| Résumé | We used the laser-induced breakdown spectroscopy (LIBS) technique to determine the concentration of gold in rock samples. 44 reference materials (mostly compressed fine powders) of various chemical compositions, with a quasi-homogeneous concentration of gold varying from 0 to 1000 ppm, were used to establish the calibration curve for the Au I 267.59 nm line. A chemometric study based on the principal component analysis (PCA) showed that ~ 83% of the LIBS spectra variations are attributable to the presence of iron in the samples. Two distinct branches were obtained in the calibration curve: one for Si-rich samples (< 5% of iron) and one for Fe-rich samples (> 13% of iron) with limits of detection of 0.8 ppm and 1.5 ppm, respectively. Different normalization schemes of the gold signal were tested in order to reduce the matrix effects. The LIBS analysis was performed on various mineral samples of practical interest, namely two Si-rich uncompressed ore powders, fine and granular, and three bulk drill cores. The fluctuations in the gold concentration measurements appear to be about two times greater in the granular powder (5–10%) than in the fine one (2–5%). A detailed mapping of the gold concentration on a solid drill core was also performed, revealing multiscale heterogeneity of the gold distribution on the surface of the sample. |
|---|
