Enhanced fluoride removal by loading Al/Zr onto carboxymethyl starch sodium: synergistic interactions between Al and Zr

Abstract

In this paper, a novel type of adsorbent was prepared by loading Al/Zr onto carboxymethyl starch sodium to generate CMS–Al, CMS–Zr or CMS–Al–Zr. The adsorbents were tested for removal of fluoride by batch adsorption experiments under various conditions. It was found that CMS–Al–Zr performed well over a considerably wide pH range of 4–10. The adsorption process could be described by a Langmuir isotherm model and Lagergren pseudo-second-order kinetic model. The maximum adsorption capacity for CMS–Al–Zr (60.61 mg g⁻¹) was higher than CMS–Al (51.44 mg g⁻¹) and CMS–Zr (40.23 mg g⁻¹) due to synergistic interactions between Al and Zr. SEM, EDS, XRD, FTIR and XPS studies revealed the Al/Zr loading process and mechanism of fluoride adsorption. The results showed that the metal complex was coordinated to the –COO group of CMS through ion exchange of sodium during the loading process.

---

1. Introduction

Fluoride is one of the essential trace elements in the human body. However, excess fluoride intake over a long period of time can lead to dental fluorosis, bone disease, and disruption of normal calcium and phosphorus metabolism.^1,2 Drinking water is the primary route of ingestion of fluoride into the human body. Consequently, the World Health Organization (WHO) has set the acceptable fluoride concentration in drinking water at 0.5–1.5 mg L⁻¹.³ However, over 200 million people worldwide are exposed to water with a fluoride concentration above 1.5 mg L⁻¹.⁴

Although several innovative methods, such as ion-exchange, reverse osmosis and electrodialysis, have been tested for water defluorination. Among these, adsorption technology stand-outs due to its low cost, high efficiency and ease of operation.⁵ Various materials, such as alumina-based adsorbents,⁶ carbon,⁷ agricultural waste,² resins,⁸ zeolite⁹ and mixed metal oxides,¹ have been employed as adsorbents for fluoride removal. Recently, carbohydrate polymer-based adsorbents, including chitosan,^10–12 cellulose,^13–16 and alginate^17–20 have been shown to be very promising adsorbents since these materials are widely available, sometimes as waste resources and are thus environmentally friendly, and possess high combined advantages.

Starch, one of the major polysaccharides in plant cells,²¹ is widely used in industry due to its low unit cost, biodegradability, wide availability and renewability.²² With these factors in mind, the potential of starch to remove metal ions in water treatment has been reported.^22,23 However, the use of starch and its derivatives as biosorbents for fluoride removal has not been reported. Carboxymethyl starch sodium (CMS), a starch derivative, is a polymer with a great importance in pharmacy, food industry, environmental protection, medicine, cosmetics and many other industrial applications,²⁴ which is due to its wide availability, renewability, low price as well as ecological technology of synthesis.²⁵ CMS is full of carboxylate groups, which make them susceptible to reactions with metal cations.²⁶

In this work, three novel adsorbents were synthesized by loading Al, Zr and Al/Zr onto CMS and tested for their abilities to remove fluoride. The synergistic interactions between Al and Zr during the fluoride adsorption were investigated by numerous adsorption experiments under various conditions: adsorbent dose, contact time, initial concentration and pH. Moreover, several isotherm and kinetic models were investigated to evaluate their performance and to understand the adsorption process. The synthesized adsorbents were characterized by SEM, EDS, XRD, FTIR and XPS to reveal the Al/Zr loading process and mechanism of fluoride adsorption.

2. Materials and methods

2.1 Materials

Carboxymethyl starch sodium (CMS; CAS: 9063-38-1), purchased from Aladdin Industrial Corporation (Shanghai, China), was used for adsorbent material without further purification. NaF, NaOH, ZrOCl₂·8H₂O, Al₂(SO₄)₃·18H₂O and HCl were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). All chemical reagents used in the work were analytical grade (A.R.).

2.2 Synthesis of CMS–Al, CMS–Zr and CMS–Al–Zr composites

The CMS–Al, CMS–Zr and CMS–Al–Zr composites were synthesized according to our previous technique with some modifications.⁶ Briefly, aluminium sulfate (0.2 M), zirconium oxychloride (0.4 M) and a mixed solution (0.4 M with Al : Zr ratio of 1 : 3) were prepared in hydrochloric acid (0.01 M). CMS (5.0 g) and deionized water (50 mL) were added to a beaker on a heated magnetic stirrer (300 rpm) at room temperature (25 °C) to a form uniform slurry. The above metal ion solutions (50 mL) were added to the CMS slurries, and the mixture was continuously stirred for 30 min. Then, sodium hydroxide solution (2.0 M) was added dropwise to adjust the solution pH to 5.0. Finally, the suspension was centrifuged and washed several times with deionized water and ethanol, and then dried under vacuum at 50 °C for 12 h.

2.3 Sample characterization

The surface structure and morphology of the sorbents were determined using scanning electron microscopy (SEM, Sirion from FEI), while at the same time the elements of the sorbents was analyzed by Energy Dispersive Spectrometry (EDS, Sirion from FEI). The structures and crystallinity of the sorbents was analyzed by X-ray diffraction (XRD, X'Pert PRO from PHILIPS). The Fourier-transformed infrared spectroscopy (FTIR) pattern of the sorbents was recorded on Nicolet 8700 (Thermo Scientific Instrument Co. USA) in the wavenumber range of 400–4000 cm⁻¹. X-ray photoelectron spectroscopy (XPS, ESCALAB 250 from Thermo-VG Scientific) was used for characterization of chemical bonding environment and adsorption site study. The specific surface area of CMS and CMS–Al–Zr were determined by the automatic surface area analyzer (Tristar II 3020 M from Micromeritics).

2.4 Adsorption experiments and fluoride measurements

All the adsorption experiments were carried out in a centrifuge tube (50 mL) containing 25 mL fluoride solution of known concentration and 0.030 g of sorbent. The centrifuge tubes were shaken in a constant temperature incubator shaker for 6 h at 25 ± 2 °C. Afterwards, the sorbent was separated from solution through centrifugation. The concentration of residual fluoride in the solution was determined by the fluoride-selective electrode 9609 BNWP.²⁷ The effect of various parameters, such as adsorbent dosage (0.4–3.2 g L⁻¹), contact time (1–360 min), initial concentration (5–200 mg L⁻¹) and pH (3.0–12.0), on adsorption capacity of fluoride were carried out by varying one parameter at a time and keeping the other parameters constant. The pH value of the solution was adjusted using 0.1 mol L⁻¹ HCl or NaOH solution. The chemical stability of the Al/Zr was investigated by measuring the concentration of dissolved Zr and Al in the solution after adsorption over a pH range of 3.0–12.0. The concentration of dissolved Al and Zr was measured by inductively coupled plasma emission spectrometer (ICP, iCAP 6300 Series from Thermo Fisher Scientific). The adsorption kinetic experiments were carried out at different time intervals (1–360 min) with 10 mg L⁻¹ initial fluoride concentration (pH 6.7 ± 0.2). The adsorption isotherm experiments were conducted at the initial fluoride concentration from 5 to 200 mg L⁻¹ (pH 6.7 ± 0.2).

The percent adsorption (%) and the amount of F⁻ adsorbed per unit weight of adsorbent (q_e, mg g⁻¹) were calculated by eqn (1) and (2), respectively.

where C₀ and C_e are the initial and equilibrium fluoride concentrations (mg L⁻¹), respectively, V is the volume of fluoride solution (mL) and m is the weight of the adsorbent (g).

3. Results and discussion

3.1 Characterization of the adsorbents

The CMS granules had the expected starch structure of a smooth, clear surface (Fig. 1), similar to that observed for other starch.²¹ Compared to the CMS granules, the surface of CMS–Al–Zr was covered by an irregular inorganic layer with many pores and abundant protuberances, and an increased surface area, which would facilitate diffusion of fluoride into the CMS–Al–Zr composite for its adsorption.^2,20 The BET surface area of sorbent increased from 0.1761 to 7.3553 m² g⁻¹ after loading process, due to Al and Zr loading. Energy Dispersive Spectrometer (EDS) spectra were carried out to further confirm the effective loading of Zr and Al ions into the CMS. In CMS, the elemental peaks of C, O, Na and Cl at energy values of 0.25, 0.5, 1.05 and 2.6 keV, respectively, were observed (Fig. 2a). After loading with Al and/or Zr, as shown in Fig. 2b–d, the Na peaks (1.05 keV) were drastically suppressed or absent and new peaks related to Al and Zr were observed at energy values of 1.5, 1.8, 2.1 and 2.35 keV, indicating that Al and Zr ions were loaded into the CMS through ion exchange of the Na of CMS during the loading process.^14,18,28 After adsorption experiments, a peak for the fluoride element could be observed (Fig. 2e), demonstrating that fluoride was adsorbed by CMS–Al–Zr.

Fig. 1 SEM images of CMS (a) and CMS–Al–Zr (b).

Fig. 2 EDS spectra of CMS (a), CMS–Al (b), CMS–Zr (c), CMS–Al–Zr (d) CMS–Al–Zr–F (e) and X-ray diffraction (XRD) for CMS–Al, CMS–Zr and CMS–Al–Zr (f).

The white powders of the adsorbents were characterized by XRD analysis in order to verify their phase identification and structural details (Fig. 2f). No distinct crystalline peak was detected in the XRD patterns, which revealed a clear amorphous structure of the adsorbents. An amorphous material structure could give rise to a greater specific surface area and more available active sites, and consequently render the material a good adsorbent.^2,29

3.2 Batch adsorption experiments

3.2.1 Effect of adsorbent dose on fluoride removal. The ability of the adsorbents to remove fluoride at different doses, ranging from 0.4 to 3.2 g L⁻¹, was tested at pH 6.7 ± 0.2 for 6 h (Fig. 3a). For a fixed fluoride concentration (10 mg L⁻¹), increasing the added adsorbent (CMS–Al, CMS–Zr or CMS–Al–Zr) from 0.4 g L⁻¹ to 0.8 g L⁻¹ increased the percentage of removed fluoride, from 66.63% to 72.90%, 50.76% to 81.43% and 79.26% to 89.53%, respectively. This might be attributable to increased surface area and greater number of surface active sites at higher does.^7,30 Thereafter, there was no significant change in fluoride removal after 0.8, 1.2 or 0.8 g L⁻¹ for CMS–Al, CMS–Zr and CMS–Al–Zr, respectively. This might be attributable to a decreased concentration of remaining fluoride available and to overlapping of active sites at higher adsorbent dose.^2,7 In order to take full advantage of the adsorbents and yet conserve resources, a 0.8 g L⁻¹ adsorbent concentration was selected as the optimum dose for further study.

Fig. 3 Effect of the adsorbent dose (a), contact time (b), initial fluoride concentration (c) and initial pH (d) on the adsorption capacity of CMS–Al, CMS–Zr and CMS–Al–Zr adsorbents.

3.2.2 Effect of contact time on fluoride removal and adsorption kinetics. The effect of contact time on fluoride adsorption capacity of CMS–Al, CMS–Zr and CMS–Al–Zr was tested (Fig. 3b) by keeping adsorbent dose at 0.8 g L⁻¹, pH at 6.7 ± 0.2 and the concentration of fluoride at 10 mg L⁻¹ at 25 ± 2 °C. The adsorption of fluoride by CMS–Al, CMS–Zr and CMS–Al–Zr occurred quickly and mostly within the first thirty minutes, which might be due to availability of a large number of adsorption sites at the initial fluoride adsorption process.^31,32 The adsorption process slowed down in the following period, which could be attributed to saturation of active sites on the face of adsorbents.³³ In addition, 85%, 89% and 82% of the equilibrium adsorption of fluoride had taken place in the first 8 min for CMS–Al, CMS–Zr and CMS–Al–Zr. The equilibrium was reached within 60 min, thus the rate of adsorption was considered fast.^34,35

To better express the mechanism of fluoride adsorption onto the surface of CMS–Al, CMS–Zr and CMS–Al–Zr, the contact time data was interpreted by two types of kinetic models:³⁶ Lagergren pseudo-first-order³⁷ and pseudo-second-order kinetic models,³⁸ as shown below:

Lagergren pseudo-first-order kinetic:

Lagergren pseudo-second-order kinetic:

where q_t and q_e denote the amount of adsorbed fluoride by absorbent at time t and at equilibrium state (mg g⁻¹), respectively; k₁ and k₂ are the rate constant of the pseudo-first-order and pseudo-second-order kinetic model, respectively.

The kinetic plots and analysis of fluoride adsorption by CMS–Al, CMS–Zr and CMS–Al–Zr along with correlation coefficients were shown in Fig. S1a and b† and Table 1. According to the values of the regression coefficients (R²), fluoride adsorption by CMS–Al, CMS–Zr and CMS–Al–Zr could be better described by the Lagergren pseudo-second-order kinetic model (R² > 0.99) rather than the pseudo-first-order kinetic model (R² < 0.9). Furthermore, the calculated theoretical adsorption capacity of CMS–Al, CMS–Zr and CMS–Al–Zr (i.e., 9.04, 9.42, 11.05 mg g⁻¹) were all close to the result-based experimental capacities (i.e., 9.03, 9.47, 11.05 mg g⁻¹), indicating that the fluoride adsorption might be via chemical adsorption or chemisorption steps involving valence forces through exchanging or sharing electrons between fluoride and adsorbents.^2,31,39,40

Table 1 Lagergren pseudo-first-order and pseudo-second-order kinetic models and the constants for adsorption of fluoride on adsorbents

Adsorbents	Experimental capacity (mg g^−1)	The pseudo-first-order kinetic model			The pseudo-second-order kinetic model
		k1 (min^−1)	qe (mg g^−1)	R^2	k2 (g mg^−1 min^−1)	qe (mg g^−1)	R^2
CMS–Al	9.03	0.01589	1.367	0.7988	0.09476	9.04	0.9999
CMS–Zr	9.47	0.01946	1.278	0.8169	0.1438	9.42	1.0000
CMS–Al–Zr	11.05	0.01946	1.625	0.6405	0.1975	11.05	1.0000

---

3.2.3 Effect of initial concentration on fluoride removal and adsorption isotherms. The effect of initial concentration of fluoride, ranging from 5 to 200 mg L⁻¹, on the adsorption capacities of CMS–Al, CMS–Zr and CMS–Al–Zr was tested by keeping the adsorbent dose at 0.8 g L⁻¹, pH at 6.7 ± 0.2 and contact time at 6 h at 25 ± 2 °C (Fig. 3c). The adsorption capacities of the adsorbents increased with the increase in initial fluoride concentration, but almost leveled off at high fluoride concentration. The low adsorption capacity at low fluoride concentration might be due to the fact that active sites on the surface of adsorbents had not reached saturation.⁴⁰ At higher initial fluoride concentration, the increase in adsorption capacity resulted primarily from the increased amount of fluoride available for adsorption and a more intensive interaction between the adsorbent and fluoride.¹⁸ The adsorption capacity reached a plateau, which could be attributed to saturation of active sites on the surface of adsorbents.² CMS–Al–Zr showed the highest adsorption across all fluoride concentrations, while CMS–Al and CMS–Zr showed a similar but much lower increase in the adsorption capacity at the same initial fluoride concentration.

An adsorption isotherm is considered essential for determining the specific relationship between adsorbent and adsorbate and predicting the maximum amount of fluoride that can be adsorbed, which is one of the most important parameters to know for an adsorption process.³¹ In this study, two typical isotherm models, namely Langmuir isotherm model⁴¹ and Freundlich isotherm model,⁴² were applied to simulate fluoride adsorption, using the mathematical forms eqn (5) and (6).

where q⁰ (mg g⁻¹) is the maximum amount of the fluoride ion corresponding to complete monolayer coverage; b (L mg⁻¹) is the adsorption equilibrium constant which relates to the adsorption energy; C_e (mg L⁻¹) is the concentration at equilibrium; q_e (mg g⁻¹) is adsorption capacity at equilibrium; k and 1/n represent isotherm constant for Freundlich that relates to adsorption capacity and degree of favorability, respectively.

The experimental fluoride adsorption data for CMS–Al, CMS–Zr and CMS–Al–Zr were plotted by both Langmuir and Freundlich isotherm models (Fig. S1c and d†). The calculated isotherm parameters along with regression coefficients (R²) obtained from fitting fluoride adsorption to the Langmuir and Freundlich isotherms were listed in Table 2. As depicted in Table 2, the Langmuir isotherm model (R² > 0.976) fitted better with the fluoride adsorption data than did the Freundlich isotherm model (R² < 0.970). This fit indicated that the adsorption of fluoride by CMS–Al, CMS–Zr and CMS–Al–Zr was by monolayer coverage on the surface of the adsorbents rather than Freundlich heterogeneous surface adsorption.⁹ In addition, a dimensionless parameter of the Langmuir isotherm model, called separation factor (R_L), is defined as 1/(1 + bC₀). The R_L values studied at different initial fluoride concentration, ranging from 5 to 200 mg L⁻¹, were given in Table S1.† The values of R_L all fall between 0 and 1, suggesting that the adsorption of fluoride was favorable.^10,13

Table 2 Langmuir and Freundlich isotherm constants for adsorption of fluoride on CMS–Al, CMS–Zr and CMS–Al–Zr adsorbents

Adsorbents	Langmuir isotherm			Freundlich isotherm
	q^0 (mg g^−1)	b (L mg^−1)	R^2	n	K (mg g^−1)	R^2
CMS–Al	51.44	0.02041	0.9924	1.659	2.165	0.9183
CMS–Zr	40.23	0.02356	0.9816	2.128	3.077	0.9670
CMS–Al–Zr	60.61	0.01944	0.9766	1.780	2.944	0.9360

---

Finally, the maximum monolayer adsorption capacity (q⁰), calculated from the Langmuir isotherm model (Table 2), for CMS–Al–Zr was 60.61 mg g⁻¹, which was much higher than those of CMS–Al (51.44 mg g⁻¹) and CMS–Zr (40.23 mg g⁻¹), indicating that a synergistic interaction between Al and Zr occurred during the adsorption process.^6,43 The adsorption capacities for fluoride of CMS–Al, CMS–Zr and CMS–Al–Zr from this study were compared to previously reported carbohydrate polymer-based adsorbents. As seen in Table S2,† CMS–Al–Zr had a much bigger adsorption capacity than the majority of the carbohydrate based adsorbents.

3.2.4 Effect of pH on fluoride removal. Both the surface properties and the distribution in solution of the adsorbents may be influenced by the initial pH via protonation and deprotonation of host layer cations.⁴⁴ Fluoride removal has been reported to be highly dependent on solution pH. Thus, the effect of initial pH on fluoride removal by CMS–Al, CMS–Zr and CMS–Al–Zr was tested from pH 3.0–11.0. The result showed that pH of the solution had a significant effect on fluoride removal (Fig. 3d). CMS–Zr recorded the minimum adsorption capacity of 1.68 mg g⁻¹ at pH 12, followed by a gradually increasing capacity with decreasing pH, to a maximum adsorption capacity of 12.02 mg g⁻¹ at pH 3. Other adsorbents modified by zirconium were reported to have a similar decreasing trend with increasing solution pH.^8,11,45

CMS–Al had the stable fluoride adsorption capacity from pH from 5 to 10. Above pH 11 or below pH 3 resulted in a significant drop in adsorption capacity. The adsorption capacity of CMS–Al–Zr showed a same tendency in the examined pH range, although the capacity was both higher and more stable. The fluoride adsorption capacity of CMS–Al–Zr increased as the pH increased from 2 to 4, remained almost constant in the pH range of 4–10, and thereafter sharply decreased at pH from 10 to 12. In an acidic solution, the adsorption capacity of CMS–Al–Zr declined, likely due not only to the leakage of metal ions from the adsorbent,⁴⁶ but also to the preferential formation of weakly ionized hydrofluoric acid and AlF_x soluble species.^5,47 In alkaline solution, fluoride removal was likely inhibited by deprotonation of CMS–Al–Zr through electrostatic repulsion and competition between anions for the active sites of the adsorbent.^1,2 The pH of drinking and natural water normally falls in the range of 6.0–8.0, thus CMS–Al–Zr could be used to defluorinate drinking water without pH adjustment.⁴⁸

It was especially interesting to find that CMS–Al–Zr possessed advantages over both CMS–Zr and CMS–Al, particularly in the pH range experiment. CMS–Al–Zr had a higher adsorption capacity than CMS–Al at any given pH value and a wider active pH range than CMS–Zr. This indicated that composite modification by Al and Zr was worthwhile and successful. The chemical stability of CMS–Al–Zr was assayed under various pH (Fig. S2†). pH had a significant influence on the chemical stability of CMS–Al–Zr. At pH values ranging from 5–10, CMS–Al–Zr resulted in low dissolved Al and Zr concentrations. Thus, the CMS–Al–Zr is stable in the range of 5–10.

3.3 Mechanism study

3.3.1 FTIR study. EDS characterization indicated that Al and Zr ions might be loaded into the CMS through ion exchange of the Na from the CMS during the loading process (in Section 3.1). Thus, FTIR spectra of CMS and CMS–Al–Zr before and after fluoride adsorption were taken to reveal the mechanisms of modification and adsorption (Fig. 4). The strong characteristic stretching vibration peaks of unmodified CMS carboxylate (–COO–) appeared around 1605 and 1422 cm⁻¹.^24,49 After loading, the peaks of carboxylate shifted to 1635 and 1423 cm⁻¹, suggesting the formation of a new band between metal ions and carboxylate.⁵⁰ A broad bank at 3200–3600 cm⁻¹ was assigned to the stretching model of banded –OH groups.⁵¹ Compared to CMS, CMS–Al–Zr displayed an apparently stronger peak, which was due to the more –OH groups on the surface of adsorbent after loading.⁵¹ The broad band at 400–800 cm⁻¹ was metal–oxygen bonds.⁵² The change of stretching frequency of carboxyl functional group in CMS–Al–Zr compared to CMS confirmed the chemical modification.⁴⁶ After fluoride adsorption, the banks at 1635 and 1423 cm⁻¹ were shifted to 1629 and 1420 cm⁻¹, respectively, indicating the adsorption of fluoride by CMS–Al–Zr was via chemical adsorption.¹⁴

Fig. 4 FTIR spectra of CMS and CMS–Al–Zr before and after fluoride adsorption.

3.3.2 XPS analyses. In order to further reveal the mechanism of carbohydrate modification and fluoride adsorption, variations in the chemical environment of the adsorbents were analyzed by XPS (Fig. S3† and 5). New peaks were observed in CMS–Al, CMS–Zr and CMS–Al–Zr samples, which were attributed to Al and Zr. Furthermore, in CMS–Al–Zr sample the peaks of Al 2p and Zr 3d shifted from 74.82 eV and 182.48 eV to 74.62 eV and 182.74 eV in comparison to CMS–Al and CMS–Zr, respectively. This indicated that a bimetallic Al–O–Zr bond might form during the loading process.^6,53 Thus, CMS–Al–Zr might not consist of a simple mixture of hydrous aluminium oxide and hydrous zirconium oxide on –COO groups.⁵⁴

Fig. 5 XPS spectra of the Al 2p (a), Zr 3d (b), F 1s (c) of adsorbents and F 1s division by XPS peak software (d).

After fluoride adsorption, the peaks of Al 2p and Zr 3d shifted from 74.62 eV and 182.74 eV to 74.49 eV and 182.56 eV, suggesting that both Al and Zr participated in the defluorination process.⁶ Additionally, after fluoride adsorption a new F 1s peak was observed at 684.86 eV (Fig. S3d† and 5c), which was higher than the NaF (684.5 eV), clearly indicating that fluoride adsorption was mainly via chemical adsorption or chemisorptions between the fluoride and CMS–Al–Zr.² The F (1s) peak was divided into four main peaks by the XPS PEAK software (Fig. 5d). The peaks at 684.5, 685.3 and 686.2 could be attributed to Na–F, Zr–F and Al–F, respectively.⁵⁵ According to the above analysis, the unknown peak at 687 eV might result from the interaction between fluoride and the bimetallic Al–O–Zr bond, indicating that a synergistic interaction between Zr and Al occurred during the defluorination process. This synergy would give CMS–Al–Zr its higher adsorption capacity.^6,54

Careful consideration of the evidence shown in this work allowed the development of proposed mechanisms of metal ion loading onto CMS and fluoride of adsorption by CMS–Al–Zr in Fig. 6.^14,34 In the first step, hydrolysis of metal ions (M) led to a complex formation surrounded by water molecules and –OH groups. During the loading process, the metal complex then became coordinated to the –COO group of CMS through ion exchange of the Na of CMS. When adsorbent was added to a fluoride solution, a final reaction occurred between the metal ion and fluoride anion. In the lower pH range, the surface of adsorbents were covered in positively charged sites, and then the negatively charged fluoride ions tended to attack them to form a strong covalent chemical bond. In the higher pH range, fluoride adsorption could occur by a ligand-exchange reaction between fluoride ions and hydroxyl.

Fig. 6 The proposed mechanisms of Al and Zr loading onto CMS and of fluoride adsorption.

4. Conclusions

In this study, a series of CMS-based adsorbents were fabricated, characterized and used to remove fluoride from aqueous solution. The metal ion became coordinated to a –COO group of CMS through ion exchange of Na of CMS during the loading process. Solution pH played an important role in fluoride removal, with CMS–Al–Zr possessing efficient fluoride removal over a wide pH range, from 4.0 to 10.0. The adsorption kinetics study showed that the Lagergren pseudo-second-order kinetic model could describe the fluoride adsorption process and that adsorption reached equilibrium within 1 h. Furthermore, we found that adsorbents followed the Langmuir isotherm model. The maximum adsorption capacities for CMS–Al–Zr (60.61 mg g⁻¹) were higher than other reported carbohydrate polymer-based adsorbents. Results of the present study demonstrated that novel adsorbents based on CMS promise to be efficient fluoride scavengers for drinking water treatment.

Acknowledgements

Financial support from the Earmarked Fund for Modern Agro-industry Technology Research System in Tea Industry (CARS-23, the Ministry of Agriculture of P. R. China), Changjiang Scholars and Innovative Research Team in University (IRT1101) and Natural Science Foundation of Anhui Province (1408085MKL38) funded this study.

Notes and references

1. T. Zhang, Q. Li, H. Xiao, Z. Mei, H. Lu and Y. Zhou, Appl. Clay Sci., 2013, 72, 117–123.
2. H. Cai, G. Chen, C. Peng, L. Xu, X. Zhu, Z. Zhang, Y. Dong, G. Shang, F. Ke, H. Gao and X. Wan, RSC Adv., 2015, 5, 29266–29275.
3. S. Jagtap, M. K. Yenkie, N. Labhsetwar and S. Rayalus, Chem. Rev., 2012, 112, 2454–2466.
4. D. Mohan, R. Sharma, V. K. Singh, P. Steele and C. U. Pittman Jr, Ind. Eng. Chem. Res., 2012, 51, 900–914.
5. A. Bhatnagar, E. Kumar and M. Sillanpaa, Chem. Eng. J., 2011, 171, 811–840.
6. H. M. Cai, G. M. Chen, C. M. Peng, Z. Z. Zhang, Y. Y. Dong, G. Z. Shang, X. H. Zhu, H. J. Gao and X. C. Wan, Appl. Surf. Sci., 2015, 328, 34–44.
7. A. K. Yadav, R. Abbassi, A. Gupta and M. Dadashzadeh, Ecol. Eng., 2013, 52, 211–218.
8. B. Pan, J. Xu, B. Wu, Z. Li and X. Liu, Environ. Sci. Technol., 2013, 47, 9347–9354.
9. Y. Sun, Q. Fang, J. Dong, X. Cheng and J. Xu, Desalination, 2011, 277, 121–127.
10. N. Viswanathan, C. S. Sundaram and S. Meenakshi, Colloids Surf., B, 2009, 68, 48–54.
11. N. Viswanathan and S. Meenakshi, J. Hazard. Mater., 2010, 176, 459–465.
12. Q. Liu, L. Zhang, B. Yang and R. Huang, Int. J. Biol. Macromol., 2015, 77, 15–23.
13. Y. Tian, M. Wu, R. Liu, D. Wang, X. Lin, W. Liu, L. Ma, Y. Li and Y. Huang, J. Hazard. Mater., 2011, 185, 93–100.
14. J. Wang, X. Lin, X. Luo and Y. Long, Chem. Eng. J., 2014, 252, 415–422.
15. J. Wang, X. Lin, X. Luo and W. Yao, RSC Adv., 2015, 5, 59273–59285.
16. M. Barathi, A. S. K. Kumar and N. Rajesh, Ultrason. Sonochem., 2014, 21, 1090–1099.
17. S. M. Prabhu and S. Meenakshi, Carbohydr. Polym., 2015, 120, 60–68.
18. Q. Zhou, X. Lin, J. Qian, J. Wang and X. Luo, RSC Adv., 2015, 5, 2100–2112.
19. C. Gao, X. Y. Yu, T. Luo, Y. Jia, B. Sun, J. H. Liu and X. J. Huang, J. Mater. Chem. A, 2014, 2, 2119–2128.
20. K. Pandi and N. Viswanathan, Carbohydr. Polym., 2014, 112, 662–667.
21. W. Yanli, G. Wenyuan and L. Xia, Carbohydr. Res., 2009, 344, 1764–1769.
22. D. K. Kweon, J. K. Choi, E. K. Kim and S. T. Lim, Carbohydr. Polym., 2001, 46, 171–177.
23. D. Soto, J. Urdaneta, K. Pernía, O. León, A. Muñoz-Bonilla and M. Fernández-García, Polym. Adv. Technol., 2015, 26, 147–152.
24. K. Wilpiszewska, A. K. Antosik and T. Spychaj, Carbohydr. Polym., 2015, 128, 82–89.
25. M. Zdanowicz and T. Spychaj, J. Appl. Polym. Sci., 2014, 131, 407193.
26. B. Grabowska, M. Sitarz, E. Olejnik and K. Kaczmarska, Spectrochim. Acta, Part A, 2015, 135, 529–535.
27. G. J. Chen, C. Y. Peng, J. Y. Fang, Y. Y. Dong, X. H. Zhu and H. M. Cai, Desalin. Water Treat., 2015 DOI:10.1080/19443994.2015.1049959.
28. H. Paudyal, B. Pangeni, K. N. Ghimire, K. Inoue, K. Ohto, H. Kawakita and S. Alam, Chem. Eng. J., 2012, 195, 289–296.
29. V. Ganvir and K. Das, J. Hazard. Mater., 2011, 185, 1287–1294.
30. V. Tomar, S. Prasad and D. Kumar, Microchem. J., 2014, 112, 97–103.
31. Q. Zhou, X. Lin, B. Li and X. Luo, Chem. Eng. J., 2014, 256, 306–315.
32. T. Wu, L. Mao and H. Wang, RSC Adv., 2015, 5, 23246–23254.
33. N. Thakur, S. A. Kumar, H. Parab, A. K. Pandey, P. Bhatt, S. D. Kumar and A. V. R. Reddy, RSC Adv., 2014, 4, 10350–10357.
34. L. Halla Velazquez-Jimenez, R. H. Hurt, J. Matos and J. Rene Rangel-Mendez, Environ. Sci. Technol., 2014, 48, 1166–1174.
35. S. K. Swain, T. Patnaik, V. K. Singh, U. Jha, R. K. Patel and R. K. Dey, Chem. Eng. J., 2011, 171, 1218–1226.
36. N. Rahman, U. Haseen and M. F. Khan, RSC Adv., 2015, 5, 39062–39074.
37. S. Lagergren, K. Sven. Vetenskapsakad. Handl., 1898, 24, 1–39.
38. Y. S. Ho, Water Res., 2006, 40, 119–125.
39. Z. Hussain, L. Daosheng, L. Xi and K. Jianxiong, RSC Adv., 2015, 5, 43906–43916.
40. S. Suriyaraj, A. Bhattacharyya and R. Selvakumar, RSC Adv., 2015, 5, 26905–26912.
41. I. Langmuir, J. Am. Chem. Soc., 1916, 38, 2221–2295.
42. H. Freundlich, Z. Phys. Chem., 1906, 385–470.
43. E. Tchomgui-Kamga, V. Alonzo, C. P. Nanseu-Njiki, N. Audebrand, E. Ngameni and A. Darchen, Carbon, 2010, 48, 333–343.
44. T. Zhang, Q. Li, H. Xiao, H. Lu and Y. Zhou, Ind. Eng. Chem. Res., 2012, 51, 11490–11498.
45. X. Dou, D. Mohan, C. U. Pittman Jr and S. Yang, Chem. Eng. J., 2012, 198, 236–245.
46. H. Paudyal, B. Pangeni, K. Inoue, H. Kawakita, K. Ohto, H. Harada and S. Alam, J. Hazard. Mater., 2011, 192, 676–682.
47. G. Zhang, Z. He and W. Xu, Chem. Eng. J., 2012, 183, 315–324.
48. N. Zhang, X. Yang, X. Yu, Y. Jia, J. Wang, L. Kong, Z. Jin, B. Sun, T. Luo and J. Liu, Chem. Eng. J., 2014, 252, 220–229.
49. H. Shi, Y. Yin and S. Jiao, J. Appl. Polym. Sci., 2014, 131, 40906.
50. K. Pandi and N. Viswanathan, Carbohydr. Polym., 2015, 118, 242–249.
51. D. Yang, Y. Li, Y. Wang and Z. Jiang, RSC Adv., 2014, 4, 49811–49818.
52. J. Zhang, N. Chen, Z. Tang, Y. Yu, Q. Hu and C. Feng, Phys. Chem. Chem. Phys., 2015, 17, 12041–12050.
53. S. Zhang, Y. Lu, X. Lin, X. Su and Y. Zhang, Appl. Surf. Sci., 2014, 303, 1–5.
54. L. Chen, B.-Y. He, S. He, T.-J. Wang, C.-L. Su and Y. Jin, Powder Technol., 2012, 227, 3–8.
55. B. Dzhurinskii, D. Gati, N. Sergushin, V. Nefedov and Y. V. Salyn, Russ. J. Inorg. Chem., 1975, 20, 2307–2314.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra18167d

---