Relationships between mineralogical, chemical, and isotopic properties of some North American kimberlites

Late Precambrian to Cretaceous kimberlites from North America have initial Sr and Nd isotopic compositions (εSr0 = −4 to 37, εNd0 = −4 to 6) which form a trend falling roughly within the mantle array. Two groups can be defined on the basis of mineralogical and geochemical characteristics: group A encompasses most of the northeast American kimberlites whose rare earth element (REE) concentrations are moderately fractionated, relative to chondrites. Nd isotopic ratios (εNd0 = 3 to 6) fall within the field of common ocean island basalts. They are also distinctive by the presence of phlogopite phenocrysts and high‐temperature garnet peridotite xenocrysts. Group B includes kimberlites from the Colorado‐Wyoming, Kansas and British Columbia areas and one sample from Pennsylvania. They lack the above quoted peridotite suite and abundant phlogopite. Their light REE concentrations show a strong enrichment, relative to chondrites, correlated with relatively unradiogenic Nd (εNd0 = −4 to 2). Kimberlites of both groups have radiogenic Pb (206Pb/204Pb = 18.17–19.41) and a wide range in 207Pb/204Pb. In both cases, melts were probably generated in the sublithospheric mantle, but group B melts interacted with an enriched component, resulting mainly from magmatic processes, and probably located within the deep lithosphere. By contrast, the Prairie Creek lamproite (Arkansas), as other ultrapotassic rocks, received its distinctive geochemical properties from a crustal component present in the mantle source.