Analysis of 232Th/ 230Th in volcanic rocks: a comparison of thermal ionization mass spectrometry and other methodologies

Abstract

Techniques for measuring Th isotopic compositions of igneous rocks are evaluated using thermal ionization mass spectrometry (TIMS) data collected at the University of Hawaii (UH). A new single collector method for unspiked 232Th/ 230Th measurement is presented and compared to UH results using other published methods (multicollector and “split-analysis” single collector, a.k.a. “bridge”). Analyses of a synthetic thorium standard (UCSC-Th-A, n=245 over 4 years), ionization from single and triple filaments, instrumental parameters affecting analytical quality, and analyses using two secondary energy filters (a 30-cm electrostatic analyzer, ESA, and a Micromass WARP©) are discussed. The TIMS technique with the best combination of accuracy (42 analysis mean=0.12%>gravimetric), precision, reproducibility and analyte size was single collector ion-counting analysis of 50-ng single filament loads. Static multicollector analysis (500 ng triple filament loads) resulted in somewhat lower overall precision at similar accuracy (with small but measurable differences between the ESA and WARP data). Bridge analyses (50-ng single filament loads) were the least precise and accurate. Low-Th-abundance volcanic rocks (10 duplicates, 2 triplicates) were also analyzed for 232Th/ 230Th by the single collector technique. Chemical blanks were demonstrably negligible and all replicate sets agreed to within the precision of the standards data. Generalizations are made about the state of the art (analysis techniques and the standards themselves) using a compilation of synthetic and natural standards data from 25 publications (15 laboratories worldwide) collected by TIMS, high resolution α-spectrometry, and other mass spectrometry types. Precision and accuracy are variable between data sets (by more than 10×), regardless of instrumentation. Not all publications give uniform analytical meta data, making it difficult to assess the causes of this variability. Perhaps contrary to popular conception, traditional α-spectrometry data can be highly accurate and only moderately imprecise relative to mass spectrometry data. Inter-laboratory means for commonly used standards (UCSC-Th-A, Table Mountain Latite, TML, and AThO) are calculated. The mean 232Th/ 230Th of the 146 UH analyses of UCSC-Th-A is 170,790 (2 σ=1.4%, 2 σ m=0.11%, 0.11%>gravimetric). Combining with available literature data (unfiltered) yields 232Th/ 230Th=170,760 (255 analyses, 2 σ=1.2%, 2 σ m=0.08, 0.09%>gravimetric).