Rutile inclusions in garnet from a dissolution‐reprecipitation mechanism

Metamorphic garnet commonly contains needle‐like rutile inclusions as well as equant rutile inclusions that surround quartz inclusions and range in size from submicrometer to nanometer. Although the origin of these equant rutile inclusions, that is, exsolution or non‐exsolution, has important implications for petrological and tectonic processes, the crystallographic characteristics of these inclusions have rarely been studied because of the small sizes and analytical difficulties involved. Here, we report the crystallographic characteristics pertinent to the genetic origin of minute equant rutile inclusions in cloudy, nearly spherically shaped garnet domains with Ti‐depleted compositions surrounding quartz inclusions in ultrahigh‐pressure garnet from several diamondiferous Erzgebirge quartzofeldspathic gneissic rock samples. TEM analyses show that the equant rutile crystals in cloudy garnet domains are partially bounded by the low‐energy {100}rt ± {110}rt ± {101}rt facets and have rather random crystallographic orientation relationships (CORs) with the garnet host, with preferential alignment of low‐energy lattice planes, for example, {100}rt//{112}grt, for some rutile crystals. Although the rather random CORs are unlikely to be attributed to solid‐state exsolution subjected to the stringent topotactic garnet lattice constraints, the characteristic subhedral {100}rt ± {110}rt ± {101}rt crystal forms of rutile can be rationalized by a metasomatic dissolution‐reprecipitation mechanism via a fluid phase. In this scenario, the quartz+fluid inclusions in garnet were first subjected to decompression microcracking during rock exhumation, followed by dissolution of Ti‐bearing garnet matrix at the crack tips or along the crack surfaces and subsequent reprecipitation of rutile, apatite, gahnite, akdalaite, and Ti‐depleted garnet. The rapid coalescence between rutile and garnet crystals in fluid or direct attachment of rutile crystals onto the dissolving crack surfaces would then yield the rather random CORs as reported here. These results, along with previous work on rutile needles, indicate rather diverse genesis of rutile inclusions in various crystal forms, thus shedding light on the controversial exsolution origin for other inclusion suite/microstructure in minerals.