The effect of Na2SO3 as a pyrite depressant in NaCl and KCl saline media and the presence of kaolinite were evaluated by zeta potential tests. Chalcopyrite was also included in the study, because pyrite usually accompanies this mineral. Subsequently, the floatability results of both minerals in the NaCl solution were optimized based on the design of experiments (DoE). The Box–Behnken DoE was applied considering the percentage of kaolinite (X1), collector dose (X2), and depressant dose (X3) as factors. The results were modeled using artificial neural networks (ANNs) to construct contour plots and to determine the optimal conditions. In particular, maximization of the mass recovery of chalcopyrite and minimization of that of pyrite were sought. The particle swarm optimization algorithm was used as an optimization technique. The results indicated that the optimal conditions to maximize the floatability of chalcopyrite were kaolinite 6.85%, collector dose 3.58 × 10–3 mol/dm3, and depressant dose 3.49 × 10–5 mol/dm3. On the contrary, the optimal conditions to minimize the floatability of pyrite were 5% kaolinite, collector dose 5 × 10–4 mol/dm3, and depressant dose 6.4 × 10–5 mol/dm3. Under these conditions, the mass recoveries of chalcopyrite and pyrite were 66.1% and 14.0%, respectively. The results also indicated that the presence of kaolinite negatively affects the flotation of chalcopyrite, while the effect of Na2SO3 is not significant. In general, the findings suggest that Na2SO3 is a viable alternative to consider as a pyrite depressant in saline environments.

Keywords: Na2SO3, IPETC, clays; flotation, saline solution, artificial neural networks