Constellation, le dépôt institutionnel de l'Université du Québec à Chicoutimi

Résumé

Counting gold grains in glacial drift is a wellproven exploration technique. However, it is limited by the size of grains that can be recovered and observed, experiencing a recuperation collapse below 50 micrometres. Since ore petrography indicates the overwhelming abundance of very small gold grains, concentrating observation on the smaller fractions will improve accordingly the efficiency of the method. A dependable counting of very small particles is beyond the realm of human skill and requires automation. First, recovery must be conducted by methods which achieve constant concentration factors in the orders of 1:100,000 with constant recovery. Second the identification process requires the use of fully automated scanning techniques based on visible light image analysis and scanning electron microscope (SEM) interoperability. The technique here described enable recovery and counting of gold grains down to a few microns, plus the acquisition of high magnification images to characterize grain shape and EDSSDD analyses. Results not only enable the detection of faint gold grains dispersion trains, but also characterized the grains in regard of their source rocks taking into account their chemical signature and morphology. A new classification of “pristine” grains in regard of their relation to host minerals is proposed.