Nano-magnetic walnut shell-rice husk for Cd(II) sorption: design and optimization using artificial intelligence and design expert

This study attempted to investigate the use of nanomagnetic activated carbon prepared from walnut shell and rice husk wastes for removal of Cd(II) from aqueous solution via application of ANN and design expert as adsorbent preparation design and optimization tools. The novel adsorbent was characterized using SEM, FTIR, EDS and BET. The result from 2-level factorial design expert revealed 78.58% Cd(II) sorption efficiency could be achieved for adsorbent prepared at optimum calcination temperature, calcination time, SS-RH mixing ratio and magnetite loading of 859.20 degrees C, 2.32 h, 2.54 and 5.56 wt% respectively. Sensitivity analysis by both proposed methodologies revealed calcination temperature as most influential factor in adsorbent preparation. Average relative errors and R-2 values of 1.2931% and 4.806%; and 0.9967 and 0.9055 obtained respectively for developed ANN model with 4-9-1 architecture and 2-level factorial design expert revealed ANN model as better prediction and optimization tool for Cd(II) sorption using NM-WS-RH-AC. Laboratory analysis revealed presence of -OH, -NH and COO- groups on adsorbent surface; presence of Cd(II) after adsorption; change in adsorbent textural and morphological structure after Cd(II) adsorption; and increase in its surface area and average pore diameter due to magnetization. Average relatively stable desorption strength of 62.74% towards Cd(II) was exhibited by adsorbent for four consecutive cycles using 0.1M HNO3. Prepared adsorbent is effective in removing Cd(II) from solution than commercial activated carbon with economically viable regeneration attribute.