Modelling the Non-Arrhenian Rheology of Silicate Melts: Numerical Considerations

Kelly Russell¹, Don Dingwell², Kai-Uwe Hess & Daniele Giordano

¹ Earth & Ocean Sciences, Univ. British Columbia, Vancouver, Canada

² Bayerisches Geoinstitut, Universität Bayreuth, Bayreuth, Germany

Accurate representations of magmatic and volcanic processes require a reliable model that can predict the viscosity ( ) of silicate melts over the range of temperatures and compositions found in nature. The task of creating such a model is complicated by the non-Arrhenian rheological behaviour of natural silicate melts. This non-Arrhenian behaviour introduces a non-linear temperature dependence which ultimately impacts on how we handle compositional controls on viscosity. This is true regardless of which of the common non-arrhenian viscosity equations is used. For example, the purely empirical Tamman-Vogel-Fulcher expression: (1) Log = A + B/(T - C). introduces three fit parameters (A, B, C) for which the compositional dependencies are at present unknown. Estimates of these model parameters are generally derived by fitting Eq. 1 to individual T-viscosity datasets. The form of Eq. 1, however, can result in strongly correlated or even non-unique estimates of the model parameters (A, B, C); this ultimately has impact on how we partition the effects of composition. To demonstrate some of these issues we have fit Eq. 1 to several experimentally-measured T-viscosity datasets (N observations)representing fragile and strong silicate melts. We have then explored the best-fit solution surface using (chi)² minimization techniques. Our results are preliminary but show that strong non-linear relationships exist between model values of A and either B or C, and that there are strong linear correlations between values of B and C. The nature of the numerical correlation between model parameters and the potential for non-unique parameters is critical to understand before compositional dependencies are assigned.

Dehydration Mechanism in Brewsterite: Single-Crystal X-ray Diffraction Study

Michele Sacerdoti¹, Giovanna Vezzalini² & Simona Quartieri³

¹ Istituto di Mineralogia, Corso Ercole I d'Este 32, 44100 Ferrara, Italy

² Dipartimento di Scienze della Terra, via S. Eufemia 19, 41100 Modena, Italy

³ Dipartimento di Scienze delle Terra, Salita Sperone 31, 98166 S. Agata di Messina, Messina, Italy

The crystal structure of the zeolite brewsterite (Schlenker et al. 1977) (s.g. P21/m, ideal formula (Sr,Ba)₂Al₄Si₁₂O₃₂·10H₂O) at different dehydration levels has been studied by single-crystal X-ray diffraction on 6 samples kept in evacuated capillaries at different temperatures and for different treatment times. The thermal behaviour of brewsterite has been previously investigated by Alberti et al. (1999) and by Ståhl and Hanson (1999) by single-crystal and in-situ powder diffraction experiments, respectively. In both studies, the dehydration induces a strong decrease of the unit cell volume, and a spreading over several sites of the exchangeable cation, which was originally hosted in only one site. However, only in the completely dehydrated form (Alberti et al., 1999), a statistical breaking of one T-O-T bridge and the formation of new 4- and 5-coordinated (Si,Al) sites were observed. This work reports the structural refinements of the following brewsterite samples: BR24h and BR96h (held at room temperature under vacuum for 24 and 96 hours, respectively); BR100 (100°C for 24 h); BR180 (180°C for 80 h); BR330f (330°C for 0.5 h); BR330s (330°C for 24 h). The following results were obtained: a) The structure refinements of the partially dehydrated brewsterite indicate that, as the dehydration advances, the cell volume decreases, the channel are squashed and the exchangeable cations spread over several sites, previously occupied by water molecules; these results are in agreement with those by Ståhl and Hanson (1999). b) The T-O-T bridge breaking is caused by the strain induced on the framework oxygens by the exchangeable cations when their coordination polyhedra loose almost all the water molecules; this phenomenon (which was not observed by Ståhl and Hanson (1999) due to the residual water content of their sample) occurs, in increasing percentages, only in BR330f and BR330s. c) Besides favoring the T-O-T breaking, a longer exposure of the single-crystal at 330°C induces a stronger spreading of the extra-framework cations.

Alberti A, Sacerdoti M, Quartieri S, Vezzalini G, Phys. Chem. Minerals, 26, 181-186, (1999).

Schlenker JL, Pluth JJ, Smith JV, Acta Crystallogr, B33, 2907-2910, (1877).

Stahl K, Hanson JC, Micro. Meso. Materials, 32, 147-158, (1999).

Phase Equilibria of Vesuvius Phonolites: Preliminary Results

Bruno Scaillet (bsaille@cnrs-orleans.fr)¹, Michel Pichavant (pichavan@cnrs-orleans.fr)¹, Raffaello Cioni (cioni@server.dst.unipi.it)², Alessandro Sbrana & Paola Marianelli

¹ ISTO, Orléans, France

² Dipartimento di Scienze Della Terra, Universita di Pisa, Italy.

Crystallisation experiments are currently performed on white phonolitic pumice samples of the Mercato (8010 BP) Avellino (3360 BP), Pompei (79 AD), and Pollena (472 AD) eruptions in order to constrain the pre-eruption conditions in the magma chamber preceding major phonolitic eruptions at Vesuvius. Experiments were done in an IHPV with an H2-membrane, at 200 MPa, 900, 850, 825 and 800°C, fO₂ between NNO and NNO+1, and aH₂O varying between 1 and 0.3. Run durations were 6-10 days. The quench was isobaric. Run products were characterized by SEM and EMPA techniques. Phases identified are: sanidine, plagioclase, leucite, nepheline, analcime, scapolite, clinopyroxene, amphibole, biotite and garnet. For the Pompei pumice the phase relationships show that at a given temperature sanidine is the first tectosilicate to appear in the crystallisation sequence, followed by plagioclase, then leucite and analcime and finally nepheneline. At H₂O-saturation sanidine crystallises slightly below 800°C and all other tectosilicates appear largely below 800°C. Clinopyroxene is the liquidus phase under H₂O-rich conditions. It is not present at 800°C where it is replaced by both garnet and amphibole. Garnet and amphibole are not stable at 850°C and above. For H₂O-saturation, the clinopyroxene precipitates at around 850°C. Preliminary results on the Mercato pumice at 800°C show neither leucite nor nepheline in the run products. The Avellino displays the lowest liquidus temperature of all three compositions, having nepheline but no leucite. This shows that significant differences exist in the phase relationships of fractionated erupted product at Vesuvius. Future experiments will be performed at 100 MPa to help constrain the depth of the magma chamber, in addition to T, and H₂O in melt.