Structural Controls on the Solubility of CO₂ in Silicate Melts

Richard A. Brooker¹, Simon C. Kohn (simon.kohn@bristol.ac.uk)¹, John R. Holloway² & Paul F. McMillan³

¹ Dept. of Earth Sciences, Wills Memorial Building, University of Bristol, BS8 1RJ, UK

² Dept of Geology, Arizona State University, Tempe, AZ 85287, USA

³ Dept of Chemistry, Arizona State University, Tempe, AZ 85287, USA

The solubility of CO₂ has been measured for the system SiO₂-Na₂O-Al₂O₃-MgO-CaO at 1.5 GPa and 1275-1400°C and for several "natural" magma compositions (Mg- and Ca-rich melilitites, andesite and phonolite) at 1.2-2.7 GPa and 1300-1600°C. For a given pressure and temperature, the solubility is a strong function of the "non-bridging oxygen" (NBO) content of the melt, expressed as the NBO/T ratio where T represents tetrahedral network-forming cations. Although the calculation of NBO/T ignores many of the complexities in silicate melt structure, it is a surprisingly useful parameter for predicting the CO₂ solubility in depolymerised melts. In highly polymerised melts, other dissolution mechanisms involving bridging oxygens become important and NBO/T is no longer the exclusive control on solubility. In Fe-bearing systems, the best correlation between CO₂ solubility and NBO/T is found when both Fe³⁺ and Fe²⁺ are assumed to be tetrahedral, indicating that these cations play a polymerising role in the melt, with respect to CO₂ dissolution. There is also evidence that some fraction of the Mg²⁺ in the melt should be assigned to a polymerising role. In addition the characteristics of carbonate groups dissolved in silicate glasses have been investigated using FTIR spectroscopy. Glasses of natural melt compositions are compared with simple analogues. This approach allows systematic investigation of the role of each major oxide component. Only Ca and Fe²⁺ bearing systems display the characteristic spectral feature which is dominant in all carbonate-bearing natural compositions, although other cations are important in producing more subtle effects. The spectra suggest that Fe³⁺, Fe²⁺ and Mg²⁺ associated with carbonate play a network forming role, consistent with the solubility systematics, and imply that carbonate is dominantly associated with Ca in natural melts.

Copper(II) Speciation in Chloride Brines and the Extraction of Thermodynamic Properties from UV-Vis-NIR Spectra

Joël Brugger (joelb@mail.earth.monash.edu.au)¹, Jay Black¹, D. C. Bear McPhail¹ & Leone Spiccia²

¹ Department of Earth Sciences, Monash University, VIC-3168 Clayton, Australia

² Department of Chemistry, Monash University, VIC-3168 Clayton, Australia

The concentrations and transport of metals in hydrothermal solutions depends on how metals combine with ligands to form aqueous complexes. The nature and thermodynamic properties of geologically relevant aqueous complexes at elevated temperatures can be measured using UV-Vis-NIR spectrophotometry; however, the quantitative interpretation of the spectra can be difficult. We combine existing interpretation methods to develop a new algorithm to interpret high quality spectra of Cu(II)-chloride complexes in 0.0 m to 18.0 m LiCl brines between 25°C and 80°C. Principal component analysis (PCA) is used to estimate quantitatively how many aqueous complexes are necessary to explain the data at each temperature. Using the results of PCA, the data are fitted using a "model-free" approach (e.g., de Juan et al. 1997). This results in spectra and concentrations of each complex without requiring any assumption on the chemistry of the system, other than the number of absorbing species present. The results from the "model-free" analysis are then used to fit equilibrium constants for Cu(II)-chloride complexes, using a non-linear least-squares approach, which includes a thermodynamic speciation model where the stoichiometry of the complexes is chosen by comparing aqueous spectra to known solid spectra. The model results are dependent on the choice of activity coefficients for charged and neutral aqueous species. Uncertainties in the fitted stability constants are estimated using maps of the residuals. The results of our study are similar to published properties of CuCl(H₂O)₅⁺ and CuCl₂(H₂O)_4(aq) at all temperatures, but diverge for CuCl₃(H₂O)₃^- and CuCl₄^2-. Moreover, the data suggest the presence of CuCl₅^3-, probably with D_3h point group, at very high salinity. The stability of this last complex decreases with increasing temperature. Our approach provides a more efficient and rigorous method to interpreting UV-Vis-NIR spectra over wide ranges of temperature, pressure and fluid composition and results in improved properties for Cu(II) complexes in chloride brines.

de Juan A, Van der Heyden Y, Tauler P & Massart DL, Anal. Chim. Acta, 346, 307-318, (1997).

Phosphates Formation by Oxidation of Phosphorus and Iron Metal in Presence of Olivine

Fabrice Brunet¹, Didier Laporte² & François Guyot³

¹ Lab. de Géologie, ENS, CNRS-UMR8538, Paris, France

² Lab. de Géologie, Univ. Blaise Pascal, CNRS-UMR6524, Clermont-Fd, France

³ Lab. de Mineralogie-Cristallographie, IPGP, Paris, France

It has been proposed that in meteorites composed of (Fe,Ni)-metal and silicates (stony-irons and some primitive achondrites), A₃(PO₄)₂ phosphates (with A = Fe, Mg, Mn, Ca) can form in presence of silicates by oxidation of the phosphorus dissolved in metal or contained in phosphides. The corresponding redox reactions are potential fO₂ indicators in these meteorites. We have investigated experimentally phosphates formation in a simplified pallasite-like system. A mixture of San Carlos olivine, pure iron-metal, hematite and FeP was held at 1000°C (0.05 and 1 GPa) for 10 days in a Fe-metal capsule. In all experiments, an Mg-dominant A₃(PO₄)₂ phosphate is recovered as well as zoned olivines with unreacted Mg-rich cores and Fa_70-75 rims. Mg-bearing wüstite often coexists with iron metal and sometimes schreibersite (Fe₃P). An increase of fO₂ is characterized by the following features in the run product: (1) Fe enrichment of both olivine and phosphate (2) Disappearance of schreibersite. (3) Decrease of the P-content in the Fe-metal grains. These results show that the reaction: Mg-olivine + [P]^Fe (or Fe₃P) + Fe° + O₂ = Mg-phosphate + Fe-olivine buffers the fO₂ in our samples. This reaction is likely to occur in natural pallasites. Consistency between thermodynamic modelling and experimental results will be discussed.

BSE image, PHOS#4 (1 GPa, 1000°C)