Comparison study of hematite bioflotation by r. erythropolis and its biosurfactant: Experiments and neural network modeling

The relevance of Biotechnology applied to mineral processing has been increasing during the last decade. Initially, microorganisms as bacteria and yeast were used as collectors due to their hydrophobic properties. However, the complexity of their cell wall makes the understanding of the adsorption mechanisms involved during the microbial cell-mineral interactions difficult. Thus, more simple substances such as their metabolic products or extracellular polymeric substances (EPS) started to be extracted and used. One of these substances, biosurfactants, has shown a great potential as collectors and also as frothers. This work presents a comparative study between R. erythropolis bacteria and its biosurfactant in the flotation of hematite. The objective was to study the floatability efficiency of the biorreagents using artificial neural network. The experimental conditions varied from 3 – 11 to pH, 0 – 150 mg/L to concentration. Recovery of hematite analysis (%) was performed to evaluate the efficiency of the process. A database was constructed with the information of the experiments, dividing them into groups of training (65%) and test (35%). The models were obtained using toolbox of the MATLAB R2017a. In the developed neural network, the data obtained by the different conditions were used as neurons in the input layer and the percentage of hematite recovered was used as the only neuron in the output layer. The performance of the neural network was evaluated by the correlation coefficient (R²⁾ and the error index (SSE). The model developed from the neural network was satisfactory, since the R² value was close to 1 and the error index values were close to 0. In addition, the angular and linear coefficients of the adjustment lines were respectively close to 1 and 0, confirming the good fit between the data tested and the developed model.