Journal of Petrology Advance Access published online on November 14, 2008
Journal of Petrology, doi:10.1093/petrology/egn053
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Formation of Highly Refractory Dunite by Focused Percolation of Pyroxenite-Derived Melt in the Balmuccia Peridotite Massif (Italy)
1Dipartimento DI Scienze Della Terra, Università di Modena E Reggio Emilia, p.Le S. Eufemia 19, I-41100 Modena, Italy
2Cnr–Istituto Di Geoscienze E Georisorse, Sezione Di Pavia, Via Ferrata 1, I-27100 Pavia, Italy
3Dipartimento Di Scienze Della Terra, Università Degli Studi Di Firenze, Via La Pira 4, I-50121 Firenze, Italy
4Cnr–Istituto Di Scienze Marine, Sezione DI Bologna, Via Gobetti 101, I-40129 Bologna, Italy
Received February 29, 2008; Revised typescript accepted October 3, 2008
 |
ABSTRACT |
|---|
A 50 m thick and 150 m long dunite body occurs as a subconcordant, tabular structure in the Balmuccia Massif, an Alpine peridotite thought to represent part of the subcontinental mantle. The contacts with the host spinel-facies depleted lherzolite are sharp. The dunite body is composed of spinel-rich dunite containing centimetre-size lenses of relict Cr-diopside websterite, spinel-poor granoblastic dunite and virtually monomineralic Cr-spinel layers exhibiting flow structures. Orthopyroxene is a minor, relict phase in all the lithologies; clinopyroxene is intergranular and amphibole is a minor accessory phase. Overall the dunite body is fairly refractory (Fo in olivine: 90·7–93·8). Strontium and neodymium isotope ratios of clinopyroxene separates from the dunitic body resemble those of a Cr-diopside websterite suite that forms a series of dykes cutting the main peridotite host. It is proposed that the dunites were generated in a part of the mantle veined by early Cr-diopside websterites by a three-stage process involving partial melting of pyroxenite, infiltration of the pyroxenite-derived melt into the depleted lherzolite and its consequent open-system partial melting and focused flow of the resultant partial melts leading to the production of reactive dunite channels through both peridotite and pyroxenite. This process has been simulated using pMELTS assuming that the pyroxenite partially melts at 1·5 GPa and focused melt transport occurs at pressures greater than 0·7 GPa. The results show that, depending on the focusing factor assumed, dunite can form from peridotite at P < 1·2 GPa and from pyroxenite at P < 1·1 GPa, in both cases over a large pressure range. The model accounts for specific characteristics of the dunite, such as its refractory composition, the presence of orthopyroxene relics, the occurrence of relict websterite lenses in the spinel-rich dunites and the flow structures in the Cr-spinel layers. The proposed mechanism allows dunite formation to occur well within the spinel stability field, and therefore at greater depth than dunites in ophiolites, which generally formed within the plagioclase stability field. The aggregated model melts extracted from the segments where dunite forms are high-Mg alkali basalts resembling, after olivine fractionation, the compositions of enriched-type mid-ocean ridge basalt from slow- and ultraslow-spreading ocean ridges.
KEY WORDS: Balmuccia; dunite; focused flow; pyroxenite melting; subcontinental mantle
|
INTRODUCTION |
|---|
Tabular dunite bodies occurring in ophiolites are held to be the result of the reaction between the ambient mantle peridotite and melts flowing in high-porosity channels (Bodinier, 1988
; Kelemen & Dick, 1995; Suhr, 1999; Morgan & Liang, 2005). In the mantle system the olivine primary field expands at the expense of pyroxenes at decreasing pressure (O’Hara, 1968; Stolper, 1980). Therefore, dunites are the expected result of the reaction between melts infiltrating mantle levels shallower than those of the melt source (Kelemen et al., 1997; Morgan & Liang, 2005). This is the reason why most ophiolitic dunites occur at relatively shallow (plagioclase-facies) mantle levels. These dunites are generally equilibrated with mid-ocean ridge basalts (MORB) consistent with their being conduits for MORB magma transport (Kelemen et al., 1995, 1997; Piccardo et al., 2007). Other dunites, however, exhibit geochemical features consistent with magma types other than MORB (Kelemen et al., 1997, p. 293), such as alkali basalts, possibly related to ocean island basalts (OIB), or highly refractory boninitic melts, possibly representative of subduction-related magmas (e.g. the highly refractory dunites in the Bay of Islands ophiolite; Suhr et al., 2003).
In this study, we examine anhydrous, tabular dunites and associated Cr-spinel layers occurring in the Balmuccia Peridotitic Massif (Western Italian Alps). On the basis of the field relationships, petrology and modelling, we propose that these dunites were generated from the interaction of an ambient peridotite–pyroxenite sequence with melts derived from a deeper peridotite source that had been previously fertilized by melt migration and the formation of pyroxenite veins. Various workers have interpreted certain characteristics of MORB and OIB geochemistry as a signature of pyroxenite-derived melts (e.g. Hirschmann & Stolper, 1996; Lundstrom et al., 2000; Hirschmann et al., 2003; Keshav et al., 2004; Seyler et al., 2007). The composition of these melts has been constrained by a number of experimental studies (Kogiso & Hirschmann, 2001; Hirshmann et al., 2003; Kogiso et al., 2003, 2004; Pertermann & Hirschmann, 2003a, 2003b; Keshav et al., 2004). However, as far as we know, natural examples of direct interaction between pyroxenite melts and peridotite are lacking and the results of this study may bridge a gap between the experimental results and the specific characteristics of those mantle-derived melts attributed to pyroxenite influence.
|
THE BALMUCCIA PERIDOTITE MASSIF AND ITS DUNITES |
|---|
 TOPABSTRACTINTRODUCTION THE BALMUCCIA PERIDOTITE MASSIF... FIELD CHARACTERISTICSPETROGRAPHYANALYTICAL METHODSANALYTICAL RESULTSTRACE ELEMENTS IN CLINOPYROXENENd-Sr ISOTOPESDISCUSSIONCONCLUSIONSSUPPLEMENTARY DATAREFERENCES |
|---|
Balmuccia is one of the subcontinental mantle peridotite bodies occurring in the Southern Domain of the Western Italian Alps (Lensch, 1971
; Rivalenti et al., 1975; Ernst, 1978). It consists of a 4·5 km long, 0·8 km wide lens crossing the Val Sesia just east of the Insubric line (Fig. 1). The massif is associated with the mafic–ultramafic Ivrea–Verbano complex, which shows an igneous contact at its eastern margin and a tectonic one at its western margin, where septa of highly depleted metasediments of the Kinzigitic Formation also appear (Rivalenti et al., 1975). The Balmuccia peridotite has a dominant NNE–SSW-trending foliation determined by flattening and elongation of olivine and pyroxenes and by preferential concentrations of minerals (Rivalenti et al., 1975; Shervais, 1979). In places, this foliation evolves to a porphyroclastic texture with a secondary foliation parallel to the peridotite contact at its eastern margin.
|
|
|
|
The dominant peridotite type is lherzolite. The lherzolites are discordantly cut by abundant websterite dykes, distinguished on the basis of the clinopyroxene composition into an older Cr-diopside suite (websterites, ol-websterites and orthopyroxenites) and a later Al-augite suite (websterites and late gabbros; Shervais, 1979
). Both dyke suites have sharp contacts against the peridotite. The Al-augite dykes may exhibit a thin, pyroxenitic reaction rim, whereas, occasionally, centimetre- to decimetre-wide dunitic bands occur at the contact between the Cr-diopside dykes and the host rock. Both the Cr-diopside and Al-augite suites comprise several dyke generations, distinguished on the basis of their mutual intersections, deformation patterns and composition (Capedri et al., 1977; Comin-Chiaramonti et al., 1982; Sinigoi et al., 1983; Shervais & Mukasa, 1991). These dykes therefore represent a prolonged period of melt intrusion, possibly related to the accretion of the peridotite from the asthenosphere to the lithosphere and, finally, to the Permian intrusion of the mafic–ultramafic Ivrea–Verbano complex, when the peridotite body had already been incorporated in the deep crust (Peressini et al., 2007).
The lherzolites have experienced moderate degrees of melt extraction. None of the websterites appear to be representative of a melt; rather, they are interpreted as mineral segregations from melts flowing into the peridotite (Sinigoi et al., 1983; Ottonello et al., 1984; Shervais & Mukasa, 1991; Mukasa & Shervais, 1999).
Reactive dunites occurring within the Balmuccia Massif have been studied previously by Rivalenti et al. (1995) and Mukasa & Shervais (1999). These dunites never exceed 40 cm in thickness and are systematically in contact with the websterite dykes. Those workers concluded that the dunites originated by pyroxene-dissolving reactions caused by melt infiltration into the wall-rock from the dyke conduit.
We have recently found two other dunite types of much larger dimensions: (1) anhydrous dunites, 10 cm to 40–50 m thick, characterized by the presence of discontinuous layers of Cr-rich spinel; (2) dunite lenses 15–20 m thick and up to 60 m long containing pods and veins rich in amphibole, phlogopite, plagioclase and rutile. As these two dunite occurrences, in addition to their lithological association, also have marked geochemical and petrological differences, we discuss here only the petrochemistry of the anhydrous dunites; the hydrous lithologies will be discussed elsewhere.
|
FIELD CHARACTERISTICS |
|---|
The anhydrous dunites appear as lenses or tabular bodies concordant with the peridotite foliation. The thinnest ones (
10 cm) contain Cr-rich spinel trails at their centre. To date, the largest body found is 50 m thick, can be followed along strike over a distance of about 150 m, and is the focus of this study. This dunite is cut by the Sesia River and its location is shown in Fig. 1. Sampling locations and internal structures are illustrated in Figs 2 and 3
, respectively. The contact between the dunite and the dominant lherzolite rock type is sharp. The internal structure of the dunite body consists of an association of various rock types occurring in irregular, but persistent bands. The different bands may fade laterally into each other or have sharp contacts. The lithological types constituting the bands are as follows.
(1) Massive spinel layers. One main, almost monomineralic layer, 10 cm thick, occurs in the central part of the dunite (Fig. 3a) and can be followed over a distance of about 30 m. Along strike, this layer fades out into a spinel-rich dunite. It exhibits flow structures (Fig. 3b), laterally grading into dunite through a thin region where spinel streaks alternate with lensoidal bodies of dunite rimmed by spinel trails.
(2) Spinel-rich dunites containing Cr-diopside websterite lenses. These dunites contain the massive spinel layers described above. They are pervasively, but randomly (Fig. 3b) spinel-enriched with respect to the ‘normal’ dunite described below. Cr-diopside websterite lenses, generally not exceeding 10 cm in length and 5 cm in width, are always present but constitute less than 1% of this dunite type (Fig. 3c). These lenses are elongated in the foliation plane defined by the spinel bands and the lithological contacts between the various rock types.
(3) ‘Normal’ dunites. These are intercalated with the spinel-rich ones and are distinguished by their low modal spinel and absence of websterite lenses. They are granular rocks in which foliation is nearly absent.
(4) Cr-diopside websterite dykes. Two dykes, 10 cm thick, discordantly cut the dunites and their structures and thus were clearly emplaced after the formation of the dunite. The contacts between the dunites and Cr-diopside dykes are sharp.
A number of ‘ghost’ Cr-diopside websterite dykes were observed within the dunite near the southern margin of the body. These dykes are 1 cm thick, folded, and cut the dunite foliation at a high angle. The original clinopyroxene has largely been replaced by spinel and olivine (Fig. 3d).
|
PETROGRAPHY |
|---|
Estimated modal proportions are reported in Table 1. These values are only indicative in view of the fine-scale modal variations illustrated in Fig. 3.
The two dunite types (‘normal’ and spinel-rich containing Cr-diopside websteritic lenses) have a granoblastic texture, often equigranular, with 120° triple junctions between the olivine crystals and rectilinear or slightly curved grain boundaries (Fig. 4a). Kink bands in olivine are common. Olivine may contain grains of euhedral spinel and relic orthopyroxene (Fig. 4b and c). A granoblastic rim also separates the Cr-diopside lenses from the dunite. In general, clinopyroxene occurs as an interstitial phase at grain boundaries or triple junctions. Normally, it is a clean, unstrained phase rarely showing exsolution lamellae. Clinopyroxene may enclose euhedral olivine and spinel. The normal dunites have low modal clinopyroxene (<3 vol.%) and spinel (<5 vol.%), whereas in the dunites associated with Cr-diopside websteritic lenses the modal clinopyroxene may be as high as 12% and spinel 36%. Orthopyroxene is rare, usually highly deformed, showing abundant exsolution lamellae and may occur in olivine (Fig. 4c). When present as an intergranular phase, it has lobate contacts against the surrounding crystals. Spinel is euhedral or subhedral, located at the grain boundaries and triple junctions. In some cases it contains inclusions of olivine or clinopyroxene. Amphibole and, in one case, phlogopite (BC6) are very minor accessory interstitial phases. Many samples contain veins of fine-grained neoblasts (olivine and minor clinopyroxene ± amphibole) at the grain boundaries (Fig. 4d).
|
The Cr-diopside websteritic lenses are dominantly composed of clinopyroxene, less abundant orthopyroxene, and minor olivine and spinel. Both pyroxenes are highly deformed and have abundant exsolution lamellae.
The Cr-spinel layers have, in addition to high modal spinel and olivine, relatively high modal interstitial clinopyroxene (up to 11%). Clinopyroxene, and more rarely orthopyroxene, may also occur as inclusions in spinel.
The websterite dykes cutting the dunites are Cr-diopside rich, like those described by Rivalenti et al. (1995) and Mukasa & Shervais (1999). Orthopyroxenes are highly exsolved. Olivine and spinel occur in variable amounts.
|
ANALYTICAL METHODS |
|---|
Bulk-rock analyses were carried out according to methods described by Rivalenti et al. (2004
) at the Dipartimento di Scienze della Terra, University of Modena and Reggio Emilia. Trace elements in clinopyroxene and amphibole were analysed by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) at the Istituto di Geoscienze e Georisorse-CNR, Unit of Pavia, consisting of an Elan DRC-e mass spectrometer coupled with a Q-switched Nd:YAG laser source (Quantel Brilliant). The fundamental emission of this laser (1064 nm) was converted to 266 nm by two harmonic generators. Helium was used as carrier gas, mixed with Ar downstream of the ablation cell. The spot diameter was 50–60 µm. NIST SRM 610 glass was used as an external standard, with 44Ca as the internal standard. Routine analyses consisted of 1 min background acquisition and 1 min of sample ablation. Precision and accuracy, both better than 10% for concentrations at the ppm level, were assessed from repeated analyses of NIST SRM 612 and BCR-2g standards.
Electron microprobe analyses of the mineral phases were performed using the automated Cameca-Camebax SX100 electron microprobe of the CAMPARIS microanalysis centre (Unversity of Paris VI). Analytical conditions are 15 kV acceleration voltage, 15 nA beam current and focused beam for 10–20 s counting times. Detection limits are within 0·01–0·03 wt %. Matrix correction was carried out using methods given by Bence & Albee (1968).
Sr and Nd isotope analyses were performed at the Department of Earth Sciences, University of Firenze following the procedure outlined by Avanzinelli et al. (2005). The samples were ground and clinopyroxene in the 0·1–0·2 mm size fraction was concentrated magnetically and was then further purified by handpicking using an optical binocular microscope. About 200–300 mg of each separated clinopyroxene fraction was first leached in warm 6N HCl in an ultrasonic bath for 1 h, then repeatedly rinsed with Milli-Q water and finally dissolved in a HF–HNO3–HCl mixture. Strontium and neodymium fractions were separated following standard chromatographic techniques using AG50x8 and Ln Spec resins with HCl as eluent. The total procedural blank was <300 pg for Sr and <150 pg for Nd, making blank correction negligible.
Mass spectrometric analyses were performed on a Thermo Finnigan Triton-Ti thermal ionization mass spectrometer equipped with nine movable collectors. Sr and Nd isotope compositions were measured in dynamic mode and are reported normalized to 86Sr/88Sr = 0·1194 and 146Nd/144Nd = 0·7219, respectively. Mass fractionation correction for both Sr and Nd used the exponential power law. Uncertainties in measured isotopic ratios refer to the least significant digits and represent ± 2
run precision. The external precision of NIST SRM987 was 87Sr/86Sr = 0·710251 ± 10 (2SD, n = 39), and that of the La Jolla standard was 143Nd/144Nd = 0·511845 ± 5 (2SD, n = 11).
|
ANALYTICAL RESULTS |
|---|
Bulk-rocks
Major element analyses are reported in Table 1 and shown in Fig. 5 in comparison with data for the Balmuccia lherzolites and Cr-diopside websterites from Rivalenti et al. (1995
). The ‘normal’ dunites cluster at MgO 49·0–50·5 wt % and have Mg-numbers [molar MgO x 100/(MgO + total FeO)] in the range 90·7–91·6. They have very low concentrations of CaO (<1 wt %), Al2O3 (<1 wt %) and TiO2 (<0·03 wt %) and relatively high Cr2O3 (up to 1·5 wt %). The compositions of the spinel-rich dunites containing the Cr-diopside lenses partially overlap with those of the ‘normal’ dunites, but extend the MgO range down to 41 wt % (Mg-number = 90·7). At low MgO concentrations, two compositional trends appear controlled by the olivine–spinel and olivine–clinopyroxene composition, respectively. When compared with the Balmuccia lherzolites, the ‘normal’ dunites and the spinel-rich dunites have higher MgO and NiO, lower Al2O3, TiO2 and CaO contents, comparable Cr2O3 concentration, and similar or slightly higher FeO. As Ni concentrations in olivine from both dunite and lherzolite are similar (see next section), the higher bulk-rock NiO content is a consequence of higher modal olivine in the dunites.
|
Mineral phases
Olivine
Olivine in the dunites (selected analyses are given in Table 2 and the complete dataset is given in Electronic Appendix 1, available for downloading at http://www.petrology.oxfordjournals.org/) is Fo 90·7–93·8 in composition and is more magnesian than that in lherzolites not in contact with websterites (Fo 88·3–90·8; Rivalenti et al., 1995
). It is also more magnesian than olivine in most ophiolitic dunites (Nicolas, 1989, pp. 227–232; Suhr et al., 1998; Piccardo et al., 2007). Most dunites have NiO concentrations in olivine in the same range as that of olivine in the lherzolites (NiO 0·20–0·53 wt %, Fig. 6). Olivine is slightly zoned, with rims about 1% Fo richer than cores. Olivine in the Cr-diopside lenses has a composition similar to that of the host dunite, whereas in the Cr-diopside websterite dykes the olivine composition is in the range of Fo 87·1–89·7 (Rivalenti et al., 1995).
|
|
Clinopyroxene
Clinopyroxene in the dunite body has the highest MgO concentrations (up to 17·4 wt %) and Mg-number (92·3–95·2) so far found in the Balmuccia Massif (Table 3, Electronic Appendix 2, and Rivalenti et al., 1995
). The major element variations with respect to Mg-number result in dispersed arrays, with TiO2, Al2O3, Cr2O3 and Na2O decreasing with increasing Mg-number (Fig. 7). The major cause of the observed scattering is the different composition of the clinopyroxene in the various lithotypes of the dunite transect. At a given Mg-number the clinopyroxenes of the massive spinel layers have higher Cr2O3, TiO2, Al2O3 and, with a few exceptions, Na2O than those of the other rock types. Clinopyroxenes of the Cr-diopside-bearing dunites plot between those of the massive spinel layers and of the ‘normal’ dunites. Importantly, the clinopyroxenes of the Cr-diopside lenses plot together with those of their host dunites and reach Mg-number values as high as 94·2, being, in this respect, different from the clinopyroxenes of the Cr-diopside dykes (Mg-number <92).
|
|
Lherzolite clinopyroxenes have on average lower Mg-number values (90·5–93·3) and are distinct from those of the dunite transect in having higher Al2O3 and Na2O, whereas TiO2 and Cr2O3 concentrations are comparable.
Another characteristic of the clinopyroxenes is their large compositional variability at the thin-section scale and their strong zoning (Electronic Appendix 2). This is particularly marked in the pyroxenes of the massive spinel regions. Zoning invariably consists of an increase of MgO and Mg-number, and decrease of all the other oxides, including Cr2O3, from core to rim. This zoning is probably caused by element diffusion between clinopyroxene and the contact spinel during sub-solidus re-equilibration.
Spinel
Spinel compositions are reported in Table 4 and Electronic Appendix 3. In an Mg-number–Cr-number diagram [Cr-number = atomic Cr x 100/(Cr + Al)] the ‘normal’ dunites define a negative variation trend where the low Cr-number end-member is represented by the spinels of the lherzolites (Fig. 8a). With respect to this array, the spinels of the massive spinel layers and of the spinel-rich dunites cluster at higher Cr-number values for a comparable Mg-number range. Spinels in the Cr-diopside websterites form a variation trend at lower Cr-number values with respect to the ‘normal’ dunites.
|
|
Spinels in the dunite transect are enriched in TiO2 with respect to those of the lherzolites, whose variation range is within the field of abyssal peridotites (Fig. 8b). TiO2 concentration increases from the spinel layers to the ‘normal’ dunites. Compared with spinel from dunites in other geological settings, these spinels have lower Cr-number and TiO2 than those from the Samail ophiolite in Oman, as well as the Atlantic and Pacific ophiolites. They are similar only to those of the dunites from the South West Indian Ridge (Bullard, Marie Celeste and Vulcan Fracture Zones, Kelemen et al., 1997
). Similar to clinopyroxene, spinel is zoned (Electronic Appendix 3), especially in the massive spinel layers. Rims are both enriched and depleted in the chromite component, depending on the surrounding silicates (Electronic Appendix 3). In particular, spinel rims are richer in Mg-number and lower in Cr-number at the contact with olivine, whereas the reverse happens at the contact with clinopyroxene.
Orthopyroxene
Orthopyroxene (Table 5) in the dunites and the Cr-spinel layers has much higher MgO concentrations (34·8–35·8 wt %) and Mg-number values (91·5–92·6) than that in the host lherzolites (MgO 32·4–34·1 wt %, Mg-number = 88·8–91·3, Rivalenti et al., 1995), similar to all the other silicate phases in the dunite. Compared with the lherzolites, orthopyroxene in the dunite is depleted in Al2O3 and FeO, enriched in TiO2 and has comparable Na2O, Cr2O3 and NiO concentrations. Orthopyroxene compositions correlate very poorly, or not at all, with those of the coexisting clinopyroxene, thus suggesting that they are not in chemical equilibrium, consistent with the textural relationships. The lack of textural and chemical equilibrium between clino- and orthopyroxene does not permit geothermometric estimates from these phases.
|
Amphibole
Amphibole (Table 6) is edenite in composition in the late Cr-diopside dykes and in the BC34 pyroxenite lens and pargasite in the dunites. Mg-number values range from 87·1 (Cr-diopside websterite dyke BC44) to 89·9 (dunite BC3). Because of its negligible amount and sporadic occurrence, amphibole will not be discussed further in the paper.
|
|
TRACE ELEMENTS IN CLINOPYROXENE |
|---|
Trace element concentrations in clinopyroxene are reported in Table 3 and illustrated in Fig. 9. Clinopyroxenes in the Cr-spinel layers and in the Cr-diopside-bearing dunites have rare earth element (REE) patterns depleted in light REE (LREE) (Lan/Ybn = 0·18–0·36) and almost flat from middle REE (MREE) to heavy REE (HREE) (Gdn/Ybn = 0·93–0·96, Fig. 9b; normalization to estimated Primitive Mantle (PM), Hofmann, 1988
). In normalized diagrams, they are also depleted in Nb, Zr, Hf and Ti, and are slightly enriched in Sr with respect to the adjacent REE (Fig. 9a).
|
The ‘normal’ dunites display two types of REE pattern: (1) LREE-depleted (Lan/Ybn = 0·17–0·27) with a linear trend (BC3, BC48, BC50, Fig. 9d); (2) spoon-shaped patterns (Lan/Ndn = 1·02–1·40; Gdn/Ybn = 0·45–0·56; BC14, BC17 and BC32; Fig. 9f). These clinopyroxenes are all depleted in Nb and Zr, weakly depleted in Hf and Ti, and enriched in Sr (Fig. 9c and e).
Similar to the clinopyroxenes of the Cr-spinel layers, those from the Cr-diopside lenses BC34 and BC45 are LREE depleted (Lan/Ybn = 0·23–0·26, Fig. 9h), and they exhibit a flat MREE to HREE pattern or slight HREE enrichment (Gdn/Ybn = 0·82–0·99), and a positive Sr spike and negative Nb, Zr, Hf and Ti spikes (Fig. 9g).
Clinopyroxene in the websterite dyke BC24 differs from those of the websterite lenses in its relative enrichment in MREE and consequent upward-convex pattern (Lan/Ybn = 0·24 and Gdn/Ybn = 1·09, Fig. 9h). Also, it does not display any Sr anomaly (Fig. 9g), being, therefore, compositionally similar to the clinopyroxenes of the Cr-diopside websterites reported by Rivalenti et al. (1995) and Mukasa & Shervais (1999).
Figure 9 also compares the trace element clinopyroxene patterns from this study with those of the lherzolites and Cr-diopside dykes reported by Rivalenti et al. (1995). The trace element profiles of the clinopyroxenes from the dunite transect plot between those of the pyroxenes in lherzolites and Cr-diopside dykes and approach the concentrations observed in lherzolites for the more compatible trace elements. The HREE concentrations of the dunitic clinopyroxene are on average similar to, or slightly higher than, those in lherzolite, whereas their LREE concentrations approach those observed in the Cr-diopside dykes. The trace element patterns of the dunitic clinopyroxenes also differ in having positive Sr anomalies, which are negative both in the lherzolites and dykes. Finally, their Zr, Ti and Y concentrations are similar to those of websterite dykes, whereas Sc is definitely higher.
|
Nd–Sr ISOTOPES |
|---|
Strontium and neodymium isotopic data for clinopyroxene separates from two Cr-spinel layers (BC1A and BC1C), one websterite dyke (BC24) and four dunites (BC3, BC32, BC48 and BC50) are reported in Table 7. On the 147Sm/144Nd vs 143Nd/144Nd diagram (Fig. 10a) these samples are compared with clinopyroxene and whole-rock data for lherzolites away from any websterite dykes, peridotites close to Cr-diopside websterite dykes, and Cr-diopside websterites. Only literature data with specific information on the relative position between websterite dykes and peridotite have been considered (Voshage et al., 1988
; Obermiller, 1994); data lacking this information are shown only for comparison (Voshage et al., 1987; Shervais & Mukasa, 1991; Mukasa & Shervais, 1999). Peridotites away from dykes have 143Nd/144Nd > 0·5134 and 147Sm/144Nd > 0·4, whereas the Cr-diopside websterite dykes and the peridotites close to dykes have 143Nd/144 Nd < 0·5132 and 147Sm/144Nd < 0·3. Samples in our study plot together with the Cr-diopside dykes and the peridotites that are in close proximity to the Cr-diopside websterites.
|
|
The isotopic data on clinopyroxene separates from the Balmuccia peridotite samples of Obermiller (1994
) define a ‘reference line’ corresponding to an age of 433 ± 75 Ma, with an 
Ndi (CHUR) = + 4·8. Obviously, this line is not an isochron and the data distribution in the Sm/Nd plot clearly suggests the effects of open-system processes in the Balmuccia peridotite massif and, particularly, the dunite body considered here.
In a present-day Sr–Nd isotope diagram (Fig. 10b) the dunites plot between the peridotites far from any dykes and the Cr-diopside websterites and peridotites close to the dykes. With respect to the MORB field (Hofmann, 1997) the peridotites far from dykes represent a more depleted mantle, whereas the dunites plot in the MORB field.
|
DISCUSSION |
|---|
A model for the Balmuccia dunites
Dunites occurring in ophiolite sequences are thought to represent focused flow channels of reactive melts derived from deeper mantle levels, which become olivine-saturated and pyroxene-undersaturated at the level of dunite occurrence (Kelemen et al., 1997
; Braun & Kelemen, 2002, and references therein). Because this process is robustly based on the polybaric variations of the primary phase boundaries in any mantle system, it can equally apply to the subcontinental mantle. Our discussion, therefore, regards mainly the origin of the melts whose focused flow produced dunites from both lherzolite and pyroxenite within the Balmuccia Massif.
The Balmuccia dunites have isotopic characteristics resembling those of the Cr-diopside websterite suite, which also occurs as depleted relics (Cr-diopside lenses) within the spinel-rich dunite domains. This suggests that the melts involved in the dunite genesis were kindred with the Cr-diopside websterites. On the basis of the trace element geochemistry, Rivalenti et al. (1995) inferred that the Cr-diopside websterites were generated by fractional crystallization of alkaline melts derived from a garnet-facies mantle source. Alkaline melts would have been potentially able to generate reactive dunites at mantle levels shallower than their source. Clear evidence of this is provided by the widespread presence of centimetre- to decimetre-thick dunitic rims often observed at the dyke–wall-rock interfaces, interpreted as the result of the infiltration into the wall-rock of the melts fractionating to form the websterite (Rivalenti et al., 1995; Mukasa & Shervais, 1999). These ‘wall’ dunites differ from those studied here in several aspects, including their volumes (which are comparatively small), and their Mg-number values, which are similar to, or lower than those of the host lherzolite (Rivalenti et al., 1995). Moreover, field evidence indicates that transformation into dunite also affected the websterite dykes, leaving spinel-rich relics of Cr-diopside lenses and ghost spinel-rich dykelets, similar to that illustrated in Fig. 3d. Although it is possible for a melt to react with, and reabsorb, its early cumulates, the system would rapidly evolve to clinopyroxene saturation; however, such a process is not observed in the Balmuccia Massif. Rather, the studied dunites point to an extreme pyroxene-dissolution reaction that leaves only refractory phases, olivine and spinel, as residual minerals. Therefore, diffuse percolation of an alkaline melt appears unlikely for the genesis of the large dunite bodies discussed here.
An alternative hypothesis for the dunite genesis arises from field evidence that the Cr-diopside websterite suite consists of several dyke generations. Partial melting of early Cr-diopside websterites is proved by the occurrence of the previously described spinel-rich ghost websteritic dykelets (Fig. 3d). It is worth noting that only the earliest dyke generation underwent partial melting after its emplacement, attesting to a later thermal pulse affecting the intruded mantle sequence. Because of the lower solidus temperature of pyroxenite with respect to peridotite (e.g. Hirschmann & Stolper, 1996) this thermal pulse triggered, at a given depth, websterite, but probably not peridotite, partial melting. We hypothesize that dunites may be the final result of this partial melting event and the subsequent melt focusing episode, through a two-stage process schematically illustrated in Fig. 11. The model we propose consists of the following steps: (1) the depleted lherzolite subcontinental mantle was veined by early Cr-diopside websterites; (2) a later thermal perturbation induced melting of the pyroxenites, whereas the lherzolite remained below the solidus; (3) the pyroxenite-derived melts pervasively moved through the host lherzolite, lowering its solidus temperature and thus triggering open-system partial melting; (4) melts released from the fertilized peridotite were progressively focused into channels and, by polybaric reaction, induced dunitization in the surrounding peridotite, even partially dissolving any pyroxenite layers encountered during ascent.
|
This model differs from a similar one proposed by Lundstrom et al. (2000
) in that the flowing melts do not derive from mafic lithologies, but from portions of mantle peridotite refertilized by pyroxenite-derived melts. The reason to prefer our approach is the extremely high pyroxenite volume required if the composition of the reactively flowing melt is limited to pyroxenite derivation, as demonstrated by the modelling below.
Numerical constraints to the model
In the absence of any experimental work on the Balmuccia pyroxenites, we have constrained our model by using the thermodynamic pMELTS program (Ghiorso et al., 2002) in the implemented version Adiabat-1ph (Smith & Asimow, 2005). However, the pMELTS approach (Ghiorso et al., 2002) may overestimate the solidus temperature by 60°C. Other problems are a systematic overestimation of the stability of orthopyroxene relative to olivine, an underestimation of garnet stability relative to clinopyroxene and a large dependence of modal spinel on the Cr2O3 contents (Asimow et al., 1995). Nevertheless, the resulting modal offsets do not exceed a few wt % and are close to the inaccuracy in the estimation of residual modes from melting modes (Smith & Asimow, 2005). Except for uncertainties derived from the above points, the pMELTS results fit well with experimental results of peridotite and pyroxenite partial melting (Hirschmann et al., 1998, 2003; Smith et al., 2003). Our model considers a mantle segment at 1·5–0·7 GPa, which is within the spinel peridotite facies, just below the plagioclase stability field.
Selection of input data
The Balmuccia Cr-diopside websterite suite consists of several dyke generations with a large compositional range even within the same dyke generation (Capedri et al., 1977; Comin-Chiaramonti et al., 1982; Rivalenti et al., 1995; Mukasa & Shervais, 1999). The Cr-diopside pyroxenites directly related to the dunite genesis survive only as relics, possibly residual from the melting process, as revealed by the very high Mg-number of clinopyroxene (Table 3). As there is no way to estimate the initial composition of the former websterites, we have chosen two pyroxenites from this study, MO375 and BM90-36Cr (Table 8), because of their contrasting compositions (Fig. 12). The composition of the host peridotite predating dunite formation used in the model corresponds to the average of the lherzolites away from pyroxenite dykes (Rivalenti et al., 1995; Table 8).
|
|
Pyroxenite partial melting and induced open-system partial melting of the ambient peridotite
The anhydrous solidus temperatures for selected potential source rocks are reported in Table 8. The lowest solidus is that of websterite MO375, which would, therefore, be the first to melt during any isobaric temperature increase. We have assumed that at 1·5 GPa the temperature conditions are such that the peridotite remains in the subsolidus and produces only low-degree partial melts in the MO375 pyroxenite, corresponding to F = 3·5% (F is the melt mass fraction). At 1·5 GPa, this pyroxenite produces nepheline-normative liquids up to F = 39%. The melt corresponding to F = 3·5% (Table 9), which contains 13·4% normative nepheline, infiltrates the host lherzolite. Melt–peridotite reaction decreases the peridotite solidus and is characterized by strong orthopyroxene dissolution because of the silica undersaturation of the melt. Such reaction results in increased melt volume and peridotite porosity. This, in turn, could accelerate melt flow and promote further peridotite dissolution. The process has been simulated under isothermal conditions by adding 5% of MO375 melt to the peridotite with each P decrease of 100 bar over a pressure range of 1·5–1·47 GPa. The final modal and chemical composition of the hybrid peridotite is shown in Table 9 and the full numerical run can be found in Electronic Appendix 4.
|
Focused flow through peridotite
Because of the reactions and melting during the hybridization process, the peridotite has been impregnated with 13·4% melt, which is nepheline-normative (11% nepheline, Table 9). Such a melt volume is sufficiently large for melt segregation to occur (Toramaru & Fujii, 1986
); focusing may nucleate at this level according to mechanisms thoroughly discussed by Aharanov et al. (1995), Kelemen et al. (1995, 1997), Spiegelman et al. (2001) and Morgan & Liang (2005).
Melt impregnation has been thermodynamically modelled by Asimow & Stolper (1999) by fluxing a melt originating from a host peridotite [the primitive upper-mantle source composition of Hart & Zindler (1986)] with a composition similar to that of the wall-rocks lining the flow channel. We have run the model under isentropic conditions between 1·47 GPa (pressure of the enriched peridotite) and 0·7 GPa with 100 bar stepwise pressure decrease, and with the oxygen fugacity fixed at the FMQ (fayalite–magnetite–quartz) buffer. According to Asimow & Stolper (1999), focusing is simulated by assuming variable ‘focusing factors’ 
d (d indicates discrete) values, in our case between 1·001 and 1·1, where d is the ratio of melt flux infinitesimally immediately above and below a given level. When = 1 focusing will not occur. Continuous focusing is characterized by the integrated focusing factor l = dm, where m is the number of incremental focusing events between the root and the end of the column.
The main model results are reported in Table 10 and the modal variations are illustrated in Fig. 13. The detailed numerical calculation is given in Electronic Appendices 5–8. Results from the model may be summarized as follows:
|
|
(1) Dunites are formed for all the assumed
d values, but the pressure at which they first appear (e.g. modal olivine >90%) increases with increasing d;
(2) Orthopyroxene is the last phase to be consumed and the opx-out pressure increases with d, so that the dunites formed with d = 1·001 always contain some residual orthopyroxene. No focusing (e.g. d = 1) produces only harzburgites;
(3) The gap between the opx-out and cpx-out pressure decreases along with the increase of the opx-out pressure (i.e. at increasing d);
(4) Compositional variations in the residue and transient melt after dunite formation are negligible, indicating that dunites may be an efficient way for isochemical melt transport to occur.
In all cases the l value at the opx-out point (Table 10) is much lower than that estimated by Asimow & Stolper (1999) for orthopyroxene dissolution. This discrepancy is probably a consequence of the use of MELTS calibration in their model instead of p-MELTS as in the present case, and of the composition of the infiltrating melt. Besides, in the Asimow & Stolper model the infiltrating melt is silica-saturated, whereas the melt in our numerical simulations is silica-undersaturated and alkaline.
Our model shows that under the postulated conditions, dunites can be formed between 1 and 1·2 GPa. The presence of relict orthopyroxene in the actual dunites may indicate that this mineral represents the final stage of dunite formation, just before the opx-out stage. The presence of clinopyroxene, which should disappear before orthopyroxene, can be explained by cooling-induced late fractional crystallization of the upwelling melts and/or closed-system crystallization of trapped melts in the dunite in the last phases of the process. Accordingly, it appears in textural and chemical disequilibrium with orthopyroxene.
Focused flow through the pyroxenite dykes
Pyroxenites are randomly distributed throughout the Balmuccia host peridotites as tabular bodies centimetres to several tens of metres long. They do not appear to be interconnected in three dimensions. It is therefore unlikely that the focused melt flow has remained trapped in each pyroxenite dyke in the entire Balmuccia mantle section. A focused channel randomly passing from peridotite to pyroxenite and vice versa would, however, impose severe difficulties on the modelling. We have, therefore, tested the effect of melt percolation in a large pyroxenitic body by simulating focused flow into two pyroxenite-host channels, represented by the composition of samples MO375 and BM90-36Cr, respectively. Supporting this unconstrained assumption is the virtual concordance of the dunite with the dominant peridotite foliation, which suggests that the focused channels had this orientation as well.
In either model dunite generation requires high focusing factors (Table 10 and Electronic Appendices 9–11). In MO375, dunite-in occurs at 0·87 GPa with d = 1·05 and at 1·07 GPa with d = 1·1 (Fig. 13). BM90-36Cr gives similar results, but slightly lower pressure for dunite-in (0·98 GPa at d = 1·1) (Fig. 13). In the initial dunites olivine coexists with clinopyroxene, and in the more developed dunites with large amounts of spinel (Fig. 13). This behaviour appears only over a limited pressure range (1· 07–1· 01 GPa in MO375 and 0·98–0·86 in BM90-36Cr at d = 1·1).
In summary, pyroxenite-derived dunites require much higher focusing factors than the peridotite-derived ones. This is consistent with a lower porosity and melt interconnectivity in the pyroxenite with respect to peridotite. It is also likely to depend on the reactivity with the matrix of the melt flowing in the column. The melt derived from the peridotite open-system melting is, in our model, dominated by the composition of the pyroxenite-derived contaminant. It is, therefore, more reactive against peridotite than pyroxenite. It is furthermore evident that the transformation of an olivine-free (or olivine-poor) pyroxenite into dunite requires more extreme reactions (i.e. a higher volume of percolating melt) compared with peridotite, which is obviously already olivine-rich.
One consequence is that when a melt generated from a peridotite fertilized with a pyroxenite-derived melt flows into a heterogeneous sequence of peridotite and pyroxenite layers, it will possibly reduce the encountered lherzolite to dunite, leaving residues of partially resorbed websterite accompanied by Cr-spinel enrichments. The example shown in Fig. 3d accounts for local Cr-spinel enrichment over a ghost Cr-diopside that is almost completely dissolved, in good agreement with the results of our model.
Major element composition of the model dunites
The compositions of some model dunites (Electronic Appendices 6, 7, 8, 10 and 11) are compared in Fig. 14 with the compositional range of the actual dunites. The model dunites that most closely resemble the natural ones are those resulting from peridotite reaction at d = 1·05 and 1·07 and from pyroxenite BM90-36Cr at d = 1·05. Minor differences exist in the overall lower Al2O3 and higher SiO2 in the model trends. The CaO and MgO concentrations and Mg-number values of the modelled dunites from both peridotite and pyroxenite at pressures 
1 GPa are within the range of the natural dunites. Figure 14 also shows that olivines from peridotite- and BM90-36Cr-derived dunites at pressures < 0·9 GPa have Mg-number values comparable with those of the measured olivines.
|
In summary, there is good agreement between the observed and modelled compositions. Most of the observed differences are within error of the major element analyses.
Trace elements in the model clinopyroxenes
The Adiabat_1ph program also includes an estimate of the trace element variations in the modelled process. The results, using the partition coefficient set of Ionov et al. (2002), are reported in Electronic Appendices 4–11. Model dunites may contain clinopyroxene only at the initial dunite-in stages. The model melts coexisting with the dunites, however, do contain CaO (Electronic Appendices 5–11), which would form clinopyroxene at temperatures below the solidus. In agreement with the textural evidence, we assume that clinopyroxene and additional granoblastic olivine derive from crystallization of residual interstitial melt. The trace element concentration in the potential clinopyroxene in equilibrium with the melts in the dunite range was estimated by using the consistent cpx/melt partition coefficients of Ionov et al. (2002). The results compare favourably with the characteristics of the LREE-depleted natural clinopyroxenes (Fig. 15).
|
Our models cannot explain the spoon-shaped patterns, which may result from weak chromatographic element fractionation related to the last phases of melt percolation during porosity reduction. Alternatively, these shapes may be generated by the channelized melt transport itself (Spiegelman & Kelemen, 2003
). In this case, strong chemical variability may occur between the centre of the channels, dominated by trace element enriched melts from depth, and the edges of the channels, which transport depleted melts extracted from the interchannel region. Element diffusion across the width of the channel may result in a large variety of trace element patterns, from depleted to spoon-shaped as in the samples evaluated in this study.
Composition of the extracted melts
A complete set of the major element and REE composition of the aggregated melts throughout the process of focused flow is reported in Electronic Appendices 5–11. The model melts are high-Mg alkali basalts. Normative nepheline concentration increases with increasing focused flow (i.e. at decreasing pressure) up to a virtually steady concentration in the dunite segment. Normative nepheline is in the range of 4·8–3·8% and 1·8–0·3% in the peridotite runs (at d = 1·1 and 1·05, respectively), 3·9–3·7% and 6·9–6·7% in MO375 (at d = 1·1 and 1·05, respectively) and 8·2–8·1% in BM90-36Cr (at d = 1·1). In all cases, the REE patterns are typically enriched: Lan/Ybn is 2·0 and 1·9 (at d = 1·1 and 1·05, respectively) in the peridotite, 2·9 and 2·6 in MO375 (at d = 1·1 and 1·05, respectively) and 3·2 in BM90-36Cr (at d = 1·1).
The average melt composition in the pressure range where dunite forms is reported in Table 11 for the various lithologies undergoing focused melt percolation, and for the considered d factors. The melts in the dunite segment represent the highest melt volume of the whole percolating region and are, therefore, those that most probably will be extruded to the surface.
|
Basalts with geochemical features resembling those of the melts resulting from our numerical modelling are well known along slow- and ultraslow-spreading sectors of mid-ocean ridges. These regions are characterized by reduced magma production, thus allowing the compositional signal of pyroxenite-derived melts to be preserved in the erupted lavas. Figure 16a compares our melts from the dunite segments, recalculated to MgO 8 wt % by olivine fractionation (Table 11), with the composition of lavas from the ultraslow-spreading South-West Indian Ridge (SWIR) basalts (Standish et al., 2008
) and E-type (enriched) MORB lavas from the Mid-Atlantic Ridge (MAR) (14–15°N, Hémond et al., 2006). The MgO 8% melt composition is within the range for these E-MORB lavas. REE patterns of the model liquids also resemble E-type MORB lavas (Fig. 16b) from the MAR (14–15°N, Hémond et al., 2006), the Central Indian Ridge (Nauret et al., 2006), the MARK area at 23°20'N on the MAR (Donnelly et al., 2004) and the East Pacific Rise (EPR; Niu & Batiza, 1997).
|
Standish et al. (2008
) proposed that the SWIR lavas are derived from melting of mafic lithologies and diffuse or focused (as our case) flow into the overlying mantle. Hémond et al. (2006) suggested the derivation of E-MORB along the MAR by melting of a mantle fertilized by alkali basalt or (Donnelly et al., 2004) by melts of subducting oceanic crust. These hypotheses are consistent with the model proposed here, in which the alkaline melts are formed by focused flow of liquids generated in a mantle source previously fertilized by pyroxenite-derived alkaline partial melts.
Finally, is there any evidence in the Ivrea crust of melts like those resulting from our model? The only suitable occurrences are represented by amphibolite dykes derived from the metamorphism of alkali basalts cropping out in the northwestern, structurally deeper, part of the volcano-sedimentary Kinzigitic Formation (Sills & Tarney, 1984; Mazzucchelli & Siena, 1986). The metamorphism of this formation predates the Permian intrusion of the igneous Ivrea–Verbano complex. The protoliths were of early Palaeozoic, possibly Devonian age. A relationship with the process recorded in the Balmuccia mantle, if demonstrated, would have very important implications for the geodynamic history of this region, but to constrain this hypothesis needs much more detailed geochemical work, which is, at present, lacking.
Final considerations
In agreement with the field observations, the models indicate that it is possible for dunites to form from different lithologies, both peridotite and pyroxenite, infiltrated by the same initial melt. The pressure range of dunite formation is largely model-dependent (e.g. by assuming a peridotite less fertile than that used here, pyroxenite-derived dunites would form at higher pressure). Our modelling indicates only partial overlap in the pressure ranges of peridotite- and pyroxenite-derived dunites, which is consistent with the notion that those in peridotites form at greater depth. Fertilization by pyroxenite-derived melts allows dunite formation at comparatively greater depths than would be inferred on the basis of ophiolite occurrences. In fact, most of the ophiolite-hosted dunites are interpreted as resulting from processes at depths corresponding to the plagioclase stability field (e.g. Morgan & Liang, 2005). The mechanism we have discussed here allows dunite formation to occur well within the spinel stability field, providing evidence for deep channelling that may also be active in the sub-ridge oceanic system. It could be significant that in MOR segments characterized by low extents of mantle melting the occurrence of alkaline silica-undersaturated melts is widely reported and their composition is similar to that of the model melts discussed here.
The preservation of dunite bands with different characteristics is consistent with the fact that the melts percolating through the two lithologies (peridotite and pyroxenite) have different compositions and viscosities (Electronic Appendices 5–11) and under laminar flow conditions would hardly mix.
The numerical simulations also account for other field observations: (1) the higher modal spinel in dunites containing websterite lenses, which is a model result of the pyroxenite BM90-36Cr runs; (2) the flow structures in the spinel layers, which is consistent with the flux of high melt volumes during the dunite formation stage (Electronic Appendices 5–11) and consequent mineral segregation.
|
CONCLUSIONS |
|---|
The dunites and associated Cr-spinel layers occurring as tabular bodies within the Balmuccia lherzolite have refractory bulk-rock and mineral phase compositions. The constant association between centimetre-size Cr-diopside websterite lenses, massive spinel layers and dunites that have high modal spinel abundances and the isotopic similarity of the dunites to the websterites of the Cr-diospide suite suggest that Cr-diopside pyroxenites were involved in dunite genesis. The Adiabat_1ph thermodynamic program has proved to be a powerful tool for modelling the dunite-forming process. We have assumed that the pyroxenite solidus is intersected by the P–T path at 1·5 GPa and that pyroxenite incipient melts pervasively migrate through the host peridotite. The addition of pyroxenite-derived melts lowers the peridotite solidus temperature, inducing open-system partial melting. The derived hybrid melts trigger focused flow into the overlying mantle. Dunites are formed by assuming flow through peridotite and into pyroxenite, and the depth at which dunite starts to appear is a function of the flow rate (the higher the flow rate, the deeper the dunite first appearance) and of the composition of the host rocks. The latter point is especially evident for the two modelled pyroxenites, one of which (MO375) produces relatively iron-rich dunites at greater depth, whereas the other (BM90-36Cr) produces spinel-rich dunites compositionally comparable with the actual ones. Modelled trace element compositions and REE patterns correspond fairly well to the observed ones, except for a few features (e.g. spoon-shaped REE patterns) that are probably related to very late migration of small volumes of extremely evolved residual melts. Another result of the models is the virtual constancy of the melt compositions at the dunite stage and their large volume, both indicating that dunite channels are efficient pathways for large melt volumes. These melts are similar in composition to those sampled along the slow- and ultraslow-spreading sectors of the MOR system, thus suggesting the possibility that this process is significant for melt generation beneath mid-ocean ridges.
Dunite formation was a relatively old event in the history of the subcontinental mantle represented by the Balmuccia Massif. Although the dunites postdate the development of the regional depleted lherzolite composition, their close association with early Cr-diopside websterites, represented only by relics, indicates that they predate the intrusion of the main Cr-diopside dyke swarms that cut the dunite. Therefore, after the intrusion of the early Cr-diopside websterites, the Balmuccia mantle is inferred to have undergone a heating event that induced melting of these dykes and triggered dunite formation. The age of this thermal event is unconstrained. The Sm/Nd ‘reference line’ of 433 ± 75 Ma and the Permian age of emplacement of the mafic–ultramafic Ivrea–Verbano complex suggest a Palaeozoic age for the accretion of the Balmuccia peridotite from the asthenosphere to the lithosphere. In this scenario dunite formation could represent a relatively early stage.
|
SUPPLEMENTARY DATA |
|---|
Supplementary data for this paper are available at Journal of Petrology online.
|
ACKNOWLEDGEMENTS |
|---|
The authors thank P. Asimow, M. Grégoire, S. B. Mukasa and P. Smith for thorough and constructive reviews, and M. Wilson for editorial handling and reviewing. Simone Tommasini is acknowledged for his help with the isotopic analytical methods and for having made available the isotopic facilities of the Dipartimento di Scienze della Terra, Università di Firenze. L. Sardisco, M. Saraceno and T. Giovanardi are gratefully thanked for their valuable help in the isotopic laboratory. M. Fialin and F. Couffignal are thanked for their technical support at CAMPARIS microprobe. E.B. thanks the Università di Firenze for a post-doctoral fellowship. This research has been financially supported by MURST (COFIN 2005 and 2007).
*Corresponding author. Telephone: +39-059-2055820. Fax: +39-059-2055887. E-mail: maurizio.mazzucchelli{at}unimore.it
|
REFERENCES |
|---|
Aharanov E, Whitehead JA, Kelemen PB, Spiegelman M. Channeling instability of upwelling melt in the mantle. Journal of Geophysical Research (1995) 100:20433–20450.[CrossRef]
Asimow PD, Stolper EM. Steady-state mantle–melt interactions in one dimension: I. Equilibrium transport and melt focusing. Journal of Petrology (1999) 40:475–494.[CrossRef][Web of Science]
Asimow PD, Hirschmann MM, Ghiorso MS, O’Hara MJ, Stolper EM. The effect of pressure-induced solid–solid phase transitions on decompression melting of the mantle. Geochimica et Cosmochimica Acta (1995) 59:4489–4506.[CrossRef][Web of Science]
Avanzinelli R, Boari E, Conticelli S, Francalanci L, Guarnieri L, Perini G, Petrone CM, Tommasini S, Ulivi M. High precision Sr, Nd, and Pb isotopic analyses using new generation thermal ionisation mass spectrometer: aims and perspective for isotope geology applications. Periodico di Mineralogia (2005) 74:147–166.
Bence AE, Albee AL. Empirical correction factors for the electron microanalyses of silicates and oxides. Journal of Geology (1968) 76:382–403.[Web of Science]
Bodinier JL. Geochemistry and petrogenesis of the Lanzo peridotite body, Western Alps. Tectonophysics (1988) 149:67–88.[CrossRef][Web of Science]
Braun MG, Kelemen PB. Dunite distribution in the Oman ophiolite: implications for melt flux through porous dunite conduits. Geochemistry, Geophysics, Geosystems (2002) 3:10.[Medline]
Capedri S, Garuti G, Rivalenti G, Rossi A. The origin of the Ivrea–Verbano basic formation—pyroxenitic and gabbroic mobilisates as product of partial melting of mantle peridotite. Neues Jahrbuch für Mineralogie, Monatshefte (1977) 4:168–179.
Comin-Chiaramonti P, Demarchi G, Sinigoi S, Siena F. Relazioni tra fusione e deformazione nella Peridotite di Balmuccia (Ivrea–Verbano). Rendiconti della Società Italiana di Mineralogia e Petrologia (1982) 38:685–700.
Donnelly KE, Goldstein SL, Langmuir CH, Spiegelman M. Origin of enriched ocean ridge basalts and implications for mantle dynamics. Earth and Planetary Science Letters (2004) 226:347–366.[CrossRef][Web of Science]
Ernst WG. Petrochemical study of lherzolitic rocks from the western Alps. Journal of Petrology (1978) 19:341–392.
Ghiorso MS, Hirschmann MM, Reiners PW, Kress VC. The pMELTS: a revision of MELTS for improved calculation of phase relations and major element partitioning related to partial melting of the mantle to 3 GPa. Geochemistry, Geophysics, Geosystems (2002) 3:10.[Medline]
Hart SR, Zindler A. In search of a bulk-earth composition. Chemical Geology (1986) 57:247–267.[CrossRef][Web of Science]
Hémond C, Hofmann AW, Vlastélic I, Nauret F. Origin of MORB enrichment and relative trace element compatibilities along the Mid-Atlantic Ridge between 10° and 24°N. Geochemistry, Geophysics, Geosystems (2006) 7:10.
Hirschmann MA, Stolper EM. A possible role for garnet pyroxenite in the origin of the garnet signature in MORB. Contributions to Mineralogy and Petrology (1996) 124:219–229.[CrossRef][Web of Science]
Hirschmann MA, Ghiorso MS, Wasylenki LE, Asimow PD, Stolper EM. Calculation of peridotite partial melting from thermodynamic models and melts. I. Review of methods and comparison with experiments. Journal of Petrology (1998) 39:1091–1115.[CrossRef][Web of Science]
Hirschmann MA, Kogiso T, Baker MB, Stolper EM. Alkalic magmas generated by partial melting of garnet pyroxenite. Geology (2003) 31:481–484.
Hofmann AW. Chemical differentiation of the Earth: The relationship between mantle, continental crust, and oceanic crust. Earth and Planetary Science Letters (1988) 90:297–314.[CrossRef][Web of Science]
Hofmann AW. Mantle geochemistry: the message from oceanic volcanism. Nature (1997) 385:219–229.[CrossRef][Web of Science]
Ionov DA, Bodinier JL, Mukasa SB, Zanetti A. Mechanisms and sources of mantle metasomatism: major and trace element compositions of peridotite xenoliths from Spitsbergen in the context of numerical modelling. Journal of Petrology (2002) 43:1–41.
Kelemen PB, Dick H. JB. Focused melt flow and localized deformation in the upper mantle: juxtaposition of replacive dunite and ductile shear zones in the Josephine peridotite, SW Oregon. Journal of Geophysical Research (1995) 100:423–438.[CrossRef]
Kelemen PB, Whitehead JA, Ahranov E, Jordahl KA. Experiments on flow focusing in soluble porous media, with applications to melt extraction from the mantle. Journal of Geophysical Research (1995) 100:475–496.[CrossRef]
Kelemen PB, Hirth G, Shimizu N, Spiegelman M, Dick H. JB. A review of melt migration processes in the adiabatically upwelling mantle beneath oceanic spreading ridges. Philosophical Transactions of the Royal Society of London, Series A (1997) 355:283–318.[CrossRef]
Keshav S, Gudfinsson GH, Sen G, Fei Y. High-pressure melting experiments on garnet clinopyroxenite and the alkalic to tholeiitic transition in ocean-island basalts. Earth and Planetary Science Letters (2004) 223:365–379.[CrossRef][Web of Science]
Kogiso T, Hirschmann MM. Experimental study of clinopyroxenite partial melting and the origin of ultra-calcic melt inclusions. Contributions to Mineralogy and Petrology (2001) 142:347–360.[Web of Science]
Kogiso T, Hirschmann MM, Frost DJ. High-pressure partial melting of garnet pyroxenite: possible mafic lithologies in the source of oceanic island basalts. Earth and Planetary Science Letters (2003) 216:603–617.[CrossRef][Web of Science]
Kogiso T, Hirschmann MM, Pertermann M. High-pressure partial melting of mafic lithologies in the mantle. Journal of Petrology (2004) 45:2407–2422.
Lensh G. Die Ultramafitite der Zone von Ivrea. Annales Universitatis Saraviensis (1971) 9:5–146.
Lundstrom CC, Gill J, Williams Q. A geochemically consistent hypothesis for MORB generation. Chemical Geology (2000) 162:105–126.[CrossRef][Web of Science]
Mazzucchelli M, Siena F. Geotectonic significance of the metabasites of the Kinzigitic Series, Ivrea–Verbano Zone (Western Italian Alps). Tschermaks Mineralogische und Petrographische Mitteilungen (1986) 35:99–116.[CrossRef][Web of Science]
Morgan Z, Liang Y. An experimental study of the kinetics of lherzolite reactive dissolution with applications to melt channel formation. Contributions to Mineralogy and Petrology (2005) 150:369–385.[CrossRef][Web of Science]
Mukasa SB, Shervais JW. Growth of subcontinental lithosphere: evidence from repeated dike injections in the Balmuccia lherzolite massif, Italian Alps. Lithos (1999) 48:287–316.[CrossRef][Web of Science]
Nauret F, Abouchami W, Galer S. JG, Hofmann AW, Hémond C, Chauvel C, Dyment J. Correlated trace element-Pb isotope enrichments in Indian MORB along 10°–20°S, Central Indian Ridge. Earth and Planetary Science Letters (2006) 245:137–152.[CrossRef][Web of Science]
Nicolas A. Structures of Ophiolites and Dynamics of Oceanic Lithosphere (1989) 367 pp: Dordrecht: Kluwer Academic.
Niu Y, Batiza R. Trace element evidence from seamounts for recycled oceanic crust in the Eastern Pacific mantle. Earth and Planetary Science Letters (1997) 148:471–483.[CrossRef][Web of Science]
Obermiller WA. Chemical and isotopic variations in the Balmuccia, Baldissero and Finero peridotite massifs (Ivrea Zone, N-Italy). PhD thesis (1994) Max Planck Institut Mainz. 192.
O’Hara MJ. The bearing of phase equilibria studies in synthetic and natural systems on the origin and evolution of basic and ultrabasic rocks. Earth-Science Reviews (1968) 4:69–133.
Ottonello G, Ernst WG, Joron JL. Rare earth and 3d transition element geochemistry of peridotitic rocks: I. Peridotite from the Western Alps. Journal of Petrology (1984) 25:343–372.
Peressini G, Quick JE, Sinigoi S, Hofmann AW, Fanning M. Duration of a large mafic intrusion and heat transfer in the lower crust: a SHRIMP U–Pb zircon study in the Ivrea–Verbano Zone (Western Alps, Italy). Journal of Petrology (2007) 48:1185–1218.
Pertermann M, Hirschmann MM. Partial melting experiments on a MORB-like pyroxenite between 2 and 3 GPa: constraints on the presence of pyroxenite in basalt source regions from solidus location and melting rate. Journal of Geophysical Research (2003a) 108. doi:10.129/2000JB000118.
Pertermann M, Hirschmann MM. Anhydrous partial melting experiments on MORB-like eclogite: phase relations, compositions and mineral–melt partitioning of major elements at 2–3 GPa. Journal of Petrology (2003b) 44:2173–2201.
Piccardo GB, Zanetti A, Müntener O. Melt/peridotite interaction in the Southern Lanzo peridotite: field, textural and geochemical evidence. Lithos (2007) 94:181–209. doi:10.1016/j.lithos.2006.07.002.[CrossRef][Web of Science]
Quick JE, Sinigoi S, Snoke AW, Kalakay TJ, Mayer A, Peressini G. Geologic map of the southern Ivrea–Verbano Zone, northern Italy. US Geological Survey Geologic Investigations Series Map (2003) I-2776, scale 1 : 25 000.
Rivalenti G, Garuti G, Rossi A. The origin of the Ivrea–Verbano basic formation (western Italian Alps). Whole rock geochemistry. Bollettino della Società Geologica Italiana (1975) 94:1149–1186.
Rivalenti G, Mazzucchelli M, Vannucci R, Hofmann AW, Ottolini L, Obermiller W. The relationship between websterite and peridotite in the Balmuccia peridotite massif (NW Italy) as revealed by trace element variations in clinopyroxene. Contributions to Mineralogy and Petrology (1995) 121:275–288.[CrossRef][Web of Science]
Rivalenti G, Mazzucchelli M, Laurora A, Ciuffi S. IA, Zanetti, A. Vannucci R, Cingolani CA. The backarc mantle lithosphere in Patagonia, South America. Journal of South American Earth Sciences (2004) 17:121–152.[CrossRef][Web of Science]
Seyler M, Lorand J, Dick H, Drouin M. Pervasive melt percolation reactions in ultra-depleted refractory harzburgites at the Mid-Atlantic Ridge, 15°20'N: ODP Hole 1274A. Contributions to Mineralogy and Petrology (2007) 153:303–319.[CrossRef][Web of Science]
Shervais JW. Ultramafic layers in the Alpine-type lherzolite massif at Balmuccia, NW Italy. Memorie di Scienze Geologiche, Università di Padova (1979) 33:135–145.
Shervais JW, Mukasa SB. The Balmuccia orogenic lherzolite massif, Italy. Journal of Petrology, Special Lherzolite Issue (1991) 155–174.
Sills JD, Tarney J. Petrogenesis and tectonic significance of amphibolites interlayered with metasedimentary gneisses in the Ivrea Zone, Southern Alps, Northwest Italy. Tectonophysics (1984) 107:187–206.[CrossRef][Web of Science]
Sinigoi S, Comin-Chiaramonti P, Demarchi G, Siena F. Differentiation of partial melts in the mantle: evidence from the Balmuccia peridotite, Italy. Contributions to Mineralogy and Petrology (1983) 82:351–359.[CrossRef][Web of Science]
Smith PM, Asimow PD. Adiabat_1ph: a new public front-end to the MELTS, pMELTS and pHMELTS models. Geochemistry, Geophysics, Geosystems (2005) 6. doi:10.1029/2004GC000816.
Smith PM, Asimow PD, Stolper EM. Thermodynamic modelling of melting in chemically heterogeneous mixtures of peridotite and pyroxenite. Goldschmidt Conference, Abstracts Volume (2003) A440.
Spiegelman M, Kelemen PB. Extreme chemical variability as a consequence of channelized melt transport. Geochemistry, Geophysics, Geosystems (2003) 4. doi:10.1029/2002GC000336.
Spiegelman M, Kelemen PB, Aharanov E. Causes and consequences of flow organization during melt transport: the reaction infiltration instability. Journal of Geophysical Research (2001) 106:2061–2078.[CrossRef]
Standish JJ, Dick H. JB, Michael PJ, Melson WG, O’Hearn T. MORB generation beneath the ultraslow spreading Southwest Indian Ridge (9–25°E): Major element chemistry and the importance of process versus source. Geochemistry, Geophysics, Geosystems (2008) 9. doi:10.1029/2008GC001959.
Stolper E. A phase diagram for mid-ocean ridge basalts: preliminary results and implications for petrogenesis. Contributions to Mineralogy and Petrology (1980) 74:13–27.[CrossRef][Web of Science]
Suhr G. Melt migration under oceanic ridges: inferences from reactive transport modelling of upper mantle hosted dunites. Journal of Petrology (1999) 40:575–599.[CrossRef][Web of Science]
Suhr G, Seck HA, Shimizu M, Günther D, Jenner G. Infiltration of refractory melts into the lowermost oceanic crust: evidence from dunite- and gabbro-hosted clinopyroxenes in the Bay of Islands ophiolite. Contributions to Mineralogy and Petrology (1998) 131:136–154.[CrossRef][Web of Science]
Suhr G, Hellebrand E, Snow JE, Seck HA, Hofmann AW. Significance of large, refractory dunite bodies in the upper mantle of the Bay of Islands Ophiolite. Geochemistry, Geophysics, Geosystems (2003) 4. doi:10.1029/2001GC000277.
Toramaru A, Fujii N. Connectivity of melt phase in a partially molten peridotite. Journal of Geophysical Research (1986) 91:9239–9252.
Voshage H, Hunziker JC, Hofmann AW, Zingg A. A Nd and Sr isotopic study of the Ivrea zone, Southern Alps, N Italy. Contributions to Mineralogy and Petrology (1987) 97:31–42.[CrossRef][Web of Science]
Voshage H, Sinigoi S, Mazzucchelli M, Demarchi G, Rivalenti G, Hofmann AW. Isotopic constraints on the origin of ultramafic and mafic dikes in the Balmuccia peridotite (Ivrea Zone). Contributions to Mineralogy and Petrology (1988) 100:261–267.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||