Our Technology

Mineralogy: The Next Challenge


Conventional wisdom says that separation recovery is a function of feed grade, but it is minerals that are floated, not elements. For simple deposits grade was a “proxy” for recovery but today mines need to understand mineralogy to optimise recovery.

For any separation process, mineralogy dictates the optimal operational conditions which impacts the separation efficiency. However, with current technologies, mineralogy and liberation analysis generally require at least one shift to measure.

...


Because ore has already been processed the system cannot be optimized according to the ore characteristics. This significantly impacts the resulting recovery and grade.

The MinDet Solution


Current commercial offerings are not suitable for real-time measurement. Therefore, we decided to revisit the conventional technique for mineral recognition which was microscopy. Microscopy has been used for centuries to study mineral crystals in thin sections, polished sections and first using visible light, and then electrons in Scanning Electron Microscopy. However, these techniques required sample preparation and only work under dry or vacuum conditions respectively.

To make the process real-time, we designed an optical microscope apparatus which allows direct imaging of minerals in the process slurry as small as 5 µm .

...


By comparing the same slurry images under proven technologies like SEM/EDS and XRD, we were able to develop an AI algorithm to determine the mineral grains and particle sizes.

...

Peer-Reviewed Results


Feasibility studies of the methodology were conducted on a polymetallic sulphide ore from an Australian deposit and published in a peer-reviewed journal1.

...


1Edwin J.Y. Koh, Eiman Amini, Carlos A. Spier, Geoffrey J. McLachlan, Weiguo Xie, Nick Beaton, A mineralogy characterisation technique for copper ore in flotation pulp using deep learning machine vision with optical microscopy, Minerals Engineering, Volume 205, 2024, 108481, ISSN 0892-6875, https://doi.org/10.1016/j.mineng.2023.108481.