Application of Taguchi Methods for Comparing Three Analytical Methods of Trace and Ultra-trace Gold in Geological Samples

There are currently three principle methods to measure trace and ultra-trace gold in geological samples， including activated carbon enrichment-Emission Spectrometry, foam enrichment-Graphite Furnace Atomic Absorption Spectrometry (GFAAS), and activated carbon enrichment-Inductively Coupled Plasma-Mass Spectrometry (ICP-MS). Of these, the Emission Spectrometry uses domestic equipment, with low cost and a lengthy process; the GFAAS is easy to operate but is costly as it employs an imported graphite tube. The ICP-MS method has a low detection limit but demands higher environmental conditions with high cost. The detection limits of these three methods are from 0.1 to 0.3 ng/g. As the composition of the ore mineral is very complicated, satisfactory results for the analysis of geological samples are provided by adopting a suitable analysis method based on the effective enrichment and separation. To ensure sensitivity and accuracy of sample requirements, selection of the analysis method should follow the specific sample and laboratory conditions. According to the signal-to-noise ratio (S/N) concept proposed by Doctor Genichi Taguchi in measurement engineering, the S/N is equal to the magnitude of the mean of the process compared to its variation. The bigger the S/N, the greater the anti-interference ability, the higher sensitivity, the better the linearity, the safer, and the more reliable is the measuring system. In this paper, a description of how the Taguchi measurement quality control method was conducted to evaluate three common analysis methods for trace and ultra-trace gold measurement in geological samples for the first time is given. The measurement data obtained by the three common methods for three national standard materials were processed to calculate the S/N values and the relative standard uncertainties, respectively. By comparing the three S/N values and the relative standard uncertainties, a more stable method was selected. The experimental results show that all three analytical methods met the requirements of the trace gold measurement in geological samples, whose Au content is greater than 1 ng/g, however, the activated carbon enrichment-ICP-MS is more suited for the ultra-trace gold in samples whose Au content is less than 1 ng/g.