ABSTRACT
The Indaiá-I and Indaiá-II intrusions are hypabyssal, small-sized ultrabasic bodies belonging to the Cretaceous magmatism of the Alto Paranaiba Alkaline Province (southeast-central western Brazil). While Indaiá-I is classified as an archetypal group-I kimberlite, Indaiá-II (its satellite intrusion) presents several petrographic and chemical distinctions: (1) an ultrapotassic composition (similar to kamafugites), (2) lower volumes of olivine macrocrysts, (3) diopside as the main matrix phase (in contrast with the presence of monticellite in Indaiá-I), (4) high amounts of phlogopite, and (5) abundant felsic boudinaged and stretched microenclaves and crustal xenoliths. Disequilibrium features, such as embayment and sieve textures in olivine and clinopyroxene grains, are indicative of open-system processes in Indaiá-II. Mineral reactions observed in Indaiá-II (e.g., diopside formed at the expense of monticellite and olivine; phlogopite nearby crustal enclaves and close to olivine macrocrysts) point to an increase in the silica activity of the kimberlite magma; otherwise partially melted crustal xenoliths present kalsilite, generated by desilification reactions. The high Contamination Index (2.12–2.25) and the large amounts of crustal xenoliths (most of them totally transformed or with evidence of partial melting) indicate a high degree of crustal assimilation in the Indaiá-II intrusion. Calculated melts (after removal of olivine xenocrysts) of Indaiá-II have higher amounts of SiO2, Al2O3, K2O, slightly higher Rb/Sr ratios, lower Ce/Pb and Gd/Lu ratios, higher 87Sr/86Sr, and lower 143Nd/144Nd than those calculated for Indaiá-I. Crustal contamination models were developed considering mixing between the calculated melts of Indaiá-I and partial melts modeled from the granitoid country rocks. Mixing-model curves using major and trace elements and isotopic compositions are consistent with crustal assimilation processes with amounts of crustal contribution of ca. 30%. We conclude that (1) Indaiá-II is representative of a highly contaminated kimberlitic intrusion, (2) this contamination occurred by the assimilation of anatectic melts from the main crustal country rocks of this area, and (3) Indaiá-I and Indaiá-II could have had the same parent melt, but with different degrees of crustal contamination. Our petrological model also indicates that Indaiá-II is a satellite blind pipe linked to the main occurrence of Indaiá-I.