Artificial neural network-based fungal chitin production for submicron-chitosan synthesis: effects on bioremediation for heavy metal pollution.

This study focused on optimizing fungal chitin (CT) production from a newly identified Fusarium incarnatum (GenBank: OL314753) for subsequent synthesis of submicron chitosan (sm-CS) tailored for enhanced heavy metal removal. Initial attempts to optimize CT yield using mixed nitrogen sources (potassium nitrate, peptone, yeast extract) via Box-Behnken Design (BBD) were insufficient for predicting optimal conditions. Consequently, Artificial Neural Networks (ANN) successfully modeled the BBD experimental data, the optimal nitrogen concentrations to 4.89, 5.70, and 4.25 g/L, respectively, achieving an experimental CT yield of 3.998 g/L. The fungal CT was deacetylated to chitosan (CS), subsequently processed into sm-CS using ionotropic gelation. Characterization (FTIR, Raman, SEM, HR-TEM, EDX) confirmed the successful formation of sm-CS, demonstrating reduced particle size and increased surface area. Batch adsorption experiments revealed the superior heavy metal sequestration capacity of sm-CS, removing 80-90 % of Fe, Mn, Cu, and Zn, which outperformed CS (20-60 % removal). CS preferentially adsorbed Mn, whereas sm-CS showed affinity for Cu. sm-CS demonstrated excellent reusability, maintaining >85 % efficiency through five adsorption-desorption cycles, markedly surpassing CS regeneration. This integration of specific fungal strain, ANN optimization, and sm-CS, offers an efficient material for addressing complex multi-metal contamination, potentially mitigating nanotoxicity risks associated with some nanocomposite adsorbents.