H₂O and CO₂ Contents of Cordierite Coexisting with Melt and the P-T Dependence of CO₂ Incorporation in Cordierite

Pauline Thompson (pauline.thompson@glg.ed.ac.uk)¹, Simon L Harley (sharley@glg.ed.ac.uk)¹ & Damian P Carrington (damian.carrington@bbc.co.uk)²

¹ Department of Geology and Geophysics, University of Edinburgh, West Mains Road, Edinburgh, EH9 3JW, UK

² News Online, BBC, Shepherds Bush, London, UK

Experiments to determine the distribution of H₂O and CO₂ between synthetic peraluminous granitic melt and natural Fe-Mg cordierite have been conducted at 3-7 kbars and 800-1000°C. The H₂O and CO₂ contents of saturated cordierites and H₂O contents of the coexisting melts equilibrated with fluids of varied H₂O/CO₂ were determined using Secondary Ion Mass Spectrometry.

Melt H₂O contents decrease sympathetically with cordierite H₂O contents whilst an antithetic relation is seen with the cordierite CO₂ contents. D_W (= wt%H₂O^(melt)/wt%H₂O^(cordierite)) for fluid saturated but CO₂ bearing conditions (3.5-4.5) are comparable to the pure H₂O system (Carrington and Harley, 1996, Harley and Carrington, submitted) when considered in terms of D_W for a specified cordierite H₂O content at fixed P-T. At a given X_CO2 in cordierite, H₂O and CO₂ both increase with increasing pressure.

Experimental a_H2O conditions have been calculated based on the measured cordierite and melt H₂O data using independent calibrations of the P-T-a_H2O relations in each phase (Harley and Carrington, submitted). These been coupled with the known mixing properties of H₂O-CO₂ to extract a_CO2 estimates for the experiments. The results are consistent with a_CO2 being proportional to m/(1-m) in cordierite at specified P-T, where m is the number of moles of per unit formula. Maximum CO₂ contents in cordierite at high temperature and pressure are then given by the relation:

lnK_C = [962 (±150)/T] - 11.91 (±0.75) - (0.2020)P

where K_C = (m_sat/(1-m_sat))/f_CO2(P,T). The cordierite CO₂ contents at saturation are significantly lower, at given P, T and X_CO2, than previously modelled by most other workers (e.g. Kurepin, 1985), but are in reasonable agreement with the cordierite B dataset of Johannes and Schreyer (1981). New experiments in the salt system place strong constraints on maximum salinities under which cordierite and melt are co-stable and show that, like CO₂, salinity does not strongly affect D_W.

Carrington DP & Harley SL, Geology, 24, 647-650, (1996).

Harley SL & Carrington DP, submitted to J. Petrol, (in prep).

Kurepin VA, Geochemistry International, 22, 148-156, (1985).

Johannes W & Schreyer W, Am. J. Sci, 281, 299-317, (1981).

T.E.M. Characterisation of Dislocation Burgers Vectors and Slip Systems in Mg₂SiO₄ Wadsleyite

Elina Thurel (elina.thurel@univ-lille1.fr)¹ & Patrick Cordier (patrick.cordier@univ-lille1.fr)^1,2

¹ LSPES, ESA 8008. Université des Sciences et Technologies de Lille, France

² Bayerisches Geoinstitut. Universität Bayreuth, Germany.

The transition zone of the mantle corresponds to the transformation of olivine in its high-pressure polymorphs, beta (wadsleyite) and gamma (ringwoodite). The role played by the transition zone in the global convection of the mantle depends largely in the relative rheological properties of these minerals. However, the deformation mechanisms of wadsleyite and ringwoodite under mantle conditions are still poorly known.

We present here an experimental investigation of plastic deformation of wadsleyite. Mg₂SiO₄ wadsleyite was synthesised with a multianvil apparatus, from a forsterite powder which was annealed 240 min at 18 GPa and 1600-1800°C. After the run, the wadsleyite sample was recovered and placed in a second high-pressure assembly designed to induce plastic deformation. This deformation experiment was carried out at 15 GPa and 1000°C during 90 min.

The deformation microstructures have been characterised by Transmission Electron Microscopy (TEM). The specimen is found to deform by dislocation creep. Most of the dislocations exhibit crystallographically controlled dislocation segments which points to glide motion. No subgrain boundaries are observed. The dislocations Burgers vectors have been determined by using Large Angle Convergent Beam Electron Diffraction (LACBED). This technique which uses a defocused beam is very well adapted to the study of electron beam sensitive materials like high-pressure phases. Four different Burgers vectors have been identified: [100] (0.57 nm), 1/2[111] (0.76 nm), [101] (0.10 nm) and [010] (0.11 nm). Dislocations with [100] and [010] Burgers vectors glide in the (001) plane. Dislocations with [010] Burgers vectors are not stable according to the Frank criterion. They are probably nucleated at the beginning of the experiment to relax the high stresses. As the experiment proceeds, the stress decreases (behaviour of a relaxation test) and the screw segments dissociate into two 1/2[111] dislocations gliding in {101} planes.