Antimicrobial study and biosorption of Pb2+ ions onto chitosan-walnut composites: mechanistic studies and neuro-fuzzy modeling approach

Abstract

The upsurge in the discharge of lead ions (Pb2+) into the environs resulting from various anthropogenic activities vis-vis its adverse effect on public health is a call for great concern. However, the adsorption technique, amongst other heavy metal removal methods, is regarded as the most promising. The present study synthesized a walnut shell-chitosan composite (WNS-CH) as an efficient biosorbent for Pb2+ uptake and biofilter of Bacillus subtilis and Klebsiella pneumoniae. WNS-CH was characterized using SEM and FTIR. Furthermore, an intelligent and cost-effective machine learning model, an adaptive neuro-fuzzy model clustered with the grid-partitioning (GP), and fuzzy c-means (FCM) technique were developed to predict the adsorption of Pb2+ based on relevant input parameters. The batch adsorption was carried out by varying operating parameters such as contact time, temperature, pH, adsorbent dose, and initial adsorbate concentration. The SEM images of WNS-CH showed a homogenous regular hollow ellipsoidal morphologies, while FTIR spectra showed the presence of O-H, N-H, C-N, and C-O. Under the conditions of initial pH 10, dosage 45 mg, and temperature of 40 °C, an adsorption efficiency of 94 % was obtained. The thermodynamic parameters, ∆H° and ∆G°, showed an endothermic and spontaneous process for Pb2+ uptake. Antibacterial activities of the WNS-CH composite showed bioactivity against Bacillus subtilis and Klebsiella pneumoniae with a mean ZI of 5.3±1.16 and 6.0 ±1.00, respectively. The experimental data was best described by Freundlich isotherm (R2= 0.9509) and pseudo-first-order kinetic (R2= 0.9674) models indicating chemisorption and multilayer adsorption process. The best prediction of Pb2+ adsorption was obtained with the optimal GP-clustered ANFIS model using a triangular membership function (triMF), giving Root Mean Square Error (RMSE), Mean Absolute Deviation (MAD), Mean Absolute Percentage Error (MAPE), and correlation determination (R1) values of 1.217, 0.563, 1.698, and 0.9966 respectively at the testing phase. The GP-ANFIS model shows good agreement with experimental results. This study revealed that WNS-CH composite could be regarded as a promising biosorbent for the remediation of Pb2+-polluted wastewater. The cost analysis demonstrated that the WNS-CH composite could serve as an alternative to commercial activated carbon.