Deeper Insights into the Bohart–Adams Model in a Fixed-Bed Column

Abstract

The Bohart–Adams model was one of the most widely used breakthrough models in column experiments. However, it usually provided a poor fit for the modeling of an asymmetric breakthrough curve. This work proposed the n-order Bohart–Adams and fractal-like Bohart–Adams models. The former indicated a nonlinear decay process of the concentration of the adsorbate or residual capacity of the adsorbent, while the latter reflected a diffusion-limited process on the heterogeneous surfaces. The Bohart–Adams and modified Bohart–Adams models were mathematically equivalent. The applicability of the n-order Bohart–Adams and fractal-like Bohart–Adams models was validated by norfloxacin and Cu(II) adsorption in a fixed-bed column. Compared with the Bohart–Adams model, the two new models had better fitting performance with higher R² and lower χ² values, and all of the residuals were randomly distributed. The fractal-like Bohart–Adams and modified dose–response models provided the best fitting quality for the adsorption of Cu(II) (R² = 0.9956 and χ² = 7.56 × 10^–4) and norfloxacin (R² = 0.9991 and χ² = 1.37 × 10^–4), respectively. This work may provide a practical method for the modeling of the asymmetric breakthrough curves.