Poisson and non-Poisson uncertainty estimations of 10Be/9Be measurements at LLNL–CAMS

Abstract

We quantify the routine performance and uncertainties of 10Be measurements made on the CAMS FN accelerator mass spectrometer in combination with the CAMS high-intensity cesium sputter source. Our analysis compiles data from 554 primary and secondary standard targets measured on 47 different wheels in nine different run campaigns over a 1-year interval (September 2009–September 2010). The series includes 87, 86, and 85 measurements of each of three different secondary standards and 296 measurements of our primary standard, KNSTD3110 (01-5-4). The average initial 9Be3+ beam current is 22±3μA (1 standard deviation). Secondary standard targets, which are measured as unknowns in each of the wheels, have average statistical uncertainties based on counting statistics of 1.8%, 1.3%, and 0.8% (1σ) (September 2009–March 2010) and 1.3%, 1.0%, and 0.6% (April 2010–September 2010) for standard materials with 10Be/9Be=5.35×10−13, 9.72×10−13, and 8.56×10−12, respectively. The mean measured ratio for each of the secondary standards (normalized to the primary standard) falls within the 1.1% uncertainties of the reported values for each standard material. The weighted standard deviation around the mean of this large number of runs is 2.5%, 2.0%, and 1.2% (September 2009–March 2010) and 1.5%, 1.1%, and 1.2% (April 2010–September 2010) for each secondary standard. These data indicate an additional source of uncertainty, 0.9–1.8% (April 2010–September 2010) and 0.2–1.0% (April 2010–September 2010), above that calculated from counting statistics alone. These 10Be AMS results demonstrate the precision and accuracy of the LLNL–CAMS system.