Holtz, Francois

A correct description and quantification of the geochemical behaviour of REE+Y (rare earth elements and Y) and HFSE (high field strength elements) is a key requirement for modeling petrological and volcanological aspects of magma dynamics. In this context, mafic alkaline magmas (MAM) are characterized by the ubiquitous stability of clinopyroxene from mantle depths to shallow crustal levels. On one hand, clinopyroxene incorporates REE+Y and HFSE at concentration levels that are much higher than those measured for olivine, plagioclase, and magnetite. On the other hand, the composition of clinopyroxene is highly sensitive to variations in pressure,temperature, and melt-water content, according to exchange-equilibria between jadeite and melt, and between jadeite/Ca-Tschermak and diopside-hedenbergite. As a consequence, the dependence of the partition coefficient on the physicochemical state of the system results in a variety of DREE+Y and DHFSE values that are sensitive to the magmatic conditions at which clinopyroxenes nucleate and grow. In order to better explore magma dynamics using clinopyroxene chemical changes, an integrated P-T-H2Olattice strain model specific to MAM compositions has been developed. The model combines a set of refined clinopyroxene-based barometric, thermometric and hygrometric equations with thermodynamically-derived expressions for the lattice strain parameters, i.e., the strain-free partition coefficient (D0), the site radius (r0), andthe effective elastic modulus (E). Through this approach, it is found that the incorporation of REE+Y and HFSE into M2 and M1 octahedral sites of clinopyroxene is determined by a variety of physicochemical variables that may or may not change simultaneously during magma differentiation. The applicability of the P-T-H2O-lattice strain model to natural environments has been verified using clinopyroxene-melt pairs from a great number of volcanic eruptions at Mt. Etna volcano (Sicily, Italy). DREE+Y and DHFSE values recovered by the model have been used as input data to quantify fractional crystallization processes in natural MAM compositions. Results from calculation illustrate that the concentration of REE+Y and HFSE in the magma is primary controlled by the geochemical evolution of clinopyroxene in terms of major cation exchange-equilibria and trace cation lattice strain properties.

We present results from phase equilibria experiments conducted on the most primitive pahoehoe "cicirara" trachybasaltic lava flow ever erupted at Mt. Etna Volcano. This lava is characterized by a pahoehoe morphology in spite of its high content of phenocrysts and microphenocrysts (> 40 vol%) with the occurrence of centimetre-sized plagioclases (locally named cicirara for their chick-pea-like appearance). Our experiments have been performed at 400 MPa, 1100-1150 degrees C and using H2O and CO2 concentrations corresponding to the water-undersaturated crystallization conditions of Etnean magmas. Results show that olivine does not crystallize from the melt, whereas titanomagnetite is the liquidus phase followed by clinopyroxene or plagioclase as a function of melt-water concentration. This mineralogical feature contrasts with the petrography of pahoehoe cicirara lavas suggesting early crystallization of olivine and late formation of titanomagnetite after plagioclase and/or in close association with clinopyroxene. The lack of olivine produces MgO-rich melt compositions that do not correspond to the evolutionary behaviour of cicirara magmas. Moreover, in a restricted thermal path of 50 degrees C and over the effect of decreasing water concentrations, we observe abundant plagioclase and clinopyroxene crystallization leading to trace element enrichments unlikely for natural products. At the same time, the equilibrium compositions of our mineral phases are rather different from those of natural cicirara phenocrysts and microphenocrysts. The comparison between our water-undersaturated data and those from previous degassing experiments conducted on a similar Etnean trachybasaltic composition demonstrates that pahoehoe cicirara lavas originate from crystal-poor, volatile-rich magmas undergoing abundant degassing and cooling in the uppermost part of the plumbing system and at subaerial conditions where most of the crystallization occurs after the development of pahoehoe surface crusts.

The purpose of this review study is to reappraise in a more comprehensive form the thermodynamic principles behind the partitioning of trace elements between clinopyroxene and melt. The original corollary is that the partitioning energetics controlling the crystal-melt exchange are described by two distinct but complementary contributions: ΔGpartitioning = ΔGstrain + ΔGelectrostatic. ΔGstrain is the excess of strain energy quantifying the elastic response of the crystallographic site to insertion of trace cations with radius different from that of the major cation at the site. ΔGelectrostatic is the excess of electrostatic energy requiring that an electrostatic energy penalty is paid when a trace cation entering the lattice site without strain has charge different from that of the resident cation. Lattice strain and electrostatic parameters for different isovalent groups of cations hosting the same lattice site from literature have been discussed in comparison with new partitioning data measured between Tschermak-rich clinopyroxenes and a primitive phonotephritic melt assimilating variable amounts of carbonate material. Through such a comparatively approach, we illustrate that the type and number of trace cation substitutions are controlled by both charge-balanced and -imbalanced configurations taking place in the structural sites of Tschermak-rich clinopyroxenes. A virtue of this complementary relationship is that the control of melt composition on the partitioning of highly charged cations is almost entirely embodied in the crystal chemistry and structure, as long as these crystallochemical aspects are the direct expression of both ΔGstrain and ΔGelectrostatic. A size mismatch caused by cation substitution is accommodated by elastic strain in the surrounding lattice of clinopyroxene, whereas the charge mismatch is enabled via increasing amounts of charge-balancing Tschermak components, as well as the electrostatic work done on transferring the trace cations from melt to crystallographic sites, and vice versa. The influence of the melt chemistry on highly-charged (3+ and 4+) cation partitioning is greatly subordinate to the lattice strain and electrostatic energies of substitutions, in agreement with the thermodynamic premise that both these energetic quantities represent simple-activity composition models for the crystal phase. The various charge-balanced and -imbalanced configurations change principally with aluminium in tetrahedral coordination and the clinopyroxene volume change produced by heterovalent cation substitutions. In contrast, for low-charged (1+ and 2+) cations, the role of melt chemistry cannot be properly deconvoluted from the structural changes of the crystal lattice. The incorporation of these cations into the clinopyroxene lattice depends on the number of structural sites critically important to accommodating network-modifying cations in the melt structure, implying that the partitioning energetics of monovalent and divalent cations are strictly controlled by both crystal and melt properties. We conclude that the competition between charge-balanced and charge-imbalanced substitutions may selectively change the ability of trace elements to be compatible or incompatible in the clinopyroxene structure, with important ramifications for the modeling of natural igneous processes in crustal magma reservoirs which differentiate under closed- and open-system conditions.

The effect of pressure, temperature, and melt composition on CO2 and H2O solubilities in aluminosilicate melts, coexisting with CO2-H2O fluids, is discussed on the basis of previously published and new experimental data. The datasets have been chosen so that CO2 and H2O are the main fluid components and the conclusions are only valid for relatively oxidizing conditions. The most important parameters controlling the solubilities of H2O and CO2 are pressure and composition of melt and fluid. On the other hand, the effect of temperature on volatile solubilities is relatively small. At pressures up to 200 MPa, intermediate compositions such as dacite, in which both molecular CO2 and carbonate species can be dissolved, show higher volatile solubilities than rhyolite and basalt. At higher pressures (0.5 to 1 GPa), basaltic melts can incorporate higher amounts of carbon dioxide (by a factor of 2 to 3) than rhyolitic and dacitic melts. Henrian behavior is observed only for CO2 solubility in equilibrium with H2O-CO2 fluids at pressures <100 MPa, whereas at higher pressures CO2 solubility varies nonlinearly with CO2 fugacity. The positive deviation from linearity with almost constant CO2 solubility at low water activity indicates that dissolved water strongly enhances the solubility of CO2. Water always shows non-Henrian solubility behavior because of its complex dissolution mechanism (incorporation of OH-groups and H2O molecules in the melt). The model of Newman and Lowenstern (2002), in which ideal mixing between volatiles in both fluid and melt phases is assumed, reproduces adequately the experimental data for rhyolitic and basaltic compositions at pressures below 200 MPa but shows noticeable disagreement at higher pressures, especially for basalt. The empirical model of Liu et al. (2004) is applicable to rhyolitic melts in a wide range of pressure (0-500 MPa) and temperature (700- 1200°C) but cannot be used for other melt compositions. The thermodynamic approach of Papale (1999) allows to calculate the effect of melt composition on volatile solubilities but needs an update to account for more recent experimental data. A disadvantage of this model is that it is not available as a program code. The review indicates a crucial need of new experimental data for scarcely investigated field of pressures and fluid compositions and new models describing evident non-ideality of H-C-O fluid solubility in silicate melts at high pressures.

been experimentally quantified via cooling-induced solidification approach. GFA is measured by the critical cooling rate Rc, the rate at which a melt solidifies ≤2 area% of crystals. Cooling rates of 9000, 1800, 180, 60, 7 and 1 °C/h have been run between 1300 °C (super-liquidus region) and 800 °C (quenching temperature), at air fO2 and ambient P for six silicate melts with compositions ranging from basalt (B) to rhyolite (R) (i.e., B100, B80R20, B60R40, B40R60, B20R80 and R100) and water contents comprised between 53 (B100) and 384 (B20R80) ppm. The ranges of cooling rates and chemical compositions used in this study are the broadest ever investigated in the Earth sciences. The phase proportions (area%) were determined by image analysis on about 500 back-scattered electron images collected over different magnifications. Phases are glass, clinopyroxene (cpx), spinel (sp) and plagioclase (plg). Sp is ubiquitous with abundance of few area% and nucleates earlier than silicate crystals. Cpx solidifies in all runs except in R100 and its abundance follows asymmetric broad Gaussian-like trends (with tails towards low rates) as a function of cooling rate. Moving from B100 to B40R60 these trends conserve their shape but shift progressively to lower cooling rates and mineral abundances. Plg crystallises only at low cooling rates and in SiO2-poor compositions. Run-products with low amounts of crystals (≤5 area%) clearly show that cpx preferentially nucleates on surfaces of sp, whereas a significant crystallisation of cpx (N5 area%) is observed with decreasing cooling rate and with changing composition from B100 to B20R80. The crystallisation of silicate crystals is related to the chemical diffusivity of components in the melt. Also the initial crystallisation of plg occurs preferentially on cpx. In general, the amount of crystals decreases as the cooling rate increases; however, in some cases, the amount of crystals remains constant or even decreases for B80R20 with decreasing cooling rate. Rc values change over 5 orders of magnitude being b1, 7, 620, 3020, 8020 and 9000 °C/h for R100, B20R80, B40R60, B60R40 and B80R20 and B100, respectively. The variation of Rc can be modelled through NBO/T (nonbridging oxygen per tetrahedron) parameter by the following equation: Rc=a / {1+e−[(NBO/T − b)/c]},where a, b and c are fitting parameters equal to 9214, 0.297 and 0.040, respectively. Similarly to other glass-forming liquids (network, metallic and molecular systems), Rc for natural sub-alkaline silicate melts is inversely related to the reduced glass transition parameter Trg (Trg=Tg / Tm) and can be quantified with the equation Rc= a × Trg−b, where a and b are 1.19 × 10−4 and 28.7, respectively. These results may be used to retrieve the solidification conditions of aphyric, degassed and oxidised lavas; in addition, our data provide general constrains on the crystallisation kinetics of natural crystal-bearing silicate melts erupted on Earth (e.g. lavas with phenocrysts). The relationship between crystal content and cooling rate suggests that the solidification path induced by degassing can be also complex and nonlinear. The growth of crystalswith size up to 1 mm from a nearly anhydrous superheated silicate melt indicates that variable cooling conditions of lavas have to be accounted to discriminate amongminerals formed before, during and after eruptions.Moreover, our results can be used to design glass-ceramics from naturally available easy to find, low-cost starting materials.

The 3930 BP Fall Stratified (FS) eruption at Mt. Etna is a rare example of a highly explosive eruption of primitive (picritic) magma directly from the mantle. The eruption produced ash plumes up to an estimated 20 km height, leading to a volcanic explosivity index (VEI) 4 (subplinian). Given its volatile-rich and primitive nature, the FS magma may have ascended rapidly from great depths to avoid fractionation and mixing within the extensive plumbing system beneath Etna. To determine the pressures from which the FS magma derived, we perform rehomogenization experiments on melt inclusions hosted in Fo90–91 olivines to resorb shrinkage bubbles and determine the initial H2O and CO2 in the melt. With measured CO2 concentrations of up to 9600 ppm, volatile solubility models yield magma storage pressures of 630–800 MPa. These correspond to depths of 24–30 km, which are comparable to the seismologically estimated Moho. Therefore, the magma’s high CO2 concentration must come from carbon in the mantle (likely from subducted carbonates), as opposed to assimilation of shallow (<10 km) crustal carbonates. Diffusion modeling of H2O and forsterite zonation profiles in clear, euhedral, and crystallographically oriented olivines indicates rapid ascent of magma directly from its source region to the surface. Forsterite profiles exhibit a narrow rim of growth zoning but no detectable diffusional zoning, reflecting maximum ascent times of 1–5 days. Eighteen measured H2O profiles result in remarkably uniform decompression rates of 0.47 MPa/s (95% confidence interval of 0.16–1.28 MPa/s), which is among the fastest measured for basaltic-intermediate magmas. These decompression rates indicate that the final stage of magma ascent over the region in which H2O degasses (between the surface and ~ 15 km) occurred extremely fast at ~ 17.5 m/s. This eruption may provide a link between primary magma composition and eruption intensity: we propose that the unusually explosive nature of this picritic eruption was driven by high H2O and CO2 concentrations, which led to continuously rapid ascent without stalling, all the way from the Moho.

The viscosity of an iron-bearing melt with composition similar to Unzen andesite was determined experimentally in the high (109-1010.5 Pa·s) and low (5-1000 Pa·s) viscosity range using a parallel plate viscometer and the falling sphere method, respectively. Falling sphere experiments were carried out in an internally heated argon pressure vessel and in a piston cylinder apparatus at 1323 to 1573 K and 200 to 2000 MPa. Creep experiments were performed in the temperature range of 747 - 845 K at 300 MPa. The water content of the melt varies from nominally dry to 6.2 wt% H2O. The Fe2+/Fetot ratio was determined for each sample in the quenched glass using a colorimetric method. Pressure has minor influence on the viscosity compared with the effect of temperature, water content (main compositional parameter controlling the viscosity) or with the Fe2+/Fetot ratio (especially important at low water content of the melt). Based on our new viscosity data and literature data with measured Fe2+/Fetot ratio we propose a new empirical equation to estimate the viscosity η (in Pa·s) of andesitic melts as a function of temperature T (in K), water content w (in wt%) and Fe2+/Fetot ratio. The derived relationship reproduces the experimental data (87 in total) in the viscosity range from 100.5 to 1013 Pa·s with a 1σ standard deviation of 0.17 log units. However, application of this calculation model is limited to Fe2+/Fetot>0.3 and to temperatures above Tg. Moreover, in the high viscosity range the variation of viscosity with water content is constrained only by few experimental data and needs verification by additional measurements. The viscosity data are used to interpret mixing processes in the Unzen magma chamber prior to 1991-1995 eruption. We demonstrate that the viscosities of the rhyolite and andesite melts from the two end-member magmas are nearly identical prior and during mixing, enabling efficient magma mixing.

Changes in magmatic assemblages and crystal stability as a response of CO2-flushing in basaltic systems have rarely been directly addressed experimentally, making the role of CO2 in magma dynamics still controversial and object of scientific debate. We conducted a series of experiments to understand the response of magmas from Etna volcano to CO2 flushing. We performed a first experiment at 300 MPa to synthesize a starting material composed of crystals of some hundreds of µm and melt pools. This material is representative of an initial magmatic assemblage composed of plagioclase, clinopyroxene and a water-undersaturated melt with 1.6 wt% H2O. In a second step, the initial assemblage was equilibrated at 300 and 100 MPa with fluids having different XCO 2 fl (CO2/(H2O + CO2)). At low XCO 2 fl (< 0.2 to 0.4), plagioclase is completely dissolved and clinopyroxene show dissolution textures. For relatively high XCO 2 fl (0.9 at 300 MPa), the flushing of a CO2-rich fluid phase leads to an increase of the amount of clinopyroxene and a decrease of the abundance of plagioclase at 300 MPa. This decrease of plagioclase proportion is associated with a change in An content. Our experiments demonstrate that flushing basaltic systems with fluids may drastically affect crystal textures and phase equilibria depending on proportions of H2O and CO2 in the fluid phase. Since texture and crystal proportions are among the most important parameters governing the rheology of magmas, fluid flushing will also influence magma ascent to the Earth’s surface. The experimental results open new perspectives to decipher the textural and compositional record of minerals observed in volcanic rocks from Mt. Etna, and at the same time offer the basis for interpreting the information preserved in minerals from other basaltic volcanoes erupting magmas enriched in CO2.

This review highlights the rheological and phase proportions variation induced by cooling events from superliquidus temperature (melt) to subliquidus temperatures. It provides a comprehensive view of the rheological response of magmatic systems undergoing dynamic cooling and shear deformation. The two main parameters which are of importance to model the rheological properties of such crystallizing systems and which are simultaneously poorly investigated so far are crystallization and strain rates. The response to relatively high deformation rates results in shear thinning behavior in partly crystallized systems under variable shear rate and it should be considered in magmatic processes. Due to the sluggish crystallization of SiO2-rich melts, data are mainly available for mafic systems, which does not allow a general reappraisal. An attempt to model available literature data for less evolved systems in dynamic scenarios and a comparison with MELTS algorithm approach (thermodynamic equilibrium conditions) is provided. Since there are difficulties in comparing experimental data gained using different methodologies, we focus mainly on data obtained with the concentric cylinder technique. This highlights the fact that a general experimental protocol is needed in order to compare and model viscosity data to predict the dynamic rheological evolution for volcanic rocks.

The viscosity of shoshonitic melts from Vulcanello Peninsula (Vulcano Island, Italy) is experimentally determined at temperatures between 733 K and 1673 K. The water content of the melts varies from 0.03 to 4.75 wt% H2O. The micropenetration technique is employed at ambient pressure in the high viscosity range (109-1012 Pa·s). Falling sphere(s) experiments are performed at 500 and 2000 MPa in the low viscosity range (100.5-103 Pa·s). Results show a decrease of about 2 orders of magnitude in viscosity if ~ 3 wt% of water is added to the dry melt at 1300 K. At high temperature the viscosity of Vulcanello melts is intermediate between that of andesitic and basaltic melts. In contrast, at low temperatures (≤1050 K), the shoshonitic melt is characterized by a lower viscosity with respect to the two previous melts. Based on our new data set, a calculation model is proposed to predict the viscosity of the shoshonitic melts as a function of temperature and water content. The viscosity data are used to constrain the ascent velocity of shoshonitic magmas from Vulcanello within dikes. Using petrological data (temperature and crystal content of the magma) and volcanological information (geometrical parameters of the eruptive fissure and depth of magma storage), we estimate the time scale for the ascent of magma from the main reservoir to the surface. Results show time scales in the order of hours to few days. We conclude that the rapid ascent of poorly evolved melts from Moho depths should be taken into account for the hazard assessment of Vulcano Island.