Modeling arsenic(V) removal in iron oxide impregnated activated carbon columns

Iron oxide was impregnated onto an activated carbon, creating an adsorbent (FeAC) with a high arsenic removal capacity that can be used in existing activated carbon column systems. Objectives of this research were to: (1) characterize the FeAC surface, (2) model the FeAC surface acidity and determine the most appropriate acid-base surface site representation, (3) determine if the surface complex formation (SCF) model could describe As(V) removal at equilibrium, and (4) determine if the SCF model coupled with the homogeneous surface diffusion mode (HSDM) could describe As(V) removal in a fixed-bed adsorber. The FeAC surface was characterized by measurement of the surface area, scanning electron microscopy (SEM), electron dispersive spectroscopy (EDS), X-ray diffraction (XRD), and differential thermal analysis (DTA). Potentiometric titrations of FeAC were performed at three ionic strengths (I). Equilibrium adsorption experiments were carried out at different FeAC concentrations, rapid small-scale column tests (RSSCT) were conducted at various empty bed contact times (EBCT), utilizing a scaling factor of 9.82. Based on SEM images and confirmed by EDS analysis, impregnation resulted in pore plugging of the FeAC by Fe oxides. XRD and DTA analyses revealed that the species at the carbon surface were a mixture of Fe oxides with hematite (α-Fe₂O ₃) predominating. The diprotic and two-monoprotic surface site representations coupled with either the diffuse layer or triple layer EDL models adequately fit the data. The triple layer EDL model (TLM) predicted titration data better compared with the diffuse layer model for I = 1.94 × 10−3 and 10−1. The diprotic representation used in conjunction with the TLM was chosen to represent the FeAC surface acidity. Using the diprotic-TLM in conjunction with two As(V)-FeAC surface reactions could not describe adsorption over the entire pH range examined in equilibrium experiments; however, the inclusion of three As(V)-FeAC surface reactions with the diprotic-TLM (3-RXN model) resulted in a significant improvement of equilibrium model predictions. Effluent As(V) breakthrough (10 μg/L) was incipient for the 0.2-minute EBCT experiment (scaled EBCT = 2.0 minutes), whereas ∼2300 bed volumes (BV) of water contaminated with 1 mg/L As(V) were treated at an EBCT of 2.1 minutes (scaled EBCT = 21 minutes) before As(V) breakthrough was detected. The tailing nature of the As(V) breakthrough curve was caused by intraparticle diffusion and an increase removal capacity as the adsorption zone pH decreased from 7 to ∼4.5. The 3-RXN model coupled with the HSDM was useful for evaluating mass transfer coefficients and predicted As(V) removal accurately in the RSCCT. Model simulations were performed to determine the effect of EBCT on removal as well as column performance at lower influent As(V) concentrations. FeAC appears to be an effective adsorbent for the removal of As(V) in continuous flow systems and the coupled HSDM-SCF model can accurately predict As(V) breakthrough in fixed-bed adsorbers.