english version  

recherche  

 

2005

Demaiffe, D., Féménias, O., Berger, J. 2005. Variscan Moho beneath the French Massif central: a xenolith perspective from Puy Beaunit. Goldschmidt Moscow, Idaho, USA.

Puy Beaunit maar carries both mantle-derived and mafic-ultramafic plutonic xenoliths [1] highlighting the complexity of the shallow upper mantle and lower crust beneath the French Massif Central.
The population of plutonic xenoliths commonly shows asymmetric, mm to cm thick, layering. Layers are pyroxenitic to gabbro-noritic, less commonly peridotitic and anorthositic. They presumably derive from a layered intrusion (Beaunit Layered Complex, BLC), located at the crust–mantle boundary and emplaced at 257±6 Ma (SIMS U-Pb zircon age, [2]). BLC belongs to the large Permian within-plate mafic (high-Mg) calc-alkaline magmatic event recognized in W Europe and spatially controlled by post-Variscan transcurrent basin tectonics in an intracontinental setting [2]. Cumulate phases are ol, opx, cpx, gt [3], am and pl. Rare intercumulus accessory phases (phl, ilm, ap, ru, armacolite, srilankite and zircon) are observed in the most differentiated layers [4]. The trace element contents of the cumulates are similar to those of the Bushveld Lower Zone.
Mantle xenoliths range from fertile spinel lherzolites to refractory dunites. Fertile peridotites have registered a modal (amphibole-bearing) and cryptic (LILE and Pb enrichment, negative Nb and Ta anomalies) metasomatic event that took place before the Permian melting episode. Depletion processes are related to two melt extraction episodes. The first melting and metasomatic event is attributed to a fluid/liquid derived from a pre-Variscan subduction. It is sub-contemporaneous with the texture acquisition and deformation of the uppermost mantle (lithospheric delamination). The second melting event produced high-Mg basalts that gave rise to the BLC [5].
Sr and Nd isotopic data on the BLC (87Sr/86Sr257Ma: 0.7027 to 0.7062 and eNd257Ma: +6.2 to -6.4) and on the mantle suite (87Sr/86Sr257Ma: 0.7033 to 0.7053 and eNd257Ma: +5.4 to -2.4) are largely overlapping, confirming their genetic relationship. These values plot outside the French Massif Central mantle array and possibly record the pre-variscan subduction-related mantle metasomatic enrichment.
References: [1] Féménias et al. 2001, CRAS, 535-542. [2] Féménias et al. 2003, Chem Geol 199, 293-315. [3] Berger et al. Eu J Min (submitted). [4] Féménias et al. Lithos (accepted). [5] Féménias et al. 2004, CMP 148, 13-28.
top...

2004

Féménias, O., Nkono, C., Diot, H., Berza, T., Demaiffe, D. 2004. Flowage differentiation in an andesitic dyke of the Motru Dyke Swarm (Southern Carpathians, Romania): Evidence from AMS, CSD and geochemical investigations. RST, Strasbourg, France.

From a regional survey of a large Pan-African calc-alkaline dyke swarm (of basaltic-andesitic-dacitic-rhyolitic composition) of the Alpine Danubian window from South Carpathians (Romania), two populations of dykes have been described: thick (from 1 to 30 meters) N-S trending dykes and thin (less than 1 m) E-W dykes. These two populations crosscut the country rocks without simple chronological relations between them.
The thick dykes display asymmetrical fabrics that involve a relatively long history of emplacement and important distance of flow (Féménias et al., JSG in press). The petrological implications of such long time flowing are not well known. To better understand these implications, two dykes of different thickness (5.5m for TJ31 and 23m for TJ34) have been studied with physical methods (crystal size distribution –CSD- and anisotropy of magnetic susceptibility -AMS) coupled to whole rock geochemistry and mineral chemistry.
All the physical and chemical variations observed along dyke’s width point to concordant results and show that the variations of both the modal proportion and size of the amphibole and biotite micro-phenocrysts inside the dykes, deduced from the classical CSD measurements, are the result of a mechanical segregation of suspended crystals during magmatic transport. These variations cannot be linked to different pulses of flow during ascent of the magma. Despite a pene-contemporaneous regional tectonic, the flow-induced differentiation in the large dyke, characterised by the concentration of pre-existing crystals (Ti-rich pargasite-tschermakite, clinopyroxene and plagioclase) in the core and the concomitant extraction of differentiated liquid near the walls, contributed to a slight chemical differentiation. As a result, the core has an andesitic basalt composition whereas the margins are andesitic. The chilled margins appear as a slightly more evolved liquid with a Newtonian behaviour when compared to the average composition of the dyke. The localisation of the liquid on both sides of the dyke has certainly facilitated the ascent of the central part of the dyke that behaved as a Binghamian mush.

top...

2003

Féménias, O., Diot, H., Berza, T., Gauffriau, A., Demaiffe, D. 2003. Asymmetrical to symmetrical magnetic fabric of dykes: Paleo-flow orientations and Paleo-stresses recorded on feeder-bodies from the Motru Dyke Swarm (Romania). EUG XII, Nice, France.

Puy Beaunit maar carries both mantle-derived and mafic-ultramafic plutonic xenoliths [1] highlighting the complexity of the shallow upper mantle and lower crust beneath the French Massif Central.
The population of plutonic xenoliths commonly shows asymmetric, mm to cm thick, layering. Layers are pyroxenitic to gabbro-noritic, less commonly peridotitic and anorthositic. They presumably derive from a layered intrusion (Beaunit Layered Complex, BLC), located at the crust–mantle boundary and emplaced at 257±6 Ma (SIMS U-Pb zircon age, [2]). BLC belongs to the large Permian within-plate mafic (high-Mg) calc-alkaline magmatic event recognized in W Europe and spatially controlled by post-Variscan transcurrent basin tectonics in an intracontinental setting [2]. Cumulate phases are ol, opx, cpx, gt [3], am and pl. Rare intercumulus accessory phases (phl, ilm, ap, ru, armacolite, srilankite and zircon) are observed in the most differentiated layers [4]. The trace element contents of the cumulates are similar to those of the Bushveld Lower Zone.
Mantle xenoliths range from fertile spinel lherzolites to refractory dunites. Fertile peridotites have registered a modal (amphibole-bearing) and cryptic (LILE and Pb enrichment, negative Nb and Ta anomalies) metasomatic event that took place before the Permian melting episode. Depletion processes are related to two melt extraction episodes. The first melting and metasomatic event is attributed to a fluid/liquid derived from a pre-Variscan subduction. It is sub-contemporaneous with the texture acquisition and deformation of the uppermost mantle (lithospheric delamination). The second melting event produced high-Mg basalts that gave rise to the BLC [5].
Sr and Nd isotopic data on the BLC (87Sr/86Sr257Ma: 0.7027 to 0.7062 and eNd257Ma: +6.2 to -6.4) and on the mantle suite (87Sr/86Sr257Ma: 0.7033 to 0.7053 and eNd257Ma: +5.4 to -2.4) are largely overlapping, confirming their genetic relationship. These values plot outside the French Massif Central mantle array and possibly record the pre-variscan subduction-related mantle metasomatic enrichment.
References: [1] Féménias et al. 2001, CRAS, 535-542. [2] Féménias et al. 2003, Chem Geol 199, 293-315. [3] Berger et al. Eu J Min (submitted). [4] Féménias et al. Lithos (accepted). [5] Féménias et al. 2004, CMP 148, 13-28.

top...
S. Brassinnes, E. Balaganskaya and D. Demaiffe. 2003. MAGMA GENESIS AND EVOLUTION IN THE
KOLA ALKALINE-CARBONATITE PROVINCE. A Pb-Sr-Nd ISOTOPE STUDY.
EUG XII, Nice, France.

The Kola Peninsula (NW Russia) is well known for its late Devonian (380-360Ma) ultramafic, alkaline and carbonatitic magmatic province, that comprises more than 20 intrusions including the two giant massifs of agpaitic nepheline syenite of Khibina and Lovozero and numerous smaller annular differentiated complexes (i.e. : Kovdor, Vuorijarvi, ...). These massifs intrude the Archean to Proterozoic crustal terranes. Whole rock Pb isotope data are reported for the first time on these intrusions. Representative samples of the Kovdor massif (carbonatites, ultramafic cumulates, ijolitesmelteigites) have been analysed. Two U-rich pyrochlore-bearing carbonatites have extremely high measured 206Pb/204Pb and 207Pb/204Pb ratios (94.1 to 253.6 and to 19.6 to 28.2 respectively). Together with 5 other samples, these carbonatites plot along a linear array in the 207Pb/204Pb-206Pb/204Pb diagram. Interpreted as a secondary isochron, this array gives an age indication of 392±15 Ma (MSWD = 22), in agreement with the U-Pb baddeleyite age of 380±4 Ma. Crustal contamination or source heterogeneity are hardly detectable with such radiogenic Pb composition. Some samples of other massifs of the Kola region (Khibina, Lovozero, Sebljavr and Ozernaya Varaka) have also been analysed, they plot reasonably close to the Kovdor Pb-Pb isochron, confirming that all the complexes of the province are grossly contemporaneous. Although most rocks plot in the depleted mantle quadrant of the Sr-Nd diagram, the data show some dispersion. The Kovdor ultramafic cumulates and carbonatites have overlapping compositions ("Nd(t): +0.5 to +6 ; (87Sr/86Sr)I : 0.7031 to 0.7039) while the ijolites-melteigites tend to have lower "Nd(t) (+0.5 to -3.4) and higher (87Sr/86Sr)I (0.7035 to 0.7047) values. This dispersion can be explained either by heterogeneities of the mantle (or mixing between different geochemical reservoirs) or by open system behaviour during complex emplacement history and hydrothermalism. Crustal contamination is difficult to ascertain by Sr and Nd isotopes because most rocks have higher Sr and REE contents than crustal materials. By contrast, Pb isotopes are more sensitive to crustal contamination. The initial 207Pb/204Pb and 206Pb/204Pb ratios (recalculated at 380Ma) define a short linear trend, that could represent a mixing line between a HIMU (or FOZO) type of reservoir (with µ > 10, 87Sr/86Sr < 0.703 and "Nd > 5) and average lower crustal material (µ < 10, 87Sr/86Sr > 0.7045 and "Nd < -4). These magmas have possibly undergone crustal contamination during their ascent through the thick (55km) cratonic crust of Kola.
top...
S. Brassinnes, E. Balaganskaya, D. Demaiffe. 2003. Magmatic evolution of the differentiated ultramafic-alkaline and carbonatite intrusion of Vuorijarvi (Kola Region, Russia): a LA-ICP-MS study of apatite. EUROCARB, Canary.

The small, elliptical Vuorijarvi massif (3.5 x 5.5 km) belongs to Kola Alkaline and Carbonatitic Province. It is a complex multi-phased intrusion classically interpreted as resulting from successive magma pulses. The core of the massif is mainly represented by ultramafic cumulates while alkaline silicate rocks of the ijolite-melteigite series are found at the periphery. Two carbonatitic stocks with associated phoscorites occur at Vuorijarvi: an early one in the eastern part of the massif and a later thick stock (c.a. 800m) in the south-eastern part. Apatite is a common liquidus phase in the three kinds of rocks (ultramafic cumulates, ijolites-melteigites and carbonatites/phoscorites). Apatites from representative clinopyroxenite, ijolite and carbonatites have been analysed by LA-ICP-MS. All apatites are rich to extremely rich in REE (SREE: 730 to 13800 ppm), with LREE enrichment: La/Yb ratio varies from 70 up to 467. Apatite REE composition does not define a regular geochemical evolution trend with emplacement sequence. The crystallisation/fractionation of other REE-rich accessory phases (pyrochlore, perovskite, …) has partly controlled the REE pattern of the apatites. There is an increase in REE content from the clinopyroxenites to the early carbonatites. Intensive crystallisation of apatite and other REE-rich phases deplete the residual melt in REE, the late magnesiocarbonatite are indeed poor in REE (SREE: 120 ppm) and devoid of apatite. REE partition coefficients for apatite crystallising in a carbonatite melt are poorly known. They have been tentatively estimated for some African carbonatites (Homa mountain, Kalkfeld, Ondurakorume, Otjisazu, Okorusu and North Ruri) by comparing the trends of the apatite and of the whole rocks (Bühn et al, 2001). The partition coefficients increase regularly from La to Yb. We use the same approach in this study by combining whole rocks analyses and LA-ICP-MS on apatites. The Vuorijarvi carbonatites display consertal (intergrowth) and polygonal textures. The carbonatites occur as cm to m thick veins suggesting quick cooling of the melt. Calcite crystals (1-2mm) are zoned (CaO: 54.5-56.3wt%; MgO: 0.23-1.9 wt%; SrO: 1.06-0.14wt%) suggesting rapid growth leading to closed-system behaviour. As the zoning of calcite is easily erased at high to medium T° (800-500°C; Cherniak, 1998), the presence of zoned calcite at Vuorijarvi testifies a rapid cooling. Moreover, deformation in calcite involves mechanical twinning rather than pressure-dissolution, so the deformation process has not homogeneised the composition of the grains. We will consider that the REE composition of the liquid in equilibrium with the apatite can be approximated by subtracting the apatite contribution to the whole rocks carbonatite REE content. The apatite/carbonatite liquid partition coefficients for the Vuorijarvi carbonatites are given in Table 1, they range from 7.6 to 3.5 for La, 8.4 to 4.9 for Eu and 5.2 to 1.0 for Yb. The REE distribution coefficient patterns are convex-upwards as the apatite/silicate liquid partition coefficients (Watson and Green, 1981). The REE patterns obtained for Vuorijarvi are completely different from those by Bühn et al (2001). This discrepancy can be explained either by variations of melt composition or by the fractionation of a co-existing LREE enriched phase (i.e. perovskite, pyrochlore).Yttrium and holmium contents are high in apatite from early silicate carbonatites (775 and 33 ppm respectively) and lower in later carbonatites (72 and 2.8 ppm). As Y and Ho have the same charge and nearly the same ionic radius, they do not commonly fractionate. Nevertheless, the calculated DY/DHo ratio (0.89 and 0.68) implies a slight fractionation. Whole rocks analyses plot along a well-defined linear array (R2=0.9884), corresponding to an Y/Ho ratio of 27.5, close to the chondritic value (28.7). Apatite crystallisation would result in an increase of the Y/Ho ratio of the residual melt. The apatite Y/Ho ratio decreases in the sequence clinopyroxenite (27)-ijolite (22.5)-carbonatite (19). The nearly constant Y/Ho ratio of whole rocks implies that apatite alone does not control the Y/Ho fractionation. Two other processes could be invoked: 1) the fractionation of an Y-rich mineral (i.e. monazite) and/or 2) the depletion of Y by exsolution of CO2/H2O fluids (Bau, 1996; Bühn et al, 2001).The apatites of the early clinopyroxenites display slight negative Ce anomaly (Ce/Ce*: 0.7 to 1.0), this anomaly decreases progressively during differentiation, suggesting a decrease in oxygen fugacity. For all the analysed apatites, the Ca (p.f.u) content decreases while the La+Ce+Na (p.f.u) content increases, this feature has already been interpreted in terms of isomorphic substitution 2Ca2+ = REE3++Na+ (Brassinnes et al, 2003), related to fractional crystallisation of a single batch of melt from the ultramafic cumulates to the carbonatites. This is confirmed by the similar Sr and Nd isotopic compositions (0.70303-0.70318; eNd380Ma: +3.5 to 6.1 of all the Vuorijarvi rocks.
REFERENCES:
Brassinnes S.et al (2003) Period. di Mineral, in press; Bau M. (1996) Contrib. Minera. Petrol 123, 323-333; Bühn B. et al (2001), Contrib. Mineral. Petrol. 141, 572-592; Cherniak D.J. (1998), Earth Planet. Sci. Lett. 160, 273-287; McDonough, W.F. and Sun, S.-s. (1995), Chemical Geology 120, 223-253; Watson E.B. and Green T.H. (1981), Earth Planet. Sci. Lett. 56, 405-421.
top...

S. Brassinnes, O. Féménias, N. Coussaert, V. Nivin, D. Demaiffe 2003. Two stages of nepheline crystallisation in the layered part of the Lovozero agpaitic syenite complex (Kola Peninsula, Russia): Implications for the layering and cooling processes. EUROCARB, Canary.

The Lovozero alkaline massif (Gerasimovsky et al, 1966; Kogarko et al, 1995) is composed of three major petrographic units; the main central unit (80% of the volume) is a differentiated layered complex that consists of numerous (29) rhythms composed of several layers: the ideal sequence is urtite-juvite-foyaite-lujavrite. A detailed textural, mineralogical and geochemical study of the complete rhythm II-7 allows to better constrain the mode of emplacement and the differentiation process.
Two distinct types of nepheline have been recognised: 1) inclusion-free euhedral nepheline occurs mainly in urtites and as a minor phase in foyaites and 2) anhedral to subhedral nepheline containing numerous inclusions of aegirine and arfvedsonite occurs in all rock types. The inclusion-free nephelines display rather homogeneous compositions (Ne56-68Ks14-19Q15-25), characterised by high SiO2 (44 to 47.5 %) and Fe2O3 (1.5 to >2 %) contents in comparison with natural nephelines. The nepheline inclusions in aegirine and alkali feldspar are comparable to the inclusion-free nephelines. The inclusion-rich nephelines also show quite homogeneous composition (Ne69-79Ks20-25Q0-11) but are clearly distinct from the inclusion-free nephelines; they are more potassic and less enriched in SiO2 (<43.5 %) and Fe2O3 (< 0.55%). The excess of Si over that required by stoichiometry can be used as a temperature indicator. The inclusion-free nephelines crystallise at significantly higher temperatures (800-1100°C) than the inclusion-rich nephelines (<700°C). The high T° of crystallisation (already noted by Kogarko et Romanchev, 1977) of the inclusion-free nepheline is probably higher than expected for slowly cooled plutons and is attributed to the rapid cooling in a subvolcanic environment (Woolley and Platt, 1986).
The morphological and textural characteristics of each mineral phase in the urtite-lujavrite rocks sequence can be discussed in terms of order of crystallisation and cumulate (solid) and intercumulate (liquid) origin. Alkali feldspar laths are ubiquitous in the sequence; they are dominantly euhedral but have been variously affected by a deformation that occurred contemporaneously with the layering development. Alkali feldspar can be considered, without ambiguity, as a cumulate phase. Inclusion-free high T° nepheline is also a cumulate phase whereas subhedral to anhedral inclusion-rich low T° nepheline is of intercumulate origin. In the rhythm II-7, high T° nepheline is rare, less than 10% of the total nepheline modal proportion; the small (0.2-0.4mm) crystals with cubic-like morphology suggest that nepheline results from the destabilisation of a high T° polymorphic phase (carnegieite?). The inclusion-rich nepheline is common in all the rocks of the rhythm II-7; it represents the main mineral phase of the basal urtite layer of the rhythm. Aegirine-augite constitutes the core of the large poikilitic augite crystals that contain aegirine inclusions, it growths after alkali feldspar and inclusion-free nepheline but before inclusion-rich nepheline. It is tentatively interpreted as a cumulate phase, characterised by a high nucleation rate (numerous crystals have been observed) but a low growth rate. Aegirine and arfvedsonite crystallise lately around the aegirine-augite cores or as large interstitial or poikilitic grains.
Inside the rhythm II-7, strong variations of the cumulate-liquid (L) proportions and of the nature of the liquid have been observed. The liquid proportion, estimated by the proportion of intercumulate phases, decreases dramatically from the basal urtite (L ~ 90%) to the uppermost lujavrite (L ~ 10%). In the urtite, the cumulate phases are represented by rare feldspar laths and inclusion-free nepheline; in lujavrite, euhedral feldspar laths are abundant. The intercumulate assemblage of the urtite is represented by the inclusion-rich nepheline and rare aegirine and arfvedsonite, whereas in the lujavrite, the proportion of the latter increases and are roughly equal to the inclusion-rich nepheline. The modal proportion of alkali feldspar and their size increase from the bottom to the top in the urtite-juvite-foyaite-lujavrite sequence of rhythm II-7. Structural features suggest that each rhythm was formed by in situ differentiation of a sill-like sheet of liquid. The layering could result from a sill-in-sill mode of emplacement rather than from magma chamber processes.

REFERENCES:
Gerasimovsky V. et al (1966) Geochemistry of the Lovozero alkaline massif, Nauka (Moscow), 365pp; Kogarko L. and Romanchev B. (1977) Geokh., 2, 199-216; Kogarko L. et al (1995) Alkaline rocks and carbonatites of the world. Pt2. Former USSR, Chapman et Hall, 225pp; Woolley A.R. and Platt R.G. (1986), Mineral. Mag. 50, 597-610.
top...

2002

Féménias, O., Demaiffe, D., Diot, H., Berza, T., Tatu, M., Liégeois, J.P., Duchesne, J.C. 2002. Petrological data from the Late Pan-African high-K calc-alkaline Motru dyke swarm (Romania): evidence for the evolution in a deep magma chamber from sub-surface observations. Colloque en l’honneur de J.-C. Duchesne. Modelling of Magma Chambers and Implications for the Evolution of the Continental Crust, Liège. Belgique.

The Pan-African basement of the Danubian window (S. Carpathians, Romania) is crosscut by several major dyke swarms. Among these, the subvolcanic Motru swarm intruded between the Late Cambrian (it crosscuts the Tismana pluton dated at 560 Ma) and the Silurian (stratigraphic relation). This swarm represents a large magmatic episode strongly controlled in a geotectonic environment by a late- to post-Pan-African transcurrent tectonics. The swarm consists of 1 to 3 m thick dykes composed of high-K calc-alkaline volcanic to subvolvanic rocks defining a complete differentiated series from andesitic basalt to rhyolite (50–72 wt% SiO2).The most common rocks are aphyric to microgranular porphyritic andesite, with euhedral brown tschermakitic amphibole and/or plagioclase phenocrysts and porphyritic dacite with euhedral green magnesiohornblende and/or corroded quartz phenocrysts. Plagioclase is often deeply retrogressed to secondary phases and appears as ghost crystals. The groundmass is composed of quartz, plagioclase, ferro-magnesian phases (amphiboles and rarely pyroxenes), Ti-magnetite and devitrified glass. Many dykes contain complexly zoned brown amphibole or plagioclase phenocrysts. Back scattered electron images and detailed microprobe investigations on selected amphibole and plagioclase phenocrysts allow to reconstruct the several stages of complex growth and resorbtion history. P-T conditions during this history are recorded by the amphibole chemistry. P estimate is based on the Al-content empirical barometer (Schmidt, 1992) and T estimates are deduced from Ti-content (Otten, 1984). In all rock types, amphibole phenocrysts were equilibrated at about the same pressure, 6±1 kbar, but their temperature of crystallisation range from 900-1000°C for the andesitic basalts to 600-700°C for dacites. In rhyolites, magnesiohornblende phenocrysts present a continuous range of P-T conditions from 700°C/6 kbar to 600°C/1 kbar. These data suggest that a calc-alkaline magma was ponding in a chamber at mid-crustal level (6 kbar, Conrad discontinuity?) and started crystallising. Episodically, dykes were emplaced at higher crustal levels (subvolcanic environment) carrying pre-emplacement phenocrysts of amphibole and plagioclase. At the scale of the Danubian window (90 x 30 km) the regional variations in composition of the dykes suggest that several small chambers evolved independently. Preliminary isotopic compositions (87Sr/86Sr550: 0.7062 – 0.7065 and eNd550: -3 to -4) suggest a juvenile (= probably mantle-derived) magma contaminated by an old crustal component during the ponding and the beginning of crystallisation at mid-crustal level. Our data demonstrate that deep magma chamber processes can be tackled from sub-surface magmatic activity. top...

2001

2000