ABSTRACT Recent studies have suggested various mechanisms to generate melt during exhumation of deeply subducted crust, including dehydroxylation of omphacite, fluid‐absent omphacite‐ and/or phengite‐breakdown melting and fluid‐present melting. However, the space–time relationships among these mechanisms remain uncertain. Here, we report microstructural evidence of the reactions responsible for, and compositions of, low‐volume melts frozen in situ (as leucosome pockets) within weakly deformed granitic veins that cross‐cut foliated UHP eclogite boudins in gneisses from the central Sulu belt, China. Phengite in the granitic veins records crystallization pressures of 3.4–2.7 GPa, and the granitic veins have whole‐rock Sr–Nd isotope compositions and trace element patterns consistent with derivation from the eclogite. The granitic veins likely crystallized from a solute‐rich supercritical fluid or hydrous melt generated by dehydroxylation of nominally anhydrous minerals during the early stage of exhumation (decompression). Subsequently, leucosome pockets and grain‐boundary films of melt formed in the granitic veins by the successive breakdown of omphacite and phengite. Based on the mineral modes and chemical compositions of 55 leucosome pockets (cf. microgranitoids), we distinguish those with (1) high Na/K ratios mainly composed of plagioclase and euhedral amphibole with skeletal omphacite and (2) low Na/K ratios predominantly composed of K‐feldspar and plagioclase and which contain phengite with corroded margins and fine‐grained biotite. High Na/K leucosome pockets (HLPs) are consistent with a melting reaction involving mostly omphacite‐breakdown, whereas low Na/K leucosome pockets (LLPs) are inferred to have formed by a reaction consuming variable proportions of phengite and omphacite. We argue that the reactions to form HHP then LLP initiated at approximately 1.5 and 1.1 GPa, respectively, and document two different mechanisms to generate melt during exhumation of deeply subducted crust in which breakdown of omphacite occurs before (deeper than) phengite. At higher temperatures, these melt‐producing reactions would generate a larger volume of melt that potentially could facilitate exhumation and increase crust–mantle interactions, thereby increasing the compositional heterogeneity of orogenic mantle.
Feng et al. (Fri,) studied this question.