A reconnaissance view of tungsten reservoirs in some crustal and mantle rocks: Implications for interpreting W isotopic compositions and crust-mantle W cycling

High-precision measurements of W isotopic ratios have enabled increased exploration of early Earth processes. However, when applying W isotopic data to understand the geological processes, it is critical to recognize the potential mobility of W and hence evaluate whether measured W contents and isotopic compositions reflect the primary petrogenetic processes or instead are influenced by the effects of secondary inputs/mobility. Furthermore, if we are to better understand how W is partitioned between different minerals during melting and metasomatic processes it is important to document the likely sinks for W during these processes. In addition, an understanding of the main hosts for W in the crust and mantle is critically important to constrain how W is cycled and stored in the crust-mantle geochemical cycle. As a first step to investigate these issues, we have carried out in situ concentration measurements of W and other HFSEs in mineral phases within a broad spectrum of crustal and mantle rocks, along with whole-rock concentration measurements. Mass balance shows that for tonalitic gneiss and amphibolite, the major rock-forming minerals can adequately account for the bulk W budget, and for the pristine ultramafic rocks, olivine and orthopyroxene are the major controlling phases for W whereas for metasomatized ultramafic rocks, significant W is hosted in Ti-bearing trace phases (e.g., rutile, lindsleyite) along grain boundaries or is inferred to reside in cryptic W-bearing trace phases. Formation or decomposition of these phases during secondary processes could cause fractionation of W from other HFSEs, and also dramatically modify bulk W concentrations in rocks. For rocks that experienced subsequent W enrichment/alteration, their W isotopic compositions may not necessarily represent their mantle sources, but could reflect later inputs. The relatively small suite of rocks analyzed here serves as a reconnaissance study but allows some preliminary speculations on their significance for crust-mantle HFSE and siderophile element budgets - to be tested in future studies. The significant concentration of W, as well as Nb and Ta hosted in rutile and titanite has interesting implications for the budget of W during crust-mantle recycling. Crust-mantle recycling models invoking the recycling of rutile-bearing eclogites to satisfy the mantle Nb/Ta ratio carry the penalty that the very high W/U and W/Th ratios of these rocks results in a concomitant large deviation from the primitive mantle-like ratios estimated for bulk continental crust. Similarly, data from the single amphibolite sample investigated in this study are inconsistent with models implicating the partial melting of amphibolite-bearing subducted slabs as a major process for formation of continental crust in the Earth's early history. Either the current widely accepted estimates for bulk continental crust W/U and W/Th ratios are in error, or partial melting or other processes lowers the W/U or W/Th of melt residues during their return to the mantle. The present small dataset cannot properly evaluate this, requiring further investigation. Finally, the lithospheric mantle has the potential to store substantial amounts of W, for example via infiltration by W-rich melts/fluids, and thus may act as a source for W mineralization in the crust. (C) 2017 Elsevier Ltd. All rights reserved.