Alumina nanotube–ZIF nanocomposites for high-capacity crystal violet dye adsorption: Experimental and machine learning investigations

The presence of crystal violet (CV) dye in wastewater poses serious environmental and health risks, highlighting the need for efficient removal strategies. In this study, alumina nanotubes (ANTs), ZIF-8, ZIF-67, and ANTs-based ZIF nanocomposites (ANTs@ZIF-8 and ANTs@ZIF-67) were synthesized for the first time herein and comprehensively characterized by XRD, FTIR, SEM, and BET analyses. Among the tested materials, ANTs@ZIF-67 demonstrated the highest CV uptake under screening conditions (100 mg/L CV, 50 mg/L adsorbent dose, pH 6.4, 25 °C), achieving 1649.4 mg/g. The equilibrium isotherm was well described by the Langmuir (q max = 3391.8 mg/g, R 2 = 0.9827) and Redlich–Peterson (R 2 = 0.9830, g = 0.97) adsorption isotherm models, indicating Langmuir-type behavior under the tested conditions. Additionally, adsorption kinetics were best represented by the Avrami model (R 2 = 0.9973), while thermodynamic parameters (ΔH = +35.63 kJ/mol; ΔG < 0) confirmed the endothermicity and spontaneity of the CV adsorption process. The composite retained high performance across five adsorption–desorption cycles, underscoring its stability and reusability. In parallel, ensemble machine learning models trained on the experimental dataset identified Extra Trees as the most accurate predictor of adsorption capacity (R 2 = 0.998, RMSE = 25 mg/g). This work represents the first report of ANTs/ZIF nanocomposites, establishes their potential in dye remediation, and highlights the integration of experimental results with data-driven prediction of adsorption performance. • ANTs@ZIF-67 nanocomposite provided an ultra-high adsorption of CV (q max > 3300 mg/g) • CV Adsorption is endothermic and spontaneous process, with spontenioty increases with T • ANTs@ZIF-67 nanocomposite is robust adsorbent with excellent reusability • ML modelling revealed that Extra Trees provides the best prediction of CV adsorption