Integrating experimental and predictive modeling approaches for phenolic pollutants removal using bisphenol-based hyper-cross-linked resin

The widespread occurrence of phenolic pollutants, bisphenol A (BPA) and para-nitrophenol ( p- NP), poses serious environmental and health concerns due to their persistence and toxicity. In this study, we synthesize a novel bisphenol-based hyper-cross-linked resin (BPA-AM ) and systematically use it for the efficient removal of these pollutants from aqueous solutions. Adsorption experiments show that pH, contact time, and adsorbent dosage significantly influence removal efficiency, with maximum adsorption capacities of 93.8 mg/g for BPA and 100.9 mg/g for p -NP. Following the DFT calculations, the resin shows stronger binding affinity for p -NP (ΔG = −31.24 kcal/mol) than for BPA (ΔG = −21.57 kcal/mol), driven by enhanced hydrogen bonding (H-bonding) and π–π interactions. Additionally, Artificial Neural Network (ANN) and Adaptive Neuro-Fuzzy Inference System (ANFIS) models were developed to predict adsorption efficiency, with the ANFIS model providing superior accuracy (R² = 0.9846 for BPA and 0.9774 for p -NP). Furthermore, the resin shows an excellent anti-resistance behavior toward heavy metal ions and salt anions. However, the presence of organic co-pollutants, p -chlorophenol and p -chloroaniline, inhibits their adsorption performance, although the resin maintains an efficiency at >40%. The resin retained over 90% of its adsorption capacity after five reuse cycles, confirming excellent stability and reusability. The current resin shows strong potential for removing persistent phenolic pollutants, such as BPA and p -NP, from water. These findings highlight BPA-AM's significant potential for practical and sustainable wastewater treatment applications.