Synthesis and Characterization of Ti-bearing Garnets

Daniele Borrini (dborrini@steno.geo.unifi.it)¹, Andrea Orlando (dborrini@steno.geo.unifi.it)², Giuseppe Pedrazzi (mosslab@ipruniv.cce.unipr.it)³, Emanuela Schingaro (schingaro@geomin.uniba.it)⁴ & Fernando Scordari (scordari@lgxserver.uniba.it)⁴

¹ Centro di Studi per la Minerogenesi e la Geochimica Applicata, C.N.R., Firenze, Italy

² Dipartimento di Scienze della Terra, Università di Firenze, Italy

³ Istituto di Scienze Fisiche, Università di Parma, Italy

⁴ Dipartimento Geomineralogico, Università di Bari, Italy

The synthesis of Ti-bearing garnets under controlled physico-chemical conditions was undertaken in order to characterize the crystal chemical behaviour of cations in the garnet structure. Ti-andradite saturated in Ti (nominal composition Ca₃Fe₂Ti_1.42 Si_1.58O₁₂ ) was prepared starting from a gel following the procedure described by Ito & Frondel (1967). The run products were: Ti-andradite (~95 wt%), perovskite (~4 wt%) and probably larnite, Ca₂SiO₄, (~1 wt%) (Yamaguchi et al.(1957)). The sample was then analysed by chemical (EPMA), structural (XRD) and spectroscopic (Mössbauer ) methods. EPMA results revealed little amount (~ 1 wt%) of unexpected Al₂O₃. Rietveld refinement yielded cell parameter a = 12.232 Å and main geometric parameters: <X-O> = 2.450, Y-O = 2.003, Z-O- = 1.728 Å. The Mössbauer investigation showed two different iron doublets, ascribable to Fe³⁺(Z) (I.S.= 0.230(4), Q.S.= 1.227(1) mm/s, area = 49.8%) and to Fe³⁺(Y) (I.S. = 0.344(5), Q.S.= 0.66(1) mm/s, area = 50.2%). The experimental cation distribution is the following. Ca_2.99(Al_0.12Ti⁴⁺_0.94Fe³⁺_0.94)(Si_1.84Fe³⁺_0.94,Ti⁴⁺_0.22)O₁₂. Electrostatic balance considerations indicate that only schorlomite- substitution

(Ti⁴⁺(Y)+ Fe³⁺(Z)´<­> Si⁴⁺(Z) + M³⁺(Y), M=Fe, Al)

affect the compound structure. From the Mössbauer indications about the iron distribution two different Ti⁴⁺ environments are also expected, in agreement with Weber et al. (1974). The XPS analysis is in progress to verify the proposed distribution.

Ito J & Frondel, C, Am. Mineral, 52, 773-781, (1967).

Weber HP, Virgo D & Huggins, FE, Carnegie Inst. Washington Year Book, 74, 575-579, (1974).

Yamaguchi G, Miyabe H, Amano K & Komatsu S, J. Ceram. Soc. Japan, 65, 99-103, (1957).

Lava Rheology: Effect of Carbon Dioxide on the Viscosity of a Simple Synthetic Silicate Melt

Emmanuelle Bourgue (bourgue@ipgp.jussieu.fr) & Pascal Richet (richet@ipgp.jussieu.fr)

Laboratoire des Géomatériaux.Institut de Physique du Globe de Paris, Tour 14-15 3ème, 4 place Jussieu, 75252 Paris Cedex, France

Temperature, pressure and chemical composition are known to exert a strong influence on physical properties of molten silicates. Among compositional variables, the volatile content is especially important. Recent work has for instance shown the tremendous influence of dissolved water on the density and the viscosity of silicate melts. Although carbon dioxide is the second most abundant volatile in magmas, its effect on melt properties are poorly known. Part of the reason for such a lack of data is the high pressure required to dissolve significant amount of CO₂. To obviate this problem in this preliminary study we have investigated a synthetic potassium silicate liquid, KS_1.3 (56.9 mol% SiO₂ - 43.1 mol% K₂O), in which CO₂ can be kept metastably thanks to a very slow rate of exsolution when the material is prepared from silica and potassium carbonate. For a series of 8 samples with up to 3.5 wt% CO₂, we have observed that the density of quenched glasses decreases with increasing CO₂ content. In addition, measurements made with a creep apparatus indicate that addition of 3.5 wt% of CO₂ causes the melt viscosity to decrease by a factor 2 above the glass transition range whereas Couette experiments at higher temperatures show little variations with the CO₂ content. As will be discussed, these effect parallel those observed for water and point to a significant role of dissolved CO₂ on magma buoyancy and rate of flow.

Solubility Mechanism and Diffusion of CO₂ in Aluminosilicate Melts

Melanie Braune-Frehse (m.braune-frehse @mineralogie.uni-hannover.de)¹, Marcus Nowak (m.nowak@mineralogie.uni-hannover.de)¹ & Hans Keppler (hans.keppler@uni-bayreuth.de)²

¹ Institut für Mineralogie, Welfengarten1, 30167 Hannover, Germany

² Bayerisches Geoinstitut, Universität Bayreuth, 95440 Bayreuth

CO₂ is the second most important volatile component in terrestrial magmas and plays an important role in degassing processes and magmatic phase equilibria. During ascent of a magma, CO₂ is often the first volatile to reach saturation in the melt. The formation of bubbles is therefore controlled by the solubility and diffusion of CO₂.

Data on the speciation of CO₂ in glasses at room temperature show that CO₂ is preferentially incorporated as molecular CO₂ in quenched melts with a high degree of polymerization (e.g. rhyolite), while CO₃^2- groups dominate in more depolymerized systems (e.g. basalt; Blank and Brooker, 1994).

To investigate the effect of polymerization and CO₂-speciation on CO₂ diffusion, we performed diffusion couple experiments with a variety of melt compositions. We synthesised glasses with different degrees of polymerization starting with albitic glass as a fully polymerized endmember (the ratio of nonbridging oxygen to tetrahedrally coordinated cation, NBO/T, is 0). By adding 2 wt% of Na₂O to the albitic melt, the network becomes slightly depolymerized (NBO/T = 0.020). Albite with 4 wt% Na₂O has a NBO/T = 0.086. FTIR spectra of the quenched glasses show that the degree of polymerization controls the speciation of CO₂. In albitic glass we observe a constant ratio of molecular CO₂ to CO₃^2-, independent of total CO₂ content. With increasing degree of depolymerization the ratio of molecular CO₂ to CO₃^2- decreases.

We performed one-dimensional semi-infinite diffusion couple experiments in the three aluminosilicate melts in an IHPV at 1100 - 1250°C and 5 kbar. The resulting symmetric diffusion profiles were measured with FTIR-micro-spectroscopy and fitted by an error function. We observe a change in the bulk diffusivity of CO₂ at constant p,T-conditions with increasing degree of depolymerization starting at albite with logD_CO2 = -11.4, passing through a minimum at albite + 2 wt% Na₂O (logD_CO2 = -11.6), and ending at albite + 4wt% Na₂O with logD_CO2 = -11.1. This change may be due to a combined effect of the bulk polymerization of the melt and a change in the speciation of CO₂.

Blank JG & Brooker RA, Rev Mineral, 30, 157-186