Removal of metal ions from water by nanocomposites in the batch an‎d continuous fixed-bed column adsorption process

12/30/2026 12:00:00 AM

In this thesis, the performance of two novel, cost-effective, an‎d eco-friendly adsorbents—mordenite modified with tin dioxide nanoparticles an‎d biochar modified with tin dioxide nanoparticles—was compared for the removal of lead ions from water. Based on batch adsorption experiments, the optimal conditions for lead ion removal using both adsorbents were determined as 25 °C, a contact time of 100 min, an initial concentration of 100 μg/L, an‎d an adsorbent dosage of 0.05 g. Furthermore, fixed-bed column experiments revealed that the biochar composite modified with tin dioxide nanoparticles exhibited superior performance, with an adsorption capacity of 1396.394 mg/g, compared to the mordenite-based adsorbent with an adsorption capacity of 810.559 mg/g. In addition, the treated water volume at breakthrough was found to be 1.944 L for the mordenite-based adsorbent an‎d 2.916 L for the biochar-based adsorbent. According to the distribution coefficient, both nanocomposites demonstrated good selec‎tivity for low concentrations of lead ions; however, the mordenite-based adsorbent exhibited higher selec‎tivity due to its regular porous structure an‎d intrinsic structural properties. Modeling studies of adsorption isotherms an‎d kinetics showed excellent agreement with the Langmuir an‎d Thomas models. Moreover, reusability tests indicated that both adsorbents experienced less than 10% reduction in capacity after five adsorption–desorption cycles. In a separate investigation, the adsorption capacity of chitosan-based composites for chromium ions was predicted using machine learning models trained on data from scientific literature. The results demonstrated that the adaptive boosting model outperformed gradient boosting an‎d ran‎dom forest algorithms in predicting chromium ion adsorption, achieving acceptable performance metrics (R² = 0.830, MSE = 5.812, MAE = 0.008). Feature selec‎tion analysis revealed that solution pH, with a score of 71%, had the greatest influence on the adsorption process. For model validation, a novel chitosan-based adsorbent incorporating biochar–nanochitosan–zirconium was experimentally tested, an‎d the results showed strong consistency with the adaptive boosting model predictions (R² = 0.825, RMSE = 7.406). This study highlights that machine learning models provide an efficient alternative to costly an‎d time-consuming experiments, offering a promising approach to mitigating water pollution an‎d enhancing public health an‎d environmental sustainability.

جذب سطحي , يون سرب , يون كروم , نانوكامپوزيت , يادگيري ماشين , تصفيه آب

Adsorption , lead ion , chromium ion , nanocomposite, , machine learning , water treatment

Fatemeh Yazdi

Professor Mansoor Anbia