Revisiting the Nd-142 deficits in the 1.48 Ga Khariar alkaline rocks, India

• Published in: Chemical geology 2014-10, Vol.386, p.238-248
• Main Authors: Roth, Antoine S. G., Scherer, Erik E., Maden, Colin, Mezger, Klaus, Bourdon, Bernard
• Format: Article
• Language: English
• Subjects: Sciences of the Universe
• ISSN: 0009-2541
• DOI: 10.1016/j.chemgeo.2014.06.022

The Sm-146-Nd-142 systemplays a central role in tracing the silicate differentiation of the Earth prior to 4.1 Ga. After this time, given its initial abundance, the Sm-146 can be considered to be extinct. Upadhyay et al. (2009) reported unexpected negative Nd-142 anomalies in 1.48 Ga rocks of the Khariar nepheline syenite complex (India) and inferred that an early enriched, low-Sm/Nd reservoir must have contributed to the mantle source rocks of the Khariar complex. As Sm-146 had been effectively extinct for about 2.6 billion years before the crystallisation of the Khariar samples, this Nd signature should have remained isolated from the convective mantle for at least that long. It was thus suggested that the source rock of Khariar samples had been sequestered in the lithospheric root of the Indian craton. Using a different chemical separation method, and a different Thermal Ionization Mass Spectrometry (TIMS) analysis protocol, the present study attempted to replicate these negative Nd-142 anomalies, but none were found. To determine which data set is correct, we investigated three possible sources of bias between them: imperfect cancellation of Faraday collector efficiencies during multidynamic TIMS analysis, rapid sample fractionation between the sequential measurement of Nd-146/Nd-144 and Nd-142/Nd-144, and nonexponential law behaviour resulting from so-called "domain mixing." Incomplete cancellation of collector efficiencies was found unlikely to cause resolvable biases at the estimated level of variation among collector efficiencies. Even in the case of highly variable efficiency and resolvable biases, there is no reason to suspect that they would reproducibly affect only four rocks out of 10 analysed by Upadhyay et al. (2009). Although domain mixing may explain apparent "reverse" fractionation trends observed in some TIMS analyses, it cannot be the cause of the apparent negative anomalies in the study of Upadhyay et al. (2009). It was determined that rapid mass fractionation during the course of a multidynamic TIMS analysis can bias all measured Nd ratios. After applying an approximate correction for this effect, only one rock from Upadhyay et al. (2009) retained an apparent negative Nd-142 anomaly. This, in conjunction with our new, anomaly-free data set measured at fractionation rates too low to cause bias, leads to the conclusion that the anomalies reported by Upadhyay et al. (2009) are a subtle and reproducible analytical artefact. The absence of