During Subduction Zone Metamorphism, Could the Transition from Aqueous Solutions to Silicate-Rich Hydrous Fluids (or Melt) be Supercritical in Nature?

Pascal Philippot CNRS, Laboratoire de Pétrologie, T26-E3, Université Paris 7, 4 place Jussieu, 75252 Paris 05, France


Segregation veins are common features of low-temperature eclogitic and high-grade blueschist rocks (1-3 GPa and 400-700°C) of the Alps. The veins contain the same mineral assemblage that the rock they contact implying that the material filling veins was deposited from saturated solutions derived from the host rocks. Vein minerals contain primary fluid inclusions. The fluid inclusions consist of mildly- to strongly-saline (0 to 50 wt % equivalent NaCl) aqueous brines locally in association with CO2 ± N2. The brines contain a large variety of daughter phases including silicate (quartz, omphacite, zoisite, kyanite, glaucophane, albite, talc, white mica, sphene), chloride (NaCl, KCl), carbonate (calcite, dolomite, magnesite), sulfate (barite, anhydrite/gypsum), sulphide (pyrite), oxide (baddeleyite, rutile, ilmenite, magnetite) and phosphate (monazite, Mg-phosphate, apatite). This argues for the high solubility of a variety of major (Si, Al, Mg, Fe, Ca, Na, K) and trace (Zr, Ti, Ba, Ce, La, Th...) elements at high pressures. Such fluids are best described in terms of variably hydrated salty-soup of SiO2, alkalis and metals.

A compilation of available experimental results on the solubility of various silicates in H2O and H2O-CO2 fluids at pressures above 1 GPa is presented (see Philippot (1996) for references). For a geothermal environment comparable to the one prevailing during subduction of Alpine rocks (~7°C.km-1), the data indicates that the total solute content of an aqueous fluid in equilibrium with a terrigeneous sedimentary component may exceed 50 weight % at ~60 km depths. Comparable level of silicate-material saturation can occur in aqueous fluids coexisting with lherzolite at a depth of ~150 km. Solubility contours of omphacitic clinopyroxene (i.e., basalt system) are subparallel to the trace of a typical prograde P-T path in a P-T diagram. This implies that any small temperature variation arising in a rock on the burial path should be accompanied by a marked change in omphacite solubility (from 25 to 30 weight% dissolved solutes for a temperature change of ~50 °C at 2.5 GPa).

A major difference between the experimental results and natural observations lies in the alkaline and/or siliceous character of the experimental fluids compare to the complex compositions of natural fluids. The majority of experimental data were performed for simple mineral aggregates in presence of H2O or H2O-CO2 fluids. The situation for natural systems is far more complex, however. The suite of daughter phases found in Alpine fluid inclusions indicates that a variety of complexing ligands (aluminosilicate, chloride, bicarbonate, sulfate, sulfide and phosphate) were available to concentrate major and trace elements in solutions. Competing ligands may explain that a variety of elements other than alkalies were soluble in the fluid. For example, the common presence of cm-scale rutile grain in eclogitic veins and of rutile and sphene daughter phases in fluid inclusions clearly shows that Ti was highly soluble in high-pressure fluids. The fact that this observation is in apparent contradiction with experimental results (Ayers and Watson, 1993) may lie in the much larger diversity of complexing ligands in the natural fluid than in the experimental one (where Ti dissolves as the neutrally-charged hydrolisis product Ti(OH)4). In addition, the presence of NaCl in natural fluids can affect major-element solute chemistry because alkali, alkaline earth and iron are well know to dissolve in fluids to form chloride complexes. Alternatively, addition of NaCl to aqueous fluid can result in an electrolyte effect that is likely to increase the solubilities of non-volatile phases (Ayers and Eggler, 1995). This effect will occur if NaCl is partially dissociated. Dissociation of solutes is largely dependent on the density of the fluids and therefore of pressure. In opposition, increasing temperature will favour association of solute species and the formation of stable complexes. Accordingly, it is to be expected that the high-pressure and low-temperature conditions prevailing during subduction zone metamorphism are likely to greatly facilitate dissociation of solutes and associated increase in element solubilities.

Although experimental studies on the solubility of minerals in the system H2O-CO2-NaCl are needed, the available data indicates that the transition from aqueous fluids to silicate-rich hydrous solutions (or melts) can be supercritical in nature at high-pressures. The volatile phases can extract substantial amounts of high-field strength elements (HFSE, Ti, Zr, Nb, Ta) from within high-pressure rocks. This together with the fact that they can form more easily in cold subducted slab than in hot-slab environment provide support for element recycling models involving a mantle column fluxed by a mobile phase with positive HFSE anomalies (Kelemen et al., 1990).


Ayers, J.C. & Watson, E.B., Contrib. Mineral. Petrol. 114, 321-330 (1993).

Ayers, J.C. & Eggler, D.H., Geochim. Cosmochim. Acta 59, 4237-4246 (1995).

Kelemen, P.B., Johnson, K.T.M., Kinzler, R.J. & Irving, A.J., Nature 345, 521-524 (1990).

Philippot, P., Lithos (1996, in press).