Template-free synthesis of 3D hierarchical amorphous aluminum oxide microspheres with broccoli-like structure and their application in fluoride removal

Abstract

3D hierarchical amorphous aluminum oxide microspheres with broccoli-like structures were synthesized successfully via a facile method without using any template. A multistep crystal adsorption-splitting growth mechanism is proposed to understand the formation of this material. It displayed an excellent adsorption performance for fluoride, and the adsorption capacity can reach 126.9 mg g⁻¹.

---

Introduction

Over the past few years, hierarchically structured metal oxides have exhibited many interesting shape- and size-dependent phenomena and special properties. They have been widely investigated from their basic scientific research to important technological applications in various fields including catalysis,^1–3 adsorption,^4–6 drug delivery,^7,8 sensor,^9,10 energy conversion and storage systems,^11–13 and many others.^14–17 The ever-growing needs for hierarchically structured metal oxides with novel chemical and physical properties, especially for large-scale practical applications (e.g., water treatment) have prompted numerous efforts dedicated to develop methods for rational the design and synthesis of these materials. So far, significant advances have been achieved with most hierarchically structured metal oxides synthesized by various synthetic methods which are mainly based on the mechanisms of soft- and hard-templating, Kirkendall effect,¹⁸ Ostwald ripening,^19,20 crystal splitting,^21,22 galvanic replacement²³ and chemically induced self-transformation.²⁴ However, most of the synthetic methods still need the use of surfactants, toxic organic solvents and complex chemical procedures, which may prevent them from being used in large-scale applications, particularly, in adsorption and catalysis. Overall, it remains a great challenge but highly desirable to fabricate hierarchically structured metal oxides with designed chemical components, tunable size and novel morphologies by a facile approach.

Most recently, hierarchically structured metal oxides with high surface area and abundant active sites have emerged as attractive materials for water treatment. Among them, aluminum oxide has received much attention in view of its low cost, natural abundance and environment-friendly properties.^25–29 Up to now, various hierarchically structured aluminum oxide, such as nanotubes,³⁰ nanofibers,³¹ hollow microspheres,^32,33 spindle-like,²⁶ flower-like,^34,35 and jellyfish-like³⁶ have been prepared by different methods. However, the obtained aluminum oxides generally exhibit crystalline phase such as γ-Al₂O₃ (ref. 34 and 37–39) and AlOOH,^25,36,40 rather than amorphous phase. As we know, the crystalline materials used as adsorbent often suffer from disadvantages related to low adsorption capacity and high cost because of the high crystallinity and less porous structure, while the amorphous adsorbents have the high adsorption capacity due to the abundant porous and highly hydrated structure.⁴¹ To the best of our knowledge, there is no study on synthesis and characterization of the 3D hierarchical amorphous aluminum oxide used to remove fluoride from aqueous solution.

Herein, 3D hierarchical amorphous aluminum oxide microspheres with broccoli-like structures for the first time are obtained by a facile method without using any template, surfactant or toxic materials. A possible formation mechanism of multistep crystal adsorption-splitting growth has been proposed based on systematic investigation of the total crystal splitting process. This is the first report about fabricating hierarchically structured amorphous aluminum oxide using this method. The formation of such microspheres with uniform primary building blocks, well-defined morphology and developed porous structure is very unusual. By controlling the reaction time, the morphology of the precursor can be tailored from 1D nanorods, 2D dumbbell-like bundles to 3D hierarchical broccoli-like microspheres. The as-obtained 3D hierarchical broccoli-like amorphous aluminum oxide microspheres with high surface area exhibit very promising performance when tested as an adsorbent for fluoride removal in water treatment.

Results and discussion

Characterization of aluminum oxide microspheres

The aluminum oxide microspheres were achieved by a facile solvothermal reaction by using acetic acid (HAc) as solvent. Aluminium isopropoxide (Al(OiPr)₃), as metal precursor, was dissolved in HAc, and then the solution was sealed into a Teflon-lined stainless steel autoclave and maintained at 200 °C for 2 h. After cooling, the as-synthesized precursor was collected and washed with deionized water and absolute ethanol several times by a centrifugation–redispersion process. The final product can be obtained by calcination of the precursor in air. The morphology of the precursor can be revealed by the field-emission scanning electron microscopy (FESEM) (Fig. 1a and b) and transmission electron microscopy (TEM) (Fig. 1c and d) observations. As shown in Fig. 1a, the as-prepared precursor consists of discrete and uniform broccoli-like microspheres with a diameter of about 10 μm. The enlarged magnification SEM image (Fig. 1b) further reveals the detailed morphology of these broccoli-like microspheres, which are composed of uniform nanorods. TEM images given in Fig. 1c and d show the hierarchical structure assembled from large continuous nanorods with relatively small thickness. These nanorods have smooth surface and width of 15–20 nm. Both the SEM and TEM images indicate that the 3D hierarchical broccoli-like microspheres with diameters about 10 μm were successfully prepared via one-step solvothermal reaction. The white powders of the precursor were characterized using X-ray diffraction (XRD) to verify their structure. The result shown in Fig. 1e demonstrates that all the reflections of the samples are in good agreement with the standard pattern of the pure orthorhombic C₄H₇AlO₅ phase (JCPDS card no. 13-0833). There were no other phases or impurities detected in the patterns, which indicates the high purity of the precursor. Nitrogen adsorption–desorption isotherm of the precursor (Fig. S1†) shows mainly H3-type hysteresis loop, indicating the presence of abundant mesopores. The average pore diameter and surface area of the precursor are calculated to be 2.769 nm and 62 m² g⁻¹ (Table S1†), by using the density functional theory (DFT) (inset in Fig. S1†) and Brunauer–Emmett–Teller (BET) method.

Fig. 1 (a) Low-magnification SEM image; (b) enlarged SEM image; (c) low-magnification TEM image; (d) HR-TEM image of the precursor and (e) XRD pattern of the precursor.

To shed light on the formation process of the 3D hierarchical microspheres, the evolution of the morphology with reaction time was investigated. As shown in Fig. 2, it can be seen that 1D nanorods (Fig. 2a) can be produced after reacting for 1 h. When the reaction time increased to 1.5 h, the dumbbell-like structures (Fig. 2b) can be obtained. The structure of the precursor was further changed into the 3D hierarchical microspheres with broccoli-like structures after reaction for 2 h, as shown in Fig. 2c. However, the shape of the precursor finally evolved into dumbbell-like structures again (Fig. 2c) after further increasing the reaction time to 4 h. Therefore, the reaction time plays an important role in the formation of the 3D hierarchical microspheres. With increasing the reaction time from 1 h to 2 h, the morphology of the precursor underwent from 1D nanorods, 2D dumbbell-like bundles to 3D hierarchy broccoli-like microspheres.

Fig. 2 SEM images of the precursor collected for different time intervals of (a) 1 h, (b) 1.5 h and (c) 2 h.

Based on the above observation and analysis, it was found that 1D nanorods can gradually assemble into dumbbell-like bundles and broccoli-like microspheres through a crystal splitting mechanism, which is similar to the growth of Bi₂S₃ and L-cysteine–Pb.^42–44 Our observations of the effect of time on the morphology of the precursor reveal not only its fast growth nature, but also complex splitting behaviors of the materials.²¹ Thus, we propose a multistep crystal adsorption-splitting growth mechanism to understand the formation of the broccoli-like microspheres (Scheme 1). Firstly, after the formation of the nuclei in the reaction system (Scheme 1a), the 1D nanorods can be obtained because of the fast growth of the precursor as shown in Scheme 1b; secondly, splitting growth can only take place at both tips except in the middle of the nanorods. Thus, branched nanorods could be produced in the first generation of crystal splitting as shown in Scheme 1c. With prolonging reaction time, dumbbell-like bundles of the precursor can be produced in the subsequent splitting growth as shown in Scheme 1c. The branched nanorods can be adsorbed on the middle part of the dumbbell-like bundles, and then the continuous splitting growth can occur not only from the seeded bundles, but also from tips of the freshly adsorbed branched nanorods (Scheme 1d). The branched nanorods repeated to be adsorbed, split and grow along the radial direction in the whole structure, and then immature splitting microspheres can be produced from the aggregated nanorods packed together as shown in Scheme 1e. With the increase of reaction time, the immature splitting microspheres became more perfect and regular into the microspheres with broccoli-like structures which loosely packed outside and densely packed inside (Scheme 1f). This process suggested the aluminum oxide had a strong splitting ability, and the branched nanorods packed together to make up a broccoli-like microsphere.

Scheme 1 Scheme illustrating the crystal adsorption-splitting growth mechanism of the precursor with 3D hierarchical structure.

Up to now, the intrinsic nature of crystal splitting phenomena is not very well established presently. Crystal splitting may take place due to a variety of reasons which are unidentified and varies depending on the system. However, crystal splitting is generally related to several factors as following: (1) crystal splitting is usually associated with fast crystal growth, while splitting only occurs if the oversaturation exceeds a certain critical level, especially, for some specified condition.^21,45,46 Although, the crystal growth is mainly subject to the solution oversaturation, it could also be dramatically affected by reaction parameters, such as reaction temperature, time and solvent, which is unique to each material. During the fast crystal growth, a high density of crystal defects can be produced, which always act as nuclei sites to gradually develop branches on the stem, leading to different degrees of splitting. (2) Crystal splitting is likely to occur in a crystal which possesses small lateral adhesion energy.⁴⁷ (3) Crystal splitting is also favored in a situation where the organic surfactant is a very potent surface stabilizer.²² These factors can synergistically affect the crystal splitting process and promote a number of subforms of split nanocrystal generated. The driving force for the splitting growth is to form nanocrystals with low energy planes as exposed facets to minimize interfacial energy and obtain stable structure for final product. Thus the crystal splitting process should be the consequence of seeking a balance between the mismatch energy and the interfacial energy.

TGA of the precursor was measured and shown in Fig. 3a. The TGA curve exhibited mainly two decomposition steps. The first step thermal decomposition located below 200 °C was ascribed to losses of adsorbed and crystal water, being their weight loss about 5%. The second step, located between 200 and 450 °C, was attributed to the dehydroxylation from the precursor, with consequent evolution of water and CO₂, corresponding to about 62% mass loss. According to the TGA results, we know a fact that the precursor will undergo the following reaction in the presence of air and high temperature in the thermal treatment process.

C₄H₇AlO₅ + O₂ → aluminum oxide + CO₂ + H₂O

Fig. 3 (a) TGA curve; (b) XRD patterns of aluminum oxide calcined at deferent temperatures; (c) low-magnification SEM image; (d) high-magnification SEM image; (e) low-magnification TEM image and (f) HR-TEM image of the final product.

This is accompanied by a phase transformation, which can also be verified by the XRD characterization results of the precursor calcined at different temperatures (Fig. 3b). After calcination, it is amazing to find that a phase transformation from crystalline C₄H₇AlO₅ to amorphous aluminum oxide was occurred, as shown in XRD pattern (Fig. 3b). As shown in Fig. 3c, the SEM image of the final product shows that the final product inherit the 3D hierarchical structure of the precursor completely. The high-magnification SEM image of the final product (Fig. 3d) further demonstrates that the 3D hierarchical microspheres consist of 1D nanochains which grow radically to form broccoli-like structures. Additionally, the 1D nanochains, with rough surfaces, are composed by nanoparticles. Therefore, a porous structure was formed due to the oxidation and decomposition of the final product. The porous structure of the microsphere is further proved by TEM observations. The low-magnification TEM image (Fig. 3e) shows that the 3D hierarchical structure of the broccoli-like microspheres is perfectly retained after calcination, and the microsphere display a porous structure. The HR-magnification TEM image (Fig. 3f) taken on a part view of the microsphere shows that the 3D hierarchical structure is composed of 1D nanochains build up by nanoparticles with diameters ranging from 15 to 20 nm. The template-free formation of such a 3D hierarchical structure with 1D primary building blocks, uniform secondary morphology and developed porous structure is very unusual. SEM-EDX spectrum (Fig. S2a†) of the final product clearly suggests the presence of Al and O elements, which indicates the formation of pure aluminum oxide. To further investigate the elements distribution on the broccoli-like microspheres, elemental mapping characterization was carried out as shown in Fig. S2b.† The different color images indicate Al and O-enriched areas of the microspheres, respectively. It can be found that the element of Al is well dispersed on the surface of the microspheres. The average pore diameter and surface area of the final product are changed from 2.769 nm and 62 m² g⁻¹ to 1.410 nm and 376 m² g⁻¹ after calcination at 300 °C, which indicate a porous structure formed due to the generation and release of gases in the process of thermal treatment (Fig. S3†). This result is good agreement with TEM characterization. Moreover, it is found that the BET surface area of the as-prepared product calcined above 300 °C decreases with the increasing of the calcination temperature, and the maximum BET surface area is obtained at 300 °C (Table S1†).

Surface compositions of the final product are obtained by XPS analysis (Fig. S4a†). It can be seen that only C, O and Al can be observed in the spectrum, and the binding energy of Al 2s and 2p are centred at 119.2 and 74.7 eV, respectively. The C 1s spectrum of the high-resolution XPS (Fig. S4b†) can be divided into three main peaks at 284.8, 286.3 and 288.8 eV, which can be attributed to C–C, C–O and C=O, respectively.^48,49 Fig. S4c† displays the high-resolution XPS spectrum of O 1s. It can be deconvoluted into three peaks located at 530.8, 531.8 and 533.1 eV, which can be assigned to O₂⁻ of aluminum atoms (denoted as Al–O), metal hydroxides or hydroxyl groups (denoted as OH⁻) and chemically or physically adsorbed water (denoted as H₂O), respectively.^50,51 The FTIR spectrum of the final product is given in Fig. S5.† Vibration bands of O–H can be observed at 3437 cm⁻¹. Absorption peak observed at 1584 cm⁻¹, could be assigned to the stretching vibrations of ν(COO⁻) group. Absorption peaks are observed in the range of 1000–1600 cm⁻¹, which correspond to ρ(CH₂), ν(C–O) and ν(C–C) bands. Absorption peaks located at 856 and 692 cm⁻¹ can be attributed to Al–O bonds vibration. ²⁷Al NMR spectrum of the final product (Fig. S6†) presents three resonance signals at 58, 32, and 1 ppm, which can be assigned to Al³⁺ ion in tetrahedral (AlO₄), pentahedral (AlO₅), and octahedral (AlO₆) coordination, respectively.^52,53

Adsorption property

Considering that materials with high BET surface area and hierarchical porous structure are advantageous to toxic pollutants removal, we investigate the adsorption property of the final product for fluoride removal in water treatment. Calcination is an important factor which should be considered for materials used as adsorbents. So the adsorption performance of the precursor before and after calcination at different temperatures was investigated in our study (Fig. S7†). It can be seen that the adsorption capacity decreased with the increasing of the calcination temperature. The precursor calcined at 300 °C obtained the highest adsorption capacity (110.3 mg g⁻¹) among all the samples. Therefore, the calcination temperature of 300 °C is the optimal choice for fluoride removal.

Adsorption isotherms for fluoride removal are shown in Fig. S8a.† The calculated isotherm parameters along with correlation coefficients are given in Table S2.† The Langmuir model can fit the experimental data well according to the high correlation coefficient (R² > 0.98), indicating that the adsorption process is a monolayer adsorption process. Moreover, it can be clearly seen that the maximum adsorption capacity of as-prepared amorphous aluminum oxide (126.9 mg g⁻¹) is much higher than that of commercially available Al₂O₃ (5.7 mg g⁻¹), which is also remarkably higher than the reported values in most literatures presented in Table 1. Especially compared with other ‘non-templated’ nano-structured materials of aluminum-involved complex metal oxides, such as the Al–Ce hybrid adsorbent reported by Liu et al.,⁵⁴ the as-prepared material still showed distinct advantage. The better adsorption performance could be attributed to high BET surface area, developed porous structure and abundant active sites supplied by the 3D hierarchical structure for fluoride ion adsorption and diffusion. To the best of our knowledge, the pure aluminum oxide is usually used as adsorbent to remove heavy metal ions³⁵ and organic pollutants³⁴ from aqueous solution, however, there is rare study on the pure aluminum oxide used as adsorbent for fluoride removal. On the other hand, the aluminum also prepared as complex metal oxides and used as adsorbent to removal fluoride,⁵ heavy metal ions³⁶ and other pollutants.²⁷

Table 1 Maximum adsorption capacities of different absorbents for fluoride

Adsorbents	Qmax (mg g^−1)	C0 (mg L^−1)	pH	Ref.
Aluminum-modified hydroxyapatite (Al-HAP)	32.6	5–50	7.0	55
Manganese-oxide-coated alumina (MOCA)	2.9	2.5–30	7.0	56
Fe3O4@Al(OH)3 NPs	88.5	1–160	6.5	57
Mg-doped nano ferrihydrite	64.0	10–150	5.8	58
Mg/Fe layered double hydroxides	50.9	5–75	7	59
Al–Ce hybrid adsorbent	91.4	2–15	6	54
Iron(III)–aluminum(III)–chromium(III) ternary mixed oxide	31.9	10–80	5.6	60
Mg–Al–LDH nanoflake impregnated magnetic alginate beads (LDH-n-MABs)	32.4	2–70	5.0	5
Titanium and lanthanum oxides impregnated activated carbon (TLAC)	27.8	0–50	7.0	61
Mesoporous alumina	14.3	20–250	6.0	62
Commercially available Al2O3	5.7	5–150	7.0	This work
Amorphous aluminum oxide microspheres	126.9	5–150	7.0	This work

---

The adsorption kinetic experiment for fluoride removal using the as-prepared materials was carried out, in order to enable the adsorbent to be used in practical application. As shown in Fig. S8b,† it can be seen that a fast adsorption rate for fluoride adsorbed onto the adsorbent occurred in the initial 15 min and achieved equilibrium in about 600 min in the whole adsorption process. The fitting curve of the pseudo-second-order model was shown in the inset of Fig. S8b† and the fitting parameters of kinetic studies were listed in Table S3.† A high value of correlation coefficient (R² > 0.98) indicates that the pseudo-second-order model can well describe the adsorption process of fluoride on the as-prepared materials. A low value of k₂ indicates that the rate of adsorption process for fluoride is fast. Additionally, it was found that the theoretical adsorption capacities were in better agreement with the experimental data.

IEP is an important parameter, and the IEPs of the as-prepared product and commercially available Al₂O₃ are examined to be around 8.0 and 6.8, respectively, as shown in Fig. S9.† The difference between the IEPs of the two materials may be related to surface OH group originated from the organic residue. The higher IEP is helpful for the fluoride adsorption at neutral pH. So, a conclusion can be given that the organic residue played an important role in the fluoride adsorption. In order to give more detailed information about the adsorption performance of the as-prepared material for the fluoride, adsorption envelope was obtained in batch experiments by varying pH (pH 2–11). The adsorption envelope for fluoride on the as-prepared material is shown in Fig. S10.† It is obviously observed that the fluoride adsorption capacity increased dramatically with pH increase below pH 6.8 and then decreased sharply with pH increase, which indicated that the solution pH plays an important role in the adsorption process. At low solution pH (<IEP = 8.0), protonation will encourage the formation of positive charged surface, and the electrostatic attraction can promote the fluoride ions adsorption on adsorbent surface. However, too low pH can reduce the adsorbents stability, and thus decreased the adsorption capacity. While at high solution pH (>IEP = 8.0), deprotonation will promote the formation of negative charged surface, which can repel the fluoride ions adsorbed on the adsorbent surface. Meanwhile, the excessive amount of hydroxyl ions could compete for the active sites in the adsorption process, resulting in low adsorption capacity.

The regeneration of the used adsorbent was also investigated, as shown in Fig. S11.† Although the fluoride removal efficiency gradually decreased with the increasing regeneration times, the removal efficiency can still be maintained above 80% after five regeneration times, which indicated that the as-prepared material has a good recyclability. All of these merits make the as-prepared material as a good candidate for fluoride removal in the practical application.

Conclusions

In summary, 3D hierarchical amorphous aluminum oxide microspheres with broccoli-like structure were synthesized successfully via a facile one-pot method. It was found that the reaction time could dramatically affect the growth of the microspheres. A multistep crystal adsorption-splitting growth mechanism is proposed to understand the formation of the 3D hierarchical porous structures. When evaluated as adsorbent for fluoride removal in water treatment, the materials displayed an excellent adsorption performance, and the adsorption capacity can reach 126.9 mg g⁻¹ for fluoride. The excellent adsorption performance of the materials can be attributed to their unique 3D hierarchical porous structures, which supplied high BET surface area and abundant micropores and mesopores for fluoride ions adsorption and diffusion. Moreover, the present synthetic method may be extended to fabricate other functional metal oxides with desirable structures and compositions for various applications.

Acknowledgements

This work was supported by the National Basic Research Program of China (973 Program, no. 2011CB933700) of Ministry of Science and Technology of China, and the National Natural Science Foundation of China (21407152). The authors would like to acknowledge Prof. Ning Zhao and Prof. Li Yao for their help with paper revising.

Notes and references

1. H. Guo, Y. He, Y. Wang, L. Liu, X. Yang, S. Wang, Z. Huang and Q. Wei, J. Mater. Chem. A, 2013, 1, 7494–7499.
2. D. S. Wang and Y. D. Li, Adv. Mater., 2011, 23, 1044–1060.
3. J. Liu, W. Liu, Q. Sun, S. Wang, K. Sun, J. Schwank and R. Wang, Chem. Commun., 2014, 50, 1804–1807.
4. J. Yang, H. W. Zhang, M. H. Yu, I. Emmanuelawati, J. Zou, Z. G. Yuan and C. Z. Yu, Adv. Funct. Mater., 2014, 24, 1354–1363.
5. 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.
6. O. Amiri, H. Emadi, S. S. Mostafa Hosseinpour-Mashkani, M. Sabet and M. M. Rad, RSC Adv., 2014, 4, 10990–10996.
7. X. Fang, X. Zhao, W. Fang, C. Chen and N. Zheng, Nanoscale, 2013, 5, 2205–2218.
8. L. Bi and G. Pan, J. Mater. Chem. A, 2014, 2, 3715–3718.
9. H. Pang, F. Gao, Q. Chen, R. Liu and Q. Lu, Dalton Trans., 2012, 5862–5868.
10. T. Y. Wang, Z. Peng, Y. H. Wang, J. Tang and G. F. Zheng, Sci. Rep., 2013, 3, 2693.
11. C. Chen, H. Jian, X. X. Fu, Z. M. Ren, M. Yan, G. D. Qian and Z. Y. Wang, RSC Adv., 2014, 4, 5367–5370.
12. C. An, G. Liu, L. Li, Y. Wang, C. Chen, Y. Wang, L. Jiao and H. Yuan, Nanoscale, 2014, 6, 3223–3230.
13. G. Q. Zhang and X. W. Lou, Adv. Mater., 2013, 25, 976–979.
14. C. M. Lan, S. E. Liu, J. W. Shiu, J. Y. Hu, M. H. Lin and E. W. G. Diau, RSC Adv., 2013, 3, 559–565.
15. R. Gao, Y. Cui, X. Liu, L. Wang and G. Cao, J. Mater. Chem. A, 2014, 2, 4765–4770.
16. D. H. Chen, L. Cao, T. L. Hanley and R. A. Caruso, Adv. Funct. Mater., 2012, 22, 1966–1971.
17. H. J. Yin, S. L. Zhao, J. W. Wan, H. J. Tang, L. Chang, L. C. He, H. J. Zhao, Y. Gao and Z. Y. Tang, Adv. Mater., 2013, 25, 6270–6276.
18. Z. Qin, H. Sun, Z. Jiang, X. Jiao and D. Chen, CrystEngComm, 2013, 15, 897–902.
19. L. Cao, D. Chen and R. A. Caruso, Angew. Chem., Int. Ed., 2013, 52, 10986–10991.
20. X. W. Lou, Y. Wang, C. L. Yuan, J. Y. Lee and L. A. Archer, Adv. Mater., 2006, 18, 2325–2329.
21. Y. H. Xiao, S. J. Liu, F. Li, A. Q. Zhang, J. H. Zhao, S. M. Fang and D. Z. Jia, Adv. Funct. Mater., 2012, 22, 4052–4059.
22. Q. Liu, Y. Zhou, Z. Tian, X. Chen, J. Gao and Z. Zou, J. Mater. Chem., 2012, 22, 2033–2038.
23. M. H. Oh, T. Yu, S.-H. Yu, B. Lim, K.-T. Ko, M.-G. Willinger, D.-H. Seo, B. H. Kim, M. G. Cho, J.-H. Park, K. Kang, Y.-E. Sung, N. Pinna and T. Hyeon, Science, 2013, 340, 964–968.
24. J. Yu, H. Guo, S. A. Davis and S. Mann, Adv. Funct. Mater., 2006, 16, 2035–2041.
25. X. L. Yang, X. Y. Wang, Y. Q. Feng, G. Q. Zhang, T. S. Wang, W. G. Song, C. Y. Shu, L. Jiang and C. R. Wang, J. Mater. Chem. A, 2013, 1, 473–477.
26. W. Cai, J. Yu and M. Jaroniec, J. Mater. Chem., 2010, 20, 4587–4594.
27. S. I. Lo, P. C. Chen, C. C. Huang and H. T. Chang, Environ. Sci. Technol., 2012, 46, 2724–2730.
28. C. Y. Cao, P. Li, J. Qu, Z. F. Dou, W. S. Yan, J. F. Zhu, Z. Y. Wu and W. G. Song, J. Mater. Chem., 2012, 22, 19898–19903.
29. J. Zhang, S. J. Liu, J. Lin, H. S. Song, J. J. Luo, E. M. Elssfah, E. Ammar, Y. Huang, X. X. Ding, J. M. Gao, S. R. Qi and C. C. Tang, J. Phys. Chem. B, 2006, 110, 14249–14252.
30. D. B. Kuang, Y. P. Fang, H. Q. Liu, C. Frommen and D. Fenske, J. Mater. Chem., 2003, 13, 660–662.
31. A. Mahapatra, B. G. Mishra and G. Hota, Ind. Eng. Chem. Res., 2013, 52, 1554–1561.
32. W. Cai, J. Yu, S. Gu and M. Jaroniec, Cryst. Growth Des., 2010, 10, 3977–3982.
33. L. M. Zhang, T. A. Liu, X. Zhao, N. Qian, P. Xiong, W. J. Ma, W. C. Lu, Y. F. Gao and H. J. Luo, CrystEngComm, 2014, 16, 6126–6134.
34. J. Xiao, H. Ji, Z. Shen, W. Yang, C. Guo, S. Wang, X. Zhang, R. Fu and F. Ling, RSC Adv., 2014, 4, 35077–35083.
35. Y.-X. Zhang, Y. Jia, Z. Jin, X.-Y. Yu, W.-H. Xu, T. Luo, B.-J. Zhu, J.-H. Liu and X.-J. Huang, CrystEngComm, 2012, 14, 3005–3007.
36. Y.-X. Zhang, X.-Y. Yu, Z. Jin, Y. Jia, W.-H. Xu, T. Luo, B.-J. Zhu, J.-H. Liu and X.-J. Huang, J. Mater. Chem., 2011, 21, 16550–16557.
37. P. J. Chupas, K. W. Chapman and G. J. Halder, J. Am. Chem. Soc., 2011, 133, 8522–8524.
38. L. Zhang and Y. J. Zhu, J. Phys. Chem. C, 2008, 112, 16764–16768.
39. G. R. Jenness, M. A. Christiansen, S. Caratzoulas, D. G. Vlachos and R. J. Gorte, J. Phys. Chem. C, 2014, 118, 12899–12907.
40. E. Yu, H. J. Lee, T.-J. Ko, S. J. Kim, K.-R. Lee, K. H. Oh and M.-W. Moon, Nanoscale, 2013, 5, 10014–10021.
41. Y. Su, H. Cui, Q. Li, S. A. Gao and J. K. Shang, Water Res., 2013, 47, 5018–5026.
42. J. Tang and A. P. Alivisatos, Nano Lett., 2006, 6, 2701–2706.
43. X.-F. Shen and X.-P. Yan, Angew. Chem., Int. Ed., 2007, 46, 7659–7663.
44. L. S. Li, N. J. Sun, Y. Y. Huang, Y. Qin, N. Zhao, J. N. Gao, M. X. Li, H. H. Zhou and L. M. Qi, Adv. Funct. Mater., 2008, 18, 1194–1201.
45. J. Sheng, L. H. Hu, W. X. Li, L. E. Mo, H. J. Tian and S. Y. Dai, Sol. Energy, 2011, 85, 2697–2703.
46. C. Pilapong, T. Thongtem and S. Thongtem, J. Alloys Compd., 2010, 507, L38–L42.
47. H. Deng, C. Liu, S. Yang, S. Xiao, Z.-K. Zhou and Q.-Q. Wang, Cryst. Growth Des., 2008, 8, 4432–4439.
48. X. Zhang, Y. Chen, H. Liu, Y. Wei and W. Wei, CrystEngComm, 2013, 15, 6184–6190.
49. V. Leon, M. Quintana, M. A. Herrero, J. L. G. Fierro, A. d. l. Hoz, M. Prato and E. Vazquez, Chem. Commun., 2011, 47, 10936–10938.
50. G. Leftheriotis, S. Papaefthimiou, P. Yianoulis, A. Siokou and D. Kefalas, Appl. Surf. Sci., 2003, 218, 276–281.
51. R. Azimirad, O. Akhavan and A. Z. Moshfegh, Vacuum, 2011, 85, 810–819.
52. Q. Yuan, A.-X. Yin, C. Luo, L.-D. Sun, Y.-W. Zhang, W.-T. Duan, H.-C. Liu and C.-H. Yan, J. Am. Chem. Soc., 2008, 130, 3465–3472.
53. A. Düvel, E. Romanova, M. Sharifi, D. Freude, M. Wark, P. Heitjans and M. Wilkening, J. Phys. Chem. C, 2011, 115, 22770–22780.
54. H. Liu, S. B. Deng, Z. J. Li, G. Yu and J. Huang, J. Hazard. Mater., 2010, 179, 424–430.
55. Y. L. Nie, C. Hu and C. P. Kong, J. Hazard. Mater., 2012, 233, 194–199.
56. S. M. Maliyekkal, A. K. Sharma and L. Philip, Water Res., 2006, 40, 3497–3506.
57. X. L. Zhao, J. M. Wang, F. C. Wu, T. Wang, Y. Q. Cai, Y. L. Shi and G. B. Jiang, J. Hazard. Mater., 2010, 173, 102–109.
58. M. Mohapatra, D. Hariprasad, L. Mohapatra, S. Anand and B. K. Mishra, Appl. Surf. Sci., 2012, 258, 4228–4236.
59. D. Kang, X. Yu, S. Tong, M. Ge, J. Zuo, C. Cao and W. Song, Chem. Eng. J., 2013, 228, 731–740.
60. K. Biswas, K. Gupta, A. Goswami and U. C. Ghosh, Desalination, 2010, 255, 44–51.
61. C. Jing, J. Cui, Y. Huang and A. Li, ACS Appl. Mater. Interfaces, 2012, 4, 714–720.
62. G. Lee, C. Chen, S. T. Yang and W. S. Ahn, Microporous Mesoporous Mater., 2010, 127, 152–156.

---

Footnote

† Electronic supplementary information (ESI) available: Experiment details. SEM-EDX, mapping, BET, XPS, FTIR and ²⁷Al NMR characterization results of the as-prepared 3D hierarchical amorphous aluminum oxide microspheres. Adsorption capacity of the precursor before and after calcination at 300 °C, 400 °C and 500 °C for fluoride. Adsorption envelope for fluoride and regeneration test for the as-prepared material BET surface areas of the as-prepared precursor calcined at different temperatures. IEPs of the final product and the commercially available Al₂O₃. BET surface areas of the as-prepared precursor calcined at different temperatures. Fitting parameters of Langmuir and pseudo-second-order model for fluoride adsorption on the as-prepared materials and the commercially available Al₂O₃. See DOI: 10.1039/c4ra13688h

---