Article pubs.acs.org/ac
Evaluation of New Geological Reference Materials for Uranium-Series Measurements: Chinese Geological Standard Glasses (CGSG) and Macusanite Obsidian J. S. Denton,*,† M. T. Murrell, S. J. Goldstein, A. J. Nunn, R. S. Amato, and K. A. Hinrichs Nuclear and Radiochemistry (C-NR), Los Alamos National Laboratory, MS J514, PO Box 1663, Los Alamos, New Mexico 87545, United States ABSTRACT: Recent advances in high-resolution, rapid, in situ microanalytical techniques present numerous opportunities for the analytical community, provided accurately characterized reference materials are available. Here, we present multicollector thermal ionization mass spectrometry (MC-TIMS) and multicollector inductively coupled plasma mass spectrometry (MC-ICP-MS) uranium and thorium concentration and isotopic data obtained by isotope dilution for a suite of newly available Chinese Geological Standard Glasses (CGSG) designed for microanalysis. These glasses exhibit a range of compositions including basalt, syenite, andesite, and a soil. Uranium concentrations for these glasses range from ∼2 to 14 μg g−1, Th/U weight ratios range from ∼4 to 6, 234U/238U activity ratios range from 0.93 to 1.02, and 230Th/238U activity ratios range from 0.98 to 1.12. Uranium and thorium concentration and isotopic data are also presented for a rhyolitic obsidian from Macusani, SE Peru (macusanite). This glass can also be used as a rhyolitic reference material, has a very low Th/U weight ratio (around 0.077), and is approximately in 238U−234U−230Th secular equilibrium. The U−Th concentration data agree with but are significantly more precise than those previously measured. U−Th concentration and isotopic data agree within estimated errors for the two measurement techniques, providing validation of the two methods. The large 238U−234U−230Th disequilibria for some of the glasses, along with the wide range in their chemical compositions and Th/U ratios should provide useful reference points for the U-series analytical community.
U
These glasses are called the Chinese Geological Standard Glasses (CGSG) and exhibit a range of chemical compositions including basalt, syenite, andesite, and a soil. These glasses were fused from existing powders at 1500 to 1600 °C as described in Hu et al.27 Na2CO3 and Li2B4O7 were used in the fusion of one of the glasses (CGSG-5). The homogeneous distribution of major and trace elements in the CGSG materials was established by conducting electron probe micro analyzer (EPMA) profiles across glass splits (10 μm spot, 30 μm distance between spots) and a large number of 120 μm spot LA-ICP-MS analyses at different locations on glass fragments respectively.27 The major element distribution was found to be homogeneous within EPMA analytical uncertainty. RSD values for the lithophile trace elements were reported to be well within the repeatability field of LA-ICP-MS.27,28 Good large-scale homogeneity has also been demonstrated with RSD values of often better than 5% for lithophile elements from LA-ICP-MS analyses of three different splits of each CGSG.27 Some heterogeneity was reportedly found in the siderophile/ chalcophile trace elements, possibly due to the loss of volatile
ranium-series analytical measurements are widely used in geochemistry1−4 and geochronology5,6 and can be applied to a wide variety of fields including paleoclimatology,7,8 volcanology,9,10 environmental risk assessment,11,12 and archeology.13,14Recent advances in high-resolution, rapid, in situ, microanalytical techniques, for example, laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS), secondary ion mass spectrometry (SIMS), and micro X-ray fluorescence (μ-XRF) present numerous opportunities for the analytical community.15−20 Reference glasses are essential in microanalysis and are commonly used as samples for calibration, method development, quality control/assurance, and for inter-laboratory comparison.21 About 20−25 reference glasses are currently available for microanalysis.22 Some are more suitable for U-series measurements than others. For example, while most natural samples have natural 235U/238U ratios, the synthetic glasses from NIST and the USGS are depleted in 235U.23−25 The natural glass reference materials produced by the USGS and MPI-DING yield a number of glasses that contain an analytically appropriate concentration of U and Th. Some of these have been previously characterized by us for uranium and thorium isotopics.26 Recently, four new reference glasses were produced by the National Research Center for Geoanalysis (NRCG), Beijing.27 © 2013 American Chemical Society
Received: June 7, 2013 Accepted: September 4, 2013 Published: September 4, 2013 9975
dx.doi.org/10.1021/ac4017117 | Anal. Chem. 2013, 85, 9975−9981
Analytical Chemistry
Article
components and/or loss to the Pt crucible.27 Pt was also found to be heterogeneous in all the glasses due to the use of Pt crucibles in the fusing process. In addition to the new matrices made available by these Chinese glasses, a possible substitute reference material for the rhyolitic glass ATHO-G is the macusanite obsidian from SE Peru.29−31 Heterogeneity in some elements has been found in the ATHO-G glass.29 The macusanite obsidian is an important archeological marker in SE Peru.32 In this study, we present multi collector thermal ionization mass spectrometer (MC-TIMS) and multi collector inductively coupled plasma mass spectrometer (MC-ICP-MS) results for uranium and thorium isotopic ratios and elemental concentrations (by isotope dilution) for each CGSG and the macusanite obsidian. These high-precision and high-accuracy ratios from a suite of reference materials that comprise both basaltic and non-basaltic compositions complement data from existing glasses and expand the catalogue of reference materials that are appropriate for in situ U-series work. These results can be used to evaluate instrumental mass and elemental fractionation in a broader range of material compositions, assess the performance of different microanalytical techniques, and facilitate inter-laboratory comparison of data within the broader analytical community.
(99.998% ThO2) (Oct. 2011). Total propagated uncertainties (2 × the standard error of the mean, SEM) for the spike calibrations were estimated to be 0.22% for uranium and 0.33% for thorium. All reagents were Optima grade (Fisher Scientific, Pittsburgh, PA, USA); all water was sub-boiling Teflon distilled Milli-Q (TDMQ) water (Millipore, Billerica, MA, USA, 18 MΩ cm). All labware was pre-cleaned and acid-leached. All resins were pre-cleaned and stored in chloride form. All sample digestion, chromatography, and analysis took place in a class 100 clean laboratory environment at LANL. Chemical Methods. The reference materials were received as centimeter-sized glass fragments and crushed into a powder using an agate pestle and mortar. To aid digestion, glass chips larger than a few millimeters were removed by hand-picking from the resulting powder. The powders were cleaned by ultrasonication in TDMQ water to remove any surface contamination or dust particles. The powders were weighed and dissolved in HCl, HF, HClO4, and H3BO3 acids to form stock solutions. Aliquots containing approximately 500 ng of uranium and 1000 ng of Th were obtained from each stock solution. Because of the low Th/U ratio in the macusanite glass,31 an aliquot with a much smaller amount of thorium was analyzed to reduce the amount of uranium that was handled. Duplicate aliquots of the stock solution for each material were spiked, purified, and analyzed separately. Aliquots were spiked with 233U and 229Th spikes and allowed to equilibrate overnight with a small amount of HClO4. After fuming the spiked solutions with HClO4 and evaporation to dryness, uranium and thorium were separated by passing the samples through an HCl−HI anion exchange column using AG MP1, 100−200 mesh resin. For additional purification, the uranium fraction was then passed through two anion exchange columns (AG MP1, 100−200 mesh resin) with HCl−H2SO4 and HNO3, respectively. The thorium fraction was then passed through a cation exchange column (Dowex 50WX8, 200−400 mesh resin) with HCl, HBr, and HNO3−HF and then a nitric acid anion exchange column (AG MP1, 100−200 mesh resin) for purification. Each fraction was then split for sequential analysesMC-TIMS followed by MC-ICP-MS analysis. The ICP-MS aliquot was approximately 10% of the TIMS aliquot. Further details of the chemical separations are provided in Matthews et al.26 and references therein. Instrumentation and Data Acquisition. MC-TIMS. Measurements were carried out using a Sector 54 MC-TIMS (VG Instruments) in dynamic multi collector mode with a single Daly and multiple Faraday collectors (FC). The purified uranium fractions were loaded onto tantalum side filaments and run in a triple configuration with a rhenium center filament. Thorium fractions were loaded onto a rhenium side filament and run in a double configuration with a rhenium center filament. Rhenium center filaments were constructed using ceramic holders to reduce alkali signals. Spiked procedural blanks were measured on the Daly collector. The total U and Th contributions measured in the blanks produced negligible corrections (≤0.02%). The macusanite thorium blank correction was larger (∼0.1%) because of the lower concentration of Th in the aliquot, as described in Chemical Methods. The data were corrected for these blanks. The uranium data were collected using a dynamic routine with eight sequences plus baselines. The thorium data were collected in two stages: at lower intensity in a dynamic-Daly-only routine
■
MATERIALS AND METHODS Materials. The CGSG test portions were obtained from X.C. Zhan of the National Center of Geoanalysis in Beijing, China. The glasses were melted and fused (as described in 27) and named CGSG-1, CGSG-2, CGSG-4, and CGSG-5 (Table 1). The sampling locations of the original materials are shown in Figure 1a. The materials consist of a basalt, a syenite, a soil, and an andesite, respectively (Table 1). Table 1. Information Regarding the Geological Reference Materials Evaluated in This Studya reference material
chemical classification
age
location
reference
CGSG-1 CGSG-2
alkali basalt nepheline syenite
unknown ∼220−240 Ma
27 27,48,50,51
CGSG-4 CGSG-5
soil andesite
unknown ∼120 Ma
Macusanite
rhyolitic obsidian glass
∼5 Ma
Tibet Saima alkaline complex, Fengcheng, Liaoning, China Beijing, China Meishan, Nanjing, China Macusani, SE Peru
27 27,47,50,51 29−32
a
Each CGSG was fused from existing natural powders in platinum crucibles.27
The macusanite reference material was obtained from A. Borisova of GET, Observatoire Midi-Pyrenees (Table 1) and is sourced from SE Peru (Figure 1b). The glass characterized here has a translucent-green matrix. For the isotope dilution method we used a 229Th and a 233U spike. The 233U spike was local to Los Alamos National Laboratory (LANL) and was calibrated against the NBS U-960 (purified uranium, 99.975 ± 0.006%) standard (July 2011). The high purity 229Th spike was from Oak Ridge National Laboratory (ORNL). It was calibrated against dissolved 232Th metal (≥99.98% Th) from the Ames Laboratory as well as a solution prepared from Johnson Matthey Puratronic ThO2 9976
dx.doi.org/10.1021/ac4017117 | Anal. Chem. 2013, 85, 9975−9981
Analytical Chemistry
Article
Figure 1. Maps showing (a) the sampling location of the original CGSG materials (1, 2, 4, and 5) in China27,50 and (b) the location of the macusanite obsidian (M) in Peru.32
Thorium (and uranium) concentration accuracy was determined by comparisons with the previously published data. MC-ICP-MS. Uranium and thorium measurements were carried out using a Plasma II MC-ICP-MS (Nu Instruments) with Ni cones. Sample introduction was carried out using an Aridus II desolvating nebulizer (Cetac Technologies) to reduce oxides and improve sensitivity, which was equipped with the QuickWash module to reduce washout times. Prior to each data collection sequence, the instrument was tuned for a robust signal with respect to signal strength and stability. RF power and gas flows did not deviate from standard instrument settings for U and Th measurements throughout the analysis period. Uranium and thorium data were collected in dynamic mode using Faraday cups and one of the ion counters (IC). The IC has a filter for better abundance sensitivity (
