Unravelling the anatectic history of the lower continental crust through the petrology of melt inclusions and lu-hf garnet geochronologyA case study from the western alpujárrides (Betic Cordillera, s. Spain)

  1. Barich, Amel
Dirigida por:
  1. Antonio Acosta Vigil Codirector/a
  2. Carlos J Garrido Marín Codirector/a

Universidad de defensa: Universidad de Granada

Fecha de defensa: 21 de diciembre de 2015

Tribunal:
  1. María Teresa Gómez Pugnaire Presidente/a
  2. Fernando Gervilla Linares Secretario/a
  3. Robert Anczkiewicz Vocal
  4. Carlos Villaseca González Vocal
  5. Bernardo Cesare Vocal

Tipo: Tesis

Resumen

The main aim of this thesis is to uncover the processes of lower crustal anatexis through the study of nanogranites inclusions in metamorphic minerals in lower-crust gneisses, and their potential relationships with lithospheric mantle tectono-magmatic processes. With this aim, in this Ph.D. Thesis I investigated gneisses from the polymetamorphic Jubrique crustal sequence ¿a complete though strongly thinned crustal section in the westernmost Alpujárrides (Betic Cordillera, S. Spain). These gneisses overlie the ultramafic Ronda peridotite massif ¿the largest exposure of subcontinental mantle on Earth¿ and provide a unique opportunity to investigate the nature and age of crustal melting events and their timing with lithospheric mantle processes in the westernmost Mediterranean. The study of the composition of primary melts during anatexis of high-pressure granulitic migmatites is relevant to understand the generation and differentiation of continental crust. Peritectic minerals in migmatites can trap droplets of melt that form via incongruent melting reactions during crustal anatexis. These melt inclusions commonly crystallize and form nanogranitoids upon slow cooling of the anatectic terrane. To obtain the primary compositions of crustal melts recorded in these nanogranitoids ¿including volatile concentrations and information on fluid regimes¿ they must be remelted and rehomogenized before analysis. In this thesis, I report a new occurrence of melt inclusions in polymetamorphic granulitic gneisses of the Jubrique unit. The gneissic sequence is composed of mylonitic gneisses at the bottom and in contact with the peridotites, and porphyroblastic gneisses on top. Mylonitic gneisses are strongly deformed rocks with abundant garnet and rare biotite. Except for the presence of melt inclusions, microstructures indicating the former presence of melt are rare or absent. Upwards in the sequence garnet decreases whereas biotite increases in modal proportion. Melt inclusions are present from cores to rims of garnets throughout the entire sequence. Most of the former melt inclusions are now totally crystallized and correspond to nanogranites, whereas some of them are partially made of glass or, more rarely, are totally glassy. They show negative crystal shapes and range in size from ¿5 to 200 micrometers, with a mean size of ¿30-40 micrometers. Daughter phases in nanogranites and partially crystallized melt inclusions include quartz, feldspars, biotite and muscovite; accidental minerals include kyanite, graphite, zircon, monazite, rutile and ilmenite; glass has a granitic composition. Melt inclusions are mostly similar throughout all the gneissic sequence. Some fluid inclusions, of possible primary origin, are spatially associated with melt inclusions, indicating that at some point during the suprasolidus history of these rocks granitic melt and fluid coexisted. Thermodynamic modeling and conventional thermobarometry of mylonitic gneisses provide peak conditions of ¿850 ºC and 12-14 kbar, corresponding to cores of large garnets with inclusions of kyanite and rutile. Post-peak conditions of ¿800-850 ºC and 5-6 kbar are represented by rim regions of large garnets with inclusions of sillimanite and ilmenite, cordierite-quartz-biotite coronas replacing garnet rims, and the matrix with oriented sillimanite. Previous conventional petrologic studies on these strongly deformed rocks have proposed that anatexis started during decompression from peak to post-peak conditions and in the field of sillimanite. The study of melt inclusions shows, however, that melt was already present in the system at peak conditions, and that most garnet grew in the presence of melt. To further constrain the primary composition and the P-T conditions of formation of nanogranitoids, I have carried out a detail experimental study. Nanogranitoids within separated chips of cores and rims of large garnets from were remelted at 15 kbar and 850, 825 or 800 ºC and dry (without added H2O), during 24 hours, using a piston cylinder apparatus. Although all experiments show glass (former melt) within melt inclusions, the extent of rehomogenization depends on the experimental temperature. Experiments at 850-825 ºC show abundant disequilibrium microstructures, whereas those at 800 ºC show a relatively high proportion of rehomogenized nanogranitoids, indicating that anatexis and entrapment of melt inclusions in these rocks occurred likely close to 800 ºC. Electron microprobe and NanoSIMS analyses show that experimental glasses are leucogranitoid and peraluminous, though define two distinct compositional groups. Type I corresponds to K-rich, Ca- and H2O-poor leucogranitic melts, whereas type II represents K-poor, Caand H2O-rich granodioritic to tonalitic melts. Type I and II melt inclusions are found in most cases at the cores and rims of large garnets, respectively. We tentatively suggest that these former migmatites underwent two melting events under contrasting fluid regimes, possibly during two different orogenic periods. This study demonstrates the strong potential of melt inclusions studies in migmatites and granulites in order to unravel their anatectic history, particularly in strongly deformed rocks where most of the classical anatectic microstructures have been erased during deformation. One of the main issues that prevents establishing reliable P-T-t paths to better constrain the potential geodynamic scenarios for the tectonic and metamorphic evolution of Alborán basement high-grade gneisses is the difficulty of relating zircon and monazite thermochronological ages with the crystallization of major mineral assemblages and, in particular, garnet and its MI. Another important issue to unravel the geodynamic evolution of this area is to establish the temporal links between the P-T-t paths of the mantle section with those of the rocks from its crustal envelope. To shed some light on the age of crystallization, we have analyzed Lu-Hf in whole rock and several garnet core and rim fractions, and have obtained isochrons in five samples of porphyroclastic and porphyroblastic gneisses of the Jubrique unit, and in three garnet pyroxenites from the garnet-spinel mylonites and the spinel tectonite domain of the Ronda peridotite. Multipoint Lu-Hf isochrons of whole rock and different garnet core and rim fractions in mylonitic and porphyroblastic gneisses of Jubrique unit yield ages mostly comprised between the Early Permian and Early Jurassic Periods. Multipoint isochrones of garnets in Jubrique mylonitic gneisses provide Early Permian (289 Ma) to Middle- Late Triassic (236 Ma) ages, except for the combination of whole rock and rims in a sample that yields an age of 193 Ma. Multipoint isochrones of garnets in porphyroblastic gneisses provide Early-Middle Triassic (248 Ma) to Early Jurassic (191 Ma) ages, except for a two-point isochron made of whole rock and a single garnet from JU-20, yielding an age of 129 Ma. We interpret the oldest ages as recording garnet formation during de Variscan orogeny. We interpret the Lu-Hf ages between ¿ 260-190 Ma and ¿129 Ma as due to partial reset of the Lu-Hf system by post-Variscan tectonometamorphic events. The Lu-Hf geochronology of Jubrique gneisses, together with previous thermochronological U-Pb studies in accessory minerals, highlights the polymetamorphic nature of high-grade rocks in the crustal envelope of the Betic-Rif peridotites. The Lu-Hf ages of the cores of garnet porphyroclasts and porphyroblasts confirm that the growth of garnet in Jubrique gneisses occurred in the Early Permian during the latest stages of the Variscan orogeny. This inference poses major problems to the interpretation of the Alpine decompression P-T-t paths in the high-grade Jubrique gneisses, which rely on peak metamorphic conditions registered in the inclusions and the mineral chemistry of garnet cores. The presence of MI in garnet cores ¿which we have demonstrated experimentally to be formed c. 800 ºC¿ in equilibrium with kyanite and rutile strongly supports a high-P and high-T (1.2¿1.4 GPa and ¿ 850 °C) melt-present environment during the growth of garnet in the Hercynian. These conditions are akin to that of granulitic rocks formed at the base of thickened continental crust during continental collision. As deduced by previous authors in equivalent gneisses of the Rif belt, we interpret that the anatexis and crystallization of garnet cores in the Jubrique unit took place in the Early Permian in a context of continental collision and overthickened continental crust. Despite the U-Pb-Th thermochronological evidence for a sequence of Alpine events in the Jubrique units, we have found no Alpine Lu-Hf ages indicating either that this event is not resolvable with the present sampling and dating techniques, or that the Lu-Hf of garnet was not equilibrated during the Alpine orogeny. A possible interpretation of the Lu-Hf isochrons is that most of the garnet and fabrics of the Jubrique gneissic sequence is of Permian and/or Jurassic age. However, this interpretation is at odd with U-Pb-Th dating of zircon and monazite in garnet rims and matrix of the Jubrique gneissic sequence that yielded Alpine ages ranging from 34 to 20 Ma, or younger. These results indicate that garnet growth in Jubrique gneisses is polymetamorphic and grew in at least in two different geodynamic events; the latest event also involved anatexis of the sequence as attested by the presence of Ca-rich inclusions in garnet rims. The slow diffusion of REE and Hf in garnet may have preserved primary garnet growth ages without being reset in the Alpine orogeny at least at the scale we sampled them. Further detailed studies relating microstructure with high spatial resolution thermochronology are needed to unravel the role of the Alpine orogenic event in the development of the magmatic and ductile microstructures of the Jubrique gneissic sequence. The Lu-Hf whole rock and garnets isochrons of Ronda garnet pyroxenites provide ages of the Jurassic-Cretaceous limit (144 Ma), the Lower-Middle Paleogene (53 Ma) and the Lower Neogene (21 Ma) Periods. We interpret Early Miocene ages as recording the waning stages of the extensional-related thermal event leading to the melting of the base of the lithospheric mantle peridotites, the development of the Ronda peridotite recrystallization front, and the final intracrustal emplacement of the Betic-Rif peridotites. In this interpretation, the Lu-Hf ages of garnet pyroxenites in the Ronda garnet-spinel mylonite domain were not reset during this thermal event because this domain was substantially cooler than the rest of the peridotite massif, which allowed the preservation of Lu-Hf ages from early orogenic events since the Mesozoic. A striking difference between the geochronological results in peridotites and rocks from their crustal envelop is that the Lu-Hf ages in rocks of the mantle section do not record any Variscan event, which is otherwise widespread in the Lu-Hf and U-Pb-Th thermochronological record of the overlying high-grade crustal rocks. This may reflect either that this Lu-Hf ages might have been reset by later mantle events or that garnet in mantle rocks grew in geodynamic events later than the Variscan. On the other hand, the coincidence of U-Pb-Th age in zircon mantle pyroxenites and of garnet Lu-Hf ages in the overlying crustal sections most likely reflects extraordinary thermal events that partially melted garnet pyroxenite in the mantle lithosphere and induced anatexis in the overlying lithospheric crustal section during the Variscan orogeny.