LAUSR

A comprehensive petrographic and in situ major and trace element study of rutile, ilmenite and Ti-magnetite was undertaken in six lower crustal xenoliths of metabasaltic (?underplate) and metasedimentary (subduction) origin from the Diavik kimberlites (central Slave Craton, Canada). The aims of the study were to improve our understanding of trace element incorporation into these Ti-minerals, and to use these systematics to obtain insights into lower continental crust formation and evolution. Abundant (oxy)exsolution of titanomagnetite lamellae, blocky rutile, as well as minor pleonaste and zircon in ilmenite from metabasaltic granulites are proposed to reflect cooling from magmatic or metamorphic temperatures and subsequent secular mantle cooling. This explains the large spread in Zr-in-rutile temperatures (>200 C) and may partly be responsible for the substantial heterogeneity of other trace element concentrations in rutile and ilmenite. Even after accounting for trace element heterogeneity and modal uncertainties, mass-balance calculations indicate that both Ti and Nb in lower crustal granulites are largely controlled by rutile and ilmenite. Rutile U–Pb data define discordia arrays that yield upper intercept ages broadly coincident with the 1 27 Ga giant Mackenzie dike swarm event, suggesting reheating of the lower crust above the rutile U–Pb closure temperature, whereas lower intercept ages roughly correspond to the age of Cretaceous to Eocene kimberlite magmatism. Subsequent cooling led to partial resetting and data spread along the concordia. Closer inspection reveals that inter-grain concentrations of elements that are compatible in rutile (Nb, Ta, W, U), but highly incompatible in the abundant silicate minerals (in equilibrium with melt), are heterogeneous and contrast with the more homogeneous concentrations of the transition metals (NiO, V). This may indicate that local reaction partners for diffusive homogenization of these element concentrations were absent. Nb/Ta is also highly variable at the sample scale. This may be explained by prograde growth from high-Nb/Ta mineral precursors (e.g. biotite) in the metasedimentary granulites and crystallization of the protoliths to the metabasaltic granulites from a mafic magma that had experienced fractionation of ilmenite with low Nb/Ta in a crustal magma chamber. Thus, (Fe)–Ti minerals represent high field strength element ‘islands’ in the granulite silicate matrix. The lack of homogenization and persistence of high-energy grain boundaries, such as exsolution lamellae, further indicate that the lower continental crust remained essentially dry and did not recrystallize, possibly since Neoarchaean metamorphism.