Iron–silver oxide nanoadsorbent synthesized by co-precipitation process for fluoride removal from aqueous solution and its adsorption mechanism

Abstract

Fe–Ag magnetic binary oxide nanoparticles (Fe–Ag MBON) are prepared with co-precipitation of ferric and ferrous chloride solutions, and used for the adsorption of fluoride from aqueous solution. The surface morphology of the adsorbent was characterized by XRD, SEM, TEM, FTIR, XPS, EDX, BET, DLS and VSM techniques. Batch method was followed to optimize the conditions for the removal of fluoride. The results showed maximum removal occurred at pH 3.0 and adsorption equilibrium was achieved within 20 min. Chemical kinetics of the adsorption were well fitted by pseudo-second order models (R² > 0.968) and the adsorption process followed the Langmuir isotherm model well (R² > 0.976). The fluoride adsorption capacity of Fe–Ag MBON was 22.883 mg g⁻¹, and decreased with increasing the temperature. Thermodynamic values revealed that the fluoride adsorption process was spontaneous and exothermic. Regeneration experiments were carried out for six cycles and the results indicate a removal efficiency loss of <22%.

---

1. Introduction

Fluoride (F⁻) has been recognized as a serious problem in surface/ground water.¹ Different minerals (e.g., fluorite, biotites, topaz) are the natural geological sources of fluoride which can be released into groundwater.² Moreover, various industries such as glass and ceramic production, fertilizer and semiconductor manufacturing also contribute to fluoride pollution to a large extent. The effluents of these industries may reach to thousands of mg per L, which is higher than natural water.³ Depending on the concentration and water temperature, the effect of fluoride in drinking water can be beneficial or harmful to mankind.⁴ The presence of small quantities of fluoride in ingested water is often considered to have a beneficial effect on human health and helps in the normal mineralization of bones and dental formation.⁵ On the contrary, excessive intake of fluoride leads to osteoporosis, Alzheimers syndrome, skeletal fluorosis, cancer, infertility, and thyroid disorder.⁶ There is some literature indicating that excess concentrations of fluoride can interfere in DNA synthesis and mineral metabolism.⁷ WHO (1984) reported that >260 million people in the world consume drinking water with a fluoride content more than acceptable limits (1.5 mg L⁻¹) and in the tropical countries fluoride concentration can be as high as 35 mg L⁻¹ naturally.⁸ Under these circumstances for the safeguard the environmental and human health, it is necessary to find out suitable material as well as methodology for the sustainable removal to bring down the fluoride levels to acceptable limits. The defluoridation techniques can be mainly classified into three categories, membrane process (verse osmosis, nanofiltration, dialysis and electro-dialysis), chemical process, and adsorption techniques.⁹ However, the shortcomings of membrane and chemical process are high operational and maintenance costs, require to pretreatment, regeneration, waste disposal, and secondary pollution.¹⁰ Among these available approaches, adsorption method is widely used and offers satisfactory results due to its cost effective, high efficiency, simplicity of design, reusability and reliable alternatives.¹¹ Studies of late focus on searching for various sorbents with low cost and high efficient for fluoride removal including activated alumina,¹² zeolite,¹³ hydroxyapatite,¹⁴ clay,¹⁵ modified chitosan beads.¹⁶ More and more attentions has been paid on the development of new sorbents with high selectivity toward fluoride ion, synthetic materials is strong candidates in this field.¹⁷ Iron oxides are regarded as well-known adsorbent due to their high affinities toward inorganic pollutant, high selectivity in sorption processes, low-cost and environmental friendliness.^18,19 In recent years, increasing efforts have been devoted to the synthesis the composite adsorbents (containing iron oxides) for removal of inorganic pollutant due to their extensive potential applications. For instances, Fe–Al–Ce,²⁰ Fe–Cu,²¹ Fe–Al,²² Fe–Zr,²³ Fe–Ti²⁴ have been reported for fluoride removal. These composites exhibited superior adsorption performance for fluoride removal than their individual components (their parent materials).²⁵ Previous study showed that Ag oxides derived sorbent had selective potential and good performance in phosphorus,²⁶ poly vinyl pyrrolidone (PVP),²⁷ and color removal from water over a wide pH range.²⁸ However, applied pure silver oxides as sorbent it is not economical due to their high cost. To save costs, some composite adsorbents comprising silver oxides have therefore been developed. However, up to now, removal of fluoride by Fe–Ag as an adsorbent has not been reported in the literature, although they are maybe an excellent adsorbent because used from both Fe and Ag oxides simultaneous benefits. Besides, combination the silver oxide with high cost (US ∼ $5000–15 000 per ton) and iron oxide with low price (US ∼ $500–1000 per ton) would remarkably lower the adsorbent cost. Therefore, the main objectives of this study summarized in (1) to synthesize Fe–Ag MBON with different Fe/Ag molar ratios and its characterization by XRD, SEM, TEM, FTIR, XPS, EDX, BET, DLS and VSM techniques; (2) investigate the efficiency of newly synthesized Fe–Ag MBON for fluoride removal from the aqueous solution with considering impact of various experimental parameters such as pH, temperature, adsorbent dose, contact time and initial fluoride concentrations; (3) the equilibrium isotherm and kinetic modeling and thermodynamics of fluoride adsorption process in batch system and finally (4) to investigate the mechanism for fluoride removal.

2. Experimental

2.1. Materials and instruments

Ferrous sulfate heptahydrate (FeSO₄·7H₂O, 99%), ammonia (NH₃), AgNO₃ (>99.0%) were purchased from Merck Company and used as received without further purification. In this study pH adjustment was conducted by hydrochloric acid (HCl, 35–37%) and sodium hydroxide (NaOH, 93%). Sodium fluoride (NaF, 99% Sigma-Aldrich) used for preparation fluoride stock solution in deionized water. All chemicals used in this study were analytical grade. pH meter (HACH-HQ-USA) and incubator shaker (VWR1535) were used to measure of pH solutions and adjust temperature and stirring rate, respectively.

2.2. Synthesis and characterization of Fe–Ag MBON

A series of Fe–Ag MBON were synthesized under various Fe/Ag molar ratios by co-precipitation method base on the procedure described by Chen et al. with little difference.²⁴ Predetermined amount of FeSO₄·7H₂O and AgNO₃ were dissolved in 500 mL deionized water. Under vigorous mechanical stirring, 12.5% ammonia solution was added dropwise to reach the solution pH to 9.5. The Fe/Ag molar ratio was regulated to the predetermined value by changing in the amount of FeSO₄·7H₂O or AgNO₃ added. After reaction, the black suspension (Fe–Ag) was stirred at 80 °C for 1 h. Then Fe–Ag MBON were separated from the solution by external magnetic field and washed several times with distilled water. Fe–Ag MBON dried at 45 °C for 12 h and stored in a desiccator.

The obtained material appeared in the tiny powder form named as Fe/Ag A : B (A : B = 1 : 0, 3 : 1, 1 : 1, 1 : 3 and 0 : 1. A : B means the molar ratio of FeSO₄·7H₂O and AgNO₃).

2.3. Adsorbent characterization

The surface morphological characteristics of Fe–Ag MBON were observed by scanning electron microscope (SEM, PHILIPS, S360, and Mv2300). Structural analysis of an adsorbent and the information of various functional groups in the adsorbent were conveyed by Fourier transform infrared spectroscopy with spectral grade KBr in an agate mortar (FTIR, FTS-165, BIO-RAD, USA). The elemental and structural composition were analyzed by energy dispersive X-ray (EDX, PHILIPS, S360, and Mv2300) before and after adsorption. The binding energies of the elements and specific functional groups before and after adsorption were obtained by using X-ray Photoelectron Spectroscopy with a monochromatic Al K-alpha X-ray source (1486.6 eV) (XPS, AXIS 165 (Kratos)). Particles size, shape and size distribution analysis of synthesized adsorbent analyzed by transmission electron microscopy (TEM, PHILIPS, EM 208 S 100KV). X-ray diffraction (XRD, Quanta chrome, NOVA2000) and vibrating sample magnetometer (VSM, Lakeshore 7307) used for the study of composition crystalline structures of Fe–Ag MBON and magnetic separation ability of sorbent, respectively. Dynamic light scattering particle size analysis was performed to determine the particle size distributions of Fe and Ag oxides (DLS, SZ-100 nanoparticles series). Nitrogen adsorption and desorption analysis was used to study the specific surface area and pore structure of Fe–Ag MBON samples (NOVA 2200 – Quantachrome Corp. USA). The point of zero charge (pzc) for Fe–Ag MBON was determined by plotting as final pH–initial pH (pH_f–pH_i) versus pH, yielded the pzc as the pH where pH_f–pH_i = 0 according to the method described by Zhang et al.²⁹

2.4. Batch adsorption experiments

Fluoride adsorption experiments by prepared Fe–Ag MBON carried out in batch condition and experimental scale in 100 mL flask while shaking on a shaker-incubator at 180 rpm. The stock solution of fluoride was prepared by dissolving 2.2 g NaF in 1000 mL deionized water and with dilution of stock solution the other concentrations of fluoride were prepared. The pH of solution (3–13), Fe–Ag MBON dosage (0.1–1 g L⁻¹), fluoride concentration (5–30 mg L⁻¹) and contact time (0–60 min) selected as variables in this research. The pH of solutions was controlled on desired ranges via 0.1 N HCl and NaOH solutions. After the adsorption process and separating the adsorbent, the supernatant solution was collected for the residual fluoride concentration and analysis in the UV-visible scope at the maximum adsorption (λ = 580 nm) by using spectrophotometer UV-Vis 7400CE CECIL in presence of SPANDS reagents (procedure of 4500 F⁻ standard method). All experiments were duplicating and average value is presented to ensure the reliability of the results. The amount of fluoride adsorbed onto the prepared MBON and the removal percentage (%) was calculated through the eqn (1) or (2):here: C₀ and C_e are initial and equilibrium concentrations of F⁻ (mg L⁻¹) respectively, V the volume of solution (L), and M is the mass of adsorbent (g) requirements.

For the kinetic study of fluoride adsorption on Fe–Ag MBON, six initial fluoride concentration was used in optimized conditions at room temperature, and the contact time varied from 0 min to 60 min. For the equilibrium adsorption isotherm study of fluoride adsorption on synthesized adsorbent, the initial fluoride concentration ranged from 5 to 30 mg L⁻¹ and other parameters have been set at optimum condition. Thermodynamic studies as well as carried out in optimized conditions at temperature ranged 293 to 323 K.

2.5. Desorption experiments

Six cycles of sorption and desorption were performed to examine the reusability of the Fe–Ag MBON. For sorption experiment, 500 mg of Fe–Ag MBON as sorbent was added into 1 L fluoride solution of 10 mg L⁻¹. The solution was stirred for 24 h at 200 rpm and room temperature. The pH of the solution was set in the neutral range during the sorption process. For desorption tests, the Fe–Ag MBON adsorbing fluoride were dispersed in 100 mL NaOH 0.1 M solution. The blend was stirred for 1 h, then the sorbents were separated from NaOH solution and washed several times with DI-water and used for next cycle of adsorption–desorption. Follows formula was used for calculating the desorption percentage:

2.6. Effects of co-existing anions in experiments

The interfere effects of common coexisting ions including Cl⁻, PO₄³⁻, HCO₃⁻, SO₄²⁻ and NO₃⁻ on the fluoride sorption were investigated by adding the salt of mentioned anions. The solution pH was adjusted in the neutral range. 500 mg of Fe–Ag MBON was added in each of the Erlenmeyer flask containing 50 mL fluoride solution of 10 mg L⁻¹ and were mixed in 200 rpm for 24 h at room temperature. After sorbent separation by magnet the residual fluoride was analyzed.

3. Results and discussion

3.1. Surface morphology of Fe–Ag MBON

The morphology of Fe–Ag MBON before and after adsorption are respectively shown in Fig. 1. As shown in Fig. 1a, the pure iron oxide was aggregates formed compactly by ball-like nanoparticles of 10–25 nm while, Fig. 1e shows that the pure Ag oxide was aggregates of nanoflakes of 5–20 nm. Fig. 1b–d show that with increasing of Fe/Ag ratio the size of Fe–Ag MBON decreased and their structures became firmer. Furthermore, comparing of Fig. 1f with another one implied that the surface of adsorbent before sorption process have highly porous structure which was useful for achieving appropriate adsorption efficiency whereas the porous adsorbent surface changed significantly after the adsorption process, that indicated that fluoride ion was adsorbed into the pores of Fe–Ag MBON and developed a layer of on its surface.³⁰

Fig. 1 SEM micrographs of Fe–Ag MBON. (a) Iron oxide, (b) Fe/Ag 3 : 1, (c) Fe/Ag 1 : 1, (d) Fe/Ag 1 : 3, (e) silver oxide and (f) Fe/Ag 3 : 1 after fluoride adsorption.

3.2. Specific surface area of Fe–Ag MBON

The N₂ adsorption and desorption isotherms of Fe–Ag MBON were shown in Fig. 2. The N₂ gas adsorption and desorption isotherms bass on the International Union of Pure and Applied Chemistry classification indicated that Fe–Ag MBON at a ratio of 3 : 1 is placed in type I while the other ratio of Fe–Ag MBON are type IV. Fe–Ag with 3 : 1 molar ratio was micro porous (average pore diameters < 2 nm) and other materials with different molar ratio were mesoporous (2 nm < average pore diameters > 50 nm) since they showed a hysteresis loop for the desorption isotherm.

Fig. 2 N₂ adsorption/desorption isotherms of the Fe–Ag MBON with different molar ratios.

The specific surface area, average pore diameter, and average pore volume of Fe–Ag MBON were analyzed and their results summarized in Table 1.

Table 1 BET analyze, total pore volume and average pore diameter of the Fe–Ag MBON

Sorbent	BET area (m^2 g^−1)	Total pore volume (cm^3 g^−1)	Average pore diameter (Å)
Iron oxide	94.96	0.252	5.84
3 : 1	254.24	0.678	1.31
1 : 1	189.57	0.391	6.91
1 : 3	145.74	0.479	6.21
Silver oxide	74.34	0.224	4.21

---

It can be seen increasing of Ag content have a negative effect on the specific surface area of Fe–Ag MBON and led to it's decrease, however, increased the pore diameter and the pore volume. For example, the surface area of Fe–Ag MBON 3 : 1 was over 254.24 m² g⁻¹, while the specific surface area of pure Ag and alone Fe oxide nanoparticles were 94.96 m² g⁻¹ and 74.34 m² g⁻¹, respectively. This increase might be attribute to reduce of particle size and change in shape of the particle from nano fragment to ball-like nanoparticle in the binary nano oxide. In simple terms, can be expressed that adsorbents with ball-like nanoparticle has the largest specific surface area compare than nano fragment materials.²⁵ Some researcher believed that replacement of Fe²⁺ with Ag⁺ to form Fe–Ag solid solutions due to induce crystal defects could causing the promotion of their specific surface areas. This further illustrates that these samples were not simple mixtures of FeO and Ag₂O nanoparticles and a noteworthy synergistic effect existed in this binary oxide system. Among Fe–Ag MBON with different molar ratio, Fe_0.75Ag_0.25O₂ (3 : 1) nanoparticles with 254.24 m² g⁻¹ have largest specific surface area which may be relevant to the smallest crystallite size of them. Largest specific surface along with highest total pore volume (0.678 cm³ g⁻¹) and average pore diameter 8.31 nm has led that Fe_0.75Ag_0.25O₂ converted to the adsorbent with well adsorption performance.

3.3. Energy dispersive X-ray of Fe–Ag MBON

The EDX spectrum of Fe–Ag MBON before and after adsorption process are recorded in Fig. 3. The element peaks of O and Ag were observed at the energy values of 0.5 keV, 3 keV, and energy values of 1.25 and 6.6 keV attributed to Fe (Fig. 3a), this indicates that Fe and Ag were successfully synthesized. Moreover, the adsorption of fluoride on the surface of Fe–Ag MBON was ascertained from the EDX spectrum which shows the presence of fluorine along with the other major peaks.

Fig. 3 EDX spectra of Fe–Ag MBON (a), before and (b) after adsorption process.

3.4. X-ray diffraction spectroscopy of Fe–Ag MBON

Fig. 4a shows the XRD patterns of Fe oxide (Fe₃O₄) nanoparticles with no silver component existed. XRD diffraction peaks appeared at 2θ of 35.4° and 77.2° which corresponds to (110) and (211) Bragg reflection respectively, in accordance with the typical XRD diffraction peaks of Fe₃O₄ cubic fluorite structure bass on PDF 34-0394.^31,32 Fig. 4b present the silver oxide nanoparticles in the absence of Fe₃O₄ nanoparticles component. XRD diffraction peaks appeared at 2θ of 37.8°, 64.7° and 73.9° attributable to the indices (111), (220) and (311) indicates existence of Ag³³ and in consistent with the typical XRD diffraction peaks of Ag₂O tetragonal phase base on PDF 49-1642. These peaks were also observed in the pattern of synthesized Fe–Ag MBON, indicating that the cubic phase of Fe₃O₄ is maintained after Ag loading. (Can be seen in Fig. 4c.)

Fig. 4 X-ray diffraction patterns of (a) Fe, (b) Ag and (c) Fe–Ag MBON in ranges of 2θ = 30–110 by using Cu kα radiation, (d) crystallite sizes of Fe–Ag MBON with the changes in amount of Ag content.

The Scherrer's formula applied to the calculated crystallite size of Fe–Ag MBON from the strongest XRD peak and the results presented in Fig. 4d

D = 0.9λ/β cos θ

where: D (nm): mean size of the crystalline domains, λ: the average X-ray wavelength, β: the line broadening at half the maximum intensity, θ: the Bragg angle.

With the increase of silver content to 20%, the crystallite size decreased slightly from ∼10.25 nm to ∼9.8 nm, however, with the further increase of the silver content from 20 to 100%, the particle size began to increase from ∼9.8 nm to ∼11.47 nm. The replacement of heavier Ag⁺ by lighter Fe²⁺ in Fe_0.75Ag_0.25O₂ could induce crystal defects and cause reduces the size of the nanoparticles degree. According to the evidence it can be found that with increasing the replacement/substitution rate of Ag⁺ : Fe²⁺ the size of nanoparticles began to decrease.

3.5. Dynamic light scattering and transmission electron microscopy of Fe–Ag MBON

Structure and shape of Fe–Ag MBON was analyzed by TEM micrographs at 90 keV (Fig. 5 (left)). According this analysis, synthesized absorbent were sphere shaped structure, almost uniform and intertwined.³⁴ DLS results are depicted in Fig. 5 (right). DLS curves strongly indicate the small Ag NPs (∼5 nm) distributed on the Fe (∼10 nm) surface which well accordance with the TEM finding.

Fig. 5 TEM images of Fe–Ag MBON (left) DLS curve of Fe, and Ag (right).

3.6. Vibrating sample magnetometer

For determining the magnetic properties of Fe–Ag MBON, VSM analysis was performed in magnetic field of ±100 kOe at 25 °C. VSM curve showed the maximum saturation magnetization for Fe–Ag MBON was equal to 80 emu g⁻¹. The achieved results depict that Fe–Ag MBON is super-paramagnetic and can easily separate from samples by an external magnetic field without secondary pollution (Fig. 6).

Fig. 6 Point of zero charge (pzc) determination for Fe–Ag MBON (left); VSM analysis of Fe–Ag MBON and magnetic separation of it from aqueous solution (insert) (right).

3.7. Point of zero charge (pzc)

The pzc values of Fe–Ag MBON were shown in Fig. 6 (left). The pzc of iron oxide, Fe/Ag 3 : 1, Fe/Ag 1 : 1, Fe/Ag 1 : 3 and silver oxide were 5.91, 6.03, 6.7, 6.1 and 7.1, respectively. Clearly, the pzc value raised as Ag content increased in these oxides.

3.8. Effect of Ag loading on removal efficiency

The effect of the Fe²⁺, Ag⁺, Fe₃O₄, Ag₂O and Fe–Ag with various molar ratios (3 : 1, 1 : 1, and 1 : 3) on the removal efficiency is shown in Fig. 7. The Fe–Ag adsorbent reached the highest adsorption capacity of 20.61 mg g⁻¹, which was much higher than of other samples, at the Fe–Ag ratio of 3 : 1. These results completely comply with the results of BET analyze and show the synergistic effect of Fe and Ag on fluoride adsorption capacity at a molar ratio of 3 : 1.

Fig. 7 Effect of the Fe²⁺, Ag⁺, Fe₃O₄, Ag₂O and Fe–Ag MBON on removal efficiency.

3.9. Solution pH effect on fluoride adsorption by Fe–Ag MBON

The solution pH is a significant factor that effects on defluorination at the solid/liquid interface. This parameter by changing the surface charges and functional groups can influence the surface properties and active sites of the sorbents. Fluoride removal by different pH as a function of time were studied and the results are shown in Fig. 8 (left).

Fig. 8 (left) Effects of solution pH on defluorination by Fe–Ag MBON. Initial fluoride concentration: 10 mg L⁻¹, adsorbent dose: 200 mg L⁻¹ at room temperature; (right) effects adsorbent dose on fluoride removal.

Fig. 8 (left) shows clearly that fluoride adsorption on Fe–Ag MBON was pH-dependent. Under the acidic pH (pH = 3), the adsorption performance increased to 100%, whileit dropped much faster at neutral and alkaline condition. These results may be relevant to acidic status effects which led to an increase of positive protons (H⁺) that can dominate the surface of Fe–Ag MBON and cause an induced a positive charge on the adsorbent surface. Thus the electrostatic attractive force between the positive protons attached on the surface of Fe–Ag and negative fluoride ions lead to an increased adsorption capacity.³⁵ Decreasing fluoride adsorption in an alkaline status is due to elevated hydroxide ion production. The electrostatic repulsion force between OH⁻ ions (hydroxide) and negative charges of fluoride prevent the diffusion of fluoride ions and caused a decrease of adsorption capacity. Since that maximum efficiency of fluoride removal was occurred at pH = 3, this pH chosen for future experiment. Similar phenomena reported by other researchers in F⁻ adsorption by various binary oxides.^36,37 In other hand, the surface charge of Fe–Ag MBON is related to the pH_pzc and pH of the solution. When pH > pH_pzc, the surface charge is negative and with getting bigger pH_pzc compare pH, positive charge induced on surface of sorbent.

M–OH + H₂O → M–OH₂⁺ + OH⁻ (pH < pH_pzc) (4)

M–OH → M–O⁻ + H⁺ (pH > pH_pzc) (5)

where, M indicates metal ions like Fe²⁺, Ag.

In this work the point zero charge (pH_pzc) calculated for the synthesized Fe–Ag MBON and depicted in Fig. 6 (left). The results showed pH_pzc for our sorbent was 5.6 and the best defluorination was achieved at pH equal 5 (Scheme 1).

Scheme 1 Effects of solution pH on defluorination by Fe–Ag MBON.

So, it can be concluded, at pH < 5.6 the surface of the adsorbent has positively charged which confirms the above contents. In addition base on the significant relationship between pH and pH_pzc it can be concluded that the mechanism of fluoride adsorption followed a non-specific adsorption pattern.³⁸ Reactions (6) and (7) depicted the fluoride sorption at acidic condition:

M–OH + H⁺ → M–OH₂⁺ (6)

M–OH₂⁺ + F⁻ → M⁺–F⁻ + H₂O (7)

3.10. Influence of adsorbent dose on fluoride removal

Adsorbent dose appears to have an important influence on adsorption capacity and removal efficiency. Adsorption of 10 mg L⁻¹ fluoride concentration by different sorbent dosages (0.1–1 g L⁻¹) were studied at room temperature. Fig. 8 (right) shows that increasing of adsorbent dosage from 0.1 to 1 g L⁻¹, the removal percentage increased from 26.21% to 100% and adsorption capacity (q) decreased from 26.21 to 10 mg g⁻¹. It can be explained that increasing the amount of sorbent provides greater number of surface area (available active sites) for the fluoride binding.³⁹ Wei Ma et al. in study of fluoride removal by Mg–Al–Fe hydrotalcite-like compound reported that increasing of adsorbent dosage from 0.1 g L⁻¹ to 2 g L⁻¹ led to elevation of adsorption performance from 40% to 85%.⁴⁰ In the present work the highest efficiency of fluoride adsorption achieved to 1 g L⁻¹ adsorbent, however, with considering the creating adsorbent cost and worthless effect of fluoride removal with increasing sorbent from 0.5 to 1 g L⁻¹, using 0.5 g L⁻¹ adsorbent will be economical. Therefore, 0.5 g L⁻¹ adsorbent dosage was chosen to future experiment.

3.11. Effect of contact time and kinetics equilibrium studies

The kinetics of the fluoride adsorption from aqueous solutions onto Fe–Ag MBON with five initial fluoride concentrations (5, 10, 15, 20, 25 and 30 mg L⁻¹), at the optimized condition was examined the results presented in Fig. 9. This figure shows in all concentration the most fluoride adsorption and the highest rate of defluoration occurred within the first 20 min of the contact time. For example, 100% fluoride was removed after 20 min when the initial fluoride concentration was less than 10 mg L⁻¹. The observed rapid defluoration could be attributed to following reasons:

Fig. 9 Effects of contact time on defluorination by Fe–Ag MBON. pH = 3, adsorbent dose: 0.5 g L⁻¹ and room temperature.

(i) The existing many vacant adsorption sites in adsorbent surface in initial time, which are filled with time passed.

(ii) Fe–Ag MBON have a nano size, thus, the distance of fluoride for diffusion from the bulk solution onto active sites is shortened.

(iii) Prepared sorbent in nano scale cause large external surface area and proper pore size/volume which provided well contact efficiency with fluoride.

When the initial fluoride concentration increased to 15, 20, 25 and 30 mg L⁻¹, fluoride removal decrease to 84.1%, 60%, 50.2% and 30.3% after 20 min, respectively. Fixed vacant adsorption sites versus increasing concentrations of contaminants is reason of relatively lower fluoride removal ratio, while with increasing the initial fluoride concentration from 5 to 30 mg L⁻¹, adsorbed fluoride amount increased from 10.0 mg g⁻¹ to 18.9 mg g⁻¹. It is notable that vacant adsorption sites and external surface area with further increase of the adsorption time filling and be caused adsorption rate slowed down and gradually reached the equilibrium state. Adsorption of fluoride from aqueous solution using meso-structured zirconium phosphate was studied by Swain (2011) which confirm these results. Swain reported that laps of the contact time led to filling of vacant sites by fluoride that the removal efficiency will be reduced.⁴¹

In the present study, we have been using pseudo-first-order and pseudo-second-order models to establish the best fitted model for experimental data and describe the kinetic behavior of fluoride sorption on the Fe–Ag MBON. The pseudo-first and second order model equations are defined as:

ln(q_e − q_t) = ln q_e − K₁t (8)

where, q_e (mg g⁻¹): adsorption capacity of the sorbent in equilibrium time, q_t (mg g⁻¹): adsorption capacity of the adsorbent at time, K₁ (L min⁻¹): coefficients of reaction rate for the pseudo-first-order models, K₂ (mg g⁻¹ min⁻¹): coefficients of reaction rate for the pseudo-second-order models. The related parameters of pseudo-first and second order kinetic can be calculated from the plots log q_e − q_t and q_t/t versus t, respectively. The regression coefficient (R²) and compatibility between q_e (experiment) with q_e (calculate) used to choose the best kinetic models. The fitting curves of the fluoride adsorption kinetic data presented in Fig. 10 and their parameter values are presented in Table 2. The results showed the pseudo-second-order model with R² > 0.968 in all concentration can be better described adsorption of fluoride, that exhibits the rate-limiting is chemisorption and sharing of electrons from sorbent and adsorbate involving covalent forces.²³ In addition q_e,cal at the pseudo-second-order compared to other models were more closer with q_e,exp. In a similar study, Swain et al. illustrated that the pseudo-second order model with R² = 0.95 describe the kinetics of fluoride sorption on alginate entrapped Fe(III)–Zr(IV) binary.²³ These findings comply with the reported results for the fluoride sorption on trimetal-oxide adsorbent,⁴ nano zirconium chitosan composite⁴² and composite material of biopolymer alginate entrapped mixed metal oxide nanomaterials⁴³

Fig. 10 Kinetics of fluoride removal by Fe–Ag MBON at pH 3 and 25 ± 1 °C.

Table 2 Pseudo-second-order kinetic and pseudo-first order model parameters for binary metal oxide

Models	Parameters	Concentration (mg L^−1)
		10	15	20	25	30
Pseudo-first order	qe,cal (mg g^−1)	21.6	27.9	32.3	41.8	46.3
	K1 (min^−1)	0.137	0.123	0.119	0.103	0.089
	R^2	0.860	0.865	0.865	0.884	0.896
Pseudo-second order	qe,cal (mg g^−1)	23.61	36.54	41.27	52.3	58.91
	K2 (g mg^−1) (min^−1)	4	4.35	5.56	5.84	6.01
	R^2	0.978	0.985	1	0.968	0.999
Experimental		20	21.54	24.27	26.3	28.69

---

3.12. Adsorption isotherm studies

Analysis of the isotherms equilibrium is a vital stage in designing of the sorption systems. Langmuir, Freundlich, Temkin and Dubinin–Radushkevich models used to investigate the behavior of fluoride sorption and determine the maximum adsorption capacity of Fe–Ag MBON at room temperatures and 10 mg L⁻¹ of the initial concentration. Linear equation of the isotherm models were used for data analysis based on the following formula:

q_e = K_F(C_e)^1/n (10)

q_e = (q_mK_LC_e)/(1 + K_LC_e) (11)

q_e = q_m ln K_T + q_m ln C_e (12)

ln q_e = ln q_m − Dε² (13)

where, K_F ((mg g⁻¹) (L g⁻¹)^1/n): Freundlich constant, 1/n: Freundlich exponent, K_L (L mg⁻¹): Langmuir isotherm constant, K_T (L mg⁻¹): Temkin isotherm constant, D (mol² kJ⁻²): Dubinin–Radushkevich isotherm constant. The relative parameters of Langmuir, Freundlich, Temkin and Dubinin–Radushkevich isotherm models were calculated from the plots between [C_e/q_e vs. C_e], [log(q_e) vs. log(C_e)], [q_e vs. log(C_e)] and [ln q_e vs. ε²]. Fig. 11 presents the isotherms resulted of the fluoride adsorption onto Fe–Ag MBON. Furthermore, Table 3 reports the other isothermal parameters obtained in fitting the experimental data. As shown in Table 3, the Langmuir model with R² > 0.976 have a higher regression coefficient compare than the other models, so, this model is very well fitting for describe the sorption behavior of fluoride on Fe–Ag MBON. Similar findings were reported in the literature for the removal of fluoride by different types of adsorbents.^44,45 The Langmuir adsorption capacity (q_max) was determined as 20.57 mg g⁻¹ at 298 K. Table 4 compared the reported q_max of other binary oxide on fluoride adsorption with prepared Fe–Ag MBON in our study. It was found that the adsorption capacity of Fe–Ag MBON was premiered to many other adsorbents. It may be due to increase of the porous structure in the adsorbents. It was about 27% higher than Fe–Zr binary oxide while it was 55% lower than Fe–Mn binary oxide in aqueous environment. Therefore, synthesized adsorbent could be considered as a new adsorbent for fluoride removal.

Fig. 11 Fluoride isotherms adsorption onto Fe–Ag MBON.

Table 3 Langmuir, Freundlich, Dubinin–Radushkevich and Temkin isotherm constants for fluoride adsorption by Fe–Ag MBON

Models	Parameters	Linear method	Non-linear method
Langmuir	qm (mg g^−1)	20.91	21.57
	KL (L mg^−1)	1.38	1.41
	R^2	0.976	0.998
	RL	0.2	0.31
Freundlich	KF (mg g^−1) (L mg^−1)^1/n	19.87	18.96
	n	15.82	15.53
	R^2	0.911	0.931
Temkin	KT (mg g^−1) (L mg^−1)^1/n	13.79	14.12
	bT	210.06	211.45
	R^2	0.903	0.903
D–R	q (mol g^−1)	1.14 × 10^−5	1.17 × 10^−5
	D (mol^2 kJ^−2)	0.0047	0.0049
	E (kJ mol^−1)	14.89	13.93
	R^2	0.944	0.964

---

Table 4 Comparison of various binary oxides for fluoride removal

Adsorbent	Solution pH	Max. sorption capacity (mg g^−1)	Ref.
Fe–Ag (25 °C)	5	20.57	This work
Fe–Mn	5.6	79.5	46
Fe–Zr	4	13.6	23
Al–Zr	5	5.8	47
Fe–Ti	ND	27.5	24
Ti–Al	3	2.22	48
Ce–Zr	5.8 ± 0.2	19.5	49

---

All the information implied that the Langmuir model can be suggested for fluoride sorption by Fe–Ag MBON that refers to the adsorption process occur on homogeneous surface and fluoride adsorption taken in a monolayer adsorption manner.

The R_L parameter is the dimensionless constant factor which indicates the affinity of adsorbed material to the adsorbent and calculated using eqn (14):

Based on R_L factor the nature of the isotherm to be classified as follows: R_L > 1: unfavorable isotherm; R_L = 1: linear isotherm; 0 < R_L < 1: favorable isotherm; R_L < 1: irreversible isotherm.

The R_L value in this experiment is between 0 and 1 that indicate the fluoride molecules are desirably adsorbed on adsorbent.

Some literature used from Dubinin–Radushkevich isotherm to obtain the mean energy of adsorption as given in equation:

E = 1/(2D)^0.5 (15)

Chemisorption process happened when 8 < E < 16 kJ mol⁻¹, while in E < 8 kJ mol⁻¹ physical adsorption process occurred. In present study E value was determined at 14.89 kJ mol⁻¹, thus, the fluoride adsorption onto Fe–Ag MBON was a chemisorption process.

3.13. Effect of temperature and thermodynamic study of fluoride removal

To determine the impact of the temperature on the fluoride adsorption at various initial concentrations, temperature varied from 293 to 323 K while the contact time was selected as 20 min. The result revealed that increasing the temperature had a negative effect on the adsorption process (Fig. 12) and lower temperature is favored to removal of fluoride.

Fig. 12 (left) Effect of temperature on the adsorption of fluoride onto Fe–Ag MBON (pH = 3, adsorbent dose: 0.5 g L⁻¹, and contact time 20 min); (right) van't Hoff curve.

The distribution coefficient constant (K_d) used for determining the thermodynamic values and the slope and intercept of the van't Hoff plot (ln K_d versus 1/T) are used for calculating the numerical values of ΔH⁰ (the change in the enthalpy (kJ mol⁻¹)) and ΔS⁰ (the amount of energy change (J mol⁻¹ K⁻¹)), respectively which defined as:

ΔG⁰ = ΔH⁰ − (TΔS⁰) (17)

where: K_d is the distribution coefficient (L g⁻¹), R is the universal gas constant (8.314 J mol⁻¹ K⁻¹) and T is solution temperature (K). The van't Hoff graph was shown in Fig. 12 (right) and the fluoride adsorption thermodynamic parameters on Fe–Ag MBON are represented in Table 5. A negative values in standard free energy acquired in our studies exhibits that the sorption process is a spontaneous reaction. The value of ΔH⁰ is negative that confirms the adsorption process is exothermic in nature. The value of ΔS⁰ was found from 0.026 to 0.12, which indicated that the randomness increased in solid/liquid interface during the fluoride sorption process with increasing temperature.

Table 5 Thermodynamic parameters for fluoride adsorption by Fe–Ag MBON

Con. (mg L^−1)	Temperature (K)	ln Kd	ΔG^0 (kJ mol^−1)	ΔH^0 (kJ mol^−1)	ΔS^0 (kJ mol^−1 K^−1)
5	293	7.60	−18.51	−237.86	0.12
	308	7.90	−20.23
	323	8.22	−22.10
10	293	7.40	−18.03	−201.23	0.11
	308	7.70	−19.71
	323	7.93	−21.29
20	293	1.58	−3.86	−167.34	0.053
	308	1.85	−4.72
	323	2.02	−5.43
30	293	0.51	−1.24	−92.48	0.026
	308	0.65	−1.67
	323	0.75	−2.02

---

3.14. Coexisting anions effect on fluoride adsorption by Fe–Ag MBON

Foreign anions such as SO₄²⁻, NO₃⁻, Cl⁻, HCO₃⁻ and PO₄³⁻ usually presence in drinking water that can compete with fluoride for adsorption on available sites. So the impact of mention ions as competing parameter on defluorination were examined. In order to evaluate the effect of co-existing anions, fluoride concentration was kept at 10 mg L⁻¹ and initial concentration of co-ions with dissolved of their sodium salts were studied from 0 to 130 mg L⁻¹. Fig. 13a can be shown that NO₃⁻, Cl⁻, SO₄²⁻ have a slight effect on defluorination processes. However, the presence of PO₄³⁻ and HCO₃⁻ have a significantly impact on defluorination processes. As can be seen in Fig. 13a, the fluoride sorption efficiency decreased approximately 30 and 52% by increasing the PO₄³⁻ and HCO₃⁻ ions respectively to 120 mg L⁻¹. Generally would be expected that PO₄³⁻ and HCO₃⁻ to occupy the active adsorption site have a high interest to compete with fluoride which will eventually lead to reduce the defluorination efficiency. Furthermore, the impact of studied anions towards sorption process may be due to their affinity towards the sorbent substance.

Fig. 13 Effect of co-anions on fluoride adsorption efficiency using Fe–Ag MBON (a), adsorption/desorption of binary metal oxide (b).

3.15. Fluoride desorption studies

Adsorption/desorption of fluoride by Fe–Ag MBON were performed in batch condition and shown in Fig. 13b. Mention figure shows that reusability of Fe–Ag MBON slightly reduced after the six successive sorption–desorption cycles. So we suggested that Fe–Ag MBON can be repeatedly used for fluoride sorption without many losses in initial sorption efficiency. Furthermore, >62.81% adsorbed fluoride could be desorbed/recovered in the presence of NaOH in the sixth cycle which can be used in different cases like industrial applications.

To test whether Fe and Ag leach out into the solution from the Fe–Ag MBON surface, the samples were analyzed by applying ICP-MS after each reaction cycle. Based on the results, the residues of Fe and Ag were not detected. In previously conducted studies, it has been reported that relatively higher amount of the residues of Cu and Pd was detected during the removal of contaminates by bimetallic adsorbent and catalysts which were supported by activated carbon (18.1%),⁵⁰ carbon nanotube (15.9%),⁵¹ silica oxide (10.6%), alumina (30.7%), and zirconium oxide (26.0%).^52,53 Considering the result of the present study, it can be speculated that Fe–Ag MBON has acceptable stability and durability to strongly bind the bimetals to each other; which can be regarded as an advantage for its recycling in continuous fluoride removal (Scheme 2).

Scheme 2 Mechanism of fluoride desorption on Fe–Ag MBON.

3.16. Mechanism of fluoride adsorption

The FTIR spectra of the Fe oxide, Ag oxide and Fe–Ag oxide before and after adsorption are revealed in Fig. 14. For three samples, the created band at 3440 cm⁻¹ was allocated to the carbonate adsorbed, it is because the sorption experiments were perform open to the atmosphere. The common peak in the Ag oxide and Fe–Ag MBON at 1398 cm⁻¹ was apparent that not present of Fe oxide. In the FTIR spectra the peaks at near 1120 cm⁻¹ in the all three oxides can be apportioned to the bending vibration of the hydroxyl (OH⁻) of metal oxides (metal–OH). It was discovered that tow peaks at 1398 and 1120 cm⁻¹ can be regarded as the identify factors for the active sites of the adsorbents. The peak at 587.8 cm⁻¹ was assigned to Fe oxide and created peak at 615.2 cm⁻¹ assigned to Ag oxide. However, the Fe–Ag oxide had its distinctive peak at a wave number of 448.0 cm⁻¹, which was intelligibly lower than those of Fe and Ag oxide. From Fe oxide (blue curve) shifting in range of 1398 to 1120 cm⁻¹, it was deduced that Fe–O₂–Ag bonds were formed in the Fe–Ag MBON adsorbent. Generally, with considering to morphologies, XRD and above analysis can be inferred which Fe and Ag have a synergistic effect on the structure of Fe–Ag MBON and was not a simple combination between Ag oxide and Fe oxide. After fluoride adsorption clearly observed that large changes happened on the FTIR spectrum of Fe–Ag MBON. Vibration peaks near 1120 cm⁻¹ (metal-OH) got largely weakened that implied the loss and decrease of surface hydroxyl groups during the fluoride uptake. In other word, surface hydroxyl groups (M–OH) were replaced by the adsorbed fluoride and can play an important role in the adsorption of fluoride.

Fig. 14 FT-IR spectra of Fe oxide, Ag oxide and Fe–Ag oxide before and after adsorption.

The mechanism of fluoride adsorption on the surface of Fe–Ag as adsorbent before and after the sorption process were analyzed by XPS technique and the results are shown in Fig. 15. Comparing XPS analysis before and after the adsorption process shows that created a new peak with a binding energy ∼ 685 eV which was assigned to F1s. Besides, the relative atomic ratios of XPS analysis (%) for Ag, O, Fe and F in adsorbents can confirm above context.

Fig. 15 XPS spectra of Fe–Ag MBON before and after fluoride adsorption (left); XPS spectra of F1s (right).

As seen in Table 6 amount of fluoride on the surface of Fe–Ag MBON before sorption process equal to 0 while after adsorption increased to 8.51%, which confirmed that fluoride had been uptake.

Table 6 Atomic ratios of XPS analysis for the optimized synthesized sorbent before and after fluoride sorption

Atomic ratios (%)	Ag	Fe	O	F
Fe–Ag oxide	14.28	7.56	81.21	0
Fe–Ag oxide–F	19.41	7.75	69.31	8.51

---

The O1s spectra region of the optimized Fe–Ag MBON before and after the sorption process presented in Fig. 16a and b. Two peaks in the O1s spectra region are observed, one at ∼532 eV related to O₂⁻ that bonding with metal and other at ∼533 eV dedicated to OH⁻ groups on the surface of sorbent. As it can be seen the OH⁻ groups on the surface of Fe–Ag MBON decreased after fluoride sorption. The summarized details of O1s spectra region are given in Table 7.

Fig. 16 XPS (O1s spectra) analyze of the optimized Fe–Ag and the fitted distribution of OH⁻ and O²⁻ before (a) and after (b) fluoride adsorption.

Table 7 Distribution of O²⁻ and OH⁻ in XPS (O1s spectra) analyze for optimized Fe–Ag MBON as adsorbents before and after fluoride sorption

Sample	Peak	BE (eV)	FWHM (eV)	Percent (%)
Fe–Ag oxide	O^2−	532.2	1.72	41.65
	OH^−	533.21	1.65	57.61
Fe–Ag oxide–F	O^2−	531.80	2.38	97.87
	OH^−	532.82	1.83	1.53

---

Table 7 shows after defluorination the relative area of hydroxyl dropped sharply from 57.61% to 1.53% and the area ratio for the peak at ∼532 eV allocated to O₂⁻ increased from 41.65% to 97.87%. The decrease of OH⁻ group percentage attributed to the substitution of hydroxyl by fluoride during the adsorption process. Thus, the XPS analysis also suggested OH⁻ groups onto the surface of Fe–Ag MBON played a significant role in fluoride adsorption, which was in accommodation with the results of the FTIR studies.

4. Conclusions

In this work, several experiments were performed to investigate the performance of Fe–Ag MBON as an adsorbent for removal of fluoride. The Fe–Ag MBON was synthesized in a molar ratio of 3 : 1 at ∼10 nm diameter (micro porous) and a suitable magnetic properties. The optimum condition for defluoridation (removal efficiency = 100%) obtained in pH = 3, 20 min contact time and 0.5 g L⁻¹ sorbent dosage at 293 K. Therefore synthesized sorbent can be used in the water and wastewater advance treatment confidently with no need of further filtering and centrifugation. The adsorption equilibrium data followed well with Langmuir isotherm models. The kinetic study shows that system has better coordination with pseudo second-order models. Thermodynamic study illustrates that sorption of fluoride is feasible, exothermic and spontaneous in nature. The presence co-existing anions except PO₄³⁻ and HCO₃⁻ have not a significant effect on F⁻ removal.

Acknowledgements

This study was done by financial support of Tehran University of Medical Sciences and Iranian Nano Technology Initiative Council. Also, the authors would like to thanks Baqiyatallah University of medical Sciences.

References

1. H. Jin, Z. Ji, J. Yuan, J. Li, M. Liu, C. Xu, J. Dong, P. Hou and S. Hou, Research on removal of fluoride in aqueous solution by alumina-modified expanded graphite composite, J. Alloys Compd., 2015, 620, 361–367.
2. A. Bhatnagar, E. Kumar and M. Sillanpää, Fluoride removal from water by adsorption—a review, Chem. Eng. J., 2011, 171, 811–840.
3. M. Habuda-Stanić, M. E. Ravančić and A. Flanagan, A Review on Adsorption of Fluoride from Aqueous Solution, Materials, 2014, 7, 6317–6366.
4. J. Wang, D. Kang, X. Yu, M. Ge and Y. Chen, Synthesis and characterization of Mg–Fe–La trimetal composite as an adsorbent for fluoride removal, Chem. Eng. J., 2015, 264, 506–513.
5. T. Nur, P. Loganathan, T. Nguyen, S. Vigneswaran, G. Singh and J. Kandasamy, Batch and column adsorption and desorption of fluoride using hydrous ferric oxide: Solution chemistry and modeling, Chem. Eng. J., 2014, 247, 93–102.
6. S. Jagtap, M. K. Yenkie, N. Labhsetwar and S. Rayalu, Fluoride in drinking water and defluoridation of water, Chem. Rev., 2012, 112, 2454–2466.
7. J.-Q. Zhu, Y.-J. Si, L.-Y. Cheng, B.-Z. Xu, Q.-W. Wang, X. Zhang, H. Wang and Z.-P. Liu, Sodium fluoride disrupts DNA methylation of H19 and Peg3 imprinted genes during the early development of mouse embryo, Arch. Toxicol., 2014, 88, 241–248.
8. V. Hernandez-Montoya, L. A. Ramírez-Montoya, A. Bonilla-Petriciolet and M. A. Montes-Moran, Optimizing the removal of fluoride from water using new carbons obtained by modification of nut shell with a calcium solution from egg shell, Biochem. Eng. J., 2012, 62, 1–7.
9. N. A. Ingle, H. V. Dubey, N. Kaur and I. Sharma, Defluoridation techniques: Which one to choose, Journal of Health Research and Reviews, 2014, 1, 1.
10. N. K. Labhsetwar, S. J. Lunge, A. K. Bansiwal, P. K. Labhasetwar, S. S. Rayalu and S. R. Wate, Technological Developments in Water Defluoridation, Aquananotechnology: Global Prospects, 2014, p. 317.
11. I. Saha, A. Ghosh, D. Nandi, K. Gupta, D. Chatterjee and U. C. Ghosh, β-Cyclodextrin modified hydrous zirconium oxide: synthesis, characterization and defluoridation performance from aqueous solution, Chem. Eng. J., 2015, 263, 220–230.
12. D. Dayananda, V. R. Sarva, S. V. Prasad, J. Arunachalam, P. Parameswaran and N. N. Ghosh, Synthesis of MgO nanoparticle loaded mesoporous Al₂O₃ and its defluoridation study, Appl. Surf. Sci., 2015, 329, 1–10.
13. S.-F. Chen, H.-Z. Zhang, Z.-L. Wang, X.-L. Tang and L.-N. Pan, Removal of Fluoride from Water by Activated Natural Zeolite Coated with Chitosan, Environ. Sci. Technol., 2012, 7, 018.
14. K. Pandi and N. Viswanathan, In situ precipitation of nano-hydroxyapatite in gelatin polymatrix towards specific fluoride sorption, Int. J. Biol. Macromol., 2015, 74, 351–359.
15. A. D. Atasoy and M. O. Sahin, Adsorption of fluoride on the raw and modified cement clay, Clean: Soil, Air, Water, 2014, 42, 415–420.
16. J. Zhang, N. Chen, Z. Tang, Y. Yu, Q. Hu and C. Feng, A study of the mechanism of fluoride adsorption from aqueous solutions onto Fe-impregnated chitosan, Phys. Chem. Chem. Phys., 2015, 17, 12041–12051.
17. S. M. Prabhu and S. Meenakshi, Chemistry of defluoridation by one-pot synthesized dicarboxylic acids mediated polyacrylamide–zirconium complex, Chem. Eng. J., 2015, 262, 224–234.
18. E. Kumar, A. Bhatnagar, W. Hogland, M. Marques and M. Sillanpää, Interaction of inorganic anions with iron–mineral adsorbents in aqueous media—A review, Adv. Colloid Interface Sci., 2014, 203, 11–21.
19. J. García-Sánchez, M. Solache-Ríos, M. Alarcón-Herrera and V. Martínez-Miranda, Removal of fluoride from well water by modified iron oxides in a column system, Desalin. Water Treat., 2014, 1–9.
20. X. Wu, Y. Zhang, X. Dou, B. Zhao and M. Yang, Fluoride adsorption on an Fe–Al–Ce trimetal hydrous oxide: characterization of adsorption sites and adsorbed fluorine complex species, Chem. Eng. J., 2013, 223, 364–370.
21. G. Li, S. Gao, G. Zhang and X. Zhang, Enhanced adsorption of phosphate from aqueous solution by nanostructured iron(III)–copper(II) binary oxides, Chem. Eng. J., 2014, 235, 124–131.
22. N. Chen, C. Feng and M. Li, Fluoride removal on Fe–Al-impregnated granular ceramic adsorbent from aqueous solution, Clean Technol. Environ. Policy, 2014, 16, 609–617.
23. S. Swain, T. Patnaik, P. Patnaik, U. Jha and R. Dey, Development of new alginate entrapped Fe(III)–Zr(IV) binary mixed oxide for removal of fluoride from water bodies, Chem. Eng. J., 2013, 215, 763–771.
24. L. Chen, B.-Y. He, S. He, T.-J. Wang, C.-L. Su and Y. Jin, Fe–Ti oxide nano-adsorbent synthesized by co-precipitation for fluoride removal from drinking water and its adsorption mechanism, Powder Technol., 2012, 227, 3–8.
25. W. Zhang, J. Fu, G. Zhang and X. Zhang, Enhanced arsenate removal by novel Fe–La composite (hydr)oxides synthesized via coprecipitation, Chem. Eng. J., 2014, 251, 69–79.
26. Z. Marková, K. n. M. Šišková, J. Filip, J. Čuda, M. Kolář, K. r. Šafářová, I. Medřík and R. Zbořil, Air stable magnetic bimetallic Fe–Ag nanoparticles for advanced antimicrobial treatment and phosphorus removal, Environ. Sci. Technol., 2013, 47, 5285–5293.
27. W. Al-Saidi, H. Feng and K. Fichthorn, Adsorption of polyvinylpyrrolidone on Ag surfaces: Insight into the workings of a structure-directing agent, Nano Lett., 2012, 12, 997–1001.
28. M. Ghaedi, B. Sadeghian, A. A. Pebdani, R. Sahraei, A. Daneshfar and C. Duran, Kinetics, thermodynamics and equilibrium evaluation of direct yellow 12 removal by adsorption onto silver nanoparticles loaded activated carbon, Chem. Eng. J., 2012, 187, 133–141.
29. W. Zhang, J. Fu, G. Zhang and X. Zhang, Enhanced arsenate removal by novel Fe–La composite (hydr)oxides synthesized via coprecipitation, Chem. Eng. J., 2014, 251, 69–79.
30. S. Luo, S. Yang, C. Sun and J.-D. Gu, Improved debromination of polybrominated diphenyl ethers by bimetallic iron–silver nanoparticles coupled with microwave energy, Sci. Total Environ., 2012, 429, 300–308.
31. Z. Zheng, S. Yuan, Y. Liu, X. Lu, J. Wan, X. Wu and J. Chen, Reductive dechlorination of hexachlorobenzene by Cu/Fe bimetal in the presence of nonionic surfactant, J. Hazard. Mater., 2009, 170, 895–901.
32. Z. Markova, K. Siskova, J. Filip, J. Cuda, M. Kolar, K. Safarova, I. Medrik and R. Zboril, Air Stable Magnetic Bimetallic Fe–Ag Nanoparticles for Advanced Antimicrobial Treatment and Phosphorus Removal, Environ. Sci. Technol., 2013, 47, 5285–5293.
33. J. P. T. Baú, C. E. Carneiro, I. G. de Souza Junior, C. M. de Souza, A. C. da Costa, E. di Mauro, C. T. Zaia, J. Coronas, C. Casado and H. de Santana, Adsorption of adenine and thymine on zeolites: FT-IR and EPR spectroscopy and X-ray diffractometry and SEM studies, Origins Life Evol. Biospheres, 2012, 42, 19–29.
34. K. J. Carroll, D. M. Hudgins, S. Spurgeon, K. M. Kemner, B. Mishra, M. I. Boyanov, L. W. Brown III, M. L. Taheri and E. E. Carpenter, One-pot aqueous synthesis of Fe and Ag core/shell nanoparticles, Chem. Mater., 2010, 22, 6291–6296.
35. B. Kakavandi, R. R. Kalantary, M. Farzadkia, A. H. Mahvi, A. Esrafili, A. Azari, A. R. Yari and A. B. Javid, Enhanced chromium(VI) removal using activated carbon modified by zero valent iron and silver bimetallic nanoparticles, Journal of Environmental Health Science and Engineering, 2014, 12, 115.
36. J. Ma, Y. Shen, C. Shen, Y. Wen and W. Liu, Al-doping chitosan–Fe(III) hydrogel for the removal of fluoride from aqueous solutions, Chem. Eng. J., 2014, 248, 98–106.
37. H.-m. Cai, G.-j. Chen, C.-y. Peng, Z.-z. Zhang, Y.-y. Dong, G.-z. Shang, X.-h. Zhu, H.-j. Gao and X.-c. Wan, Removal of fluoride from drinking water using tea waste loaded with Al/Fe oxides: A novel, safe and efficient biosorbent, Appl. Surf. Sci., 2015, 328, 34–44.
38. A. B. Albadarin, Z. Yang, C. Mangwandi, Y. Glocheux, G. Walker and M. Ahmad, Experimental design and batch experiments for optimization of Cr(VI) removal from aqueous solutions by hydrous cerium oxide nanoparticles, Chem. Eng. Res. Des., 2014, 92, 1354–1362.
39. K. Chu, Removal of copper from aqueous solution by chitosan in prawn shell: adsorption equilibrium and kinetics, J. Hazard. Mater., 2002, 90, 77–95.
40. W. Ma, N. Zhao, G. Yang, L. Tian and R. Wang, Removal of fluoride ions from aqueous solution by the calcination product of Mg–Al–Fe hydrotalcite-like compound, Desalination, 2011, 268, 20–26.
41. S. Swain, T. Patnaik, V. Singh, U. Jha, R. Patel and R. Dey, Kinetics, equilibrium and thermodynamic aspects of removal of fluoride from drinking water using meso-structured zirconium phosphate, Chem. Eng. J., 2011, 171, 1218–1226.
42. K. S. Prasad, Y. Amin and K. Selvaraj, Defluoridation using biomimetically synthesized nano zirconium chitosan composite: Kinetic and equilibrium studies, J. Hazard. Mater., 2014, 276, 232–240.
43. S. Swain, T. Patnaik and R. Dey, Efficient removal of fluoride using new composite material of biopolymer alginate entrapped mixed metal oxide nanomaterials, Desalin. Water Treat., 2013, 51, 4368–4378.
44. S. K. Nath and K. G. Bhattacharyya, Adsorption of Arsenite and Fluoride on Untreated and Treated Bamboo Dust, Management of Natural Resources in a Changing Environment, Springer, 2015, pp. 167–180.
45. D. Wan, Y. Liu, S. Xiao, J. Chen and J. Zhang, Uptake fluoride from water by caclined Mg–Al–CO₃ hydrotalcite: Mg/Al ratio effect on its structure, electrical affinity and adsorptive property, Colloids Surf., A, 2015, 469, 307–314.
46. S. Deng, H. Liu, W. Zhou, J. Huang and G. Yu, Mn–Ce oxide as a high-capacity adsorbent for fluoride removal from water, J. Hazard. Mater., 2011, 186, 1360–1366.
47. M. Barathi, A. S. K. Kumar and N. Rajesh, Efficacy of novel Al–Zr impregnated cellulose adsorbent prepared using microwave irradiation for the facile defluoridation of water, J. Environ. Chem. Eng., 2013, 1, 1325–1335.
48. D. Thakre, S. Jagtap, N. Sakhare, N. Labhsetwar, S. Meshram and S. Rayalu, Chitosan based mesoporous Ti–Al binary metal oxide supported beads for defluoridation of water, Chem. Eng. J., 2010, 158, 315–324.
49. A. Ghosh, S. Chakrabarti, K. Biswas and U. C. Ghosh, Agglomerated nanoparticles of hydrous Ce(IV) + Zr(IV) mixed oxide: Preparation, characterization and physicochemical aspects on fluoride adsorption, Appl. Surf. Sci., 2014, 307, 665–676.
50. L. Calvo, M. A. Gilarranz, J. A. Casas, A. F. Mohedano and J. J. Rodriguez, Denitrification of water with activated carbon-supported metallic catalysts, Ind. Eng. Chem. Res., 2010, 49, 5603–5609.
51. O. S. G. Soares, J. J. Órfão and M. F. R. Pereira, Nitrate reduction in water catalysed by Pd–Cu on different supports, Desalination, 2011, 279, 367–374.
52. Y. Yoshinaga, T. Akita, I. Mikami and T. Okuhara, Hydrogenation of nitrate in water to nitrogen over Pd–Cu supported on active carbon, J. Catal., 2002, 207, 37–45.
53. J. Jung, S. Bae and W. Lee, Nitrate reduction by maghemite supported Cu–Pd bimetallic catalyst, Appl. Catal., B, 2012, 127, 148–158.

---