RICERCA ITALIANA     

Physical and chemical properties of volatile-bearing silicate melts: experiments, modeling and application to volcanic degassing.

Università degli Studi Roma Tre

Abstract

Magmatic processes are governed by transport and thermodynamic properties of the silicate melts and minerals constituting the magma. The properties that most affect the fluid-dynamic behaviour of magmas are viscosity, density and calorimetric properties and all the first (enthalpy, entropy, V) and second (Cp, adiabatic compressibility, isobaric thermal expansivity) order thermodynamic variables associated with them.
Although the experimental data base for many properties is growing, these data are often from chemically comparatively simple systems. Therefore, direct experimental determination and modelling of such properties for natural magma compositions to understand magmatic processes is of extreme importance.
The project has three main objectives:
1)Modeling of transport and thermodynamic properties of natural magmas and of the associated volatile phases. To do so, the following tasks will be addressed:
a.Modelling of the rheology of magmas as a function of T, volatile and crystal content and deformation regime:
b.Implementation of the current construction of PVT equations of state for silicate melts to include for the effect of volatiles;
c.Implementation of the current parameterization of the Cp and Cpconf of silicate glasses and liquids to include for volatile components.
2)Development of a computational tool able to depict the thermo-chemical features of chemically complex melts in a wide P-T regime. In particular, the polymeric model of melt reactivity will be applied to the modelling of:
a.Volumetric properties in a wide P-T regime;
b.Viscosity of silicate liquids in terms of Sconf;
c.Multicomponent solubility;
d.Volatile diffusivity in silicate melts.

3)Development of a model of volatile degassing during magma ascent, which includes solubility (2a, 2b), volatile diffusivity (2d) and nucleation and growth models. The modelling will be based on the polymeric model HPA, extended to minor and trace elements (obj. 2) and on the experimental work (obj. 1).

The originality of this project stems from the application of several non conventional geological geochemical and statistical methodologies. Its success is therefore related to its multidisciplinary character and the cooperation and synergy among all of the different units.
Five units will participate in the project accounting for the following different methodological approaches:

-Experimental techniques at high T and P (Dilatometry, Calorimetry, Rheometry, ultrasonic interferometry ) and high P - T syntheses apparatus (IHPV)
-Spectroscopic techniques (FTIR, KFT, EPMA, SEM)
-Experimental methodologies of in situ gas analyses (Multi-GAS and Filter Pack)
-Statistical methodologies
-Thermodynamics, with numerical calculation techniques (mixing properties (Gmix) and gas-melt equilibria).
-Computational techniques for differential equations of nucleation and bubble growth.

The following results will be delivered:
Unit 1 (Romano). Parameterization of the transport and thermodynamic properties of silicate melts and of the associated volatiles. This task will be achieved through a careful experimental work aimed to investigate: (1) viscosity of magmatic liquids and mixtures (liquid+solid; liquid+bubbles) and multiphase rheology of volatile-bearing volcanic materials as they undergo viscous and brittle deformation at different deformation regimes, (2) Cp of volatile-free and volatile-bearing glasses and liquids and (3) thermal expansivity and compressibility of glasses and liquid in a wide P-T regime.
Unit 3 (Ottonello) Development of a computational tool able to depict the thermochemical features of chemical complex silicate melts in a wide P-T regime. This work will benefit from the new thermodynamic data, at different P and volatile content, from UR 1 for the interactions between components in the liquid fase (Gmix).
Unit 2 (Moretti) Development of a degassing model in the melt+gas±crystals system. The paramaterization will include: (1) modelling of the volatile solubility and diffusivity by use of the theoretical model of UR 3 and the experimental database of UR 1 and UR 5; (2) modeling of bubble nucleation and growth on the basis of new experimental data of UR 1 (multiphase rheology) and UR 2 (chemical-textural data on natural samples) and UR 5 (plume degassing).
Unit 4 (Buccianti) Achievement of new and innovative methods for modelling and representing theoretical and experimental multivariant data in the appropriate sample space, or simplex, to be applied to analyses of saturation PT conditions of volatiles (UR 2, 5) and the parameterization of the transport and thermodynamic properties (UR 1) of volatile-bearing melts.
Unit 5 (Valenza). Validation of the general model for gas-melt interactions by integrating petrologic, geochemical and theoretical data on basaltic volcanoes, and allowing for the composition of the volcanic gas phase emitted to be quantitatively interpreted in terms of dynamics of magma, ascent, accumulation and eruption. <<<

Principal Investigator

Claudia Romano Università degli Studi ROMA TRE

Research Objectives

This research project has three main objectives:

1)Modeling of the transport and thermodynamic properties of natural magmas and of the associated volatile phases, which govern volcanic and magmatic processes. To do so, the following tasks will be addressed:
a.Modelling of the rheology of magmas as a function of T, volatile and crystal content and deformation regime:
b.Implementation of the current contruction of PVT equations of state for silicate melts to include for the effect of volatiles both at 1 atm and at high P, low and high T;
c.Implementation of the current parameterization of the Cp and Cpconf of silicate glasses and liquids to include for the volatile components.

2)Development of a computational tool able to depict the thermo-chemical features of chemical complex silicate melts in a wide thermo-baric regime. In particular, the polymeric model of silicate melt reactivity will be applied to the modelling of:
a.Volumetric properties in a wide thermo-baric regime;
b.Viscosity of silicate liquids in terms of Sconf;
c.Solubility in the multicomponent system CO2-H2O-H2S-Cl-F-silicate melt;
d.Diffusivity of volatiles in silicate liquids.

3)Development of a model of volatile degassing during magma ascent, which includes multicomponent solubility (2a, 2b), volatile diffusivity (2d) and nucleation and growth models. This modelling will be based on the polymeric model HPA, extended to minor and trace elements (obj. 2) and on the experimental work (obj. 1).

The achievement of these objectives will require close collaboration among all URs. Integration of the various methodologies represents a key element to the success of the project. Each UR brings its own expertise to at least one of the methodological approaches, highlighting the importance of each unit’s participation as well as the need for coordination and teamwork between units in order to realize the overall aims of the project. We would also like to stress the close connection between the proposed objectives and the individual activities of the research units. For instance, the development of the polymeric model of silicate melts (obj. 2) requires the experimental database of obj. 1. The degassing model (obj. 3) is based on the implementation and refinement of the general theoretical model of obj. 2 and on the volumetric, solubility and viscosity data which will be produced by UR 1 and 5.

Unit 1 (Romano) has as its main objective the parameterization of the transport and thermodynamic properties of silicate melts and associated volatiles phases. This objective will be reached through careful experimental work dedicated to investigate: (1) the viscosity of magmatic liquids, magmatic mixtures (liquid + solid; liquid + bubbles) and the multiphase rheology of volatile-bearing volcanic materials as they undergo viscous (flow) and brittle (fracturing) deformation at different deformation regimes, (2) Cp of volatile-free and volatile-bearing glasses and liquids and (3) thermal expansivity and compressibility of glasses and liquids in a wide thermo-baric regime. The experimental work will be carried out using a series of innovative methodologies (low and high T rheometry, dilatometry, calorimetry, ultrasonic interferometry), available at the newly born Laboratory of Experimental Volcanology at the University of Rome 3 and in collaboration with the INGV (Rome) and the University of Munich (Prof. Don B. Dingwell).

Unit 3 (Ottonello) has as its general objective the development of a computational tool able to depict the thermo-chemical features of chemically complex silicate melts in a wide thermo-baric regime, with particular attention given to mixing properties (Gmix). The theoretical model of silicate melt reactivity based on the polymeric approach, at present restricted to eight major components, will be extended to 13 minor components and to 20 trace elements. The implemented model will then be applied to the parameterization of the PVT equations of state for volatiles in a wide thermo-baric regime (see 1b) and the viscosity of silicate melts in terms of Sconf (see 1a). This work will benefit from new thermodynamic data, at different P and volatile content, obtained by UR 1.

Unit 2 (Moretti) has as its main objective the development of a degassing model either under equilibrium (gas pressure = external pressure) or non-equilibrium conditions (onset of overpressures) in the silicate melt+gas±crystals system, as a function of dP/dT/dt gradients, phase proportion, relative amount and concentration of volatiles, and redox conditions. The activities of UR 2 consist of two tasks: (1) modelling of the volatile solubility in the system CO2-H2O-H2S-Cl-F-silicate melt by the use of the theoretical model constructed by UR 3. This parameterization will benefit from the experimental measurements of partial molar volumes of volatiles and the configurational contributions of Cp for the H-bearing species (UR1) as well as Cl and F solubility data obtained by UR 5 (Valenza); (2) modeling of bubble nucleation and growth on the basis of (a) new experimental data on the multiphase rheology (liquid+crystal+bubbles) of natural melts (UR 1); (b) new volatile diffusivity model based on the polymeric approach and (c) new chemical-textural data on natural samples obtained by UR 2 and UR 5 (Valenza). In parallel, UR 2 will also carry out chemical (volatile concentrations and stable isotope analyses) and textural analyses of natural samples in order to investigate correlations between degassing, decompression and crystallization.

Unit 4 (Buccianti) has as its main objective to propose and implement new and innovative methods for the mathematical-statistical modelling of the compositional variations in natural silicate melts and their associated volatile phases within an appropriate sample space (simplex) given the experimental data and theoretical constraints imposed by URs 1,2,3 and 5. The understanding of the mechanisms that drive natural phenomena is highly related to the adoption of a correct statistical modelling procedure. The definition of saturation PT conditions of volatiles (UR 2, 5) and the parameterization of the transport and thermodynamic properties (UR 1) will be made easier with the algebraic techniques developed by this unit.

Unit 5 (Valenza) has the specific objective of testing and validating the general model for gas-melt interactions by integrating petrologic, geochemical and theoretical data, and allowing for the composition of the volcanic gas phase emitted by volcanoes to be quantitatively interpreted in terms of magma dynamics, ascent, accumulation and finally eruption. In order to achieve these objectives, UR 5 will coordinate its activities according to the following three steps:
(1) Acquire a comprehensive and systematic data-set of volcanic gas observations, by means of direct field surveys, on-site measurements and laboratory measurements.
(2) perform new laboratory experiments for the derivation of Cl and F partitioning between a vapor fluid phase and a basaltic melt;
(3) Review melt inclusion records of pre-eruptive halogen contents in natural basaltic magmas, with special reference to Mount Etna (in collaboration with UR 2).
The solubility measurements of step (2) will be used directly by UR 2 to build and validate the CO2-H2O-H2S-Cl-F-solubility model for silicate melts.

The bulk of the data regarding composition and vesiculation from natural samples (UR2) will be compared and integrated with the results from the study of volcanic plumes and fumaroles (UR 5) in order to validate the thermodynamic model of volatile degassing produced by UR 2, allowing for i) definition of the main dynamic processes responsible for gas transport to the surface and ii) budgeting of the bulk of volatiles involved in the different volcanic systems of interest. <<<

First Results

As described above, the general expected results from this research consist of (1) a parameterization of the thermodynamic and transport properties of magmas, (2) a general model for the chemical reactivity of silicate melts, and (3) on the theoretical basis of this model, a model for the degassing of volatiles during magma ascent along a volcanic conduit. These expected results, in a sense, are the product of the individual results of each research unit. Additionally, the various individual results of each unit and the combined product have potentially strong applications. Attainment of these results will require a strongly concentrated collaboration and cooperation between all of the research units involved in the project.

UR 1 (Romano) will provide, as a result of their study, a general parameterization of transport and thermodynamic properties of silicate melts and of the associated volatile phases. To achieve this task, this unit will execute a careful experimental work aimed at investigating: (1) the viscosity of magmatic liquids and magmatic mixtures (liquid + solid; liquid + bubbles) and (2) the multiphase rheology of volatile-bearing volcanic materials as they undergo viscous (flow) and brittle (fracturing) deformation at different deformation regimes; (3) Heat capacity of volatile-free and volatile-bearing glasses and liquids; (4) thermal expansivity and compressibility of glasses and liquids in a wide thermobaric regime. The experimental work will be performed using innovative and original methodologies (low and high-T rheometry, dilatometry, calorimetry, and ultrasonic interferometry) available at the newly born Laboratory of Experimental Volcanology at the University of Rome 3 and in collaboration with the INGV (Rome) and the Department of Earth and Enviromental Science of the University of Munich (Prof. Dingwell). As a result of this experimental work, this unit will provide some important constitutive equations, which will be of general interest and application to the earth sciences for understanding the physical and chemical behaviour of magmas and many geological processes, such as partial melting, magma rising through the crust, differentiation processes such as crystallization or mixing, advective heat transport, exsolution and degassing, fragmentation and eruption style. In detail, the unit will provide at the end of the project:

1)Constitutive equations of the rheology of magmas as a function of T, volatile and crystal content and deformation regime:
2)Implementation of the current construction of PVT equations of state for silicate melts to include for the effect of volatiles both at 1 atm and at high P, low and high T;
3)Implementation of the current parameterization of the Cp and Cpconf of silicate glasses and liquids to include for the volatile components.

We emphasize that the results from these experimental investigation will not only be important to the intended application of volatile degassing in natural silicate liquids, but will also have potential implications to other areas in the geosciences as well as in the materials sciences.

Unit 3 (Ottonello) has as its general objective the development of a computational tool able to depict the thermochemical features of chemically complex silicate melts in a wide thermo-baric regime.
The theoretical model of reactivity of silicate melt based on the polymeric approach, at present restricted to the eight major components, will be extended to 13 minor components and to 20 trace elements. This work will benefit from new thermodynamic data, at different P and volatile content obtained by UR 1.This implemented model will then be applied to the parameterization of the PVT equation of states for volatiles in a wide thermobaric regime (see 2) and the viscosity of silicate melts in terms of Sconf (see 1). The polymeric model, besides a strong application within this project, has a strong general feature which aims at a quantitative understanding of all the energetic and reactive properties of silicate melts. The quantitative appraisal of thermodynamic properties of silicate melts has been a fundamental research task joining the efforts of metallurgists and geochemists for more than half a century and also bears a strong application in the material sciences. From a practical standpoint, this research unit will provide public domain software for the quantification of the thermodynamic and reactive properties of complex aluminosilicate melts as well as graphic visualization of the results. The resulting software will be accessible via the internet, allowing fruition of the model to the entire scientific community.

Unit 2 (Moretti) will provide as the main result the development of a degassing model either under equilibrium or non-equilibrium conditions in the silicate melt+gas±crystals system, as a function of dP/dT/dt gradients, phase proportion, relative amount and concentration of volatiles, and also redox conditions. This modelization will be completed in two steps:
The whole activity on UR 2 is made up of two tasks: (1) modelling of the volatile solubility in the system CO2-H2O-H2S-Cl-F-silicate melts by the use of the theorethical model constructed by UR 3. This parametrization will benefit of the experimental measurements of UR1 and of solubility of Cl and F data performed by UR 5. (2) modeling of bubble nucleation and growth on the basis of (a)new experimental data on the multiphase rheology (liquid+crystal+bubbles) of natural melts (UR 1); (b) new volatile diffusivity model based to the polymeric approach and (c) new chemical-textural data on natural samples obtained by UR 2 and UR 5 (Valenza).

The chemical-textural information gathered from analyses of natural samples contribute to the understanding of the characteristics of eruptive magmas in an explosive manner, not only towards eruptive dynamics but also towards the study of processes during magma ascent, as well as characterizing the variation of rheological properties of the magma during eruption. Thus, these studies will provide implications for the definition and mitigation of volcanic risk. We also highlight the importance of microtextural studies, which represent an approach not only applicable to the earth sciences but also to fields of materials science and engineering.

Unit 4 (Buccianti). The main result of this unit will be the acquisition of new and innovative methods for modelling and representing theoretical and experimental multivariant data in the appropriate sample space, called a simplex. In fact, the understanding of the mechanisms that drive natural phenomena is highly related to the adoption of a correct statistical modelling procedure. The definition of saturation PT conditions of volatiles (UR 2, 5) and the parameterization of the transport and thermodynamic properties (UR 1) will be made easier with the algebraic techniques developed by this unit. The theoretical advancements made in this field of research have numerous applications throughout the earth sciences by its treatment of compositional data with multivariant techniques.

Unit 5 (Valenza). The results of this unit will comprise: i) definition of the main dynamic processes responsible of gas transportation to surface and ii) budgeting of the bulk of volatiles involved in the different volcanic systems of interest
In order to provide these results, the UR 5 will coordinate its activities among the following three steps:
(1) Acquire a comprehensive and systematic data-set of volcanic gas observations, by means of direct field surveys, on-site measurements and laboratory measurements.
(2) perform new laboratory experiments for the derivation of Cl and F partitioning between a vapor fluid phase and a basaltic melt;
(3) Review melt inclusion records of pre-eruptive halogen contents in natural basaltic magmas, with special reference to Mount Etna (in collaboration with UR 2);
The solubility measurements of step (2) will be directly used by UR 2 to build and validate the solubility model CO2-H2O-H2S-Cl-F-silicate melts.
Results from the study of volcanic plumes and fumaroles of Etnean volcanism will be directly applied to the validation of the thermodynamic model of volatile degassing produced by the joint work by IR 1-4 so allowing for the i) definition of the main dynamic processes responsible of gas transportation to surface and to ii) budgeting of the bulk of volatiles involved in the different volcanic systems of interest.

The final obtained model obtained will be applicable to all compositional cases recognized in nature, at pressures up to approximately 10 kbar, thus promising enormous applicative potential to terrestrial magmatism. It will be possible to simulate both equilibrium and kinetic volatile fractionation, to determine the budget of volatiles involved in any volcanic scenario or in any geodynamic setting in general. Additionally, it will permit the calculation of timescales associated with magmatic processes (magma ascent, replenishment of magmatic reservoirs, and overpressure timescales) with obvious implications for hazard prevention. Other important implications concern regional and global environmental issues (e.g. volcanic forcing), with scenarios of enhanced injection of volcanic gases into the stratosphere upon eruption. Fluorine, chlorine, bromine and sulfur in particular, exhausted by volcanoes, affect the absorption of solar energy in the stratosphere through formation of sulfate aerosols and therefore may affect the atmospheric circulation and the average surface temperatures. Thus modelizations from a regional scale to the global scale can be made, not only for an eruptive scenario, but also for persistent degassing which also affects the atmospheric environment. For example, when released into the troposphere by passively degassing quiescent volcanoes, volcanogenic S and halogens contribute to produce acid rains and air pollution at the local and regional scales, and may give rise to anomalous wet and dry depositions adversely impacting the local ecosystems. This is particularly critical in the Italian region, densely populated by volcanoes, with relatively high gas emissions. Thus, results of this research will constitute a comprehensive model of how volcanoes degas before and during eruptions and will provide a tool to better understand effects, which can be coupled to numerical models of plume formation, evolution, and dispersion in the atmosphere. We strongly believe that a magmatic degassing model is a preliminary but unconditional step on the long path to the evaluation of volcano-driven climate change

Finally, we strongly emphasize that the expected products of this research project can be utilized beyond the disciplines of the earth sciences, but also have potential application to industrial sectors, such as the glass and metallurgic sciences. Modern techniques on refining glassy materials based on foaming are very similar to the process of degassing observed in nature, necessitating the same approaches used here to the solubility and diffusivity of components in relation to different compositional types, i.e. their mixing properties must be known. An example is that of the stripping of CO2 for SO2 production, which is strongly correlated to the redox conditions of the system. A similar example is found in metallurgy for which steel refinement is highly sensitive to the slag type, which must be a strongly basic, depolymerised silicate melt. In fact, the subtraction of sulfur from steels, which is able to “tune” the physical and mechanical properties of the material, is a familiar problem in the industry, related to S partitioning between the different equilibrium phases present in the slag. A slag of high S solubility not only makes the quality of the steel higher, but the refinement process more efficient for achieving a given type of material. <<<

Timescale

24 months

National and international background

Magmatic and volcanic processes (e.g., partial melting, formation and ascent of magma, crystallization/degassing kinetics, fragmentation and eruptive style) are governed by transport properties (e.g., density, viscosity) and properties defining the energy exchanges among the phases of magmatic silicate mixtures. In particular, the density and the rheological properties (e.g., viscosity) control the buoyancy forces, fluid-dynamics of transport, eruptive style and rates of physicochemical processes (degassing, crystallization) in natural magmas [1-3]. On the other hand, the calorimetric properties of first (H, S, V) and second order (Cp, adiabatic compressibility, thermal expansivity) give us information about the energy budget and their variation with thermodynamic conditions associated to the magmatic mixtures.
As a consequence, understanding the mechanisms that control the above mentioned processes, and their impact on the environment, depends, first, on our ability to describe the physico-chemical properties by the proper computational tools and secondly, on our capability to use our model to evaluate and forecast the process evolution.
The investigation of the chemical, physical and structural properties for natural systems is extremely complex and frequently such kinds of analyses are limited to the characterization of simplified systems in a narrow range of P-T-composition (X)-volatile content and type (G) conditions.
Composition, T, solid and gas bubble content, dissolved volatiles, P and stress-strain regimes are all parameters that affect the rheology of melts. Despite the recent increase of experimental studies and modelling [4-7] describing the rheological behaviour of silicate liquids [8], a general model describing the viscosity of silicate liquid as a function of composition and volatile content still does not exist. Only two recent studies [9-10] model the effect of composition and water content on the liquid viscosity, but the role of other volatiles, such as C-oxide species, F-, Cl- and S-species, is essentially unknown.
As far as the effect of phases other than the liquid is concerned, although a number of experiments and models have examined the rheological properties of natural magmas at subliquidus T [11-12], a generalized method for predicting the rheological properties of multiphase systems has yet to be proposed. This is because of the numerous factors influencing the rheology of multiphase systems, such as the changing chemistry of residual melt, the concentration, shape, mean size, size distribution and maximum packing concentration of suspended particles and the rheological properties of the continuous fluid.
As for calorimetric properties, the experimental data from simple silicate systems have led to the development of models that assume ideal mixing of oxide components, each of which has an associated partial molar Cp [13-14]. However, for aluminium-bearing liquids Cp is not a linear function of composition at fixed T [15-16]; thus, the assumption of ideal mixing of oxide components would not appear to be a general feature of silicate liquids as well as the T independence of Cp liquid [17-19]. With exception to a single contribution [20] for hydrous phonolites, Cp data for volatile bearing multicomponent systems are lacking.
Finally, the lack of knowledge about volumetric properties (thermal expansivity, adiabatic compressibility, molar volum of oxides and volatile elements) is even more evident. The existing data are obtained on simple synthetic compositions, and it is difficult to extrapolate them to multicomponent natural melts. Only one study is reported on dry natural compositions [21]. As far as the evaluation of partial molar volume of H2O-bearing melts is concerned, the proposed values for the partial molar volume of H2O, mostly derived from thermodynamic calculation, range from 0 to 25 cm3/mol. The studies performed on silicate liquids [e.g., 21] confirm the strong effect of H2O in decreasing the density of melts. The database is however small and entirely based on synthetic simple systems, so it is unclear whether the same conclusion would also apply at higher T in natural melts. Only one work has been published on natural hydrous phonolite liquids [22] which points to a very strong difference between the expansion behaviour of hydrated glasses and liquids reflecting the existence of configurational contributions to the liquid expansivities, nonexistent in glasses.
On the other hand, several conceptually valid approaches exist that allow to establish a relationship between the physical and the atomistic properties for simplied silicate melt mixtures. Amongst them the Flory Huggins (e.g., [23]), the quasi-chemical (e.g., [24]) and the polymeric models (e.g., [25-29]) are the most commonly used in material sciences. These allow to describe the chemical interactions by topologically accurate phase diagrams. The most recent formulation of a “Hybrid Polymeric” Approach (HPA), capable of describing the properties of chemically complex melts [27], was recently used to predict energy of all phases at their melting point in the four-components CaO-MgO-Al2O3-SiO2 [28].
Finally, because the polymeric approach is able to quantify the extension and disorder in the anionic sublattice, it also furnishes a practical tool in depicting the intrinsic significance of viscosity, along the guidelines of the Adams-Gibbs theory [30] and the effect of the intensive variables on this property[28].
The identification of the mathematical solutions providing the numerical coefficients necessary to adequately describe the equations of state of the investigated natural systems is facilitated by adopting the SIMPLEX numerical method [31]. This step is dependent by the knowledge about the reactive properties of silicatic melts in different thermo-baric regimes. This method allow to topologically limit the interval of solutions thorugh rapresentation of the considered systems within “closed” diagrams (e.g., ternary or tetragonal spaces of petrological relevance), called in geometry simplex.
Starting from this point several papers have been published in which partial solutions to the problem of closed data were proposed, solutions that, however, have not been validated from mathematical nor statistical points of view. A new phase was based on the realization by Aitchison in the 1980s (e.g., [32-346]) that compositions provide information about relative, not absolute, values of components, that therefore every statement about a composition can be stated in terms of ratios of components.
The calibration of the HPA model to multicomponent natural system is never attempted so far, but its structure is such to allow also calibration under pressurized conditions.
The final step of our program is such that we want to test the model in the context of magmatic degassing.
Magmatic degassing depends primarily on the pressure gradient exerted on the gas-silicate melt system, hence on the solubility of gaseous species [35-36]. We can identify three main phases [37] i) exsolution, or the formation of a gas phase, ii) expansion, due to gas-phase volume increase upon continuous decompression (closed system degassing) and iii) gas-phase separation, that is the physical decoupling of the exsolved phase (open system degassing), that tends to evolve as an independent thermodynamic system. The degassing phenomenon is then affected by relative velocities of rising melts and gas bubbles. The resulting degassing style (closed vs open) determines relative enrichments of some volatile components to detriment of some others. Among volatiles dissolved into magmas commonly occurring in Nature, H2O is the most abundant, then CO2, S, and halogens (F, Cl) follow [38]. Volatiles, H2O particularly, drive eruptive dynamics [39] and influence magma rheology and thermodynamic properties, modifying its reactivity and shifting phase boundaries [40]. The quantification by analytical techniques (e.g., titrations, spectroscopic studies) of the volatile content contained in the glassy matrix and fluid inclusions within crystal phases as well as textural analysis [41-42] allow to evaluate phase partitioning of the different volatile species (exsolved or dissolved) in a magmatic mixture [43] and the degassing paths [41-42], respectively. Volatile solubility and diffusivity, but also viscosity and density of the magmatic mixtures strongly affect the evolution of the degassing process [UR Romano]. The shape, size and abundance of crystals in the groundmass depends on undercooling, hence depressurization and degassing style. Usefulness of textural and compositional studies of pyroclastic rocks has been widely shown to recognize the interrelations among volatile content, degassing style and crystallization [44-48]. For equilibrium degassing, we omit here the details about the approaches adopted for the study of the solubility of H2O [49-52] and CO2 [52-54]. Although they reproduce satisfactorily a large solubility database [34-36], they do not fully explain the chemical reactivity of volatiles in melts. An example is given by the speciation of H2O; its amphoteric behavior of (both acidic and base), has not yet been realized. The polymeric approach to S solubility has recently been merged with the recently calibrated and performing H2O-CO2 saturation model [55], therefore obtaining a code for the saturation of CO2-H2O-SO2-H2S [56].
For halogens, very little is known about their solubility law, and no approach based on solid physical-chemistry arguments has been proposed, also because of the additional difficulties associated with the eventual coexistence of melts and aqueous brines [57].
On the kinetic side, bubble nucleation and growth models [37, 58-61] give satisfactory results and reproduce quite well the physics behind diffusivity-limited and viscosity-limited bubble growth processes. Still on the kinetic side, we see that the literature reports many Arrhenian 1/T dependencies for the diffusivities of H2O, CO2, S [62], and Cl and F [63]. However, general models for compositional dependencies are lacking.
Among the gaseous volatiles emitted by volcanoes, halogens have been the focus of growing concern, for multi-fold reasons. In particular because the presence of halogens have a significant effect on silicate melt properties (e.g., [64]), and phase equilibria (e.g.,[65]). In addition halogens are exsolved and released from ascending magmas at relatively-shallow depths (typically at pressures lower than 50 MPa in basaltic magmas; [66]), implying that their study in volcanic gas emissions can allow the retrieval of significant insights into pre-eruptive and syn-eruptive volcanic degassing, thus contributing to volcano monitoring ([67]).


BIBLIOGRAPHY
[1] Sparks RSJ (2004) Volcanic Degassing. Geological Society, London, Special Publications 213, 5-22;
[2] Klug C, Cashman KV (1996) Bull. Volc. 58, 87-100;
[3] Papale P (1999) Nature 397, 425-428;
[5] Giordano D, Romano C, Papale P, Dingwell DB (2004) Chem. Geol. 213, 49-61;
[6] Richet P, Lejeune AM, Holtz F, Roux J (1996) Chem. Geol. 128, 185-197;
[7] Giordano D, Dingwell DB (2003) Earth Planet. Sci. Lett. 208, 337-349; plus Erratum 221, 449;
[8] Russell JK, Giordano D, Dingwell DB (2003) Am. Min. 88, 1390-1394;
[9] Hui H, Zhang Y. (2007) GCA 71, 403-416;
[10] Giordano D, Russell JK, Dingwell DB (2007) Earth Planet. Sci. Lett. in press;
[11] Ishibashi H, Sato H (2007) J. Volcanol. Geoth. Res. 160, 223-238;
[12] Costa A (2005) Geophys. Res. Lett. 32,
[13] Lejeune AM, Richet P (1995) J. Geophys. Res. 100, B3, 4215 – 4229;
[14] Toplis MJ, Gottsmann J, Knoche R, Dingwell DB (2001) GCA 65, 1985-1994;
[15] Stebbins, J.F., Carmichael, L.S.E., Moret, L.K. (1984). Con. Min. Pet. 86, 131-148;
[16] Richet, P. and Bottinga, Y. (1985). GCA 49, 471-486;
[17] Bouhifd MA, Courtial P, Richet P (1998). J. of non-cryst. sol. 231, 169-177;
[18] Navrotsky 1995, Rev Mineral. 32, 121-143;
[19] Bouhifd MA, Whittington A, Roux J, Richet P (2006) GCA0, 711-722;
[20] Potuzàk M, Dingwell DB (2006). Chem. Geol. 229, 10-27;
[21] Bouhifd MA, Whittington A, Richet P (2007) Contr. Min. Petr. 142, 235-243;
[22] Ochs FA, Lange RA (1999) Science 283, 1314-1317;
[23] Flory Huggyns (1953).
[24] Pelton, A.D. and Blander, M. (1986) Metall. Trans. B. 17B, 805-15
[25] Toop, G.W. and Samis, C.S. (1962a) Can. Met. Quart. 1, 129-152
[26] Toop, G.W. and Samis, C.S. (1962b) Trans. Met. Soc. AIME 224, 878-887
[27] Ottonello, G. (2001) J. Non-Cryst. Solids 282, 72-85
[28] Ottonello G. (2005) Ann. Geophys., 48, 561-581
[29] Moretti R. and Ottonello G. (2003) Metall. Mater. Trans. B, 34B, 399-410
[30] Adam G. and Gibbs J.H. (1965) J. Chem. Phys., 43, 139-146
[31] Aitchison J., 1986. Chapman and Hall Ltd., London, 416 p.
[32] Aitchison J. , 1982. J. R. Stat. Soc. ser. B (Stat. Methodol.) 44, 139.177.
[33] Aitchison J., 1984. Math. Geol., 16, 531-564.
[34] Aitchison J., 1992. Mathematical Geology, 24(4), 365-379.
[35] Goranson RW (1931) Am. J. Sci., 22, 481-502
[36] Goranson, RW (1936) YB – Carn. Inst. Wash., 99-101
[37] Sparks RSJ (1978) JVGR, 3, 1-37
[38] Holloway J.R: (1981) In: Thermodynamics of minerals and melts (Newton RC, Navrotsky A, Wood BJ eds; Saxena SK series ed.), 272-293, Springer-Verlag
[39] Eichelberger J.C., Carrigan C.R., Westrich H.R., Price R.H. (1986) Nature, 323, 598-602
[40] Burnham C.W. (1979) In: The Evolution of Igneous Rocks: 50th anniv. Persp. (ed. H.S. Yoder), pp 439-482, Princeton Univ. Press.
[41] Polacci M (2005) Ann. Geophys., 48, 731-738
[42] Cashman K. and Blundy J. (2000) Phil. Trans. R. Soc. Lond. A, 358, 1487-1513.
[43] Papale P (2005) JGR-Solid Earth, 110 B3, 13 pp.
[44] Polacci M (2005) Ann. Geophys., 48, 731-738
[45] Polacci M, Baker DR, Mancini L, Tromba G, Zanini F (2006) GRL, 33, L13312
[46] Piochi M., Polacci M., De Astis G., Zanetti A., Mangiacapra A., Vannucci R. and Giordano D. (2007) G-CUBED, IN PRESS
[47] Cashman K. and Blundy J. (2000) Phil. Trans. R. Soc. Lond. A, 358, 1487-1513.
[48] Hammer J.E., Cashman K.V., Hoblitt R. and Newman S. (1999) Bull. Volc., 60, 355-380.
[49] Burnham CW and Davis NF (1979) Am. J. Sci., 270, 54-79
[50] Silver LA and Stolper EM (1989) J. Petrol., 30, 667-709
[51] Papale P (1997) Contrib. Mineral. Petrol., 126, 237-251
[52] Papale P. (1999) Am. Mineral., 84, 477-492
[53] Fine GJ and Stolper EM (1986) EPSL, 76, 263-278
[54] Newman S and Lowenstern JB (2002) Comp. Geosci., 28, 597-604
[55] Papale P, Moretti R., Barbato D. (2006) Chem. Geol., 229, 79-85
[56] Moretti R., Papale P. and Ottonello G. (2003) In: Volcanic Degassing (Oppenheimer C., Pyle D.M., Barclay J. eds.), 213, 81-101, Geol. Soc. Lond. Spec. Publ.
[57] Bottinga Y and Javoy M (1990) JGR, 95, B4, 5125-5131
[58] Proussevitch AA and Sahagian DL (1998) JGR-Solid Earth, 103, 18223-18251
[59] Lensky NG, Navon O and Lyakhowsky V (2004) JVGR, 129, 7-22
[60] Toramaru A (1995) JGR-Solid Earth, 100, 1913-1931
[61] Paonita A. and Martelli M. (2006) EPSL, 241, 138-158
[62] Baker DR, Freda C, Brooker R and Scarlato P (2005) Ann. Geophys., 48, 699-717
[63] Freda C, Baker DR and Scarlato, P (2005) GCA, 69, 5061-5069
[64] Giordano D., Romano C., Dingwell D. B., Poe B. and Behrens H. (2004). Geochim. Cosmochim. Acta 68, 5159-5168.
[65] Manning D.A.C. (1981) The effect of fluorine on liquidus phase relationships in the system Qz-Ab-Or with excess water at 1kb. Contrib. Mineral. Petrol. 76, 206-215.
[66] Spilliaert, N., Allard, P., Metrich, N., and Sobolev, A.V. (2006). J. Geophys. Res. 111, doi: 10.1029/2005JB003934.
[67] Aiuppa, A., Federico, C., (2004) Geophys. Res. Lett., 31, L14607, doi:10.1029/2004GL020458 <<<