Separation and preconcentration of Pb(II) and Cd(II) from aqueous samples using hyperbranched polyethyleneimine-functionalized graphene oxide-immobilized polystyrene spherical adsorbents

Accurate determination of metal ions in wastewater is essential for environmental monitoring and assessment. In aqueous media, the presence of metal ions at trace concentrations and the interference of matrices pose difficulty in its quantification. Hence, a radical way to eliminate the matrix components is to preconcentrate the target metal ions. Graphene oxide, due to their exceptional higher surface area, has received significant attention and research interest in the field of adsorption science. However, direct use of graphene oxide in preconcentration of metal ions faces irreversible aggregation and leakage from column. To overcome these challenges, we synthesized a polyethylenimine-functionalized graphene oxide-immobilized polystyrene adsorbent and employed for the column preconcentration of Pb(II) and Cd(II) from wastewater samples. The optimum pH for Pb(II) and Cd(II) preconcentration were found to be 6 and 8, respectively. The method showed a preconcentration factor of 833 and 900 and a detection limit of 1.5 ± 0.2 ng L⁻¹ and 1.8 ± 0.3 ng L⁻¹ for Pb(II) and Cd(II), respectively. The preconcentration limit for Pb(II) and Cd(II) were found to be 0.80 μg L⁻¹ and 0.74 μg L⁻¹, respectively. The PS@GOA adsorbent can be used repeatedly up to 72 cycles without any loss of the extraction efficiency. The prepared adsorbent is chemically stable, showing no leaching of graphene oxide from the column. The accuracy of the method against the systematic and constant errors was resolute by analyzing SRM 1572 and NIES 10C (<5% RSD). The method can be suitably applied for the routine analyses of Pb(II) and Cd(II) in wastewater samples. The amount of Pb(II) and Cd(II) in river water, industrial wastewater and tap water samples were found to be 3.8, 28.6, 1.2 ng mL⁻¹ and 5.6, 2.2, and 3.2 ng mL⁻¹, respectively. The recoveries of the analytes ranged from 98.0 to 104%. The developed method showed good reproducibility with an average error of ±4%.