Carbonate melts are critically important for the deep carbon cycle, mantle melting, redox reactions, and transport of highly incompatible elements. The presence of carbonate melts in the cratonic mantle has been inferred from experimental studies, metasomatic transformations, and melt/fluid inclusions in xenoliths, kimberlites, and diamonds. However, the exact composition of such melts is difficult to determine due to their ephemeral nature and highly reactive properties. Once formed, they migrate away from the source and react with silicate mantle minerals, especially orthopyroxene, causing mantle metasomatism. Wehrlite is one of the products of interaction between the carbonate melt and peridotitic mantle and hence is an excellent candidate for locating in situ carbonate melts. Here, we report petrological, geochemical, and melt inclusion data for a garnet wehrlite xenolith in the Majuagaa kimberlite dike, West Greenland. The xenolith, which last equilibrated with the mantle at 4.5 GPa and 1000 °C, contains abundant melt pools composed of dolomite, calcite, serpentine, spinel, apatite, and phlogopite. Although the original magmatic mineralogy was largely destroyed by low-temperature alteration, remnants of the crystallized carbonatitic melt are preserved as primary melt inclusions in the liquidus Ti-Mg-Fe spinel. These melt inclusions, composed of carbonates, alkali carbonates, periclase/brucite, and minor halides, K-sulfide, apatite, and phlogopite, are the first direct evidence for in situ alkali-carbonate melt in the deep cratonic mantle. Compositionally, they are very similar to primary Na-dolomite melt found in experiments and in fluid inclusions within diamonds.
Read full abstract