International Journal of Mass Spectrometry 309 (2012) 109–117 Contents lists available at SciVerse ScienceDirect International Journal of Mass Spectrometry j our na l ho me page: www.elsevier.com/locate/ijms High precision tungsten isotope measurement by thermal ionization mass spectrometry Mathieu Touboul ∗ , Richard J. Walker Department of Geology, University of Maryland, College Park, MD 20742, USA a r t i c l e i n f o Article history: Received 21 June 2011 Received in revised form 31 August 2011 Accepted 31 August 2011 Available online 10 September 2011 Keywords: Tungsten isotope Extinct nuclide 182 W Thermal ionization Triton N-TIMS a b s t r a c t We describe a new technique for measuring the isotopic abundance of 182 W with improved precision in natural silicate samples. After chemical purification of W through a four-step ion exchange chromato- graphic separation, the W isotopic composition is measured as WO 3 - by negative thermal ionization mass spectrometry using a Thermo-Fisher Triton instrument. Amplifier biases are cancelled by using amplifier rotation, and Faraday cup biases are monitored by using a cup configuration that allows two-line data acquisition. Data are initially corrected for oxide interferences, assuming a predefined O isotope composi- tion, and for mass fractionation, by normalization to 186 W/ 184 W or 186 W/ 183 W, using an exponential law. Despite these corrections, isotopic ratios exhibit small but strongly correlated variations. This second- order effect may reflect a mass dependent change of O isotope composition in the measured W (and Re) oxides, and is corrected by normalization to 183 W/ 184 W using a linear law. Repeated analysis of an Alfa Aesar W standard (n = 39), and of three dissolutions of a La Palma (Canary Islands) basalt, applying the double normalization procedure, demonstrate external reproducibility of 182 W/ 184 W within ±4.5 ppm (2 SD). Repeated measurement of a gravimetrically prepared mixture of a natural W standard and a 182 W enriched spike shows that differences in 182 W/ 184 W of ∼10 ppm can be well resolved using this method. The external reproducibility of ±4.5 ppm is ∼5 times more precise than conventional W iso- tope measurements by MC-ICP-MS. The new technique constitutes an ideal tool for investigating the W isotope composition of terrestrial rocks for potential contributions from the core, and late accreted extraterrestrial materials. Published by Elsevier B.V. 1. Introduction Over the past two decades, the short-lived 182 Hf– 182 W chronometer (T 1/2 = 8.9 Myr; [1]) has been widely used for dat- ing early Solar System processes, due to the unique geochemical properties of the system. Tungsten is a siderophile (iron-loving) element, and as such, it is largely (but not completely) extracted from the silicate mantles of planetary bodies during segregation of metallic cores [2–8]. Hafnium, in contrast, is lithophile (silicate- loving) and is wholly retained in the silicate portion of planetary bodies. Therefore, determination of the abundance of the daughter nuclide, 182 W, relative to other stable, non-radiogenic W isotopes (e.g., 184 W) is of special interest for constraining the timing of plan- etary core formation. All terrestrial rocks investigated to date [9–13] have been characterized by 182 W/ 184 W ratios that are ∼200 ppm more radio- genic than primitive, undifferentiated meteorites, e.g., chondrites ∗ Corresponding author. E-mail addresses: mtouboul@umd.edu (M. Touboul), rjwalker@umd.edu (R.J. Walker). [14–17]. The difference between the terrestrial rocks and chondritic meteorite compositions has been interpreted as evidence of the early formation of the Earth’s core, less than 30 Myr after the ini- tial formation of the Solar System [14,16,18,19]. The difference also implies that the Earth’s core is a W-rich reservoir with a 182 W/ 184 W that is likely ∼350 ppm lower than terrestrial silicates. Because Hf and W are also fractionated by magma ocean crystallization, par- tial melting of the silicate mantle, and subsequent crystal-liquid fractionation processes [20], the absence of significant W isotopic variations among terrestrial rocks suggests either that the differ- entiation of the Earth’s mantle occurred more than 60 Myr after Solar System formation, after 182 Hf became extinct, or that traces of primitive crustal or mantle reservoirs that formed earlier have been erased by convective mixing. The late formation of the Moon (>50 Myr after Solar System formation), ostensibly by a giant impact [21–23] provides supporting evidence for late-stage homogeniza- tion of W isotopes in the Earth’s mantle. Despite the likely homogenization of W isotopes in the mantle following the putative giant impact that generated the Moon, small W isotope heterogeneities might have been generated after 182 Hf was no longer extant as a result of the late accretion of materials with either more 182 W-depleted or enriched isotopic compositions 1387-3806/$ – see front matter Published by Elsevier B.V. doi:10.1016/j.ijms.2011.08.033