Raw ilmenite (FeTiO3) is only weakly magnetic and can't be cleanly pulled away from silicate gangue with a simple magnet. A magnetizing (reduction) roast exposes hot ore to CO/H2 gas, driving:
3 FeTiO3 + CO → Fe3O4 (magnetite) + 3 TiO2 + CO2
Magnetite is strongly ferromagnetic, so the fraction of each grain that has converted controls how hard a magnet can pull on it. Conversion follows first-order Arrhenius kinetics:
k = k0 · gasFactor · exp(-Ea / R·T)
X(t) = 1 − exp(−k·t)
Individual grains scatter around the mean conversion (uneven heat and gas contact in a real kiln), so the roasted stream is never uniform. In the separator, a grain's magnetic pull scales with its own conversion; the field strength sets how weak a pull is still enough to deflect a grain into the magnetic concentrate.
- Higher temperature / longer time / more reducing gas — faster, more complete conversion to magnetite, so more grains become strongly magnetic.
- Higher field strength — captures more grains into the concentrate (higher Fe recovery) but also drags in barely-converted, still-titanium-rich grains, diluting the concentrate's magnetic grade. This is the real recovery-vs-grade tradeoff every magnetic separation plant has to tune.
- Under-roasting (low T, short time, weak gas) leaves most grains non-magnetic — raising the field just floods the concentrate with unconverted gangue instead of fixing the problem.