| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology Volume 42 Number 1 Pages 55-67 2001
© Oxford University Press 2001
Fluid and Element Cycling in Subducted Serpentinite: a Trace-Element Study of the Erro–Tobbio High-Pressure Ultramafites (Western Alps, NW Italy)
1DIPARTIMENTO PER LO STUDIO DEL TERRITORIO E DELLE SUE RISORSE, UNIVERSITÀ DI GENOVA, CORSO EUROPA 26, 16132 GENOVA, ITALY
2DIPARTIMENTO DI SCIENZE DELLA TERRA, UNIVERSITÀ DI MILANO, VIA BOTTICELLI 23, MILANO, ITALY
Received November 16, 1999; Revised typescript accepted July 4, 2000
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONGEOLOGICAL AND PETROLOGICAL...SAMPLE DESCRIPTION AND...BULK-ROCK COMPOSITIONSMINERAL COMPOSITIONSDISCUSSIONSUMMARYREFERENCES |
|---|
The oceanic serpentinization of peridotites and the influence of such an alteration on element cycling during their subduction dewatering are here investigated in a mantle slice (Erro–Tobbio peridotite), first exposed to oceanic serpentinization and later involved in alpine subduction, partial dewatering and formation of a high-pressure olivine + titanian-clinohumite + diopside + antigorite assemblage in the peridotites and in veins. Previous work indicates that high-pressure veins include primary brines, representing a residue after crystallization of the vein assemblage and containing recycled oceanic Cl and alkalis. To reconstruct the main changes during oceanic peridotite serpentinization and subsequent subduction, we analysed samples in profiles from serpentinized oceanic peridotites to high-pressure serpentinites, and from high-pressure ultramafites to veins. Here we present results indicating that the main features of the oceanic serpentinization are immobility of rare earth elements (REE), considerable water increase, local CaO decrease and uptake of trace amounts of Sr, probably as a function of the intensity of alteration. Sr entered fine-grained Ca phases associated with serpentine and chlorite. Trace-element analyses of mantle clinopyroxenes and high-pressure diopsides (in country ultramafites and veins), highlight the close similarity in the REE compositions of the various clinopyroxenes, thereby indicating rock control on the vein fluids and lack of exotic components carried by externally derived fluids. Presence of appreciable Sr contents in vein-forming diopside indicates cycling of oceanic Sr in the high-pressure fluid. This, together with the recognition of pre-subduction Cl and alkalis in the vein fluid, indicates closed-system behaviour during eclogitization and internal cycling of exogenic components. Diopside and Ti-clinohumite are the high-pressure minerals acting as repositories for REE and Sr, and for high field strength elements (HFSE), respectively. The aqueous fluid equilibrated with such an assemblage is enriched in Cl and alkaline elements but strongly depleted in REE and HFSE (less than chondrite abundances). Sr is low [(0·2–1·6) x chondrites], although selectively enriched relative to light REE.
KEY WORDS: eclogite facies; fluid; trace elements; serpentinite; subduction
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL AND PETROLOGICAL...SAMPLE DESCRIPTION AND...BULK-ROCK COMPOSITIONSMINERAL COMPOSITIONSDISCUSSIONSUMMARYREFERENCES |
|---|
Cycling of crustal materials in subduction zones is accompanied by production of fluids and/or melts whose transfer from the subducting plate to the overlying mantle controls the onset of magmatism at convergent plate margins. Distinctive signatures of arc basalts include depletion in high field strength elements (HFSE) and selective enrichment in large ion lithophile elements (LILE), light rare earth elements (LREE), 10Be, B and Cl. This indicates contamination of their mantle sources by solute-rich agents containing several exogenic components, e.g. 10Be and Cl (Tera et al., 1986
; Philippot et al., 1998). Although some experimental studies constrain the elemental fluxes in subduction zones through element partitioning between minerals and fluids or melts (Brenan et al., 1995a, 1995b; Ayers et al., 1997; Stalder et al., 1998), little is known about the actual properties and compositions of natural agents and about their release during subduction. Studies of fluid inclusions in alpine eclogite-facies rocks recognize saline aqueous solutions as potential carriers of major and trace elements (Philippot & Selverstone, 1991; Scambelluri et al., 1997). However, these fluids are often residues after rock dehydration, fluid–rock exchange and precipitation of eclogitic veins (Scambelluri & Philippot, 2000), and at present there is little evidence of pristine eclogitic fluids, and there are few analyses of such fluids.
Concerning the process of fluid release at convergent plate margins, subduction of hydrous oceanic mantle is an important variable, as it represents a major source of deep dehydration water. Serpentinized peridotites are widespread in oceanic basins, and can contain 10–13 wt % bulk H2O fixed in serpentine and associated hydrous phases after interaction with seawater. Experiments and petrology of alpine peridotites indicate that serpentine stability to very high pressures allows considerable fluxing of surface waters into the mantle (Scambelluri et al., 1995; Ulmer & Trommsdorff, 1995). Formation of high-pressure veins with primary fluid inclusions is compelling evidence of deep fluid release in subducted serpentinites. The trapped fluid dissolves up to 50 wt % chlorine and alkalis, i.e. components formerly hosted in oceanic hydrous minerals and then inherited by the subduction-zone fluid (Scambelluri et al., 1997). The compositional relationships between oceanic and high-pressure assemblages thus become crucial to define deep fluid and element cycling by serpentinized peridotites and to assess the control of pre-subduction alteration on the composition of subduction fluids. Allied questions concern the other components involved in the transfer process, and the control exerted by the high-pressure minerals on the trace-element composition of associated fluids.
We investigate the above features in the alpine Erro–Tobbio peridotite, a slice of hydrous oceanic mantle involved in alpine subduction and eclogitization, by studying the compositional features related to oceanic serpentinization and later subduction of these rocks.
|
GEOLOGICAL AND PETROLOGICAL BACKGROUND |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL AND PETROLOGICAL...SAMPLE DESCRIPTION AND...BULK-ROCK COMPOSITIONSMINERAL COMPOSITIONSDISCUSSIONSUMMARYREFERENCES |
|---|
The Erro–Tobbio peridotite (Ligurian Alps) corresponds to subcontinental mantle first exhumed and hydrated during opening of the Jurassic Tethyan ocean, and then involved in alpine subduction and high-pressure recrystallization during subsequent closure of the oceanic basin (Hoogerduijn Strating et al., 1993
; Scambelluri et al., 1995). During subduction, ductile deformation was focused in serpentinite shear zones (high-strain domains) surrounding volumes of serpentinized peridotite unaffected by the plastic deformation (low-strain domains). The latter still preserve mineralogical and textural records of both mantle and oceanic history (Fig. 1).
|
In low-strain domains, granular spinel lherzolites are overprinted by tectonitic to mylonitic lherzolites showing decompressional recrystallization to spinel-, plagioclase- and hornblende-bearing assemblages (Hoogerduijn Strating et al., 1993). Subsequent intrusion of mid-ocean ridge gabbros and basalts indicates the early shallow exposure of this lithospheric mantle in an oceanic setting. Later peridotite serpentinization (and local rodingitization of the mafic dykes) caused development of chlorine-bearing chrysotile and lizardite, chlorite, magnetite, minor brucite, and chlorine- and alkali-bearing phyllosilicates. The low-grade nature of such an assemblage and the mineral compositions point to peridotite interaction with Cl- and alkali-bearing solutions, presumably seawater derived (Scambelluri et al., 1997). In low-strain domains, subduction led to patchy replacement of mantle and hydrothermal assemblages by radial aggregates of antigorite, fine-grained olivine, titanian-clinohumite, magnetite and diopside (Fig. 1). The peridotite volumes recording this recrystallization display olivine + titanian-clinohumite-bearing veins, coeval with an eclogitic foliation in the associated mafic dykes. This indicates a high-pressure origin of the olivine-bearing assemblages in ultramafites, and production of eclogitic vein fluids.
In the high-strain domains, serpentinite mylonites display an antigorite foliation enclosing boudins of eclogitized metagabbro and metarodingite, and cut by olivine-bearing veins and later olivine + antigorite shear bands (Fig. 1). The olivine assemblage thus represents the highest grade achieved during subduction: 2–2·5 GPa and 550–600°C based on the eclogitic paragenesis in metagabbros (Messiga et al., 1995). The altered Erro–Tobbio peridotites thus underwent high-pressure partial deserpentinization producing peak olivine + titanian-clinohumite + fluid in presence of stable antigorite. The high-pressure minerals lack Cl and alkalis, and veins contain primary hypersaline fluid inclusions with sodium and potassium chlorides as main daughter crystals (Scambelluri et al., 1997). This highlights partitioning of chlorine and alkalis into the eclogitic fluid and cycling of oceanic substances at eclogite facies.
|
SAMPLE DESCRIPTION AND ANALYTICAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL AND PETROLOGICAL...SAMPLE DESCRIPTION AND...BULK-ROCK COMPOSITIONSMINERAL COMPOSITIONSDISCUSSIONSUMMARYREFERENCES |
|---|
To study the compositional variations during serpentinization and the element redistribution during deep dewatering and veining, the analytical work was focused on: (1) two profiles from cores of peridotites with variable oceanic serpentinization (with relict mantle assemblages and no high-pressure minerals), to high-pressure serpentinite mylonite shear zones (Fig. 1; profiles A and B); (2) several vein to host-rock profiles in low- and high-strain domains (Fig. 1; profiles C and D, respectively).
In profiles A and B, samples were collected at regular distances of about 15 m. Profile A is composed of ETA46, ETA47 and ETA51. ETA46 is a serpentinized peridotite tectonite core consisting of olivine, clinopyroxene, orthopyroxene and spinel porphyroclasts partly recrystallized into plagioclase-facies assemblages. Olivine is overgrown by oceanic chrysotile and lizardite, plagioclase is replaced by chlorite ± epidote, and thin carbonate veinlets are locally present. No alpine high-pressure minerals are present. In ETA47 (border of the preserved lherzolite), the same mantle and oceanic textures and assemblages as in ETA46 are statically overgrown by the peak alpine assemblage, i.e. antigorite, olivine, titanian-clinohumite and diopside. ETA51 (the alpine serpentinite mylonite) shows a prograde foliation of prevalent antigorite, chlorite, magnetite, subordinated diopside and titanian-clinohumite. Porphyroclasts of relict mantle clinopyroxene are diffusely preserved. This foliation is cut by shear bands with fine-grained olivine, antigorite and titanian-clinohumite. Profile B is composed of ETF1, ETF2, ETF3, ETF4, ETF6 and ETF7. ETF1, ETF2 and ETF3 represent a preserved spinel tectonite core. ETF1 (the inner part) has a tectonite foliation of olivine, clinopyroxene, orthopyroxene and spinel porphyroclasts partly recrystallized in granoblastic aggregates containing the same spinel-bearing assemblage. The mantle minerals are overgrown by oceanic chrysotile and lizardite. ETF2 and ETF3 have the same structure and mineralogy as ETF1 and display an increasing intensity of serpentinization (50 vol. % chrysotile and lizardite in ETF2, 55 vol. % in ETF3, based on X-ray diffraction). In ETF3 fine-grained Ca phases coexist in veins with chlorite and serpentine. Samples ETF4, ETF6 and ETF7 form a shear-zone surrounding the spinel peridotite (ETF7 being the inner part of the mylonite zone). They contain some mantle clinopyroxene relics in a foliated matrix of high-pressure antigorite, chlorite, magnetite and minor diopside. This foliation is cut by olivine + antigorite + titanian-clinohumite shear bands.
Samples ETA71, ETA74, ETA75, ETA75A, ETA76, ET42 and ET42A are veins and their wall- and host-rocks in low-strain domains. ETA71, ETA75 and ET42 are olivine + titanian-clinohumite + magnetite + diopside ± chlorite veins with variable mineral abundances: ETA71 contains diopside and accessory Ti-clinohumite, olivine and magnetite; ETA75 and ET42 are richer in titanian-clinohumite and olivine, and display comparable mineralogies and modal proportions. ETA74, ETA75A and ET42A are peridotites collected at centimetre distances from the veins (i.e. wall-rocks), ETA76 is at metre distance (i.e. host-rock). Wall- and host-rocks do not display significant petrographic differences; they preserve the mantle textures and some mantle clinopyroxene and olivine. Their major constituents are peak olivine + titanian-clinohumite + antigorite + diopside grown as static pseudomorphic replacements. ETF9, ETF10 and ETF11 constitute a profile from vein (ETF10) to host-rock serpentinite (ETF9 and ETF11, 
1 m from the vein) in a mylonite zone. These host-rocks display an antigorite + chlorite foliation (with relict porphyroclasts of mantle olivine and clinopyroxene) cut by olivine + titanian-clinohumite + antigorite shear bands. Vein ETF10 contains abundant chlorite.
Whole-rock major-element compositions were determined by X-ray fluorescence (XRF). Trace and rare earth element (REE) concentrations were measured by inductively coupled plasma-mass spectrometry (ICP-MS) with a VG PQ2 instrument, at the Institut des Sciences de la Terre, de l’Eau et de l’Espace de Montpellier [the analytical procedure has been described by Ionov et al. (1992)].
The major-element mineral analyses were performed by energy-dispersive spectrometry with a Philips SEM 515 at the Dipartimento di Scienze della Terra, Genoa, using accelerating potential of 15 kV, beam current of 20 nA, counting times of 100 s, and natural mineral standards. Mineral analyses were also measured by wavelength-dispersive spectrometry using an ARL SEMQ electron microprobe at 15 kV, sample current of 15 nA, at the Dipartimento di Scienze della Terra, Milan. Trace-element mineral analyses were carried out with a Cameca IMS 4F ion microprobe at the CNR-CSCC, Pavia [the analytical procedure has been described by Bottazzi et al. (1991)]. Trace-element abundances have been normalized to an average C1 chondrite composition (Anders & Ebihara, 1982).
|
BULK-ROCK COMPOSITIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL AND PETROLOGICAL...SAMPLE DESCRIPTION AND...BULK-ROCK COMPOSITIONSMINERAL COMPOSITIONSDISCUSSIONSUMMARYREFERENCES |
|---|
Profiles from serpentinized peridotites to serpentinite mylonites
Most peridotite samples in profiles A and B (Fig. 1; Table 1) display coherent variations in Al2O3 and CaO (Al2O3 = 1·87–2·68 wt %; CaO = 1·97–2·38 wt %), and preserve Al/Ca ratios comparable with mantle values (Hofmann, 1988
). One exception is the serpentinite ETA51, which has 2·42 wt % Al2O3 coupled with rather low CaO (1·67 wt %), probably resulting from a higher intensity of serpentinization and Ca loss caused by partial clinopyroxene dissolution.
|
The trace-element compositions of serpentinized peridotites and serpentinite mylonites in both profiles (Table 1; Fig. 2) do not display significant REE variability. As classically observed in orogenic peridotites, the studied samples show overall depletion of LREE relative to heavy REE (HREE) (CeN/YbN = 0·015–0·076). All analysed peridotites display negative Zr anomalies with respect to adjacent REE: comparable negative Zr anomalies observed in other depleted peridotites (e.g. the Internal Liguride peridotites of the Northern Apennines) were ascribed to different solid–melt partition coefficients of these elements during fractional melting (Rampone et al., 1996). Thus, the Zr anomaly is probably inherited from earlier mantle melting. On the other hand, Fig. 2 illustrates that ultramafites in profiles A and B display increasing Sr from serpentinized oceanic peridotite to serpentinite mylonites (e.g. samples ETA46 to ETA51). Comparable positive Sr anomalies are not normally observed in depleted mantle peridotites from the Alpine–Apennine chain, where negative Sr anomalies are acquired as a result of fractional melting (Rampone et al., 1996). The measured positive Sr anomalies were thus acquired during secondary, post-melting, enrichment.
|
Profiles from high-pressure veins to host-rocks
The major-element compositions of veins and host-rocks are reported in Table 2. Whole wall- and host-rocks display Al2O3 and CaO variations comparable with those of ultramafites from profiles A and B. In particular, the wall- and host-rock samples ETA75 and ETA76 resemble the serpentinite ETA51 in being depleted in CaO (0·66 wt % and 0·79 wt %, respectively) relative to Al2O3 (1·72 wt % and 1·87 wt %). The veins display bulk major-element compositions significantly different from those of the surrounding peridotites. Moreover, veins display a compositional variability that reflects their variable mineral abundances: vein ETA71, with the highest CaO (13·21 wt %), has the highest diopside content (see Table 2 for modal estimates); ETA75 and ETA72 have higher TiO2 (1·23 wt % and 1·6 wt %) as a result of higher titanian-clinohumite content. Vein ETF10 is H2O and Al2O3 rich because of appreciable modal chlorite.
|
The absolute REE concentrations in wall- and host-rocks are lower than chondrite values (Table 2; Fig. 3) and are comparable with those of ultramafites in profiles A and B. The absolute REE concentrations of veins are higher than those of wall- and host-rocks, with the exception of vein ETF10 (Fig. 3b). In accordance with the major-element compositions, the trace-element abundances in veins are highly variable and are controlled by the different mineralogies. This REE variability is controlled by the diopside abundance, with vein ETA71 (the richest in diopside) having the highest REE values. On the other hand, Ti variability depends on modal titanian-clinohumite: Ti is, in fact, very low in veins having trace amounts of this mineral (e.g. ETA71, ETF10). Despite this heterogeneity in absolute concentrations, the trace-element patterns of veins are strikingly similar to those of their wall- and host-rocks (Fig. 3). Similar to REE, Sr is much more concentrated in veins than in peridotites (except for sample ETF10; Fig. 4) and is broadly correlated with the diopside abundance.
|
|
|
MINERAL COMPOSITIONS |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL AND PETROLOGICAL...SAMPLE DESCRIPTION AND...BULK-ROCK COMPOSITIONSMINERAL COMPOSITIONSDISCUSSIONSUMMARYREFERENCES |
|---|
The major-element compositions of mantle and alpine minerals have been documented in previous studies (Hoogerduijn Strating et al., 1993
; Scambelluri et al., 1997). In this study major and trace-element analyses mainly concern clinopyroxene, the most effective repository for trace elements in ultramafites. Analyses were performed on (1) mantle clinopyroxene relics in profile A and in wall- and host-rocks ETA74 and ETA75A, and (2) high-pressure diopside. Their compositions are reported in Tables 3 and 4, together with some representative analyses of spinel- and plagioclase-facies Erro–Tobbio clinopyroxenes. Major and trace elements were also analysed in high-pressure titanian-clinohumite (Table 5) coexisting with diopside, to assess its potential role as HFSE carrier.
|
|
|
Clinopyroxenes
The Erro–Tobbio mantle clinopyroxenes display typical decrease in Al and Na, and increase in Ti and Cr from spinel- to plagioclase-facies clinopyroxenes (Table 3; Hoogerduijn Strating et al., 1993). Concerning the trace-element compositions, the shaded fields of Fig. 5a represent typical clinopyroxene compositions for Erro–Tobbio spinel- and plagioclase-facies peridotites (Table 4). As a whole, these clinopyroxenes have similar REE patterns (CeN/YbN = 0·0704–0·1168 and 0·0826–0·1265, respectively) and share negative Sr, Zr and Ti anomalies. The plagioclase-facies clinopyroxenes are characterized by higher middle REE (MREE) to HREE concentrations, higher Zr (13–24 ppm), Y, V (205–287 ppm), Sc (53–78 ppm) and Ti (2491–4158 ppm), and display more pronounced Sr depletion (0·91–1·79 ppm). Comparable trace-element variations were observed in clinopyroxenes from ophiolitic peridotites and ascribed to interphase redistribution during metamorphic transition from spinel- to plagioclase-facies assemblages, and clinopyroxene re-equilibration with plagioclase (Rampone et al., 1993). The trace-element compositions of relict mantle clinopyroxenes preserved in profile A, as well as in host- and wall-rocks (Table 4; Fig. 5a), have the same variability as mantle clinopyroxenes defining the shaded fields in Fig. 5a.
|
The high-pressure diopside displays significantly different compositions. Compared with mantle clinopyroxenes, diopside has very low Na, Al, Ti and Cr, and higher Si and Mg values. Concerning trace elements (Fig. 5b), the REE and Sc concentrations in diopside are similar to those of mantle clinopyroxene, although diopside displays a slight HREE depletion (DyN/YbN = 1·4–1·8). The most striking difference between metamorphic diopside and mantle clinopyroxenes is the extreme HFSE depletion (Ti 11–618 ppm; Zr 0·2–1·8 ppm) and Sr enrichment (13–119 ppm) of diopside.
Titanian-clinohumite
Titanian-clinohumite (Table 5) has 3·4–4·9 wt % TiO2 and F is absent. Abundances of most REE (not reported in Table 5) are below the detection limits, with the exception of Yb (0·07–0·35 ppm; Fig. 6a). Titanian-clinohumite also incorporates appreciable amounts of other HFSE, such as Nb (0·2–1·2 ppm) and Zr (1·05–12·2 ppm). Figure 6a compares the Erro–Tobbio titanian-clinohumites with those from Val Malenco [Central Alps; data after Weiss (1997)], formed at much lower pressures (0·4–0·7 GPa): this figure indicates that titanian-clinohumite is a major repository of HFSE in meta-peridotites and that the HFSE partitioning is not pressure dependent.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL AND PETROLOGICAL...SAMPLE DESCRIPTION AND...BULK-ROCK COMPOSITIONSMINERAL COMPOSITIONSDISCUSSIONSUMMARYREFERENCES |
|---|
Compositional variations during serpentinization.
Profiles A and B (Fig. 1) include (1) mantle peridotites with variable oceanic serpentinization, (2) peridotites with static high-pressure recrystallization, and (3) high-pressure serpentinite mylonites. Their analysis thus enables us to discuss the compositional changes that occurredEZ[4]occurred during oceanic alteration and subduction of the peridotites. In spite of their variable serpentinization or high-pressure overprint, the analysed samples show overlapping REE patterns, thus proving REE immobility during the overall history. Also, peridotites affected by oceanic low-grade alteration display considerable water increase and ubiquitous Sr enrichment. Compared with unaltered Northern Apennine depleted peridotites (Rampone et al., 1996
), the Erro–Tobbio serpentinized peridotites have significantly higher Sr (up to 5 ppm compared with 0·08–0·15 ppm Sr; Fig. 7). Moreover, oceanic serpentine and phyllosilicates occurring in these rocks can contain appreciable chlorine and alkali contents (Scambelluri et al., 1997). Comparable REE behaviour and Sr enrichment were observed by Menzies et al. (1993) after peridotite–seawater experiments at 300°C. According to Menzies et al. (1993) serpentinization of precursor lherzolite with Sr < 8 ppm brings about an increase of bulk Sr to levels of 20–60 ppm. Comparable Sr concentrations were reported for abyssal serpentinites (Bonatti et al., 1970). The Sr enrichment of the Erro–Tobbio serpentinized peridotites therefore records an interaction with seawater-derived solutions during oceanic alteration. At this stage Sr was probably incorporated in Ca-rich phases (carbonate, tremolite, epidote and diopside) replacing mantle plagioclase and clinopyroxene. The undeformed Erro–Tobbio high-pressure metaperidotites display Sr contents (2–6 ppm) similar to those of the serpentinized peridotites with no high-pressure imprint (Fig. 7): they could represent formerly altered peridotites that preserved an oceanic Sr imprint during static eclogitization. Transition to serpentinite mylonites is accompanied by further increase in water and Sr (10–15 ppm), and two possible explanations can be given for this enrichment: (1) it is an oceanic inheritance, related to high water/rock ratios causing higher Sr uptake in the strongly serpentinized horizons; (2) it is an effect of Sr redistribution during high-pressure metamorphism, caused by channelling of Sr-bearing fluids into the shear zones. Although there is no definitive argument against the second hypothesis, several lines of evidence suggest that the Sr came primarily from an oceanic source. On the basis of the oxygen isotope analysis of the same rock samples as studied in this work, Vallis (1997) and Frueh-Green et al. (2001) demonstrated that high-pressure metaperidotites and serpentinite mylonites display widespread oxygen isotope heterogeneities reflecting oceanic pre-subduction signatures. Comparable heterogeneities are common in eclogites (Philippot et al., 1998) and suggest that the high-pressure recrystallization was not associated with isotopic re-equilibration caused by large-scale fluid fluxing. Fluid redistribution was probably limited to centimetre to metre scales. Also, the Sr values of our samples are well within the ranges of Sr concentrations measured in abyssal serpentinites and obtained experimentally. Consequently, the Sr and water variability of Fig. 7 probably reflects pre-subduction heterogeneities in serpentinization and in water distribution within the oceanic lithosphere.
|
Our study thus indicates that the main changes during serpentinization consist of uptake of trace amounts of strontium, chlorine and alkalis, associated with significant addition of water. This exerted an important control on the composition of fluids produced during subsequent subduction and high-pressure recrystallization of the Erro–Tobbio peridotite.
The high-pressure recrystallization
Petrography and compositions of the Erro–Tobbio high-pressure ultramafites and veins provide constraints on the mobility and composition of fluids released during partial deserpentinization at eclogite facies. Veins and host- and wall-rocks have the same high-pressure mineralogy, indicating control by the country peridotite on the components dissolved in the fluid. Similarity between the REE patterns of clinopyroxenes in the veins and in the host-rock also reflects such a rock-control on the fluid. In particular, the bulk REE concentrations in veins are dominated by diopside, whose REE absolute concentrations and patterns are, in turn, very similar to those of rock-forming mantle clinopyroxene relics and high-pressure diopside in peridotites. These comparable geochemical signatures in clinopyroxene suggest a lack of infiltration of external fluids transporting exotic components during the high-pressure metamorphism. Rather, internally produced fluids equilibrated with, and were compositionally controlled by, the surrounding ultramafites. Further indication of local derivation of rock components in the fluid is the occurrence of Sr-bearing diopside in the veins: this also represents compelling evidence for cycling of marine Sr in the high-pressure fluid. These data, together with previous recognition that pre-subduction chlorine and alkalis were released in the dehydration fluid (Scambelluri et al., 1997), indicate closed-system behaviour of the Erro–Tobbio ultramafites during eclogitization and internal cycling of oceanic water and components. The data presented here and the stable isotope heterogeneities diffusely documented in high-pressure and very high pressure rocks [Philippot et al. (1998), and reference cited therein), point to restricted fluid mobility during eclogitization. This contradicts the large-scale mobility required to flux slab fluids into the mantle, and suggests that fluids may remain entrapped in the slab, until they are injected into the upper mantle at depths greater than those attained by most exposed eclogites. The latter may not represent the levels at which fluids become extracted from the slab and higher-pressure rocks need to be exhumed (or discovered) to find the regions of important fluid loss from the slab. Watson et al. (1990) proposed that transition from immobile to mobile fluids in the mantle is related to concurrent deformation, and to decrease of wetting angles below a critical value of 60°, which occurs at higher pressure and temperature. This points to a depth-related change in fluid mobility with continuing subduction. Alternatively, recent experiments of Watson & Wark (1997) indicate that SiO2 diffusion into an immobile fluid can occur over kilometre-scale distances at rates comparable with those of thermal conductivity. These results have been taken by Philippot & Rumble (2000) to explain both limited fluid circulation and mantle metasomatism in deep subduction zones. Whether or not the eclogitic fluids may flux the overlying mantle is thus matter of debate and remains an open problem.
Ti-clinohumite–diopside partition coefficients (Fig. 6b) indicate Sr partitioning into the high-pressure pyroxene and the affinity of titanian-clinohumite for HFSE (mainly Ti and Zr); their crystallization was thus accompanied by incorporation of Sr and HFSE into the respective mineral phases. Sr was originally held in Ca-rich oceanic phases. Negative HFSE in the alpine diopside, despite the lack of negative HFSE anomaly in the whole rocks, indicates that these elements, primarily hosted in mantle clinopyroxene that survived subduction, were released during its breakdown in presence of fluid and partitioned in titanian-clinohumite. This mineral thus plays a role in fractionating HFSE from LILE at high pressures, similarly to other Ti phases such as rutile and ilmenite (Brenan et al., 1995b; Ayers et al., 1997). Capability of titanian-clinohumite to store HFSE was firstly envisaged by Weiss & Muentener (1996) for the low-pressure titanian-clinohumites from Val Malenco peridotite (Central Alps); our study confirms this hypothesis and shows that clinohumite acts as a repository for HFSE independent of pressure conditions.
The vein mineralogies indicate that the dehydration fluid dissolved appreciable amounts of silicate components from surrounding rocks, and probably corresponded to an aqueous solute-rich solution. Presence of diopside, titanian-clinohumite and hypersaline fluid inclusions in the veins indicates that Sr, HFSE, Cl and alkalis were also soluble in this fluid. Precipitation of vein minerals consumed silicate components from the fluid, and left a residual solution enriched in excess water, which became trapped in the primary fluid inclusions inside vein minerals (Scambelluri et al., 1997). Crystallization of diopside and titanian-clinohumite thus fixed Sr and HFSE, respectively, and controlled the trace-element composition of the coexisting aqueous residue. High Cl and alkali concentrations in the primary hypersaline fluid inclusion indicate partitioning of these large ions in the fluid residue, rather than in the compact structures of vein minerals (Scambelluri et al., 1997). Additional information concerning the trace-element composition of such a fluid can be achieved using experimental clinopyroxene–fluid partition coefficients, which are available for a range of P–T conditions. Partitioning data determined at 2 GPa and 900°C (Ayers et al., 1997; Brenan et al., 1995a, 1995b), the conditions closest to P–T estimates of the Erro–Tobbio high-pressure recrystallization, indicate that most trace elements preferentially enter the clinopyroxene. The coexisting aqueous fluid has extremely low REE and HFSE contents in the range (0·01–0·1) x chondrite abundances. Sr is also low—although selectively enriched relative to LREE—and ranges from (0·2–0·4) x chondrites (Brenan et al., 1995a) to (0·8–1·6) x chondrites (Ayers et al., 1997). On the other hand, the fluid is strongly enriched in Cl and alkaline elements (Scambelluri et al., 1997).
|
SUMMARY |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL AND PETROLOGICAL...SAMPLE DESCRIPTION AND...BULK-ROCK COMPOSITIONSMINERAL COMPOSITIONSDISCUSSIONSUMMARYREFERENCES |
|---|
Our study of the alpine Erro–Tobbio eclogitized peridotite indicates that pre-subduction serpentinization and alteration had a relevant control on the composition of fluids evolved during later burial and high-pressure recrystallization. Serpentinization was accompanied by REE immobility, uptake of trace amounts of Cl, Sr and alkalis, and by significant addition of water. Formation of a high-pressure paragenesis made up of olivine, antigorite, diopside and Ti-clinohumite in country ultramafites and veins was the result of one major pulse of fluid released during subduction. The close similarity in the REE compositions of mantle clinopyroxenes and high-pressure diopside indicates rock control on the vein fluids and absence of exotic components carried by externally derived fluids. Presence of appreciable Sr contents in vein-forming diopside indicates cycling of oceanic Sr in the high-pressure fluid. This, together with the presence of pre-subduction Cl and alkalis in brines trapped as inclusions in vein minerals, indicates internal cycling of exogenic components into the fluid phase.
Diopside and Ti-clinohumite are the high-pressure minerals acting as the most effective trace-element repositories: Ti-clinohumite hosts HFSE, diopside mainly incorporates REE and Sr, and lacks HFSE. The brine inclusions hosted in vein diopside and Ti-clinohumite represent a fluid residue after mineral deposition: they correspond to the aqueous fluid phase that equilibrated with such an assemblage and that was eventually evolved at this stage of deep peridotite recrystallization. On the basis of our trace-element analysis and clinopyroxene–fluid partition coefficients available in the literature (Brenan et al., 1995a; Ayers et al., 1997), this fluid has very low REE and HFSE content, is selectively enriched in Sr relative to LREE, and is Cl and alkali rich.
|
ACKNOWLEDGEMENTS |
|---|
We are grateful to Alberto Zanetti and Riccardo Vannucci (Pavia) for ion probe analyses, Jean-Louis Bodinier (Montpellier) for ICP-MS analyses, and Danilo Biondelli (Milan) and Laura Negretti (Genoa) for assistance during electron microprobe and scanning electron microscope analyses. We thank Jeffrey Alt, Marguerite Godard and Ian Parkinson for careful and constructive reviews. Special thanks go to Jean-Louis Bodinier for helpful comments and for the editorial handling of this paper. This work was funded by the Italian MURST within the project ‘Transformations in subducted materials and mass transfer to the mantle wedge’.
|
FOOTNOTES |
|---|
*Corresponding author. Telephone: +39-010-3538315. Fax: +39-010-352169. E-mail: msca{at}dipteris.unige.it
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONGEOLOGICAL AND PETROLOGICAL...SAMPLE DESCRIPTION AND...BULK-ROCK COMPOSITIONSMINERAL COMPOSITIONSDISCUSSIONSUMMARYREFERENCES |
|---|
Anders, E. & Ebihara, M. (1982). Solar system abundances of the elements. Geochimica et Cosmochimica Acta 46, 2363–2380.
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]
Bonatti, E., Honnorez, J. & Ferrara, G. (1970). Equatorial Mid-Atlantic Ridge: petrologic and Sr isotopic evidence for alpine-type rock assemblage. Earth and Planetary Science Letters 9, 247–256.
Bottazzi, P., Ottolini, L. & Vannucci, R. (1991). Determination of rare earth elements in sixteen silicate reference samples by secondary ion mass spectrometry using conventional energy filtering technique. Geostandards Newsletters 15(1), 1.
Brenan, J. M., Shaw, H. F. & Ryerson, F. J. (1995a). Experimental evidence for the origin of lead enrichment in convergent-margin magmas. Nature 380, 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]
Frueh-Green, G. l., Scambelluri, M. & Vallis, F. (2001). O–H isotope ratios of high-pressure ultramafia rocks: implications for fluid sources and mobility in the subducted hydrous mantle. Contributions to Mineralogy Petrology (in press).
Hofmann, A. W. (1988). Chemical differentiation of the Earth: the relationships between mantle, continental crust, and oceanic crust. Earth and Planetary Science Letters 90, 297–314.[Web of Science]
Hoogerduijn Strating, E. H., Rampone, E., Piccardo, G. B., Drury, M. R. & Vissers, R. L. M. (1993). Subsolidus emplacement of mantle peridotites during incipient oceanic rifting and opening of the Mesozoic Tethys (Voltri Massif, NW Italy). Journal of Petrology 34, 901–927.
Ionov, D. A., Savoyant, L. & Dupuy, C. (1992). Application of the ICP-MS technique to trace element analysis of peridotites and their minerals. Geostandards Newsletters 16(2), 2.
Menzies, M. A., Long, A., Ingram, G., Tatnel, M. & Janecky, D. (1993). MORB–peridotite–sea water interaction: experimental constraints on the behaviour of trace elements, 87Sr/86Sr and 143Nd/144Nd ratios. In: Prichard, H. M., Alabaster, T., Harris, N. B. W. & Neary, C. R. (eds) Magmatic Processes and Plate Tectonics. Geological Society, London, Special Publications 76, 309–322.
Messiga, B., Scambelluri, M. & Piccardo, G. B. (1995). Formation and breakdown of chloritoid–omphacite high-pressure assemblages in mafic systems: evidence from the Erro–Tobbio eclogitic metagabbros (Ligurian Western Alps). European Journal of Mineralogy 7, 1149–1167.[Web of Science]
Philippot, P. & Rumble, D. III (2000). Fluid–rock interactions during high pressure and ultrahigh pressure metamorphism. International Geology Review 42, 312–327.[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., Agrinier, P. & Scambelluri, M. (1998). Chlorine cycling in the subducted oceanic lithosphere. Earth and Planetary Science Letters 161, 33–44.[Web of Science]
Rampone, E., Piccardo, G. B., Vannucci, R., Bottazzi, P. & Ottolini, L. (1993). Subsolidus reactions monitored by trace element partitioning: the spinel- to plagioclase-facies transition in mantle peridotites. Contributions to Mineralogy and Petrology 115, 1–17.[Web of Science]
Rampone, E., Hofmann, A. W., Piccardo, G. B., Vannucci, R., Bottazzi, P. & Ottolini, L. (1996). Trace element and isotope geochemistry of depleted peridotites from an N-MORB type ophiolite (Internal Liguride, N Italy). Contributions to Mineralogy and Petrology 123, 61–76.
Scambelluri, M. & Philippot, P. (2000). Deep fluids in subduction zones. In: Andersen, T., Burke, E. A. J., Diamond, L. & Frezzotti, M. L. (eds) Fluid Inclusions: Phase Systematics Methods—Applications. Lithos 780, in press.
Scambelluri, M., Muentener, O., Hermann, J., Piccardo, G. B. & Trommsdorff, V. (1995). Subduction of water into the mantle: history of an alpine peridotite. Geology 23, 459–462.
Scambelluri, M., Piccardo, G. B., Philippot, P., Robbiano, A. & Negretti, L. (1997). High salinity fluid inclusions formed from recycled seawater in deeply subducted alpine serpentinite. Earth and Planetary Science Letters 148, 485–500.
Stalder, R., Foley, S. F., Brey, G. P. & Horn, I. (1998). Mineral–aqueous fluid partitioning of trace elements at 900–1200°C and 3·0–5·7 GPa: new experimental data for garnet, clinopyroxene, and rutile, and implications for mantle metasomatism. Geochimica et Cosmochimica Acta 62, 1781–1801.[Web of Science]
Tera, F., Brown, L., Morris, J., Sacks, I. S., Klein, J. & Middleton, R. (1986). Sediment incorporation in island-arc magmas: inferences from 10Be. Geochimica et Cosmochimica Acta 50, 535–550.
Vallis, F. (1997). Studio degli isotopi stabili (H–O) nelle rocce ultrafemiche di ambiente eclogitico dell’Unità Erro–Tobbio (Alpi Occidentali): implicazioni sul rilascio di fluidi durante la subduzione. Ph.D. thesis, University of Genoa.
Ulmer, P. & Trommsdorff, V. (1995). Serpentine stability to mantle depths and subduction-related magmatism. Science 268, 858–861.
Watson, E. B. & Wark, D. A. (1997). Diffusion of dissolved SiO2 in H2O at 1 GPa, with implications for mass transport in the crust and upper mantle. Contributions to Mineralogy and Petrology 130, 66–80.
Watson, E. B., Brenan, J. M. & Baker, D. R. (1993). Distribution of fluids in the continental mantle. In: Menzies, M. A. (ed.) Continental Mantle. Oxford: Clarendon Press, pp. 111–125.
Weiss, M. (1997). Clinohumites: a field and experimental study. Ph.D. thesis, ETH-Zurich.
Weiss, M. & Muentener, O. (1996). Crystal chemistry of titanian-clinohumite: implications for storage of HFSE in the mantle. Sixth Goldschmidt Conference. Journal of Conference Abstracts 1, 665.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  N. A. Stroncik and K. M. Haase Chlorine in oceanic intraplate basalts: Constraints on mantle sources and recycling processes Geology, November 1, 2004; 32(11): 945 - 948. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Tenthorey and J. Hermann Composition of fluids during serpentinite breakdown in subduction zones: Evidence for limited boron mobility Geology, October 1, 2004; 32(10): 865 - 868. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||