Abstract
Selective flotation of low-grade ilmenite ores is complicated by the similarity in surface properties between ilmenite and silicate gangue, while reagent performance depends jointly on several interacting process variables. The feed ore contained 6.8 wt.% TiO2 and consisted predominantly of silicate gangue, confirming its low-grade character and the need for selective flotation. This study presents a response surface methodology (RSM) model for ilmenite flotation using a single modified lignosulphonate collector, LS-O2, obtained by permanganate oxidation. A three-factor, three-level Box-Behnken design was used to relate pulp pH, reagent dosage, and temperature to ilmenite recovery and the TiO2 grade of the concentrate. Second-order polynomial models for both responses were statistically significant (p<0.001), with no significant lack of fit, and explained 99.6% and 98.4% of the variance in recovery and TiO2 grade, respectively. Reagent dosage was the dominant factor for both responses, followed by pH, while temperature had a comparatively minor effect. For TiO2 grade, the linear temperature term and temperature-related interactions were not significant, although the quadratic temperature term remained significant. Joint optimisation using a composite desirability function identified an optimum at pH 7.31, an LS-O2 dosage of 0.331 kg/t, and a temperature of 28°C, predicting 76.6% recovery and a TiO2 grade of 21.4%. Model predictions were verified at five independent operating points, including a boundary combination and an extrapolated condition, with deviations not exceeding 2.52%.