Decoupling functional group effects on uranium adsorption in UiO-66 through density functional theory and neural...

The vast uranium resources in seawater provide a sustainable foundation for nuclear energy development. These abundant resources also offer a potential solution to long-term uranium supply challenges caused by dwindling terrestrial reserves and increasing ecological pressures. Therefore, developing high-performance adsorbent materials is essential for the efficient and selective extraction of U(VI) from seawater. Herein, pristine UiO-66 and a series of functionalized UiO-66, denoted as 2X-UiO-66 (X = H, OH, NH₂, SH, and AO), were synthesized to systematically investigate the effects of different functional groups on U(VI) adsorption. Among them, 2AO-UiO-66 exhibited the highest adsorption capacity (344.5 mg/g) and favorable selectivity toward U(VI) in the presence of competing metal ions. X-ray photoelectron spectroscopy and density functional theory calculations revealed the superior affinity of amidoxime-functionalized UiO-66 toward U(VI). This enhanced adsorption performance originated from N,O dual-site synergistic chelation, in contrast to the monodentate binding modes observed for hydroxyl-, amino-, and thiol-functionalized UiO-66. Furthermore, a Stacking ensemble model demonstrated superior predictive performance on an independent test set (R² = 0.9964, RMSE = 5.0768 mg/g), outperforming all individual models in forecasting the U(VI) adsorption capacity of 2X-UiO-66. SHapley Additive exPlanations (SHAP) analysis further quantified the feature contributions and identified contact time and average pore size as the dominant factors affecting U(VI) adsorption. Overall, this work establishes an integrated framework combining experiments, quantum-chemical calculations, and interpretable neural network modeling, providing valuable guidance for the rational design of adsorbents for U(VI) extraction from seawater.