Precision and accuracy of glass analyses by electron microprobe (EMP) at Uppsala University

Abstract: In situ analysis of glass requires matrix-matched standards for timely assessment of the accuracy and precision of the results. We provide data to constrain the precision and accuracy of glass analyses by FEG-EMP at Uppsala University. Seven of the DING glasses that represent a broad spectrum of compositions were used to test for accuracy and precision (Jochum et al., 2006). Multiple analyses were performed on each glass; ML3B (n=30), STHs 6/80 (n=43), ATHO (n=54), KL2 (n=30), GOR-132 (n=30), GOR-128 (n=38) and T1 (n=36).
Glass analysis was performed using a field emission gun JEOL JXA-8530F Hyperprobe electron probe microanalyser at Uppsala University. The analytical conditions were 15 kV accelerating voltage and 4 nA probe current with a defocused beam of 5 to 10 micron diameter. Measurements were conducted with counting times of 10 s on peak and 5 s on lower and upper background and the PRZ correction method was applied to the data. Detection limits are typically ≤150 ppm.
Despite different volatile contents, the DING glasses gave totals of 97.2±2.4 (2s.d.; n=241) and have been normalized to further assess the data quality. The precision of the analyses is reported as uncertainties at the 95% confidence level for major element oxides. When concentrations are >30 wt% the uncertainties are 0.7 to 1.6% (2s.d.), for concentrations of 5 to 30 wt% uncertainties are <7% (2s.d.) and at contents of 1 to 5 wt% uncertainties are <10% (2s.d.). Comparison to the preferred values reported by Jochum et al. (2006) reveals a correlation of 1.0079.

Other Description: Jochum et al., (2006) MPI-DING reference glasses for in situ microanalysis: New reference values for element concentrations and isotopes ratios. Geochemistry, Geophysics, Geosystems Q02008, doi: 10.1029/2005GC001060