Experimental constraints on mantle metasomatism caused by silicate and carbonate melts

Abstract

Metasomatic processes are responsible for many of the heterogeneities found in the upper mantle. To better understand the metasomatism in the lithospheric mantle and to illustrate the differences between metasomatism caused by hydrous silicate and carbonate-rich melts, we performed various interaction experiments: (1) Reactions between hydrous eclogite-derived melts and peridotite at 2.2–2.5GPa and 900–1000°C reproduce the metasomatism in the mantle wedge above subduction zones. (2) Reactions between carbonate-rich melts and peridotite at 2.5GPa and 1050–1000°C, and at 6GPa and 1200–1250°C simulate metasomatism of carbonatite and ultramafic silicate–carbonate melts in different regions of cratonic lithosphere. Our experimental results show that partial melting of hydrous eclogite produces hydrous Si- and Al-rich melts that react with peridotite and form bi-mineralic assemblages of Al-rich orthopyroxene and Mg-rich amphibole. We also found that carbonate-rich melts with different compositions react with peridotite and form new metasomatic wehrlitic mineral assemblages. Metasomatic reactions caused by Ca-rich carbonatite melt consume the primary peridotite and produce large amounts of metasomatic clinopyroxene; on the other hand, metasomatism caused by ultramafic silicate–carbonate melts produces less clinopyroxene. Furthermore, our experiments show that ultramafic silicate–carbonate melts react strongly with peridotite and cause crystallization of large amounts of metasomatic Fe-Ti oxides. The reactions of metasomatic melts with peridotite also change the melt composition. For instance, if the carbonatite melt is not entirely consumed during the metasomatic reactions, its melt composition may change dramatically, generating an alkali-rich carbonated silicate melt that is similar in composition to type I kimberlites.