Solubility of Noble Gases in H₂O-CO₂ Bearing Magmas: Experimental Investigation and Modelling

Antonio Paonita (paonita@cfta.math.unipa.it)¹, P. Mario Nuccio (nuccio@cfta.math.unipa.it)¹ & Raffaello Trigila²

¹ Dipartimento di Chimica e Fisica della Terra ed Applicazioni, Università degli Studi di Palermo, Via Archirafi 36, 90123 Palermo, Italy

² Dipartimento di Scienze della Terra, Università degli Studi La Sapienza, P.le Aldo Moro 5, I-00185 Roma, Italy

The study of the noble gas solution in silicate melts provides fundamental constraints to a number of questions about Earth Science. In natural systems, noble gases always occur as minor or tracer constituents of the gaseous phase, which will be mostly composed by H₂O and/or CO₂. Nevertheless, literature data and related models on the noble gas solubility in silicate liquids only concern gas phase formed by one or two noble gases. We have developed a model of solubility of noble gases in any silicate magmas having a gas phase mostly composed by H₂O and CO₂. The model is based on the relationship between concentration of dissolved noble gas and ionic porosity of melt, found by Carroll and Stolper (1993) for H₂O-CO₂ free melts. It evaluates the H₂O and CO₂ effect on the melt ionic porosity and, consequently, on Henry's constant of noble gas. The fugacities of the noble gases in H₂O-CO₂-noble gas mixtures are also considered in our equilibrium calculations. Our model predicts a clearly positive dependence of inert gas solubility on dissolved H₂O in melt, which becomes less important when water concentration is higher than 3 wt%. Oppositely, noble gas solubility decreases as a consequence of CO₂ addition in both basaltic and rhyolitic melts. Model predictions for helium solubility were tested by designing a new experimental method to investigate the solubility of inert gases in H₂O-CO₂ bearing silicate melts. The peculiarity of the method concerns the capability to load, in accurately known proportions (even lower than ppm), volatiles having a gaseous state at the room conditions. Gas was loaded as a weighted amount of a gas-bearing glass, which was previously prepared by using the same gas as pressure medium of a superliquidus run. Finally, comparison between experimental data performed by the above method and calculated values revealed a good agree.

Carroll M & Stolper EM, Geochim. Cosmochim. Acta, 57, 5039-5051, (1993).

Investigation of the Tremolite-Mn Cummingtonite Join by FTIR Spectrometry

Arnaud Papin (papin@cnrs-orleans.fr) & Jean-Louis Robert (jlrobert@cnrs-orleans.fr)

ISTO, CNRS-Univ. d'Orléans, 1A, rue de la Férollerie, 45071 Orleans Cedex 2, France

Amphiboles have been prepared along the tremolite Ca₂Mg₅Si₈O₂₂(OH)₂) - Mn cummingtonite Mn₂Mg₅Si₈O₂₂(OH)₂) join at 700°C, 1 kbar PH₂O, under fO₂ conditions set by the NNO buffer. Almost complete solid solution exists along the investigated join, in these conditions, only trace amounts of diopside appear for compositions close to the tremolite end-member. The evolutions of cell parameters are not linear towards Mn cummingtonite end-member. The a unit cell parameter and the ß angle decrease, b increases, c remains constant, as the Mn content increases. This evolution of cell parameters reflects complex Ca, Mg, Mn distributions within the amphibole structure, along the tremolite-Mn cummingtonite join. Theoretically, in synthetic Mn cummingtonite, Mn should exclusively occupy M4 sites, replacing Ca of tremolite. By analogy with tremolite a single band is expected for Mn cummingtonite, by FTIR in the OH-stretching region. But, this amphibole shows a complex spectrum with a "tremolite-like" band at 3675 cm^-1 and 6 additional bands at lower wavenumbers (3671 to 3645 cm^-1). The presence of these bands reflects various cationic environments, around the OH group, that is, a Mn - Mg distribution over M4 and M(1,2,3) sites. Along this join, each additional band can be assigned to different first, second, third Mn, Mg (eventually Ca) distributions around the OH group. As a conclusion, the tremolite -Mn cummingtonite does not reflect a simple Ca²⁺ - Mn²⁺ substitution at M4, but a complete rearrangement of divalent cations over M sites.

Mn and Zn in Synthetic Micas, Under Different Oxygen Fugacities

Arnaud Papin (papin@cnrs-orleans.fr) & Jean-Louis Robert (jlrobert@cnrs-orleans.fr)

ISTO, CNRS-Univ. d'Orléans, 1A, rue de la Férollerie, 45071 Orleans Cedex 2, France

Structural determinations show that Mn can occupy M(1,2) octahedra sites in micas. A positive Mn-Zn correlation has been observed in micas and clinoamphiboles from the Franklin furnace deposit, New Jersey. We present results of an experimental and crystal-chemical study of the behaviour of (Mn,Zn) in phlogopite KMg₃(Si₃Al)O₁₀(OH)₂, chosen as a models of micas. Experiments were performed at 600°C, 1 kbar, under HM, NNO and MW fO₂ conditions. An almost complete solid solution is observed along the Mg₃ - (MnZn) join, under NNO and MW fO₂ conditions, with only trace amounts of willemite and orthoclase at the (Mn,Zn) end-member. With HM, the solid solution range is more restricted (to 70%) along the Mg₃-(MnZn) join. FTIR absorption spectroscopy in the OH-stretching region shows a pronounced decrease of band intensities, as the (Mn,Zn) content increases, for all three fO₂ conditions. This indicates a progressive loss of protons along the join. The (MnZn) end-member, stable with NNO and MW, is completely deprotonated. The most likely explanation for this is the presence of manganese under the trivalent state instead of the expected divalent one. The presence of Mn³⁺ can be interpreted in terms of dimensional constraints, since Mn²⁺ is by far a too large cation to occupy the M(1,2) octahedra in significant proportions. Large amounts of Mn³⁺ in a mica are already known in norrishite KLiMn₂₊³⁺Si₄O₁₀O₂. On the whole, the mica stability in the system studied implies a combination of Mg, Mn and Zn, resulting in an average ionic radius close to that of ^[6]Fe²⁺ H.S. (0.78 Å).