Competition & UV254 projection in odorants vs natural organic matter adsorption onto activated carbon surfaces: Is the chemistry right?

Abstract

Powdered activated carbon (PAC) adsorption remains an indispensable method for addressing odor problems in drinking water. While natural organic matter (NOM) is ubiquitous and competes strongly in deteriorating odorant adsorption capacity, it can also serve as a promising indicator for predicting odorant adsorption through online measurement. However, the impact of PAC surface chemistry on NOM competition and feasibility of prediction across various adsorbents are not well understood. Here, we examined the role of PAC properties (pore structure and surface chemistry) in the competitive adsorption between odorants and NOM components, aligned with the applicability assessment of using NOM optical properties for odorant adsorption projection across various PAC samples. Chemical oxidation and thermal treatment achieved considerable changes in surface functional group composition, alongside minimal changes in pore structure, of two typical PAC products with microporous/mesoporous pore characteristics. The effect of NOM interference on the reduction of odorant adsorption exhibited a similar level regardless of the PACs with different pore structure (average pore size of 1.7 nm vs. 4.2 nm). Surface modification increased the equilibrium adsorption capacity (q_e50) of odorants by 15.1% to 146.4% (thermal treatment) or decreased by -81.3% to -34.1% (chemical oxidation), respectively, but minimal changes in odorant-NOM selectivity. For various odorants, hydrophobicity (log D) influenced the adsorption capacity while the structural flexibility (reflected by the rotatable bonds) affected the vulnerability of odorant adsorption to NOM competition. It was found for the first time that four-parameter Richards model (RMSE = 2.6%) is superior to the linear model (RMSE = 12.5%) or logarithmic model (RMSE = 77.6%) to describe the S-shape UV₂₅₄ projection curves associated with odorant adsorption on PAC. Moreover, the feasibility was confirmed to use UV₂₅₄ projection curves of pristine PAC fitted with the Richard model to predict the odorant adsorption on surface-modified PAC in two different surface waters (RMSE 9.2% and 7.4%, respectively). This study provides insight into the role of PAC surface chemistry and pore characteristics in odorant adsorption in NOM-containing waters and enhances the feasibility of the NOM surrogate model for odorant monitor and control during PAC adsorption.