Abstract:
The growing presence of antibiotics in urban water systems poses serious risks to
environmental and human health. This thesis presents a comprehensive approach
to remove ciprofloxacin (CIP), a widely used and persistent antibiotic, using a TiO₂-
coated CuWO₄ (CuWO₄@TiO₂) core-shell nanocatalyst. The nanomaterial was
synthesized via a two-step hydrothermal method and thoroughly characterized
using FTIR, XRD, FESEM, XPS, and BET techniques, confirming its core-shell
architecture, surface properties, and active sites. Batch adsorption experiments,
optimized through Response Surface Methodology (RSM), demonstrated over 99%
CIP removal under neutral pH conditions with optimal parameters including 10
mg/L initial CIP concentration, 300 mg/L adsorbent dose, and 40 minutes of contact
time. The adsorption process followed the Langmuir isotherm and pseudo-secondorder
kinetic models, with a maximum theoretical adsorption capacity of 109.58
mg/g. The material showed strong reusability, retaining high removal efficiency
after five adsorption–desorption cycles. Complementing the experimental study, a
multilayer perceptron (MLP) neural network model was developed using
experimental data to predict CIP removal efficiency. Hyperparameter tuning via
Bayesian and Random Search, coupled with Particle Swarm Optimization (PSO),
yielded highly accurate predictions, with R² values up to 0.9985 for training and R²
0.9856 for testing. Feature importance analysis identified contact time and
adsorbent dosage as the most influential factors. This work provides a robust and
transparent machine learning-integrated framework for optimizing nanomaterialbased
antibiotic removal, contributing to the advancement of sustainable water
treatment technologies.
