Nonprecious mixed oxide catalysts Co3AlO and Co2NiAlO derived from nanoflowerlike cobalt-based hydrotalcites for highly efficient oxidation of nitric oxide

Abstract

Nonprecious mixed oxide catalysts Co3AlO-T and Co2NiAlO-T (T = 500, 800) were obtained from thermal decomposition of nanoflowerlike Co3Al-HT and Co2NiAl-HT hydrotalcite-like precursors at 500 and 800 °C, which were synthesized by a facile co-precipitation method, and systematically characterized by combinational techniques. The mixed oxide catalysts mainly consisted of homogeneous and stable non-stoichiometric cobalt-based spinel phases Co(Co,Al)₂O₄ and Ni(Co,Al)₂O₄. The 500 °C calcined catalysts showed nanoflowerlike morphology, while the 800 °C calcined catalysts presented disorderly stacked states consisting of severely agglomerated irregular nanoparticles. The catalysts Co3AlO-500 and Co2NiAlO-500 presented excellent catalytic NO oxidation performance, with the maximum conversion efficiency ca. 88.8% and 87.6% (285 °C), respectively, which is much higher than those of Co3AlO-800 and Co2NiAlO-800 (55.6% and 66.6% at 350 °C). The excellent catalytic NO oxidation performance of Co3AlO-500 and Co2NiAlO-500 could be attributed to the much larger amount of surface active sites endowed by these catalysts with smaller crystallite sizes of cobalt-based spinel phases, higher specific surface areas, and mesoporous structure. The active sites and NO oxidation pathways of the present catalysts were tentatively proposed. Although Co2NiAlO-500 possessed a slightly smaller amount of active sites (Co³⁺ and Ni³⁺ associated with surface adsorbed oxygen), compared to Co3AlO-500 with active sites (Co³⁺ associated with surface adsorbed oxygen), catalyst Co2NiAlO-500 possessed higher reducibility, leading to their similar desorption amount of NO₂, thus similar NO oxidation performance. Moreover, the lower cost of catalyst Co2NiAlO-500 gives it a greater potential for practical applications.

---

1. Introduction

Diesel and lean-burn gasoline engines with high air/fuel ratios have drawn much attention because of their higher fuel efficiency and lower CO₂ emission than conventional gasoline engines.^1–4 However, the nitrogen oxide (NO_x) pollutants emitted from the lean-burn engines cannot be effectively removed by the traditional three-way catalysts due to the high oxygen content in the lean-burn exhaust, which can cause severe environmental problems such as photochemical smog, acid rain, and ozone layer depletion.^1,5,6 In order to abate the NO_x pollutants and meet the increasingly stringent NO_x emission regulations, various techniques have been developed, among which the NO_x storage and reduction (NSR),^7–9 selective catalytic reduction (SCR),^10–12 and continuously regenerating trap (CRT) for simultaneous NO_x and soot removal^2,13,14 are the main solutions. It is known that the NO_x pollutants emitted from the lean-burn engines mainly consist of NO (∼90%), while NO₂ is considered as an important intermediate species for NO_x conversion in all these techniques.^1,2 The oxidation of NO to NO₂ is the first step in the NSR technology and NO₂ is much easier to be absorbed by the storage materials than NO.^7–9 The reaction rate of NH₃-SCR can be greatly improved at low temperatures (200–300 °C) with equal amount of NO and NO₂ in the exhaust during the so-called fast-SCR process, which is achieved by setting an oxidation catalyst at the upstream of the SCR unit.^10–12 Furthermore, in the CRT technology, NO₂ shows stronger oxidizability than O₂, and can oxidize soot particles collected on a particulate filter for the diesel vehicles at lower temperature.^2,13,14 Therefore, it is of great importance to develop the effective catalysts for oxidizing NO to NO₂.

Among the previously reported catalysts for NO oxidation, platinum based noble metal catalysts have been widely used due to their good performance but are limited by their high cost and poor thermal stability under lean-burn conditions.^15,16 Transition metal oxide catalysts such as Co-related oxides are favorable alternatives to noble metal catalysts because of their low cost, high catalytic activity, and potential practicality.^2,10,17–20 Ren et al.¹⁹ reported two Co₃O₄ nanowires-array catalysts, grown on ceramic cordierite honeycomb substrates using nitrate and acetate as cobalt sources followed 300 °C calcinations, showing the highest catalytic NO oxidation property with similar conversion efficiency as 80% at 275 °C. Wang et al.²⁰ reported ordered mesoporous catalysts Co₃O₄, NiO and NiCo₂O₄, prepared by nanocasting method using mesoporous KIT-6 as a hard template. The catalyst NiCo₂O₄ showed the best low-temperature H₂-SCR activity and the highest NO-to-NO₂ conversion (∼60%, 350 °C), which was attributed to the higher surface adsorption oxygen concentration in NiCo₂O₄ than in Co₃O₄ and NiO, and the mesoporous structure of NiCo₂O₄ rendered it larger surface area and extra surface oxygen vacancies. However, the synthesis of the hierarchical Co₃O₄ nanowires-array catalysts needs the usage of ceramic cordierite honeycomb substrates and the hydrothermal condition. While the preparation of the ordered mesoporous catalyst NiCo₂O₄ uses the mesoporous KIT-6 template in the nanocasting method. Obviously, the preparation procedures of these Co-based oxide catalysts are complicated and time consuming, which may restrict the large scale synthesis of these catalysts in the practical applications.

Hydrotalcite-like compounds, also known as layered double hydroxides (LDHs) with the general formula [M_1−x²⁺M_x³⁺(OH)₂](A_x/nⁿ⁻)·mH₂O, are a class of 2D layered clay materials consisting of positively charged hydroxide layers with M²⁺ and M³⁺ metal ions arranged in a uniform and ordered manner and interlayer hydrated anions for charge balance.^21–23 Hydrotalcite-like compounds can contain more than two types of highly dispersed metal cations and their morphology and particle sizes are tunable through different synthesis methods, and these unique structural characters can provide a good platform for the smart design of LDH-based NO_x abatement catalysts.^6,7,24–27 Yu et al.^24,25 reported a series of Co-based oxide catalysts xCoMgAlO and xCaCoAlO obtained by thermal decomposition hydrotalcite precursors at 800 °C in air for the effective abatement of NO_x in the NSR technology, while one demerit of these catalysts was the relatively low surface area (∼20 m² g⁻¹) due to the high calcination temperature. Li et al.¹⁴ reported a K-containing (4.5 wt%) CoMgAlO catalyst calcined at 600 °C, showing the best catalytic performance in soot combustion and simultaneous soot-NO_x removal with the highest NO_x reduction (by soot) percentage of 32% among a series of K-promoted CoMgAlO catalysts due to the high surface K/Co atomic ratio and the strong interaction between K and Co. Wang et al.²⁶ reported a series of hydrotalcites derived oxide catalysts MgAlO, CoMgAlO, CuMgAlO, and CuCoMgAlO with large surface areas and highly dispersed redox components, exhibiting high performance in NO_x storage and soot combustion, however, the NO oxidation activities of these catalysts were relatively low. Recently, our group²⁷ reported a 3D oxide nanosheet array catalyst CoMgAlO-array with small-sized active Co₃O₄ species (5.7 nm) highly dispersed on a Mg/Al-oxide matrix and relatively larger surface area, showing much higher NO_x storage capacity and catalytic soot combustion activity compared to traditional CoMgAlO catalyst. However, although the above hydrotalcite-like precursors derived oxide catalysts are reported for effective abatement of NO_x in the NSR and simultaneous soot-NO_x removal techniques, the NO oxidation activities of these catalysts are relatively low probably due to the containing of alkali metals (K) or alkaline-earth metals (Mg, Ca). Up to now, there are few reports on the specially designed hydrotalcite-like precursors derived cobalt-based mixed oxide catalysts for NO oxidation. Therefore, it is highly desired and urgent for the special design and fabrication of cobalt-based mixed oxide catalysts with high catalytic NO oxidation activity and low cost using hydrotalcite-like precursors with adjustable metal cations.

Herein, nonprecious mixed oxide catalysts Co3AlO-T and Co2NiAlO-T (T = 500, 800) have been synthesized via a facile and economical thermal decomposition of hydrotalcite precursor method and firstly used for catalytic oxidation of NO. The crystal structure, textural properties and redox property of the samples were systematically characterized by XRD, EXAFS, SEM/HRTEM, BET, XPS, H₂-TPR and in situ FT-IR etc. techniques. Furthermore, the active sites and oxidation pathways of the NO oxidation reaction for these catalysts were also carefully studied and the relationship of the structure and activity for the catalysts was also well explained.

2. Experimental

2.1 Materials

All reagents were of analytical pure grade. Cobalt nitrate hexahydrate (Co(NO₃)₂·6H₂O), nickel nitrate hexahydrate (Ni(NO₃)₂·6H₂O) and aluminum nitrate nonahydrate (Al(NO₃)₃·9H₂O) were purchased from Xi Long Chemical Corporation. Sodium hydroxide and sodium carbonate were purchased from Beijing Chemical Works.

2.2 Preparation of the catalysts

Co3Al and Co2NiAl hydrotalcite-like compounds with M²⁺/Al³⁺ molar ratio fixed at 3.0 were prepared with a constant-pH coprecipitation method. Typically, a mixed salt solution (100 mL) and a mixed basic solution (200 mL) were simultaneously dropwise added into 100 mL distilled water for ∼1 h under vigorous mechanical stirring at room temperature, keeping a constant pH as 10 ± 0.2. The mixed salt solution with a total metal ion concentration of 0.2 M contains suitable amounts of Co(NO₃)₂·6H₂O and Al(NO₃)₃·9H₂O. The mixed basic solution contains NaOH and Na₂CO₃ with [OH⁻] = 1.6([M²⁺] + [M³⁺]) and [CO₃²⁻] = 2.0[M³⁺]. The resulting suspension was kept at 65 °C for 4 h under stirring in static air and the resulting precipitate was filtered, washed thoroughly with distilled water and dried at 60 °C for 12 h. The obtained hydrotalcite-like precursor was denoted as Co3Al-HT. Precursor Co2NiAl-HT was prepared using a mixed salt solution containing Co(NO₃)₂·6H₂O, Ni(NO₃)₂·6H₂O and Al(NO₃)₃·9H₂O with molar ratio of 2/1/1 and other conditions unchanged. The precursor samples were ground to fine powders and then calcined at 500 °C and 800 °C for 4 h in static air, respectively, to get the mixed oxide catalysts Co3AlO-T and Co2NiAlO-T (T = 500, 800). The oxide catalysts were pressed to a disk, crushed and sieved to particles with 20–40 mesh size and kept in a desiccator.

2.3 Catalytic activity tests

The tests of catalytic NO oxidation were conducted in a fixed-bed flow microreactor under atmospheric pressure. Typically, 0.3 g catalyst (sieve fraction of 20–40 mesh) was placed in a quartz reactor (i.d. = 6 mm, L = 540 mm), the reactant gas mixture (500 ppm NO, 8% O₂ and N₂ balance) was fed to the reactor with a total flow rate of 500 mL min⁻¹, corresponding to a gas hourly space velocity (GHSV) of 100 000 mL g⁻¹ h⁻¹. Before the test, the catalyst was pre-treated in a gas flow of O₂/N₂ (8% O₂) at 450 °C at a constant flow rate of 500 mL min⁻¹ for 0.5 h and cooled down to the experiment temperature (150 °C), then the catalyst was pre-adsorbed the reaction gas mixture at 150 °C for 2 h to avoid errors caused by NO adsorption. The steady-state tests were conducted from 150 °C to 350 °C and the gas products (after 0.5 h reaction) were analyzed by a Chemiluminescence NO–NO₂–NO_x analyzer (model EC 9841, Ecotech Corporation). The NO conversion is defined as:

Experiments for testing the effect of the NO concentration on the NO conversion were made at NO concentration ranging from 300–700 ppm (300, 400, 500, 600, 700 ppm) in a feed containing 8% O₂. Experiments for testing the effect of the O₂ concentration on the NO conversion were made at O₂ concentration ranging from 2–9% (2%, 4%, 6%, 8%, 9%) in a feed containing 500 ppm NO. All of these experiments were conducted at 300 °C with N₂ as balance gas and GHSV of 100 000 mL g⁻¹ h⁻¹. Experiments for measuring the reaction orders of NO and O₂ were made at the same conditions as those of the experiments for testing the effects of NO and O₂ concentration on the NO conversion, except the reaction temperature was set at 150 °C.

2.4 Characterizations

Powder X-ray diffraction (PXRD) patterns were taken on a Shimadzu XRD-6000 diffractometer using Cu Kα radiation (λ = 1.5418 Å, 40 kV, 30 mA) in 2θ 3–70° with a scan speed of 10° min⁻¹. Fourier transform infrared spectra (FT-IR) were recorded on a Bruker Vector-22 FT-IR spectrophotometer using the KBr pellet method (sample/KBr = 1/100) in the range of 4000–400 cm⁻¹ with a resolution of 4 cm⁻¹. Scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDX) measurements were carried out on an Oxford Instrument INCAx-act EDX detector attached to a Zeiss Supra 55 field emission scanning electron microscope using a 20 kV electron beam and 60 s acquisition. The high resolution transmission electron microscopy (HRTEM) images were recorded on a JEM-2010 high resolution transmission electron microscope with an accelerating voltage of 200 kV. Raman spectra were recorded on a Jobin Yvon Horiba Raman spectrometer model HR800 using a 532 nm line of Ar⁺ ion laser as the excitation source at room temperature. The content of metal components Co, Ni and Al in the hydrotalcite precursors was measured by inductively coupled plasma (ICP) emission spectroscopy on a Shimadzu ICPS-7500 instrument. The sample (0.025 g) was solubilized using concentrated nitric acid (1 mL) and the solution was diluted to 25 ppm. Thermogravimetric (TG) analysis was performed on a Mettler Toledo TGA/DSC1 STAR^e system under static air atmosphere at a heating rate 10 °C min⁻¹ from room temperature to 900 °C.

Low temperature N₂ adsorption–desorption measurement was conducted at 77 K using a Quantachrome Autosorb-1C-VP system. Before the measurement, the samples were outgassed under vacuum at 300 °C for 6 h. The specific surface area was determined by the Brunauer–Emmett–Teller (BET) method upon the adsorption branch and the pore size distribution by the Barrett–Joyner–Halenda (BJH) method upon the desorption branch of the isotherm.

X-ray photoelectron spectra (XPS) were obtained on a Thermo VG ESCALAB 250 X-ray photoelectron spectrometer at a base pressure of 2 × 10⁻⁹ Pa in the analysis chamber using Al Kα X-ray (1486.6 eV) as excitation source and the binding energies were calibrated by the C 1s line at 284.9 eV.

Temperature-programmed reduction (TPR) measurements of the samples were carried out on a Thermo Scientific TPDRO 1100 instrument equipped with a thermal conductivity detector (TCD). In each TPR run, 20 mg sample was placed in a quartz tube and sandwiched between two quartz wool plugs. Prior to each run, the sample was heated to 300 °C with Ar flushing (20 mL min⁻¹) for 30 min, and then cooled down to room temperature. Afterward, the sample was heated to 1000 °C at a ramp of 10 °C min⁻¹ and reduced in a reducing environment (10 vol% H₂ in Ar at a flow rate of 30 mL min⁻¹). The H₂ consumption was monitored by the TCD during the heating.

Temperature-programmed desorption (TPD) tests were conducted on the fixed-bed flow microreactor. Prior to the TPD test of NO adsorption alone at 50 °C, 0.3 g sample was placed in the quartz tube and pre-treated at 450 °C for 0.5 h under 8% O₂/N₂ gas flow (500 mL min⁻¹) and then cooled down to 50 °C. After the sample was exposed in 500 ppm NO/N₂ (500 mL min⁻¹) at 50 °C for 2 h and flushed by N₂ (500 mL min⁻¹) for 1 h, the desorption profiles of NO and NO₂ were obtained by heating the sample to 400 °C at a heating rate of 5 °C min⁻¹ in a N₂ flow (500 mL min⁻¹). Prior to the TPD test of NO and O₂ co-adsorption at 300 °C, 0.3 g sample was placed in the quartz tube and pre-treated at 450 °C for 0.5 h under 8% O₂/N₂ gas flow (500 mL min⁻¹) and then cooled down to 300 °C. Subsequently, the sample was exposed in 500 ppm NO/8% O₂/N₂ (500 mL min⁻¹) at 300 °C for 2 h and cooled down to 50 °C, then flushed by N₂ (500 mL min⁻¹) for 1 h, and the desorption profiles of NO and NO₂ were obtained by heating the sample to 400 °C at a heating rate of 5 °C min⁻¹ in a N₂ flow (500 mL min⁻¹). The NO and NO₂ desorption signals were continuously recorded as a function of increasing temperature using a on-line Chemiluminescence NO–NO₂–NO_x analyzer (model EC 9841, Ecotech Corporation).

Extended X-ray absorption fine structure (EXAFS) measurements were carried out at 1W2B beamline of the Beijing Synchrotron Radiation Facility (BSRF) operating at about 2.5 GeV and 200 mA and using a Si (111) double-crystal monochromator. The absorption spectra of Co K-edge and Ni K-edge of the samples and the reference compounds Co₃O₄ and NiO were collected in the transmission mode at room temperature. Data analysis was performed using IFEFFIT program packages, the back-subtracted EXAFS function was converted into k space and weighted by k³, then the Fourier transforming of the k³-weighted EXAFS data was performed in the range of k = 3–14 Å⁻¹ with a Hanning function window and the radial structure function was obtained.

In situ FT-IR spectra were obtained on a Nicolet 380 FT-IR spectrometer in the range of 650–4000 cm⁻¹ after 64 scans at a resolution of 4 cm⁻¹. Prior to the recording of IR spectra, the sample (10 mg, pellets with diameter of 13 mm) was pre-treated in pure N₂ at 300 °C for 1 h with a flow rate of 30 mL min⁻¹ to eliminate the impure species on the surface of the sample. After the background spectrum was recorded with the flowing of N₂ and was subtracted, the sample was exposed to 500 ppm NO/N₂ for 60 min and followed by exposed to 500 ppm NO and 8% O₂ in N₂ for 60 min and the time-dependent FT-IR spectra were recorded.

3. Results and discussion

3.1 Crystal structure and textural properties

The XRD patterns of the precursors Co3Al-HT, Co2NiAl-HT and the derived mixed oxide catalysts Co3AlO-500, Co2NiAlO-500, Co3AlO-800 and Co2NiAlO-800 are shown in Fig. 1. The crystal structure parameters of precursors and catalysts are summarized in Tables S1† and 1, respectively. All the precursors show the characteristic reflections of a pure hydrotalcite phase (JCPDS 22-0700),^24,28,29 with sharp and symmetric reflections at 2θ ∼ 11.8°, 23.5°, 35° for the basal (003), (006) and (009) planes, and broad and asymmetric reflections at ∼35°, 39°, 47° for the non-basal (012), (015) and (018) planes. The broad signal between 32° and 38° may be resulted from the overlap of (009) and (012) reflections. Moreover, the reflections of (110) and (113) can be clearly seen at 2θ around 60–62°, implying a good dispersion of metal cations in the hydrotalcite layers.²⁴ These phenomena suggest that the hydrotalcite phase possesses well-formed layered structure and is highly crystallized. The reflections appeared at d₀₀₃ = 0.76 nm indicate that the interlayer anions are CO₃²⁻, which is confirmed by the sharp and strong FT-IR absorption peak at ca. 1371 cm⁻¹ due to the ν₃ (asymmetric stretching) vibrational mode of CO₃²⁻ ions associated with a shoulder at 857 cm⁻¹ (Fig. 3A).^30,31 Compared to Co3Al-HT, the absence of other crystalline phase in Co2NiAl-HT suggests that Ni²⁺ ions isomorphically substitute partial Co²⁺ ions in the brucite-like layers due to the similar ionic radii of Co²⁺ ions (0.74 Å) and Ni²⁺ ions (0.69 Å),³⁰ while the lower values of d₁₁₀ and D₁₁₀ in Table S1† for Co2NiAl-HT may be attributed to the slightly smaller ionic radii of Ni²⁺ ions than Co²⁺ ions and the synergetic effect between Co²⁺ ions and Ni²⁺ ions.^30,32

Fig. 1 XRD patterns of the hydrotalcite precursors: Co3Al-HT (a₀) and Co2NiAl-HT (b₀) and the oxide catalysts: Co3AlO-500 (a), Co2NiAlO-500 (b), Co3AlO-800 (a₁) and Co2NiAlO-800 (b₁).

Table 1 Crystallite sizes, lattice parameters, textual parameters, and molar ratios of metal elements of the catalysts

Samples	Crystallite size of spinel^a (nm)	a (nm)	SSA (m^2 g^−1)	Total pore volume (cm^3 g^−1)	(Co + Ni)/Al^b	Co/Ni^b
a Calculated from the diffraction peaks at 2θ ∼ 59° (511) in the XRD (Fig. 1) upon Scherrer formula D = 0.89λ(β cos θ) (λ is the X-ray wavelength (0.1542 nm), θ is the diffraction angle and β is the full width at half maximum (in radian)).b Obtained from the ICP analysis on the as-prepared Co3Al-HT and Co2NiAl-HT precursors.
Co3AlO-500	16.5	0.8063	96.2	0.69	2.83	—
Co2NiAlO-500	11.0	0.8082	88.6	0.78	2.86	2.11
Co3AlO-800	50.0	0.8090	19.3	0.07	2.83	—
Co2NiAlO-800	40.8	0.8091	17.2	0.10	2.86	2.11

---

After calcination at 500 °C and 800 °C, the characteristic XRD reflections of the hydrotalcite-like compounds completely disappear, indicating that the hydrotalcite phase has changed to mixed metal oxide phase, supported by the TG results without obvious weight loss after 500 °C (Fig. S1†). While the strong diffraction peaks for all the catalysts at 2θ ∼ 19°, 31°, 37°, 39°, 45°, 55°, 59° and 65° can be indexed to the (111), (220), (311), (222), (400), (422), (511) and (440) planes of cobalt-based spinel phases (Co₃O₄, JCPDS 42-1467; Co₂AlO₄, JCPDS 38-0814; CoAl₂O₄, JCPDS 44-0160; NiCo₂O₄, JCPDS 73-1702),^24,28 and the additional diffraction peaks appeared at 2θ ∼ 37°, 43° and 63° for Co2NiAlO-800 can be attributed to the (111), (200) and (220) planes of cubic NiO phase (JCPDS 73-1523). From the XRD analysis, it is impossible to distinguish the cobalt-containing spinel-type oxides (Co₃O₄, Co₂AlO₄, CoAl₂O₄ and NiCo₂O₄) due to the almost identical positions of the diffractions. Compared to Co3AlO-500 and Co2NiAlO-500, the obvious increase in the intensity and sharpness of the XRD diffraction peaks for Co3AlO-800 and Co2NiAlO-800 may be attributed to the growth of the spinel phases and sintering of the particles, implying that the increase in the calcination temperature can enhance the crystallinity of the spinel phases.³³ The lattice parameter a values (Table 1) of the catalysts are lower than those of the pure cobalt-containing spinel-type oxides (Co₃O₄ 0.8084 nm, Co₂AlO₄ 0.8086 nm, CoAl₂O₄ 0.8104 nm and NiCo₂O₄ 0.8114 nm), indicating the formation of homogeneous and stable non-stoichiometric spinel-like oxides, probably in the form of anion vacancies and/or dissolution of Al³⁺ ions in the Co₃O₄ and/or NiCo₂O₄ spinel phases.^28,33,34 And these solid solution phases are denoted as Co(Co,Al)₂O₄ and Ni(Co,Al)₂O₄.

Based on the strong and isolated diffraction peaks at 2θ ∼ 59°, the crystallite sizes of spinel phase for the catalysts were calculated by the Scherrer equation and listed in Table 1. The crystallite size of Co2NiAlO-500 is smaller than that of Co3AlO-500, suggesting that the incorporation of Ni in the cobalt-based spinel phase may inhibit the nanocrystal growth. Similar inhibiting effect has been reported by Tang et al. in the formation of Co–Mn solid solution.³⁵ The crystallite sizes of Co3AlO-800 and Co2NiAlO-800 are much larger than those of the 500 °C calcined catalysts, confirming the further growth of the spinel phases due to the higher calcination temperature, while the formation of NiO in Co2NiAlO-800 may also inhibit the growth of the spinel phase to some extent.

EXAFS measurement was performed to confirm the mainly existed Co and Ni phases in the oxide catalysts. The radial structure functions (RSFs) of Co K-edge of the oxide catalysts Co3AlO-500 and Co2NiAlO-500, compared to Co₃O₄, are shown in Fig. 2A. It has been reported that the Co environment in Co₃O₄ corresponds to a spinel atomic arrangement with Co²⁺ and Co³⁺ ions located in tetrahedron (T_d) and octahedron (O_h) coordination, giving an average environment around the Co atoms of 5.3 O at 1.91 Å (first shell), 4Co at 2.85 Å (second shell) and 8Co at 3.36 Å (third shell).³⁶ Thus, the first three peaks in the RSF at 1.496 Å, 2.399 Å, 2.963 Å (not corrected by phase scattering shift) can be ascribed to the above mentioned three coordination shells, and the fourth peak at 4.633 Å corresponds to higher Co–Co coordination shell.^14,37 It can be clearly seen that the RSF of Co K-edge of the oxide catalyst Co3AlO-500, very different from that of CoAl₂O₄,¹⁴ is similar to that of Co₃O₄, implying that Co3AlO-500 mainly existed as Co₃O₄-like spinel phase. While the RSF of Co K-edge of Co2NiAlO-500 (Fig. 2A) is resemble to those of Co₃O₄ and NiCo₂O₄ reported by J. F. Marco et al.³⁸ and its RSF of Ni K-edge (Fig. 2B), different from that of NiO, is similar to that of NiCo₂O₄.³⁸ These results suggest that the oxide catalyst Co2NiAlO-500 mainly existed as NiCo₂O₄-like spinel phase, in line with the above XRD results.

Fig. 2 (A) Co K-edge and (B) Ni K-edge radial structure functions of the catalysts: Co3AlO-500 (a), Co2NiAlO-500 (b) and the references Co₃O₄ and NiO.

The obvious bands around 670 and 568 cm⁻¹ of the oxide catalysts in FT-IR spectra (Fig. 3A), resemble the band position of Co₃O₄ (672 and 590 cm⁻¹), correspond to the ν₁ and ν₂ vibrations of spinel phase.³⁹ The obvious shift of ν₂ band (characteristic of Co³⁺ in octahedral coordination) of the oxides may be ascribed to some dissolution of Al³⁺ in Co₃O₄ to form Co(Al)O solid solution and/or the formation of non-stoichiometric spinel phases,^33,40,41 in line with the XRD results. Moreover, the Raman spectra (Fig. 3B) of the oxide catalysts exhibit the strongest bands at ∼676 cm⁻¹ and very weak bands at ∼186 cm⁻¹, which are different from those of the CoAl₂O₄ with additional bands at 753 cm⁻¹ and 412 cm⁻¹,⁴² but similar to those of the reference Co₃O₄ and the Co–Al mixed oxides derived from hydrotalcites.⁴³ These results further confirm the dissolution of Al³⁺ in Co₃O₄ and NiCo₂O₄ spinel phases and the formation of non-stoichiometric cobalt-based spinel phases.

Fig. 3 (A) FT-IR and (B) Raman spectra of the hydrotalcite precursors Co3Al-HT (a₀) and Co2NiAl-HT (b₀), Co₃O₄ and the catalysts Co3AlO-500 (a), Co2NiAlO-500 (b), Co3AlO-800 (a₁) and Co2NiAlO-800 (b₁).

The (Co + Ni)/Al molar ratios (Table 1) of the oxide catalysts Co3AlO-T and Co2NiAlO-T (T = 500, 800) obtained from ICP data of the corresponding hydrotalcite precursors Co3Al-HT and Co2NiAl-HT are 2.83 and 2.86, respectively, similar to the feeding values (3.0) during the synthesis, indicating the complete coprecipitation of the metal cations in the hydrotalcite precursors. While the Co/Ni molar ratios of Co2NiAlO-T (T = 500, 800) (2.11) are nearly the same as the feeding ratios (2.0), further confirming the substitution of partial Co²⁺ ions by Ni²⁺ ions in the well-crystallized Co2NiAl-HT precursor compared to Co3Al-HT. Thus, the obtained mixed oxide catalysts mainly consist of homogeneous and stable non-stoichiometric cobalt-based spinel phases. In detail, both of Co3AlO-500 and Co3AlO-800 are composed of Co(Co,Al)₂O₄ phase, Co2NiAlO-500 consists of Ni(Co,Al)₂O₄ phase, while Co2NiAlO-800 consists of Co(Co,Al)₂O₄ and NiO phases.

The morphological properties of the hydrotalcite precursors and derived oxide catalysts are revealed by the SEM images (Fig. 4). For the precursors, the high-magnification SEM images (Fig. 4(a₀ and b₀)) show the typical “sand-rose” morphology, characteristic to the hydrotalcite-like materials.^44,45 In detail, the SEM image of Co3Al-HT clearly show the flowerlike ensembles (∼830 nm) consisting of interlaced petal-like nanoplates with thickness of ca. 16 nm and width of ca. 200 nm. While for Co2NiAl-HT, much smaller flowerlike ensembles (∼420 nm) consisting of decreased interlaced petal-like nanoplates (thickness ca. 16 nm and width ca. 116 nm) can be clearly seen, in line with the smaller D₁₁₀ values (Table S1†).

Fig. 4 SEM images of the hydrotalcite precursors: Co3Al-HT (a₀) and Co2NiAl-HT (b₀) and the oxide catalysts: Co3AlO-500 (a), Co2NiAlO-500 (b), Co3AlO-800 (a₁) and Co2NiAlO-800 (b₁) and the EDX of the precursors Co3Al-HT and Co2NiAl-HT (inset in a₀ and b₀).

After calcination at 500 °C, the SEM images of Co3AlO-500 and Co2NiAlO-500 (Fig. 4(a and b)) also show the similar nanoflowerlike morphology with no observable size changes of the ensembles and nanoplates, well inheriting the morphology of the precursors. Compared to Co3AlO-500, the smaller size of nanoplates in Co2NiAlO-500 can be obviously seen, consistent with its smaller crystallite size of spinel phase revealed by XRD. Many stack holes (∼60–100 nm) are formed by the self assembly interlaced petal-like nanoplates in the nanoflowerlike oxide catalysts. However, after calcination at 800 °C, the SEM images of Co3AlO-800 and Co2NiAlO-800 (Fig. 4(a₁ and b₁)) depict the formation of irregular nanoparticles which are disorderly stacked and severely agglomerated, completely different from the nanoflowerlike morphology of the precursors. The EDX results of the precursors (inset in Fig. 4(a₀ and b₀)) reveal the existence of Co, Al, O, C and Co, Ni, Al, O, C elements in the hydrotalcite precursors Co3Al-HT and Co2NiAl-HT, respectively. The mapping analysis of the samples (Fig. S2†) shows that all the precursors and catalysts possess nearly uniformly distributed metal elements (Co, Ni and Al) due to the atomic-scale uniform distribution of metal cations in the hydrotalcite layers and topotactic transformation nature of the hydrotalcite thermal decomposition to mixed oxides.^22,23 Thus, it can be deduced that the nanoflowerlike morphology of cobalt-based mixed oxide catalysts may avoid the severe aggregation of the particles and disperse the active redox components, and the presence of many stack holes in the nanoflowerlike oxide catalysts may lead to the relatively high surface areas and pore volumes.

The textural structures of the oxide catalysts are studied by low temperature N₂ adsorption analysis. The N₂ adsorption–desorption isotherms of Co3AlO-500 and Co2NiAlO-500 (Fig. 5) exhibit type IV isotherms with H3 type hysteresis loop upon the IUPAC classification,⁴⁶ typical of mesoporous materials, implying the probable existence of slit-shaped pores.^18,26,47,48 The BJH pore size distribution curves of Co3AlO-500 and Co2NiAlO-500 (inset in Fig. 5) show that the pore sizes at the maximum probability of the samples are 8.6 nm, suggesting that the samples possess relatively uniform mesoporous structure.³⁷ The BET surface area and total pore volume of the oxide catalysts are summarized in Table 1. The catalysts Co3AlO-500 and Co2NiAlO-500 possess larger specific surface area of 96.2–88.6 m² g⁻¹ and total pore volume of 0.69–0.78 m³ g⁻¹, compared to those of Co3AlO-800 (19.3 m² g⁻¹, 0.07 m³ g⁻¹) and Co2NiAlO-800 (17.2 m² g⁻¹, 0.10 m³ g⁻¹), in agreement with the smaller spinel crystallite sizes revealed by XRD and the nanoflowerlike morphology shown by SEM, which is benefit for the exposure of the active sites. The nanoflowerlike morphology and larger surface area can also facilitate the diffusion of gas reactants and products and the gas–solid interaction.

Fig. 5 N₂ adsorption–desorption isotherms and pore size distribution curves of Co3AlO-500 (a) and Co2NiAlO-500 (b).

The morphology and microstructure of the mixed oxide catalysts are further revealed by the HRTEM analysis. The HRTEM images of Co3AlO-500 and Co2NiAlO-500 (Fig. 6(a and b)) show relatively small spherical-like nanoparticles with sizes of 6–22 nm, close to the crystallite sizes of spinel phases (11.0–16.5 nm) revealed by XRD. One can also recognize the slit-like pores formed by these small nanoparticles, confirming the mesoporous characteristic of these catalysts. However, the HRTEM images for Co3AlO-800 and Co2NiAlO-800 (Fig. 6(c and d)) show typical hexagonal sheet-like nanoparticles with sizes of 50–80 nm, slightly bigger than the crystallite sizes of spinel phases (40.8–50 nm). Compared to Co3AlO-500 and Co2NiAlO-500, the greatly increased particles sizes and crystallinity, and the greatly decreased surface area and pore volume for Co3AlO-800 and Co2NiAlO-800 can be ascribed to the sintering effect that promotes small particles to merge and recrystallize into larger particles¹⁹ and also leads to the disappear of mesoporous structure.

Fig. 6 HRTEM images, lattice images and corresponding fast Fourier transform (FFT) patterns of the oxide catalysts: Co3AlO-500 (a, a′, inset in a′), Co2NiAlO-500 (b, b′, inset in b′), Co3AlO-800 (c, c′, inset in c′) and Co2NiAlO-800 (d, d′, inset in d′).

The morphology and structure of these nanoparticles have also been analyzed by HRTEM shown in Fig. 6(a′–d′). The catalysts Co3AlO-500, Co2NiAlO-500 and Co2NiAlO-800 are revealed to be assembled by spherical or rod-like nano-crystallites with random orientation and high dispersion.³⁷ While Co3AlO-800 only shows a well-defined large plate-like nano-crystallite with high crystallinity. The observed distinct lattice fringes in HRTEM images (Fig. 6(a′, b′ and d′)) of Co3AlO-500, Co2NiAlO-500 and Co2NiAlO-800 with interplanar distances of ca. 0.47 nm, ca. 0.28 nm and ca. 0.24 nm correspond to the crystallographic planes (111), (220) and (311), respectively, of the cubic cobalt-based spinel phases,^27,43 and the corresponding fast Fourier transform (FFT) patterns (inset in Fig. 6(a′, b′ and d′)) also show diffraction spots due to these planes of the cubic cobalt-based spinel phases, in line with the XRD results. For Co3AlO-800, the special interplanar angles between the planes with interplanar distances of 0.473 nm, 0.283 nm and 0.241 nm indicate that the observed planes are (111), (2−20) and (3−11) planes.⁴⁹ In addition, the lattice fringes in HRTEM image of Co2NiAlO-800 (Fig. S3†) with the interplanar distance of 0.212 nm correspond to the (200) plane of cubic NiO phase,⁵⁰ in line with the XRD results. Thus, the nanoflowerlike mixed oxide catalysts Co3AlO-500 and Co2NiAlO-500 possess highly dispersed active cobalt-based spinel phases with smaller crystallite sizes, larger surface areas and mesoporous structure, which may give more surface active sites and are expected to exhibit higher catalytic activity.

3.2 Catalytic NO oxidation performance

The catalytic activities of the oxide catalysts Co3AlO-500, Co2NiAlO-500, Co3AlO-800 and Co2NiAlO-800 towards NO oxidation are shown in Fig. 7. The thermodynamic equilibrium curve of NO to NO₂ under given conditions shows that the equilibrium conversion decreases with the increasing reaction temperature because the reaction is exothermic and reversible.¹⁷ In the temperature range 150–350 °C, the catalysts Co3AlO-500 and Co2NiAlO-500 show similar catalytic activity towards NO oxidation, much higher than Co3AlO-800 and Co2NiAlO-800. In detail, at lower temperatures, the NO conversion for all the catalysts are relatively low since the NO oxidation is kinetically limited.¹ With the temperature rising, the NO conversion over Co3AlO-500 and Co2NiAlO-500 first increase to the maximum ∼88% at 285 °C and then drop down due to the thermodynamic limitations, while the NO conversion of Co3AlO-800 and Co2NiAlO-800 only increase to the maximum 55.6% and 66.6% at 350 °C, respectively. The larger specific surface areas of the catalysts Co3AlO-500 and Co2NiAlO-500 enabled by the smaller crystallite sizes and mesoporous structure improve their low temperature NO conversion. Similar phenomenon has been reported in the Co₃O₄ nanowires-array catalysts by Ren et al.¹⁹ From the catalytic oxidation performance of the oxide catalysts listed in Table 2, it is noted that the NO conversion for Co3AlO-500 and Co2NiAlO-500 are close to those of the previously reported Co₃O₄ nanowire-arrays prepared from cobalt nitrate and cobalt acetate¹⁹ and La_0.9Ba_0.1CoO₃ perovskites¹⁷ and higher than that of Co₃O₄/CeO₂ (ref. 18) and much higher than those of the Co₃Al-oxide²⁵ and CoMgAlO-array²⁷ derived from hydrotalcites precursors.

Fig. 7 NO conversion over the catalysts Co3AlO-500, Co2NiAlO-500, Co3AlO-800 and Co2NiAlO-800 as a function of temperature. Reaction conditions: 500 ppm NO, 8% O₂, N₂ balance, GHSV = 100 000 mL g⁻¹ h⁻¹.

Table 2 Comparison of the catalytic performance of the present cobalt-based oxide catalysts and related oxide catalysts for the NO oxidation

Catalysts	TCalc. (°C)	SSA (m^2 g^−1)	Temp. (°C)	Conv.NO (%)	Reactant gas mixture	Space velocity/mL g^−1 h^−1	Ref.
Co3AlO-500	500	96.2	285	88.8	500 ppm NO, 8% O2, N2 balance	100 000	This work
Co2NiAlO-500	500	88.6	285	87.6	500 ppm NO, 8% O2, N2 balance	100 000	This work
Co3AlO-800	800	19.3	350	55.6	500 ppm NO, 8% O2, N2 balance	100 000	This work
Co2NiAlO-800	800	17.2	350	66.6	500 ppm NO, 8% O2, N2 balance	100 000	This work
Co3O4 nano-arrays	300	122.6	275	80	500 ppm NO, 10% O2, N2 balance	50 000 h^−1	19
La0.9Ba0.1CoO3	700	7.48	265	93	400 ppm NO, 10% O2, N2 balance	180 000	17
Co3O4/CeO2	500	21	325	∼67	390 ppm NO, 8% O2, N2 balance	20 000	18
Co3Al-oxide	800	22	300	48.8	800 ppm NO, 8% O2, N2 balance	30 000 h^−1	25
CoMgAlO-array	500	116	300	46	500 ppm NO, 8% O2, N2 balance	100 000	27

---

The effect (Fig. 8A) of NO concentration (300–700 ppm) on NO conversion of the catalysts Co3AlO-500 and Co2NiAlO-500 was studied. It is noted that the conversion curves for Co3AlO-500 and Co2NiAlO-500 are indistinguishable, similar to that observed in Fig. 7, confirming the similar NO oxidation performance for these two catalysts. Cai et al.⁵¹ reported that in the modified Cr/Ce_0.2Zr_0.8O₂ catalyst with rigid benzene-muti-carboxylate ligands, the NO conversion decreased from 63% to 47% with the increase of NO concentration from 300 ppm to 570 ppm, and then almost stayed a constant with the increase of NO concentration from 570 ppm to 660 ppm. Interestingly, in our case, the NO conversion (∼86%) of Co3AlO-500 and Co2NiAlO-500 almost maintained the same as the NO concentration increased from 300 ppm to 700 ppm, different from the phenomenon reported by Cai et al., suggesting the neglectable effect of the NO concentration (300–700 ppm) on the NO conversion over these oxide catalysts and the varied reaction mechanism between our catalysts and the modified Cr/Ce_0.2Zr_0.8O₂ catalysts. These results also mean that the cobalt-based mixed oxide catalysts can keep high NO conversion with a relatively wide NO concentration range, which is very appealing in the NO_x removal application.

Fig. 8 (A and B) The effect of NO and O₂ concentration on NO conversion over catalysts Co3AlO-500 and Co2NiAlO-500 at 300 °C. (C) The stability tests over catalysts Co3AlO-500 and Co2NiAlO-500 at 300 °C for 20 h (reaction conditions: 500 ppm NO, 8% O₂, N₂ as balance gas, GHSV = 100 000 mL g⁻¹ h⁻¹). (D and E) Reaction rates r measured at 150 °C as a function of NO or O₂ concentration over catalysts Co3AlO-500 and Co2NiAlO-500. (F) The Arrhenius plots for catalysts Co3AlO-500 and Co2NiAlO-500 in the temperature range of 150–190 °C.

The effect of O₂ concentration (2–9%) on NO conversion of the catalysts Co3AlO-500 and Co2NiAlO-500 was also studied. As Fig. 8B shows, the NO conversion increased slowly from 75.1% to 86.6% with the increase of O₂ concentration from 2% to 8% and almost kept the same with the further increase of O₂ concentration from 8% to 9%, similar to the phenomenon of the Co₃O₄ nanowires-array catalysts.¹⁹ The relatively high conversion in the whole O₂ concentration range (2–9%) of these catalysts suggests that they are reliable to work under both lean-burn and rich-burn conditions, which is quite important for the NO_x removal of the internal combustion engines.^13,19 The long term isothermal catalytic stability tests of catalysts Co3AlO-500 and Co2NiAlO-500 were studied by monitoring the NO conversion change at the temperature 300 °C. The NO conversion was quantified every 2 hours using the Chemiluminescence NO–NO₂–NO_x analyzer. Fig. 8C demonstrated stable NO conversion without any degradation during the 20 hours' test and both catalysts maintained the high NO conversion of ∼86%. Moreover, the surface areas and SEM images (Fig. S4†) of the used catalysts Co3AlO-500-u and Co2NiAlO-500-u after the stability tests were also obtained. The surface area of Co3AlO-500-u (58.8 m² g⁻¹) was lower than Co3AlO-500 (96.2 m² g⁻¹) along with aggregated nanoplates (Fig. S4a†) instead of flowerlike morphology of fresh Co3AlO-500. While the surface area of Co2NiAlO-500-u (88.0 m² g⁻¹) was well-kept compared to fresh one (88.6 m² g⁻¹), in line with its flowerlike morphology (Fig. S4b†). These results imply that Co2NiAlO-500 may maintain the high NO oxidation activities for a longer time than Co3AlO-500.

It is reported that the NO oxidation reaction is a reversible reaction and NO₂ formation may inhabit the NO conversion.^15,17,19 While in our catalytic tests, the NO oxidation can be reasonably considered as irreversible since the O₂ concentration (8%) is high enough compared with the NO concentration (500 ppm). Thus, the reaction rate only depends on the concentration of NO and O₂. For the catalysts Co3AlO-500 and Co2NiAlO-500, the reaction orders of NO (1.37 and 1.28) and O₂ (0.50 and 0.56) were obtained by the kinetic analysis (Fig. 8(D and E)) through only changing the concentration of NO or O₂, respectively. These results indicate that the NO oxidation over the catalysts Co3AlO-500 and Co2NiAlO-500 follows the Eley–Rideal mechanism where NO is gaseous and O₂ is at adsorbed state, similar to the previously reported Co₃O₄ based catalysts for NO oxidation.^10,19 The calculated results of the apparent activation energy (E_a) and pre-exponential factor (A) of the catalysts Co3AlO-500 and Co2NiAlO-500 are given in Table S2.† The E_a and A values were obtained by the Arrhenius plot, in which the reaction rate constant k was calculated upon the measured reaction rate r and reaction orders a and b. The E_a values for catalysts Co3AlO-500 and Co2NiAlO-500 obtained in the temperature range of 150–190 °C (Fig. 8F) were 29.12 and 29.84 kJ mol⁻¹, respectively. The similarity of the E_a values between Co3AlO-500 and Co2NiAlO-500 indicates that the active sites and reaction mechanism of the NO oxidation reaction over these catalysts may be similar.^19,52 In addition, the catalyst Co3AlO-500 possesses slightly larger pre-exponential factor (A) value (19.44 mol g⁻¹ s⁻¹), compared to Co2NiAlO-500 (18.39 mol g⁻¹ s⁻¹), suggesting the slightly more active sites of Co3AlO-500 exposed.¹⁹ Compared to Co3AlO-800 and Co2NiAlO-800, the excellent catalytic NO oxidation performance of Co3AlO-500 and Co2NiAlO-500 can be assigned to the cobalt-based spinel phases with smaller crystallite sizes, larger surface areas and mesoporous structure resulting from topotactic transformation of the hydrotalcite-like precursors at suitable temperature, which may give larger amount of surface active sites.

3.3 Catalytic mechanism

After the microstructures and the catalytic performance of the cobalt-based mixed oxide catalysts have been carefully analyzed, a noticeable problem is why the different catalysts Co3AlO-500 and Co2NiAlO-500 exhibit similar NO oxidation performance. XPS was used as an effective tool to gain information of the surface element compositions and chemical states of the catalysts Co3AlO-500 and Co2NiAlO-500 and the used ones Co3AlO-500-u and Co2NiAlO-500-u after stability tests. The XPS spectra of Co 2p, Ni 2p, Al 2p and O 1s are illustrated in Fig. 9. The binding energies (BE) of Co 2p_3/2, Ni 2p_3/2, Al 2p and O 1s and the molar ratios of the surface Co³⁺/Co²⁺, Ni³⁺/Ni²⁺, O_ads/O_latt, Co/Ni and (Co + Ni)/Al are summarized in Table 3. The molar ratios of Co³⁺/Co²⁺, Ni³⁺/Ni²⁺ and O_ads/O_latt were obtained by quantitative calculation of the corresponding deconvoluted peak areas.

Fig. 9 XPS spectra of Co 2p, Ni 2p, Al 2p and O 1s for the oxide catalysts Co3AlO-500 (a), Co2NiAlO-500 (b) and used ones Co3AlO-500-u (c) and Co2NiAlO-500-u (d).

Table 3 XPS binding energies (BE) of the peak-fitted Co 2p_3/2, Ni 2p_3/2, O 1s and BE of Al 2p and the surface molar ratios of Co³⁺/Co²⁺, Ni³⁺/Ni²⁺, O_ads/O_latt, Co/Ni and (Co + Ni)/Al of the catalysts Co3AlO-500, Co2NiAlO-500 and used ones Co3AlO-500-u and Co2NiAlO-500-u

Catalysts	Binding energy (eV)							Surface molar ratio
	Co^2+	Co^3+	Ni^2+	Ni^3+	Olatt	Oads	Al2p	Co^3+/Co^2+	Ni^3+/Ni^2+	Oads/Olatt	Co/Ni	(Co + Ni)/Al
Co3AlO-500	781.3	779.8	—	—	530.1	531.5	73.8	1.22	—	0.77	—	0.84
Co2NiAlO-500	780.9	779.2	854.1	855.8	529.7	531.7	73.6	1.30	1.24	1.05	1.31	1.24
Co3AlO-500-u	781.6	780.1	—	—	530.3	531.7	73.5	1.25	—	0.82	—	0.86
Co2NiAlO-500-u	781.1	779.5	854.1	855.7	529.8	531.4	73.6	1.31	1.26	1.12	1.25	1.20

---

The high resolution Co 2p spectra of Co3AlO-500 and Co2NiAlO-500 present two distinct peaks centered at ∼780 eV and ∼795 eV attributed to Co 2p_3/2 and Co 2p_1/2, respectively.^19,32,53 The spin–orbit splitting value is 15.3–15.4 eV, implying the coexistence of Co²⁺ and Co³⁺ species.³⁷ By performing the peak-fitting deconvolution, the Co 2p_3/2 peaks can be divided into two characteristic peaks at ∼779.2 eV and ∼780.9 eV, assigned to Co³⁺ and Co²⁺ in the Co₃O₄-like and NiCo₂O₄-like spinel phases.^19,32,53 Moreover, it is noted that the BE values of the two deconvoluted peaks of Co 2p_3/2 for Co2NiAlO-500 (779.2 eV, 780.9 eV) are both slightly lower than those of Co3AlO-500 (779.8 eV, 781.3 eV). The surface Co³⁺/Co²⁺ molar ratio of catalyst Co2NiAlO-500 is 1.30 (Table 3), slightly higher than that of Co3AlO-500 (1.22), suggesting that the Co³⁺ species are dominant on the surface of these catalysts and the oxidation state of Co can be changed via the addition of Ni due to the strong interaction between Co and Ni.

The high resolution Ni 2p spectrum of Co2NiAlO-500 exhibits two strong peaks at 854.9 eV for Ni 2p_3/2 and 872.9 eV for Ni 2p_1/2, and two strong satellite peaks at 861.2 eV and 879.3 eV. The Ni 2p_3/2 peak is deconvoluted into two peaks at 854.1 and 855.8 eV, ascribed to Ni³⁺ and Ni³⁺, respectively,^20,38 and the surface Ni³⁺/Ni²⁺ molar ratio (Table 3) is 1.24. These results suggest that the Co2NiAlO-500 possesses a surface composition of Co²⁺, Co³⁺, Ni²⁺ and Ni³⁺, in line with the literature observations for NiCo₂O₄.^54,55 It can be seen that the surface (Co + Ni)/Al molar ratios are 0.84 and 1.24 for Co3AlO-500 and Co2NiAlO-500, respectively, much lower than the bulk (Co + Ni)/Al molar ratios of 2.83 and 2.86, suggesting the surface enrichment of Al on these catalysts.³³ While Co2NiAlO-500 possesses higher surface (Co + Ni)/Al molar ratio than Co3AlO-500, indicating that the addition of Ni may prohibit the surface enrichment of Al and increase the surface content of Co and Ni. Moreover, the surface Co/Ni molar ratio (1.31) of Co2NiAlO-500 is much lower than the bulk Co/Ni ratio (2.11), indicating the surface enrichment of Ni on the catalyst. The Al 2p spectra of Co3AlO-500 and Co2NiAlO-500 show a single peak at ∼73.6 eV. The BE values of Al 2p (Table 3) are slightly lower than those of Al₂O₃ (74.1 eV) as well as CoAl₂O₄ and NiAl₂O₄ (73.93 eV),³⁰ confirming the existence of non-stoichiometric cobalt-based spinel phases Co(Co,Al)₂O₄ and Ni(Co,Al)₂O₄.

The asymmetrical O 1s peak of catalysts Co3AlO-500 and Co2NiAlO-500 can be deconvoluted into two peaks at BE values of ∼530.1 eV and ∼531.5 eV, ascribed to the lattice oxygen (O²⁻) and surface adsorbed oxygen (O⁻, O₂⁻, O₂²⁻), respectively.^1,32,51,53 It has been reported that the adsorption of gaseous O₂ on oxygen vacancies can lead to the formation of surface adsorbed oxygen on the oxide catalysts.^20,27 As listed in Table 3, the catalyst Co2NiAlO-500 possesses a much higher surface O_ads/O_latt molar ratio (1.05) than Co3AlO-500 (0.77), implying the formation of additional oxygen vacancies due to the partial substitution of octahedral Co³⁺ by Ni²⁺.⁵⁶ When these additional oxygen vacancies are filled up with gaseous O₂ leading to the increased O_ads/O_latt molar ratio, the neighboring Co²⁺ may be oxidized into Co³⁺ to balance the charge, resulting in the higher Co³⁺/Co²⁺ molar ratio.¹ Generally, the surface adsorbed oxygen is much more reactive in the oxidation reaction than lattice oxygen due to its higher mobility.^27,53,57

However, considering the surface Co atomic content of Co3AlO-500 and Co2NiAlO-500 is 13.21% and 5.85%, respectively, thus the surface Co³⁺ content of Co3AlO-500 and Co2NiAlO-500 is 7.26% and 3.31%, respectively. While the surface Ni atomic content of Co2NiAlO-500 is 4.46%, then the surface Ni³⁺ content is 2.47%. Thus, the total surface Co³⁺ and Ni³⁺ content (5.78%) of Co2NiAlO-500 is slightly lower than the surface Co³⁺ content (7.26%) of Co3AlO-500. While the surface O atomic content of Co3AlO-500 and Co2NiAlO-500 is 50.4% and 41.2%, respectively, thus, the surface adsorption oxygen content of Co3AlO-500 (21.9%) is close to that of Co2NiAlO-500 (21.1%). Moreover, the obtained BE values of Co²⁺, Co³⁺, Ni²⁺, Ni³⁺, O_ads and O_latt and the molar ratios of Co³⁺/Co²⁺, Ni³⁺/Ni²⁺ and O_ads/O_latt for the used catalysts Co3AlO-500-u and Co2NiAlO-500-u are similar to those of the corresponding fresh catalysts, indicating the almost unchanged surface composition of the used catalysts. These results are in good agreement with the stable NO conversion of Co3AlO-500 and Co2NiAlO-500 during the 20 hours' test.

With regard to the active sites for the NO oxidation reaction, Ren et al.¹⁹ reported that the Co³⁺ on the surface acted as the active sites of the Co₃O₄ nanowire-array catalysts, while Feng et al.¹ reported that some of the surface adsorbed oxygen species served as the active site over the La-modified SmMn₂O₅ mullite oxide (La_xSm_1−xMn₂O_δ) catalysts, and Chen et al.⁵² reported that Mn⁴⁺ ion associated with O ligand served as the active sites over the LaMnO₃ perovskite catalysts. If we presume the Co³⁺ ions associated with surface adsorbed oxygen as the active sites for Co3AlO-500, then the amount of (Co³⁺ and Ni³⁺) ions associated with surface adsorbed oxygen active sites for Co2NiAlO-500 may be a little lower than that of Co3AlO-500. However, the catalytic activity of NO oxidation catalyst is not simply determined by the amount of the active sites, but the reducibility of the catalysts and the adsorbed species formed on the surface of the catalysts may also play crucial roles.

H₂-TPR experiment was carried out to investigate the reducibility of the catalysts Co3AlO-500 and Co2NiAlO-500. As shown in Fig. 10, the H₂-TPR profile of Co3AlO-500 exhibited a weak peak centered at 347 °C and a strong one at 668 °C. The former can be attributed to the reduction of Co³⁺ to Co²⁺ in the Co³⁺-related spinels, while the latter assigned to the reduction of Co²⁺-related species to Co⁰.^14,20,27,58 The H₂-TPR profile of Co2NiAlO-500 presented one weak peak at 292 °C and one strong and asymmetric peak in the temperature range of 300–800 °C. The former is ascribed to the Co³⁺ to Co²⁺, similar to Co3AlO-500. While the shoulder (∼500 °C) of the latter peak is probably related to the reduction of Ni²⁺-related spinels,³² and the main peak centered at 631 °C can be assigned to the reduction of Co²⁺-related species, the two reduction peaks are overlapped to form the strong and broad peak,²⁰ which is different from that of Co3AlO-500. It should be mentioned that the two reduction peaks of Co2NiAlO-500 shift to lower temperatures compared to Co3AlO-500 and the total H₂ consumption of Co2NiAlO-500 (7.40 mmol g⁻¹) is slightly higher than that of Co3AlO-500 (7.24 mmol g⁻¹). These results suggest that Co2NiAlO-500 possesses higher reducibility than Co3AlO-500 due to the existence of strong interaction between Co and Ni in the NiCo₂O₄-like spinel phase.²⁰

Fig. 10 H₂-TPR profiles of Co3AlO-500 (a) and Co2NiAlO-500 (b).

NO-TPD and in situ FT-IR analyses are combined to reveal the adsorbed species formed on the surface of the catalysts during the NO oxidation process. The NO-TPD profiles of Co3AlO-500 and Co2NiAlO-500 (Fig. 11A and B) for NO adsorption alone at 50 °C showed two distinct NO desorption peaks. The strong and broad NO desorption peaks centered at 152 °C are assigned to the decomposition of nitrite species, while the slightly lower and narrower NO desorption peaks around 302–324 °C are assigned to the decomposition of nitrate species.⁵⁹ Moreover, the weak and broad NO₂ desorption peaks around 288–314 °C are also observed, assigned to the decomposition of nitrates species. Although NO alone is adsorbed over Co3AlO-500 and Co2NiAlO-500, nitrates species are formed surprisingly, indicating the existence of active oxygen species on the surface of these catalysts after 8% O₂/N₂ pre-treatment. Compared to Co3AlO-500, it can be seen that the desorption peaks' areas of NO and NO₂ for Co2NiAlO-500 are much larger, indicating that larger amount of nitrite/nitrate species are formed on the catalyst Co2NiAlO-500 at low adsorption temperature. The NO-TPD profiles of Co3AlO-500 and Co2NiAlO-500 (Fig. 11C and D) for NO and O₂ co-adsorption at 300 °C show one stronger NO desorption peak around 316–329 °C and one weaker NO₂ desorption peak around 280–307 °C, assigned to the decomposition of nitrate species.⁵⁹ Compared to Co3AlO-500, although the desorption peaks' area of NO for Co2NiAlO-500 is much larger, the desorption peaks' area of NO₂ for Co2NiAlO-500 is only slightly larger, indicating the similar desorption amount of NO₂ over the two catalysts at high adsorption temperature.

Fig. 11 NO-TPD profiles of Co3AlO-500 (A and C) and Co2NiAlO-500 (B and D) for NO adsorption at 50 °C (A and B) and NO + O₂ co-adsorption at 300 °C (C and D).

The adsorbed nitrate species at high adsorption temperature over the catalysts Co3AlO-500 and Co2NiAlO-500 were further revealed by the in situ FT-IR analysis. Fig. 12 showed the in situ FT-IR spectra of the catalysts after the introduction of 500 ppm NO at 300 °C and followed 8% O₂ was introduced. NO adsorption on these catalysts was performed first. For Co3AlO-500 (Fig. 12A), after 5 min of exposure, the broad peak at 1577 cm⁻¹ is ascribed to the ν(N=O) of the chelating bidentate nitrate while the smaller shoulder at 1630 cm⁻¹ is assigned to the ν(N=O) of the bridged bidentate nitrate and weakly adsorbed NO₂.^24,60,61 After 5 min of exposure, the catalyst Co2NiAlO-500 (Fig. 12B) exhibits the main peak at 1630 cm⁻¹ and a shoulder at 1598 cm⁻¹, indicating the much larger amount of the bridged bidentate nitrate and weakly adsorbed NO₂. In addition, two very weak peaks are observed at 1906 and 1853 cm⁻¹ for Co2NiAlO-500, may be assigned to linear Ni²⁺–NO species and Ni²⁺(NO)₂ geminal species.⁶² It is noticed that the intensities of all these peaks are not obviously increased with time, implying that the formation of these nitrates is very fast and the amount of formed nitrates over these catalysts is very limited due to the limited surface active oxygen species.⁴⁹ When O₂ supply is switched on, the intensities of the above mentioned peaks are increased rapidly, and almost reach the strongest after the NO + O₂ feeding for 5 min, implying that the introduction of O₂ can further promote the oxidation of NO to NO₂ and the storage as nitrate species probably due to the complement of the surface active oxygen species. Some new peaks could also be observed. In detail, the weak peaks at 1286–1289 cm⁻¹ and 1034–1037 cm⁻¹ are ascribed to the ν_asym(NO₂) and ν_sym(NO₂) of the chelating bidentate nitrate, while the weak peaks at 1236–1234 cm⁻¹ and 999–1005 cm⁻¹ are ascribed to the ν_asym(NO₂) and ν_sym(NO₂) of the bridged bidentate nitrate.⁶⁰ A very weak shoulder at 1195 cm⁻¹ for Co3AlO-500 may be attributed to the ν(N–O) vibration of anionic NO_x surface species such as NO⁻ and/or bidentate nitrites.⁶³

Fig. 12 The in situ FT-IR spectra of Co3AlO-500 (A) and Co2NiAlO-500 (B).

It is noted that the peaks at 1598 cm⁻¹ for chelating bidentate nitrate species in the in situ FT-IR spectra of Co2NiAlO-500 are broader than those of Co3AlO-500 at 1577 cm⁻¹, and accompanied by much higher shoulders, in line with the larger amount of released NO in the NO-TPD profiles. These results suggest that the NiCo₂O₄-like spinel phase can capture more NO stored as chelating bidentate nitrate than Co₃O₄-like spinel phase, which may be related to the strong synergistic effect of Co and Ni originating from the hydrotalcite precursor method. Besides the obviously existed chelating bidentate nitrate species in these catalysts, it should be noted that the intensity of the peaks at 1630 cm⁻¹ corresponding to less stable bridged bidentate nitrate and weakly adsorbed NO₂ species of Co2NiAlO-500 is much higher than that of Co3AlO-500, while the appearance of the shoulders at 1648 cm⁻¹ of Co3AlO-500 also indicate more production of NO₂.⁶¹ These phenomena suggest that the two catalysts possess similar amount of less stable bridged bidentate nitrate and weakly adsorbed NO₂ species, in line with the NO-TPD results for the co-adsorption of NO and O₂ at 300 °C.

By combining the evidence from XPS, H₂-TPR, NO-TPD and in situ FT-IR, the reaction pathways for NO oxidation to NO₂ over the catalysts Co3AlO-500 and Co2NiAlO-500 are tentatively proposed and depicted in Scheme 1. In brief, the NO oxidation pathways mainly include: (I) the NO oxidation by the Co³⁺ and/or Ni³⁺ associated with surface adsorbed oxygen (Co³⁺/Ni³⁺–O_ads) species, and (II) the NO oxidation by the Co³⁺ and/or Ni³⁺ associated with lattice oxygen (Co³⁺/Ni³⁺–O_latt) species. In Scheme 1, two kinds of surface reactive species are ideally representative by the cuboids with different colours (Co³⁺/Ni³⁺–O_ads species: yellow colour; Co³⁺/Ni³⁺–O_latt species: light gray colour), while the bulk of the catalysts is representative by the big cuboids with dark gray colour.

Scheme 1 The reaction pathways for NO oxidation on the surface of the cobalt-based mixed oxide catalysts.

For pathway (I), NO was first oxidized by the Co³⁺/Ni³⁺–O_ads species on the surface of the catalysts, and formed the less stable bridged bidentate nitrate and weakly adsorbed NO₂ species, followed by their decomposition into NO₂ to generate oxygen vacancies, and finally the gaseous O₂ adsorbed onto the oxygen vacancies to replenish the surface adsorbed oxygen and complete the redox cycle. The pathway (I) is in line with the above mentioned Eley–Rideal mechanism. While for pathway (II), the NO was first oxidized by the Co³⁺/Ni³⁺–O_latt species on the surface of the catalysts, and then formed the relatively stable chelating bidentate nitrate species, which were prone to decompose into NO, supported by the large amount of the decomposed NO from the nitrate species in the NO-TPD profiles. So pathway (II) is not considered as an effective NO oxidation route. Therefore, the Co³⁺ and Ni³⁺ associated with surface adsorbed oxygen and Co³⁺ associated with surface adsorbed oxygen species are believed to serve as the active sites over the present Co3AlO and Co2NiAlO catalysts, respectively, in the NO oxidation reaction.

Although Co2NiAlO-500 possesses slightly smaller amount of active sites (Co³⁺ and Ni³⁺ associated with surface adsorbed oxygen), compared to Co3AlO-500 with active sites (Co³⁺ associated with surface adsorbed oxygen), catalyst Co2NiAlO-500 exhibits higher reducibility, leading to their similar desorption amount of NO₂, thus similar excellent NO oxidation performance. While the lower cost of Co2NiAlO-500 renders it a greater potential for practical applications. Moreover, the exploration of the relationship of the excellent catalytic NO oxidation activity and the microstructure of the facilely synthesized nanoflowerlike oxide catalysts may provide some new insights for the design of hierarchical Co-based oxide catalysts in the NO_x removal area.

4. Conclusions

Nonprecious cobalt-based mixed oxide catalysts Co3AlO-500, Co2NiAlO-500, Co3AlO-800 and Co2NiAlO-800 have been fabricated by thermal decomposition of nanoflowerlike hydrotalcite-like precursors Co3Al-HT and Co2NiAl-HT at 500 °C and 800 °C in air. The obtained oxide catalysts mainly consisted of homogeneous and stable non-stoichiometric cobalt-based spinel phases Co(Co,Al)₂O₄ and Ni(Co,Al)₂O₄. The catalysts Co3AlO-500 and Co2NiAlO-500 both showed nanoflowerlike morphology, while catalysts Co3AlO-800 and Co2NiAlO-800 both presented disorderly stacked states consisting of severely agglomerated irregular nanoparticles. The catalysts Co3AlO-500 and Co2NiAlO-500 presented excellent NO oxidation performance, with the maximum conversion efficiency ca. 88.8% and 87.6% (285 °C), respectively, which were much higher than the values of Co3AlO-800 and Co2NiAlO-800 (55.6% and 66.6% at 350 °C). The excellent catalytic NO oxidation performance of catalysts Co3AlO-500 and Co2NiAlO-500 could be attributed to the much larger amount of surface active sites endowed by these catalysts with smaller crystallite sizes of cobalt-based spinel phase, higher surface areas, and mesoporous structure. The mechanism study revealed the active sites and oxidation pathways during the NO oxidation reaction over the present catalysts. Although Co2NiAlO-500 possessed slightly smaller amount of active sites (Co³⁺ and Ni³⁺ associated with surface adsorbed oxygen), compared to Co3AlO-500 with active sites (Co³⁺ associated with surface adsorbed oxygen), the reducibility of Co2NiAlO-500 was higher, leading to the similar desorption amount of NO₂, thus similar excellent NO oxidation performance of nanoflowerlike catalysts Co3AlO-500 and Co2NiAlO-500. Moreover, catalyst Co2NiAlO-500 has lower cost than Co3AlO-500, rendering it a greater potential for practical applications.

Acknowledgements

The authors greatly appreciate the financial support by the National Nature Science Foundation of China (21276015, 21576013, 21521005).

References

1. Z. J. Feng, J. Q. Wang, X. Liu, Y. W. Wen, R. Chen, H. F. Yin, M. Q. Shen and B. Shan, Catal. Sci. Technol., 2016, 6, 5580–5589.
2. C. H. Kim, G. S. Qi, K. Dahlberg and W. Li, Science, 2010, 327, 1624–1627.
3. J. R. Theis and E. Gulari, Appl. Catal., B, 2007, 74, 40–52.
4. X. Auvray, T. Pingel, E. Olsson and L. Olssona, Appl. Catal., B, 2013, 129, 517–527.
5. S. F. Zhong, Y. Y. Sun, H. C. Xin, C. P. Yang, L. Chen and X. B. Li, Chem. Eng. J., 2015, 275, 351–356.
6. M. Jabłońska and R. Palkovits, Catal. Sci. Technol., 2016, 6, 49–72.
7. J. J. Yu, J. Cheng, C. Y. Ma, H. L. Wang, L. D. Li, Z. P. Hao and Z. P. Xu, J. Colloid Interface Sci., 2009, 333, 423–430.
8. G. Liu and P. X. Gao, Catal. Sci. Technol., 2011, 1, 552–568.
9. Y. Peng, W. Z. Si, J. M. Luo, W. K. Su, H. Z. Chang, J. H. Li, J. M. Hao and J. Crittenden, Environ. Sci. Technol., 2016, 50, 6442–6448.
10. M. F. Irfan, J. H. Goo and S. D. Kim, Appl. Catal., B, 2008, 78, 267–274.
11. H. Hu, S. X. Cai, H. R. Li, L. Huang, L. Y. Shi and D. S. Zhang, ACS Catal., 2015, 5, 6069–6077.
12. M. F. Fu, C. T. Li, P. Lu, L. Qu, M. Y. Zhang, Y. Zhou, M. G. Yu and Y. Fang, Catal. Sci. Technol., 2014, 4, 14–25.
13. W. C. Wang, G. McCool, N. Kapur, G. Yuan, B. Shan, M. Nguyen, U. M. Graham, B. H. Davis, G. Jacobs, K. Cho and X. H. (Kelly) Hao, Science, 2012, 337, 832–835.
14. Q. Li, M. Meng, N. Tsubaki, X. G. Li, Z. Q. Li, Y. N. Xie, T. D. Hu and J. Zhang, Appl. Catal., B, 2009, 91, 406–415.
15. J. Després, M. Elsener, M. Koebel, O. Kröcher, B. Schnyder and A. Wokaun, Appl. Catal., B, 2004, 50, 73–82.
16. E. Lira, L. R. Merte, F. Behafarid, L. K. Ono, L. Zhang and B. R. Cuenya, ACS Catal., 2014, 4, 1875–1884.
17. C. Zhou, Z. J. Feng, Y. X. Zhang, L. J. Hu, R. Chen, B. Shan, H. F. Yin, W. G. Wang and A. S. Huang, RSC Adv., 2015, 5, 28054–28059.
18. Y. Yu, L. Zhong, J. Ding, W. Cai and Q. Zhong, RSC Adv., 2015, 5, 23193–23201.
19. Z. Ren, Y. B. Guo, Z. H. Zhang, C. H. Liu and P. X. Gao, J. Mater. Chem. A, 2013, 1, 9897–9906.
20. X. Y. Wang, W. Wen, J. X. Mi, X. X. Li and R. H. Wang, Appl. Catal., B, 2015, 176–177, 454–463.
21. F. Cavani, F. Trifirò and A. Vaccari, Catal. Today, 1991, 11, 173–301.
22. P. J. Sideris, U. G. Nielsen, Z. H. Gan and C. P. Grey, Science, 2008, 321, 113–117.
23. G. L. Fan, F. Li, D. G. Evans and X. Duan, Chem. Soc. Rev., 2014, 43, 7040–7066.
24. J. J. Yu, Z. Jiang, L. Zhu, Z. P. Hao and Z. P. Xu, J. Phys. Chem. B, 2006, 110, 4291–4300.
25. J. J. Yu, X. P. Wang, L. D. Li, Z. P. Hao, Z. P. Xu and G. Q. Lu, Adv. Funct. Mater., 2007, 17, 3598–3606.
26. Z. P. Wang, P. Lu, X. M. Zhang, L. G. Wang, Q. Li and Z. L. Zhang, RSC Adv., 2015, 5, 52743–52753.
27. L. G. Dou, T. Fan and H. Zhang, Catal. Sci. Technol., 2015, 5, 5153–5167.
28. J. Pérez-Ramírez, G. Mul, F. Kapteijn and J. A. Moulijn, J. Mater. Chem., 2001, 11, 821–830.
29. V. Rives, O. Prieto, A. Dubey and S. Kannan, J. Catal., 2003, 220, 161–171.
30. H. H. Yi, H. Y. Wang, X. L. Tang, P. Ning, L. L. Yu, D. He and S. Z. Zhao, Ind. Eng. Chem. Res., 2011, 50, 13273–13279.
31. E. Genty, R. Cousin, S. Capelle, C. Gennequin and S. Siffert, Eur. J. Inorg. Chem., 2012, 2802–2811.
32. S. D. Li, S. P. Mo, J. Q. Li, H. D. Liu and Y. F. Chen, RSC Adv., 2016, 6, 56874–56884.
33. S. Kannan and C. S. Swamy, Catal. Today, 1999, 53, 725–737.
34. M. Gabrovska, R. Edreva-Kardjieva, K. Tenchev, P. Tzvetkov, A. Spojakina and L. Petrov, Appl. Catal., A, 2011, 399, 242–251.
35. W. X. Tang, X. F. Wu, S. D. Li, W. H. Li and Y. F. Chen, Catal. Commun., 2014, 56, 134–138.
36. F. Morales, D. Grandjean, A. Mens, F. M. F. de Groot and B. M. Weckhuysen, J. Phys. Chem. B, 2006, 110, 8626–8639.
37. J. Y. Luo, M. Meng, X. Li, X. G. Li, Y. Q. Zha, T. D. Hu, Y. N. Xie and J. Zhang, J. Catal., 2008, 254, 310–324.
38. J. F. Marco, J. R. Gancedo, M. Gracia, J. L. Gautier, E. Ríos and F. J. Berry, J. Solid State Chem., 2000, 153, 74–81.
39. J. Preudhomme and P. Tarte, Spectrochim. Acta, Part A, 1971, 27, 1817–1835.
40. L. Obalová, K. Jirátová, F. Kovanda, K. Pacultová, Z. Lacný and Z. Mikulová, Appl. Catal., B, 2005, 60, 289–297.
41. C. Gennequin, S. Siffert, R. Cousin and A. Aboukaïs, Top. Catal., 2009, 52, 482–491.
42. B. Jongsomjit, J. Panpranot and J. G. Goodwin Jr, J. Catal., 2001, 204, 98–109.
43. D. L. Li, Y. Y. Ding, X. F. Wei, Y. H. Xiao and L. L. Jiang, Appl. Catal., A, 2015, 507, 130–138.
44. M. Adachi-Pagano, C. Forano and J. P. Besse, J. Mater. Chem., 2003, 13, 1988–1993.
45. G. Carja, H. Obata, Y. Kameshima and K. Okada, Microporous Mesoporous Mater., 2007, 98, 150–155.
46. S. J. Gregg and K. S. W. Sing, Adsorption, Surface Area and Porosity, Academic Press, New York, 1982.
47. Y. Gou, X. Liang and B. H. Chen, J. Alloys Compd., 2013, 574, 181–187.
48. C. H. Wang, X. Zhang, D. C. Zhang, C. Yao and Y. W. Ma, Electrochim. Acta, 2012, 63, 220–227.
49. K. Song, E. Cho and Y. M. Kang, ACS Catal., 2015, 5, 5116–5122.
50. Y. F. Zhao, X. D. Jia, G. B. Chen, L. Shang, G. I. N. Waterhouse, L. Z. Wu, C. H. Tung, D. O'Hare and T. R. Zhang, J. Am. Chem. Soc., 2016, 138, 6517–6524.
51. W. Cai, Q. Zhong, W. Zhao and Y. F. Bu, Appl. Catal., B, 2014, 158–159, 258–268.
52. J. H. Chen, M. Q. Shen, X. Q. Wang, G. S. Qi, J. Wang and W. Li, Appl. Catal., B, 2013, 134–135, 251–257.
53. H. Hu, S. X. Cai, H. R. Li, L. Huang, L. Y. Shi and D. S. Zhang, J. Phys. Chem. C, 2015, 119, 22924–22933.
54. B. Cui, H. Lin, Y. Z. Liu, J. B. Li, P. Sun, X. C. Zhao and C. J. Liu, J. Phys. Chem. C, 2009, 113, 14083–14087.
55. F. X. Ma, L. Yu, C. Y. Xu and X. W. (David) Lou, Energy Environ. Sci., 2016, 9, 862–866.
56. Z. Ren, Z. L. Wu, W. Q. Song, W. Xiao, Y. B. Guo, J. Ding, S. L. Suib and P. X. Gao, Appl. Catal., B, 2016, 180, 150–160.
57. G. I. Panov, K. A. Dubkov and E. V. Starokon, Catal. Today, 2006, 117, 148–155.
58. P. Arnoldy and J. A. Moulijn, J. Catal., 1985, 93, 38–54.
59. L. Guo, H. Xian, Q. F. Li, D. Chen, Y. S. Tan, J. Zhang, L. R. Zheng and X. G. Li, J. Hazard. Mater., 2013, 260, 543–551.
60. F. Prinetto, G. Ghiotti, I. Nova, L. Lietti, E. Tronconi and P. Forzatti, J. Phys. Chem. B, 2001, 105, 12732–12745.
61. N. Tang, Y. Liu, H. Q. Wang and Z. B. Wu, J. Phys. Chem. C, 2011, 115, 8214–8220.
62. M. Mihaylov and K. Hadjiivanov, Langmuir, 2002, 18, 4376–4383.
63. B. Azambre, L. Zenboury, A. Koch and J. V. Weber, J. Phys. Chem. C, 2009, 113, 13287–13299.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra23704e

---