Reproducibility of electron-microprobe bulk analyses of fine-grained media: A case study using modern bone china

The reproducibility of data on the bulk composition of very fine-grained me dia, and specifically the concentration of the major elements as establishe d by electron-microprobe analysis, have been statistically assessed in orde r to establish criteria for analysis. Two sets of spot analyses, one using a beam diameter of 20 mu m, the other of 40 mu m, were undertaken on each o f three fragments cut from a sample of modern bone china. Student's t-tests were run to determine if there was a significant difference (at 95% precis ion) between the two datasets. The results show that as few as 35 spot anal yses need be averaged to provide a reliable indication of bulk composition, provided that the beam diameter is at least as large as the coarsest grain s, and much larger than the average grain-size of the medium being analyzed .