Search

International audience; The rheological and thermodynamic properties of silicate melts control fluid-dynamics of transport, eruption style and rates of physico-chemical processes (degassing and crystallization) in natural magmas. In this study we investigated the effect of H2O and CO2 on the liquid viscosity and heat capacity on several multicomponent systems. Measurements were conducted on four series of melts, obtained by remelting and homogenization of natural pantelleritic (Khaggiar dome, Pantelleria), trachytic (Agnano Monte Spina eruption, Phlegrean Fields), latitic (Fondo Riccio eruption, Phlegrean Fields) and trachybasaltic (Etna 1992 eruption) magmas. CO2 or H2O synthesis experiments were conducted in piston cylinder apparatus. The volatile -bearing samples were measured with a differential scanning calorimeter (DSC) and a vertical dilatometer (micropenetration technique). Water and CO2 content were measured by Karl Fisher Titration and FTIR spectroscopy. Compositions were analyzed over a range of water contents up to 5.3 wt% and CO2 content up to 4000 ppm. Viscosity ranged from 108 to 1013 Pa s with decreasing temperature from 630 to 1100 K. Viscosity is strongly affected by H2O and CO2. The effect of CO2 on viscosity appears to be a function of speciation and chemical composition. The heat capacity of glasses and liquids and the glass transition interval were also investigated. Base chemical composition have a strong influences on Tg: high alkali contents can lower Tg of low NBO/T liquids. This behavior is demonstrated by Pantellerite samples. Glass transition temperatures are also strongly affected by H2O and CO2. The CO2 effect, such as water, is to decrease Tg and it appears to be a function of chemical composition. We present data for partial molar CpH2Omol and CpOH- and derive a simple expression to evaluate the relative contributions of different H-bearing species to the total heat capacity of hydrous melts. Experimental viscosity and calorimetric data were fitted according to the Adam and Gibbs theory in which configurational entropy (Sconf) is the main factor controlling the viscosity of melts. From calorimetric measurements, and assuming that the vibrational contribution to the liquid Cp remains constant above Tg, we determined the configurational contribution to Cpliq and thus calculated the variation of the Sconf as a function of T, H2O and CO2 content in the liquid state. Combining viscosity measurements with the configurational entropies for our liquids, we parameterized the variation of viscosity as a function of temperature and volatiles content within the framework of the Adam and Gibbs theory of structural relaxation.

International audience; The main objective of this experimental study is to constrain and quantitatively model the complex solidification process that transforms a magma in a solid material. Of major interest are crystal nucleation and growth driven by isothermal decompression of hydrous magmas, and comparison with results from more abundant crystal growth/nucleation data obtained in isobaric cooling experiments. This research concerns two different volcanic systems, Pantelleria (peralkaline rhyolite) and Stromboli (basalt), to better understand how crystallization kinetics can affect different magma compositions. For Stromboli volcanic system TZM apparatus has been used to perform decompression runs at Bayerisches Geoinstitut in Bayreuth (DE). As for Pantelleria composition, cooling experiments has been done using IHPV devices at ISTO of Orléans (FR), on the basis of previous phase equilibrium work (Di Carlo et al., 2010). First obtained results for Stromboli case show high rates of nucleation and crystal growth during the initial stages of crystallization which were followed by crystal growth at approximately constant number densities as equilibrium was approached. Shapes of crystals growing in melts are controlled by the kinetics of crystallization and may provide information about the degree of undercooling experienced by batches of magma en route to the surface (Lofgren, 1980). The study of crystallization kinetics through phases growth rates (Couch et al., 2003), together with the calculation of nucleation density and nucleation rates (Hammer et al., 1999) represent a step toward the estimation of the time scales of magmatic processes in volcanic systems and the interpretation of shallow magmatic processes. The results for Stromboli suggest average crystal growth timescales on the order of weeks, but this is complicated by clear evidence that some crystals have experienced repeated periods of both dissolution and growth (Landi et al., 2004).

International audience; We present a freely available and easy-to-use web application called SIGMELTS allowing the calculation of the electrical conductivity of geomaterials at relevant conditions for the Earth's crust and mantle. By compiling previous results of electrical measurements in laboratory, this software enables to discriminate between the effect of different parameters on the bulk conductivity of silicate melts, carbonatites, fluids, minerals and mantle materials, such as the temperature (T), the pressure (P), the composition, the water content, the oxygen fugacity (fO2) and the crystal content. Different existing geometrical models are proposed to calculate the bulk conductivity of a two-phase mixture. Based on the electrical conductivity value of a mantle anomaly, an application has also been developed to determine the corresponding melt fraction at defined conditions (T, P, composition). This web application aims at improving the accessibility to laboratory data in order to precise the interpretation of MT profiles. Although there are examples of where the laboratory data have been used to interpret field data, there are also many instances where there are disconnects between those interpreting field MT data and the laboratory results. SIGMELTS also underlines that new electrical measurements in laboratory are needed to enlarge the present electrical database, particularly at high pressure conditions. An illustration of the use of SIGMELTS will be presented, in which calculations are applied to subduction zone related volcanic zone in the Central Andes. Along with petrological considerations, field and laboratory electrical data allow discrimination between the different hypotheses regarding the formation and rise from depth of melts and fluids and to quantify their storage conditions.

International audience; Bubble nucleation, growth, and coalescence control the ascent, degassing, and eruption of all types of magma. Nucleation and growth are consequences of magma decompression, and are satisfactorily described by solubility and physical models. Coalescence is a more complicated issue, and one that bears heavily on the eruptive fate of an ascending magma. If coalescence is efficient, the resulting permeable networks of interconnected void space may lead to efficient outgassing, and in turn, a reduction in explosive potential. Alternatively, if coalescence is inefficient, bubbles may remain isolated and overpressured, leading to cataclysmic explosive eruptions. Vesicle textures in pyroclasts preserve the only natural evidence of bubble-bubble dynamics. However, most (if not all) pyroclast textures either have lost their gas by bubble collapse, or record syn- to post-fragmentation processes. In both cases, natural products may not always be good indicators of processes at depth. There is thus a gap in our understanding of deep conduit coalescence processes. We present an x-ray computed microtomographic (µ-cT) study of coalescence at the single bubble-bubble pair scale in experimentally and naturally decompressed, low viscosity magmas. Experimentally decompressed phonolites allow us to essentially look "behind" fragmentation, by reproducing deep conduit conditions, and controlling decompression and quenching. Initial results indicate there is no single law of coalescence that can explain all the interactions observed in these low-viscosity magmas, where bubble buoyancy and mobility are key issues. Instead, we identify four different processes, acting concurrently. (1) Traditional capillary +/- gravitational drainage along flattened interbubble films. (2) Static "dimpling," or penetration of one bubble into an adjacent bubble, driven by growth dispersion-controlled differences in internal bubble pressures. (3) Dynamic "handshaking," similar to dimpling, but occurring between buoyantly moving bubbles. (4) "Suction," or entrainment of small trailing bubble into faster-rising larger bubbles. Initial analytical models indicate that in low viscosity magmas, all of these processes act on timescales shorter than the typical timescales of ascent and decompression. Examination of a range of natural submarine and subaerial pyroclasts permits discussion of how each coalescence mechanism is represented in nature, and on the caveats of using textural information in natural pyroclasts to quantify coalescence mechanisms.

International audience; The replenishment by a hot, basic magma is supposed to upset a differentiated magma chamber by magma mixing. The changes in temperature, convection patterns, extent of crystallization, or the vesiculation of the re-heated felsic magma may be all factors that ultimately trigger eruption. However, only little experimental work has been performed so far on mixing/mingling processes, and the existing data base concerns essentially atmospheric pressure. In order to characterize magma mixing, hybridization and related events, we performed experiments juxtaposing two contrasted magma compositions. We have chosen to use a synthetic haplotonalitic composition (SiO2: 68.9%, Al2O3: 19.6%, Na2O: 8.0%, CaO: 3.5%) as the felsic end-member, and natural basalt from Santorini Volcano (Greece). The geometries adopted were (1) a mafic core plug inserted into an outer hollow felsic cylinder and (2) two basaltic disks separated by a felsic one. Then, we simulated a simple shear deformation using a Paterson HP-HT apparatus, at 300 MPa confining pressure, and temperatures from 900 to 1200°C. The temperature range allowed us to get the lowest viscosity contrast at which magma mixing is enhanced. Under these conditions calculated relative viscosities of both felsic and mafic end-members range from Log η = 8.2 to Log η = 1.3 Pa.s, depending on temperature. However, crystal content changes considerably the apparent viscosity. Applied shear strain rates range from γr = 7x10-5 s-1 to γr = 1x10-3 s-1. Finite shear strains (γ) range between 0 (static experiments) and 5.4. SEM images on run products show textures and particular fabrics such as enclave formation, mafic material dislocation, whereas microprobe analyses reveal basalt melting and mixing with the felsic end-member producing andesitic melts. In addition, close to the contact between the two materials, basaltic phenocrysts are replaced by a new generation of crystals, different from plagioclases growing from the haplotonalite only. The comparison with natural textures from e.g. Santorini volcano will constrain the conditions required for magma mixing at this volcano. Experiments under hydrous conditions relevant to the Santorini context are in progress. These experiments will help us to find out the consequences of such a common process partly responsible of explosive eruptions in volcanic arcs.