Cobalt supported on metal-doped ceria catalysts (M = Zr, Sn and Ti) for NO oxidation

Abstract

A series of ceria oxides doped with 20 mol% of Zr, Sn and Ti were prepared by a citric sol–gel method throughout the thermal decomposition of the corresponding metallic propionates. Furthermore, cobalt oxides were loaded on the supports by a one-step sol–gel method. The results showed a solid solution when doping with Zr and Ti and SnO₂ surface segregation in the case of Sn. We demonstrated that cobalt dispersion states were influenced by the doping metals. The Ce–Co metal-oxide catalyst possessed a higher amount of finely dispersed cobalt species, more oxygen vacancies and excellent redox ability. These features may be responsible for improving the catalytic activity of Ce–Co metal-oxides for NO oxidation.

---

1. Introduction

Development of methods to control NO_x emission has been an active research area¹ because NO_x are associated with acid deposition, photochemical smog, and ozone depletion. The catalytic oxidation of NO to NO₂ is a key step for many NO_x removal techniques, such as fast selective catalytic reduction (fast SCR),² NO_x storage-reduction (NSR),³ continuously regenerating trap (CRT),⁴ and the wet scrubbing method.⁵ Therefore, many efforts have been focused on developing eligible catalysts for NO oxidation.^6–10

Ceria (CeO₂) is one of the most reactive rare-earth oxides and attracts much attention due to its unique application in catalysts.^6,8,9 The catalytic properties of ceria are considered to originate from the Ce⁴⁺/Ce³⁺ redox cycle. It is affected by various structural factors, such as composition modification, large surface area, preferential exposure of reactive facets and the oxygen vacancies concentration on its surface.^11–14 However, it is well known that sintering effect often occurs when pure CeO₂ is used at high temperature, resulting in the decrease of specific surface area and oxygen storage capacity (OSC).¹⁵ Many researches have been performed by incorporating foreign metal cations into the lattice of CeO₂ to restrain this effect. These metal cations mainly contain three categories: (i) M²⁺ (Ca²⁺,¹⁶ Mg²⁺ (ref. 17)), (ii) M³⁺ (Al³⁺, Ga³⁺, In³⁺ (ref. 18 and 19)) and (iii) M⁴⁺ (Zr⁴⁺,⁶ Sn⁴⁺,²⁰ Ti⁴⁺,²¹ Mn⁴⁺,²² Hf⁴⁺ (ref. 23)). Recently, our lab²⁴ prepared a series of Cr/Ce_xZr_1−xO₂ catalysts used for NO oxidation and found that the Zr-rich samples showed higher activity than Ce-rich sample; Dobrosz-Gómez et al.²⁵ prepared a series of Au/Ce_xZr_1−xO₂ used them in CO oxidation and the Au/Ce_0.75Zr_0.25O₂ exhibited the best catalytic performance; Ayastuy et al.²⁶ reported that Ce_0.95Sn_0.05O₂ had the highest activity in CO oxidation reaction among the Ce_xSn_1−xO₂ catalysts. According to the above-mentioned literatures, we could find that the previous works were mainly focused on investigating the catalytic property of the catalyst by modulating the ratio of Ce–M. However, there were almost few reports which focus on changing the doped metal cation of Co₃O₄/Ce_xM_1−xO₂ for a comparative study.

It was widely reported that cobalt oxide based catalysts were potential candidate for NO oxidation. Wen et al. studied series of La_1−xCe_xCoO₃ perovskite oxide catalysts for nitrogen monoxide oxidation, and reported approximately 80% conversion on La_0.8Ce_0.2CoO₃ at 300 °C.²⁷ TiO₂ and SiO₂ supported Co₃O₄ catalysts were studied for NO oxidation by Irfan et al.²⁸ They reported maximum of 69% conversion on Co₃O₄/SiO₂ at 300 °C with high space velocity condition. Recently, Kim et al. studied NO oxidation on supported cobalt oxide synthesized by conventional wet impregnation method using various supports including CeO₂, SiO₂, ZrO₂, TiO₂ and SiO₂, and reported that CeO₂ supported catalyst exhibits better activity. They reported approximately 70% conversion at about 270 °C on Co-ceria and found that surface area of CeO₂ played an important role on catalytic activity.²⁹ As it was reported in the previous literatures, the support was crucial for the dispersion of active species, however, few reports discussed the different role of cobalt species in the process of NO oxidation.

In the present work, ceria doped with Zr⁴⁺, Sn⁴⁺, Ti⁴⁺ were used as supports to prepare cobalt-based catalysts and the obtained samples have been studied systematically by means of XRD, LRS, UV-vis DRS, TEM, N₂-physisorption, H₂-TPR, XRF, XPS and NO + O₂ model reaction. The study was mainly focused on understanding the influence of dopant on the structure, texture, reduction, dispersion of cobalt species, oxygen vacancies, adsorption properties and catalytic performance of the obtained samples.

2. Experimental section

2.1. Preparation of the catalysts

The Ce_0.8M_0.2O₂ (M = Zr⁴⁺, Sn⁴⁺ and Ti⁴⁺) solid solutions (Ce–M = 4 : 1 molar ratio) were prepared by sol–gel method. The Ce, Zr, Sn and Ti sources were Ce(NO₃)₃·6H₂O, ZrOCl₂, SnCl₄ and TiCl₄. The requisite quantity of Ce and M sources with citric acid (citric acid/(Ce + M) = 2) were added into distilled water to form solution. The above two solutions (Ce and M) together with citric acid were mixed together and stirred for 2 h. After that the mixture was heated at 90 °C under stirring until it became a viscous gel and dried at 120 °C for 12 h then the obtained solid was calcined in air at 500 °C for 4 h. In addition, pure CeO₂ was prepared via the same procedure for comparison. These synthesized samples were CeO₂, Ce_0.8Zr_0.2O₂ (hereafter denoted as Ce–Zr), Ce_0.8Sn_0.2O₂ (hereafter denoted as Ce–Sn), Ce_0.8Ti_0.2O₂ (hereafter denoted as Ce–Ti) supports.

The CoO_x/S (S = CeO₂, Ce–Zr, Ce–Sn, Ce–Ti) catalysts were prepared by one-step sol–gel method. The Co source was Co(NO₃)₂·6H₂O. The Ce, M (M = Zr, Sn and Ti) and Co solution with citric acid (citric acid/(Ce + M + Co) = 2) were mixed together and stirred for 2 h. After that the mixture was heated at 90 °C under stirring until it became a viscous gel and dried at 120 °C for 12 h then the obtained solid was calcined in air at 500 °C for 4 h. These catalysts were denoted as Ce–Co, Ce–Zr–Co, Ce–Sn–Co and Ce–Ti–Co, respectively.

2.2. Sample characterization

The powder XRD patterns were recorded on a Beijing Purkinjie general instrument XD-3 X-ray diffraction using Cu-Kα radiation at 36 kV and 20 mA (2θ from 5° to 80°). The scanning speed is 8° min⁻¹ and the step value is 0.04°. The average grain sizes of the catalysts were calculated using the Scherer's formula, eqn (1):where D is the average crystallite size; K is a constant; λ is the wavelength of the X-ray radiation (Cu Kα = 0.1541 nm); β is the full width at half-maximum (FWHM) of the diffraction peak (the corrected β equaling to the measured width (β_M) minus the instrument width (β_S)); θ is the diffraction angle. After correction, the formula is as shown in the following equation:

Laser Raman spectra (LRS) were recorded on a Renishaw Invia Raman Microscope with Ar⁺ radiation (514 nm). The laser light was focused onto the samples by using a microscope equipped with a 6100 objective lens.

Diffuse reflectance spectroscopy (DRS) was carried out on a Shimadzu UV-2550 UV-vis spectrophotometer. BaSO₄ was the reference sample and the spectra were recorded in the range of 200–800 nm.

The micromorphology of the catalysts was examined on a JEOL JEM-2100 transmission electron microscope (TEM), and the sample was deposited on a copper mesh by means of dipcoating. The acceleration voltage is 200 kV.

Specific surface areas of the different catalysts were determined by N₂ adsorption–desorption measurements at −196 °C by employing the Brunauer–Emmet–Teller (BET) method (Gold App V-sorb 2800 p), and the pore volume and pore size of the samples were calculated by Barrett–Joyner–Halenda (BJH) method. The systematic error calculated for BET surface area, pore volume and pore size was given with the accuracy at ±5 m² g⁻¹, ±5 mm³ g⁻¹ and ±0.5 nm. Prior to N₂ adsorption, the sample was outgassed at 200 °C for 12 h to desorb moisture adsorbed on the surface and inside the porous network.

Hydrogen temperature programmed reduction (H₂-TPR) was performed in a quartz U-tube reactor on an automated chemisorption analyzer (Quantachrome Instruments) by the GC method. About 100 mg sample was pretreated in N₂ stream at 600 °C for 0.5 h. As the sample was cooled downed to 50 °C, switched N₂ to H₂–N₂ mixture gas (10% H₂, v/v) at a flow rate of 70 mL min⁻¹. H₂-TPR was performed by heating the sample from 50 to 700 °C, at the same time, the consumption of H₂ was detected by a thermal conductivity detector (TCD).

X-ray photoelectron spectra (XPS) was performed on a Thermo ESCALAB 250 (USA) apparatus with Al Kα X-rays (hv = 1486.6 eV) radiation operated at 150 W to investigate the surface atomic concentrations and the oxidation state distribution of the elements in the samples. The samples were compensate for charging with low-energy electron beam, and the peak of C 1s (binding energy = 284.4 eV) was used to correct for sample charging. This reference gave BE values with an accuracy ±0.1 eV. And the atomic surface ratios of the corresponding species were given with the accuracy at ±0.1%. The penetration depth of the XPS probe is 10 nm.

X-ray fluorescence spectra (XRF) were carried out on an FL3-TCSPC apparatus to determine the bulk composition of these catalysts. The X-ray tube was operated at 60 kV and 20 mA. The accurate measurement mode was adopted to determine the contents of Co, Ce, Zr, Sn and Ti elements of these catalysts.

2.3. Catalytic testing

The catalytic oxidation of NO was performed in a fixed-bed flow microreactor under atmospheric pressure. Typically, 300 mg sample (sieve fraction of 40–60 mesh) was placed in a quartz reactor (6.8 mm i.d.); the reactant gas mixture (390 ppm NO, 8% O₂, N₂ balance) was fed to the reactor with a total flow rate of 100 mL min⁻¹, corresponding to a gas hour space velocity (GHSV) of 35 400 h⁻¹. The steady-state tests were conducted isothermally every 25 °C from 200 to 400 °C and the gas products (after 90 min reaction) were analyzed by a Ecom-JZKN flue gas analyzer (Germany). The NO conversion is defined as:

NO conversion = (NO_in − NO_out)/NO_in × 100%.

3. Result and discussion

3.1. Catalytic activity of ceria-based mixed oxides (NO oxidation)

The profiles for NO conversion as a function of reaction temperature over these catalysts were present in Fig. 1a and b. For the Co-containing samples (Fig. 1a), the sequence of catalytic activity expressed as NO conversion is Ce–Co > Ce–Zr–Co > Ce–Sn–Co > Ce–Ti–Co. In contrast, the pure CeO₂ and CeO₂–MO_x (M = Zr⁴⁺, Sn⁴⁺ and Ti⁴⁺) samples were also prepared to evaluate the catalytic activity. Although the catalytic activity of all the supports is low, modified supports have relative higher activity than pure CeO₂. In general, the Co-containing samples shows higher catalytic activity than the supports, indicating Co species play an important role in the NO oxidation process.

Fig. 1 Comparing NO conversion over the supports and catalysts.

3.2. Structural characteristics (XRD, LRS, UV-vis DRS and TEM)

Fig. 2 shows the XRD patterns of these samples. It could be seen that all the diffraction peaks of these samples except Ce–Sn were characteristic of typical cubic fluorite phase of ceria (PDF-ICDD 34-0394). The characteristic lines attributed to ZrO₂ and TiO₂ were absent, the diffraction peaks of Ce–Zr and Ce–Ti shifted to high angle direction slightly and the peaks became wider compared to CeO₂, which indicated that Zr⁴⁺ and Ti⁴⁺ were successfully doped into the lattice of CeO₂ to form uniform solid solutions. As for Ce–Sn, the diffraction peak of SnO₂ was observed, showing the absence of solid solution or little solid solution forming. In contrast, the Ce–Sn–Co sample showed no SnO₂ peak and the (1 1 1) peak of CeO₂ shifted to higher angle, this phenomenon indicated the adding Co could contribute to the forming of solid solution. The grain sizes of these samples are summarized in Table 1. The grain sizes of Ce–Zr, Ce–Sn and Ce–Ti were smaller than CeO₂, suggesting that the introduction of Zr⁴⁺, Sn⁴⁺ and Ti⁴⁺ inhibited the crystal growth of the cubic phase.²¹ After introducing of cobalt, the grain sizes decreased slightly compared to their corresponding supports except for Ce–Ti–Co. This phenomenon was probably due to the incorporation of monolayer cobalt into the surface/subsurface layers of these supports.³⁰ In addition, all the Co supported catalysts showed no peaks for crystalline cobalt oxides, indicating that cobalt oxide species were highly dispersed and/or present as small clusters and/or form solid solution which were difficult to be detected due to the limitation of XRD.

Fig. 2 XRD patterns of the synthesized samples.

Table 1 Grain size, the position, FWHM and the peak area ratio of the main Raman line of the samples

Sample	Grain size (nm)	Position of Raman line (cm^−1)	FWHM of Raman line (cm^−1)	AII/AI
CeO2	16.4	464	9.6	—
Ce–Zr	7.7	470	52.3	—
Ce–Sn	15.4	464	15.5	—
Ce–Ti	5.5	432	69.9	—
Ce–Co	7.7	431	77.2	0.3113
Ce–Zr–Co	5.2	445	94.9	0.2098
Ce–Sn–Co	5.1	437	62.5	0.1327
Ce–Ti–Co	6.6	438	56.9	0.1024

---

Laser Raman spectroscopy is an intriguing tool to analyze the structural properties of oxide materials particularly nanostructured ceria-based solid solutions. The LRS of the supports are shown in Fig. 3a. The band at 464 cm⁻¹ was the characteristic of the cubic fluorite structure of CeO₂.³¹ The observed shift and broadening of the F_2g peaks in the 464 cm⁻¹ was attributed to the change in the M–O vibration which accounted for the difference in the ionic radius of the dopants.³² In our case, the observed decrease in the ceria particle size in the doped samples calculated by the Scherrer equation (Table 1) could agree with the shift and/or broadening of the F_2g peak. The Raman lines assigned to MO₂ (M = Zr⁴⁺, Sn⁴⁺ and Ti⁴⁺) were not observed. This results also confirmed that the doped cations entered the crystal lattice of CeO₂ which was in line with XRD results.

Fig. 3 LRS of the supports (a) and the catalysts (b).

For the catalysts, the Raman lines of cobalt species were absent in all of the samples. In addition, the main band of these catalysts were red or blue shift in contrast to their support. This phenomenon could be explained be two reasons. One is that the existence of an interaction between support and the active species.³³ The other reason may be the incorporation of monolayer cobalt into the surface/subsurface layers of these supports, which was in consistent with XRD results.

Apart from the main peak of CeO₂, a new band II appeared at ca. 570 to 630 cm⁻¹, ascribing to the intrinsic oxygen vacancies.³⁴ They were very essential for absorption/dissociation of oxygen molecule during the catalytic oxidation reactions, which in turn could affect the catalytic activity of the catalysts.³⁵ Since the band II was related to oxygen vacancies, and the band I was the main band of these samples, the area ratio of A_II/A_I could reflect the concentration of oxygen vacancies. We could see from Table 1 that the Ce–Co had higher amount of oxygen vacancies.

UV-vis DRS technique is effectively used for the study of metal oxides to obtain information about the charge transfer transitions. Ceria exhibited strong absorption below 400 nm caused by charge transfer bands. DR spectra of these samples are presented in Fig. 4. Three characteristic bands of ceria corresponding to the O²⁻ → Ce³⁺ and O²⁻ → Ce⁴⁺ charge transfer bands and inter-band transitions in the wavelength range 250–340 nm were observed for all samples.^36,37 The absorption band observed at ca. 470 nm can be attributed to the ¹A_1g → ¹T_1g forbidden d–d transitions of Co³⁺ in its octahedral sites.³⁸ The peak observed at ca. 750 nm corresponded to Co²⁺ ions in its tetrahedral geometry.³⁹ It was confirmed that the evidence of the Co³⁺ and Co²⁺ species from DRS study was in accordance with H₂-TPR and XPS results.

Fig. 4 UV-vis spectra of these catalysts.

TEM was used to investigate the morphology and microstructure of the sample. Fig. 5 shows the TEM and HR-TEM imagines of these catalysts. From TEM imagines, we could clearly see that the particles of all the catalysts were irregular. It could be seen from the HRTEM imagines that these irregular imagines were comprised of many small particles with a crystallite size of 3–5 nm, and there were clear voids with diameter 3–5 nm among the small particles, revealing the mesostructure of the CeO₂ particles. Furthermore, the absence of Co-related particles' lattice in HRTEM images suggested a good dispersion of CoO_x crystallites. From HRTEM, the interplanar distance of lattice fringes was 0.312 nm corresponding to (1 1 1) crystallographic plane of CeO₂.⁴⁰ In contrast, the fringes of Ce–M–Co (M = Zr⁴⁺, Sn⁴⁺ and Ti⁴⁺) were little lower than 0.312 nm, indicating the forming of Ce–M solid solution. The results were consistent with XRD and Raman results.

Fig. 5 TEM and HRTEM patterns of the samples. (A) Ce–Co; (B) Ce–Zr–Co; (C) Ce–Sn–Co; (D) Ce–Ti–Co.

3.3. Textural characterization (N₂-physisorption)

The N₂ adsorption–desorption isothermal plots and the corresponding BJH pore size distribution curves of these samples were displayed in Fig. 6. It could be seen from Fig. 6a that all these samples were of classical type IV as defined by IUPAC and exhibited hysteresis loops of type H3, which indicated that these samples contained mesopores (2–50 nm) with narrow slit-like shapes.⁴¹ The pore size distribution curves suggested that all the samples had defined mesoporosities. The pore size at the maximum probability of the catalysts was 6.9, 11.3, 10.8 and 13.8 nm for Ce–Co, Ce–Zr–Co, Ce–Sn–Co and Ce–Ti–Co, respectively. The BJH pore size distribution plots of the various catalysts possessed nonuniform mesopore size distribution.

Fig. 6 N₂ adsorption–desorption isotherms (a) and pore size distributions (b) of the samples.

The BET surface area, pore volume and pore size of these catalysts are summarized in Table 2. Compared with Ce–Co, when the foreign metal cations were doped into the lattice of CeO₂ (formed Ce–Zr, Ce–Sn, and Ce–Ti solid solutions), both the specific surface area and pore volume increased. These changes may be related to the grain size of the samples to some extent. In other words, the incorporation of Zr⁴⁺, Sn⁴⁺ and Ti⁴⁺ into the lattice of CeO₂ could improve the texture property effectively. In order to explore the thermal stability, all of the four supports were calcined in air at 800 °C for 4 h in flowing air, and then, BET measurements were carried out, and the specific surface areas were summarized in Table S1.† Compared with pure CeO₂, the degree of sintering (i.e. S_BET(500) − S_BET(800)/S_BET(500)) was smaller when Zr⁴⁺, Sn⁴⁺ and Ti⁴⁺ incorporated into the lattice of CeO₂. It indicated that the ability of anti-sintering of CeO₂ was enhanced with the introduction of these foreign metal cations.

Table 2 The surface area and pore structure parameters of the samples

Sample	Ce–Co	Ce–Zr–Co	Ce–Sn–Co	Ce–Ti–Co
BET (m^2 g^−1)	21	46	65	39
Pore size (nm)	6.9	11.3	10.8	13.8
Pore volume (mm^3 g^−1)	66	192	188	257

---

3.4. Reduction properties of the catalysts (H₂-TPR)

H₂-TPR tests were performed to study the reducibility of these catalysts. Fig. 7 shows that the shape of these TPR profiles is very different, indicating the support has an important effect on the dispersion of Co₃O₄. With respect to the reduction steps of Co₃O₄, they were rather controversial in literature. Arnoldy and Moulijn⁴² observed a single step for the reduction of Co₃O₄, while many others^43–45 reported that the reduction of Co₃O₄ was a two-step process, involving in the reduction to CoO. The results here clearly showed that the reduction behaviors of Co₃O₄ strongly depended on preparation, catalyst composition and dispersion behavior, consistent with the finding of Spadaro et al.⁴⁶ Large particles of Co₃O₄ were often reduced directly to metallic Co in a single step, while fine particles of Co₃O₄ that interacted with ceria seemed to get reduced in two steps. The first step was generally promoted by component interaction between them, possibly ascribed to the lengthening in the Co–O bond, while the second step was perhaps delayed by the stabilization effect of ceria to cobalt ions, similar assignment has ever reported by Martínez-Arias et al. for CuO/CeO₂ system.⁴⁷ On the basis, the following assignments were made for the reduction of cobalt species in respect of the temperature ranges:

Fig. 7 H₂-TPR profiles of these catalysts.

Range I (below 320 °C): reduction of the surface adsorbed oxygen species (indicated as peak α for Ce–Co catalyst);

Range II (320–420 °C): reduction of Co³⁺ at the interface between Co₃O₄ and CeO₂ to Co²⁺ (peak β);

Range III (420–500 °C): reduction of independent Co₃O₄ that weakly interacts with CeO₂ directly to Co (peak γ);

Range IV (500–700 °C): reduction of Co²⁺ interacting with CeO₂ to Co (peak θ).

Since the peak in Range II were related to the finely Co₃O₄ species and Range III were independent or bulk-like Co₃O₄ phase, we could infer the dispersion state of Co₃O₄ phase from the intensity of this peak. As shown the dotted box in Fig. 7, the peak area ratio of A_β/A_γ was 1.8, 0.5, 0.4 and 0.3, referring to Ce–Co, Ce–Zr–Co, Ce–Sn–Co and Ce–Ti–Co, respectively. The reduction temperature of peak β and peak γ shifted to higher temperature in contrast with Ce–Co. Thus it could be seen that the support had significant effect on the reduction properties and component dispersion. Combined with the activities performance, we could infer that highly dispersed Co₃O₄ possessed better catalytic activity. Similar conclusions were reported by Luo et al.,¹³ they found that the finely dispersed CuO species had higher activity compared to bulk CuO species.

3.5. Bulk and surface analysis of the catalysts (XRF and XPS)

The bulk composition of these catalysts were characterized by XRF. As shown in Table 3, the chemical composition of the catalysts matched well the intended one, evidencing that the thermal decomposition of the propionates method allowed obtaining materials with a controlled stoichiometry. XPS was performed in order to illuminate the surface composition and the chemical state of the elements existing in the catalysts. The deconvoluted photoelectron spectra of the Ce 3d, Co 2p, O 1s, Zr 3d, Sn 3d and Ti 2p are shown in Fig. 8. Ce 3d core level spectra (Fig. 8a) for all the samples were deconvoluted into eight contributions. The spin–orbit splitting of Ce 3d_5/2 and Ce 3d_3/2 was 18.5 eV for all the samples, which was in good agreement with literature.⁴⁸ The u′′′ satellite peak at about 916.5 eV was the fingerprint of Ce⁴⁺ state, and its high intensity and area suggested that the main part of the ceria was in Ce⁴⁺ oxidation state. After deconvolution the appearance of bands labeled u′ and v′ were typical for Ce³⁺ ions, which suggested that both oxidation states Ce⁴⁺ and Ce³⁺ coexist in the surfaces of samples.⁴⁸ The enhancement of unhomogeneity of the Ce and M (M = Zr, Sn and Ti) atoms could improve the valence change of the Ce (Ce⁴⁺ → Ce³⁺).⁶ Another reason was probably due to the substitution of Ce⁴⁺ by M (M = Zr⁴⁺, Sn⁴⁺ and Ti⁴⁺) ions partly in the calcination process. The spontaneous transformation of Ce⁴⁺ (r_Ce⁴⁺ = 0.97 Å) ions to the larger Ce³⁺ (r_Ce³⁺ = 1.10 Å) could compensate for this lattice contraction. The relative intensities of u′ to v′ to the whole eight bands were 12.88%, 20.35%, 23.58% and 30.42%, referring to Ce–Co, Ce–Zr–Co, Ce–Sn–Co and Ce–Ti–Co, respectively, which was shown in Table 3. Obviously, the surface ratios of Ce³⁺ to (Ce³⁺ + Ce⁴⁺) on the catalysts were higher after introduction M ions, indicating the introduction of M could cause charge imbalance and forming the Ce–M solid solution. It was obvious that the binding energy of Ce shifted to the high field after introduction of other cations. The reason could be explained by the density of electron cloud decreased as the addition of strong electrophilic M cations (Electronegativity: Zr = 1.33, Sn = 1.96 and Ti = 1.54) than Ce (1.10).

Table 3 The bulk and surface composition of these catalysts (M = Zr, Sn, Ti)

Sample	Atomic ratio (%)
	XRF (bulk)		XPS (surface)
	Ce/M	Co/(Ce + M + Co)	Ce/M	Co/(Ce + M + Co)	Co^3+/(Co^3+ + Co^2+)	Ce^3+/(Ce^3+ + Ce^4+)	Oβ/(Oα + Oβ)
Ce–Co	—	15.08 (15.00)	—	10.96	61.41	12.88	27.21
Ce–Zr–Co	4.08 (4.00)	14.98 (15.00)	4.42	16.99	55.20	20.35	47.26
Ce–Sn–Co	4.02 (4.00)	15.03 (15.00)	4.62	26.28	53.35	23.58	70.09
Ce–Ti–Co	4.01 (4.00)	15.01 (15.00)	4.51	15.33	49.28	30.42	71.12

---

Fig. 8 XPS spectra of (Ce 3d, Co 2p, O 1s, Zr 3d, Sn 3d and Ti 2p) of the catalysts.

The Co 2p spectra are displayed in Fig. 8b. A set of Co 2p peaks located at ca. 780 eV and 795 eV corresponding to Co 2p_3/2 and Co 2p_1/2 respectively. Each peak was the summation of signals from Co²⁺ and Co³⁺. In Co₃O₄, Co²⁺ has higher binding energy than Co³⁺.⁴⁹ From the deconvoluted spectra in Fig. 8b, it was concluded that the peaks at 779 eV and 794 eV with a small satellite signal around 788 eV correspond to Co³⁺. The peaks at 781 eV and 797 eV were characteristic of Co²⁺.⁵⁰ Detailed calculation from the deconvoluted peaks reveals that the intensity ratio of Co³⁺ and Co²⁺ was different in different catalysts. The surface atomic ratio of (Co³⁺)_xps to (Co³⁺ + Co²⁺)_xps was in following order Ce–Co > Ce–Zr–Co > Ce–Sn–Co > Ce–Ti–Co (Table 3). Since this sequence was consistent with the NO conversion performance, the Co³⁺ on the surface therefore acted as the active sites and its quantity determined the capability of catalytic NO conversion into NO₂. Combined with H₂-TPR results, we could concluded that Co³⁺ was more favorable to form in the highly dispersed Co₃O₄ species.

O 1s spectra were mainly composed of two components. The peak of O 1s named as O_α, with a binding energy of 528–530 eV, contributed to the lattice oxygen in the metal oxides.³⁶ The high binding 531–532 eV energy was attributed to the surface chemisorbed oxygen (O_β).³⁶ We could find that the O 1s ionization features for all of the catalysts are very similar (Fig. 8c). Obviously, the ratios of chemisorbed oxygen to the whole type of oxygen were higher compared to Ce–Co (Table 3), which indicated that the doping of other cations was beneficial for the formation of chemisorbed oxygen.

In addition, the narrow spectra of Zr 3d, Sn 3d and Ti 2p are shown in Fig. 8d. The binding energy of Zr 3d_5/2 (181.8 eV) in the Ce–Zr–Co was consistent with ZrO₂ (181.8 eV), indicating that Zr was mainly in a +4 oxidation state.⁵¹ As for the XPS spectra of Sn 3d, the most of Sn was mainly in a +4 oxidation state because the binding energy of Sn 3d_5/2 (486.2 eV) in Ce–Sn–Co was uniform in SnO₂ (486.2 eV).⁵² For Ce–Ti–Co, the binding energy of Ti 2p_3/2 was 458.0 eV, which was lower that in TiO₂ (458.3 eV) but higher that in TiO (454.8 eV), indicating that Ti was in a +4 oxidation state and a small quantity on a +3 oxidation state.⁵³

3.6. Relationship between catalytic performance and catalysts properties

It was interesting that different supports could lead to different catalytic performance. In order to understand this question clearly, we attempted to correlate the activity with the above-mentioned characterizations, and the relationship between NO conversion (at 300 °C) and textural/structural/reducible properties was plotted in Fig. 9. As far as we know, for some important catalytic reactions, the increase of the specific surface area of support was beneficial to the dispersion of active species, which could lead the activity to increase; the activity enhanced with decreasing the grain size of catalysts. However, it could be noted from Fig. 9a, the activity couldn't relate to the BET surface area and the grain size very well, we took the dispersion states of Co₃O₄ and the oxygen vacancies into account. In our previous work,³⁵ we demonstrated that oxygen vacancies were important to the activation of oxygen. Oxygen turned into superoxide ions on oxygen vacancies, superoxide ions could oxide NO for its' strong oxidation ability.⁵⁴ Furthermore, the dispersion states of the oxides was also a vital factor, Luo et al.¹³ found the finely dispersed CuO species had the highest catalytic activity among the finely dispersed CuO, the bulk CuO and the Cu²⁺ in the CeO₂ lattice. Highly dispersed Co₃O₄ species provided more NO adsorption sites and oxygen vacancy were responsible for O₂ activation. The NO-TPD profiles are shown in Fig. S1.† The results demonstrated the highly dispersed Co₃O₄ species were favorable for the NO adsorption. In H₂-TPR results, the peak β was ascribed to the reduction of Co³⁺ to Co²⁺, the lower reduction temperature may lead to increase the activity. Interestingly, these properties are well in agreement with the activity (Fig. 9b).

Fig. 9 The relationship between the catalytic activity of NO oxidation (at 300 °C) and (a) textural and (b) structural/reducible properties.

4. Conclusion

This work focused on different doped metal cations on the structure, texture, reduction, dispersion states of cobalt species, oxygen vacancies, adsorption properties and catalytic performance of the Co₃O₄/Ce_0.8M_0.2O₂ (M = Zr⁴⁺, Sn⁴⁺ and Ti⁴⁺) samples for NO oxidation. The catalytic performance results indicated that Co was the active sites. Combined with the above analysis, the catalytic activity of Ce–Co oxides was due to the higher amount of highly dispersed Co₃O₄, the larger amount of oxygen vacancies and the excellent redox ability.

Acknowledgements

This work was financially supported by the Assembly Foundation of The Industry and Information Ministry of the People's Republic of China 2012 (543), the National Natural Science Foundation of China (51408309), Science and Technology Support Program of Jiangsu Province (BE2014713), Natural Science Foundation of Jiangsu Province (BK20140777), Scientific Research Project of Environmental Protection Department of Jiangsu Province (2013003), Industry-Academia Cooperation Innovation Fund Projects of Jiangsu Province (BY2014004-10), Science and technology project of Nanjing (201306012).

References

1. Z. B. Wu, R. B. Jin, Y. Liu and H. Q. Wang, Catal. Commun., 2008, 9, 2217.
2. G. Madia, M. Koebel, M. Elsener and A. Wokaun, Ind. Eng. Chem. Res., 2002, 41, 3512.
3. W. S. Epling, L. E. Campbell, A. Yezerets, N. W. Currier and J. E. Parks II, Catal. Rev.: Sci. Eng., 2004, 46, 163.
4. B. J. Cooper, H. J. Jung and J. E. Thoss, US Patent 4, 902, 487, 1990.
5. H. Takeuchi, M. Ando and N. Kizawa, Ind. Eng. Chem. Process Des. Dev., 1977, 16, 303.
6. W. Cai, Q. Zhong, W. Zhao and Y. F. Bu, Appl. Catal., B, 2014, 158–159, 258.
7. J. Giménez-Maňogil, A. Bueno-López and A. García-García, Appl. Catal., B, 2014, 152–153, 99.
8. I. Atribak, A. Bueno-López, A. García-García and B. Azambre, Phys. Chem. Chem. Phys., 2010, 12, 13770.
9. S. K. Megarajana, S. Rayalu, Y. Teraoka and N. Labhsetwar, J. Mol. Catal. A: Chem., 2014, 385, 112.
10. Z. Ren, Y. B. Guo, Z. H. Zhang, C. H. Liu and P. X. Gao, J. Mater. Chem. A, 2013, 1, 9897.
11. M. Sanchez-Dominguez, L. F. Liotta, G. D. Carlo, G. Pantaleo, A. M. Venezia, C. Solans and M. Boutonnet, Catal. Today, 2010, 158, 35.
12. L. Qi, C. J. Tang, L. Zhang, X. J. Yao, Y. Cao, L. F. Liu, F. Gao, L. Dong and Y. Chen, Appl. Catal., B, 2012, 127, 234.
13. M. F. Luo, Y. P. Song, J. Q. Lu, X. Y. Wang and Z. Y. Pu, J. Phys. Chem. C, 2007, 111, 12686.
14. L. Meng, J. J. Lin, Z. Y. Pu, L. F. Luo, A. P. Jia, W. X. Huang, M. F. Luo and J. Q. Lu, Appl. Catal., B, 2012, 119–120, 117.
15. J. Z. Shyu, W. H. Weber and H. S. Gandhi, J. Phys. Chem., 1988, 2, 4964.
16. M. Yan, T. Mori, J. Zou, F. Ye, D. R. Ou and J. Drennan, Acta Mater., 2009, 57, 722.
17. F. Chen, J. J. Zhu, Y. Y. Zhan, X. Y. Lin, G. H. Cai, K. M. Wei and Q. Zheng, Appl. Catal., A, 2009, 363, 208.
18. T. Yuzhakova, V. Rakić, C. Guimon and A. Auroux, Chem. Mater., 2007, 19, 2970.
19. B. Bonnetot, V. Rakic, T. Yuzhakova, C. Guimon and A. Auroux, Chem. Mater., 2008, 20, 1585.
20. S. Somacescu, P. Osiceanu, J. M. C. Morenoa, L. Navarrete and J. M. Serra, Microporous Mesoporous Mater., 2013, 179, 78.
21. M. F. Luo, J. Chen, L. S. Chen, J. Q. Lu, Z. C. Feng and C. Li, Chem. Mater., 2001, 13, 197.
22. T. Y. Li, S. J. Chiang, B. J. Liaw and Y. Z. Chen, Appl. Catal., B, 2011, 103, 143.
23. W. T. Chen, K. B. Chen, M. F. Wang, S. F. Weng, C. S. Lee and M. C. Lin, Chem. Commun., 2010, 46, 3286.
24. W. Cai, Q. Zhong, S. L. Zhang and W. Zhao, Chem. Eng. J., 2014, 36, 223.
25. I. Dobrosz-Gómez, I. Kocemba and J. M. Rynkowski, Appl. Catal., B, 2008, 83, 240.
26. J. L. Ayastuy, A. Iglesias-González and M. A. Gutiérrez-Ortiz, Chem. Eng. J., 2014, 244, 372.
27. Y. Wen, C. Zhang, H. He, Y. Yu and Y. Teraoka, Catal. Today, 2007, 126, 400.
28. M. F. Irfan, J. H. Goo and S. D. Kim, Appl. Catal., B, 2008, 78, 267.
29. D. S. Kim, Y. H. Kim, J. E. Yie and E. D. Park, Korean J. Chem. Eng., 2010, 27, 49.
30. L. J. Liu, Z. J. Yao, B. Liu and L. Dong, J. Catal., 2010, 275, 45.
31. A. Martínez-Arias, M. Fernández-García, L. N. Salamanca, R. X. Valenzuela, J. C. Conesa and J. Soria, J. Phys. Chem. B, 2000, 104, 4038.
32. Q. Q. Jiang, G. L. Zhou, Z. X. Jiang and C. Li, Sol. Energy, 2014, 99, 55.
33. B. Bonnetot, V. Rakic, T. Yuzhakova, C. Guimon and A. Auroux, Chem. Mater., 2008, 20, 1585.
34. J. R. McBride, K. C. Hass, B. D. Poindexter and W. H. Weber, J. Appl. Phys., 1994, 76, 2435.
35. S. L. Zhang, H. Y. Li and Q. Zhong, Appl. Catal., A, 2012, 435–436, 156.
36. R. Si, Y. W. Zhang, S. J. Li, B. X. Lin and C. H. Yan, J. Phys. Chem. B, 2004, 108, 12481.
37. B. M. Reddy, P. Bharali and P. Saikia, J. Phys. Chem. C, 2008, 112, 11729.
38. Á. Kukovecz, Z. Kónya, D. Mönter, W. Reschetilowski and I. Kiricsi, J. Mol. Struct., 2001, 563–564, 403.
39. L. Poul, N. Jouini and F. Fiévet, Chem. Mater., 2000, 12, 3123.
40. L. J. Liu, Y. Cao, W. J. Sun, Z. J. Yao, B. Liu, F. Gao and L. Dong, Catal. Today, 2011, 175, 48.
41. 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.
42. P. Arnoldy and J. A. Moulijn, J. Catal., 1985, 93, 38.
43. M. Kang, M. W. Song and C. H. Lee, Appl. Catal., A, 2003, 251, 143.
44. C. W. Tang, C. C. Kuo, M. C. Kuo, C. B. Wang and S. H. Chien, Appl. Catal., A, 2006, 309, 37.
45. L. F. Liotta, G. Di Carlo, G. Pantaleo, A. M. Venezia and G. Deganello, Appl. Catal., B, 2006, 66, 217.
46. L. Spadaro, F. Arena, M. L. Granados, M. Ojeda, J. L. G. Fierro and F. Frusteri, J. Catal., 2005, 234, 451.
47. A. Martínez-Arias, M. Fernández-García, O. Gálvez, J. M. Coronado, J. A. Anderson, J. C. Conesa, J. Soria and G. Munuera, J. Catal., 2000, 195, 207.
48. E. Béche, P. Charvin, D. peranau, S. Abanades and G. Flamant, Surf. Interface Anal., 2008, 40, 264.
49. M. Yung, E. Holmgreen and U. Ozkan, Catal. Lett., 2007, 118, 180.
50. Y. Ding, L. Zhu, A. Huang, X. Zhao, X. Zhang and H. Tang, Catal. Sci. Technol., 2012, 2, 1977.
51. S. Velu, K. Suzuki, C. S. Gopinath, H. Yoshida and T. Hattori, Phys. Chem. Chem. Phys., 2002, 4, 1990.
52. D. Shuttleworth, J. Phys. Chem., 1980, 84, 1629.
53. B. M. Reddy, P. Lakshmanan and A. Khan, J. Phys. Chem. B, 2004, 108, 16855.
54. Z. Zhao, X. G. Yang and Y. Wu, Appl. Catal., B, 1996, 8, 281.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra15439h

---