Prediction of Heavy Metal Adsorption by Biomass Ash Using Artificial Neural Network

An invasive Prosopis tree biomass ash was used as an adsorbent for heavy metals from a mine leachate. The adsorption experiments were carried out by varying adsorbent solid loading, adsorption time and temperature. The biomass ash was capable of neutralising the acidic mine leachate using a solid loading of 2.5% m/v after 1 h of adsorption. The adsorption capacity of the biomass ash was 0.216, 40.267, 0.681, 0.470 and 0.101 mg/g for Cr (III), Fe (III), Cu (II), Zn (II) and Pb (II) respectively, with over 96% metal ion removal. The adsorption process could be modelled well using a three parameter Sips isotherm and pseudo first order kinetic model. Physisorption was the main mechanism of adsorption. A 3,5,5 feed forward multilayer perceptron artificial neural network architecture could predict the biomass ash performance with a correlation coefficient of 0.99. This study therefore provides opportunities for the circular economy handling of biomass ash which then satisfies the dictates of sustainability.