Cr(VI) removal by cationic cellulose nanofiber aerogels: batch and fixed-bed adsorption, reduction mechanism, and machine learning breakthrough predictions

• Cationic cellulose nanofibers were used for Cr(VI) adsorption. • Both batch and continuous experiments were performed for Cr(VI) uptake. • The adsorbent displayed excellent adsorption capacity, exceeding 100 mg/g. • Properties like color, conductivity and pH were correlated with Cr concentration. • Random Forest was effectively employed to predict outlet Cr concentration. A nanocellulose-based aerogel functionalized with quaternary ammonium groups was studied for Cr(VI) adsorption from water in both batch and continuous systems. The adsorbent performed stably across a pH of 1 to 8, with optimal performance at pH 2, where Cr(VI) removal was enhanced by in-situ reduction to Cr(III). Adsorption was rapid, reaching equilibrium within five minutes, and followed the Freundlich model, indicating multilayer uptake. The adsorbent displayed excellent adsorption capacity, exceeding 100 mg/g. Incorporation of alkyl ketene dimer (AKD) increased Cr(VI) uptake, but slowed the adsorption rate. Kinetics studies revealed pseudo-second-order behavior for adsorption and first-order for reduction. In continuous-mode, 0.5% AKD provided the best performance, with earlier breakthrough at higher AKD concentrations. Increasing inlet Cr(VI) concentrations led to faster saturation and higher uptake, while higher flowrates reduced breakthrough time and adsorption capacity. During continuous-flow experiments, direct properties (color, conductivity, pH) were recorded to evaluate correlations with outlet Cr concentrations. Based on these, Random Forests were effectively employed to predict Cr(VI) and total Cr concentrations using operational parameters (model A) or effluent characteristics (model B). Both yielded high accuracy (R² > 0.9), with Model A offering potential for process optimization, while Model B showed potential for real-time malfunction detection.