Influence of Process Variables on Ilmenite Leaching in Binary Solution: Experimental Evaluation and Artificial Neural Network Modeling

Abstract

This research is focused on evaluating the suitability of an unconventional binary lixiviant in the leaching of iron from ilmenite. Batch dissolution studies were conducted to examine the influence of some independent process variables: particle size, acid concentration, oxidant concentration, solution temperature, stirring speed and liquid-to-solid ratio on the % iron recovery from ilmenite. A three-layered (5:n:1) feedforward architecture of artificial neural network trained by the Levenberg–Marquard back-propagation algorithm were developed to model and predict the leaching process parameters. SEM, XRF and XRD characterizations of the ilmenite suggests that the sample is crystalline with titanium and iron as dominant metals primarily existing as FeTiO3. The batch dissolution study results reveal that examined process variables significantly affected the response variable. A maximum % iron recovery of 96.08% was achieved at 75 μm particle size, 1M acid concentration, 0.6M oxidant concentration, 75˚C solution temperature, 300rpm stirring speed and 30L/g liquid-to-solid ratio. A machine learning model, artificial neural network (ANN) results show the robustness of ANN model in capturing the nonlinear behaviour of the leaching system. The outcome of this study reveal that the unconventional lixiviant, HCl-KClO3 has the potential of leaching iron from ilmenite.