The main purpose of this work was to determine adsorption characteristics of fluoride, nitrate and sulfate ions on the AFN membrane. The sorption isotherms for , and ions on the AFN membrane were investigated in the range of 0.05–1 mol.L–1 at 298 K. The suitability of the Langmuir, Dubinin–Astakhov (D-A) and Redlich–Peterson adsorption models to the equilibrium data was investigated. The sorption parameters of the studied models were determined by non-linear regression and discussed. Equilibrium data obtained in this study were found to best fit the Langmuir isotherm. The ΔG° values deduced from the Langmuir isotherm suggest that the affinity order of the AFN membrane for the studied anions is: . In order to improve their selectivity towards monovalent ions, the AFN membrane was modified by adsorption of polyethyleneimine on its surface. Adsorption parameters values of the studied models were determined for the modified AFN membrane. The D-A model provides the best fit to the experimental points. In fact, removal of fluoride and nitrate ions by adsorption on the modified AFN membrane was more effective than the adsorption on the unmodified one. Desorption of fluoride from the modified AFN membrane by nitrate and sulfate ions was investigated.

Symbols

     
  • sorption capacity of the membrane at equilibrium (mmol g–1)

  •  
  • initial anions concentrations (mmol L–1)

  •  
  • concentrations at equilibrium (mmol L–1)

  •  
  • V

    volume of the solution (L)

  •  
  • m

    mass adsorbent (g)

  •  
  • maximum adsorption capacity (mmol g–1)

  •  
  • Langmuir constant (L mol–1)

  •  
  • E

    adsorption energy (kJ mol–1)

  •  
  • nD

    homogeneity parameter

  •  
  • ε

    Polanyi potential

  •  
  • βRP

    degree of heterogeneity

  •  
  • A and B

    Redlich–Peterson parameters

  •  
  • T

    absolute temperature (K)

  •  
  • R

    universal gas constant (J mol–1 K–1)

  •  
  • R2

    correlation coefficient

  •  
  • χ2

    chi-square test

  •  
  • ΔG°

    Gibbs free energy (kJ mol–1)

INTRODUCTION

Various anthropogenic and natural substances can be found dissolved in water. Some are harmful but others are beneficial. When it comes to humans, the latter category must be at a specific dose. In terms of harmful substances or contaminants, fluoride, nitrate and sulfate in water have been recognized as serious problems worldwide (Chabani et al. 2009; Dron & Dodi 2011; Ranjana 2012; Samadi et al. 2014). Several processes were used to remove these contaminants such as ion exchange, electrodialysis, reverse osmosis, nanofiltration, microbiological denitrification, chemical treatment, biological treatment, adsorption, etc. (Tehobanoglous 1979; Du Preez et al. 1992; Schoeman & Steyn 2001; Bae et al. 2002; Sehn 2008; Haghsheno et al. 2009; Solangi et al. 2009; Dolar et al. 2011). Among them the adsorption process is simpler and more effective for fluoride, nitrate and sulfate removal. Adsorption on ion exchangers is considered to be highly stable and hence is considered as one of the most promising adsorbents (Haghsheno et al. 2009; Sachin et al. 2011; Samadi et al. 2014).

Extensive investigations (Chia-Hung et al. 2007; Meenakshi & Viswanathan 2007; Jeng-Shiou et al. 2008; Xu et al. 2012; Abdelwahab et al. 2013; Hannachi et al. 2014) were carried out to study the adsorption of contaminant ions onto ion exchange resins and membranes. Investigators found that removal of ions using resins and membranes has a good potential for water treatment. In fact, Chabani et al. (2006) studied the adsorption of nitrate on Amberlite IRA 400 resin. They found that this resin is an effective adsorbent for the removal of nitrate from aqueous solution.

Recently, Sachin et al. (2011) investigated the removal of nitrate by adsorption onto anion exchange Indion NSSR resin. Freundlich, Langmuir and Dubinin–Radushkevich models were used to fit the equilibrium isotherms. The results showed that nitrate removal had followed the Langmuir adsorption isotherm and that Indion NSSR resin is effective for nitrate removal.

Removal of sulfate anions from Sarcheshmeh copper complex wastewater by adsorption on Lewatit K 6362 anion exchange resin has been studied by Haghsheno et al. (2009) in order to determine the adsorption equilibrium data and adsorption isotherms. Their results showed that isothermal data best fit a Freundlich adsorption isotherm and the adsorption of sulfate ions on the Lewatit K 6362 ion exchange resin followed reversible first-order kinetics. Samadi et al. (2014) investigated equilibrium and kinetic parameters for the removal of fluoride ions from water by adsorption on anion exchange resin. They found that the removal of fluoride was high at natural pH and was improved for increasing contact time and adsorbent dosage.

Recently, Hannachi et al. (2013, 2014) investigated adsorption of fluoride, nitrate and sulfate ions on AFN and AMX anion exchange membranes at various temperatures. They showed that the adsorption linear pattern on the AMX membrane followed Langmuir isotherms. It was also found that at 298 K the AMX membrane has more affinity for sulfate than nitrate than fluoride, while for the AFN membrane the affinity order is .

On the other hand, studies on separation using ion exchange membranes have resulted in difficult separations. One of the unsolved problems of ion exchange membrane is the separation of ions with the same charge and different valences. Several studies have shown that the efficiency of separation of ions from water using ion exchange membranes is dependent of the selectivity of these membranes for monovalent ions against multivalent ones. In this context, modification of ion exchange membranes was carried out showing the improvement of the selectivity (Sata & Mizutani 1979). Several studies (Cortina et al. 1997; Biesuz et al. 2001; Serarols et al. 2001; Amara & Kerdjoudj 2003, 2007; Gohil et al. 2006) have been realized to modify the properties of ion exchange resins and membranes in order to improve their selectivity towards ions which are different in size and valence.

Therefore, in the present investigation, attempts were made to determine adsorption characteristics of fluoride, nitrate and sulfate ions on the unmodified and the modified AFN membranes. The latter was modified by adsorption of polyethyleneimine (PEI) on its surface. The adsorbed amount of PEI on membrane surface was determined as a function of the following parameters: initial PEI concentration, pH and molecular weight of the PEI. To investigate the influence of these parameters an experimental design methodology by means of a full factorial design (23) was adopted and the different data were analyzed with MINITAB software.

The adsorption of the ions was investigated through the most commonly used single-component adsorption isotherms: Langmuir, Dubinin–Astakhov (D-A) and Redlich–Peterson (R-P). The adsorption parameters of the experimental models were determined from the non-linear regression.

EXPERIMENTAL PROCEDURE

Membrane characteristics

The commercial anion exchange membrane evaluated in this investigation is the AFN membrane provided by Tokuyama Soda. The base polymer of this membrane is styrene and divinylbenzene, and the ionic fixed sites are quaternary ammonium groups, rendering the membrane a strong anion exchanger. The main characteristic of the AFN are presented in Table 1.

Table 1

Main characteristics of the AFN membrane

Ion exchangerAFN
Ion exchange capacity (meq.g–12–3.5 
Humidity percentage in Cl form (%) 45.35 
Thickness (mm) 0.15–0.2 
Resistance (Ω.cm20.4–1.5 
Membrane density (g.cm31.09 
Transport number of Cl 0.92 
Ion exchangerAFN
Ion exchange capacity (meq.g–12–3.5 
Humidity percentage in Cl form (%) 45.35 
Thickness (mm) 0.15–0.2 
Resistance (Ω.cm20.4–1.5 
Membrane density (g.cm31.09 
Transport number of Cl 0.92 

Adsorption experiments

To obtain adsorption isotherms of , and , different samples of the AFN membrane (25 cm2) in form were immersed in 200 mL of fluoride, nitrate or sulfate ions solutions at different initial concentrations (from 0.05 to 1 mol.L–1). All experiments were carried out at 298 K. The solution and membrane were then shaken at 150 rpm in a thermostatic bath until equilibrium was achieved.

The nitrate, fluoride, and sulfate ions concentration in the solution were determined, at equilibrium, by ion chromatography using a Metrohm 761 compact IC with conductivity detector and chemical suppression, 4.6 × 250 mm Metrosep A Supp 1 (6.1005.300), an eluent of 3 mmol.L–1 sodium carbonate at 1 mL.min–1, an injection volume of 20 μL, and a pressure of 8–9 MPa.

The sorption capacity (mmol.g–1) of the membrane at equilibrium was given by the following equation: 
formula
where (mmol.L–1) and (mmol.L–1) are the initial anion concentration and the concentration at equilibrium, respectively, V is the volume of solution and m is the mass adsorbent.

Surface modification

The AFN membrane surface was modified by adsorption of PEI which was obtained from Sigma-Aldrich. The PEI used in this investigation is a branched chain polymer that contains primary, secondary and tertiary amine groups in a ratio of approximately 1/4, 1/2 and 1/4, respectively.

The methodology used to modify the AFN membrane has been described extensively in our previous studies (Guesmi et al. 2012, 2013). The AFN membrane (25 cm2) was immersed in 250 mL of a stirred PEI aqueous solution of a given concentration at 25 °C for 6 days. At the equilibrium, the membrane was rinsed with ultrapure water to eliminate the excess of PEI. Several parameters affecting the modification of ion exchange membrane and the initial concentration of PEI, the molecular weight of PEI and the pH of the solutions were evaluated in this investigation.

The polyethyleneimine concentration was determined using the UV absorption spectrum of copper-PEI complex. The solution containing PEI was mixed with a cupric sulfate solution and then analyzed by UV absorption spectrum. The maximum absorption of the copper-PEI complex was located at 326 μm.

RESULTS AND DISCUSSION

Adsorption isotherms onto unmodified AFN membrane

Adsorption isotherms establish the relationship between the amount of anion adsorbed per gram of membrane, qe (mmol.g–1), and the amount of anions left in equilibrium solution, Ce (mmol.L–1). They can give information about the distribution of the adsorbate anions between the solution and the membrane phases when the adsorption process reaches the equilibrium state (Vasiliu et al. 2011). Several isotherm models were used to describe adsorption characteristics. The adsorption of fluoride, nitrate and sulfate ions onto the AFN membrane were analyzed using the Langmuir (1918), D-A (Dron & Dodi 2011) and Redlich & Peterson (1959) models.

The Langmuir model is the most frequently used isotherm to describe the adsorption equilibrium. Indeed, it assumes a monolayer adsorption phenomenon onto a completely homogenous surface. On the other hand, the D-A model is based in the Polanyi adsorption potential theory which defines an adsorption potential and is related to the adsorption energy. It is also used in order to distinguish between physical and chemical adsorption processes. The R-P is an empirical isotherm used to represent adsorption equilibrium over a wide concentration range. It has features of both the Langmuir (β = 1) and Freundlich (at higher concentrations) isotherms, and consequently, it can be used either in heterogeneous or homogenous systems.

The non-linear forms of the studied models are given in Table 2.

Table 2

Isotherm constants for parameter models by non-linear regression

ModelsNon-linear formCalculated parameters
Langmuir   and  
D-A  E and  
R-P  A, B and  
ModelsNon-linear formCalculated parameters
Langmuir   and  
D-A  E and  
R-P  A, B and  

The characteristic parameters, for the adsorption of , and on the unmodified AFN membrane were determined from the non-linear regression using the OriginPro software. Isotherms for the studied anions are shown in Figure 1.
Figure 1

Experimental data and adsorption isotherms determined by non-linear regression of fluoride (a), nitrate (b) and sulfate (c) ions onto unmodified AFN membrane at 298 K.

Figure 1

Experimental data and adsorption isotherms determined by non-linear regression of fluoride (a), nitrate (b) and sulfate (c) ions onto unmodified AFN membrane at 298 K.

Table 3 displays the different parameters of the Langmuir, D-A and R-P models along with regression coefficients (R2) obtained from the non-linear regression.

Table 3

Langmuir, D-A and R-P parameters determined from non-linear regression for the adsorption of , and on the unmodified AFN membrane

 Langmuir
D-A
R-P
Systemsq0 (mmol.g–1)KL (L.mol–1)R2q0 (mmol.g–1)nDE (kJ.mol–1)R2ABβRPR2
 1.51 47.09 0.994 1.43 3.05 8.12 0.967 0.07 0.05 0.994 
 2.66 116.46 0.934 2.63 3.06 9.88 0.943 0.31 0.12 0.934 
 1.42 54.7 0.967 1.37 3.11 8.44 0.950 0.08 0.05 0.967 
 Langmuir
D-A
R-P
Systemsq0 (mmol.g–1)KL (L.mol–1)R2q0 (mmol.g–1)nDE (kJ.mol–1)R2ABβRPR2
 1.51 47.09 0.994 1.43 3.05 8.12 0.967 0.07 0.05 0.994 
 2.66 116.46 0.934 2.63 3.06 9.88 0.943 0.31 0.12 0.934 
 1.42 54.7 0.967 1.37 3.11 8.44 0.950 0.08 0.05 0.967 

Results show that for the Langmuir isotherm, the value of reflects quantitatively the affinity of the membrane towards fluorides, nitrates and sulfates. The unmodified AFN membrane exhibited high selectivity for nitrate () than sulfate () and fluoride ().

The adsorption energy of a solute on the sorbent is determined from the Langmuir isotherms as the corresponding change in Gibbs free energy ΔG°, deduced from KL as follows: 
formula
where R is the universal gas constant (8.314 J.mol–1.K–1), T (K) is absolute temperature.

For the adsorption of fluoride, sulfate and nitrate on the unmodified AFN membrane, the ΔG° values were found to be –9.54 kJ.mol–1, –9.91 kJ.mol–1 and –11.79 kJ.mol–1, respectively. Results show that the standard free energy during the sorption process at 298 K were negative for the studied ions, and indicate that the adsorption process onto the unmodified AFN membrane is a favorable and spontaneous process in the standard conditions. The ΔG° values suggest the affinity order of the AFN membrane for the studied anions. Results show that the affinity order is: .

For the R-P model the magnitudes for the adsorption of , and ions were found to be equal to 1. Hence, in this study, the R-P isotherm is approaching the Langmuir but not the Freundlich isotherm because the values of are close to unity. It can be concluded that the surface of the unmodified AFN membrane is homogenous for fluoride, nitrate and sulfate ions adsorption.

For the D-A model, the maximum adsorption capacities values are almost in agreement with the values determined from the Langmuir isotherm. The magnitude of the adsorption energy, E, deduced from D-A model is useful for estimating the adsorption process type. Results (Table 3) show that the E values lies between 8 and 16 kJ.mol–1, thus it is concluded that the adsorption of fluoride, nitrate and sulfate on the unmodified AFN membrane is an ion exchange process. Adsorption energies, E, determined from the D-A are consistent with the ΔG° values obtained from the Langmuir, and suggest the same affinity order: .

For the D-A model the heterogeneity parameter was also estimated. This parameter is dependent on the adsorbent surface properties (Dron & Dodi 2011). It was mentioned that higher values of () indicate a greater homogeneity of the materials. In this study the values were determined to be in the range indicating the high homogeneity of the unmodified AFN membrane.

In order to evaluate the best fit of isotherm equations for the studied adsorption models to the experimental data, the chi-square test, , was used and it is given (Abdelwahab et al. 2013) by the following equation: 
formula
where and are the experimental and calculated equilibrium capacities (mmol.g–1).

The best isotherm was that with the highest regression coefficient, , and lowest chi-square statistic . Obtained values are given in Table 4.

Table 4

Regression coefficients, and chi-square statistic

 Langmuir
D-A
R-P
SystemsR2χ2R2χ2R2χ2
 0.994 0.14 10−3 0.967 0.94 10−3 0.994 0.17 10−3 
 0.934 1.44 10−3 0.943 1.50 10−3 0.934 1.73 10−3 
 0.967 0.60 10−3 0.950 1.12 10−3 0.967 0.72 10−3 
 Langmuir
D-A
R-P
SystemsR2χ2R2χ2R2χ2
 0.994 0.14 10−3 0.967 0.94 10−3 0.994 0.17 10−3 
 0.934 1.44 10−3 0.943 1.50 10−3 0.934 1.73 10−3 
 0.967 0.60 10−3 0.950 1.12 10−3 0.967 0.72 10−3 

As seen from Table 4, the Langmuir model describes the best fit to the experimental data.

SURFACE MODIFICATION OF THE AFN MEMBRANE

Factorial design study

The modification of the AFN membrane by the adsorption of PEI on its surface was determined as a function of the several parameters such as: initial PEI concentration, pH solution and molecular weight of the PEI. To investigate the influence of these parameters an experimental design methodology by means of a full factorial design (23) has been adopted.

The objective of the factorial design was to find parameters that have a significant influence on the adsorption of PEI on the surface of the AFN membrane. In this study a full factorial design of the type 23 was used to investigate the influence of three parameters on the surface modification of the AFN membrane. These factors were: initial PEI concentration (X1), molecular weight of the PEI (X2) and solution pH (X3).

The factorial levels of the coded factor are presented in Table 5.

Table 5

Experimental range and levels investigated for the modification of the AFN membrane

Experimental field
Coded variables (Xi)Factors (Ui)UnitLow valueHigh value
X1 (PEI) g.L–1 0.5 1.5 
X2 M (PEI) g.mol–1 1,300 106 
X3 pH – 8.8 10.8 
Experimental field
Coded variables (Xi)Factors (Ui)UnitLow valueHigh value
X1 (PEI) g.L–1 0.5 1.5 
X2 M (PEI) g.mol–1 1,300 106 
X3 pH – 8.8 10.8 

The different data have been analyzed with MINITAB software. The relative importance of the various investigated factors and their interactions are illustrated in the Pareto chart of the effects (Figure 2). The vertical line in the chart indicates minimum statistically significant effect at a 95% confidence level.
Figure 2

Pareto chart of the effects.

Figure 2

Pareto chart of the effects.

According to the Pareto chart, the initial concentration and the pH seem to be highly influent on the experimental response since the absolute values of their coefficients are higher than the t value (t = 12.71, for α = 0.05).

The main effects of each parameter on the adsorption amount of PEI are given in Figure 3.
Figure 3

Main effects plots for PEI adsorption.

Figure 3

Main effects plots for PEI adsorption.

It can be seen that the initial concentration of PEI is the most important variable on the adsorbed amount of PEI and has a positive effect. The effect of pH factors is negative, it means that increasing pH value of PEI solution leads to a decrease in the amount of PEI adsorbed on the AFN membrane. The effect of the molecular weight is very negligible.

Therefore, to modify the studied membrane, the initial concentration and the pH of the PEI solution were fixed, respectively, at the highest value (1.5 g.L–1) and the lowest value (pH = 8.8). A similar finding has been reported by Amara & Kerdjoudj (2007), they found that the modification of anion exchange membrane is more effective at high concentration and a low pH of the polyelectrolyte solution.

Adsorption study on the modified AFN membrane

Equilibrium adsorption isotherms of Langmuir, D-A and R-P models established for the adsorption of fluoride, nitrate and sulfate ions using the modified AFN membrane are given in Figure 4.
Figure 4

Experimental data and adsorption isotherms determined by non-linear regression of fluoride (a), nitrate (b) and sulfate (c) ions onto modified AFN membrane at 298 K.

Figure 4

Experimental data and adsorption isotherms determined by non-linear regression of fluoride (a), nitrate (b) and sulfate (c) ions onto modified AFN membrane at 298 K.

The different adsorption parameters along with regression coefficients (R2) and chi-square statistic obtained from the non-linear regression have been evaluated and compiled in Table 6.

Table 6

Langmuir, D-A and R-P parameters determined from nonlinear regression for the adsorption of , and on the modified AFN membrane

Model parametersCl/FCl/NO3Cl/SO42−
Langmuir 
(mmol.g−15.34 3.38 1.27 
(L.mol−186.62 86.20 51.48 
 0.978 0.982 0.959 
 3.33 × 10–3 0.99 × 10–3 0.63 × 10–3 
D-A 
(mmol.g−15.28 3.35 1.25 
 E (kJ.mol−19.49 9.45 8.33 
 2.80 2.82 2.61 
 0.980 0.985 0.964 
 3.65 × 10–3 0.97 × 10–3 0.68 × 10–3 
R-P 
 A 0.62 0.39 0.08 
 B 0.13 0.13 0.07 
 0.98 0.98 0.98 
 0.963 0.962 0.952 
 6.68 × 10–3 2.47 × 10–3 0.90 × 10–3 
Model parametersCl/FCl/NO3Cl/SO42−
Langmuir 
(mmol.g−15.34 3.38 1.27 
(L.mol−186.62 86.20 51.48 
 0.978 0.982 0.959 
 3.33 × 10–3 0.99 × 10–3 0.63 × 10–3 
D-A 
(mmol.g−15.28 3.35 1.25 
 E (kJ.mol−19.49 9.45 8.33 
 2.80 2.82 2.61 
 0.980 0.985 0.964 
 3.65 × 10–3 0.97 × 10–3 0.68 × 10–3 
R-P 
 A 0.62 0.39 0.08 
 B 0.13 0.13 0.07 
 0.98 0.98 0.98 
 0.963 0.962 0.952 
 6.68 × 10–3 2.47 × 10–3 0.90 × 10–3 

Based on the values of the regression coefficient () and the chi-square test obtained for the three studied anions, the experimental results of the adsorption of fluorides, nitrates and sulfates on the modified AFN membrane fitted well with the D-A model. The adsorption energies, E, determined from this model are between 8 and 16 kJ.mol–1, which indicates that the adsorption of fluoride, nitrate and sulfate ions on the modified AFN membrane is an ion exchange process.

To better illustrate the efficiency of the surface modification on the adsorption capacity of the AFN membrane, results were presented with the best fit modelled line both for the modified and the unmodified membrane as given in Figure 5.
Figure 5

Adsorption isotherms of fluoride (a), nitrate (b) and sulfate (c) ions on the modified and the unmodified AFN membrane at 298 K.

Figure 5

Adsorption isotherms of fluoride (a), nitrate (b) and sulfate (c) ions on the modified and the unmodified AFN membrane at 298 K.

The change in the adequacy of the isotherm model from the Langmuir model for the adsorption on the unmodified AFN membrane to the D-A model for the adsorption on the modified membrane can be explained by the decrease of membrane homogeneity after modification. In fact, for the unmodified AFN membrane the better correlation between the experimental data and the Langmuir model fit can be explained by the high homogeneity of the membrane surface. This result is in agreement with the heterogeneity parameter deduced from the R-P model which is equal to 1. As seen in Table 6, the values of the heterogeneity parameters, , determined from the D-A model, decreased () for the modified AFN membrane, indicating a decrease in the homogeneity of the AFN membrane after surface modification by PEI.

The D-A constants q0, which represent the maximum adsorption capacities, for the adsorption of , and on the modified AFN membrane, were compared with those obtained with the unmodified one deduced from the Langmuir fitting. In fact, for the unmodified membrane the maximum adsorption capacities were 1.51 mmol.g–1, 2.66 mmol.g–1 and 1.42 mmol.g–1, respectively, for the adsorption of , and ions. It is shown, from the results given in Table 6, that after surface modification of the AFN membrane the maximum adsorption capacities of fluoride and nitrate ions increase. The maximum adsorption capacity of sulfate on the unmodified membrane, however, is larger than that of the modified one.

For the modified AFN membrane, the E values determined from D-A model were found to be 9.49, 9.45 and 8.33 kJ.mol−1, respectively, for fluorides, nitrates and sulfates. Results show that the affinity order is: This finding can be attributed to the layer of the PEI formed in the surface of the membrane which plays a role of a repulsive barrier towards ions of high valences.

Desorption of fluoride from the modified AFN membrane

In order to study the possibility of the desorption of fluoride from the modified AFN membrane by nitrate and sulfate ions, adsorption equilibrium isotherms of Langmuir, D-A and R-P models were established for , and systems using the modified AFN membrane. All experiments were carried out at initial concentrations from 0.05 to 1 mol.L–1 and constant temperature of 298 K. The different adsorption parameters determined from the non-linear regression are illustrated in Table 7.

Table 7

Langmuir, D-A and R-P parameters for , and systems determined by non-linear regression

Model parametersF/NO3F/SO42−NO3/SO42−
Langmuir 
(mmol.g–11.97 1.32 1.26 
(L.mol–119.41 12.61 17.95 
 0.953 0.934 0.958 
 6.01 × 10–3 5.81 × 10−3 2.16 × 10–3 
D-A 
(mmol.g–11.87 1.22 1.19 
 E (kJ.mol–16.06 5.32 6.06 
 2.69 2.78 2.53 
 0.974 0.976 0.964 
 4.08 × 10–3 2.54 × 10–3 2.21 × 10–3 
R-P 
 A 0.04 0.02 0.03 
 B 0.02 0.02 0.02 
 0.98 0.96 0.97 
 0.941 0.914 0.948 
 9.09 × 10–3 9.02 × 10–3 3.18 × 10–3 
Model parametersF/NO3F/SO42−NO3/SO42−
Langmuir 
(mmol.g–11.97 1.32 1.26 
(L.mol–119.41 12.61 17.95 
 0.953 0.934 0.958 
 6.01 × 10–3 5.81 × 10−3 2.16 × 10–3 
D-A 
(mmol.g–11.87 1.22 1.19 
 E (kJ.mol–16.06 5.32 6.06 
 2.69 2.78 2.53 
 0.974 0.976 0.964 
 4.08 × 10–3 2.54 × 10–3 2.21 × 10–3 
R-P 
 A 0.04 0.02 0.03 
 B 0.02 0.02 0.02 
 0.98 0.96 0.97 
 0.941 0.914 0.948 
 9.09 × 10–3 9.02 × 10–3 3.18 × 10–3 

For the modified AFN membrane in form, the adsorption energies of and were found to be 9.45 kJ.mol–1 and 8.33 kJ.mol–1, respectively. As seen in Table 7, the values of the adsorption energies of and on the same membrane in form decreased. In fact, nitrate and sulfate ions are favorably adsorbed on the modified membrane in form. These results show that the anions previously exchanged on the modified AFN membrane are difficult to be removed by and ions.

CONCLUSIONS

The adsorption of fluoride, nitrate and sulfate ions on the AFN membrane was studied. The suitability of the Langmuir, D-A and R-P adsorption models to the equilibrium data was investigated. Adsorption parameters were determined from non-linear regression. Results were found to best fit Langmuir adsorption isotherm. In order to improve the selectivity of the AFN membrane towards anions, modification of the surface of this membrane with PEI has been attempted. An experimental design methodology was adopted in this study to determine the optimal experimental conditions for the modification of the AFN membrane. Factorial design results showed that the main influent parameters for the adsorption of PEI on the surface of the membrane were the initial concentration and the pH of the PEI solution. Modified AFN membrane had higher adsorption capacities for fluoride and nitrate than the unmodified membrane, while the adsorption capacity for sulfate ions decreased after the surface modification of the membrane. According to these results it can be concluded that the modified AFN membrane is an effective adsorbent for the removal of fluorides from aqueous solutions.

REFERENCES

REFERENCES
Abdelwahab
O.
Amin
N. K.
El-Ashtoukhy
E.-S. Z.
2013
Removal of zinc ions from aqueous solution using a cation exchange resin
.
Chem. Eng. Res. Des.
91
,
165
173
.
Chabani
M.
Amrane
A.
Bensmaili
A.
2006
Kinetic modelling of the adsorption nitrates by ion exchange resin
.
Chem. Eng. J.
125
,
111
117
.
Chabani
M.
Amrane
A.
Bensmaili
A.
2009
Equilibrium sorption isotherms for nitrate on resin Amberlite IRA 400
.
J. Hazard. Mater.
165
,
27
33
.
Chia-Hung
L.
Jeng-Shiou
W.
Hsin-Chieh
C.
Shing-Yi
S.
Hoong Chu
K.
2007
Removal of anionic reactive dyes from water using anion exchange membranes as adsorbers
.
Water Res.
41
,
1491
1500
.
Du Preez
L. A.
Odendaal
J. P.
Maree
J. P.
Ponsonby
M.
1992
Biological removal of sulfate from industrial effluents using producer gas as energy source
.
Environ. Technol.
13
,
875
882
.
Jeng-Shiou
W.
Chia-Hung
L.
Hoong Chu
K.
Shing-Yi
S.
2008
Removal of cationic dye methyl violet 2B from water by cation exchange membranes
.
J. Membr. Sci.
309
,
239
245
.
Meenakshi
S.
Viswanathan
N.
2007
Identification of selective ion exchange resin for fluoride sorption
.
J. Colloid Interface Sci.
308
,
438
450
.
Ranjana
A.
2012
Nitrate contamination in ground water samples of Gangapurcity town (Sawai Madhopur District) Rajasthan
.
J. Chem. Bio. Phys. Sci.
2
,
511
513
.
Redlich
O. J.
Peterson
D. L.
1959
A useful adsorption isotherm
.
J. Phys. Chem.
63
,
1024
1026
.
Sachin
N. M.
Jayshri
V. P.
Shilpi
K.
Amit
B.
Tapan
C.
Rajesh
B. B.
2011
Equilibrium isotherm and kinetic modeling of the adsorption of nitrates by anion exchange Indion NSSR resin
.
Desalination
276
,
38
44
.
Samadi
M. T.
Zarrabi
M.
Sepehr
M. N.
Ramhormozi
S. M.
Azizian
S.
Amrane
A.
2014
Removal of fluoride ions by ion exchange resin: kinetic and equilibrium studies
.
Environ. Eng. Manage. J.
13
,
205
214
.
Solangi
I. B.
Memon
S.
Bhanger
M. I.
2009
Removal of fluoride from aqueous environment by modified Amberlite resin
.
J. Hazard. Mater.
171
,
815
819
.
Tehobanoglous
G.
1979
Wastewater Engineering: Treatment Disposal Reuse
. 2nd edn.
McGraw-Hill Inc.
,
New York
.
Xu
X.
Gao
B.
Zhao
Y.
Chen
S.
Tan
X.
Yue
Q.
Lin
J.
Wang
Y.
2012
Nitrate removal from aqueous solution by Arundo donax L. reed based anion exchange resin
.
J. Hazard. Mater.
203–204
,
86
92
.