Ali Fakhri^{*} and Saeideh Adami  
Department of Chemistry, ShahreQods Branch, Islamic Azad University, Tehran, Iran  
Corresponding Author :  Ali Fakhri Department of Chemistry, ShahreQods Branch Islamic Azad University, Tehran, Iran Tel: +98(21)22873079 Fax: +98(21)22873079 Email: [email protected] 

Received April 18, 2013; Accepted June 12, 2013; Published June 14, 2013  
Citation: Fakhri A, Adami S (2013) Response Surface Methodology for Adsorption of Fluoride Ion Using Nanoparticle of Zero Valent Iron from Aqueous Solution. J Chem Eng Process Technol 4:161 doi: 10.4172/21577048.1000161  
Copyright: © 2013 Fakhri A, et al. This is an openaccess article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.  
Related article at Pubmed Scholar Google 
Visit for more related articles at Journal of Chemical Engineering & Process Technology
This study is removal of fluoride ion from aqueous solution by nano zero valent iron (nZVI). Effects of the factor variables (temperature, nZVI dose and pH) and their interactions on adsorption of fluoride ion were investigated by response surface methodology (RSM) based on BoxBehnken design (BBD). Optimized values of temperature, nZVI dose and pH for fluoride sorption were found as 313K, 0.5 g, and 4, respectively. The effect of initial fluoride concentration on the adsorption amount was investigated by a batch experiment. For study the fluoride removal mechanism, various adsorption isotherms such as Langmuir, Freundlich and Temkin were fitted. The results showed that the Freundlich isotherm gave high fit for fluoride adsorption. The time of adsorption reaction was rapid and subordinated pseudosecondorder kinetics. The values of thermodynamic parameters indicated that adsorption was spontaneous and endothermic in nature. The nano zero valent iron (nZVI) could be used as a potential adsorbent for fluoride ion containing aqueous solution.
Keywords  
Adsorption; Fluoride; BoxBehnken design; Nanoparticle zero valent iron; Response surface methodology  
Introduction  
Fluoride is a health affecting substance. The physiologic effects of fluoride ingestion on public health have been studied extensively [1]. The acceptable fluoride concentration in drinking water is generally in the range of 0.5 to 1.5 mgl^{1} [2]. High concentration has the effects on the metabolism of elements such as Ca, P in human body and lead to dental and skeletal fluorosis. The fluoride content of soils varies from under 20 to several thousand ppm, the higher records being mostly from areas with bedded phosphate on fluoride deposits [3]. The natural presence of fluoride generally occurs through soil and rock formation in the form of fluorapatite, fluorspar and amphiboles, geochemical deposits, natural water systems and earth crust [4,5]. In addition to this fluoride can also be found in various Industrial work, chiefly semiconductor, electroplating, glass, steel, ceramic and fertilizers industries [6]. Because of these reasons the pollution of ground water by fluoride contamination has been a major concern. The problems in connection with fluoride ion pollution could be reduced or minimized by ultrafiltration, precipitation, reverse osmosis, electrodedeposition, etc., but these processes have flaws such as high cost, generation of secondary pollutants and low removal efficiency. Adsorption experiment has been found to be an effective and economic method with high potential for the removal, recovery and recycle of fluoride ions from aqueous solution [7], although desorption is an issue.  
Zero valent iron (ZVI) was proposed as a reactive material in permeable reactive barriers (PRGs) due to its great ability in reducing and the stabilization pollutants different [8]. Nowadays, there has been an increasing interest in synthesizing this material on nanoscale for to enhance its ability restore by probity of the increase in the surface area and surface reactivity of the particles [9]. Nano zero valent iron (nZVI), a recently discovered technology, is being used to successfully treatment various organic compounds (e.g. [1012]) and various heavy metal ions in aqueous solutions (e.g. [1318]).  
Design of experiments (DOE) and response surface methodology (RSM) is largely used for modeling mechanism parameters, especially in adsorption or removal process [1932]. Because RSM contains a lower number of experiments, it is flaws over conventional methods available. It is suitable for multifactor experiments and searches the common connection between various factors for the determined of most favorable or unfavorable conditions of the processes. Response surface methodology has different model types such as central composite design (CCD), Doehlert matrix (DM) and BoxBehnken design (BBD). The objective of this study was the application of the RSM combined with BoxBehnken design as a statistic method in optimizing adsorption mechanism of fluoride ion using nano zero valent iron. The Langmuir, Freundlich and Temkin isotherm models were used to define the equilibrium data. The adsorption mechanisms of fluoride ion from aqueous solutions onto nZVI were also evaluated in terms of kinetics and thermodynamic parameters.  
Materials and Methods  
Raw materials  
Sodium fluoride salt (NaF) (molecular weight, 41.98871 g/mol) was supplied by Merck Co. (Germany) (maximum purity available > 99%). Doubly distilled deionized water (HPLC grade 99.99% purity) was obtained from Sigma Aldrich Co. (Germany).  
Synthesis of nanoparticle zero valent iron (nZVI)  
The nZVI material was synthesized by drop wise addition of 1.6 M NaBH_{4} (sodium borohydride) aqueous solution to a Ne gaspurged 1 M FeCl_{3}.6H_{2}O (Ferric chlorid hexahydrate) aqueous solution at 23°C with magnetic stirring as described by Wang and Zhang [33]. Ferric iron (Fe^{3+}) was reduced according to the reaction [34]: 4Fe^{+3}+ 3BH_{4} + 9H_{2}O → 4Fe^{0} + 3H_{2}BO_{3}^{} + 12H^{+} + 6H_{2}.  
The solution was stirred for 20 min and centrifuged at 6000 rpm for 2 min, and the supernatant solution was replaced by acetone. Acetonewashing prevented the immediate rusting of nZVI during purification leading to a fine black powder product after freezedrying. Xray diffractometer (XRD, Philips X’Pert) was used to investigate the material structure of zero valent iron nanoparticles. In addition, the surface morphology of nZVI was determined using transmission electron microscopy (TEM, JEM2100F HR, 200 kV).  
Adsorption experiment  
The adsorption of fluoride onto nZVI was investigated using batch experiments. In these studies 1000 mg/L stock solution was prepared by dissolving 1 g of NaF in 1000 mL distilled water. Different concentrations (20200 mg/L) of fluoride solutions were prepared by this stock solution. Solutions were evacuated to flasks of 100 ml. Then adsorbent in the range of dosage 0.050.5 g was added and placed in the water bath shaker after pH adjustments made in the range of 210. The suspensions were shaken at 2000 rpm for 12 min at room temperature. Samples from shaker were filtered with filter paper, and then remaining fluoride levels were measured using a fluoride electrode (Orion, 9606BNWP). The final adsorption amount was calculated from the equation  
(1)  
where, q_{e} (mg/g) is the equilibrium adsorption amount, C_{e} is the fluoride concentration at equilibrium (mg/l), V is the volume of solution (l) and w is the weight of adsorbent (g).  
Response surface methodology  
The threelevel, threefactorial BoxBehnken experimental design with categorical factor of 0 was engaged to optimize the treatment process based on the adsorption amount of the nZVI for fluoride ion. The design was composed of three levels (low, medium and high, being coded as (1, 0 and +1) and a total of 17 runs were carried out in repetitious to optimize the level of chosen variables, such as temperature, nZVI dosage and pH. For the aim of statistical calculations, the three factor variables were denoted as X_{1}, X_{2}, and X_{3}, respectively. According to the preparatory experiments, the range and levels used in the experiments are selected and listed in Table 1. The main effects and in connection between factors were determined. Figure 1 illustrates the mean of the experimental results for the respective low, medium and high levels of temperature, pH and nZVI dosage of scattering. The experimental design matrix by the BoxBehnken design is tabulated in table 2 and corresponding experiments were performed. The results were analyzed by applying the response plots and analysis of variance (ANOVA). For RSM, the most commonly used secondorder polynomial equation developed to fit the experimental data can be written as:  
(2)  
where Y represents the predicted response, i.e. the adsorption amount for fluoride ion by the nZVI (mg/g), β_{0}, the constant coefficient, β_{i}, the i^{th} linear coefficient of the input factor x_{i}, β_{ii}, the i^{th} quadratic coefficient of the input factors x_{i}, β_{i,j}, the different interaction coefficients between input factors x_{i} and x_{j} (i = 1–3, j = 1–3 and i / = j), and ε, the error of the model [35]. The equation expresses the in connection between the predicted response and factor variables in coded values according to Tables 1 and 2.  
Results and Discussion  
Characterization of nZVI  
Figure 2 shows the transmission electron microscope image of freshly synthesized iron nanoparticles. Surface morphology shows that there exist two layers in the nZVI particle. The layer of intrant core Indicative the Fe^{0}, and the external layer surrounding on the Fe^{0} was ironoxide(s). Figure 2B shows that the fluoride molecules into the nZVI surface are covered. The XRay diffraction of nZVI surface composition under ambient conditions is shown in Figure 3. The wide peak reveals the being of an amorphous phase of iron. The characteristic broad peak at 2θ of 45° indicates that the zero valent iron is predominantly present in the sample.  
Statistical analysis  
The optimum terms for adsorption of fluoride onto nZVI surface were determined by means of the BBD under RSM. The results were displayed in Tables 3. In this manner, the fluoride uptake by nZVI could be described using the following equation: Y = 22.474 + 2.476 X_{1} + 1.241 X_{2} – 2.445 X_{3} – 2.650 X_{1}X_{2} + 0.147 X_{1}X_{3} – 2.472 X_{2}X_{3} + 9.095 X_{1}^{2} – 0.879 X_{2}^{2} + 6.873 X_{2}^{3} (3)  
The state of the fitted model was declared by the coefficient of determination. The R^{2} coefficient gives the relation of the total variation in the response predicted by the model and a high R^{2} value (close to 1) is desirable. Eq. (3) Indication that the model is well fitted, considering the determination coefficient (R^{2}= 97.71%). The estimated effects and coefficients for model are listed in Table 3. Model terms were evaluated by the Pvalue (probability) with 95% confidence level. The Pvalues were used to estimate whether F was large enough to indicate statistical significance and used to check the significance of each coefficient. The Pvalues lower than 0.05 indicated that the model and model terms were statistically significant. All the factors and their square interactions (P < 0.05) except for interaction of temperature–temperature (X_{1}^{2}) and pHpH (X_{3}^{2}) were significant at the 95% confidence level. nZVI dose was the most significant factor that affect the removal of fluoride. Also, the quadratic effect of nZVI dose  nZVI dose (X_{2}^{2}) was found larger than effect of nZVI dose, and the removal of fluoride significantly decreased. Figure 4 shows that the data were well distributed near to a straight line (R^{2}=0.9726), which proposed a good relationship between the experimental and predicted values of the response, and the underlying assumptions of the above analysis were appropriate.  
3D response surface plot  
The 3D response surface plot are useful in investigation both the main and interaction effects of the factors [36,37]. The response surface plots are presented in Figure 5. This figure also shows the estimated Y parameter as a function of the normalized factor variables, the height of the surface represents the value of Y. After executing a screening of factors using a BBD, the surface plots of the response (Y) indicated the same results as observed in the interaction plot (Figure 5).  
Adsorption isotherms  
The mechanism of fluoride adsorption from aqueous solutions onto nZVI is evaluated using adsorption isotherm. In this study, isotherm data were applied to four adsorption models and the results of their linear regressions were used to find the model with the best fit. Values of resulting parameters and regression coefficients (R^{2}) are listed in Tables 4 and 5.  
The R^{2} value for the Freundlich isotherm was 0.9990, which is higher than the values obtained from the Langmuir and Temkin isotherm models. The experimental data fit very well to this isotherm model, and indicates that fluoride adsorption occurs on heterogeneous surfaces, which is similar to the conclusion reached for nZVI [3840].  
Adsorption kinetics  
Various models have been used to investigation of the adsorption mechanisms and potential rate. Effects of contact time on adsorption are investigated, as shown in Figure 6. The adsorption process was quite rapid and reached equilibrium in 30 min (Figure 6). In this study, the different kinetics models such as pseudofirstorder, pseudosecondorder, and intraparticle diffusion models were used.  
The pseudofirstorder rate equation is given as [41]:  
(4)  
where q_{e} and q_{t} are the amounts of fluoride adsorbed (mg/g) at equilibrium and at time t (min), respectively, and k_{1}(L/min) is the adsorption rate constant of firstorder adsorption. A straight line of log(q_{e}q_{t}) versus t (Figure not shown) suggests the applicability of this kinetic model. q_{e} and k_{1} were determined from the intercept and slope of the plot which were shown in Table 6. From the data, q_{e} (calculated) and q_{e} (experimental) values are not in agreement with each other. Therefore, that indicates the adsorption of fluoride on nZVI was not a firstorder reaction.  
In addition, the experimental data was also applied to the pseudosecond order kinetic model Equation [42]:  
(5)  
where k_{2} is the rate constant of pseudosecondorder chemisorptions (g/(mg min)). The plot t/q_{t} versus t giving a straight line which is shown in Figure 7 and the constant calculated from the slop and intercept of the plots are given in Table 6. Figure 7 shows that R^{2} values are higher than those obtained from the firstorder kinetics. In addition, theoretical and experimental q_{e} values are in agreement. Therefore, it is possible to prove that the adsorption process using nZVI followed the secondorder kinetic model.  
The intraparticle diffusion equation can be described as:  
q = k_{i} t^{1/2} + C (6)  
where k_{i} is the intraparticle diffusion rate constant (mg/g min). The data (Figure not shown) shown have an initial curved portion, followed by a linear portion. The curved portion of the plot is due to the diffusion of fluoride through the solution to the external surface of nZVI, or boundary layer diffusion. However, the extrapolated linear regions at different initial concentrations did not pass through the origin and that suggests that the intraparticle diffusion was not the ratecontrolling step [43].  
Thermodynamic of adsorption studies  
Thermodynamic parameters connected to the adsorption reaction, i.e., Gibbs free energy change (ΔG°, kJ mol^{1}), enthalpy change (ΔH°, kJ mol^{1}), and entropy change (ΔS°, J mol^{1} K^{1}) are defined by the following equations: ΔG° = RTlnK_{c (7)}  
ΔG°= ΔH° = ΔH° – TΔS° (8)  
where K_{c} is the equilibrium constant, which can be obtained from Langmuir isotherm, R is the universal gas constant, 8.314 Jmol^{1} K^{1}, and T is absolute temperature (K). ΔH° and ΔS° were obtained from the slope and intercept of the plot of Gibbs free energy change, ΔGº vs. temperature, T (Figure 8).  
The negative values of ΔG° (5.746, 6.793 and 8.270 KJ mol^{1} for 283, 297 and 313K, respectively) confirm the feasibility of the process and the spontaneous nature of adsorption with a high preference for fluoride onto nZVI. The standard enthalpy and entropy changes determined from plot in Figure 8. The value of ΔH° is positive (18.165 KJ mol^{1}), indicating that the adsorption reaction is endothermic. The positive value of ΔS° (84.300 J mol^{1}K^{1}) reflects an increase in the randomness at the solid/solution interface during the adsorption process [44].  
Conclusions  
The statistical design of the experiments was applied in optimizing the conditions of maximum adsorption of the fluoride onto nZVI. The result data from ANOVA demonstrates that the model was highly significant. nZVI dose was the most significant factor affecting fluoride removal. Therefore, it is apparent that the response surface methodology not only gives valuable information on interactions between the factors but also helps to the recognition of possible optimum values of the studied factors. Adsorption experiments show that the adsorption equilibrium can be achieved within 30 min. The kinetics studies of fluoride on nZVI indicated that the adsorption kinetics of fluoride on nZVI followed the pseudosecond order at different initial concentration. The results of Isotherm data showed that the removal of fluoride followed Freundlich isotherm. Thermodynamic adsorption studies indicated that the adsorption fluoride using nZVI aqueous solution was a spontaneous, endothermic.  
Acknowledgements  
The authors gratefully acknowledge supporting of this research by the Islamic Azad University ShahreGods Branch.  
References  

Table 1  Table 2  Table 3  Table 4  Table 5  Table 6 
Figure 1  Figure 2  Figure 3  Figure 4  
Figure 5  Figure 6  Figure 7  Figure 8 