The Pacific Centre for Isotopic and Geochemical Research (PCIGR) at the University of British Columbia has undertaken a systematic analysis of the isotopic (Sr, Nd, and Pb) compositions and concentrations of a broad compositional range of U.S. Geological Survey (USGS) reference materials, including basalt (BCR‐1, 2; BHVO‐1, 2), andesite (AGV‐1, 2), rhyolite (RGM‐1, 2), syenite (STM‐1, 2), granodiorite (GSP‐2), and granite (G‐2, 3). USGS rock reference materials are geochemically well characterized, but there is neither a systematic methodology nor a database for radiogenic isotopic compositions, even for the widely used BCR‐1. This investigation represents the first comprehensive, systematic analysis of the isotopic composition and concentration of USGS reference materials and provides an important database for the isotopic community. In addition, the range of equipment at the PCIGR, including a Nu Instruments Plasma MC‐ICP‐MS, a Thermo Finnigan Triton TIMS, and a Thermo Finnigan Element2 HR‐ICP‐MS, permits an assessment and comparison of the precision and accuracy of isotopic analyses determined by both the TIMS and MC‐ICP‐MS methods (e.g., Nd isotopic compositions). For each of the reference materials, 5 to 10 complete replicate analyses provide coherent isotopic results, all with external precision below 30 ppm (2 SD) for Sr and Nd isotopic compositions (27 and 24 ppm for TIMS and MC‐ICP‐MS, respectively). Our results also show that the first‐ and second‐generation USGS reference materials have homogeneous Sr and Nd isotopic compositions. Nd isotopic compositions by MC‐ICP‐MS and TIMS agree to within 15 ppm for all reference materials. Interlaboratory MC‐ICP‐MS comparisons show excellent agreement for Pb isotopic compositions; however, the reproducibility is not as good as for Sr and Nd. A careful, sequential leaching experiment of three first‐ and second‐generation reference materials (BCR, BHVO, AGV) indicates that the heterogeneity in Pb isotopic compositions, and concentrations, could be directly related to contamination by the steel (mortar/pestle) used to process the materials. Contamination also accounts for the high concentrations of certain other trace elements (e.g., Li, Mo, Cd, Sn, Sb, W) in various USGS reference materials.

. Reproducibility is better than 50 ppm for the granitoid compositions and better than 40 ppm for the basaltic/andesitic compositions. For the isotopic analyses acquired early in this project on glass columns, Hf isotopic analyses from several of the reference materials were significantly less reproducible than Nd and Sr isotopic analyses determined from the same sample dissolution. The 176 Hf/ 177 Hf ratios for relatively radiogenic compositions (BCR-1, 2; BHVO-1, 2; RGM-1) were shifted systematically toward lower values by 100-150 ppm when a borosilicate primary column was used. Although systematic, the shift for felsic compositions was generally within analytical error, except for GSP-2, which has a very low Hf isotopic ratio, where the shift was to higher Hf and high Hf concentrations of the borosilicate glass column (16 ppm) and frit material (22 ppm) indicate that only small amounts of such unradiogenic material could cause significant contamination of small samples. For the basaltic (BCR-1, 2; BHVO-1, 2) and rhyolitic (RGM-1) samples, approximately 3 ng of Hf from the column or frit would be enough to produce the observed 100-150 ppm shift. Accurate, high-precision 176 Hf/ 177 Hf data can only be acquired if samples are processed using all PTFE Teflon 1 labware, or quartz and polypropylene.

[1] The recent development of multiple-collector inductively coupled plasma mass spectrometry (MC-ICP-MS) and increasing use of the technique have created the need for well-characterized rock standards, especially for isotopic systems where no internal fractionation correction can be applied. This paper presents a careful leaching experiment on the U.S. Geological Survey (USGS) reference materials BHVO-1 and BHVO-2 (Hawaiian basalts) and documents the evidence for contamination of the rock powders during processing. This contamination accounts for the difference in Pb isotopic ratios of BHVO-1 and BHVO-2 as well as for their lack of homogeneity both in Pb isotopic compositions and in some trace element contents.

The United States Geological Survey granitic and granodioritic reference materials G‐2 and GSP‐2 were decomposed in high‐pressure bombs using both HF‐HNO3 and HF‐HNO3‐HClO4 in order to evaluate the feasibility of characterising the entire suite of geologically relevant trace elements through direct analysis with a high‐resolution inductively coupled plasma‐mass spectrometer (HR‐ICP‐MS). The digested samples were diluted to the appropriate levels and analysed at low, medium and high resolution depending on the required sensitivity and potential interferences for each element. Memory effects during analysis of the high field strength elements (HFSE) were negligible when analysed using an all‐Teflon, uncooled sample introduction system and combined with adequate wash times with 4% v/v aqua regia + 0.5% v/v HF between samples. The concentration of the remaining lithophile elements was determined with a conventional, cooled, Scott‐type spray chamber using a wash solution of 1% v/v HNO3. Total procedural blanks contributed between 0.01 to 0.5% to final sample concentrations and blank subtractions were typically unnecessary. Abundances for Li, Hf, Ba, Zr, Ga, Rb, Sr, La, Ce, Th and U were systematically higher, while those for the heavy rare earth elements (HREEs), Cu and Y were systematically lower in this study compared to USGS values for G‐2 and GSP‐2. This is likely to be related to, respectively, higher recoveries from more efficient digestion of refractory phases (i.e., zircon, tourmaline), and better resolution of interferences when using a HR‐ICP‐MS. Sample digestion experiments also showed that perchloric acid digestion in high pressure bombs resulted in superior recoveries and better precision for the bulk of the trace elements analysed. The concentration of the remaining elements overlapped within uncertainty with recommended reference values and with values determined in other studies using isotope‐dilution TIMS, ICP‐MS and XRF. Concentrations for the elements Cd, Sn, Sb, Ta, Bi, Tb, Ni and Mo are also reported for G‐2 and GSP‐2 reference materials. Our study shows therefore that it is feasible to determine thirty‐nine geologically relevant trace elements accurately and directly in granitoid sample digests when using a HR‐ICP‐MS, thereby negating the need for ion exchange or isotopic spiking.

No abstract