Comment on “The beginnings of hydrous mantle wedge melting”, CB Till, TL Grove, AC Withers, Contributions to Mineralogy and Petrology, DOI 10.1007/s00410-011-0692-6

Abstract

The melting of the Earth’s upper mantle is the basic process behind volcanism. In the previous decades, many experimental studies on mantle melting have been performed, and a consensus has been reached that water lowers the solidus temperature dramatically and that water solubility in silicate melts depends strongly on pressure. In a recent article, Till et al. (2012) discuss their findings from high-pressure experiments and place the solidus of a primitive mantle peridotite between 800 and 820 C at pressures between 3 and 6 GPa (corresponding to depths of approximately 100–200 km). This result is in contrast to many previous publications, where the solidus was placed between 1,000 and 1,100 C (Kushiro et al. 1968; Kawamoto and Holloway 1997; Green et al. 2010). Generally, the most severe problem for the interpretation of high-pressure melting experiments is the unequivocal distinction between aqueous fluid (existing in water bearing systems below the solidus) and hydrous silicate melts (occurring above the solidus). The reason for this difficulty is the tendency that aqueous fluids dissolve increasing amounts of silicate material with increasing pressure, while hydrous melts dissolve increasing amounts of water. Finally, both phases eventually converge, where a supercritical phase is formed (Kennedy et al. 1962; Stalder et al. 2001; Kessel et al. 2005; Mibe et al. 2007), and the solidus is no longer defined. Criteria to discriminate between aqueous fluid and hydrous melt are either based on textural and chemical evidence of the quenched material or based on direct observation in in situ studies. The major problem of the study by Till et al. (2012) is the lack of unequivocal criteria to discriminate a fluid from a melt. The attribution of low porosity below 810 C to subsolidus and high porosity above 810 C to above solidus conditions is subject to alternate explanation (Green et al. 2012). The second main criterion applied is the shape of the quenched solute (Fig.6: spherical objects = ‘‘quenched solute from a fluid’’ and wisps = ‘‘quenched solute from a melt’’). This criterion is also used by Adam et al. (1997) and is applied here to identify 2 fluid phases (hydrous silicate melt and solute-bearing aqueous vapour). It is noteworthy that the evidence presented for two fluid phases as well as the analyses of silicate glass is from experiments at [1,000 C. In water–silicate systems, the solute content of aqueous fluids is positively correlated with temperature. Therefore, a subsolidus fluid tends to release dissolved material upon quenching. Similarly to dry quenched melts, where quench crystals and glass may be observed, amorphous material and quench crystals may be present after quenching a solute-saturated fluid phase. Consequently, the observation of differently shaped objects in the quench material does not imply that more than one phase was present before quenching. Other criteria applied in the study are not very convincing either. To me, the most prominent textural change in Fig. 2 occurs at 1,100 C, and the chemical changes documented are not unequivocal (except the Crand Ti-content of Opx above 1,100 C). Finally, the major chemical parameter of the melt, for example, the water to silicate ratio, is not quantified, and therefore, the distinction between fluid and melt in Till This comment refers to the article available at 10.1007/s00410-011-0692-6. An author’s reply to this comment is available at 10.1007/s00410-012-0797-6.