Facile Removal of Pesticides from Aqueous Solutions Using Magnetic Nanocomposites II. Adsorption experiments and kinetic study

IRINA FIERASCU, STEFAN OVIDIU DIMA, SORIN MARIUS AVRAMESCU, FLORENTIN CALOIAN, RADU CLAUDIU FIERASCU* National Institute for Research & Development in Chemistry and Petrochemistry – ICECHIM Bucharest, 202 Spl. Independentei, 060021, Bucharest, Romania 2 University of Bucharest, Research Center for Environmental Protection and Waste Management, 36-46 Mihail Kogalniceanu Blvd., 050107 Bucharest, Romania Bucharest University of Economic Studies, 6 Piata Romana, 010374 Bucharest, Romania


Analytical methods
Deltamethrin and thiamethoxam pesticides were determined and quantified by liquid chromatography using an Agilent 1260 Infinity series LC system (Agilent Technologies), equipped with a time-of-flight mass spectrometer (TOF-MS) detector, a quaternary pump and autosampler. Elution was achieved by an isocratic program with the mobile phase consisting of a mixture of ultrapure water with 0.1% acetic acid/acetonitrile (95:5) (v/v) at 0.3 mL/min flow rate.
Deltamethrin and thiamethoxam were quantitatively determined using calibration curves built with the concentrations 25, 50, 100, 200 and 400 ppm. For the highest concentration, 400 ppm, it was necessary a 50% dilution in order to be kept in the linearity domain.

Pesticide solutions and adsorption experiments
The aqueous solutions were obtained by dissolving Decis Expert 100 EC ® and Actara 25 WG in distilled water, to obtain stock concentrations of 400 mg/L (active substance). All the other concentrations used in the study were obtained by dilution. For the adsorption experiments, the batch technique was applied. A typical experiment was conducted by weighting 0.02 g of adsorbent materials and mixing with 20 mL of pesticide solution (at a concentration of 200 mg/L) into glass tubes. The test tubes were introduced in a test tube rotating shaker (GFL 3025) for different amount of time in order to evaluate the adsorption process evolution in time. The encoding of samples collected at different time are: T1 -1 hour, T2 -3 hours, T3 -5 hours, T4 -8 hours, T5 -24 hours and T6 -48 hours. In order to preserve the relevancy for real conditions, solution pH was adjusted to 7. At the end of the reaction time, tubes content was filtered through 0.45 µm syringe filters and the remaining pesticide content was evaluated using HPLC. Each experiment was carried out in triplicate, and the average values were used as results. The amount of pesticides retained on the adsorbent at a time t was calculated by the following equation: where: C0 and Ct are the initial concentration and the concentration of the pesticide at the time t in the solution (mg/L), V is the volume of the solution (L), and m is the amount of the adsorbent (g) [11]. For the evaluation of the adsorption equilibrium, 0.02 g of adsorbent and mixed with 20 mL pesticide solution (concentrations ranging from 10 to 200 mg/L) into glass tubes for 10 h in order to reach adsorption equilibrium. Adsorption vials content was filtered through 0.45 μm syringe filters and the filtrate was submitted to analysis. Each experiment was carried out in triplicate, and the average values were used as results. The amount of pesticide retained on solid surface was calculated by equation: where: C0 and Ct are the initial concentration and the concentration of the pesticide at equilibrium (mg/L), V is the volume of the solution (L), and m is the amount of the adsorbent (g) [12].

Results and discussions
According to literature and NIST mass spectra, the main five mass fragments for deltamethrin are 181, 253, 77, 93, and 91. The characteristic peak for deltamethrin was found to be 284, respectively 285 for the protonated fragment M + . For the quantitative analysis it was calculated the cumulative peak area for both m/z fragments, 284 and 285, by mathematical integration within the limits 283.4 -285.7 m/z. Based on known concentrations and calculated peak areas, it was constructed a calibration curve for deltamethrin. A similar approach was used for thiamethoxam solutions. The results (Table 1, graphically represented in Fig. 1) suggest that the adsorption experiments reached an adsorption plateau, after which continuing the adsorption process becomes inefficient.  The differences in adsorption (visible in Fig. 1) can be explained by the reduced affinity of the nitro group in the ionized form from thiamethoxam towards the carboxyl groups from chitosan.
The kinetic parameters are useful for the prediction of adsorption parameters, which are related with efficiency of adsorption. To determine the adsorption rate of thiamethoxam and deltamethrin on the prepared adsorbent three kinetic models such as Lagergren-first-order, pseudo-second-order kinetic models and intra-particle mass transfer diffusion model were applied in order to get insight on the adsorption process dynamics (a more detailed discussion regarding the kinetic models being previously presented [12]). The linear forms of these models were fitted based on the experimental data. Linear regression was used to determine the best fitting kinetic model, and the method of least squares is used for finding the parameters of the kinetic models (Fig. 2). The calculated kinetic constants and related coefficient of determination (R 2 ) are presented in Table 2. From Figure 2 it was observed that the best fit to the experimental kinetic data and closest to the theoretical equilibrium values correspond to first order model. The first order model basically relies on the sorption capacity and the good concordance with experimental data suggesting that the adsorption mechanism might be a physisorption process. Regarding the intraparticle diffusion model, two stages can be observed in the adsorption process of pesticides on chitosan adsorbents. The high value of correlation coefficients for each stage show that intraparticle diffusion model favorably approximates the adsorption process. However, as the plots does not reach origin, it means that intraparticle diffusion is not the only ratelimiting step. The first stage of the plot (Fig. 2) is typically attributed to a fast-external surface adsorption. The second slower part shows a gradual adsorption stage due to intraparticle diffusion larger pesticide molecules and as a consequence of decreasing the concentration gradient.The whole behavior of the solid-pollutant system allows us to assume a combination of intraparticle diffusion and surface adsorption occurred simultaneously. An adsorption isotherm provides a relationship between the solute concentration in the solution and the amount of pollutant adsorbed on the solid phase when the two phases are at equilibrium. The adsorption isotherm experimental data of pesticides were fitted by three common adsorption models: Langmuir, Freundlich and Sips (which is a combination of Langmuir and Freundlich) [12]. The theoretical parameters of adsorption isotherms along with regression coefficients (R 2 ) are presented in Table 3.  The Langmuir parameters were used to assess the adsorption efficiency and affinity between the sorbate and sorbent using the separation factor (RL): where C0 (mg/L) represent the initial phenol concentration and b (L/mg) is the Langmuir constant.
The values of RL at different temperature and initial concentrations for both pollutants fall between 0 and 1 ( Table 4) which demonstrate that adsorption process is favorable especially at higher concentrations. Also, from the values of Freundlich isotherm parameter (1/n) which are below unity regardless the operating conditions it results a favorable adsorption.  Figure 3 shows the fitted equilibrium data for the Sips model which show the highest correlation coefficients (R 2 ) comparing with those of Langmuir and Freundlich models. This occur for both pesticides and for all working temperatures. In addition, the qm value is 53.296 mg/g for thiamethoxam and 247.12 mg/g for deltamethrin. Obviously, the Sips isotherm, compared with the other two models was more suitable to describe the adsorption of pesticides on prepared adsorbents. According to isotherm behavior it can be observed that in case of thiamethoxam adsorption process is exothermic in nature, while in the case of deltamethrin the process is endothermic since the pesticide amount on solid increase with temperature.
When comparing the maximum adsorption capacity determined for the developed material with literature data, it can be observed superior qm values for deltamethrin. Hassan et al. [13] obtained a maximum value of 162.6 mg/g for the adsorption of deltamethrin on potassium hydroxide activated carbons prepared from pistachio nutshells (at 35°C), Al-Qodah et al. [14] obtained a value of 10.96 mg/g (at 25°C) for the adsorption on oil shale ash, while Ghafari et al. [15] registered values up to 41.7 mg/g for the adsorption on modified magnetic iron oxide nanoparticles. Regarding the adsorption of thiamethoxam, the experimental results also shows superior results, by comparison with literature data: Liu et al. [16] obtained an adsorption capacity of approx. 3 mg/g using magnetic graphene oxide-β-cyclodextrin, while Mir et al. [17] obtained a value of approx. 7.43 mg/g for the adsorption on TiO2.

Effect of temperature on the adsorption of pesticides on prepared adsorbent
Adsorption isotherms obtained allow us to assess the effect of temperature on the adsorption process. It was found that higher temperature favored the adsorption of deltamethrin which suggests that the adsorption is endothermic in nature while for the thiamethoxam we observe an opposite behavior. The magnitude of thermodynamic parameters such as the changes in free energy (ΔGo), enthalpy (ΔHo), and entropy (ΔSo) has been calculated with the help of van't Hoff plot using the following formulae: where R is gas constant (J/K mol), T is the absolute temperature (K) and Kd (L/g) is the adsorption equilibrium constant which is calculated by the ratio of amount of adsorbate adsorbed on unit mass of adsorbent to the amount remaining in the solution.
The adsorption equilibrium constant, Kc, for the binding of pesticides to prepared adsorbent surface were calculated from the slope of the qe vs. Ce curve. The Kc values are expressed in L/g. In order to obtain a dimensionless parameter these values must be multiplied by water density (1000 g/L) [18]. Hence, by plotting lnKc versus 1/T the values ΔS 0 and ΔH 0 can be obtained from the intercept and slope (Table 5). It is known that if ΔGo values goes up to -20 KJ gmol -1 , the process is in the range of physical adsorption while higher values of ΔGo fall in the chemical adsorption type of processes.
Moreover, DH 0 and DS 0 were obtained from the slope and intercept of Van-Hoff plots of lnKc versus 1/T (Fig. 4) and the values are presented in Table 5. The negative enthalpy change (ΔH, kJ/mol) confirmed for thiamethoxam adsorption that in this case the removal is exothermic while in the case of deltamethrin the process is endothermic since ΔH have positive values.

Conclusions
Previously presented magnetic composites were presented to be good adsorbents for the targeted pesticides. The Sips isotherm was found to be most adequate to describe the equilibrium process. Moreover, the adsorption experiments revealed that the used materials possess superior adsorption capacity when compared with literature data regarding the up-take of the selected active substances.