Tracking Gold Losses Using TIMA-X Automated Mineralogy
Compared to other ores, gold ores are considered of low concentration with commercially viable deposits typically containing 2 to 10ppm, with only very few containing hundreds or thousands of ppm. Dues to this low concentration, the beneficiation of the gold is often complex and the precious gold can easily be lost at any stage during the process.
Precise control over the entire process is thus the key to minimizing losses and maximizing profits of a gold mining operation. The only viable tool to suitably maintain such a process is an SEM-based automated mineralogy system like the TIMA-X. The TIMA-X can determine the gold content in tailings as well as the relationship to other minerals. This understanding allows the mine operator to understand why gold losses are occurring.
The following case study describes this scenario.
The TESCAN TIMA-X is an automated mineralogy solution that can automatically identify and quantify minerals present in a sample. It can also provide additional valuable textural information such as grain size, mineral association and liberation degree. The TIMA-X determines mineral classifications by using EDS (Energy Dispersive Spectroscopy) and BSE (Backscattered Electrons) signals.
In this case study data was acquired using:
- Accelerating voltage = 25kV,
- Beam current = 5nA
- Working distance = 20mm
- Applied pixel spacing = 2µm
- High vacuum mode
The sample was obtained from a European gold producer and came in the form of loose powder and was representative of a week average tailings from the plant. The original 0.5kg sample was reduced to 10g using a laboratory splitter and made into 3 epoxy mounts. The mounts were produced under vacuum to remove any gases from the epoxy. The resultant mounts were ground and polished and coated with 10nm of carbon to aid SEM imaging and EDS mapping.
Bright Phase Search
The TIMA-X has three modes for data acquisition and processing:
- Modal analysis
- Liberation analysis
- Bright Phase Search
The former two are suited to most geological samples providing information about mineral abundance and related textural data.
BPS is suited to scenarios where the mineral of interest is only present in a minor proportion and is embedded in minerals of significantly different molecular weight to provide BSE contrast. This makes it a suitable solution for characterizing such things are platinum group metals (PGM), gold and REE ores.
BPS search does more than identify the mineral of interest. It scans the entire particle that the mineral of interest is contained in providing data on mineral locking as well. This can all be automatically distinguished from the epoxy matrix that the particles are embedded using BSE thresholding.
In this particular case, the BSE threshold can be set higher (compared to Modal and Liberation analyses) thus also avoiding minerals with lower BSE signals, reducing acquisition times from hours to minutes.
Gold tailings typically contain sulphides such as galena, sphalerite, pyrite, chalcopyrite etc. The BSE threshold was originally set to 90 to distinguish gold from galena. The sample was found to be rich in silver (up to 20%) which resulted in the BSE signal dropping below the limit. Hence the BSE threshold was dropped to 80 and a phase filter which uses X-ray spectra in conjunction with BSE was applied (see figure 2a and 2b). This enabled the system to filter gold-containing particles from other minerals such as galena.
Automated mineralogy data support fire assay results which indicated that gold was only present in very small amounts (<1ppm). TIMA-X found only 5 gold-containing particles in a study of 1.5 million gangue particles.
Furthermore, gold particles that passed through the processing plant to tailings were found to be very small (1 to 5µm) and they are totally enclosed by quartz. This is not surprising as the processing plant is designed to collect sulphides which typically enclose gold in the primary ore (see figure 3). The gold forms less than 5% of gold containing particles and it belongs to the first liberation class forming less than 10% of the particles surface.
Gold content is typically carried out using fire assays of different size fractions. Gold assays are inadequate when looking to try and find reasons why gold is making it through to tailings.
Even SEM would struggle to differentiate gold from galena given their similar BSE signal. The TIMA automated mineralogy solution with BPS and phase filtering can provide valuable insights into why gold is not being separated out within a matter of minutes.
This same method can be used beyond gold tailings in areas such as other low grade ores. While the TIMA-X is automates, user intervention can be minimized by adding the sample autoloader which caters for 100 samples and allows the system to work 24/7 unattended.