Exsolution of rutile or apatite precipitates surrounding ruptured inclusions in garnet from UHT and UHP rocks

Abstract

AbstractThis study describes halos of rutile ± apatite needles and/or plates centred on quartz (inferred former coesite) inclusions or multiphase inclusions in garnet. Both types of central inclusions are surrounded by cracks. The multiphase inclusions contain mica or carbonate minerals and are interpreted to represent decrepitated fluid inclusions. Examples from two localities are examined: (i) ultrahigh‐temperature (UHT) metapelitic gneisses from the southern end of the Central Maine Terrane in northeastern Connecticut, USA (rutile only) and (ii) ultrahigh‐pressure (UHP) diamondiferous saidenbachite from the Saxonian Erzgebirge (rutile and apatite). The needles of apatite or rutile are typically oriented in three directions within garnet. Chemical zonation in garnet shows clear Ti or P depletion halos corresponding spatially with the rutile or apatite inclusion halos. The radii of the inclusion halos, when measured from the centre of the central inclusions, are about two to several times the radii of the central inclusions. We propose that the inclusion halos of rutile ± apatite formed by exsolution out of Ti‐bearing and/or P‐bearing garnet during retrogression. Because of its strength, garnet can internally maintain higher pressures than the surrounding matrix. Nonetheless, if the inclusion pressure is significantly greater than the confining lithostatic pressure, then deformation of the host garnet can occur. During exhumation, high internal pressures relative to the matrix can result from retention of fluid entrapment pressure or a phase change with a large positive volume increase (e.g. coesite → quartz). The differential stress imposed on the garnet adjacent to the central inclusion preceding and during mechanical failure would create dislocations and other crystalline defects which are ideal sites for exsolved precipitates to nucleate and grow. The strain‐induced exsolution hypothesis is consistent with observations. (i) The radii of the halos are roughly consistent with the pressure vessel model or the multi‐anvil model for an inclusion, which predict that differential stress will drop off sharply away from the boundary of the central inclusion into the host garnet. (ii) The rutile or apatite inclusions formed in Ti‐ or P‐bearing garnet adjacent to ruptured central inclusions; rutile or apatite are rare or absent elsewhere in garnet, even in areas of high‐Ti or high‐P concentration. In addition, rutile is absent in areas surrounding the central inclusion that were depleted in Ti prior to rupturing (e.g. garnet rims that lost Ti during retrogression). Thus, both elevated Ti ± P concentrations and proximity to inclusions which deformed the host garnet were necessary for halo formation. (iii) Reintegrated garnet Ti or P contents in and adjacent to the halos obtained using wide‐beam electron probe microanalysis are consistent with local derivation of the Ti or P necessary to form rutile and apatite from garnet. Elevated Ti ± P concentrations in aluminous garnet are mostly found in UHT, HP/UHP, granulite and mantle environments. Consequently, the phenomena described will most likely occur in such settings.