Impact of grinding

Does it all have to be such a grind?

A finer particle size is naturally expected to require higher flocculant dosages. Modelling for a median size of 1 µm suggests ~0.2 of a polymer flocculant chain (molecular weight 20×10⁶) would adsorb per particle at a dosage of 25 g t^-1; when the size is halved, the fraction of a chain per particle reduces by over a factor of four (Hogg, 1999). At even smaller particle sizes, dosages required to have some probability of all the particles contacting flocculant become prohibitively high. This simplistic approach doesn’t consider broad particle size distributions (PSDs) or that a single chain may interact with many particles, but highlights how fines dominate flocculation responses.

The “fines” content of a feed, typically defined as the -45 or -38 µm mass fraction from sieve analysis, is often used to indicate flocculation requirements. However, fines in the 1-10 µm range will be much harder to flocculate than those that are 10-30 µm. Sieving to lower cut sizes isn’t practical, so more detailed PSD quantification if needed to probe issues from feed variations.

Collaboration between CSIRO Mineral Resources and the University of Queensland has sought to demonstrate the likely impacts of over-grinding by generating seven different PSDs from one iron ore, with their full PSDs as measured by laser light scattering shown below (Grabsch et al., 2020). PSD 7 corresponds to only 10% of the mass being -38 µm, while PSD 1 is 59%.

PSDs for iron ore slurries, as measured by low angle laser light scattering (Grabsch et al., 2020)	Solids settling flux vs. solids concentration for flocculation at 10 g t-1 (Grabsch et al., 2020)

Most flocculation studies focus on a single solids concentration, but useful comparisons for different PSDs requires testing across multiple concentrations and flocculant dosages, with settling data then converted to solids settling flux. Flux values varied by an order of magnitude at a dosage of 10 g t^-1 across the PSD extremes (above). While tripling the dosage significantly increases fluxes for slurries containing high levels of fines at their optimum solids concentration (~100 kg m^-3), much higher fluxes are still attained with coarser PSDs at both higher concentrations and lower dosages.

Volume- or mass-sensitive measurements of broad PSDs are dominated by coarser particles, despite their contribution to particle numbers being very low. Chord length distributions acquired in situ by focused beam reflectance measurement (FBRM) give real-time indications of size at high solids concentrations, and the raw (unweighted) counts are sensitive to particle number. The solids settling flux obtained from flocculation at a fixed dosage for the seven different PSDs slurries increased as the counts <20 µm within the unflocculated slurries declined, indicating FBRM may have potential for monitoring feed properties for flocculation control.

The results described above are an extreme example to demonstrate potential impacts – the collaboration with the University of Queensland is now considering specific examples of how grinding is achieved industrially. The application of fine grinding, which involves stirred media mills that are installed either vertically (e.g. Vertical Tower Mills) or horizontally (e.g. Isa Mill), has become more common, and it is accepted stirred mills generate a different PSD relative to conventional ball mills. Over-grinding or poorly optimised grinding may have minor effects on value liberation, but may be highly detrimental to the tailings circuit.

References

• Grabsch, A.F., Yahyaei, M., Fawell, P.D., 2020. Number-sensitive particle size measurements for monitoring flocculation responses to different grinding conditions. Minerals Engineering, 145, 106088.
• Hogg, R., 1999. Polymer adsorption and flocculation, in: Laskowski, J.S. (Ed.), Polymers in Mineral Processing. Canadian Inst. Min., Met. & Petr., Quebec City, Canada, pp. 3-17.