The present study evaluates the amoxicillin adsorption from aqueous solutions using biochar derived from sugar beet pulp (SBC), in batch conditions. The biomass improved by phosphoric acid pretreatment followed by pyrolysis. Adsorption variables were optimized by the response surface methodology (RSM). In this regards, initial amoxicillin concentration (7.5 to 77.5 \(\text{mg }{\text{L}}^{-1}\) ), solution pH (2.5 to 10.5), sorbent dosages (0.05 and 1.85 \(\text{g }{\text{L}}^{-1}\) ), and adsorption time (5 to 105 min) were considered. Furthermore, the effect of temperature at 293 to 323 K were considered in optimized conditions. According to the obtained results, the specific surface area and total pore volume of the phosphoric acid-modified SBC (MSBC) exhibited significant enhancement, increasing from 4.46 to 564.25 \({\text{m}}^{2} {\text{g}}^{-1}\) and from 0.004 to 0.350 \({\text{m}}^{3} {\text{g}}^{-1}\) , respectively. The equilibrium data were well represented by the Sips and Redlich-Peterson isotherm models. The maximum adsorption capacity for amoxicillin, as determined by the Sips and Langmuir isotherms, were found to be 110.63 and 66.87 \(\text{mg }{\text{g}}^{-1}\) , respectively, at 298 K. Kinetic analysis revealed that the adsorption process followed to the pseudo-second-order kinetics model. Furthermore, thermodynamic analyses indicated that the adsorption process is spontaneous, exothermic, and primarily physical. The introduction of 0.1 M sodium, calcium, and magnesium chloride salts resulted in a 21–34% reduction in amoxicillin removal efficiency. Additionally, this research assessed the effectiveness of non-linear least squares modeling in estimating the parameters of pseudo-order equations, comparing its efficacy to that of linear modeling. The findings advocate for the adoption of non-linear optimization methods in the estimation of parameters for kinetic models to enhance accuracy and reduce potential errors.
