Zircon and titanium: four separation principles in one plant

You separate these minerals by density, then by magnetism, then by conductivity. Density alone is not enough. Ilmenite is strongly magnetic, leucoxene only weakly magnetic, rutile is non-magnetic, zircon is non-magnetic, and monazite is paramagnetic but its magnetic response varies with thorium content. In order of decreasing magnetic susceptibility, a typical dry plant feed might run: ilmenite, leucoxene, garnet, monazite, and then the non-magnetic pair rutile and zircon. Conductivity separates rutile from zircon because rutile conducts electricity and zircon does not. Rutile is a conductor, ilmenite and leucoxene are weakly conductive or non-conductive depending on alteration, and zircon and monazite are non-conductors. Rutile and zircon are close in density, far too close for a gravity table to split them cleanly.

Altered ilmenite and leucoxene blur the magnetic cut points. A grain of ilmenite that has lost iron to weathering becomes less magnetic and can report to the leucoxene product or even the non-magnetic fraction. That's why a single magnetic susceptibility number from a random grain means little. You need the distribution by size fraction, not an average for the whole sample.

Monazite must be tracked from the very first assay because it concentrates naturally occurring radioactivity. You cannot leave it as an afterthought in a table of minor minerals. It changes handling, storage and transport classification long before the dry plant starts.

Zircon and monazite both carry naturally occurring uranium and thorium. Monazite concentrates those elements more strongly, so any stream that upgrades monazite also upgrades radioactivity. The dry plant layout must plan for radioactive dust control from the first line on the drawing. That means sealed transfer points, filtered air handling, and segregated storage for the monazite product.

Once monazite is concentrated beyond a certain level, transport classification changes. A bulk container of heavy mineral concentrate may be treated differently from a drum of monazite-rich product. You need to know the activity per gramme, not just the percentage of monazite in the mass. Ask your radiation safety adviser for the transport classification thresholds before you commit to a monazite product. Zircon itself is the main source of zirconium and hafnium, which occur in it at a ratio of about 36 to 1 (USGS Mineral Commodity Summaries 2022). The leading end use for zircon was ceramics, with foundry sand, refractories and zirconium chemicals also significant (USGS Mineral Commodity Summaries 2026). That context matters, because the same concentrate that supplies a ceramic grade zircon can also produce a monazite by-product that needs careful handling.

Before you commit to a dry plant, get quantitative mineralogy by size fraction. This is the first deliverable. A bulk assay that says 3 per cent zircon tells you nothing useful. You need to know, for example, how much of the zircon is in the minus 150 plus 75 micrometre fraction, and how much rutile is in the same fraction. Electrostatic and magnetic separation only work well on narrow size fractions, so the testwork must be done on the actual fractions your dry plant will see. Quantitative mineralogy means QEMSCAN, MLA or equivalent automated mineralogy by size fraction. A petrographer's visual estimate is not enough.

Bulk sample size is not a fixed number. It depends on your top size, the number of testwork stages, and the variability of the deposit. The supplier should ask for the same mass that a pilot plant needs, not less. Think in terms of enough mass to run a full factorial of electrostatic and magnetic settings on at least three size fractions, with enough left for a pilot run. Ask any supplier for a mass balance by size and mineral, not a single grade. If they can't give you that, the testwork is not ready. Bench-scale electrostatic and magnetic tests on representative size fractions will tell you whether a dry separation sequence is feasible. After that, a pilot-scale run on a larger sample confirms the sequence before you spend capital. You'll find the testwork questions covered on the mineral processing testwork page, and the broader flowsheet logic under solutions. For more field notes like this, see insights.

That's the part the dry plant flowsheet doesn't show you. Lock the wet plant first, read the mineralogy by size fraction, and test the actual separation response before you cut steel.