The Uncertainty Budget of the Multi-Element Analysis of Glasses Using La-Icp-Ms

Yan Luo,Shan Gao,Henry P. Longerich,Detlef Guenther,Samuel Wunderli,Hong-Lin Yuan,Xiao-Ming Liu

DOI: https://doi.org/10.1039/b608010c

2007-01-01

Journal of Analytical Atomic Spectrometry

Abstract:A first attempt was made to estimate an uncertainty budget for the multi-element analysis of glasses using LA-ICP-MS, in accordance with the "Bottom-up'' approach of the EURACHEM/CITAC-Guide.(1) Analyses of NIST SRM 612, 614 and USGS glasses BCR-2G and BIR-1G were carried out using a 193 nm excimer LA-ICP-MS under routine conditions. Calibration was performed using NIST 610 with internal standardisation using Ca. The uncertainty budgets for the analytes Co, La and Th were studied. Instrumental drift and uncertainties from working values of NIST 610, as reported by Pearce et al.,(2) are the dominant sources of uncertainty for a typical individual analysis of NIST 612 and BCR-2G/BIR-1G with mass contents of Co, La and Th ranging from 6 to 52 mu g g(-1). In contrast, the uncertainty contributions from Poisson counting statistics prevail for those of NIST 614 and BIR-1G with the three elements having a lower range between 0.029 and 0.75 mu g g(-1). La was an exception. Its combined uncertainties were consistently dominated by its uncertainty from the working value of NIST 610(2) at all mass content ranges investigated, suggesting that more accurate reference values for the analyte in NIST 610,(2) and for all analytes with large uncertainties, are needed. Additionally, a z-score assessment was carried out using procedures similar to those used in the International Proficiency Test for Analytical Microprobe Geochemistry Laboratories. The z-scores in this study were in the range -2 < z < 2, indicating that there were no significant unsuspected influences in the analytical system. This suggests that the uncertainty budget reported here contains all the significant parameters.