Experimental Determination of Aluminous Clinopyroxene-Melt Partition Coefficients Bearing on the Roman Province Magmas Evolutionary Patterns

Raffaello Trigila (trigila@axrma.uniroma1.it)¹ & Bernard J. Wood (b.j.wood@bristol.ac.uk)²

¹ Università degli Studi di Roma La Sapienza, Dipartimento di Scienze della Terra, P.le A.Moro 5, 00185 Roma, Italy

² University of Bristol, Department of Earth Sciences, Bristol BS8 1RJ, United Kindom

Experiments on lava-types from 1944 Vesuvius and 1991-93 Etna eruptions, performed at 0.1-200 MPa, 1000-1200°C, oxygen fugacities corresponding to NNO buffer and moderate water activities, generate, as liquidus phase, very aluminous (Al₂O₃ up to 12 wt%) and ferric (Fe₂O₃ up to 10 wt%) clinopyroxenes. Partitioning of REE and HFSE between clinopyroxene and melt confirms the trend of increasing D_i with Al^iv observed in previous studies (Blundy et al., 1998; Lundstrom et al., 1998). The extent of this increase for Al^iv passing from 0.02 to 0.5 a.f.u. is about 1 order of magnitude for Ti and REE and 2 order of magnitude for Nb, Ta, and Zr. These data can be applied to model the evolutionary history of the Roman Province potassic magmas if the clinopyroxene compositional dependence on P, T, H₂O^liq of the magma ascent is known. In particular the new trace element partition coefficients verify the amount of fractional crystallisation of clinopyroxene which is deduced from the major element composition of the rocks. I.e. for Vesuvius historical lavas the experimental partition coefficients for the REE when combined with the observed REE of the rocks imply that the fractionating clinopyroxene contains Al₂O₃ very close to 7.5wt%. Experimental data demonstrate that Al content of clinopyroxene increases with decreasing temperature and increasing water fugacity or increasing pressure. The observed Al₂O₃ amounts are common for clinopyroxene phenocrysts suggesting precipitation within the magma chamber at low pressure (about 100 MPa) and low water activity. Considering in turn, the HFSE for which data are available, the clinopyroxene-liquid D values for Zr and Hf of 1.3 and 0.9-2.1 respectively, produce concentrations along the fractionation trend of Vesuvius historical lavas which agree with those observed from Pearce ratios analysis and are consistent with the above indicated model of magma evolution (Trigila and De Benedetti, 1993).

Blundy JD, Robinson JAC, Wood BJ, Earth Planet. Sci. Letts, 160, 493-504

Lundstrom CC, Shaw HF, Ryerson FJ, Williams Q, Gill J, Geochim. Cosmochim. Acta, 62, 2849-2862

Trigila R, De Benedetti AA, J. Volcanol. and Geoth. Res, 58, 315-343

Activity-Composition Relations of Titanite in the System CaTiSiO₄O-CaAlSiO₄F at 1000-1100°C

Peter Tropper (peter.tropper@uibk.ac.at)¹, Craig Manning (manning@ess.ucla.edu)² & Eric Essene (essene@umich.edu)³

¹ Institute of Mineralogy and petrography, University of innsbruck, Innrain 52; A-6020 Innsbruck, Austria

² Department of Earth and Space Sciences, University of California, Loa Angeles, CA-90095-1567, USA

³ Department of Geological Sciences, University of Michigan, MI-48109-1063, USA

Titanite is a common accessory mineral and could be used reliably in phase equilibrium calculations, if activity-composition relations in (Al + F)-bearing titanites were known. We present results from piston-cylinder experiments on the activity-composition relations in the system CaTiSi₄O - CaAlSiO₄F at 1000 - 1100_°C and 1.2 - 2.0 GPa. Experiments were performed by adding fluorite to the assemblage anorthite + rutile + titanite + kyanite, thus displacing the reaction anorthite + rutile = titanite + kyanite to lower pressures from the position of the pure end-member curve (Manning and Bohlen 1991) due to the (Al + F) substitution in titanite. In the presence of fluorite, the stability limits of the assemblage (Al + F) titanite + anorthite + rutile shift from 1.78 GPa to 1.375 ± 0.025 GPa at 1000_°C and from 2.01 GPa to 1.575 ± 0.025 GPa at 1100_°C. The shift constrains the activity of titanite to be 0.520 ± 0.03 at 1000°C and 0.522 ± 0.04 at 1100°C. Compared to the X_Ti of (Al + F) titanites from these experiments, 0.47 ± 0.02 at 1000°C and 0.50 ± 0.03 at 1100°C, the results indicate that the activity of titanite shows a slight positive deviation from ideality. The data also show that the activity is very close to the ideal molecular activity model (X_Ti) at these temperatures and therefore the molecular activity model is recommended for use in calculations until more data are available. Recalculations of the P-T conditions of two eclogites from the study of Manning and Bohlen (1991) illustrates the difference in pressure estimates resulting from the choice of different activity models (molecular vs. ionic) ranging from 0.05 to 2.0 GPa, depending on the (Al + F) substitution in the natural titanites and the sensitivity of reactions to changes in titanite activity.

Manning CE & Bohlen SR, Contrib. Mineral. Petrol, 109, 1-9, (1991).

Techniques of the Correction of Melt Inclusion Compositions for Boundary Layer Effects

Sergey Trousov (stup@geo.web.ru) & Pavel Pletchov (pavel@cs.ru)

Room A-402,Geological Department, Moscow State University, Vorobievi Gori, Moscow,119899,Russia, Russia

It's well known, that at growth of crystals there is the boundary layer near growing surface, distinguished from melt composition on a distance from this surface. Width of this layer is depending on melt viscosity and growth conditions from 10-15 micron (for basalt melts) up to 100-150 microns (for viscous rhiolitic melts). The melt inclusion at formation should partially grasp a melt portion from this layer (Lu et al., 1995). The obtained experimental data on melt inclusion growth (Pletchov et al.,1997), study of boundary effects in natural glasses and melt inclusions allow to take into account influence of boundary effects on the basis of multi-component diffusion model (Oishi, 1965) of a boundary layer. The account of complete model of a boundary layer is complicated and requires attraction of a numerous of uncertain parameters, such as diffusion coefficients of ionic complexes in the melt. The semi-empirical approach for correction calculation based on study of glassy melt inclusion zonation is offered. The correction on boundary effects is similar in a quantitative sense adjusted for the host mineral crystallization inside inclusion at main components calculation, but essentially differs for trace elements collecting in a boundary layer proportional diffusion ability and Kd between melt and the host-mineral. The boundary effects calculation, probably, will allow reduce significant disorder of the trace element contents in melt inclusions.

Lu F.Q., Anderson A.T., Davis A.M., Journal of Geology, 103, 591-597, (1995).

Plechov PYu, Kotelnikov AR, Experiment in Geosciences, 6, 48-51, (1997).

Oishi Y, Journ. Chem. Phys, 46, 1611-1620, (1965).