Fe₃O₄/SiO₂/C nanocomposite as a high-performance Fenton-like catalyst in a neutral environment

Abstract

The traditional Fenton system (Fe²⁺–H₂O₂) only works in an acidic environment and produces a large quantity of sludge. In this study, we reported that a Fe₃O₄/SiO₂/C nanocomposite (FSCNC) could be used as a high-performance Fenton-like catalyst for the decoloration of methylene blue (MB). To prepare FSCNC, SiO₂ was precipitated on Fe₃O₄ cores by the hydrolysis of tetraethyl orthosilicate, and the deposition of carbon was via the hydrothermal dehydrogenation of glucose. FSCNC showed much higher catalytic activity than naked Fe₃O₄ at a neutral pH of 7.5. Efficient decoloration of MB was achieved within 15 min in the FSCNC–H₂O₂ system. The FSCNC–H₂O₂ system worked well in the pH range of 3.5–9.5 and showed good resistance to radical scavengers tertiary butanol and ethanol. Higher H₂O₂ concentration and temperature were preferred to achieve faster kinetics. The regeneration of FSCNC was easily achieved by washing the catalyst and about 70% of the initial activity was retained after 8 cycles. The implication to the future applications of FSCNC as a Fenton-like catalyst is discussed.

---

Introduction

Advanced oxidation processes (AOP) are widely applied water treatment technologies that are suitable for electron-enriched organic pollutants.^1–3 Highly reactive free radicals are produced to oxidize the pollutants in AOP. Although the operating cost of AOP is more expensive than the activated sludge method, AOP could treat those very toxic or hardly biotransferred pollutants. Among the novel AOP technologies, the Fe²⁺–H₂O₂ Fenton system is the most used one, and is fast, efficient and facile.^1–3 However, the Fe²⁺–H₂O₂ system has several vital disadvantages. First, the optimized pH is 3, which requires the acidification/neutralization of water before/after the treatment. Second, a stoichiometric amount of iron is added but is very hard to remove after the reaction. Third, a large quantity of sludge is generated and further treatment of sludge is unavoidable. Therefore, alternative catalysts for the Fenton reaction are pursued to overcome the aforementioned drawbacks of the Fe²⁺–H₂O₂ system.

Heterogeneous Fenton-like catalysts attract the great interest in recent years.^4–6 In particular, nanoparticles (NPs) containing Fe or other metal elements have shown great potential in catalyzing the decomposition of H₂O₂. For instance, efficient catalysis was achieved by using Fe NPs,⁷ Fe₃O₄ NPs,⁸ Mn₃O₄ NPs,⁹ etc. In many cases, light, sonication or electricity are supplied to accelerate the reaction process.^10–12 More recently, Fe₃O₄@SiO₂ NPs was reported to catalyze the decomposition of H₂O₂ under various pHs without external energy supply.¹³ However, the reaction kinetics of Fe₃O₄@SiO₂–H₂O₂ system was slow and the regeneration of Fe₃O₄@SiO₂ was poor.

Herein, we reported the preparation of Fe₃O₄/SiO₂/C nanocomposites (FSCNC) as high-performance Fenton-like catalyst. FSCNC was prepared by the precipitation of SiO₂ on Fe₃O₄ cores via the hydrolysis of tetraethyl orthosilicate (TEOS), and then the deposition of carbon was achieved by the hydrothermal dehydrogenation of glucose. After the careful characterizations, the catalytic performance of FSCNC was compared with naked Fe₃O₄ at neutral pH of 7.5. The decoloration of methylene blue (MB) was achieved within 15 min in FSCNC–H₂O₂ system, while the decoloration was nearly unrecognizable in Fe₃O₄–H₂O₂ system. Although the decoloration was radical reaction in nature, FSCNC–H₂O₂ system showed good resistance to radical scavengers (tertiary butanol and ethanol). The influences of pH, H₂O₂ concentration and temperature were investigated. The regeneration of FSCNC was easily achieved by washing with water and about 70% of the initial activity retained after 8 cycles. The implication to the practical applications of FSCNC–H₂O₂ system in water treatment is discussed.

Experimental

Materials

FeCl₂·4H₂O was purchased from Damao Chemical Reagent Co., Ltd, China. FeCl₃·6H₂O was bought from Bodi Chemical Engineering Co., Ltd, China. TEOS was obtained from Jinshan Chemical Reagent Co., Ltd, China. Glucose was purchased from Beijing Yili Fine Chemical Co., Ltd, China. MB was obtained from Sinopharm Chemical Reagent Co., Ltd, China. Other chemicals were of analytical grade. All reagents were used without further purification.

Preparation of FSCNC

FeCl₂·4H₂O (0.816 g) and FeCl₃·6H₂O (1.684 g) were dissolved in 50 mL water and the pH value was adjusted to 12 by adding NaOH solution (3 mol L⁻¹) under vigorous stirring. The starting ratio of Fe³⁺/Fe²⁺ was 1.52, which was reported to give higher magnetic performance.¹⁴ After the co-precipitation (1 h), Fe₃O₄ NPs were collected and washed with deionized water for three times. Upon drying, part of Fe₃O₄ NPs (0.1 g for each batch) were re-dispersed in 1 mL of water, and then 10 mL of isopropanol and ammonia (90 μL) were added. The mixture was sonicated for 10 min and TEOS (0.025 g) was added. The mixture was shaken at 100 rpm under 35 °C for 5 h. The obtained Fe₃O₄/SiO₂ composite was magnetically separated, washed and dried. The dry Fe₃O₄/SiO₂ composite from each batch and 0.275 g glucose were added into 50 mL water. After transferring into a Teflon tube, the mixture was hydrothermally treated at 160 °C for 6 h. FSCNC was magnetically separated, washed by water for three times, and dried under vacuum overnight. To optimize the glucose amount, different weights of glucose were adopted to produce several FSCNC samples.

FSCNC was characterized by transmission electron microscopy (TEM, Tecnai G2 20, FEI, USA), X-ray photoelectron spectroscopy (XPS, Kratos, UK), Brunauer–Emmett–Teller (BET) technique (ASAP2010, Micromeritics, USA), infrared spectrometer (IR, Magna-IR 750, Nicolet, USA) and magnetometer (MPMS XL-7 tesla, Quantum Design, USA).

Decoloration of MB in FSCNC–H₂O₂ system

To decolorize MB, 20 mL of MB (pH 7.5, 50 mg L⁻¹) was incubated with 20 mg of FSCNC for 1 h (100 rpm at 35 °C) on a thermostat shaker. Before the addition of H₂O₂, 50 μL of the supernatant was collected for the absorbance (A₀) measurement at 664 nm on a UV-Vis spectrometer (UV1800, PGeneral, China). Then, 1.0 mL of H₂O₂ (30%) was added to the mixture and incubated at 35 °C. At each time interval of 5 min, 50 μL of the supernatant was collected for absorbance (A_t) measurement. The decoloration efficiency was calculated as (1 − A_t/A₀) × 100%. For comparison, the catalytic activity of naked Fe₃O₄ NPs was evaluated following the same protocol. As a control, the protocol was performed without adding any catalyst. The chemical oxygen demand (COD) was measured using Hach reagent (low range 3–150 mg L⁻¹) on Hach DR900. The kinetics was analyzed by eqn (1).

−ln(C/C₀) = kt (1)

Tolerance against radical scavenger

To verify the existence of ˙OH radicals, 100 mL sample solution was collected from the reaction solution (FSCNC–H₂O₂ system without adding MB) at 5 min. It was mixed with 20 mL of 0.2 mol L⁻¹ 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) to form DMPO–OH adduct. The electron spin resonance (ESR) spectrum was recorded on a JEOLJES FA200 facility with microwave bridge (receiver gain: 1 × 10⁵; modulation amplitude: 2 gauss; microwave power: 4 mW; modulation frequency: 100 kHz).

To evaluate the tolerance against radical scavenger tertiary butanol, 20 mg of FSCNC and 1.0 mL of H₂O₂ were used to decolorize 20 mL of MB (pH 7.5, 50 mg L⁻¹) at 35 °C in the presence of different concentrations of tertiary butanol (0, 0.26, 0.52 and 1.05 mol L⁻¹). FSCNC of different SiO₂ contents were prepared using different TEOS amounts and the tolerance against tertiary butanol was tested. Similarly, Fe₃O₄@C and Fe₃O₄@SiO₂ were prepared and tested following the same protocol.^13,15

To evaluate the tolerance against radical scavenger ethanol, 20 mg of FSCNC and 1.0 mL of H₂O₂ were used to decolorize 20 mL of the ethanol solution (ethanol content: 0–100 v/v%) of MB (pH 7.5, 50 mg L⁻¹) at 35 °C.

Influencing factors

To investigate the influence of pH, 20 mg of FSCNC and 1.0 mL of H₂O₂ were used to decolorize 20 mL of MB (50 mg L⁻¹) of different pH values (3.5–9.5) at 35 °C.

To evaluate the influence of H₂O₂ concentration, 20 mg of FSCNC was used to decolorize 20 mL of MB (pH 7.5, 50 mg L⁻¹) at 35 °C in the presence of different concentrations of H₂O₂ (0.025–0.91 mol L⁻¹).

To quantify the influence of temperature, 20 mg of FSCNC and 1.0 mL of H₂O₂ were used to decolorize 20 mL of MB (pH 7.5, 50 mg L⁻¹) at different temperature (0–45 °C).

Recycling

To recycle FSCNC, the used FSCNC samples were magnetically separated with a magnet. The separated FSCNC samples were washed with deionized water by stirring for 20 min. After washing for three times, the recycled FSCNC samples were dried and the catalytic activity was determined following aforementioned protocol. The procedure was repeated to reach the cycles of 8.

Results and discussion

Characterization of FSCNC

FSCNC was black powder upon drying. As shown in Fig. 1, FSCNC aggregated seriously due to the magnetic attraction. The particle diameters were in the range of 5–20 nm with the average diameter of 8.0 ± 2.3 nm (Fig. 1a). At higher resolution, the lattice planes were roughly recognized. The distance of 0.25 nm should be assigned to the (311) lattice planes of Fe₃O₄ (Fig. 1b).¹⁶ It was hardly to see core–shell structure, thus, FSCNC should be regarded as the composites of Fe₃O₄, SiO₂ and C. The small particulates were consistent with the big surface area. According to BET analysis, the BET surface area of FSCNC was 65.7 m² g⁻¹.

Fig. 1 TEM images of FSCNC at lower (a) and higher (b) resolutions.

The magnetic property of FSCNC was indicated by the magnetic hysteresis loop (Fig. 2a). The saturated magnetization of FSCNC was 29 emu g⁻¹. The value was 92.8 emu g⁻¹ for bulk Fe₃O₄ in the literature.¹⁷ The weaker magnetization of FSCNC should be due to the smaller sizes of particulates and the lower Fe₃O₄ content.¹⁸ The shape of magnetic hysteresis loop indicated that FSCNC was ferromagnetic, consistent with the results in the literature that nanosized Fe₃O₄ was ferromagnetic. Without magnetic field, the ferromagnetic property of FSCNC allowed better dispersion, which was good for the catalytic performance. After the reaction, FSCNC could be magnetically separated and easily collected from the system for recycling purpose. Thus, the magnetic property of FSCNC was proper for the Fenton-like catalysis.

Fig. 2 Magnetic hysteresis loop (a) and IR spectrum (b) of FSCNC.

According to the IR spectrum (Fig. 2b), there were meaningful oxygen containing groups on FSCNC. The broad band at 3442 cm⁻¹ should be assigned to –OH/–COOH groups on FSCNC, which might be due to Fe₃O₄, SiO₂ and C. The C–H bond was indicated by the small peak at 2938 cm⁻¹, which might come from the remnant residues of glucose. The peak at 1661 cm⁻¹ should be assigned to C=O bond, which should be the oxidized carbon during the hydrothermal treatment. The peak at 1100 cm⁻¹ was attributed to C–O bond.

The element contents were quantified by XPS (Fig. 3a). Fe (2.1 at%), Si (26.3 at%), C (7.4 at%) and O (64.2 at%) were identified in the XPS spectrum. Comparing to our previous report on Fe₃O₄@SiO₂, the Si content increased and Fe content decreased sharply. This suggested that part of Fe atoms were lost during the hydrothermal reduction. Probably, Fe atoms were converted into nonmagnetic forms, considering that black precipitates were observed after the magnetic separation of FSCNC. In addition, the C 1s XPS spectrum (Fig. 3b) suggested that carbon atoms were in the chemical stages of C–C (54.16%), C–O (34.46%), and C=O (11.37%). The oxidized forms of carbon atoms were consistent with the IR results. The results were also consistent with the literature reports that hydrothermal dehydrogenation of organic molecules usually reached partially oxidized carbon atoms.^19,20 The Fe 2p spectrum was shown in Fig. 3c. Fe 2p_1/2, Fe 2p_3/2 and satellite peaks were clearly presented. The Fe 2p_3/2 peak was divided into two components, Fe²⁺ (710.5 eV) and Fe³⁺ (712.1 eV).²¹ The ratio of Fe³⁺ : Fe²⁺ was 1.19, much smaller than the starting ratio of 1.52. This suggested that Fe³⁺ was partially reduced during the preparation of FSCNC. We still named the composite as Fe₃O₄/SiO₂/C, because the (311) lattice planes of Fe₃O₄ was observed under TEM.

Fig. 3 XPS spectra of FSCNC. (a) Total; (b) C 1s spectrum; (c) Fe 2p spectrum.

Catalytic activity of FSCNC

The catalytic activity of FSCNC was reflected by the decoloration of MB. The formulation of Fe₃O₄/SiO₂ followed our previous study.¹³ We optimized the glucose amount in our study. As shown in Fig. 4a, the mole ratio of Fe : glucose was optimized at 6.8 : 1. The decoloration was achieved within 15 min, much faster than that in Fe₃O₄@SiO₂–H₂O₂ system without carbon incorporation (120 min). Correspondingly, the kinetic constant k value reached maximum at Fe : glucose ratio of 6.8 : 1 (Fig. 4b).

Fig. 4 Influence of initial Fe : glucose mole ratio on the decoloration of MB in FSCNC–H₂O₂ system. (a) Decoloration efficiency; (b) kinetic constant.

The performance of FSCNC was compared with naked Fe₃O₄, too. As presented in Fig. 5, FSCNC showed very high catalytic activity (decoloration efficiency of 96%), while naked Fe₃O₄ had very low activity at pH 7.5 (decoloration efficiency of 21%). Without adding any catalyst, no decoloration of MB occurred. In addition, the adsorption had limited contribution to the decoloration during the catalysis (∼4%). After 1 h equilibrium, the MB concentration did not change much thereafter (Fig. S1†). The k value of FSCNC (0.126 min⁻¹) was much larger than those of naked Fe₃O₄ (0.007 min⁻¹) and Fe₃O₄@SiO₂ (0.020 min⁻¹).¹³ The kinetics of FSCNC–H₂O₂–MB system was also much faster than other systems with diverse high-performance nanocatalysts.^22,23 For instance, we reported that Ti-doped Fe₃O₄–H₂O₂–MB system had a k value of 0.0165 min⁻¹.²² Hsieh et al. found that the k value of FePt NPs–H₂O₂–MB system was 0.0033–0.023 min⁻¹ at pH 5.5.²³ The fast kinetics of FSCNC–H₂O₂ system should be attributed to the carbon incorporation. This was reflected by the faster kinetics of Fe₃O₄@C than Fe₃O₄@SiO₂.^13,22 Presumably, carbon with oxygen containing groups adsorbs MB better than SiO₂, due to the stronger electrostatic interaction.²⁴ And carbon is more conductive than SiO₂, which benefits the electron transfer from ion to MB.

Fig. 5 Comparison between FSCNC–H₂O₂ and Fe₃O₄–H₂O₂ systems. (a) Decoloration efficiency; (b) kinetic constant.

We measured the COD removal of MB in the FSCNC–H₂O₂ system. The initial COD of MB solution was 102 mg L⁻¹. After the decoloration, the COD decreased to 32 mg mL⁻¹. This oxidized suggested that most of MB was fully oxidized, while minor was into other colorless forms. Mechanically, the decoloration should be attributed to the radicals generated following eqn (2)–(4). The reactive radicals attacked MB molecules and oxidized them. The oxidation mainly occurred on the surface of FSCNC, because radical scavengers did not affect the performance (Fig. 6). That implied radicals reacted with MB molecules before entering the solution.

Fe²⁺ + H₂O₂ = Fe³⁺ + OH⁻ + ˙OH (2)

Fe³⁺ + H₂O₂ = FeOOH²⁺ + H⁺ (3)

FeOOH²⁺ = Fe²⁺ + HO₂˙ (4)

Fig. 6 Tolerance of FSCNC–H₂O₂ system against radical scavenger. (a) ESR spectrum of FSCNC–H₂O₂ system; (b) influence of tertiary butanol; (c) influence of different TEOS amounts at tertiary butanol concentration of 0.52 mol L⁻¹; (d) influence of ethanol.

Tolerance against radical scavenger

Typically, Fenton reaction involves radicals to oxidize pollutants. ESR spectrum of FSCNC–H₂O₂ system was recorded to verify the existence of ˙OH radicals (Fig. 6a). The ESR spectrum of FSCNC–H₂O₂ system had 4-fold characteristic peak of the typical DMPO–OH adduct with an intensity ratio of 1 : 2 : 2 : 1. Although the radicals were vital for the Fenton reaction in the oxidation of pollutants, FSCNC–H₂O₂ system showed very high tolerance against radical scavengers (tertiary butanol and ethanol). The decoloration became slightly slower upon adding tertiary butanol (Fig. 6b). The k value decreased very slightly along with the increase of tertiary butanol concentration (0.26, 0.52 and 1.05 mol L⁻¹). The decoloration efficiency was still 80% at the tertiary butanol concentration of 1.05 mol L⁻¹. It should be noted that coating Fe₃O₄ with SiO₂ or C alone did not result in the high tolerance against tertiary butanol (Fig. S2†). Fe₃O₄@C lost the catalytic capability at tertiary butanol concentration of 0.052 mol L⁻¹.¹⁵ Fe₃O₄@SiO₂ showed concentration-dependent loss of activity, too. At 1.05 mol L⁻¹, Fe₃O₄@SiO₂–H₂O₂ system had a decoloration efficiency of 39%,¹³ much lower than that of FSCNC–H₂O₂ system (80%). Possible reason might be that FSCNC had higher SiO₂ content and SiO₂ interacted with MB stronger than tertiary butanol. Thus, relative high MB concentration and low tertiary butanol concentration were formed on the FSCNC surface. As shown in Fig. 6c, increasing the usage of TEOS (resulting in higher SiO₂ contents) did show better tolerance against tertiary butanol. Activity decrease only occurred at very high TEOS usage.

Beyond the tolerance against tertiary butanol, FSCNC–H₂O₂ system also should good resistance to another radical scavenger ethanol (Fig. 6d). Ethanol is novel scavenger for ˙OH radicals. As shown in Fig. 6d, in the range of ethanol content of 0–100 v/v%. Except the slight decrease at 40% ethanol, the k values were all similar to that in pure water. In particular, in pure ethanol, the k value was even bigger than that in pure water. This might imply that the decoloration of MB occurred on the surface of FSCNC.^25,26 Otherwise, when the radicals entered the solution phase, they would be quenched immediately. The good resistance of FSCNC–H₂O₂ to radical scavengers made the system survive under more complicated situations, thus would benefit the practical applications in water treatment.

Influencing factors

One of the best properties of FSCNC was the high activity in a wide pH range. During our test, FSCNC–H₂O₂ system showed fast decoloration kinetics in pH 3.5–9.5 (Fig. 7a). The k value was 0.084 min⁻¹ at pH 3.5 and increased gradually along with the increase of pH. The wide applicable pH range of FSCNC was attributed to the incorporation of SiO₂. In our previous reports, we have demonstrated that carbon would accelerate the decoloration kinetics of Fe₃O₄, but could not shift the pH range toward neutral.¹⁵ Instead, SiO₂ could enlarge the applicable pH range of Fe₃O₄ (pH 3.5–8.5).²⁴ SiO₂ might enhance the catalytic performance by several approaches. First, amorphous SiO₂ would adsorb and concentrate MB around Fe₃O₄ surface. Second, SiO₂ might benefit the electron transfer. Third, the hydroxyl groups on SiO₂ might somehow chelate with Fe²⁺/Fe³⁺ on Fe₃O₄ surface. There are several papers suggesting that chelation would facilitate the catalytic performance of Fe₃O₄.^27–29 Considering that SiO₂ had many hydroxyl groups on the surface, it might be the hydroxyl groups cheating the Fe²⁺/Fe³⁺ ions. Despite the unclear mechanism, the wide applicable pH range of FSCNC would avoid the acidification of polluted water and the neutralization after the treatment. It would save a lot of operating cost and be friendly to the environment.

Fig. 7 Influence of pH (a), initial H₂O₂ concentration (b) and temperature (c) on the decoloration of MB in FSCNC–H₂O₂ system.

We also concerned the influence of H₂O₂ concentration on the decoloration performance of FSCNC–H₂O₂ system. As shown in Fig. 7b, increasing the H₂O₂ concentration resulted in faster decoloration, which was reasonable from the kinetics perspective. It should be noted that at higher H₂O₂ concentrations, the increase of decoloration speed was lower than the increase of H₂O₂ concentration. This might be due to the extinction of radicals with H₂O₂ following eqn (5) and (6).³⁰ Thus, too high H₂O₂ concentration was not recommended.

H₂O₂ + ˙OH → H₂O + ˙OOH (5)

˙OOH + ˙OH → H₂O + O₂ (6)

In addition, the temperature had significant influence on the decoloration kinetics. At 0 °C, the k value was 0.017 min⁻¹, much lower than that at 35 °C. Further increasing the temperature to 45 °C resulted in slightly decrease of kinetics. This might be due to the fast decomposition of H₂O₂. The over decomposed H₂O₂ led to too much radicals that self-extinction occurred rather than the oxidation of MB. Therefore, room temperature was recommended for practical applications.

Separation and regeneration

Recycling of catalyst is crucial for the practical applications, because recycling would reduce the operating cost and be more friendly to the environment. During the recycling test, FSCNC showed a decrease trend in the first 5 cycles (Fig. 8). At the fifth cycle, the decoloration efficiency was 55.8% after 30 min reaction. Surprisingly, the decoloration efficiency increased significantly at the sixth cycle to 93.8%. Till cycle of 8, the decoloration was still 82.0%. The regeneration of FSCNC might be due to the carbon incorporation. The sacrificial role of graphene oxide (GO) was reported to be crucial for the regeneration of GO–Fe₃O₄.³¹ Carbon, instead of Fe²⁺, was oxidized during the Fenton-like reaction. Nevertheless, our results indicated that FSCNC could be reused during the water treatment. To facilitate the decontamination, extra supplement of catalyst might be taken to keep the high performance.

Fig. 8 Recycling of FSCNC after the decoloration of MB.

Conclusions

In summary, FSCNC was prepared for high-performance Fenton-like catalysis, where FSCNC–H₂O₂ system could be used for the fast decoloration of dye solution at various pH values. There were several outstanding properties of FSCNC as Fenton-like catalyst, including the wide applicable pH range, the resistance against radical scavengers, the fast kinetics, the easy magnetic separation, and the facile regeneration. It is hoped that our results would benefit the development of heterogeneous Fenton-like catalysts and stimulate more interest on environmental nanotechnologies.

Acknowledgements

We acknowledge financial support from the Top-notch Young Talents Program of China, the China Natural Science Foundation (No. 201307101), and the Innovation Scientific Research Program for Graduates in Southwest University for Nationalities (No. CX2016SZ030).

Notes and references

1. L. Feng, E. D. van Hullebusch, M. A. Rodrigo, G. Esposito and M. A. Oturan, Chem. Eng. J., 2013, 228, 944–964.
2. J. Naumczyk, J. Bogacki, P. Marcinowski and P. Kowalik, Environ. Technol., 2014, 35, 541–548.
3. E. Neyens and J. Baeyens, J. Hazard. Mater., 2003, 98, 33–50.
4. S. Caudo, G. Centi, C. Genovese and S. Perathoner, Top. Catal., 2006, 40, 207–219.
5. A. D. Bokare and W. Choi, J. Hazard. Mater., 2014, 275, 121–135.
6. F. Fu, D. D. Dionysiou and H. Liu, A review, J. Hazard. Mater., 2014, 267, 194–205.
7. B. Yang, Z. Tian, L. Zhang, Y. Gou and S. Yan, J. Water Process Eng., 2015, 5, 101–111.
8. H. Niu, D. Zhang, S. Zhang, X. Zhang, Z. Meng and Y. Cai, J. Hazard. Mater., 2011, 190, 559–565.
9. Y. Wang, L. Zhu, X. Yang, E. Shao, X. Deng, N. Liu and M. Wu, J. Mater. Chem. A., 2015, 3, 2934–2941.
10. S. Valizadeh, M. H. Rasoulifard and M. S. Seyed Dorraji, Appl. Surf. Sci., 2014, 319, 358–366.
11. T. Zhou, X. Wu, J. Mao, Y. Zhang and T. Lim, Appl. Catal., B, 2014, 160, 325–334.
12. M. Luo, S. Yuan, M. Tong, P. Liao, W. Xie and X. Xu, Water Res., 2014, 48, 190–199.
13. S.-T. Yang, W. Zhang, J. Xie, R. Liao, X. Zhang, B. Yu, R. Wu, X. Liu, H. Li and Z. Guo, RSC Adv., 2015, 5, 5458–5463.
14. W. Jiang, K. Lai, H. Hu, X. Zeng, F. Lan, K. Liu, Y. Wu and Z. Gu, J. Nanopart. Res., 2011, 13, 5135–5145.
15. X. Zhang, M. He, J. Liu, R. Liao, L. Zhao, J. Xie, R. Wang, S.-T. Yang, H. Wang and Y. Liu, Chin. Sci. Bull., 2014, 59, 3406–3412.
16. F. Shi, Y. Li, Q. Zhang and H. Wang, Int. J. Photoenergy, 2012, 1155, 365–401.
17. Z. Zhang, H. Duan, S. Li and Y. Lin, Langmuir, 2010, 26, 6676–6680.
18. C. Luna, M. P. Morales, C. J. Serna and M. Vazquez, Mater. Sci. Eng., C, 2003, 23, 1129–1132.
19. L. Zhao, B. Yu, F. Xue, J. Xie, X. Zhang, R. Wu, R. Wang, Z. Hu, S.-T. Yang and J. Luo, J. Hazard. Mater., 2015, 286, 449–456.
20. L. Zhao, X. Chang, R. Liao, X. Zhang, J. Xie, B. Yu, R. Wu, R. Wang and S. Yang, Mater. Lett., 2014, 135, 154–157.
21. T. Yamashita and P. Hayes, Appl. Surf. Sci., 2008, 254, 2441–2449.
22. S.-T. Yang, L. Yang, X. Liu, J. Xie, X. Zhang, B. Yu, R. Wu, H. Li, L. Chen and J. Liu, Sci. China: Technol. Sci., 2015, 58, 858–863.
23. S. Hsieh and P. Lin, J. Nanopart. Res., 2012, 14, 956.
24. R. Wu, J. Liu, L. Zhao, X. Zhang, J. Xie, B. Yu, X. Ma, S.-T. Yang, H. Wang and Y. Liu, J. Environ. Chem. Eng., 2014, 2, 907–913.
25. J. He, X. Yang, B. Men, Z. Bi, Y. Pu and D. Wang, Chem. Eng. J., 2014, 258, 433–441.
26. L. Zhou, J. Ma, H. Zhang, Y. Shao and Y. Li, Appl. Surf. Sci., 2015, 324, 490–498.
27. X. Xue, K. Hanna, C. Despas, F. Wu and N. Deng, J. Mol. Catal. A: Chem., 2009, 311, 29–35.
28. S. Sun, X. Zeng, C. Li and A. T. Lemley, Chem. Eng. J., 2014, 244, 44–49.
29. N. Wang, L. Zhu, M. Lei, Y. She, M. Cao and H. Tang, ACS Catal., 2011, 1, 1193–1202.
30. L. Yu, X. Yang, Y. Ye and D. Wang, RSC Adv., 2015, 5, 46059–46066.
31. N. A. Zubir, C. Yacou, J. Motuzas, X. Zhang, X. Zhao and J. C. D. da Costa, Chem. Commun., 2015, 51, 9291–9293.

---

Footnote

† Electronic supplementary information (ESI) available: Catalysis vs. adsorption during the decoloration of MB; tolerance of FSCNC–H₂O₂, Fe₃O₄@C–H₂O₂ and Fe₃O₄@SiO₂–H₂O₂ systems against tertiary butanol. See DOI: 10.1039/c5ra22890e

---