Oxygen Fugacity Dependence of Osmium Partitioning between Metal and Melt: A Study Using the Mechanically Assisted Equilibration Technique

Sophie Fortenfant (sophie.fortenfant @uni-bayreuth.de)¹, Donald Dingwell (don.dingwell@uni-bayreuth.de)¹, Jean Louis Birck (birck@ipgp.jussieu.fr)² & Pierre Schiano²

¹ BGI, Universitaet Bayreuth, Germany

² IPGP, Université de Paris VI, France

Borisov and Palme (1997) have pioneered Os partitioning at 1 atm under controlled oxygen fugacity (fO₂) using the loop technique [3]. Their analyses of the glassy experimental products were performed by INAA. The partition coefficients appear to be independent of fO₂ and show considerable scatter. These results may be indicative of technical problems with the loop technique as previously discussed (presence of "micronuggets" in the quenched silicate liquid and uncertainty of chemical equilibration between metal and silicate) [1,3]. To improve those preliminary results, Borisov and Walker (1998) performed new analyses of the samples using the Isotopic Dilution technique [4]. For very low oxygen fugacity (log fO₂<-8), no improvement in the results is obtained. For more oxidizing conditions (log fO₂>-8), the metal-melt partition coefficients of osmium show a clearer dependence on fO₂.

To complement those data and to circumvent potential disadvantages of the loop technique, the present investigation uses the mechanically assisted equilibration technique [5]. After quenching, the glass samples are analyzed by microprobe for major elements (Si, Al, Ca, Mg) and Ni, and ID-NTIMS for Os [6].

Our results are comparable with those of Borisov and Walker (1998). The same scatter in the data has been observed for very reducing experimental conditions. This may suggest significant experimental contamination problems and a duplication of the analysis using LA-ICPMS technique is in planning. DM/S values deduced from the present work are still much too high to explain abundances of Os in the mantle, assuming metal-silicate equilibrium during core formation.

Borisov A & Palme H, G.C.A., 59, 481-485, (1994).

Borisov A & Palme H, Neues Jahrbuch für Mineralogie, 172, 347-356, (1998).

Borisov A & Walker RJ, OEM 98, poster presentation, (1998).

Dingwell DB, G. C. A, 58, 1967-1974, (1994).

Birck JL & al, Geostandards Newletter, 20, 19-27, (1997).

Compositional Evolution of Amphibole During the Retrograde Re-equilibration of Mafic Amphibolitic Rocks from NE Sardinia, Italy

Marcello Franceschelli (francmar@unica.it)¹, Gianfranco Carcangiu¹, Anna Maria Caredda (acaredda@unica.it)¹, Gabriele Cruciani¹, Isabella Memmi³ & Marco Zucca¹

¹ Dip. Scienze della Terra, Via Trentino, 51, 09127 Cagliari, Italy

² Centro Studi Geominerari e Mineralurgici C.N.R., P.zza d'Armi, 16, 09127 Cagliari, Italy

³ Dip. Scienze della Terra, Via delle Cerchia, 3, 53100 Siena, Italy

Lenses of mafic amphibolitic rocks with relics of granulite facies assemblages hosted in the Hercynian migmatite of NE Sardinia show a great amphibole variety. Based on morphological and textural features, five types of amphiboles have been distinguished: large brown (Cam1) or green (Cam2) clinoamphiboles; colourless amphiboles (Cam3) replacing olivine (Ol) and orthopyroxene (Opx); colourless orthoamphiboles (Oam1) interleaved with Mg-chlorite (Chl); pale green clinoamphiboles (Cam4), replacing or rimming Cam2 amphiboles. Texturally different amphiboles have a different chemistry. Cam1 ranges in composition from tschermakite to magnesio-hornblende; Cam2 has a composition that varies from tschermakite to magnesio-hornblende as far as actinolitic-hornblende. Cam3 with a low Ca content and X_Mg = 0.70 is dubitatively classified as a cummingtonite; Cam4 is an actinolite with X_Mg ranging from 0.60 to 0.70. Oam1 is an anthophyllite with low Al content and X_Mg = 0.75-0.80. Cam1 and Cam2 amphiboles formed through the destabilization of the garnet (Grt) - pyroxene pair (Cpx and Opx) via the following reaction: Grt + Cpx + Opx + H₂O = Cam1,2 + Spl ± Mg-Chl In the garnet-free amphibolite, the Cam2 amphibole grew on the orthopyroxene or on the clinopyroxene in textural equilibrium with the Mg-chlorite via the following reaction: Ol + Cpx + Opx + H₂O = Cam2 + Spl + Mg-Chl The late stage actinolitic amphibole, coexisting with chlorite and epidote (Ep) may be formed by the simplified reaction: Cam2 + H₂O = Ep + Mg-Fe Chl + Cam4 + Qtz The composition of the Ca-amphiboles shows great chemical variation: in particular Cam1 shows the highest Na/Na+Ca (0.28) and Al/Al+Si (0.37) ratios, decreasing gradually towards Cam2 and Cam4, where they are respectively ~0.05 and ~0.06. The compositional variation of the amphiboles can be related to the retrograde P-T evolution from the amphibolite to the greenschist stage during regional Hercynian metamorphism.

Measurement of Hydrogen Fugacity from 50 to 350°C Using Palladium Membranes; Measurement of the Permeability of Au₅₀Pd₅₀

John Frantz (jfrantz@cnrs-orleans.fr)¹ & Jacques Roux (jroux@cnrs-orleans.fr)²

¹ Geophysical Laboratory, 5251 Broadbranch Rd., NW, Washington, D.C., USA

² ISTO, CNRS-Université d'Orléans, 1A rue de la Férollerie, 45071 Orléans cedex 2, France

The use of semi-permeable membranes in the control and measurement of hydrogen fugacity in experimental geochemistry has, for the most part, been restricted to temperatures above 500°C. Considering the minimal dependence of temperature on the permeability of palladium alloys, use of membranes semipermeable to hydrogen should be feasible to much lower temperatures. A systematic study was initiated to test the applicability of these methods in the temperature range from 20 to 350°C. The permeability of Au₅₀Pd₅₀ membranes was determined between 50 and 350°C using pure hydrogen gas. Substantial mass transport of hydrogen occurred throughout the temperature range. The dependence of permeability on pressure was investigated using argon-hydrogen mixtures. The effect of water on the hydrogen flux across the membrane surface was investigated using water-hydrogen mixtures. The measured permeabilities of hydrogen were then used to theoretically predict the flux of hydrogen across gold palladium membranes as a function of hydrogen fugacity, temperature, and membrane configurations. The rate of equilibration was found to dramatically improve with decreasing hydrogen fugacity and decreasing membrane volume. The simulations were used to modify the membrane configuration used in monitoring hydrogen. The resulting design permits measurement of the hydrogen fugacity to temperatures well below 100°C. Equilibration times ranged from less than a minute to a few minutes depending on temperature and hydrogen fugacity. Using these techniques, future studies should include low-temperature mineral-fluid equilibria, reactions involving organic compounds, and monitoring of the hydrogen fugacity of natural fluids such as those emanating from hydrothermal vents.