Elucidation of Adsorption Mechanisms And Mass Transfer Controlling Resistances During Single And Binary Adsorption of Caffeic And Chlorogenic Acids


 In this work, the potential of activated carbon to remove caffeic and chlorogenic acids was investigated. The study focused on evaluating the single and binary adsorption equilibrium, as well as investigating the mass transfer resistances present during the process by applying kinetic and diffusional models for a future scale-up of the process. For both compounds, the single adsorption equilibrium was studied at pH values of 3, 5, and 7. The experimental adsorption isotherms were interpreted using the Langmuir and Freundlich models, obtaining maximum adsorption capacities of 1.33 and 1.62 mmol/g for caffeic and chlorogenic acid, respectively. It was found that the adsorption mechanisms for both compounds was derived from π-π and electrostatic interactions. Also, the binary adsorption equilibrium was performed and the experimental data were interpreted using the extended multicomponent Langmuir model. The results evidenced that the binary adsorption of caffeic acid and chlorogenic acid is antagonistic in nature. The application of the first and second order kinetic models showed that the latter interpreted better the experimental data, obtaining R2 values close to one. Finally, the experimental adsorption rate data were interpreted by a diffusional model, finding the presence of different mass transfer resistances during the adsorption process. For both compounds, intraparticle diffusion mechanisms were meaningful.

Finally, the surface charge of the GAC was determined using the acid-base titration method 172 established by Kuzin andLoskutov (1996) (Kuzin, I. A. y Loskutov 1996).

175
The concentration decay curves for CA and CGA on GAC were determined using a rotatory   The first order kinetic model is described by Equation 3 (it is solved considering that at the 208 beginning of the adsorption process, no adsorbate exists on the particle). It considers that the 209 adsorption is driven by a linear driving force generated from the difference between the 210 equilibrium capacity and the existing capacity such that the process is interrupted when 211 reaching equilibrium.

213
The second order kinetic model is expressed in Equation 4. The same driving force of the 214 first order model is considered but squared to indicate a faster adsorption process. The same 215 initial condition of the first order model is used. active site is instantaneous, and iv) the GAC particles are spherical, rigid, and isotropic. From these considerations, mass balances of the adsorbate in the bulk solution and within the 233 particle establish Equations 6-9. inside the pores of the particle, and there is continuity of the mass flux at the external surface 243 of the particle (r = R).

244
Usually, the adsorption rate on an active site is assumed to be instantaneous; therefore, The chemical properties of GAC were obtained using an acid-base titration. The 265 concentration of total acid sites was 0.090 meq/g, while that for total basic sites was 0.486 266 meq/g. Therefore, the surface of GAC has a predominantly basic character. The pH value of 267 zero charge, pHPZC, resulted equal to 9.4, which indicates that pH values below 9.4 propitiate 268 a positively charged GAC surface, while the opposite (negatively charged) occurs at pH 269 values above 9.4. This result also corroborates the basic character of the GAC surface.

270
The textural properties were obtained from the experimental data of the adsorption-271 desorption isotherm of N2 at 77 K ( Figure 2). In this isotherm, a type I-B behavior is observed, which is characteristic of microporous materials (a great amount of N2 is adsorbed at low 273 pressures). Moreover, a slight type H4 hysteresis loop can be observed, which is 274 representative of solids with pores with narrow slit. The experimental data heled determining 275 that the total pore volume for the material studied was 0.36 cm 3 g -1 (for a P/P0 value of 0.95),

Individual and binary adsorption at equilibrium 283
The adsorption equilibrium for both acids was studied at pH values of 3, 5, and 7 (at 25 °C), 284 at these pH values studied the adsorption capacity is expected to be affected through changes 285 not only in the surface charge of the adsorbent but the speciation of the acids studied.

286
Additionally, experiments at pH values above 7 were not carried out due to stability issues 315 Therefore, the electrostatic forces have influence on the adsorption over the positively 316 charged adsorbent surface; nonetheless, they are not synergistic with the π-π interactions to 317 the same level found with CA, promoting less surface saturation.
The experimental adsorption data at different pH values were described by the 319 Freundlich and Langmuir isotherm models represented by Equations 11 and 12, respectively.

320
The optimized values of the fitting parameters were determined using nonlinear regression 321 using Statistica®, Table 2 shows the results for each model. The R 2 values establish that the .
The EML model assumes that the adsorption sites on the adsorbent surface are 331 uniform; therefore, both adsorbates compete for the same adsorption sites that are 332 energetically equal. Table 3 shows the values of the optimized parameters from the model, 333 qmax reaches a value of 1.48 mmol g -1 , while the values of KEI demonstrate that the GAC 334 surface has higher (almost 2-fold) affinity towards CGA when compared to CA. The prediction of the EML model is presented in Figure 3 for both compounds where good fitting 336 to the experimental data is observed.

337
The effect of CGA presence on the adsorption capacity of CA is presented in Figure   338 5a. In this figure, the adsorption capacity of CA is drastically reduced at CGA equilibrium 339 concentrations less than 0.2 meq L -1 , whereas at higher concentrations the effect is less In Equation 14, the term in square brackets is the slope of the kinetic curve at time  As an example, Figure 9 shows the experimental data for Exp. 2 for CA and Exp. 4 Furthermore, the magnitude of each flux can be estimated using Equations 20 and 21.