Bioremediation of turquoise blue by Mangifera indica — particle swarm optimization and kinetic modeling

Treatments of dye contaminated waste water by conventional physico-chemical methods are less attractive due to various factors and bioremediation using natural plant-based sorbents has gained attention due to eco-friendliness and economical advantages. In this research, a natural adsorbent derived from Mangifera indica was employed to remove the dye turquoise blue. Central composite design technique was implemented to investigate the factors influencing the dye adsorption onto the activated Mangifera indica shell and their interactions. The Lasso and Ridge machine learning models were employed to optimize adsorption. The Ridge model accurately predicts the experimental data. Furthermore, a constrained nonlinear particle swarm optimization was employed to maximize the dye removal efficiency. It was observed that the maximum removal percentage of dye using this adsorbent was 87.44% under optimum operating conditions. The adsorption isotherm was best fitted by the Redlich–Peterson model in all the temperature conditions investigated. The parameter values computed from cftool kit in MATLAB are observed as k_RP = 2.4420 1/g, a_RP = 0.0474 1/mg^β, and b_RP = 1.2360 at the optimum temperature of 50 °C with an R² of 0.9740. The kinetic studies revealed the fit of Lagergren’s pseudo first order kinetics (k₁ = 0.1047 1/min and q₁ = 11.230 mg/g) in sorption kinetics for which the R² value was reported as 0.9980.