Acetaminophen adsorption to spherical carbons hydrothermally synthesized from sucrose: experimental, molecular, and mathematical modeling studies

Abstract

Acetaminophen (AAP) is an analgesic and non-steroidal anti-inflammatory drug and a micropollutant that has been detected in waterbodies worldwide. Here, we explore the characteristics of AAP adsorption onto spherical carbons (SCs) hydrothermally synthesized from pure sucrose as a carbon source. In one-factor-at-a-time experiments, the adsorption capacity of AAP remained relatively constant between pH 2 and 10 but became negligible at pH 12. The Raman, FTIR, and XPS spectra illustrate that hydrogen bonding, π-π interactions, and n-π^* interactions could contribute to the AAP adsorption onto the SCs. CHEM3D modeling was used to explore hydrogen-bond formation, π-π interactions, n-π^* interactions, and electrostatic repulsion between AAP and the SCs. In view of the pH_pzc of the SCs (3.1) and the pKa of AAP (10.96), electrostatic repulsion could occur between negatively charged SCs and anionic AAP above pH 10. In consideration of the average pore diameter of the SCs (1.89 nm) and the AAP molecular size (8.94 Å × 7.95 Å × 4.93 Å), a pore-filling mechanism could contribute to the adsorption. A pseudo-second-order model was best fitted to the kinetic data (equilibrium time = 6 h), whereas the Liu isotherm was most suitable for the equilibrium data (maximum adsorption capacity = 92.0 mg/g). Adsorption of AAP to the SCs was exothermic at 10-40 °C. The SCs were regenerated and reused for AAP adsorption using a methanol. Multiple-factor-at-once (MFAO) experiments (input variables: pH, temperature, adsorbent dosage, and initial AAP concentration; output: AAP adsorption capacity) were used to develop response surface methodology (RSM, quartic regression) and artificial neural network (ANN, topology 4:11:9:1) models. Analyses using additional MFAO experimental data reveal that the predictive ability of the ANN model (R² = 0.890) was better than that of the RSM model (R² = 0.764). Based on the weight values of the ANN model, the relative importance of the input variables on the output was quantified in the order of initial AAP concentration (100%) > adsorbent dosage (92.3%) > temperature (77.6%) > pH (43.6%).

Keywords: Acetaminophen; Adsorbent; CHEM3D modeling; Multiple-factor modeling; Spherical carbons.