No abstract

From measurements of Hf-Yb mixtures, we have found that the correction of isobaric interferences involving accepted Yb isotope ratios and reasonable estimates of mass bias result in a significantly under-corrected 176 Hf, which is proportional to the amount of Yb added. This can be explained by (1) a significant difference in the instrumental mass bias between Hf and Yb, and (2) that the accepted values for isotopic ratios within the Yb and/or Hf systems are incorrect. We have evaluated these possibilities by measuring mixed solutions of Yb and Hf on two MC-ICP-MS instruments and undertaking a series of REE fractionation experiments using a thermal ionisation mass spectrometer (TIMS). Our results indicate that the presently accepted abundances of the Yb isotopes are not appropriate. We present new values for Yb isotopic abundances based on the TIMS and MC-ICP-MS results. Using the newly defined Yb values, we demonstrate that Yb and Hf have similar levels of mass bias in plasma ionisation instruments, and that Hf isotope ratios can be used to correct Yb mass bias before subsequent correction of isobaric interference. A laser ablation comparison of Yb and Hf indicates that similar relationships exist, and can be applied to micro-analytical techniques where chemical separation is not possible.

No abstract

No abstract

No abstract

11The fate of crustal material returned to the convecting mantle by plate 12 tectonics is important for understanding the chemical and physical evolu-13 tion of the planet. Marked isotopic variability of Mo at the Earth's surface 14 o↵ers the promise of providing distinctive signatures of of such recycled ma-15 terial. However, characterisation of the behaviour of Mo during subduction 16 is needed to assess the potential of Mo isotope ratios as tracers for global 17 geochemical cycles. Here we present Mo isotope data for input and output 18 components of the archetypical Mariana arc: Mariana arc lavas, sediments 19 from ODP Sites 800, 801 and 802 near the Mariana trench and the altered 20 mafic, oceanic crust (AOC), from ODP Site 801, together with samples of 21 the deeper oceanic crust from ODP Site 1256. We also report new high pre-22 cision Pb isotope data for the Mariana arc lavas and a dataset of Pb isotope 23 ratios from sediments from ODP Sites 800, 801 and 802. The Mariana arc 24 lavas are enriched in Mo compared to elements of similar incompatibility 25 Email address: glxhf@bristol.ac.uk (Heye Freymuth)during upper mantle melting, and have distinct, isotopically heavy Mo (high 26 98 Mo/ 95 Mo) relative to the upper mantle, by up to 0.3 parts per thousand. 27 In contrast, the various subducting sediment lithologies dominantly host iso-28 topically light Mo. Coupled Pb and Mo enrichment in the Mariana arc lavas 29 suggests a common source for these elements and we further use Pb isotopes 30 to identify the origin of the isotopically heavy Mo. We infer that an aqueous 31 fluid component with elevated [Mo], [Pb], high 98 Mo/ 95 Mo and unradiogenic 32 Pb is derived from the subducting, mafic oceanic crust. Although the top few 33 hundred metres of the subducting, mafic crust have a high 98 Mo/ 95 Mo, as a 34 result of seawater alteration, tightly defined Pb isotope arrays of the Mariana 35 arc lavas extrapolate to a fluid component akin to fresh Pacific mid-ocean 36 ridge basalts. This argues against a flux dominantly derived from the highly 37 altered, uppermost mafic crust or indeed from an Indian-like mantle wedge. 38 Thus we infer that the Pb and Mo budgets of the fluid component are dom-39 inated by contributions from the deeper, less altered (cooler) portion of the 40 subducting Pacific crust. The high 98 Mo/ 95 Mo of this flux is likely caused 41 by isotopic fractionation during dehydration and fluid flow in the slab. As a 42 result, the residual mafic crust becomes isotopically lighter than the upper 43 mantle from which it was derived. Our results suggest that the continen-44 tal crust produced by arc magmatism should have an isotopically heavy Mo 45 composition compared to the mantle, whilst a contribution of deep recycled 46 oceanic crust to the sources of some ocean island basalts might be evident 47 from an isotopically light Mo signature.48 crust 50 2 portant tool to reconstruct paleo-redox conditions in the ocean (e.g. Siebert 53 et al., 2003; Arnold et al., 2004). Fractionation of Mo isotopes ...

No abstract

[1] The Kyushu-Palau Ridge (KPR) is a 2600 km long remnant island arc that is separated from the active Izu-Bonin-Mariana (IBM) arc by a series of spreading and rift basins. We present 40 Ar/ 39 Ar ages and geochemical data for the entire length of the Kyushu-Palau arc as well as for the conjugate arc which is stranded within the IBM fore arc. New 40 Ar/ 39 Ar ages indicate that the KPR was active between 25 and 48 Ma, but the majority of the exposed volcanism occurred in the final phase, between 25 and 28 Ma. Rifting of the Kyushu-Palau arc to form the Shikoku and Parece Vela basins occurred simultaneously along the length of the arc (circa 25 Ma), and at a similar distance from the trench. Unlike the IBM, the KPR has only limited systematic along-arc geochemical trends. Two geochemical components within the KPR indicate an origin in the suprasubduction mantle. First, EM-1-like lavas are identified in a restricted section of the arc, suggesting a localized heterogeneity. Second, EM-2-like arc volcanoes formed on juvenile West Philippine Basin crust, potentially reflecting ingress of mantle from the then active EM-2 province which lies in the west. Another geochemical heterogeneity is found at the KPR-Daito Ridge intersection where the arc developed on preexisting Cretaceous Daito Ridge crust. The geochemical characteristics at this intersection likely result from the involvement of sub-Daito Ridge lithospheric mantle. Subduction flux beneath the KPR generally matches post-45 Ma Eocene/Oligocene lavas in the IBM fore arc, involving fluids and melts derived from altered igneous crust.