(419f) Modeling Adsorption Kinetics of Tetracycline from Aqueous Solution ONTO Natural Bentonite Nanoclay. Elucidating Intraparticle Diffusion Mechanisms

The presence of a great variety of organic compounds from anthropogenic sources in surface and wastewater has attracted great concern worldwide. Pharmaceutical compounds and, in particular, antibiotics, are considered emerging pollutants and have been widely detected in surface, groundwater, drinking water and effluents from wastewater treatment plants. The occurrence of these emerging pollutants can cause that the microorganisms become resistant to antibiotics and pose a potential threat to human health.

Tetracycline (TC) is a broad-spectrum bacteriostatic antibiotic frequently used in veterinary medicine. TC has been detected in surface waters (0.11-4.2 μg/L) and effluents from wastewater treatment plants (46-1300 μg//L), and can produce severe toxic effects to human and animal health. Various processes have been implemented to remove these compounds from water systems; however, most of these processes have been shown to generate recalcitrant and highly toxic by-products. Adsorption is a very promising alternative method for treating polluted aqueous effluents. Several materials can be applied as adsorbents, from advanced nanostructured materials to natural nanometric minerals.

Bentonite (Bnt) is a natural, laminar and expandable nano clay with a high content of montmorillonite phase and belongs to the group of smectites. Its trilaminar structure (2:1) consists of an octahedral Al³⁺ layer arranged between two tetrahedral Si⁴⁺ layers. The charge deficiency is generated by isomorphic substitutions in the octahedral and tetrahedral sheets and is balanced by the exchangeable cations, mainly Na⁺ and Ca²⁺.

The design of adsorption fixed-bed columns requires information about the adsorption capacity and adsorption rate. The adsorption rate depends on the following stages: i) External mass transport, ii) Intraparticular diffusion and iii) Adsorption at an active site. The intraparticular diffusion may be due to pore volume diffusion, surface diffusion, or a combination of both mechanisms. The adsorption rate of pharmaceutical compounds on clays has been analyzed usually by simple kinetic models, which neglect the porous structure of clays. The application of diffusional models to interpret the adsorption has not been reported previously, and no information is available about the diffusion mechanism controlling the overall adsorption rate of TC on Bnt.

This work was aimed at studying and modeling the adsorption rate of TC on Bnt under different experimental conditions and elucidating the intraparticular diffusion mechanisms controlling the overall adsorption rate.

The TC used in this work was of analytical grade. The calcium Bnt nanoclay was from a mineral deposit located in San Luis Potosí, Mexico. The experimental adsorption equilibrium data of TC and potassium cations (K⁺) onto Bnt were determined in a batch adsorber consisting of a conical vial (50 mL). A volume of 40 mL of a TC or K⁺ solution with known initial concentration was poured into a batch adsorber containing a given mass of Bnt. The clay and the solutes in the solution were allowed to attain equilibrium in about 24 hours. The solution was sampled and the equilibrium concentration of TC or K⁺ was quantified by UV-Visible spectroscopic and atomic absorption spectroscopy. The mass of TC or K⁺ adsorbed on Bnt was calculated by a mass balance.

The experimental data for the adsorption kinetics of TC or K⁺ on Bnt were obtained in a stirred tank batch (STB) adsorber, and the experimental method for procuring the concentration decay curves is briefly described in the following. A specific mass of Bnt particles was contacted in the adsorber with a volume of 0.01 N HCl/NaOH buffer solution, and the suspension was continually mixed by a propeller agitator to ensure suspending the particles. Next, an aliquot of a TC or K⁺ solution was added to bring the total volume to 200 mL and attain a specific initial concentration of TC or K⁺, and then the timer was activated (t = 0). The sampling of the adsorber solution (1 or 2 mL) was conducted periodically, and each sample was later analyzed, as described earlier.

The adsorption equilibrium data of TC and K⁺ on Bnt was interpreted using the Freundlich, Langmuir and Prausnitz-Radke isotherms and the parameters of the adsorption isotherms were evaluated by a least-squares method. The Prausnitz-Radke isotherms fitted reasonaby well the experimental equilibrium data.

Diffusional models interpreted the experimental concentration decay data of the TC or K⁺ for the adsorption on Bnt. The diffusional models were derived assuming the following: 1) Bnt particles are spherical, 2) a mass transport coefficient represents the external mass transfer, 3) intraparticular diffusion is due to pore volume diffusion and surface diffusion, 4) the effective pore volume diffudion coefficient (D_ep) can be estimated using the tortuosity factor (Ï„_p), and 5) adsorption rate on an active site is instantaneous. The pore volume diffusional model (PVDM) and the general diffusional model (PVSDM) were particular cases of the diffusional models applied in this work. In the PVDM is assumed that intraparticle diffusion is solely due to pore volume diffusion, and in the PVSDM, both pore volume and surface diffusion are co-occurring.

The adsorption equilibrium data and the concentration decay data of K⁺ were evaluated to determine the tortuosity factor (Ï„_p) of the Bnt. The PVDM model was fitted to the K⁺ concentration decay curves, and the values of the external mass transfer coefficient and the effective diffusion coefficient (D_ep) were needed for solving the PVDM model. The external mass transport coefficient was computed from the slope of the concentration decay data points at t = 0 and t = 1 min, and the value of D_ep was obtained by matching the numerical solution of the PVDM model to the experimental data. The optimal value of D_ep was the value that best fitted the experimental data. The concentration decay curves of K⁺ at different conditions are depicted in Figure 1a, and the adsorption equilibrium was reached in less than 20 min. As seen in this figure, the PVDM model interpreted the experimental data satisfactorily, and Ï„_p varied between 4 and 5, and the average Ï„_p,Av was 4.57. The PVDM with D_ep = 7.3×10^-7cm²/s and Ï„_p,Av = 4.57, predicted the experimental data quite well with average percentage deviations (%D) between 0.31 and 2.18 %.

The experimental concentration decay data of TC is shown in Figure 1b, and equilibrium was attained at 240 min. The experimental data were interpreted by the PVDM model (D_ep ≠ 0) and assuming no surface diffusion (D_s = 0). The results showed that the PVDM model with Ï„_p,Av = 4.57 and D_ep = 1.53×10^-7 cm²/s did not interpret the experimental data, predicting that the adsorption rate was much slower than that of the experimental one. As seen in Figure 1b, the concentration decay predicted with PVDM was above that of the experimental one, confirming that the surface diffusion was also taking place. The PVSDM model well interpreted the concentration decay data of TC using D_ep = 1.53×10^-7 cm²/s and D_s = 2.11×10^-10 cm²/s). The optimal value of D_swas estimated by matching the numerical solution of the PVSDM model to the experimental concentration decay of TC. Hence, surface diffusion is a significant intraparticle diffusion of TC during the adsorption on Bnt. The contribution of surface diffusion represented more than 70% of the total intraparticle diffusion, revealing that surface diffusion controlled the overall adsorption rate of TC on Bnt.