Preparation of 3D pompon-like titanate nanotube microspheres and their adsorption properties on cationic dyes

Abstract

Hierarchical 3D pompon-like microspheres built with titanate nanotubes have been successfully synthesized by a simple hydrothermal method in the presence of CH₃COOH and H₂O₂. The samples have been characterized by X-ray powder diffraction, scanning electron microscopy, transmission electron microscopy, X-ray photoelectron spectroscopy and nitrogen adsorption and desorption, zeta potential etc. techniques. The pompon-like microspheres consist of titanate nanotubes twisted each other. These nanotubes are 7 nm in inner diameter and 17 nm in outer diameter, and its length can reach several tens micrometer. The hierarchical nanotube microspheres have a high specific surface area of 217.98 m² g⁻¹, and it exhibits excellent adsorption performance toward cationic dyes. 50–60% measured dyes could be removed in 10 min; after 30 min, the removal rate of neutral red, methylene blue, malachite green and crystal violet could reach 76%, 79%, 77% and 55% respectively, and the saturated adsorption capacities of these four dyes are 401.8, 382.07, 361.95, and 322.5 mg g⁻¹ at 25 °C, respectively. The adsorption kinetics has been studied, it indicates that the adsorption data fits perfectly with the pseudo-second-order kinetics, and the controlling step of adsorption process should be chemisorption.

---

1. Introduction

As important industrial materials, dyes have been widely used in textile, printing and some other industries, but they are unsustainable and environmentally unfriendly, dye pollution is harmful to flora and fauna, and threatens human health seriously.^1,2 Obviously, the removal of color synthetic organic dye stuff from waste effluents becomes environmentally important. It is rather difficult to treat these dyes due to their complex molecular structure and synthetic origin. Many conventional methods, including ion exchange,³ photocatalytic degradation,⁴ oxidation,⁵ precipitation,⁶ and adsorption, have been investigated to remove dyes from aqueous systems. Among many chemical, physical, and biological treatment methods, adsorption technology is a promising one, mainly because of its effectiveness and low cost. Many materials, such as seaweed,⁷ activated carbon,⁸ graphene,⁹ porous metal–organic frameworks,³ titanium peroxide¹⁰ and layered double hydroxides¹¹ could be used to adsorb dyes from waste water.

Titanate nanotubes (TNT) usually act as an intermediate product for the preparation of TiO₂ nanotubes via wet chemical method. Recent years, TNT itself have attracted much attention due to their potential application in the areas of photocatalysis,^12,13 wastewater treatment,¹⁴ ion exchange,¹⁵ lithium ion storage,¹⁶ solar energy conversion,¹⁷ and so on. Due to the high surface area, nanosized titanate is a good candidate for dye adsorption materials. For example, Lin et al. reported MB adsorption capacity of titanate nanotubes could reach to 133.33 mg g⁻¹,¹⁸ Cheng et al. found that titanate nanosheets could adsorb up to 184 mg g⁻¹ MB and 366 mg g⁻¹ Pd(II) ions, respectively.¹⁹

The size, morphology and microstructure of nanomaterials will affect their performance. Up to now, numerous titanate nanostructures including 1D titanate nanotubes,^18,20,21 nanowires,²² nanorods,²³ 2D titanate nanosheets,^19,24,25 nanoflowers²⁶ and layered hollow spheres²⁷ have been successfully prepared. Recently, Yu et al. have synthesized hierarchical titanate tubular structures self-assembled by nanotubes and nanosheets which demonstrated enhanced photovoltaic conversion efficiency.²⁸ 3D hierarchical nanostructures should also improve the absorption performance toward organic pollutants, due to the superior properties derived from the high specific surface area and porous structure. However, it is still a challenge to produce 3D hierarchical titanate nanomaterials with unique microstructure via a simple, fast, and inexpensive method.

Herein, we synthesized hierarchical 3D pompon-like microspheres containing titanate nanotubes by a facile hydrothermal method, and this 3D TNT microspheres exhibit excellent selective adsorption performance toward cationic dyes. To our best knowledge, this unique 3D TNT microsphere with excellent adsorption ability has not been reported yet. This study provides a ‘green’ method for preparing hierarchical 3D titanate nanotubes which will be a potential adsorption material toward dyes.

2. Experimental

2.1. Materials

All chemicals including titanic sulfate Ti(SO₄)₂, deionized water (H₂O), sodium hydroxide (NaOH), acetic acid (HAc), hydrogen peroxide (H₂O₂), hydrochloric acid (HCl), methylene blue (MB), neutral red (NR) and crystal violet (CV) malachite green (MG) were purchased from Chengdu Kelong Chemical Reagent Co. Ltd, (China), and used without further purification.

2.2. Preparation

The synthesis procedure of 3D-TNT was accomplished via a hydrothermal process. At first, 3.6 g of Ti(SO₄)₂ powders were added to 75 mL of aqueous solution containing NaOH (10.0 M) and CH₃COOH. After stirring for 15 min, 6 mL of hydrogen peroxide was added to the mixture, then the mixture was placed in a Teflon-lined autoclave, the autoclave was sealed and heated at 180 °C for 12 h, and then cooled naturally to room temperature. The precipitate was carefully collected by centrifugation and washed several times with distilled water, after being rinsed in CH₃COOH (0.1 M) for 12 h, the precipitate was washed with deionized water several times until the pH of the solution became 7. Finally, the white powder was collected for following characterization and measurement.

2.3. Characterization

The morphology and microstructure of samples were characterized by a field emission scanning electron microscopy (SEM, Hitachi S-4800) and transmission electron microscope (TEM, Tecnai G2F30 S-TWIN); the crystalline structure of the samples is identified by X-ray diffraction (XRD, X'Pert PRO MPD, PANalytical Co.) using Cu K_α radiation at a scanning rate of 5° per min in the range of 5–80°. Nitrogen adsorption and desorption isotherm was measured using Micromeritics Tristar II3020 sorptometer. The specific surface area of the sample was derived using the multipoint Brunauer–Emmett–Teller (BET) method and the pore size distribution was determined using the Barret–Joyner–Halenda (BJH) mathematical model. X-ray photoelectron spectroscopy (XPS) measurements were carried out using a Kratos XSAM 800 instrument with Al Kα X-ray radiation during XPS analysis. Absorbance spectra of products were measured by using UV1101 spectrophotometer (Techcomp). The zeta potential of all samples were measured in the pH 1–11 range with a Nano-ZS90 Zetasizer (Malvern Instruments).

2.4. Adsorption test

All the adsorption experiments were conducted under stirring conditions throughout the test at room temperature (25 °C) in the dark. 10 mg of as prepared adsorbent was added to 50 mL of dye solutions (MB, NR, MG and CV) of initial concentrations of 50, 100 and 200 mg L⁻¹ respectively with stirring continuously. At appropriate time intervals, the aliquots were withdrawn from the solution and the adsorbents were separated from the suspension via centrifugation. The concentration of residual dye in the supernatant solution was detected using a UV 1101 visible spectrophotometer at the corresponding maximum absorption wavelength. The sampling continued until the dye concentration became constant, that is, the adsorption process reaches its equilibrium. Every adsorption experiment was repeated 3 times. Table 1 lists the maximum absorption wavelengths and chemical structures of measured dyes.

Table 1 Maximum absorption wavelengths and chemical structures of measured dyes

Name	λmax/nm	Chemical structure
Methylene blue	664
Malachite green	621
Neutral red	523
Crystal violet	581

---

3. Results and discussion

3.1. Characterization results

The diffraction pattern of the sample indicated that the product has an orthorhombic H₂Ti₂O₅·H₂O (JCPDS 47-0124) layered crystal structure (Fig. 1). The diffraction peaks at 9.9°, 24.4°, 28.6° and 48.6° correspond to the (200), (110), (310) and (020) lattice planes of protonic titanate, respectively. Physical structure of as-synthesized samples was also investigated using SEM and TEM. The SEM and TEM images with different magnifications are shown in Fig. 2. As can be seen, the hierarchical 3D pompon-like microspheres consist of titanate nanotubes twisted each other. The nanotubes are 7 nm in inner diameter and 17 nm in outer diameter, and its length can reach to several tens micrometer. The specific surface area and pore size distribution are the key factors for the tubular inorganic nanomaterials for water treatment application. The nitrogen adsorption/desorption isotherm of 3D-TNT demonstrated a specific surface area of 217.98 m² g⁻¹ from BET analysis (Fig. 3a). According to the IUPAC classification, the isotherm could be assigned as type IV isotherm with H3-type hysteresis loop (P/P₀ > 0.5), indicating the presence of mesopore.²⁹ The corresponding BJH analysis (Fig. 3a inset) suggested the maximum pore volume occurring at the pore size of 11.4 nm. The high BET surface area and open mesoporous morphology are likely to benefit adsorption performance.¹³ Fig. 3b shows the zeta potential of 3D TNT microspheres, it indicates that the titanate nanotubes possessed a dominant negative charge under the neutral condition, and it suggests that the TNT microspheres are suitable for the cationic organic dye adsorption.¹⁸

Fig. 1 XRD pattern of sample.

Fig. 2 SEM (a and b) and TEM (c and d) images of sample.

Fig. 3 (a) Nitrogen adsorption/desorption isotherm and (b) zeta potential of sample in water at different pH values.

Fig. 4 shows the XPS spectra for the C1s, Ti2p and O1s region of the TNT microspheres. The two peaks at 284.853 and 288.70 eV in C1s spectrum can be attributed to hydrocarbon and C=O bonds respectively,³⁰ which maybe come from residual CH₃COOH or XPS instrument itself. The Ti2p3/2 and Ti2p1/2 were identified at binding energies 458.612 eV and at 464.278 eV, which is the feature of Ti⁴⁺ in the hydrogen titanitate.³⁰ The O1s spectrum shows a main peak at 530.284 eV with two shoulders at 531.684 eV and 535.260 eV. The peak at 530.284 eV is assigned to the Ti–O in the hydrogen titanitate, while the shoulder at 531.684 eV may be attributed to the Ti–OH.¹⁰ The shoulder at 535.260 eV indicated the existence of oxygen in hydroxyl groups (H–O) resulting mainly from the constitution water.³¹ The existence of Ti–O and Ti–OH make the TNT microspheres be negatively charged, and this result is consistent with the zeta potential analysis.

Fig. 4 XPS spectra of the C1s, Ti2p and O1s region.

3.2. Formation mechanism of 3D TNT microspheres

HAc plays an important role on the formation of 3D TNT microspheres. At first, titanic sulfate combined with acetic acid to form chelates—Ti₂(Ac)_n⁸⁻ⁿ. Here HAc can serve as a chelating agent to prevent precipitation and control the release of Ti ions, which will facilitate the formation of 1D nanotubes. After adding NaOH solution (10 M), the Ti₂(Ac)_n⁸⁻ⁿ aggregated on the water–gas interface of O₂ bubbles produced from H₂O₂ will hydrolyze and Na-titanate crystal nuclei start to form. Na-titanate crystal nuclei aggregated on the surface of O₂ bubbles will grow to form nanotubes twisted together, and 3D Na-titanate-nanotube microspheres are obtained. After the as-prepared Na-titanate-nanotube microspheres are thoroughly washed with HAc solution (0.1 M), Na⁺ is replaced by H⁺,^32–34 finally, H-titanate-nanotube microspheres can be obtained.

3.3. Adsorption experiment

Four different dyes including methylene blue, neutral red, crystal violet, and malachite green were used to study the adsorption property of 3D-TNT. As can be seen from Fig. 5, the hierarchical nanotube microspheres exhibit excellent adsorption performance toward cationic dyes. All curves exhibit the same regularity. (1) In the first 10 min, the adsorption is very fast and 50–60% cationic dyes could be removed. Combined with XPS results, there are many hydroxyl groups on 3D-TNT surface, which are negatively charged. So there is strong electrostatic interaction between positively charged dyes and negatively charged 3D-TNT, this will enhance the adsorption ability of TNT.³¹ (2) Subsequently, adsorption rate decreases significantly. After 30 min, the remove rate of NR, MG, MB, and CV could reach 76%, 77%, 79% and 55% respectively. Slower adsorption rate should be induced by the decreasing adsorption points and vacant surface, and the repulsion between adsorbed dyes molecules.¹⁰

Fig. 5 Removal efficiency of 3D TNT microspheres for MB, CV, MG and NR (initial concentration 100 mg L⁻¹, catalyst dosage 0.2 g L⁻¹, temperature 25 °C).

We studied the adsorption property of 3D-TNT at different initial concentration. As shown in Fig. 6, 3D-TNT demonstrates excellent adsorption ability toward cationic dyes at different initial concentration. In addition, the adsorption rates on NR, MG, MB and CV are different. The molecular size, molecular structure and their charged situation have an influence on adsorption. We can know from Table 1, the molecular structure and size of measured dyes are different, the smaller molecular size will induce easier and faster adsorption. The big size of CV molecule induces slower adsorption. Based on the adsorption results, the calculated saturated adsorption capacities for NR, MB, MG, and CV are 401.8, 382.07, 361.95, and 322.5 mg g⁻¹ at 25 °C, respectively.

Fig. 6 The adsorption curves of MB, NR, MG and CV at different initial concentration (catalyst dosage 0.2 g L⁻¹, temperature 25 °C).

3.4. Adsorption isotherms

Langmuir model is a common isotherm model which is widely used to describe adsorption isotherms. Fig. 7 shows the equilibrium isotherms for adsorption of dyes onto 3D-TNT microspheres and the equilibrium adsorption characteristics were analyzed by using the Langmuir isotherm model.³⁵where q_e (mg g⁻¹) is the amount of adsorption capacity at equilibrium, q_max (mg g⁻¹) is the maximum adsorption capacity of the adsorbate, C_e (mg L⁻¹) is the concentration of the adsorbate at equilibrium and K (L mg⁻¹) is the Langmuir adsorption equilibrium constant. The calculated constants and correlation coefficients are listed in Table 2. The high R² values support the fact that the Langmuir model is indeed an adequate model for this system. In addition, the q_max of NR, MB, MG and CV are 415.61, 394.96, 369.80 and 334.19 mg g⁻¹ at 25 °C, which is in accordance with the q_max acquired from the experiment.

Fig. 7 Langmuir isotherm plots for adsorption of MB, CV, MG and NR.

Table 2 Langmuir isotherm constants for adsorptions of cationic dyes onto 3D TNT microspheres

Isotherm constants	Pollutants
	MB	CV	NR	MG
K (L mg^−1)	0.3298	0.15618	0.30735	0.4824
qmax (mg g^−1)	394.96	334.19	415.61	369.80
R^2	0.99209	0.9968	0.99336	0.99601

---

3.5. Adsorption kinetics

In order to study the adsorption kinetics of cationic dyes onto 3D TNT microspheres, three kinetic models, the pseudo-first-order, pseudo-second-order and intra-particle diffusion kinetic models, were used to fit the experimental data. The Lagergren pseudo-first-order kinetic model has been widely applied for the adsorption of solid/liquid systems.^36,37 Fig. 8 shows the plots of ln(q_e − q_t) versus t for four cationic dyes at different initial concentrations ranging from 50 to 200 mg L⁻¹. The kinetic parameters obtained from the pseudo-first-order and pseudo-second-order kinetic model are listed in Table 3. It can be seen from Fig. 8, the fitting of experimental data to the pseudo-first-order model is not good, with low linear regression correlation coefficients R₁² (0.559–0.922). Moreover, a large difference of q_e between the experiment (q_e,exp) and calculation (q_e,cal) is observed. These facts indicate that it is not appropriate to use the pseudo-first-order kinetic model to predict the adsorption kinetics of cationic dyes onto 3D-TNT.^36,37

Fig. 8 Pseudo-first-order kinetic plots for MB, NR, MG and CV.

Table 3 Kinetic constants for MB, NR, MG and CV onto 3D-TNT microspheres

Kinetic model	Pseudo-first-order kinetic model	Pseudo-second-order kinetic model
Equation	ln(qe − qt) = ln qe − k1t	t/qt = (1/qe)t + 1/(qe^2k2)
Capacity term	qt, qe: the amounts of dyes absorbed (mg g^−1) at time t and at equilibrium, respectively, k1: the first-order equilibrium rate constant (min^−1), k2: the second-order equilibrium rate constant (g (mg min)^−1)

---

Parameters		qe,exp (mg g^−1)	qe,cal (mg g^−1)	k1 (min^−1)	R1^2	qe,cal (mg g^−1)	k2 (g (mg min)^−1)	R2^2
Concentration of MB (mg L^−1)	50	225.67	6.58	0.05927	0.75365	225.82	0.02364	1
	100	358.11	70.35	0.04948	0.90706	360.96	0.00121	0.99999
	200	382.07	77.47	0.05287	0.85572	385.74	0.00101	0.99997
Concentration of NR (mg L^−1)	50	227.4	11.72	0.03682	0.63919	227.57	0.00773	0.99999
	100	369.75	106.48	0.0382	0.9064	374.65	0.000559	0.99997
	200	401.8	124.6	0.03549	0.89986	406.70	0.000449	0.99975
Concentration of MG (mg L^−1)	50	237.04	37.54	0.07043	0.92248	237.89	0.00439	1
	100	347.05	95.72	0.0678	0.88499	349.63	0.00142	1
	200	361.95	77.340	0.05289	0.58897	364.92	0.00118	0.99999
Concentration of CV (mg L^−1)	50	217.4	28.04	0.04922	0.86285	218.39	0.00342	1
	100	286.05	41.81	0.05368	0.89994	287.45	0.0025	1
	200	322.5	56.63	0.04889	0.85427	324.37	0.00164	0.99999

---

The pseudo-second-order kinetics model was proposed by Ho and McKay based on the assumption that the rate-controlling step may be chemical sorption or chemisorption involving valency forces through sharing or exchange of electrons between adsorbent and adsorbate.³⁸ Fig. 9 shows the plots of pseudo-second-order kinetics model for four cationic dyes at different initial concentrations. The calculated parameters of the pseudo-second-order kinetic models are also listed in Table 3. From Table 3 and Fig. 9, it was noticed that, at all initial dye concentrations and for the entire adsorption period, the experimental data fits the pseudo-second-order model very well, and the linear regression correlation coefficient R₂² values are very high (0.999–1). Besides, there is only minute differences between the q_e,exp and q_e,cal. These all support that the adsorption data are well represented by the pseudo-second-order kinetics model for the entire adsorption period and thus imply that the rate-controlling step of cationic dyes onto 3D-TNT microspheres is chemisorption, which is in good agreement with the sorption equilibrium data represented by Langmuir isotherm equation.³⁷ The similar results have also been found in other nanosized titanate system.^19,26

Fig. 9 Pseudo-second-order kinetic plots for MB, NR, MG, and CV.

We further use the Weber's intraparticle diffusion model to analyze the adsorption kinetic data. This model can be expressed as:³⁹

q_t = k_it^0.5 + C (2)

where k_i (mg (g min^0.5)⁻¹) is the intraparticle diffusion rate constant and C (mg g⁻¹) indicates the boundary layer effect of the adsorption. If the intraparticle diffusion is the rate-controlling step, the plot should be linear and pass through the origin. As shown in Fig. 10, for all dyes adsorbed onto 3D-TNT, a two-stage adsorption plot are observed. The initial sharper portion of the plot can be attributed to the boundary layer effect, while the second portion is due to the intraparticle diffusion.^18,33 the linear plots of the second potion do not pass through the origin and the intercepts are big (225.12, 232.45, 224.85 and 212.72 mg g⁻¹ for MB, MG, NR and CV respectively), this strongly suggests that the intraparticle diffusion is not the rate-controlling step, and the bulk mass transfer is remarkable in the adsorption process of cationic dyes onto the 3D-TNT microspheres. A similar phenomenon was also observed for the adsorption of MB from aqueous solution on titanate nanoflowers¹⁴ and titanate nanotubes.¹⁸

Fig. 10 The linear plots of intraparticle diffusion model for the adsorption of MB, NR, MG, and CV (C₀ = 50 mg L⁻¹) on 3D-TNT.

4. Conclusions

In this study, hierarchical 3D pompon-like microspheres consisting of titanate nanotubes have been successfully prepared by a simple hydrothermal method in the presence of CH₃COOH and H₂O₂. It indicates that the obtained material is rich in hydroxyl groups. The 3D-TNT microspheres possess excellent selective adsorption performance toward cationic dyes and the decolorization process is very fast, the saturated adsorption capacities for NR, MB, MG, and CV are 401.8, 382.07, 361.95, and 322.5 mg g⁻¹ at 25 °C, respectively. Kinetics data of adsorption fits well with the pseudo-second-order kinetics, the rate-controlling step is chemisorption. Consequently, the 3D-TNT microspheres can be a potential adsorbent due to its high adsorption capacity and fast adsorption rate.

Acknowledgements

The authors gratefully acknowledge the financial support provided by the National Natural Science Foundation of China (No. 50872084, 51072124) and the Program for New Century Excellent Talents in University (No. NCET100605). We wish to thank the Analytical & Testing Center of Sichuan University (SCU) for the assistance in sample characterization.

References

1. S. Saber-Samandari, H. O. Gulcan and S. Saber-Samandari, Water, Air, Soil Pollut., 2014, 225, 1–14.
2. X. S. Wang, J. Liang and L. Li, Inorg. Chem., 2016, 55, 2641–2649.
3. Y. C. He, J. Yang and W. Q. Kan, J. Mater. Chem. A, 2014, 3, 1675–1681.
4. M. Amini and M. Ashrafi, Nanochemistry Research, 2016, 1, 79–86.
5. H. S. Oliveira and L. D. Almeida, Catal. Today, 2015, 240, 176–181.
6. M. X. Zhu, L. Li and H. H. Wang, J. Hazard. Mater., 2007, 149, 735–741.
7. S. L. Hii, L. L. Estrop and C. L. Wong, Int. J. Phys. Sci., 2011, 6, 7171–7178.
8. Y. S. Al-Degs, M. I. El-Barghouthi and A. H. El-Sheikh, Dyes Pigm., 2008, 77, 16–23.
9. T. Liu, Y. Li and Q. Du, Colloids Surf., B, 2012, 90, 197–203.
10. X. G. Zhao, J. G. Huang and B. Wang, J. Bacteriol., 1991, 173, 2521–2529.
11. R. Marangoni, A. Mikowski and F. Wypych, J. Solid State Electrochem., 2011, 15, 303–311.
12. M. Grandcolas, A. Louvet and N. Keller, Angew. Chem., Int. Ed., 2009, 48, 161–164.
13. Q. Q. Zhai, T. Bo and G. X. Hu, J. Hazard. Mater., 2011, 198, 78–86.
14. J. Huang, Y. Cao and Z. Liu, Chem. Eng. J., 2012, 180, 75–80.
15. N. Li, L. Zhang and Y. Chen, Adv. Funct. Mater., 2012, 22, 835–841.
16. J. H. Liu, J. S. Chen and X. F. Wei, Adv. Mater., 2011, 23, 998–1002.
17. D. V. Bavykin, J. M. Friedrich and F. C. Walsh, Adv. Mater., 2006, 18, 2807–2824.
18. L. Xiong, Y. Yang and J. Mai, Chem. Eng. J., 2010, 156, 313–320.
19. C. H. Lin, S. H. Wong and S. Y. Lu, ACS Appl. Mater. Interfaces, 2014, 6, 16669–16678.
20. L. Xiong, C. Chen and Q. Chen, J. Hazard. Mater., 2011, 189, 741–748.
21. Y. Qiu and X. J. Li, Dalton Trans., 2014, 43, 14537–14541.
22. X. Peng and A. Chen, Adv. Funct. Mater., 2006, 16, 1355–1362.
23. L. Z. Pei, H. D. Liu and N. Lin, Mater. Res. Bull., 2015, 61, 40–46.
24. P. Liu, H. Zhang and H. Liu, J. Am. Chem. Soc., 2011, 133, 19032–19035.
25. N. Sutradhar, A. Sinhamahapatra and S. K. Pahari, Chem. Commun., 2011, 47, 7731–7733.
26. J. Huang, Y. Cao and Z. Liu, Chem. Eng. J., 2012, 191, 38–44.
27. N. Sutradhar, S. K. Pahari and M. Jayachandran, J. Mater. Chem. A, 2013, 1, 9122–9131.
28. J. G. Yu, Q. L. Li and J. J. Fan, Chem. Commun., 2011, 47, 9161–9163.
29. K. S. W. Sing, Pure Appl. Chem., 1985, 57, 603–619.
30. D. Fan, W. Zhao and Z. Wu, Nanotechnology, 2008, 19, 2618–2624.
31. J. Ma, F. Yu and L. Zhou, ACS Appl. Mater. Interfaces, 2012, 4, 5749–5760.
32. H. H. Ou and S. L. Lo, Sep. Purif. Technol., 2007, 58, 179–191.
33. C. K. Lee, K. S. Lin and C. F. Wu, J. Hazard. Mater., 2008, 150, 494–503.
34. T. Kasuga, M. Hiramatsu and A. Hoson, Adv. Mater., 1999, 11, 1307–1311.
35. K. Y. Foo and B. H. Hameed, Chem. Eng. J., 2010, 156, 2–10.
36. H. Chen and J. Zhao, Adsorption, 2009, 15, 381–389.
37. K. V. Kumar, V. Ramamurthi and S. Sivanesan, J. Colloid Interface Sci., 2005, 284, 14–21.
38. Y. S. Ho and G. Mckay, Process Biochem., 1999, 34, 451–465.
39. F. Liu, S. Chung and G. Oh, ACS Appl. Mater. Interfaces, 2012, 4, 922–927.

---