New Nd (0.51261–0.51268), Pb (²⁰⁶Pb/²⁰⁴Pb: 19.24–19.26), and Sr (0.70437–0.70446) isotopic compositions from ten natrocarbonatite lavas, collected in June 1993 from Oldoinyo Lengai, the only known active carbonatite volcano, are relatively uniform, and are similar to data from the 1960 and 1988 flows. Three of the samples contain silicate spheroids, one of which has Nd and Sr isotopic ratios similar to host natrocarbonatite, consistent with an origin by liquid immiscibility or the mixing of melts with similar isotopic compositions. Pb isotope data for two samples of trona are inconsistent with its involvement in the genesis of natrocarbonatite. New Pb isotope data from silicate volcanic and plutonic blocks (ijolite, nephelinite, phonolite, syenite) from Oldoinyo Lengai are highly variable (²⁰⁶Pb/²⁰⁴Pb, 17.75–19.34; ²⁰⁷Pb/²⁰⁴Pb, 15.41–15.67; ²⁰⁸Pb/²⁰⁴Pb, 37.79–39.67), and define near-linear arrays in Pb-Pb diagrams. The isotopic data for the silicate rocks from Oldoinyo Lengai are best explained by invoking discrete partial melting events which generate undersaturated alkaline silicate magmas with distinct isotopic ratios. Pb isotope ratios from most ijolites and phonolites are predominantly lower and more variable than from the natrocarbonatites, and are attributed to interaction between silicate melts involving HIMU and EMI source components and an additional component, such as lower-crustal granulites, DMM or PREMA (prevalent mantle). Variations in Nd, Pb and Sr isotope ratios from Oldoinyo Lengai, among the largest yet documented from a single volcano, are attributed to mantle source heterogeneity involving mainly the mixing of HIMU and EMI mantle components. Based on the new isotopic data from Oldoinyo Lengai and data from other East African carbonatites, and mantle xenoliths, we propose a two-stage model in an attempt to explain the isotope variations shown by carbonatites in this area. The model involves (I) the release of metasomatizing agents with HIMU-like signatures from upwelling mantle (‘plume’) source, which in turn metasomatize the sub-continental (old, isotopically enriched, EMI-like) lithosphere, and (2) variable degrees and discrete partial melting of the resulting heterogeneous, metasomatized lithosphere.

This study reports U–Pb geochronological data for zircon obtained by laser ablation-multi-collector-ICP-MS using a new collector block design that includes three ion counters and twelve Faraday buckets. The collector configuration allows for simultaneous detection of ion signals from mass ²³⁸U to ²⁰³Tl, an important factor for the achievement of highly precise and reproducible Pb–Pb and Pb–U ratios. The main advantage of the multiple ion counting system is the capability to readily measure low Pb ion signals (<1 × 10⁶ counts per second) resulting from single spot analyses of ≤40 microns (and corresponding small sample volumes). The latter is an extremely important feature when deciphering multiple domains in complexly zoned zircon populations. A comparative study was conducted between analytical protocols involving non-Tl- and Tl-doped laser ablation runs with regards to evaluating the external reproducibility (i.e., relative standard deviation). The results indicate that the (2σ) external reproducibility for both analytical protocols varies between ∼0.3–1.0% for the ²⁰⁷Pb/²⁰⁶Pb, and ca. 1.3 to <3.0% for the ²⁰⁶Pb/²³⁸U (and ²⁰⁷Pb/²³⁵U) values. The external reproducibility for the Pb/U values are relatively constant for the Tl-doped analyses, whereas those for the non-Tl runs are inversely correlated with the total absolute (Pb) count rates. The precision and accuracy of both analytical protocols were verified with the analysis of BR266 and 91500 zircon standards and zircons previously dated by ID-TIMS. The instrument configuration and Tl-doping protocol employed here are ideally suited for studies requiring the rapid analysis of a large number (n > 50) of zircon grains.