Activation of persulfate by Co₃O₄ nanoparticles for orange G degradation

Abstract

Nano-Co₃O₄ was prepared by a precipitation method and successfully applied as a heterogeneous catalyst to activate persulfate (PS). The heterogeneous character of PS activation with nano-Co₃O₄ was more pronounced at neutral pH as indicated by the maximum degradation rate of orange G (OG) and the low concentration of dissolved cobalt ions. The increasing dosages of nano-Co₃O₄ and PS, and the higher temperature rapidly promoted the degradation kinetics of OG. Sulfate radicals (SO₄˙⁻) and hydroxyl radicals (·OH) were proved to be the primary oxidative species, although the intensity of DMPO-OH signals were much stronger than that of DMPO-SO₄ due to the fast transformation from DMPO-SO₄ to DMPO-OH. The catalyst presented an acceptable stability through using it in five consecutive runs. The degradation pathways of OG in nano-Co₃O₄/PS process were proposed for the first time based on LC-MS analysis. The effects of Cl⁻, NO₃⁻, HCO₃⁻, Co²⁺, Fe²⁺ and Mn²⁺ ions, which usually co-exist as water matrix chemicals with azo-dyes, on OG removal by nano-Co₃O₄/PS were examined intensively. The reactivity of potassium persulfate (PS), sodium persulfate (NaPS) and potassium peroxymonosulfate (PMS) in the presence of nano-Co₃O₄ followed the order of nano-Co₃O₄/PMS > nano-Co₃O₄/PS > nano-Co₃O₄/NaPS.

---

1 Introduction

As an advanced oxidation process (AOPs), a sulfate radical based AOPs has become a hotspot and been widely studied for the decomposition of various kinds of recalcitrant or hazardous organic compounds due to its high oxidizing ability. Sulfate radicals can be generated by activation of persulfate (PS) or peroxymonosulfate (PMS) with UV,¹ heat,² sono,³ base,⁴ granular activated carbon,⁵ quinones,⁶ phenols,⁷ transition metals,⁸ or magnetic spinel.^9,10 Among these methods, activation of PS with transition metal ions is commonly used and an efficient method to generate sulfate radicals. Among the transition metal ions (Co²⁺, Cu²⁺, Ni²⁺, Fe²⁺, Ag⁺, Ru³⁺) effective for activating PS or PMS, Co mediated decomposition of persulfate (Co/PS) is an efficient catalytic system that can form SO₄˙⁻ as the major oxidizing species.^11,12 The Co/PS system for the degradation of contaminants has shown a lot of interests due to its advantages such as high efficiency at wide pH range, small amounts of cobalt catalyst and high efficiency in both carbonate and phosphate buffer solutions.¹¹

Previous studies on Co/PS system mainly focused on homogeneous catalysis, where cobalt ions have to be separated at the end of the treatment by precipitation due to their toxic nature and thus need additional operational costs.^8,11,12 To avoid the drawback of the homogeneous Co/PS reagent and broaden the application of the reagent, Dionysiou group investigated heterogeneous activation of PMS with Co₃O₄ and found this system had good performance on degradation of 2,4-dichlorophenol.¹³ Recently, the composite oxides of cobalt and another metal element have attracted a great deal of research interest^14,15 in order to promote the performance of catalysts, for example the Co–Fe bimetallic oxides (CoFe₂O₄) (ref. 16) and Co–Mn oxides (Co_xMn_3−xO₄, x = 1, 2).¹⁷ Nano-scaled catalysts also draw much attention in recent years due to their particular physical and chemical properties, and excellent performance.^18,19 Hence reducing the diameter of Co₃O₄ to nano-scale may enhance the reactivity. Chen et al. proved that nano-Co₃O₄ was an effective activator for PMS to decompose Acid Orange 7 at both acid and neutral pH conditions.²⁰ Saputra et al. also reported that nano-Co₃O₄ of 24 nm could fast and completely remove phenol in about 20 min, at the conditions of 25 mg L⁻¹ phenol, 0.4 g L⁻¹ catalyst, 2 g L⁻¹ oxone and 25 °C.²¹ However, the size effect of Co₃O₄ on the catalytic performance was still unclear. Therefore, in this study both the synthetic nano-Co₃O₄ particles and the commercial Co₃O₄ powder were employed as catalysts to activate PS. It should be noted that most of the studies employed nano-Co₃O₄ as heterogeneous catalyst were performed with PMS as a source of SO₄˙⁻ rather than PS. Thus in this research, the performance discrepancy of PMS and PS in the presence of nano-Co₃O₄ was compared and discussed.

The degradation of target organics may occur on the catalyst surface (i.e., heterogeneous reaction) and/or in the bulk solution (i.e., homogeneous reaction). The contribution from heterogeneous or homogeneous reaction was still uncovered in the nano-Co₃O₄/PS system. The stability and reusability of the catalyst are critical in catalyzed reactions, especially for practical industrial applications. Tan et al. has reported that in nano-Fe₃O₄/PMS process, the stability of nano-Fe₃O₄ decreased significantly from the first run to the third run.²² Chen et al. declared that the stability of nano-Co₃O₄ remained almost unchanged in PMS solution. By analyzing their results, we found the activity of nano-Co₃O₄ decreased steady actually, but the slight decline was covered by the long enough reaction time and all the Acid Orange 7 was decomposed at the end of reaction. Therefore, the stability of nano-Co₃O₄ in consecutive runs needed to be confirmed.

In this study, a textile azo-dye, orange G (OG), was chosen as a model compound. Many studies about AOPs (i.e. Fenton, photo-Fenton and TiO₂ photocatalysis) used to employ OG as model mainly because it is a widely used dye, and resistant to conventional methods, like adsorption and physico-chemical and biological treatments, in sewage treatment plants.^23,24 To our knowledge, little information on OG degradation kinetics and mechanism in heterogeneous nano-Co₃O₄/PS process is available from previous studies. Hence this study not only intended to examine the influence of the experimental conditions (i.e. dosages of nano-Co₃O₄ and PS, catalyst particle size, pH and temperature), but also aimed to propose the degradation pathways of OG in nano-Co₃O₄/PS process. The primary reactive oxidants in nano-Co₃O₄/PS process were clarified. This study also tried to assess the stability of nano-Co₃O₄ in successive runs and to explore the causes of deactivation. Usually, a great amount of salts and heavy metals are employed in various dyeing processes and the inorganic ions in dyeing wastewater may affect the efficiency of dye degradation reaction. Thus, the effects of several inorganic ions that commonly occur in real dye-containing wastewater on the OG degradation were examined. The reactivity of three different sources of SO₄˙⁻ (i.e. potassium persulfate, sodium persulfate and Oxone®) in the presence of nano-Co₃O₄ was compared quantitatively for the first time.

2 Experimental

2.1 Reagents and materials

OG (98%), Co(NO₃)₂·6H₂O, N-cetyl-N,N,N-trimethyl ammonium bromide (CTAB), (CH₃CH₂)₃N (TEA), NaH₂PO₄ (99%), Na₂HPO₄ (98%), Na₂S₂O₈ (98%), potassium persulfate, sodium persulfate and Oxone® were all analytical grade and purchased from Sigma-Aldrich. Methanol (99.7%), ethanol (99.5%) and t-butyl alcohol (TBA, 99.5%) were both chromatographic pure and purchased from Sigma-Aldrich. Commercial CoO, Co₃O₄ and Co₂O₃ powder (97%), NaNO₃, NaCl, NaHCO₃, Co(NO₃)₂, Fe(NO₃)₂ and Mn(NO₃)₂ was from Sinopharm Chemical Reagent Co., China, 5,5-dimethyl-1-pyrrolidine N-oxide (DMPO) from J&K Chemical Co. All reagents were used as received, without further purification. The aqueous solutions were prepared by using Milli-Q water (resistivity ≥18.2 mΩ cm).

2.2 Preparation and characterization of nano-Co₃O₄

Nano-Co₃O₄ particles were prepared as described by Zhou et al.,¹⁹ as shown in Text S1 in the ESI.† The morphology of nano-Co₃O₄ was determined using transmission electron microscopy (TEM, Model JEM-2011, Japan). Its crystal structure was characterized by X-ray diffractometer (XRD, D8 Advance, Bruker, Germany) with Cu Kα radiation (λ = 1.5418) in the 2θ scanning range from 15 to 70°. Zeta potentials at different pH were determined with a zeta analyzer (Zetasizer Nano, Malvern Instruments Ltd., UK). X-ray photoelectron spectroscopy (XPS) measurements were performed on a PHI 5000C ESCA System with Al Kα source. All the binding energies were referenced to the contaminant C1s peak at 284.6 eV of the surface adventitious carbon. The BET surface areas were measured with the N₂ gas adsorption method on an ASAP analyzer (Micromeritics, USA). Prior to the adsorption–desorption measurements, the fresh catalyst was degassed at 300 °C in a N₂ flow for 2 h.

2.3 Experimental procedures

Batch experiments were conducted to explore the influences of initial pH, PS and nano-Co₃O₄ dosages, diameter of catalysts, pH and temperature on the performance of OG degradation. The experiments were performed open to the air and in a series of 500 mL borosilicate glass jars that were placed in a water bath to keep temperature constant. The solution was mixed at 300 rpm using a magnetic stirrer. With this stirring intensity, the catalyst could be evenly distributed in the solution. pH of the reaction solution was adjusted to predetermined value with NaOH or HClO₄. At the given time intervals, the samples aliquots were collected with syringes and mixed immediately with appropriate amounts of methanol to quench any further oxidation reactions.⁸ The samples were filtered through a 0.22 μm membrane filter before measurement. All experiments were run in duplicates, and all points in the figures are the mean of the results and error bars represent standard deviation of the means.

For TOC measurement, sodium thiosulfate was chosen as the quencher to minimize any interference of quenching agent in TOC analysis. In the consecutive runs, the used catalyst was collected by filtration, rinsed with Milli-Q water for several times, and then dried in vacuum at 40 °C for 12 h. Then, 2.0 mM PS was added with dry catalyst to the fresh OG-containing solution to initiate the next treatment.

The details of the experiments to assess the leaching test of nano-Co₃O₄ and the catalytic performance of dissolved Co ions were presented in Text S2 in ESI.† Quenching experiments were conducted with the addition of methanol and TBA for identifying the primary radical species. Electron paramagnetic resonance (EPR) experiments were performed using DMPO as spin-trapping agent, as shown in Text S3.†

2.4 Analytical methods

The concentration of OG was analyzed with a Purkinje TU-1902 automatic scanning UV-Vis spectrophotometers with a spectrometric quartz cell (1 cm). The maximum absorbance wavelength of OG was observed at 478 nm. The concentrations of leached Co ions were determined by ICP-AES (Agilent 4200) after microwave digestion in the mixture of nitric acid and hydrogen peroxide. Total organic carbon (TOC) was monitored using a Shimadzu TOC analyzer (model TOC-VCPN, Japan) to identify the mineralization of OG. EPR experiments were performed on a Bruker A200 spectrometer (Germany). The UPLC-QTOF-MS was used to detect the intermediates of OG degradation. In this study, the mass spectrometer was operated in the m/z 50–500 range for LC-MS. Samples were eluted at 0.5 mL min⁻¹ through a Acquity UPLC BEH C18 column (2.1 mm × 100 mm, 1.7 μm) with the following gradient: from 1/99 (acetonitrile/0.2% formic acid) to 95/5 in 10 min, which was then maintained for 3 min.

3 Results and discussion

3.1 Characterization of nano-Co₃O₄

TEM images of nano-Co₃O₄ before and after catalytic reaction in five successive runs are shown in Fig. S1(a) and (b),† respectively. The virgin catalysts with an average diameter of 50 nm were mostly quasi-spherical. The morphology of spent Co₃O₄ nanoparticles remained virtually unchanged after five successive runs. Fig. S2† presented the XRD patterns of nano-Co₃O₄ before and after use for five times. Peaks of virgin and spent nano-Co₃O₄ appeared at 2θ = 18.9°, 31.2°, 36.7°, 38.5°, 44.7°, 55.7°, 59.2° and 65.0°, which matched well with the standard XRD pattern of Co₃O₄ (JCPDS no. 47-1049),^25,26 indicating that after five successive runs the main crystalline for nano-Co₃O₄ remained unchanged. N₂ adsorption–desorption isotherms were employed to study the surface area of virgin nano-Co₃O₄ and its spent counterpart after five successive runs. The BET surface areas were 18.1 and 17.8 m² g⁻¹ for the virgin and spent nano-Co₃O₄, respectively.

The surface chemical composition and oxidation state of the virgin and spent nano-Co₃O₄ was investigated by XPS (Fig. 1). The XPS survey (0–1300 eV) in Fig. 1(A) shows the presence of oxygen and cobalt for nano-Co₃O₄. Fig. 1(B) shows the high resolution spectrum of the Co2p peak for the catalyst. One can see the presence of the two spin orbit components of Co (Co2p_3/2 at 780.1 eV and Co 2p_1/2 at 795.5 eV) and the presence of associated shake-up satellites at higher energy (789.0 and 804.5 eV) for both Co peaks.²⁹ The energy between the Co2p_1/2 and Co2p_3/2 peaks for nano-Co₃O₄ is 15.4 eV, and it is in agreement with that previously reported for Co₃O₄.^27,28 The satellite peaks in the Co2p spectra are an important signal for distinguishing the bonding valence of cobalt oxide compounds. The lower intensity of the shake-up satellites at 9 eV from the main spin–orbit components in Fig. 1(B) has identified the cobalt in as-prepared catalyst was Co₃O₄ and not CoO.^27,28 For nano-Co₃O₄, the peak of Co 2p_3/2 level was deconvoluted into two peaks concentrated on at binding energy of 780.3 eV and 785.2 eV (Fig. 1(C)), which can be assigned to surface Co(III) and Co(II) species,²⁹ respectively. The recording of a weak signal at binding energy of 789.8 eV indicates the presence of surface Co(II) species. Quantitative analyses of the Co 2p_3/2 XPS spectra can give the surface cobalt compositions of the samples. Apparently, the surface Co(III)/Co(II) molar ratio (3.68) of virgin Co₃O₄ was lower than that (3.82) used after 5 times. As shown in Fig. 1(D), the observed position around 530 eV was the typical XPS spectrum of O1s region. The O1s peak was deconvoluted into three spectral bands at 529.6, 530.6 and 532.9 eV in the virgin nano-Co₃O₄. The most intense peak at 529.6 eV was attributed to the lattice oxygen (O_latt) in the metal oxide.²⁹ The 530.6 eV of binding energy was due to the hydroxide (Co–OH) in the surface hydroxyls,²⁹ and the relatively small peak at 532.9 eV represented physically adsorbed H₂O.^30,31 Based on the XPS analysis, the intensity of adsorbed water molecules in the spent nano-Co₃O₄ increased, although the catalyst was dried in the vacuum oven for 12 h before XPS analysis.

Fig. 1 Survey spectra (A), Co 2p XPS spectra (B), Co 2p_3/2 XPS spectra (C) and O 1s XPS spectra (D) of virgin and spent nano-Co₃O₄.

3.2 Effect of reaction conditions

3.2.1 Effects of Co₃O₄ particle size. In order to compare the catalytic performance of nano-Co₃O₄ to that of commercial Co₃O₄ powder, both the nano-Co₃O₄ about 50 nm and Co₃O₄ powder with particle size of 5 μm were applied for activation of PS. Negligible OG was degraded by PS alone (2.0 mM) (Fig. S3†) and less than 2.5% of OG was removed by nano-Co₃O₄ (Fig. S4†) or Co₃O₄ alone (2.0 g L⁻¹) (data not shown) over 3 h. The addition of 0.5 g L⁻¹ nano-Co₃O₄ to PS solution resulted in a very rapid and pronounced degradation of OG with 90.7% removal in 3 h at pH 7.0, indicating that nano-Co₃O₄ could activate PS effectively to initiate OG oxidation. In Co₃O₄/PS process, only 38.7% of OG was removed (Fig. 2). The higher catalytic performance of nano-Co₃O₄ might derive from its smaller particle size and higher surface area. Therefore, the specific rate constant (k_SA), yielded from the normalization of k (s⁻¹) data to the catalyst surface area (eqn (1)), is of great help in investigating the reactivity of catalysts.

k_SA = k/S_BET (1)

where k_SA is the specific reaction rate constant (g m⁻² s⁻¹) of OG removal, and S_BET is the surface area of catalyst (m² g⁻¹). The obtained close k_SA values for Co₃O₄ and nano-Co₃O₄ are 9.18 × 10⁻⁶ and 1.23 × 10⁻⁵ g m⁻² s⁻¹ (Table S1†), respectively, indicating that the catalytic active sites available to PS depended largely on the catalyst surface area.³²

Fig. 2 Effect of catalyst particle size on OG degradation in Co₃O₄/PS and nano-Co₃O₄/PS processes. (Reaction conditions: [OG] = 0.1 mM, pH = 7.0 ± 0.1, T = 25 °C).

3.2.2 Effect of nano-Co₃O₄ dosage. Fig. 3(a) illustrated the OG degradation by nano-Co₃O₄ activated PS at various nano-Co₃O₄ dosages. The OG degradation well followed exponential rate law, indicating that the reaction was a pseudo-first-order (R² > 0.998) with respect to OG. The overall rate law for OG degradation can be expressed as eqn (2),where [OG] is the OG concentration at any specific time t and k (s⁻¹) is the pseudo-first-order degradation rate constant. The k (s⁻¹) rapidly increased linearly from 6.3 × 10⁻⁵ s⁻¹ to 8.58 × 10⁻⁴ s⁻¹ as the catalyst dosage increased from 0.1 g L⁻¹ to 2.0 g L⁻¹. The significantly enhanced degradation could be attributed to the increased production of active radical species after introduction of the catalyst.^22,33 A linear relationship was established between k (s⁻¹) and nano-Co₃O₄ dosage, as shown in the inset of Fig. 3(a). Similar linear formula were also reported in prior studies that cobalt-MCM₄₁ activated PS to degrade caffeine,³⁴ nano-Co₃O₄ activated H₂O₂ to decompose 2-chlorophenol³⁵ and Fe₃O₄ nanoparticles activated PMS to degrade acetaminophen.²²

Fig. 3 (a) Effect of nano-Co₃O₄ dosage on OG degradation in nano-Co₃O₄/PS system. Inset indicates pseudo-first-order rate constants of OG degradation versus dosage of nano-Co₃O₄. (Reaction conditions: [OG] = 0.1 mM, [PS] = 2.0 mM, pH = 7.0 ± 0.1, T = 25 °C). (b) Effect of PS dosage on OG degradation in nano-Co₃O₄/PS system. Inset indicates pseudo-first-order rate constants of OG degradation versus dosage of PS. (Reaction conditions: [OG] = 0.1 mM, [nano-Co₃O₄] = 0.5 g L⁻¹, pH = 7.0 ± 0.1, T = 25 °C).

3.2.3 Effect of PS dosage. The effect of variable PS concentration on the transformation of OG was studied in the presence of nano-Co₃O₄, as shown in Fig. 3(b). The increased dosage of PS from 0.5 to 2.0 mM and then to 4.0 mM understandably accelerated the degradation of OG with the k (s⁻¹) increasing from 7.5 × 10⁻⁵ to 3.57 × 10⁻⁴ s⁻¹. A linear relationship could be established between k (s⁻¹) and PS dosage in the range of 0.5–2.0 mM, as shown in the inset of Fig. 3(b), which indicated that the availability of PS was the limiting factor controlling the yield of radicals at a low PS dose. However, the PS contribution became considerably less important with further increasing the dosage of PS from 2.0 mM to 4.0 mM, indicating that the active sites of fixed catalyst concentration gradually became the limiting factor. Similar phenomenon was also observed in Fe₃O₄ nanoparticles activating PMS to degrade acetaminophen.²²

3.2.4 Effects of initial pH. The effect of initial pH on OG degradation in nano-Co₃O₄/PS system is presented in Fig. 4(a). The OG adsorption onto nano-Co₃O₄ (Fig. S2†) and oxidation by PS alone (Fig. S3†) was less than 5% at pH 3.0–10.0, which indicated the adsorption by nano-Co₃O₄ and the oxidation by PS alone could be ignored in the evaluation of pH effect. As pH increased from 3.0 to 9.0, the k (s⁻¹) dramatically climbed from 1.35 × 10⁻⁴ s⁻¹ to 3.18 × 10⁻⁴ s⁻¹, as shown in the inset of Fig. 4(a). With further increasing pH from 9.0 to 10.0, the k (s⁻¹) dropped from 3.18 × 10⁻⁴ s⁻¹ to 2.45 × 10⁻⁴ s⁻¹. This could be attributed to the zeta potentials and surface charges of the catalysts. The point of zero charge (PZC) was determined to be around 8.5 for nano-Co₃O₄ catalyst (Fig. S5†). When nano-Co₃O₄ was dispersed in water then the surface became cationic in nature, which would increase the more coverage of hydroxyl groups (–OH) from H₂O. Pu et al. reported that the uncharged surface hydroxyl groups of nano-Co₃O₄ were the main active sites in promoting persulfate decomposition to generate sulfate radicals SO₄˙⁻.³⁶ Thus with increasing pH from 3.0 to 9.0, the kinetics of OG degradation sharply increased and reached it maximum at the pH near PZC. In addition, at an acidic condition, the H-bond formation between H⁺ and the O–O group of S₂O₈²⁻ would be significant, thus hindering the reaction between S₂O₈²⁻ and positively charged catalyst surface.²² When the solution pH was higher than PZC, the catalyst surface became anionic and had a higher electronic force to repel the negative PS anion, so that less PS could reach the catalyst surface and OG degradation was decreased from pH 9.0 to 10.0.

Fig. 4 (a) Degradation of OG in the nano-Co₃O₄/PS system at different pH. Inset indicates pseudo-first-order rate constants of OG degradation versus pH; (b) dissolution of Co from nano-Co₃O₄ under various pH. (Reaction conditions for (a and b): [OG]₀ = 0.1 mM, [PS]₀ = 2.0 mM, [nano-Co₃O₄] = 0.5 g L⁻¹, T = 25 °C). (c) Degradation of OG with homogeneous Co²⁺/PS at various pH. The dosage of Co²⁺ was set according the maximum dissolved Co ions at the end of reactions in (b). (Reaction conditions for (c): [OG]₀ = 0.1 mM, [PS]₀ = 2.0 mM, T = 25 °C).

The degradation of OG occurs on the catalyst surface (i.e., heterogeneous reaction) and/or in the bulk solution (i.e., homogeneous reaction). Which is the main process, heterogeneous nano-Co₃O₄/PS or homogeneous Co/PS? To clarify the mechanism, the concentration of dissolved Co ions leached from nano-Co₃O₄ catalyst at different pH was measured and the results are presented in Fig. 4(b). It is shown that under acidic condition, Co dissolved slowly from nano-Co₃O₄, reaching a value of 0.59 mg L⁻¹ and 0.25 mg L⁻¹ after 3 h at pH 3.0 and 5.0. Thus, nano-Co₃O₄ was slightly unstable under acidic condition. While at neutral or alkaline condition (pH 7.0, 9.0 and 10.0), dissolved Co was always below 0.05 mg L⁻¹, thus very limited Co dissolved from nano-Co₃O₄. The results were consistent with that of Dionysiou group¹³ and Chen et al.²⁰

To evaluate the catalytic contribution from dissolved Co, homogeneous experiments with introduction of Co²⁺ into PS solution were conducted. As can be observed from Fig. 4(c), the degradation rate of OG was much faster when nano-Co₃O₄ was added. Thus it can be concluded safely that the main catalytic effect is caused by nano-Co₃O₄, not dissolved Co ions, whether at acidic or neutral pH. The heterogeneous catalysis of nano-Co₃O₄ was especially prominent under neutral pH, since its homogeneous counterpart was limited owing to the low concentration of dissolved Co ions. Most dyestuff wastewaters are neutral since they usually contain large amount of buffer salts.²⁰ Thus the nano-Co₃O₄/PS system is quite acceptable from application point of view due to its very low dissolution of Co ions and high reactivity.

3.2.5 Effect of temperature. The kinetics of OG degradation by PS in the presence of nano-Co₃O₄ at 15–35 °C is depicted in Fig. S6.† It is clearly seen that the reaction temperature positively affected the rate of OG degradation. The k (s⁻¹) increased from 1.48 × 10⁻⁴ s⁻¹ to 3.52 × 10⁻⁴ s⁻¹ with temperature increasing from 15 °C to 35 °C. Furthermore, the activation energy (E_a) was determined by plotting ln k against 1/T, according to Arrhenius equation (eqn (3)):where k is the rate constant (s⁻¹), R is the universal gas constant (8.314 J mol⁻¹ K⁻¹) and A is a constant. The E_a value for nano-Co₃O₄/PS/OG system was determined to be 42.0 kJ mol⁻¹. This E_a value was much lower than that observed in PS/Fe₃O₄ process for p-nitroaniline degradation (65.6 kJ mol⁻¹)³⁷ and that in nZVI/PS process for 2,4-dichlorophenol removal (91.5 kJ mol⁻¹).³⁸

3.3 Identification of primary reactive oxidants in nano-Co₃O₄/PS process

It has been reported that two different reactive oxidants (i.e., SO₄˙⁻ and ˙OH) can be generated for the catalyst-mediated decomposition of PS,^39,40 as shown in eqn (4)–(9). Owing to the high rate constants with SO₄˙⁻ (2.5 × 10⁷ M⁻¹ s⁻¹) (ref. 41) and ˙OH (9.7 × 10⁸ M⁻¹ s⁻¹),⁴² methanol is an effective quencher for both SO₄˙⁻ and ˙OH. Due to the high rate constant with ˙OH (6.0 × 10⁸ M⁻¹ s⁻¹) (ref. 41) and the much slower rate constant with SO₄˙⁻ (8.0 × 10⁵ M⁻¹ s⁻¹),⁴² TBA is an effective quencher for ˙OH but not for SO₄˙⁻. Based on these properties, the quenching experiments with methanol and TBA could allow us to differentiate between the contribution of SO₄˙⁻ and ˙OH.

S₂O₈²⁻ + ≡Co₃O₄ → 2SO₄˙⁻ (4)

SO₄˙⁻ + OH⁻ → ˙OH + SO₄²⁻ k₁ = 6.5 × 10⁷ (ref. 33) (5)

SO₄˙⁻ + SO₄˙⁻ → S₂O₈²⁻ k₂ = 8.1 × 10⁸ M⁻¹ s⁻¹ pH = 5.8, 5.0 × 10⁸ M⁻¹ s⁻¹ pH = 5.0, 4.8 × 10⁸ M⁻¹ s⁻¹ pH = 4.8 (ref. 43) (6)

SO₄˙⁻ + S₂O₈²⁻ → S₂O₈˙⁻ + SO₄²⁻ k₃ = 5.5 × 10⁵ M⁻¹ s⁻¹ (ref. 44) (7)

SO₄˙⁻ + ˙OH → HSO₅⁻ k₄ = 1.0 × 10¹⁰ M⁻¹ s⁻¹ (ref. 45) (8)

SO₄˙⁻ + H₂O → HSO₄²⁻ + ˙OH k₅ = 8.3 M⁻¹ s⁻¹ (ref. 44) (9)

In the presence of TBA

HO˙ + (CH₃)₃COH → product k₆ = 6.0 × 10⁸ (ref. 42) (10)

SO₄˙⁻ + (CH₃)₃COH → product k₇ = 8.0 × 10⁵ (ref. 41) (11)

In the presence of methanol

HO˙ + CH₃OH → product k₈ = 9.7 × 10⁸ (ref. 42) (12)

SO₄˙⁻ + CH₃OH → product k₉ = 2.5 × 10⁷ (ref. 41) (13)

Fig. S7† shows the inhibition effect of TBA and methanol on the degradation of OG in nano-Co₃O₄/PS process. With the addition of 10 mM methanol (100 times of the initial OG concentration), the removal of OG decreased remarkably from 90.7% to 30.7%. Meanwhile, the degradation of OG was decreased from 90.7% to 44.9% with the addition of 10 mM TBA. Based on the inhibition effect of methanol and TBA on OG degradation, it could be concluded that both the SO₄˙⁻ and ˙OH were the primary reactive oxidants in nano-Co₃O₄/PS process.^20,40

An attempt with EPR experiments was made to consolidate the presence of SO₄˙⁻ and ˙OH in nano-Co₃O₄/PS process. DMPO was selected as the spin trapping agent in EPR experiments. SO₄˙⁻ and ˙OH could be detected by measuring the signals of DMPO-OH adducts and DMPO-SO₄ adducts, respectively.^4,46 As seen in Fig. 5(a), when pure water was tested with addition of DMPO, no peaks were identified, suggesting that no spins were captured. Characteristic signals of 5,5-dimethylpyrroline-(2)-oxyl-(1) (DMPOX) with an intensity ratio of 1 : 2 : 1 : 2 : 1 : 2 : 1 were identified with the addition of 0.1 M DMPO to the PS solution, indicating the oxidation of DMPO by PS (Scheme S1†). When nano-Co₃O₄ was added together with DMPO and PS, both SO₄˙⁻ and ˙OH were identified with the characteristic peaks of DMPO-HO and DMPO-SO₄ adducts, respectively (Fig. 5(b)). The special hyperfine coupling constants (a(N) 1.49 mT, a(H) 1.49 mT, all ±0.05 mT, obtained by simulation) were consistent with that of DMPO-OH adducts, while the special hyperfine coupling constants (a(N) 1.38 mT, a(H) 1.02 mT, a(H) 0.14 mT, a(H) 0.08 mT, all ±0.05 mT, obtained by simulation) were in accordance with that of DMPO-SO₄ adducts. But the intensity of DMPO-OH signals was much stronger than that of DMPO-SO₄ due to the fast transformation from DMPO-SO₄ to DMPO-OH via nucleophilic substitution (Scheme S1†).⁴⁶

Fig. 5 EPR spectra obtained from ultrapure water, PS oxidative process and nano-Co₃O₄ activated PS oxidative process in the presence of DMPO. Reaction condition: [PS]₀ = 40 mM, [Co₃O₄]₀ = 0 or 5 g L⁻¹, [DMPO]₀ ≈ 0.1 M, pH = 3.0 ± 0.1, T = 25 °C.

3.4 Decolorization, degradation pathways and mineralization of OG in nano-Co₃O₄/PS process

The UV-Vis spectra of OG decolorization evolution by nano-Co₃O₄/PS is presented in Fig. S8.† The absorption spectra of OG were scanned in the range of 300–600 nm. The spectrum of light absorption by OG solution before reaction consists of two main peaks at 328 and 478 nm, plus a shoulder peak at 421 nm, which was consistent with the observations of Xu and Li⁴⁷ and Xiong et al.⁴⁰ The peak at 478 nm is attributed to the absorption of the π–π* transition related to the –N=N– group, while additional bands at 328 nm are assigned to its aromatic ring in the OG molecule. As the reaction proceeded, the two characteristic absorption peaks at 328 and 478 nm decreased dramatically and almost disappeared after 180 min, showing that the chromophore and conjugated system were completely destroyed.

Although efforts have been made to explore OG degradation pathways in sonolysis⁴⁸ and Fenton^49,50 processes, the mechanisms of OG degradation in nano-Co₃O₄/PS process were kept unknown. Thus in this study the byproducts were identified by LC-MS and the mechanistic pathways were depicted in Fig. 6. The retention time and formula for each byproduct was listed in Table S2.† Most of the byproducts had lower molecular weights and earlier retention times than their parent molecules. The proposed structures of degradation products revealed that SO₄˙⁻ and ˙OH initially attacks the aromatic ring, leading to the loss of characteristic fragments of 80 (SO₃H), 77 (phenyl), 16 (OH), and 104 (phenyl–N=N group). Then the further attacks by radicals resulted the formation of various hydroxyl substituted intermediates and quinone end products, as shown in Fig. 6. Previous studies have confirmed that azo chromophore is the essential functional group responsible for the color of azo dyes, thus the loss of 77 (phenyl) and 104 (phenyl–N=N group) directly lead to the decoloration of OG in nano-Co₃O₄/PS process.

Fig. 6 Proposed mechanistic pathways of orange G degradation. All the products postulated are based on the analysis of LC-MS data.

The influence of nano-Co₃O₄ on the mineralization of OG by PS was determined by varying the dosage of PS from 2.0 to 4.0 mM with a fixed nano-Co₃O₄ dosage of 0.5 g L⁻¹ and the results are shown in Fig. S9.† Without nano-Co₃O₄, no mineralization was observed when PS was dosed at 2.0 mM and 1.1% of OG was mineralized when PS was dosed at 4.0 mM. With the introduction of nano-Co₃O₄, OG mineralization rate increased remarkably from 5.6% to 16.4% in 3 h by increasing PS dosage from 2.0 mM to 4.0 mM. Therefore, the application of nano-Co₃O₄ significantly increased not only the removal rate but also the mineralization of OG.

3.5 Stability and reusability of nano-Co₃O₄

Since the stability and reusability of catalyst are critical in catalyzed reactions, the stability of nano-Co₃O₄ was investigated by reusing it in five successive experiments under the same reaction conditions and the results are shown in Fig. 7. The removal efficiency of OG dropped progressively from 100% to ∼64.2% during five successive runs after 3 h of reaction, probably due to the leaching of Co from the catalyst surface determined in the former section, which means that the degradation rate of OG was gradually reduced by repeated reuse of the catalyst, thereby prolonging the time for complete removal of OG. After 5 h of reaction, OG was almost completely removed by the reused catalyst, which indicated the possibility of using the catalyst for a longer operation time. Another cause for the decreasing performance of nano-Co₃O₄ may be the reduced number of Co(II) and Co(II)–OH on the surface of catalyst. The standard redox potentials of Co(H₂O)₆³⁺/Co(H₂O)₆²⁺ and Co(OH)₃/Co(OH)₂ are 1.92 V and 0.17 V (ref. 51) respectively, while that of ·OH/OH⁻ and SO₄˙⁻/SO₄²⁻ is 2.8 V and 2.6–3.1 V;³⁹ hence, the transfer of electrons from Co(II) to·OH and SO₄˙⁻ radicals is thermodynamically favored. The XPS results also confirmed the oxidation of Co(II) and Co(II)–OH to Co(III) and Co(III)–OH. The surface Co(III)/Co(II) molar ratio (3.68) of virgin nano-Co₃O₄ was lower than that (3.82) used after 5 times due to the gradual formation of Co(III) and Co(III)–OH from Co(II) and Co(II)–OH, as shown in Fig. 1(C). In order to consolidate this result, experiments were performed with commercial CoO, Co₃O₄ and Co₂O₃ as heterogeneous catalysts (Fig. S10†). All the commercial powders were of micrometer, so as to exclude the size effect of nano-Co₃O₄. It was interesting to observe that CoO possesses the highest catalytic capability, Co₃O₄ in the medium, and Co₂O₃ the lowest. Thus it can be safely concluded that the main catalytic sites on nano-Co₃O₄ was Co(II) and Co(III)–OH. However, it should be noted that Co₂O₃ also possesses a slight catalytic effect on OG oxidation by PS, possibly due to the conversion between Co(III) to Co(II).⁵² Therefore, the depression in the catalytic performance may be mainly associated with the reduced number of active sites on nano-Co₃O₄ by the oxidation of sulfate- and hydroxyl-radicals and the leaching of cobalt ions.

Fig. 7 Stability of nano-Co₃O₄ in consecutive runs. (Reaction conditions: [OG]₀ = 0.1 mM, [PS]₀ = 2.0 mM, [nano-Co₃O₄] = 0.5 g L⁻¹, pH = 7.0, T = 25 °C, reaction time = 3 h or 5 h).

3.6 Effect of co-existing water matrix chemicals

Textile effluents contain a variety of azo-dyes of various structures, lots of salts and various metal ions. NaCl, NaNO₃, Na₂CO₃ and NaHCO₃ salts are generally added to dye baths for improving the fixation of dyes on fabrics, and to adjust the ionic strength of the dye baths.⁵³ According to one estimate, about 30% metal complexed dyes are used in dyeing wool and 40% for dyeing polyamide.⁵⁴ Therefore, dyes are considered a major source of various metals, including Cd, Cr, Co, Cu, Hg, Ni, Mg, Fe and Mn that are discharged in the raw textile effluents.⁵³ Thus the effects of Cl⁻, NO₃⁻, HCO₃⁻, Co²⁺, Fe²⁺ and Mn²⁺ ions on OG removal by nano-Co₃O₄/PS was studied. The experiments were conducted in the concentration range of 5–20 mM for Cl⁻ and HCO₃⁻ ions, and in the concentration range of 5 μM to 10 mM for heavy metals.

The results indicated that OG degradation was not significantly influenced with 5 mM Cl⁻ ion. However, as the concentration of Cl⁻ ions increased to 10–20 mM, an inhibition was observed (Fig. 8(a)). It is thermodynamically feasible for SO₄˙⁻ (2.50 V) to oxidize chloride ions (Cl⁻) into less reactive chlorine species viz. Cl₂/2Cl⁻ (1.40 V)⁵¹ and HOCl/Cl⁻ (1.48 V).⁵⁵ The chemical reactions involved in (but not limited to) nano-Co₃O₄/PS system in the presence of Cl⁻ ions can be given as follows:⁵⁶

SO₄˙⁻ + Cl⁻ → SO₄²⁻ + Cl˙ (14)

Cl˙ + Cl⁻ → Cl₂˙⁻ (15)

Cl₂˙⁻ + Cl₂˙⁻ → Cl₂ + 2Cl⁻ (16)

Cl₂ + H₂O → HOCl + H⁺ + Cl⁻ (17)

HOCl → H⁺ + OCl⁻ (18)

Fig. 8 Effects of co-existing water matrix chemicals on OG degradation in nano-Co₃O₄/PS process. (Reaction conditions: [OG]₀ = 0.1 mM, [PS]₀ = 2.0 mM, [nano-Co₃O₄] = 0.5 g L⁻¹, pH = 7.0, T = 25 °C, reaction time = 90 min).

Therefore, the observed inhibition in the presence of 10–20 mM Cl⁻ is due to the consumption of sulfate radicals by Cl⁻ ions (eqn (14)–(18)),⁵⁷ and the formation of less reactive chlorine species Cl₂, HOCl, Cl˙ and Cl₂˙⁻.^55,57

The bicarbonate ion (HCO₃⁻) is an efficient radical scavenger (eqn (19) and (20)).⁵⁸ With increasing HCO₃⁻ from 0 to 20 mM, the OG removal dropped from 71% to 33% (Fig. 8(b)). The NO₃⁻ ions have no obvious effect on the performance of nano-Co₃O₄/PS process. With its concentration increasing from 5 mM to 20 mM, the removal of OG remained constant, as shown in Fig. 8(c).

SO₄˙⁻ + HCO₃⁻ → SO₄²⁻ + CO₃˙⁻ + H⁺ (k₁₀ = 1.6 × 10⁶ M⁻¹ s⁻¹) (19)

·OH + HCO₃⁻ → CO₃˙⁻ + H₂O (k₁₁ = 8.5 × 10⁶ M⁻¹ s⁻¹) (20)

Ferrous ions (Fe²⁺) can rapidly activate persulfate to form sulfate radicals (SO₄˙⁻) at a high rate constant (k) of 2.7 × 10 M⁻¹ s⁻¹. However, a stronger interaction (k = 4.6 × 10⁹ M⁻¹ s⁻¹) between Fe²⁺ and sulfate radicals was reported.⁵⁹ Thus, Fe²⁺ might be converted simultaneously by both PS and sulfate radicals, and the final reaction product (SO₄²⁻) remained in the system, as shown in Fig. 8(d). The reaction equations were as follows (eqn (21) and (22)).

Fe²⁺ + S₂O₈²⁻ → Fe³⁺ + SO₄²⁻ + SO₄˙⁻ k₁₂ = 2.7 × 10 M⁻¹ s⁻¹ (21)

Fe²⁺ + SO₄˙⁻ → Fe³⁺ + SO₄²⁻ k₁₃ = 4.6 × 10⁹ M⁻¹ s⁻¹ (22)

Therefore, 1.0 mM was found to be the optimized iron concentration with 87% of OG removal in 90 min (Fig. 8(d)). When the Fe²⁺ concentration was increased to 10 mM, only 60.5% of OG was removed due to the consumption of generated radicals by excess Fe²⁺ ions.

Experimental results showed that Co²⁺/PS system was more efficient for OG dissipation in relative to Fe²⁺/PS system (Fig. 8(e)). The lowest Co²⁺ concentration used for PS activation was orders of magnitude lower than that of Fe²⁺, confirming Co²⁺ was a better activator for PS. The degradation of OG increased appreciably with increasing Co²⁺ concentration. For example, removal rate was found to be 91.5%, 100% and 100% for Co²⁺ of 5 μM, 10 μM and 1 mM, respectively. It has been proposed that Co²⁺ acts as a catalyst during PS activation (eqn (23)), and Co²⁺ regeneration occurs via Co³⁺ reduction.⁵² However, it has been reported that excess Co²⁺ might also scavenge SO₄˙⁻ (eqn (24)), which was also confirmed by this study. With increasing the concentration of Co²⁺ to 10 mM, the decomposition of OG by nano-Co₃O₄/PS was suppressed.

Co²⁺ + S₂O₈²⁻ → Co³⁺ + SO₄²⁻ + SO₄˙⁻ (23)

Co²⁺ + SO₄˙⁻ → Co³⁺ + SO₄²⁻ (24)

With increasing the concentration of Mn²⁺ from 5 μM to 10 mM, the degradation of OG was sharply suppressed (Fig. 8(f)). During the reaction, Mn²⁺ was also oxidized by PS or radicals and precipitated as MnO₂.⁶⁰ Thus the Mn²⁺ played a role of scavenger for reactive radical species and competitor to the target organics in nano-Co₃O₄/PS system.

In summary, the anions and Mn²⁺ have no or some inhibition on the performance of nano-Co₃O₄/PS system due to their scavenging effect on the reactive radical species (SO₄˙⁻ or ˙OH). While both the ferrous and cobalt ions with lower concentration enhanced the degradation of OG by promoting the production of active radicals. But the higher concentration of ferrous and cobalt ions played a role of scavenger for reactive radical species and competitor to the target organics in nano-Co₃O₄/PS system.

3.7 Comparison with the nano-Co₃O₄/PMS and nano-Co₃O₄/NaPS systems

Nano-Co₃O₄ species are known to activate peroxymonosulfate and persulfate to generate SO₄˙⁻ and ˙OH for oxidation of organic compounds in water. Thus, the performances of potassium persulfate (PS), which was used throughout this study in above sections, sodium persulfate (NaPS) and Oxone® (PMS) were compared in nano-Co₃O₄ related oxidation systems, i.e., nano-Co₃O₄/PMS, nano-Co₃O₄/PS and nano-Co₃O₄/NaPS. Fig. 9 shows that nano-Co₃O₄/PMS displayed the best performance for the decomposition of OG (93% removal) after 60 min, at which the removal of OG in nano-Co₃O₄/PS and nano-Co₃O₄/NaPS processes were 56.8% and 46.1% respectively. This indicated that higher levels of radicals were generated in nano-Co₃O₄/PMS than that in nano-Co₃O₄/PS and nano-Co₃O₄/NaPS.⁶¹ However, this finding was totally contrary to the results reported by Luo et al.,⁶¹ who found PS produced higher amount of sulfate- and hydroxyl-radicals than PMS in the presence of UV irradiation at 254 nm. It should be noted that the pH for PMS/UV and PS/UV were 3.65 and 5.88 respectively in Luo's study and very different from that of 7.0 in this study. In addition, the production of radicals may also have close relationship to the activation methods. The investigation and discussion about the reactivity discrepancy of PS, NaPS and PMS was not deep enough in this study and further studies are needed.

Fig. 9 OG degradation in nano-Co₃O₄/NaPS, nano-Co₃O₄/PS and nano-Co₃O₄/PMS processes. (Reaction conditions: [OG] = 0.1 mM, [NaPS] = [PS] = [PMS] = 2.0 mM, [nano-Co₃O₄] = 0.5 g L⁻¹, pH = 7.0 ± 0.1, T = 25 °C).

4 Conclusions

Nano-Co₃O₄ exhibited good heterogeneous activity in nano-Co₃O₄/PS system and low dissolved Co ions especially at neutral or alkaline conditions. The smaller particle size of Co₃O₄, higher dosage of PS and catalyst, and higher temperature promoted the degradation of OG in nano-Co₃O₄/PS system. Both SO₄˙⁻ and ·OH were proved to be the primary oxidative species by EPR experiment and the quenching results with TBA and methanol. The byproducts of OG in nano-Co₃O₄/PS process were identified by LC-MS and the mechanistic pathways were proposed. The stability of nano-Co₃O₄ was acceptable, although a slight decline in the catalytic ability was observed due to the oxidation of surface Co(II) to Co(III) and the leaching of cobalt ions. The NO₃⁻, Cl⁻, HCO₃⁻, and Mn²⁺ have no or some inhibited effect on the performance of nano-Co₃O₄/PS system due to their scavenging effect on the reactive radical species and competition with target organics for the oxidant. Fe²⁺ and Co²⁺ with lower concentration enhanced the degradation of OG by promoting the production of active radicals. The reactivity discrepancy of PS, NaPS and PMS followed the order of nano-Co₃O₄/PMS > nano-Co₃O₄/PS > nano-Co₃O₄/NaPS.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (51508152), Natural Science Foundation of Jiangsu Province (BK20150812), the China Postdoctoral Science Foundation (2015M571660), the Fundamental Research Funds for the Central Universities (2014B12614) and the Priority Academic Program Development of Jiangsu Higher Education Institutions.

References

1. J. Y. Fang and C. Shang, Environ. Sci. Technol., 2012, 46, 8976.
2. R. L. Johnson, P. G. Tratnyek and R. O. Johnson, Environ. Sci. Technol., 2008, 42, 9350.
3. W. S. Chen and Y. C. Su, Ultrason. Sonochem., 2012, 19, 921.
4. O. S. Furman, A. L. Teel and R. J. Watts, Environ. Sci. Technol., 2010, 44, 6423.
5. S. Y. Yang, X. Yang, X. T. Shao, R. Niu and L. L. Wang, J. Hazard. Mater., 2011, 186, 659.
6. G. D. Fang, J. Gao, D. D. Dionysiou, C. Liu and D. M. Zhou, Environ. Sci. Technol., 2013, 47, 4605.
7. M. Ahmad, A. L. Teel and R. J. Watts, Environ. Sci. Technol., 2013, 47, 5864.
8. G. P. Anipsitakis and D. D. Dionysiou, Environ. Sci. Technol., 2004, 38, 3705.
9. T. Zhang, H. B. Zhu and J. P. Croue, Environ. Sci. Technol., 2013, 47, 2784.
10. Y. B. Ding, L. H. Zhu, N. Wang and H. Q. Tang, Appl. Catal., B, 2013, 129, 153.
11. G. P. Anipsitakis and D. D. Dionysiou, Environ. Sci. Technol., 2003, 37, 4790.
12. J. Fernandez, P. Maruthamuthu, A. Renken and J. Kiwi, Appl. Catal., B, 2004, 49, 207.
13. G. P. Anipsitakis, E. Stathatos and D. D. Dionysiou, J. Phys. Chem. B, 2005, 109, 13052.
14. Y. H. Guan, J. Ma, Y. M. Ren, Y. L. Liu, J. Y. Xiao, L. Q. Lin and C. Zhang, Water Res., 2013, 47, 5431.
15. C. Cai, H. Zhang, X. Zhong and L. W. Hou, J. Hazard. Mater., 2015, 283, 70.
16. S. N. Su, W. L. Guo, Y. Q. Leng, C. L. Yi and Z. M. Ma, J. Hazard. Mater., 2013, 244, 736.
17. Y. J. Yao, Y. M. Cai, G. D. Wu, F. Y. Wei, X. Y. Li, H. Chen and S. B. Wang, J. Hazard. Mater., 2015, 296, 128.
18. Y. Jiang, Y. Wu, B. Xie, Y. Xie and Y. T. Qian, Mater. Chem. Phys., 2002, 74, 234.
19. L. P. Zhou, J. Xu, H. Miao, F. Wang and X. Q. Li, Appl. Catal., A, 2005, 292, 223.
20. X. Y. Chen, J. W. Chen, X. L. Qiao, D. G. Wang and X. Y. Cai, Appl. Catal., B, 2008, 80, 116.
21. E. Saputra, S. Muhammad, H. Q. Sun, H. M. Ang, M. O. Tade and S. B. Wang, J. Colloid Interface Sci., 2013, 407, 467.
22. C. Q. Tan, N. Y. Gao, Y. Deng, J. Deng, S. Q. Zhou, J. Li and X. Y. Xin, J. Hazard. Mater., 2014, 276, 452.
23. A. El-Ghenymy, F. Centellas, J. A. Garrido, R. M. Rodriguez, I. Sires, P. L. Cabot and E. Brillas, Electrochim. Acta, 2014, 130, 568.
24. A. B. dos Santos, F. J. Cervantes and J. B. van Lier, Bioresour. Technol., 2007, 98, 2369.
25. P. Shukla, H. Q. Sun, S. B. Wang, H. M. Ang and M. O. Tade, Sep. Purif. Technol., 2011, 77, 230.
26. P. H. Shi, R. J. Su, S. B. Zhu, M. C. Zhu, D. X. Li and S. H. Xu, J. Hazard. Mater., 2012, 229, 331.
27. D. Barreca, C. Massignan, S. Daolio, M. Fabrizio, C. Piccirillo, L. Armelao and E. Tondello, Chem. Mater., 2001, 13, 588.
28. Z. L. Chen, S. H. Chen, Y. H. Li, X. L. Si, J. Huang, S. Massey and G. L. Chen, Mater. Res. Bull., 2014, 57, 170.
29. G. M. Bai, H. X. Dai, J. G. Deng, Y. X. Liu, F. Wang, Z. X. Zhao, W. G. Qiu and C. T. Au, Appl. Catal., A, 2013, 450, 42.
30. C. H. Sun and J. C. Berg, Adv. Colloid Interface Sci., 2003, 105, 151.
31. P. H. Shi, X. F. Dai, H. G. Zheng, D. X. Li, W. F. Yao and C. Y. Hu, Chem. Eng. J., 2014, 240, 264.
32. T. L. Johnson, M. M. Scherer and P. G. Tratnyek, Environ. Sci. Technol., 1996, 30, 2634.
33. Y. H. Guan, J. Ma, X. C. Li, J. Y. Fang and L. W. Chen, Environ. Sci. Technol., 2011, 45, 9308.
34. F. Qi, W. Chu and B. B. Xu, Appl. Catal., B, 2013, 134, 324.
35. Q. Chang, K. J. Deng, L. H. Zhu, G. D. Jiang, C. Yu and H. Q. Tang, Microchim. Acta, 2009, 165, 299.
36. M. J. Pu, Y. W. Ma, J. Q. Wan, Y. Wang, M. Z. Huang and Y. M. Chen, J. Colloid Interface Sci., 2014, 418, 330.
37. Y. S. Zhao, C. Sun, J. Q. Sun and R. Zhou, Sep. Purif. Technol., 2015, 142, 182.
38. R. C. Li, X. Y. Jin, M. Megharaj, R. Naidu and Z. L. Chen, Chem. Eng. J., 2015, 264, 587.
39. P. Avetta, A. Pensato, M. Minella, M. Malandrino, V. Maurino, C. Minero, K. Hanna and D. Vione, Environ. Sci. Technol., 2015, 49, 1043.
40. X. M. Xiong, B. Sun, J. Zhang, N. Y. Gao, J. M. Shen, J. L. Li and X. H. Guan, Water Res., 2014, 62, 53.
41. P. Neta, R. E. Huie and A. B. Ross, J. Phys. Chem. Ref. Data, 1988, 17, 1027.
42. G. V. Buxton, C. L. Greenstock, W. P. Helman and A. B. Ross, J. Phys. Chem. Ref. Data, 1988, 17, 513.
43. P. Neta, R. E. Huie and A. B. Ross, J. Phys. Chem. Ref. Data, 1988, 17, 1027.
44. X. Y. Yu, Z. C. Bao and J. R. Barker, J. Phys. Chem. A, 2004, 108, 295.
45. U. K. Klaning, K. Sehested and E. H. Appelman, Inorg. Chem., 1991, 30, 3582.
46. G. S. Timmins, K. J. Liu, E. J. H. Bechara, Y. Kotake and H. M. Swartz, Free Radical Biol. Med., 1999, 27, 329.
47. X. R. Xu and X. Z. Li, Sep. Purif. Technol., 2010, 72, 105.
48. M. Q. Cai, M. C. Jin and L. K. Weavers, Ultrason. Sonochem., 2011, 18, 1068.
49. A. D. Bokare, R. C. Chikate, C. V. Rode and K. M. Paknikar, Environ. Sci. Technol., 2007, 41, 7437.
50. M. Q. Cai, J. Su, Y. Z. Zhu, X. Q. Wei, M. C. Jin, H. J. Zhang, C. Y. Dong and Z. S. Wei, Ultrason. Sonochem., 2016, 28, 302.
51. J. G. Speight, Lange's Handbook of Chemistry, McGraw-Hill Education, New York, 2005.
52. Y. F. Ji, C. X. Dong, D. A. Kong and J. H. Lu, J. Hazard. Mater., 2015, 285, 491.
53. M. Imran, D. E. Crowley, A. Khalid, S. Hussain, M. W. Mumtaz and M. Arshad, Rev. Environ. Sci. Bio/Technol., 2015, 14, 73.
54. K. Hunger, Industrial dyes-chemistry, properties, application, VCH, New York, 2003.
55. Z. H. Wang, R. X. Yuan, Y. G. Guo, L. Xu and J. S. Liu, J. Hazard. Mater., 2011, 190, 1083.
56. R. X. Yuan, S. N. Ramjaun, Z. H. Wang and J. S. Liu, J. Hazard. Mater., 2011, 196, 173.
57. C. J. Liang, Z. S. Wang and N. Mohanty, Sci. Total Environ., 2006, 370, 271.
58. A. Ghauch and A. Tuqan, Chem. Eng. J., 2012, 183, 162.
59. C. S. Liu, K. Shih, C. X. Sun and F. Wang, Sci. Total Environ., 2012, 416, 507.
60. H. Akiho, S. Ito, H. Matsuda and T. Yoshioka, Environ. Sci. Technol., 2013, 47, 11311.
61. C. W. Luo, J. Ma, J. Jiang, Y. Z. Liu, Y. Song, Y. Yang, Y. H. Guan and D. J. Wu, Water Res., 2015, 80, 99.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra22457h

---