Carbonates in fresh hypabyssal kimberlites worldwide have been studied to understand their origin [i.e. primary magmatic (high T) versus deuteric (‘low T’) versus hydrothermal/alteration (‘low T’)] and identify optimal strategies for petrogenetic studies of kimberlitic carbonates. The approach presented here integrates detailed textural characterisation, cathodoluminescence (CL) imaging, in situ major- and trace-element analysis, as well as in situ Sr-isotope analysis. The results reveal a wide textural diversity. Calcite occurs as fine-grained groundmass, larger laths, segregations, veins or as a late crystallising phase, replacing olivine or early carbonates. Different generations of carbonates commonly coexist in the same kimberlite, each one defined by a characteristic texture, CL response and composition (e.g., variable Sr and Ba concentrations). In situ Sr isotope analysis revealed a magmatic signature for most of the carbonates, based on comparable ⁸⁷Sr/⁸⁶Sr values between these carbonates and the coexisting perovskite, a robust magmatic phase. However, this study also shows that in situ Sr isotope analysis not always allow distinction between primary (i.e., magmatic) and texturally secondary carbonates within the same sample. Carbonates with a clear secondary origin (e.g., late-stage veins) occasionally show the same moderately depleted ⁸⁷Sr/⁸⁶Sr ratios of primary carbonates and coexisting perovskite (e.g., calcite laths-shaped crystals with ⁸⁷Sr/⁸⁶Sr values identical within uncertainty to those of vein calcite in the De Beers kimberlite). This complexity emphasises the necessity of integrating detailed petrography, geochemical and in situ Sr isotopic analyses for an accurate interpretation of carbonate petrogenesis in kimberlites. Therefore, the complex petrogenesis of carbonates demonstrated here not only highlights the compositional variability of kimberlites, but also raises concerns about the use of bulk-carbonate C-O isotope studies to characterise the parental melt compositions. Conversely, our integrated textural and in situ study successfully identifies the most appropriate (i.e. primary) carbonates for providing constraints on the isotopic parameters of parental kimberlite magmas.

Five compositional-textural types of ilmenite can be distinguished in nine kimberlites from the Eastern Dharwar craton of southern India. These ilmenite generations record different processes in kimberlite history, from mantle to surface. A first generation of Mg-rich ilmenite (type 1) was produced by metasomatic processes in the mantle before the emplacement of the kimberlite. It is found as xenolithic polycrystalline ilmenite aggregates as well as megacrysts and macrocrysts. All of these ilmenite forms may disaggregate within the kimberlite. Due to the interaction with low-viscosity kimberlitic magma replacement of pre-existing type 1 ilmenite by a succeeding generation of geikielite (type 2) along grain boundaries and cracks occurs. Another generation of Mg-rich ilmenite maybe produced by exsolution processes (type 3 ilmenite). Although the identity of the host mineral is unclear due to extensive alteration and possibility includes enstatite. Type 4 Mn-rich ilmenite is produced before the crystallization of groundmass perovskite and ulvöspinel. It usually mantles ilmenite and other Ti-rich minerals. Type 5 Mn-rich ilmenite is produced after the crystallization of the groundmass minerals and replaces them. The contents of Cr and Nb in type 2, 4 and 5 ilmenites are highly dependent on the composition of the replaced minerals, they may not be a good argument in exploration. The highest Mg contents are recorded in metasomatic ilmenite that is produced during kimberlite emplacement, and cannot be associated with diamond formation. The higher Mn contents are linked to magmatic processes and also late processes clearly produced after the crystallization of the kimberlite groundmass, and therefore ilmenite with high Mn contents cannot be considered as a reliable diamond indicator mineral (DIM) and kimberlite indicator mineral (KIM).