Albumin-functionalized magnetic nanoparticles as an efficient sorbent for removal of Pb(II), Cd(II), Cu(II) and Cr(VI) ions from aqueous solutions

Abstract

Functionalized magnetic nanoparticles are increasingly attracting interest as a new and efficient sorbent for metallic contaminant elimination in environmental samples. Albumin modified magnetic nanoparticles (AMMNPs) were synthesized and characterized. This sorbent was used for removal of cationic and anionic species from the aqueous solutions. The sorbent showed a high adsorption capacity and short equilibrium time for Pb²⁺, Cu²⁺, Cd²⁺ and CrO₄²⁻ ions removal from aqueous solutions. The effects of different parameters on the adsorption processes were investigated and isotherm curves were constructed. This study showed that AMMNPs are a proper adsorbent for the removal of the studied hazardous ions and can also be used as an efficient sorption material for preconcentration of the studied ions in trace analyses. The adsorption process was very fast and adsorbent performance was not affected clearly by the presence of other ionic species.

---

1. Introduction

Metallic pollutants are a serious threat to the environment, food chain and human health. Their biological behavior and toxic effects can cause many problems in the areas of agriculture and health and to the environment all over the world.¹ It is therefore obvious that the study of metallic pollutant removal is necessary. There have been reported relatively large numbers of scientific studies for eliminating metal ions from water and wastewaters, including chemical precipitation,² reverse osmosis,³ electrochemical reduction,⁴ ion exchange⁵ and adsorption.⁶

Among these methods, the use of adsorbents is very attractive because of their low cost and high efficiency. The common adsorbents are activated carbon,⁷ carbon nanotubes,⁸ charcoal,⁹ clays,¹⁰ zeolites,¹¹ composite materials,^12,13 biomass,¹³ nanomaterials¹⁴ and polymeric materials.¹⁵ The efficiency of an adsorbent depends on the specific surface area, number of adsorption sites and chemical stability.¹⁶

Nanomaterials have unique features, and because of their small size, they possess a high surface area to volume ratio.¹⁷ Magnetic nanoparticles (MNPs) are a prevalent type of sorbent, which are widely used all over the world. These particles have a low cost, show sufficient magnetic properties and can be easily modified without altering their properties significantly. These features make them adsorbents with a great number of active sites and high efficiency. The most important advantage of using magnetic nanoparticles is manipulation simplicity.¹⁸

Albumin is the most abundant protein in mammalian blood serum, which is involved in the transport of hormones, fatty acids, small molecules and metal ions.¹⁹ This protein can adsorb metals because of their structural functional groups. It is expected that HSA shows a strong affinity towards transition metal ions because of the existence of amino acid residues such as histidine and cysteine.²⁰ These amino acids have S and N atoms, which bind effectively with transition metal ions. In fact, a physical adsorption can also exist because of abovementioned functional groups in the albumin structure. In this study, our aim is the synthesis and characterization of the albumin modified magnetic nanoparticles (AMMNPs) and investigation of their ability to removal of the selected metal ions from aqueous solutions. We also intend to introduce AMMNPs as an effective and reliable adsorbent for preconcentration of the selected ions in trace analyses.

2. Materials and method

2.1. Materials

Hydrochloric acid, ethanol, sodium hydroxide, ferric chloride hexahydrate, ferrous chloride tetrahydrate, 25% ammonia solution, potassium dichromate, lead(II) nitrate, cadmium nitrate, and copper(II) nitrate were obtained from Merck (Germany). Amino propyl triethoxysilane (APTES) and tetra ethyl orthosilicate (TEOS) were purchased from Fluka (USA). A 25% human serum albumin (HSA) solution was prepared from Kedrion (Lukka, Italy).

2.2. Apparatus

Characterization of the sorbents was performed using a Jasco, 6300 Fourier transform infrared spectrometer (FT-IR) in the range of 400–4000 cm⁻¹. The superparamagnetic properties of the magnetic particles were verified by the magnetization curve measured by the Alternating Gradient Force Magnetometer (AGFM). A Vega Tescan-LMU (Brno, Kohoutovice, Czech Republic) scanning electron microscope (SEM) was used to study the nanoparticles surface. The atomic absorption spectrometer was a Shimadzu, AA-670 and it was used with an air/acetylene flame for measuring all the investigated cations.

2.3. Synthesis of silica coated magnetic nanoparticles

The silica coated magnetite nanoparticles were obtained by a co-precipitation method.²¹ In a conventional synthesis, 12.3 g of FeCl₃·6H₂O and 4.60 g of FeCl₂·4H₂O were dissolved in 800 mL deionized water in a round bottomed flask under nitrogen with mechanical stirring at 80 °C. Then, 20% of 30 mL NH₃ solution was added dropwise into the solution and a color change was observed immediately. The product was rinsed with deionized water twice. The resulting nanoparticles were dispersed in deionized water and then appropriate amounts of TEOS were added, and the pH value was increased to 12–13 by adding 25% ammonia solution while it was stirred with a mechanical stirrer. The suspension was sonicated at 85 °C for 30 min, and the pH was reduced to 6 by HCl. The synthesized product was washed three times with deionized water, and a brown precipitate was achieved.

2.4. APTES functionalized silica coated MNPs

The surface of the silica coated magnetic nanoparticles was activated by using APTES and glutaraldehyde for HSA immobilization.²¹ 4 g of the magnetic nanoparticles prepared above was dispersed into 250 mL methanol in a flask; then 7 mL APTES and 3 mL deionized water was added dropwise to this solution and kept in an ultrasonic for 30 min. Subsequently, 150 mL glycerol was mixed into the solution by mechanical stirring at 90 °C for 3 h. The products were collected by a permanent magnet and washed with methanol–water (50–50) solution three times. Then, 20 mL glutaraldehyde and 0.1 M phosphate buffer were added to the flask followed by sonication for 6 h for surface activation. The collected nanoparticles were washed several times with deionized water to remove excess glutaraldehyde and other remaining reagents.

2.5. Preparation of MNP functionalized with albumin

HSA immobilized magnetic nanoparticles were synthesized with the process as follows: 10 mL HSA solution was added to the phosphate buffer solution in a round-bottomed flask. Then, 5 g of treated magnetic nanoparticles was immersed in this solution; after mixing for 4 h under vigorous stirring, the resultant materials were separated by a magnet and washed several times with deionized water.

2.6. Adsorption experiments

All the adsorption experiments were carried out using a batch adsorption technique. To study adsorption phenomena, a microcentrifuge tube was used and the volume of solution fixed at 1 mL. To perform the experiment the metal solutions were prepared and transferred to microcentrifuge and AMMNPs were added to the solution. To investigate the pH effect, the pH was varied by 0.1 M HCl and NaOH with the temperature maintained at 25 °C. pH measurements were carried out in 10 mL solutions. In all the adsorption experiments, the sorbent dose and ion concentration were fixed at 0.0014 g mL⁻¹ and 10 ppm, respectively.

AMMNPs were dispersed in solution by an ultrasonic bath for 10 min and adsorbent separation was carried out using a magnet. All the experiments were repeated twice.

An atomic absorption spectrophotometer was used to find the residual ion concentrations. To evaluation the effect of each parameter the studied parameters were fixed at the previously mentioned level and others were varied. To compare the adsorption efficiency of adsorbent in different conditions the removal percentage and adsorbed concentration per mass of adsorbent, q (mg g⁻¹) were calculated as follows:²²

where V is the solution volume, W is the mass of the adsorbent, and C_ini and C_e are initial and equilibrium concentration of the cations, respectively.

The isotherms were constructed using Langmuir and Freundlich models. The Langmuir adsorption model assumes that the all adsorption sites of the adsorbent have equal energy, and the adsorbent surface is homogeneous. The linear form of the Langmuir adsorption model is given as follows:

where q_e is the adsorption capacity of metal ions (mg g⁻¹), C_e is the equilibrium concentration (mg L⁻¹) in the liquid phase, q_m is the maximum adsorption capacity (mg g⁻¹) and b is the Langmuir constant (L mg⁻¹).

The dimensionless quantity (f) was calculated to express the adsorption efficiency of the process. The equation is as follows:

where C₀ and b are the initial concentration of ion and Langmuir isotherm constant, respectively. A f-value lower than 1.0 indicates a higher tendency for adsorption.²³

The Freundlich model illustrates the adsorption phenomena based on the heterogeneity of the sorbent surface. In addition, this model considers multilayer adsorption as follows:

q_e = K_fC_e^1/n (3)

where q_e is the amount adsorbed (mg g⁻¹), C_e is the equilibrium concentration (mg L⁻¹), and K_f and 1/n are related to the adsorbent capacity and the sorption intensity of the adsorbent, respectively. Eqn (3) can be linearized as follows:

3. Results and discussion

3.1. Characterization of MNP

3.1.1 FTIR spectroscopy. Fig. 1 shows FT-IR spectra of the silica coated MNPs, APTES modified MNPs and albumin modified MNPs. A vibration band assignment was carried out with the group frequencies of the various analytes present in the sample. As Fig. 1a shows, the presence of MNPs was observed by two strong bands at around 465 and 587 cm⁻¹.²⁴ The Fe–O bond of MNPs was observed at 576 cm⁻¹. Furthermore, an absorption band was observed at around 440 cm⁻¹. The introduction of APTES to the surface of the magnetic NPs was confirmed by the bands at 1115 and 1030 cm⁻¹, which was assigned to the Si–O groups. The band appearing at 1570 cm⁻¹ could be assigned to the NH₂ bending mode of the free NH₂ group (Fig. 2b).²⁵

Fig. 1 FT-IR spectra (A) silica coated MNPs, (B) amino modified MNPs and (C) AMMNPs.

Fig. 2 Magnetization curve of AMMNPs.

The broad peak at 3300 cm⁻¹ was due to the N–H and O–H stretching mode (amide) of albumin and silica. It was observed that asymmetric and symmetric stretching C–H vibrations of methyl and methylene groups were found around 2932–2867 cm⁻¹. The band at 1661 cm⁻¹ was assigned to C=O stretching vibrations (amide), and the vibration band at 1528 cm⁻¹ was assigned to N–H bending vibrations strongly coupled with C–N stretching of albumin (Fig. 2c).^26,27

3.1.2 Magnetic properties of MNPs. Isolation of magnetic nanoparticles sorbents from solutions is easily possible using a magnet, therefore the magnetization power of sorbents is essential. The magnetic properties of modified MNPs were determined and compared with the original nanoparticles. Fig. 2 shows the magnetization curve as a function of perpendicular magnetic field for AMMNPs at room temperature.

The S shaped behavior indicates the super paramagnetic feature of the adsorbents. In the absence of a magnetic field the magnetization value was zero, and the saturation magnetization was at 18.51 emu g⁻¹, this observation implies that these surface modified MNPs have strong magnetic properties and can be isolated conveniently from a matrix solution using an external magnet.²⁸

3.1.3 SEM images. Fig. 3 shows SEM micrographs of (a) Fe MNPs, (b) amino modified silica coated MNPs and (c) AMMNPs. It could be observed that the coating of the sorbents did not change the uniform size distribution. The adsorbents have a spherical and quasi spherical shape and a relatively high surface area. The average size of the MNPs was about 30 nm and after coating and functionalization, the average size was increased to about 85 nm. In addition, Fig. 3b and c clearly shows that the silica coated MNPs functionalization does not result in the change of morphology, but the aggregation of the noncoated MNPs is more evident than that of coated forms.

Fig. 3 SEM micrograph of MNPs (a), amino modified silica coated MNPs (b) and AMMNPs (c).

3.2. Influence of experimental conditions on the adsorption process

3.2.1 Effect of pH. One of the most effective factors of the adsorption phenomena is pH. Fig. 4 shows the significant role of pH on Cd²⁺, Cu²⁺, and Cr(VI) adsorption on the AMMNPs. As reported in the literature, the HSA isoelectric point (pI) was equal to 4.7. The results showed a distinct change in removal percentages at this pH value. The variations arose because of this fact that, potassium dichromate makes several forms of chromium ions in aqueous solutions. These ions, depending on pH of the solution, are chromate (CrO₄²⁻), dichromate (Cr₂O₇²⁻) and hydrogen chromate (HCrO₄⁻).²⁹ In acidic media, the negatively charged HCrO₄⁻ is the dominant specie. The surface functional groups of the adsorbent, such as NH₂ and OH, can be protonated and create positively charged NH₃⁺and OH₂⁺ species and therefore can serve as active sites for electrostatic adsorption of Cr(VI) ions. When pH becomes more than 5, the sorption decreased greatly because of the electrostatic repulsion among the COO⁻ groups and CrO₄²⁻ and Cr₂O₇²⁻ anions. In addition, in alkaline solution OH⁻ competes with chromate and dichromate ions to diffuse into the surface of sorbent. Therefore, the adsorption of these ions was decreased at pH > 5.

Fig. 4 Effect of pH on removal percentage of Cd²⁺, Cu²⁺, Cr(VI) and Pb²⁺.

Cd²⁺ ions were strongly adsorbed in the pH range of 5–7, the surface of the sorbent was positively charged at pH < 5 and the repulsive electrostatic forces became possible between charged species of Cd²⁺ and the charged MNPs surface, especially protonated carboxylic groups; therefore, the ability as a sorbent for Cd²⁺ adsorption was decreased, whereas the decrease in adsorption at higher pH may be attributed to the hydrolysis of cadmium ions and also because of competition with hydroxyl groups with the ions. The active sites for copper binding to albumin are N-terminal amines. As shown in Fig. 4, the maximum Cu²⁺ adsorption occurred at low pH (pH = 4) and the removal percentage was decreased concurrently with pH enhancement. When pH was more than 5, the adsorption efficiency was decreased dramatically. It was expected that increasing the pH would cause an increased adsorption because of the negative charge on N atoms and electrostatic interactions. These results do not imply these facts and this behavior may have been due to the fact that copper precipitation has occurred.³⁰ Varying the pH did not show a distinct change for removal efficiency of Pb²⁺.

3.2.2 Effect of adsorbent dose. To find the proper amounts in all experiments, different amounts of AMMNPs were tested. The effect of adsorbent dose of AMMNPs on the cadmium uptake is shown in Fig. 5. It was evident from this figure that adsorption efficiency increased very rapidly when the sorbent dose was increased up to 0.0014 g mL⁻¹.

Fig. 5 Effect of adsorbent dose on adsorption capacity.

Adsorbent dosage provides more active site and more contact surface area for metal ion adsorption. In contrast, when adsorbent dosage was above 0.0014 g mL⁻¹, the removal percentage variation was slow and therefore adsorption capacity decreased. These behaviors may have been due to adsorption site aggregation that can cause in a decrease in total adsorbent surface area.

3.2.3 Effect of temperature. Fig. 6 presents the effect of temperature on the adsorption behavior of the adsorbent. The removal of Cd²⁺ and Cu²⁺ ions from aqueous solutions decreased dramatically above of 30 centigrade. This can be explained by the fact that in higher temperatures, some chemical bonds will break, and therefore HSA to be denatured and finally it caused a reduction in the available active sites for metal ions.

Fig. 6 Effect of temperature on the removal of Cd²⁺, Cu²⁺, Cr(VI) and Pb²⁺.

The AMMNPs have a higher chromium adsorption capacity above of 30 °C, which reveals the endothermic nature of the adsorption phenomenon. In the case of Pb²⁺, the temperature did not show a sensible adsorption effect.

3.2.4 Effect of ultrasonic time. One of the effective factors for adsorption efficiency is contact time. The effect of different ultrasonic times to find the optimum time was studied. As shown in Fig. 7 for Pb²⁺, Cd²⁺ and Cu²⁺, a rapid kinetic adsorption was observed and the adsorption equilibrium time was at about 8 min. At longer times the removal percentage did not change distinctly.

Fig. 7 The relationship between ultrasonic time and percent removal at 25 °C for Cd²⁺, Cu²⁺, Cr(VI) and Pb²⁺.

The results for Cr(VI) adsorption show different behavior; the optimal time for Cr(VI) adsorption was 4–5 min. At longer contact times the removal efficiency was decreased. The reason of this phenomenon maybe was the rate of desorption of chromium ions. At times longer than 8 min, the chromium ions have enough time for desorption and ion exchange with other ions can be an important factor for decreasing the adsorption efficiency.

3.2.5 Study of interferences. The performance of an adsorbent for industrial and analytical applications needs to be investigated in the presence of interferences. For this purpose, the analyte adsorption in the solutions containing of 1000 ppm NaCl and MgSO₄ were investigated. The obtained results showed that Cu²⁺ and Cr(VI) adsorption were affected due to the presence of other ions and the adsorption amounts were decreased to below 3.6 mg g⁻¹ and 3.2 mg g⁻¹, respectively, whereas removal percentages were about 51% and 46%. Conversely, Pb²⁺ and Cd²⁺ ion adsorption in the presence of co-existing ions was high; the removal percentage for Pb²⁺ was 94.28% and for Cd²⁺ it was about 88.57%. This behavior can be attributed to the fact that interfering ions, such as Mg²⁺ and Na⁺, occupied available sites on the adsorbent and therefore the Cu²⁺ and Cr(VI) ions could not reach to active sites on the surface of adsorbent, while it seems that the active sites for Cd²⁺ and Pb²⁺ ion adsorption were different than the active sites for interfering ions.

3.3. Adsorption isotherms

Langmuir and Freundlich isotherm models were examined to choose the best fitting one to the obtained experimental data and to represent the mechanism of adsorption.

The results in the Tables 1 and 2 show that the data could be fitted better to the Langmuir model than the other isotherm (Fig. 8). The latter fact suggests the adsorption of ions on the surface formed a monolayer. The abovementioned results suggest strongly that the adsorption of metal ions on AMMNs is a chemical type.

Table 1 Langmuir parameters

Adsorbed ion	Regression equation	R^2	qmax (mg g^−1)	b (L mg^−1)
Cd^2+	y = 0.102x + 0.016	0.993	9.803	6.375
Cu^2+	y = 0.084x + 0.041	0.998	11.904	2.048
Cr(VI)	y = 0.062x + 0.104	0.986	16.129	0.596
Pb^2+	y = 0.066x + 0.013	0.995	15.151	5.077

---

Table 2 Freundlich model parameters

Ion	Kf (mg g^−1)	n (L g^−1)	R^2
Cd^2+	6.053	6.289	0.986
Cu^2+	2.392	4.931	0.972
Cr(VI)	5.308	2.840	0.767
Pb^2+	13.77	2.762	0.902

---

Fig. 8 Equilibrium Langmuir (a) and Freundlich (b) isotherms for Cd²⁺, Cu²⁺, Cr(VI) and Pb²⁺ ions adsorption on the AMMNPs at 25 °C.

3.4. Competitive adsorption experiment

To study the effect of other ions on the analyte adsorption, binary solutions were examined. In each experiment, the analyte concentration was constant, whereas the concentration of the other ion was changed from 10 ppm to 30 ppm. The relationship between the removal percentage and the initial concentration of the competitor ion was investigated. The results demonstrated that Cd²⁺, Cr(VI) and Cu²⁺ can share active sites and therefore these ions compete with each other to reach the active sites. Oxygen ligands and histidine ligands prepare active sites in HSA. Sadler and Viles demonstrated that Cu²⁺ had a higher affinity than Cd²⁺ for their active sites.³¹ These results are consistent with our observations. The removal percentage for Cu²⁺ in the presence of Cd²⁺ was decreased but this effect was not effective compared with resulting decrease of Cd²⁺ removal in the presence of Cu²⁺. In addition, the change of concentration of competitor ions caused a variation of removal percentage that can be attributed to changes of ionic strength because of the presence of interfering ions.

3.5. Comparison with other adsorbents

To evaluate the efficiency of AMMNPs a comparison with various adsorbents was made for Cr(VI) removal (Table 3). For better assessment, four parameters were compared. Table 3 shows that AMMNPs needed the lowest time for chromium ion removal. In addition, the adsorbent dose of AMMNPs was 1.4 mg mL⁻¹, which was lower than other sorbents. This feature caused the water treatment process to become economical. The most important parameter for comparing adsorbents is the maximum adsorption capacity (q_max), and the table shows that AMMNs and surface modified sand has the greatest q_max value.

Table 3 Comparison between various adsorbents used for metal ion removal

Adsorbent	Analyte	Contact time (min)	Initial concentration (mg L^−1)	Maximum adsorption capacity (mg g^−1)	Adsorbent dose (mg mL^−1)	Ref.
Bare magnetic nanoparticles	Cr(VI)	10	5–15	3.55	4	32
Surface modified sand	Cr(VI)	1440	20	16.95	10	33
Fe@Fe2O3 core–shell nanowires	Cr(VI)	—	8	7.78	1	34
Elaeagnus tree leaves	Cr(VI)	100	10–50	10.94	10	35
Amino starch	Cr(VI)	120	—	12.50	—	36
Chitosan coated resin	Cr(VI)	1440	0.4	0.03	—	37
Amorphous and mesoporous silica	Cr(VI)	60	100	3.50	6.6	38
Silica gelatin composite	Cr(VI)	160 (h)	260	9.41	2	39
Porous silica	Cr(VI)	1440	5.2	35.51	10.4	40
HSA	Cr(VI)	8	10	16.10	1.4	This work

---

3.6. Conclusions

In this study, preparation and modification of magnetic nanoparticles using albumin were performed and their characterizations were obtained. The ability of AMMNPs toward Cr(VI), Cd²⁺, Pb²⁺ and Cu²⁺ ion adsorption was investigated. The obtained results showed a rapid, effective and economically favored adsorption method for the removal of selected ions. The variation of parameters did not show a detrimental effect on Pb²⁺ adsorption behavior. Because the results presented equally for both cationic and anionic species, it is concluded that the important feature for AMMNPs is simultaneous adsorption of both anions and cations. This property is because of effective functional groups in the albumin structure that were modified on the surface of MNPs. Displaying immobilization of albumin to the surface of MNPs is of significance for magnetic applications in the removal of toxic metal ions from water and wastewaters. Moreover, this adsorbent, because of its high efficiency can be a good candidate for preconcentration of trace levels of the investigated ions prior their determination by conventional analytical methods.

References

1. G. Crini, Prog. Polym. Sci., 2005, 30, 38–70.
2. A. Luptakova, S. Ubaldini, E. Macingova, P. Fornari and V. Giuliano, Process Biochem., 2012, 47, 1633–1639.
3. N. Misdan, W. Lau and A. Ismail, Desalination, 2012, 287, 228–237.
4. G. Z. Chen, D. J. Fray and T. W. Farthing, Nature, 2000, 407, 361–364.
5. F. Fu and Q. Wang, J. Environ. Manage., 2011, 92, 407–418.
6. L. Monser and N. Adhoum, Sep. Purif. Technol., 2002, 26, 137–146.
7. K. Kadirvelu and C. Namasivayam, Adv. Environ. Res., 2003, 7, 471–478.
8. P. Ganesan, R. Kamaraj, G. Sozhan and S. Vasudevan, Environ. Sci. Pollut. Res., 2013, 20, 987–996.
9. F. Y. Wang, H. Wang and J. W. Ma, J. Hazard. Mater., 2010, 177, 300–306.
10. S. Veli and B. Alyuz, J. Hazard. Mater., 2007, 149, 226–233.
11. L. S. Jin, X. F. Chu, G. L. Lin and B. Ying, Adv. Mater. Res., 2008, 383, 6091–6096.
12. Y. Kong, J. Wei, Z. Wang, T. Sun, C. Yao and Z. Chen, J. Appl. Polym. Sci., 2011, 122, 2054–2059.
13. L. Carro, J. L. Barriada, R. Herrero and M. E. Sastre de Vicente, J. Hazard. Mater., 2011, 192, 284–291.
14. N. Savage and M. S. Diallo, J. Nanopart. Res., 2005, 7, 331–342.
15. K. Krishnapriya and M. Kandaswamy, Carbohydr. Res., 2010, 345, 2013–2022.
16. S. Singh, K. C. Barick and D. Bahadur, J. Hazard. Mater., 2011, 192, 1539–1547.
17. G. E. Fryxell and G.-Z. Cao, Environmental applications of nanomaterials: synthesis, sorbents and sensors, World Scientific, 2012.
18. A. H. Lu, E. L. Salabas and F. Schuth, Angew. Chem., Int. Ed., 2007, 46, 1222–1244.
19. M. S. F. C. Adrnlnlstratlon, Lipoproteins, Apolipoproteins, and Lipases, 1994, vol. 45, p. 153.
20. X. Jiang, L. Huang and X. Liu, Microchim. Acta, 2013, 180, 219–226.
21. X. Liu, J. Xing, Y. Guan, G. Shan and H. Liu, Colloids Surf., A, 2004, 238, 127–131.
22. B. Zhu, N. Ma, D. Wu, Y. Sun and A. W. Li, Ind. Eng. Chem. Res., 2011, 50, 11698–11705.
23. W. Jiang, M. Pelaez, D. D. Dionysiou, M. H. Entezari, D. Tsoutsou and K. O'Shea, Chem. Eng. J., 2013, 222, 527–533.
24. F. Sulek, M. Drofenik, M. Habulin and Z. Knez, J. Magn. Magn. Mater., 2010, 322, 179–185.
25. I. J. Bruce and T. Sen, Langmuir, 2005, 21, 7029–7035.
26. R. M. Dhoble, S. Lunge, A. G. Bhole and S. Rayalu, Water Res., 2011, 45, 4769–4781.
27. N. Morishita, H. Nakagami, R. Morishita, S.-i. Takeda, F. Mishima, S. Nishijima, Y. Kaneda and N. Tanaka, Biochem. Biophys. Res. Commun., 2005, 334, 1121–1126.
28. Z. Ma, Y. Guan and H. Liu, J. Polym. Sci., Part A: Polym. Chem., 2005, 43, 3433–3439.
29. S.-H. Huang and D.-H. Chen, J. Hazard. Mater., 2009, 163, 174–179.
30. N. D. Tumin, A. L. Chuah, Z. Zawani and S. A. Rashid, J. Eng. Sci. Technol., 2008, 3, 180–189.
31. P. J. Sadler and J. H. Viles, Inorg. Chem., 1996, 35, 4490–4496.
32. Y. C. Sharma and V. Srivastava, J. Chem. Eng. Data, 2011, 56, 819–825.
33. S.-M. Lee, W.-G. Kim, C. Laldawngliana and D. Tiwari, J. Chem. Eng. Data, 2010, 55, 3089–3094.
34. Z. Ai, Y. Cheng, L. Zhang and J. Qiu, Environ. Sci. Technol., 2008, 42, 6955–6960.
35. J. Zolgharnein and A. Shahmoradi, J. Chem. Eng. Data, 2010, 55, 3428–3437.
36. A. Dong, J. Xie, W. Wang, L. Yu, Q. Liu and Y. Yin, J. Hazard. Mater., 2010, 181, 448–454.
37. P. Pal and F. Banat, J. Nat. Gas Sci. Eng., 2014, 18, 227–236.
38. D. Pérez-Quintanilla, I. del Hierro, M. Fajardo and I. Sierra, Mater. Res. Bull., 2007, 42, 1518–1530.
39. F. Venditti, F. Cuomo, A. Ceglie, L. Ambrosone and F. Lopez, J. Hazard. Mater., 2010, 173, 552–557.
40. L. Bois, A. Bonhommé, A. Ribes, B. Pais, G. Raffin and F. Tessier, Colloids Surf., A, 2003, 221, 221–230.

---