (Tribuzio et al., 1996), allanite is the dominant phase for LREE whereas garnet dominates the HREE budget. Sorensen & Grossman (1989) described subduction-related garnet amphibolites in which the LREE budget is dominated by zoisite and the HREE budget by garnet.
Th and U
Because of the lack of data for allanite, the element mass budget for Th and U is uncertain. Zoisite probably controls the element budget for U and Th in CHM30b, despite the large variations in Th and U in the single analysed zoisite grain. Unanalysed zoisite with a slightly larger than measured Th content more probably balances the Th budget than zircon. The whole-rock Th/U ratio of 2·7 in sample CHM30b suggests zoisite is the dominant phase, as this is within the measured range of Th/U ratios in zoisite (0·4–5·1), whereas eclogite-facies zircons have ratios <0·02 (Creaser et al., 1997). The similarities of trace element incorporation in allanite and zoisite, as well as the ability of allanite to store large amounts of Th (as high as 5 wt %: Deer et al., 1992; up to 6000 ppm in high-grade rocks: Sorensen & Grossman, 1989), suggest that allanite is the likely carrier of Th and U in the zoisite-free sample Z6-50-13. In any case, the major phases garnet and clinopyroxene are insignificant for the Th and U budget in eclogites (both phases together incorporate always <5% of the measured whole-rock content of Th and U).
Mineral–mineral partitioning systematics
Evaluating the use of samples for trace element equilibration studies
The inner rim areas of zoisite and garnet, the core areas of amphibole, phengite (Zack et al., 2001), paragonite and apatite, as well as the large clinopyroxene grains make up the re-equilibrated eclogite stage (Trescolmen stage). Because of the mixed analyses obtained by LAM for garnet, we are not able to discuss trace element systematics of this phase. Equilibrium was approached to different degrees in the samples. The variability of clinopyroxene compositions within individual samples is used as a guide for evaluating equilibrium. Clinopyroxene compositions in samples Ad25, Z6-50-14, Z6-55-3 and Z6-59-1 are heterogeneous (Fig. 4b), whereas samples CHM30a, CHM30b, Z6-50-2, Z6-50-13, Z6-52-1, Z6-55-4 and Z8-77-5 have restricted clinopyroxene compositions on a scale >5 µm (Fig. 4a) and are therefore well suited for trace element analysis with LAM (>40 µm). The textural relationships between different phases yielded another equilibrium criterion. Of samples with restricted clinopyroxene compositions, Z6-50-2, Z6-50-13, Z6-52-1 and Z8-77-5 are foliated eclogites (Table 1). Here, hydrous phases define the same foliation as clinopyroxene, implying that they dynamically recrystallized together.
In summary, the samples that probably attained trace element equilibrium are Z6-50-2, Z6-50-13, Z6-52-1 and Z8-77-5 (in the following called preferred samples). Samples CHM30a, CHM30b and Z6-55-4 have homogeneous clinopyroxene compositions, but lack signs of dynamic recrystallization. Samples Ad25, Z6-50-14, Z6-55-3 and Z6-59-1 show major element inhomogeneities in clinopyroxene composition. An area of 1 cm2 was necessary to find enough large inclusion-free grains of each mineral for trace element analysis. Messiga et al. (1995) showed that equilibrium of the REE was restricted to microdomains within low-T, undeformed eclogites from Liguria. Although our chosen areas are larger than the equilibrium domains recognized by Messiga et al. (1995), the higher temperatures (650°C instead of 450°C) and dynamic recrystallization in some Trescolmen eclogites should have promoted larger equilibrium domains.
Despite the large number of LAM analyses (>400 single spot analyses) accumulated for this study, the value for a specific trace element in one phase per sample may be based on only one or two analyses. If we obtained three or more values, the standard deviation is a homogeneity test (Kretz et al., 1999). A threshold of 25% standard deviation is an upper limit for acceptable analytical scatter, and a lower limit of 35% standard deviation is a strong indication for trace element heterogeneity on a 40–100 µm scale. Values <0·1 ppm are regarded as maximum values (see the ‘Analytical methods’ section). In Tables 6–9, homogeneous values are shown in bold type, and inhomogeneous ones in italics. Other values have standard deviations between 25 and 35%, are below 0·1 ppm or are based on fewer than three analyses. The threshold value of 25% is larger than the estimated precision of 10% for regular LAM analyses of standard glasses under constant conditions and was chosen to compare data from different analytical set-ups and from different spot sizes. It should be noted that, in detail, we do not follow the procedure described by Kretz et al. (1999), in which threshold values are solely based on precision calculated from repeated analyses of standard glasses under constant conditions.
|
|
|
Trace elements buffered by single phases
Trace element equilibrium between different phases exists if Henry’s law behaviour can be demonstrated (e.g. Beattie, 1993). At equilibrium, trace element partition coefficients between two phases are independent of concentration, and element–element plots for a given mineral pair should define a straight line that passes through the origin. This will not occur if a trace element is buffered by an accessory phase as an essential structural constituent [see discussion for Zr in rutile by Zack et al. (2002)]. In the Trescolmen eclogites, Ti and Zr dominantly reside in rutile and zircon, respectively. In all phases except amphibole, Ti and Zr concentrations are constant (Tables 6–9), whereas the Ti and Zr contents in amphibole are correlated with the Al content of the tetrahedral site (Fig. 8). According to Skulski et al. (1994), both Zr and Ti can be substituted as a Tschermak’s component, Ca(Zr,Ti)Al2O6, in clinopyroxene. Therefore an increase in the Al content of the tetrahedral site in amphibole probably also facilitates the incorporation of both trace elements in the crystal structure. In contrast, constant concentration is expected for buffered trace elements where crystal chemistry does not vary with respect to relevant substitution mechanisms (such as the lack of measurable Tschermak’s component in eclogitic clinopyroxene), as long as PT conditions are constant and equilibrium is reached. Therefore, the absence of a linear correlation of Ti in coexisting mineral pairs in rutile-buffered systems does not indicate trace element disequilibrium, and clustering at a certain value denotes equilibrium (compare Getty & Selverstone, 1994).
|
Amphibole–clinopyroxene partitioning
Values for REE (Ce, Nd, Sm), Y, Sr and Pb vary widely in clinopyroxene and amphibole from different samples (Tables 6 and 7). Amphibole–clinopyroxene partition coefficients (Table 10) for some elements indicate an approach towards equilibrium in our samples, as shown by a good correlation for Sm between coexisting mineral pairs (Fig. 9a). Over a concentration range of two orders of magnitude, samples lie on a straight line passing through the origin, demonstrating similar DAMP/CPX partition coefficients for Sm of about unity. Amphibole and clinopyroxene in zoisite-bearing samples (Ad25, CHM30b, Z8-77-5) are poor in Sm. Because zoisite preferentially incorporates Sm under equilibrium conditions, samples of similar bulk Sm contents will have coexisting phases with different Sm concentrations depending on modal zoisite abundance. Sample Z8-77-5, with the highest zoisite and clinozoisite abundance (together 15 modal %, Table 1), displays less Sm in both clinopyroxene and amphibole than do counterparts in samples Ad25 and CHM30b (6–8 modal % zoisite). Samples Ad25 and CHM30b contain less Sm in clinopyroxene and amphibole than do zoisite-absent samples. Because all samples vary in their Sm whole-rock content by only a factor of two (1·8–4·3 ppm), zoisite evidently controls the Sm contents of both clinopyroxene and amphibole.
|
|
With two exceptions, DAMP/CPX partition coefficients for Y are about four (Fig. 9b). Only mineral pairs from Z6-55-3 and Z6-55-4 plot off the linear correlation. Although Y contents in both clinopyroxene and amphibole in sample Z6-55-4 are homogeneous, DAMP/CPX is a factor of three lower than the average value. A low standard deviation of repeated measurements on a 40–100 µm scale evidently is a less reliable indicator of equilibrium than textural criteria and electron microprobe data on a 5 µm scale.
DAMP/CPX partitioning for Sr shows clear evidence for disequilibrium processes in some samples. The preferred samples Z6-50-2, Z6-50-13, Z6-52-1 and Z8-77-5 plot close to a straight line (here D 1) that passes through the origin, which implies equilibrium. The other samples (except Z6-55-4) display smaller DAMP/CPX values, with zoisite-bearing samples Ad25 and CHM30b being most extreme. The partitioning behaviour of Pb closely resembles that of Sr (Fig. 10). Because Sr is a reliable element by the LAM technique, this correlation confirms the reliability of the Pb data, which can be affected by fractionation during ablation (Longerich et al., 1996).
|
Although crystal chemistry is thought to be controlling trace element partitioning, the large range in DAMP/CPX for Sr cannot be explained by crystal chemical controls. Because Sr is similar in ionic radius to Ca, it is probably accommodated in the M2 site of clinopyroxene (Blundy & Wood, 1994) and the M4 site of amphibole (LaTourrette et al., 1995). However, crystal chemistry of clinopyroxene and amphibole co-vary strongly, as shown by correlations between jadeite in clinopyroxene vs edenite in amphibole content and between mg-number in both phases (Heinrich, 1986). It can therefore be concluded that variation in DAMP/CPX cannot be explained by differences in crystal chemistry. The DAMP/CPX values are similar for samples CHM30a and CHM30b (Tables 6 and 7, Fig. 9b and c). These are phengite- and zoisite-rich domains, respectively, collected <40 cm apart. Although amphibole from the zoisite-rich domain is poorer in Ce, Sm and Nd than that of the phengite-rich domain, the similar DAMP/CPX suggests similar disequilibrium processes.
Zoisite–clinopyroxene partitioning
DZOI/CPX values scatter widely among the three analysed zoisite-bearing samples (Table 11). Most elements vary over an order of magnitude, indicating different degrees of equilibration between zoisite and clinopyroxene in the sample suite. Only the DZOI/CPX values of sample Z8-77-5 could possibly represent equilibrium, as the trace element disequilibrium is shown by amphibole and clinopyroxene in the other two zoisite-bearing samples (Ad25 and CHM30b). However, DZOI/CPX partition coefficients from sample Z8-77-5 are also uncertain, because the trace element contents of the clinopyroxene are small.
|
The presence of coexisting zoisite and clinozoisite in Z8-77-5 allows calculation of trace element partition coefficients for clinozoisite. In general, the trace element contents of both phases are similar (Table 9) and zoisite–clinozoisite partition coefficients vary by a factor of two. Only Th appears to be significantly richer in clinozoisite than zoisite (DZOI/CZO of 0·32). As a first-order approximation, clinozoisite partitioning behaviour can, therefore, be modelled together with zoisite.
Phengite–clinopyroxene partitioning
For most trace elements, DPHE/CPX values are based on six or seven mineral pairs. Average DPHE/CPX values yield results similar to the subset of the preferred samples (Table 12). For U, Nb, Ba, REE and Th, large uncertainties reflect extremely low element abundances in clinopyroxene (Nb and Ba), phengite (REE) or both phases (Th and U). However, the DPHE/CPX values for Be, B, Sr, Y, Zr, Pb and Th yield consistent results in all samples with standard errors between 5 and 21%, indicating a good approach towards equilibrium between clinopyroxene and phengite. We note that the average of all samples is similar to the average of the preferred samples.
|
Paragonite–clinopyroxene and apatite–clinopyroxene partitioning
In the preferred samples Z6-50-2, Z6-50-13 and Z6-52-1, trace element partitioning between paragonite and clinopyroxene as well as between apatite and clinopyroxene suggests a close approach to equilibrium (Tables 13 and 14). In all samples, DAPA/CPX decreases in the order Ce > Nd > Sm > Y. Large discrepancies between paragonite and clinopyroxene occur only for Li partitioning. For sample Z6-50-14, DPAR/CPX show a wide range for B and DAPA/CPX for REE (Ce, Sm and Nd) also varies widely. Discrepancies in sample Z6-50-14 compared with the other preferred samples reflect trace element disequilibrium or crystal chemical effects caused by the unusually high jadeite content of the clinopyroxene.
|
|
Talc
Trace element concentrations in talc were analysed in sample Z6-55-4 (Table 9). Abundances are extremely low (<100 ppb), commonly below calculated detection limits. DTLC/CPX values for sample Z6-55-4 suggest that talc does not contain significant amounts of large ion lithophile elements, high field strength elements, U and Th.
|
MECHANISMS OF TRACE ELEMENT EQUILIBRATION |
|---|
Equilibration processes
Samples Z6-50-2, Z6-50-13, Z6-52-1 and Z8-77-5 give consistent partition values for a number of phases and trace elements at different abundance levels, indicating a close attainment of trace element equilibrium. These samples display a preferred orientation of all eclogite-facies minerals, which supports the conclusion of Messiga et al. (1995) that syn-metamorphic plastic deformation enhances trace element equilibration processes. Messiga et al. favoured a process of higher surface reaction exchange rates as a result of the continuous production of crystal defects and dislocations during plastic deformation. We note that fine to medium grain size is another common feature of the four equilibrated samples (see below).
Although the presence of a free fluid phase along grain boundaries is thought to be a prerequisite for attaining equilibrium (Jamtveit et al., 1990; Dipple & Ferry, 1992; Zheng et al., 2002), it does not appear to enhance equilibration processes as efficiently as plastic deformation at Trescolmen. Evidence for mobility of Cs, Rb and Ba in the Trescolmen eclogites indicates at least metre-scale fluid transport of these elements during eclogite-facies conditions (Zack et al., 2001). Here, fluid mobility is indicated in both deformed and undeformed eclogites. Because a free fluid phase was present in both foliated and unfoliated eclogites, but trace element equilibration was more complete in deformed ones, deformation rate appears to be more important than the presence of a free fluid in promoting equilibrium. However, we correlate the growth of large, homogeneous and equilibrated omphacite (CPX2 in Fig. 5a) with the influx of fluid, as the overgrowth of strained, small omphacite by homogeneous omphacite is best explained by a dominance of mass transfer processes over dislocation creep (Godard & van Roermund, 1995). Fluid as a mass transfer medium has also been advocated by Philippot & van Roermund (1992) to explain local annealing of omphacite in a former mylonitic fabric. In any case, the interaction between deformation and fluid infiltration is very complex, as evidenced by examples of both fluid-induced deformation (Boundy et al., 1992; Erambert & Austrheim, 1993; Pennacchioni, 1996) as well as deformation-enhanced fluid infiltration (Holness & Graham, 1995; Holness, 1997).
Crystallization sequences in unequilibrated samples
Gabbroic samples CHM30a and CHM30b show disequilibrium features produced under eclogite-facies conditions. Sr partition coefficients differ significantly from those in the equilibrated eclogite samples, in particular for amphibole–clinopyroxene (0·17 vs 1·2; Table 10) and zoisite rim–clinopyroxene partitioning (18 vs 190; Table 11). In contrast, Sr partition coefficients between phengite and clinopyroxene (CHM30a), as well as zoisite rim and amphibole (CHM30b) show ratios similar to those of the well-equilibrated samples (3·5 vs 4·2 and 107 vs 108, respectively). These data reflect a crystallization sequence in the samples, in which phengite and clinopyroxene first crystallized concomitantly, and then zoisite and amphibole crystallized together. At the peak pressure conditions of 2·4 GPa phengite is stable with clinopyroxene (Schmidt & Poli, 1998). Zoisite and amphibole are interpreted to have formed during subsequent fluid infiltration at 2·0 GPa (Zack et al., 2001). Unless both zoisite and amphibole crystallized entirely from an externally derived fluid, they should have exchanged components with the clinopyroxene and garnet. However, only the largest clinopyroxenes (>40 µm) were analysed by LAM and a finer-grained population of clinopyroxene could be in equilibrium with zoisite. In this example, major element homogeneity in clinopyroxene (both small and large clinopyroxenes have been analysed by electron microprobe) might not exclude heterogeneity for certain trace elements.
Protolith grain size as an equilibrium parameter
In general, a close approach to equilibrium for the investigated trace elements was often not reached in metagabbroic samples (CHM30a, CHM30b, Ad25 and Z6-59-1), mostly because of a lack of penetrative deformation in these samples. The only exception is a gabbroic sample with a pronounced foliation defined by clinopyroxene and amphibole (Ad25), which none the less displays a strong heterogeneity both at major and trace element levels. This indicates that grain coarseness of the protolith is another important parameter for attainment of equilibrium. A fine-grained or glassy groundmass in basalts obviously provides a better starting condition than the coarse grain size typical of gabbros. It is of particular interest that studies concerning fluid mobility in eclogite-facies rocks have so far mostly been conducted on metagabbroic and undeformed metabasaltic samples. The small-scale inhomogeneities, as observed for fluid inclusion and oxygen systematics in metagabbroic eclogites from west Alpine ophiolites (Philippot & Selverstone, 1991; Barnicoat & Cartwright, 1997), point to a lack of equilibrium. Only a few studies have concentrated on fluid processes in metabasaltic eclogites (Barnicoat & Cartwright, 1995; Cartwright & Barnicoat, 1999), and as little deformation occurred in the investigated rocks (pillow structures are still preserved), it is not surprising that no sign of equilibrium by fluids has yet been documented in metabasaltic eclogites.
|
CONCLUDING REMARKS |
|---|
Partition coefficients between eclogite-facies hydrous phases (amphibole, zoisite–clinozoisite, phengite, paragonite and apatite) and clinopyroxene for a large suite of trace elements provide a basis for modelling the trace element composition of fluids liberated from dehydrating slabs. The results of this study indicate that penetrative plastic deformation is needed to attain trace element equilibration between eclogite-facies phases, where dynamic recrystallization enhances surface reaction exchange rates. Considerable shear stresses are expected at the slab–mantle interface in subducting slabs (Yuen et al., 1977), so that plastic deformation in the basaltic layer of the subducting oceanic crust is probably the rule rather than the exception. Deformation either is facilitated by fluid infiltration or has promoted fluid mobility; in both cases a free fluid phase is required to transfer trace elements between the phases. Backscattered imaging of Trescolmen eclogites has revealed several generations of clinopyroxene, annealed cracks in garnet and complex zoning patterns in zoisite. These examples show that attempts to derive partition coefficients between coexisting phases by measuring rim–rim compositions adjacent to apparently fresh grain boundaries or by analysing mineral separates cannot be treated as reliable data if microtextural data are not provided.
Mass balance calculations reveal that the minor and accessory minerals zoisite, allanite, phengite and paragonite are important carriers of many trace elements in eclogites. Zoisite and phengite are probably widespread minor phases in the subducting oceanic crust beneath volcanic arcs (Schmidt & Poli, 1998). The element budgets of Sr, Pb, LREE, Th and U are controlled by zoisite, whereas phengite is an important carrier of B, Sr, Ba and Pb (and Rb and Cs; Zack et al., 2001). Therefore DZoisite/Fluid and DPhengite/Fluid values are essential for modelling the trace element composition of fluids liberated from a dehydrating, subducting oceanic crust. Although amphibole, clinopyroxene and garnet carry only small amounts of LREE, Ba, Th and U in eclogite-facies metabasalts (<5% of the trace element mass budget), dehydration modelling for Li and B using only DClinopyroxene/Fluid and DGarnet/Fluid values (Brenan et al., 1998) would account for >80% of these elements.
Major obstacles to combining DMin/Clinopyroxene values with experimental DClinopyroxene/Fluid values are the different chemical compositions of clinopyroxene and PT conditions of the latter. The Trescolmen samples are jadeite rich (25–55 mol % Jd) and have re-equilibrated at 650°C and 2·0 GPa, whereas most experimental clinopyroxenes are diopside rich (<10 mol % Jd), and the experiments have been carried out at 900°C and 2·0 GPa (Brenan et al., 1995a, 1995b, 1998; Stalder et al., 1998). Extrapolation from experiments to our Trescolmen samples requires adjusting experimental DClinopyroxene/Fluid values by applying correction factors of potentially orders of magnitude [for variations in D values as a result of T effects, see, e.g. Brenan et al. (1994)].
Estimated temperature conditions for the Trescolmen eclogites are similar to those for subducting oceanic crust underneath volcanic arcs. Clearly, subducted crust will be subjected to even higher pressure, which results in different mineral parageneses. Although phengite, clinopyroxene and garnet are still stable phases under increased pressure, paragonite and amphibole are no longer stable and zoisite is replaced by lawsonite (Schmidt & Poli, 1998). The partition coefficients from this study must therefore be augmented by trace element partitioning studies of allanite, zircon (not measured in this study) and lawsonite (not stable in Trescolmen) in eclogite- and blueschist-facies rocks.
|
ACKNOWLEDGEMENTS |
|---|
This project was supported by grants from the DFG (Fo 181/10-1) and NSERC. We thank C. A. Heinrich and C. Meyre for discussions about Adula geology and for donating samples. We also thank H.-P. Meyer and A. Kronz (electron microprobe), I. Horn, S. Jackson, P. Sylvester and M. Tubrett (laser ablation microprobe) for helping with the analytical facilities and for fruitful discussions. P. King and G. Hartmann (XRF spectrometry) and L. Hewa (solution-ICP-MS) are thanked for conducting some of the measurements. We thank J. Ayers, H. Becker and P. O’Brien for constructive reviews on an earlier version of this manuscript, and S. S. Sorensen and M. Wilson for editorial handling.
|
FOOTNOTES |
|---|
*Corresponding author. E-mail: tzack{at}min.uni-heidelberg.de
|
REFERENCES |
|---|
Ayers, J. C. & Watson, E. B. (1993). Apatite/fluid partitioning of rare-earth elements and strontium: experimental results at 1·0 GPa and 1000°C and application to models of fluid–rock interaction. Chemical Geology 110, 299–314.[Web of Science]
Ayers, J. C., Dittmer, S. K. & Layne, G. D. (1997). Partitioning of elements between peridotite and H2O at 2·0–3·0 GPa and 900–1100°C, and application to models of subduction zone processes. Earth and Planetary Science Letters 150, 381–398.[Web of Science]
Barnicoat, A. C. & Cartwright, I. (1995). Focused fluid flow during subduction: oxygen isotope data from high-pressure ophiolites of the Western Alps. Earth and Planetary Science Letters 132, 53–61.[Web of Science]
Barnicoat, A. C. & Cartwright, I. (1997). The gabbro–eclogite transformation: an oxygen isotope and petrographic study of west Alpine ophiolites. Journal of Metamorphic Geology 15, 93–104.[Web of Science]
Beattie, P. (1993). On the occurrence of apparent non-Henry’s Law behaviour in experimental partitioning studies. Geochimica et Cosmochimica Acta 57, 47–55.[Web of Science]
Becker, H., Jochum, K. P. & Carlson, R. W. (1999). Constraints from high-pressure veins in eclogites on the composition of hydrous fluids in subduction zones. Chemical Geology 160, 291–308.[Web of Science]
Becker, H., Jochum, K. P. & Carlson, R. W. (2000). Trace element fractionation during dehydration of eclogites from high-pressure terranes and the implications for element fluxes in subduction zones. Chemical Geology 163, 65–99.[Web of Science]
Beswick, A. E. (1973). An experimental study of alkali metal distributions in feldspars and micas. Geochimica et Cosmochimica Acta 37, 183–208.[Web of Science]
Blundy, J. & Wood, B. (1994). Prediction of crystal–melt partition coefficients from elastic moduli. Nature 372, 452–454.
Boundy, T. M., Fountain, D. M. & Austrheim, H. (1992). Structural development and petrofabrics of eclogite facies shear zones, Bergen Arcs, western Norway: implications for deep crustal deformational processes. Journal of Metamorphic Geology 10, 127–146.[Web of Science]
Brenan, J. M., Shaw, H. F., Phinney, D. L. & Ryerson, F. J. (1994). Rutile–aqueous fluid partitioning of Nb, Ta, Hf, Zr, U and Th: implications for high field strength element depletions in island-arc basalts. Earth and Planetary Science Letters 128, 327–339.[Web of Science]
Brenan, J. M., Shaw, H. F. & Ryerson, F. J. (1995a). Experimental evidence for the origin of lead enrichment in convergent-margin magmas. Nature 389, 54–56.
Brenan, J. M., Shaw, H. F., Ryerson, F. J. & Phinney, D. L. (1995b). Mineral–aqueous fluid partitioning of trace elements at 900°C and 2·0 GPa: constraints on the trace element chemistry of mantle and deep crustal fluids. Geochimica et Cosmochimica Acta 59, 3331–3350.[Web of Science]
Brenan, J. M., Ryerson, F. J. & Shaw, H. F. (1998). The role of aqueous fluids in the slab-to-mantle transfer of boron, beryllium, and lithium during subduction: experiments and models. Geochimica et Cosmochimica Acta 62, 3337–3347.[Web of Science]
Cartwright, I. & Barnicoat, A. C. (1999). Stable isotope geochemistry of Alpine ophiolites: a window to ocean-floor hydrothermal alteration and constraints on fluid–rock interaction during high-pressure metamorphism. International Journal of Earth Sciences 88, 219–235.[Web of Science]
Clague, D. A., Frey, F. A., Thompson, G. & Rindge, S. (1981). Minor and trace element geochemistry of volcanic rocks dredged from the Galapagos spreading center: role of crystal fractionation and mantle heterogeneity. Journal of Geophysical Research 86, 9469–9482.
Creaser, R. A., Heaman, L. M. & Erdmer, P. (1997). Timing of high-pressure metamorphism in the Yukon–Tanana terrane, Canadian Cordillera: constraints from U–Pb zircon dating of eclogite from the Teslin tectonic zone. Canadian Journal of Earth Sciences 34, 709–715.[Web of Science]
Deer, W.A., Howie, R.A. & Zussman, J. (1992). An Introduction to the Rock-forming Minerals, 2nd edn. Harlow, UK: Addison Wesley Longman, p. 696.
Dipple, G. M. & Ferry, J. M. (1992). Metasomatism and fluid flow in ductile fault zones. Contributions to Mineralogy and Petrology 112, 149–164.
Domanik, K. J., Hervig, R. L. & Peacock, S. M. (1993). Beryllium and boron in subduction zone minerals: an ion microprobe study. Geochimica et Cosmochimica Acta 57, 4997–5010.[Web of Science]
Erambert, M. & Austrheim, H. (1993). The effect of fluid and deformation on zoning and inclusion patterns in poly-metamorphic garnets. Contributions to Mineralogy and Petrology 115, 204–214.[Web of Science]
Franz, G. & Selverstone, J. (1992). An empirical phase diagram for the clinozoisite–zoisite transformation in the system Ca2Al3Si3O12(OH)–Ca2Al2Fe3+Si3O12(OH). American Mineralogist 77, 631–642.[Abstract]
Froitzheim, N., Schmid, S. M. & Frey, M. (1996). Mesozoic paleogeography and the timing of eclogite-facies metamorphism in the Alps: a working hypothesis. Eclogae Geologicae Helvetiae 89, 81–110.[Web of Science]
Getty, S. R. & Selverstone, J. (1994). Stable isotopic and trace element evidence for restricted fluid migration in 2 GPa eclogites. Journal of Metamorphic Geology 12, 747–760.[Web of Science]
Godard, G. & van Roermund, H. L. M. (1995). Deformation-induced clinopyroxene fabrics from eclogites. Journal of Structural Geology 17, 1425–1443.[Web of Science]
Guidotti, C. V. & Sassi, F. P. (1998). Miscellaneous isomorphous substitutions in Na–K white micas: a review, with special emphasis to metamorphic micas. Rendiconti Lincei, Scienze Fisiche e Naturali 9, 57–78.
Heinrich, C. A. (1982). Kyanite–eclogite to amphibolite facies evolution of hydrous mafic and pelitic rocks, Adula nappe, Central Alps. Contributions to Mineralogy and Petrology 81, 30–38.[Web of Science]
Heinrich, C. A. (1986). Eclogite facies regional metamorphism of hydrous mafic rocks in the Central Alpine Adula Nappe. Journal of Petrology 27, 123–154.
Hickmott, D. D. & Shimizu, N. (1990). Trace element zoning in garnet from the Kwoiek Area, British Columbia: disequilibrium partitioning during garnet growth. Contributions to Mineralogy and Petrology 104, 619–630.[Web of Science]
Hickmott, D. D., Sorensen, S. S. & Rogers, P. S. Z. (1992). Metasomatism in a subduction complex: constraints from microanalysis of trace elements in minerals from garnet amphibolite from the Catalina Schist. Geology 20, 347–350.
Holness, M. B. (1997). Fluid flow paths and mechanisms of fluid infiltration in carbonates during contact metamorphism: the Beinn an Dubhaich aureole, Skye. Journal of Metamorphic Geology 15, 59–70.[Web of Science]
Holness, M. B. & Graham, C. M. (1995). P–T–X effects on equilibrium carbonate–H2O–CO2–NaCl dihedral angles: constraints on carbonate permeability and the role of deformation during fluid infiltration. Contributions to Mineralogy and Petrology 119, 301–313.[Web of Science]
Horn, I., Hinton, R. W., Jackson, S. E. & Longerich, H. P. (1997). Ultra-trace element analysis of NIST SRM 616 using laser ablation microprobe-inductively coupled plasma-mass spectrometry (LAM-ICP-MS): a comparison with secondary ion mass spectrometry (SIMS). Geostandards Newsletter 21, 191–203.[Web of Science]
Jackson, S. E., Longerich, H. P., Dunning, G. R. & Fryer, B. J. (1992). The application of laser ablation microprobe inductively coupled plasma-mass spectrometry to in situ trace element determinations in minerals. Canadian Mineralogist 30, 1049–1064.[Web of Science]
Jamtveit, B., Bucher-Nurminen, K. & Austrheim, H. (1990). Fluid controlled eclogitization of granulites in deep crustal shear zones, Bergen arcs, Western Norway. Contributions to Mineralogy and Petrology 104, 184–193.[Web of Science]
Kohn, M. J. & Valley, J. W. (1998). Effects of cation substitutions in garnet and pyroxene on equilibrium oxygen isotope fractionations. Journal of Metamorphic Geology 16, 625–639.[Web of Science]
Kretz, R., Campbell, J. L., Hoffman, E. L., Hartree, R. & Teesdale, W. J. (1999). Approaches to equilibrium in the distribution of trace elements among the principal minerals in a high-grade metamorphic terrane. Journal of Metamorphic Geology 17, 41–59.[Web of Science]
Kurz, W., Neubauer, F. & Dachs, E. (1998a). Eclogite meso- and microfabrics: implications for the burial and exhumation history of eclogites in the Tauern Window (eastern Alps) from P–T–d paths. Tectonophysics 285, 183–209.[Web of Science]
Kurz, W., Neubauer, F., Genser, J. & Dachs, E. (1998b). Alpine geodynamic evolution of passive and active continental margin sequences in the Tauern Window (eastern Alps, Austria, Italy): a review. Geologische Rundschau 87, 225–242.
LaTourrette, T., Hervig, R. L. & Holloway, J. R. (1995). Trace element partitioning between amphibole, phlogopite, and basanite melt. Earth and Planetary Science Letters 135, 13–30.[Web of Science]
Leake, B. E., Woolley, A. R., Arps, C. E. S., Birch, W. D., Gilbert, M. C., Grice, J. D., et al. (1997). Nomenclature of amphiboles: report of the subcommittee on amphibole of the International Mineralogical Association, Commission on New Minerals and Mineral Names. Canadian Mineralogist 35, 219–246.[Web of Science]
Longerich, H. P., Guenther, D. & Jackson, S. E. (1996). Elemental fractionation in laser ablation inductively coupled plasma mass spectrometry. Fresenius’ Journal of Analytical Chemistry 355, 538–542.
Melzer, S. & Wunder, B. (2000). Island arc basalt alkali-ratios: constraints from phengite–fluid partitioning experiments. Geology 28, 583–586.
Melzer, S., Gottschalk, M. & Heinrich, W. (1998). Experimentally determined partitioning of Rb between richterites and aqueous (Na, K)-chloride solutions. Contributions to Mineralogy and Petrology 133, 315–328.[Web of Science]
Messiga, B., Tribuzio, R., Bottazzi, P. & Ottolini, L. (1995). An ion microprobe study on trace element composition of clinopyroxenes from blueschist and eclogitized Fe–Ti-gabbros, Ligurian Alps, northwestern Italy: some petrologic considerations. Geochimica et Cosmochimica Acta 59, 59–75.[Web of Science]
Messiga, B., Kienast, J. R., Rebay, G., Riccardi, M. P. & Tribuzio, R. (1999). Cr-rich magnesiochloritoid eclogites from the Monviso ophiolites (Western Alps, Italy). Journal of Metamorphic Geology 17, 287–299.[Web of Science]
Meyre, C. & Frey, M. (1998). Eclogite facies metamorphism and deformation of the middle Adula nappe (Central Alps, Switzerland): excursion to Trescolmen. Schweizerische Mineralogische und Petrographische Mitteilungen 78, 355–362.[Web of Science]
Meyre, C. & Puschnig, A. R. (1993). High-pressure metamorphism and deformation at Trescolmen, Adula nappe, Central Alps. Schweizerische Mineralogische und Petrographische Mitteilungen 73, 277–283.[Web of Science]
Meyre, C., de Capitani, C. & Partzsch, J. H. (1997). A ternary solid solution model for omphacite and its application to geothermobarometry of eclogites from the Middle Adula nappe (Central Alps, Switzerland). Journal of Metamorphic Geology 15, 687–700.[Web of Science]
Meyre, C., de Capitani, C., Zack, T. & Frey, M. (1999). Petrology of high-pressure metapelites from the Adula nappe (Central Alps, Switzerland). Journal of Petrology 40, 199–213.[Web of Science]
Miller, C. & Thöni, M. (1995). Origin of eclogites from the Austroalpine Ötztal basement (Tirol, Austria): geochemistry and Sm–Nd vs Rb–Sr isotope systematics. Chemical Geology 122, 199–225.[Web of Science]
Miller, C., Stosch, H.-G. & Hoernes, S. (1988). Geochemistry and origin of eclogites from the type locality Koralpe and Saualpe, Eastern Alps, Austria. Chemical Geology 67, 103–118.[Web of Science]
Nagasaki, A. & Enami, M. (1998). Sr-bearing zoisite and epidote in ultra-high pressure (UHP) metamorphic rocks from the Su-Lu province, eastern China: an important Sr reservoir under UHP conditions. American Mineralogist 83, 240–247.[Abstract]
Najorka, J., Gottschalk, M., Franz, G. & Heinrich, W. (1999). Ca–Sr distribution among amphibole, clinopyroxene, and chloride-bearing solutions. American Mineralogist 84, 596–606.[Abstract]
Okay, A. I. (1995). Paragonite eclogites from Dabie Shan, China: re-equilibration during exhumation? Journal of Metamorphic Geology 13, 449–460.[Web of Science]
Pearce, J. A. & Cann, J. R. (1973). Tectonic setting of basic volcanic rocks determined using trace element analyses. Earth and Planetary Science Letters 19, 290–300.[Web of Science]
Pennacchioni, G. (1996). Progressive eclogitization under fluid-present conditions of pre-Alpine mafic granulites in the Austroalpine Mt Emilius Klippe (Italian Western Alps). Journal of Structural Geology 18, 549–561.[Web of Science]
Philippot, P. & Selverstone, J. (1991). Trace-element-rich brines in eclogitic veins: implications for fluid composition and transport during subduction. Contributions to Mineralogy and Petrology 106, 417–430.[Web of Science]
Philippot, P. & van Roermund, H. L. M. (1992). Deformation processes in eclogitic rocks: evidence for the rheological delamination of the oceanic crust in deeper levels of subduction zones. Journal of Structural Geology 14, 1059–1077.[Web of Science]
Santini, L. (1992). Geochemistry and Geochronology of the Basic Rocks of the Penninic Nappes of East–Central Alps (Switzerland). Lausanne: Faculté des Sciences, Université de Lausanne, p. 200.
Schilling, J. G., Zajac, M., Evans, R., Johnston, T., White, W., Devine, J. D. & Kingsley, R. (1985). Petrologic and geochemical variations along the Mid-Atlantic Ridge from 29°N to 73°N. American Journal of Science 283, 510–586.
Schmidt, M. W. & Poli, S. (1998). Experimental based water budgets for dehydrating slabs and consequences for arc magma generation. Earth and Planetary Science Letters 163, 361–379.[Web of Science]
Schumacher, J. C. (1997). Appendix 2: The estimation of the proportion of ferric iron in the electron-microprobe analysis of amphiboles. Canadian Mineralogist 35, 238–246.
Skulski, T., Minarik, W. & Watson, E. B. (1994). High-pressure experimental trace-element partitioning between clinopyroxene and basaltic melts. Chemical Geology 117, 127–147.[Web of Science]
Sorensen, S. S. & Grossman, J. N. (1989). Enrichment of trace elements in garnet amphibolites from a paleo-subduction zone: Catalina Schist, California. Geochimica et Cosmochimica Acta 53, 3155–3177.[Web of Science]
Sorensen, S. S., Grossman, J. N. & Perfit, M. R. (1997). Phengite-hosted LILE enrichment in eclogite and related rocks: implications for fluid-mediated mass transfer in subduction zones and arc magma genesis. Journal of Petrology 38, 3–34.[Web of Science]
Stalder, R., Foley, S. F., Brey, G. P. & Horn, I. (1998). Mineral–aqueous fluid partitioning of trace elements at 900°C–1200°C and 3·0 GPa to 5·7 GPa: new experimental data set for garnet, clinopyroxene and rutile and implications for mantle metasomatism. Geochimica et Cosmochimica Acta 62, 1781–1801.[Web of Science]
Stroncik, N. A. & Schmincke, H.-U. (2001). Evolution of palagonite: crystallization, chemical changes, and element budget. Geochemistry, Geophysics, Geosystems 2, paper 2000GC000102.
Thöni, M. & Jagoutz, E. (1992). Some new aspects of dating eclogites in orogenic belts: Sm–Nd, Rb–Sr, and Pb–Pb isotopic results from the Austroalpine Saualpe and Koralpe type-locality (Carinthia/Styria, southeastern Austria). Geochimica et Cosmochimica Acta 56, 347–368.[Web of Science]
Tracy, R. J. (1982). Compositional zoning and inclusions in metamorphic minerals. In: Ferry, J. M. (ed.) Characterization of Metamorphism through Mineral Equilibria. Mineralogical Society of America, Reviews in Mineralogy 10, 355–397.
Tribuzio, R., Messiga, B., Vannucci, R. & Bottazzi, P. (1996). Rare earth element redistribution during high-pressure low-temperature metamorphism in ophiolitic Fe-gabbros (Liguria, northwestern Italy): implications for light REE mobility in subduction zones. Geology 24, 711–714.
Volfinger, M. (1976). Éffet de la température sur les distributions de Na, Rb et Cs entre la sanidine, la muscovite, la phlogopite et une solution hydrothermale sous une pression de 1 kbar. Geochimica et Cosmochimica Acta 40, 267–282.[Web of Science]
Watson, E. B. (1996). Surface enrichment and trace-element uptake during crystal growth. Geochimica et Cosmochimica Acta 60, 5013–5020.[Web of Science]
Widmer, T. W. (1996). Entwässerung ozeanisch alterierter Basalte in Subduktionszonen (Zone von Zermatt– Saas Fee). Ph.D. thesis, ETH Zürich.
Wiesli, R. A., Taylor, L. A., Valley, J. W., Trommsdorff, V. & Kurosawa, M. (2001). Geochemistry of eclogites and metapelites from Trescolmen, Central Alps, as observed from major and trace elements and oxygen isotopes. International Geology Reviews 43, 95–119.
Yang, P. & Rivers, T. (2001). Chromium and manganese zoning in pelitic garnet and kyanite: spiral, overprint, and oscillatory (?) zoning patterns and the role of growth rate. Journal of Metamorphic Geology 19, 455–474.[Web of Science]
Yang, P., Rivers, T. & Jackson, S. (1999). Crystal-chemical and thermal controls on trace-element partitioning between coexisting garnet and biotite in metamorphic rocks from Western Labrador. Canadian Mineralogist 37, 443–468.[Web of Science]
Yuen, D. A., Fleitout, L., Schubert, G. & Froideveaux, C. (1977). Shear deformation zones along major transform faults and subducting slabs. Geophysical Journal of the Royal Astronomical Society 54, 93–119.[Web of Science]
Zack, T. & Foley, S. F. (1997). First laser ablation microprobe measurements of trace-element partitioning between hydrous phases in high-pressure metamorphic rocks. In: Seventh Annual V. M. Goldschmidt Conference. Lunar and Planetary Institute Contribution 921, 226.
Zack, T., Rivers, T. & Foley, S. F. (2001). Cs–Rb–Ba systematics in phengite and amphibole: an assessment of fluid mobility at 2·0 GPa in eclogites from Trescolmen, Central Alps. Contributions to Mineralogy and Petrology 140, 651–669.[Web of Science]
Zack, T., Kronz, A., Foley, S. F. & Rivers, T. (2002). Trace element abundances in rutiles from eclogites and associated garnet mica schists. Chemical Geology 184, 97–122.[Web of Science]
Zheng, Y.-F., Wang, Z.-R., Li, S.-G. & Zhao, Z.-F. (2002). Oxygen isotope equilibrium between eclogite minerals and its constraints on mineral Sm–Nd chronometer. Geochimica et Cosmochimica Acta 66, 625–634.[Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Konrad-Schmolke, P. J. O'Brien, and F. Heidelbach Compositional re-equilibration of garnet: the importance of sub-grain boundaries European Journal of Mineralogy, July 1, 2007; 19(4): 431 - 438. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. USUI, E. NAKAMURA, and H. HELMSTAEDT Petrology and Geochemistry of Eclogite Xenoliths from the Colorado Plateau: Implications for the Evolution of Subducted Oceanic Crust J. Petrology, May 1, 2006; 47(5): 929 - 964. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. M. Breeding, J. J. Ague, and M. Brocker Fluid-metasedimentary rock interactions in subduction-zone melange: Implications for the chemical composition of arc magmas Geology, December 1, 2004; 32(12): 1041 - 1044. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. ZACK, T. RIVERS, R. BRUMM, and A. KRONZ Cold subduction of oceanic crust: Implications from a lawsonite eclogite from the Dominican Republic European Journal of Mineralogy, December 1, 2004; 16(6): 909 - 916. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Frei, A. Liebscher, G. Franz, and P. Dulski Trace Element Geochemistry of Epidote Minerals Reviews in Mineralogy and Geochemistry, January 1, 2004; 56(1): 553 - 605. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||