HR: 0800h
AN: V31A-1413 [Abstracts]
TI: Experimental Study of Trace Element Partitioning Between Immiscible Magmatic Brine and Granitic
Melt
AU: Veksler, I V
EM: veksler@gfz-potsdam.de
AF: GFZ Potsdam, Telegrafenberg, Potsdam, D-14473
Germany
AU: Dorfman, A M
EM: mandorf@mega.ru
AF: The Vernadsky Institute, Kosygin str., 19, Moscow, 119588
Russian Federation
AU: Dorfman, A M
EM: mandorf@mega.ru
AF: Earth and Environmental Sciences, University of Munich, Theresienstrasse 41/III, Munich, D-80333
Germany
AU: Dulski, P
EM: dulski@gfz-potsdam.de
AF: GFZ Potsdam, Telegrafenberg, Potsdam, D-14473
Germany
AU: * Dingwell, D B
EM: Dingwell@lmu.de
AF: Earth and Environmental Sciences, University of Munich, Theresienstrasse 41/III, Munich, D-80333
Germany
AB:
Highly concentrated magmatic brines can exsolve in silicate magmas of variable composition. The exsolutions are documented in
fluid and melt inclusions trapped in magmatic minerals during growth, and the occurrences of the brines are most common in
silicic magmas of granitic-rhyolitic composition. Here we present the results of experimental study of element partitioning
between immiscible brine (chloride melt) and aluminosilicate melt of granitic composition. Starting mixture was prepared from
equal amounts of synthetic haplogranitic glass and mixed reagent-grade chlorides of alkalis and alkaline earths. The mixture
was doped by trace elements, at about 300 ppm each. Experiments were performed in an internally heated pressure vessel
mounted on a centrifuge, which allowed an effective in situ centrifuge separation of immiscible melts during the run.
Quenched run products were split along the meniscus, and the chloride and the silicate layer analysed separately by bulk
solution-based ICP MS.
Two runs were carried out: the first one at 900 C, 100 MPa, no water added, and the second at 800 C, 100 MPa, with about 10
percent of water. The two-liquid Nernst partition coefficients (D) calculated as a ratio of weight concentrations in the
chloride melt and concentrations in the granitic melt were the following (first run - second run): Rb (2.3 - 2.7); Cs (2.8 -
3.3); Sr (6.1 - 10.2); Ba (6.0 - 9.7); Y (0.75 - 0.71); La (6.4 - 10.3); Sm (3.1 - 3.4); Eu (6.1 - 8.5); Gd (2.0 - 2.1); Ho
(0.97 - 0.93); Lu (0.53 - 0.51); Zr (0.001 - 0.003); Hf (0.0008 - 0.002); Pb (0.17 - 0.79). All the other REE were also
analysed, and their D values decrease almost linearly from La to Lu. D value of Eu is anomalous (by a factor of 2.4-3),
although oxygen fugacity was not controlled and Eu is likely to be in the oxidised trivalent form.
The addition of water increased D values of Zr and Hf by a factor of 2.5 and that of Pb by a factor of 4-5, while the other
elements included in the study were not much affected. In general, the solubility of HFSE in the brine appears to be very
low. The brine-silicate melt partitioning noticeably decouples the "geochemical twins", such as Y-Ho and Zr-Hf (see the D
values above). Thus, the separation of the brine in natural magmas and brine-induced metasomatism of country rocks may
produce a distinct trace element signature in volcanic and plutonic rocks affected by the processes.
DE: 3630 Experimental mineralogy and petrology
DE: 3640 Igneous petrology
DE: 3670 Minor and trace element composition
DE: 1065 Trace elements (3670)
SC: Volcanology, Geochemistry, Petrology [V]
MN: 2004 AGU Fall Meeting