Hf Isotope Geochemistry of USGS Reference Materials and Various Labware: Insight into Potential Contaminant Sources

We have undertaken a high-precision geochemical and isotopic study of USGS reference materials by HR-ICP-MS, TIMS and MC-ICP-MS, including basalt (BCR-1,2; BHVO-1,2), andesite (AGV-1,2), rhyolite (RGM-1), syenite (STM-1,2), granodiorite (GSP-2), and granite (G-2,3). Only a few 176Hf/177Hf results are published on these materials and with the increased use of MC-ICP-MS it is critical to build a solid reference database. Standard hotplate dissolution was used, except for granitoid compositions where it involved a high-pressure bomb procedure. The reproducibility of 176Hf/177Hf is better than 100 ppm for granitoid compositions (G-2: 0.282523±8; G-3: 0.282505±20; GSP-2: 0.282059±27) and better than 65 ppm for basaltic/andesitic compositions in glassware and better than 30 ppm in teflon (BCR-2: 0.282872±9; BHVO-2: 0.283103±6). Overall, our results agree with the rare published data (BCR-1&2, BHVO-1 and RGM-1). Slight differences appear depending on the chemical procedure used to separate Hf and the type of labware used. There are systematic shifts in 176Hf/177Hf for basaltic compositions towards lower values (by 100-150 ppm) when non-teflon material is used. As a result, we then carried out a systematic trace element and isotopic study of various labware, including borosilicate glass and quartz columns and frits. Maximum concentrations (in ppm) of these materials (in the order listed above) are: Hf=16-0.3-22, Nd=0.8-0.1-23, Sr=8-0.08-16, Pb=1.4-0.5-14. The frit material appears the most variable in elemental concentration and isotopic composition, which might reflect various accumulations resulting from column chemistry. 176Hf/177Hf is 0.282198±4 in borosilicate glass and even lower in some of the frit material (<0.28195). Only a small amount of such unradiogenic material can account for the shifts observed in basaltic rocks. Our systematic study shows that careful analyses of rock reference materials with different compositional matrices are necessary, in addition to standard solution measurements, to achieve accuracy and reproducibility in MC-ICP-MS analyses.