MWCNT-Fe3O4 as a superior adsorbent for microcystins LR removal: Investigation on the magnetic adsorption separation, artificial neural network modeling, and genetic algorithm optimization

Magnetic multi-wall carbon nanotube (MMWCNT) was prepared by simple protocol and its structural features were characterized using SEM, TEM, and XRD analysis. The association between removal (%) and variables such as pH (3 − 11), adsorbent amounts (0.005, 0.1, 0.25, 0.5, 0.75, and 1 g/L), reaction time (5–180 min), and concentration of microcystins-LR (10, 25, 50, 75, and 125 μg/L) was investigated and optimized. The results of the isotherm study indicated that Langmuir offered high determination coefficients (R2 = 0.993, 0.996, and 0.998, for the three different working temperatures of 20 °C, 35 °C, and 50 °C respectively) and was the optimum isotherm to anticipate adsorption of MC-LR (microcystins-LR) by magnetic MWCNT adsorbent. The kinetic study revealed that the adsorption kinetics of MC-LR could be better defined using the pseudo-second-order model. A three-layer model of an artificial neural network was applied to forecast the MC-LR removal efficiency by magnetic MWCNTs over 66 runs. To forecast the MC-LR removal efficiency, the minimum mean squared error of 0.0011 and determination coefficient (R2) of 0.9813 were obtained. The use of the artificial neural network model achieved a good level of compatibility between the acquired and anticipated data.