EXSOLUTION OF A PHLOGOPITE-10Å PHASE SOLID SOLUTION FROM DIOPSIDE IN UHPM GARNET PERIDOTITE FROM SULU, CHINA

For many years, a poorly-defined “10Ŕ phyllosilicate has been observed in high-pressure experiments in the system Mg₂SiO₄ + H₂O. Although never described from natural rocks, this phase is believed to be important in subduction zones as a water-carrying phase with a stability field that bridges the gap in “warm” subduction zones between antigorite and phase A.

Here we describe a phyllosilicate with average composition K_0.66(Mg_2.60Fe_0.10Al_0.20Ti_0.02Cr_0.03)(Si_3.10Al_0.90)O₁₁ exsolved from diopside (topotaxy: (100)_di=(001)_mica) in a garnet peridotite from the SuLu UHPM terrane, China. The K content varies significantly within and between crystals from 0.4 to 0.8 apfu, corresponding to an approximately 60/40 molar ratio of the end-members phlogopite and 10Å phase. The abundance of the lamellae, as estimated from optical and scanning electron microscopy, varies between 0.1 to 1 vol %. HRTEM imaging and electron diffraction show that polytypes with different stacking are present, along with the most abundant 1M polymorph. Within individual crystals, domains with random stacking are observed as well as isolated occurrences of domains with periodicity corresponding to 12M and 15M polytypes. The combined chemical and diffraction information rules out all known phases except phlogopite and 10Å phase, with the paucity of K indicating a solid solution between these two phases.

The observed amount of phyllosilicate could have been exsolved from a diopside host which had incorporated homogeneously distributed hydrogen in its crystal lattice in the range of 90 to 900 ppm. The microstructural context in which this mineral is found suggests also that the diopside host in its turn was originally dissolved in a former supersilicic (majoritic) garnet whose reconstructed composition suggests depth of origin significantly exceeding 200 km.

Identification of this phlogopite - “10Å phase” solid solution strongly suggests that the supposed gap in ability of mantle peridotite to carry water to depth in warm subducting slabs (as a consequence of a gap between stability of antigorite and phase A) does not exist. Thus, in subducting lithosphere, very significant amounts of H₂O could be carried through this “gap” in a phase comparable to that described here, with the K content highly variable, dependent on the bulk chemistry of the rock.