Comparison of RSM and ANN prediction approaches for crystal violet removal by Fe-modified Cedrus libani carpel biochar

In this study, Cedrus libani a coniferous tree species native to the Mediterranean region was used as the precursor material. This species grows naturally, particularly in the Taurus Mountains of Türkiye, as well as in Lebanon and Syria. The adsorbent material (Fe-mCLCB), produced through iron modification of Cedrus libani carpel biochar, was employed for the removal of crystal violet (CV) from aqueous solutions. To evaluate the structural, morphological, surface, and functional properties of the prepared adsorbent, Fe-mCLCB was characterized by XRD, BET, SEM-EDS, and FT-IR analyses. The effects of temperature, contact time, adsorbent dosage, and initial CV concentration were optimized using response surface methodology (RSM). The adsorption kinetics were better described by the pseudo-second-order model, while the equilibrium data were best fitted by the Langmuir isotherm model, with a maximum adsorption capacity of 31.949 mg/g. Within the limitations of the dataset, the ANN model showed better agreement with the experimental data than the RSM model. These findings indicate that machine-learning-based modeling can be effectively applied to the removal of crystal violet using Fe-modified Cedrus libani carpel biochar.