Introduction

Volatile organic compounds (VOCs) emitted from the exhaust gas emission of chemical plants and motor vehicles can easily cause environmental pollution [1,2,3]. Therefore, the effective removal of toluene has been urgently demanded. Researchers studied a variety of technologies to remove VOCs, such as heat incineration, biodegradation, adsorption and catalytic oxidation [4]. Among these methods, catalytic oxidation has been widely used for its superior removal efficiency of VOCs [5,6,7]. Noble metal-based catalysts exhibit the superior efficiency in catalytic performance [8]. However, its exorbitant cost, inferior stability and susceptibility to poisoning restrain its further application [9]. As substitutes, some transition metals and rare earth metals have been generally studied in catalytic oxidation of VOCs [10, 11].

Recently, the CeO2-based materials have exhibited great potential in the catalytic oxidation of VOCs due to the excellent oxygen storage capacity (OSC) and abundant oxygen vacancies. Specifically, the reversible Ce4+/Ce3+ redox cycles promote the mobility of reactive oxygen species. Ce-M-Ox (M is the transition metal) mixed oxide material catalyst exhibits superior VOCs oxidation performance compared to the corresponding single metal oxide catalyst [12,13,14,15]. Venkataswamy et al. [16] claimed that Ce0.7Mn0.3O2–δ catalysts presented more surface oxygen defects or hydroxyl-like groups than pure CeO2, which could contribute to the improvement of catalytic activity. Li et al. [17] revealed that the roles of adsorption oxygen and lattice oxygen species over CuO/Ce0.7Mn0.3 were ambiguous in the catalytic removal of benzene. It was reported that the doping of ZrOx could enhance  the thermal stability of catalyst due to higher binding energy of Zr and O [18, 19]. Hu et al. [20] proved that the introduction of NiOx could optimize the reducibility for the total oxidation of propane. In spite of this, it was still not clear which transition metal could better improve the catalytic oxidation capacity of the catalyst. We found that the addition of different transition metals can promote the catalytic performance of Ce-based catalysts in VOCs oxidation. However, different metals exhibited different effects on their oxidation properties and chemical valence states. It is still necessary to discuss the effect of metal doping on CeO2 catalysts in a specific catalyst environment. Therefore, we selected several common metals to modify the performance of the catalyst, in order to further develop and enrich the theory of catalytic oxidation of VOCs.

In this work, the CeO2 and CeO2–MOx (M: Mn, Zr and Ni) catalysts were synthesized by co-precipitation method. In order to investigate the influence of different transition metals on the catalytic performance of CeO2, the CeO2–MOx (M: Mn, Zr and Ni) catalysts have been characterized by powder X-ray diffraction patterns (XRD), N2 adsorption–desorption, X-ray photoelectron spectroscopy (XPS), Raman spectra, H2 temperature-programed reduction (H2-TPR) and oxygen temperature-programmed desorption (O2-TPD) techniques.

Experimental

Chemicals and materials

All the reagents involved in catalyst synthesis were A.R. grade and were used directly. Concentrated NH3·H2O solution, cerium(III) nitrate hexahydrate (Ce(NO3)3·6H2O), manganese(II) nitrate (Mn(NO3)2), nickel(II) nitrate hexahydrate (Ni(NO3)2·6H2O), zirconium(IV) nitrate pentahydrate (Zr(NO3)4·5H2O) and toluene were supplied from Sinopharm Chemical Reagent Co., Ltd. (Shenyang, China).

Catalyst preparation

The procedures for synthesis of CeO2–MOx (M: Mn, Zr and Ni) by co-precipitation were as follows: the desired amount of M nitrates was dissolved in 100 mL of deionized water in beakers under stirring condition. After complete dissolution, the corresponding amount of Ce(NO3)3·6H2O was dissolved into the above solution to obtain the mixed solution (MOx/CeO2 = 25 wt%). The ammonia solution was added drop by drop to the aforementioned mixed solution with vigorous stirring, until the pH value reached 10. Then the beakers containing the precipitate were transferred to the heat-collecting thermostatic magnetic stirrer and stirred at 80  °C for 4 h. After resting 12 h at room temperature, the precursor precipitate was screened out from the supernatant by the suction filter device and washed several times with distilled water until the pH reached 7. The obtained solid precursor was dried at 100 °C for 12 h and then calcined in a muffle furnace at 550 °C for 5 h in air (the heating rate of 3 °C min−1). The obtained catalysts were denoted as CeO2–MnOx, CeO2-ZrOx and CeO2–NiOx. For comparison purpose, pure CeO2 sample was also prepared under similar conditions using cerium nitrate as the precursor and represented as CeO2.

Measurement of catalytic activity

The catalytic combustion of toluene was performed on the fixed bed reactor (i.d. = 8.0 mm) to measure catalytic performance using 0.1 mL of catalysts (40 − 60 mesh), in the temperature range 200–360 °C. The simulated gas atmosphere was 500 ppm of toluene diluted in 20 V % O2/N2 mixture (100 mL/min) resulting in gas hour space velocity (GHSV) of 60,000 h−1. In order to stabilize the system, the gaseous mixture was introduced over the catalyst at room temperature before heating and the on-line gas chromatography (GC, FULI 9790 II) equipped with FID detectors was adopted to analyze the composition of the outlet gas.

The following formulas were used to calculate the conversion of toluene (Ct) and CO2 generation rate \(\left( {C_{{CO_{2} }} } \right)\):

The conversion of toluene (Ct):

$$ C_{t} \left( {\text{\% }} \right) = \frac{{\left( {toluene} \right)_{in} - \left( {toluene} \right)_{out} }}{{\left( {toluene} \right)_{in} }} \times 100 $$

where \(\left( {toluene} \right)_{in}\) and \(\left( {toluene} \right)_{out}\) denotes the inlet and outlet concentrations of toluene.

The CO2 generation rate \(\left( {C_{{CO_{2} }} } \right)\):

$$C_{{co_{2} }} (\% ) = \frac{{(Cco_{2} )_{out}}}{{7 \times (toluene)_{in} }} \times 100$$

where \(\left( {C_{{CO_{2} }} } \right)_{out}\) was the CO2 outlet concentration.

Catalyst characterization

Textural properties of the samples were derived by nitrogen adsorption–desorption at − 196 °C using an analyzer of SSA-6000 from Beijing Builder Electronic Technology Company. Before the measurement, the catalysts were degassed at 250 °C for 3 h. The specific surface area (SBET) and porous volume of each catalyst were determined using the Brunauer–Emmett–Teller (BET) and the Barrett-Joyner-Halenda (BJH) methods.

Powder X-ray diffraction (XRD) patterns of all samples were collected on Brucker diffractometer (D8 Advance) and it adopted Cu Kα radiation at the 2θ range 10–80° with a scanning step size of 0.02° and scanning speed of 1.5 s.

Raman spectra were operated on a Raman microscope (Renishaw-2000) at a resolution of 2 cm−1 plus with an argon ion laser of 514.5 nm wavelength as exciting source.

Chemical state of surface elements was recorded by an ESCALAB 250 X-ray photoelectron spectroscopy (XPS) with Al Kα X-ray radiation for the X-ray source (Thermo Fisher Scientific, USA). C1 s core excitation at 284.6 eV was adopted to calibrate the electron binding energy (BE) scale. The operation process was as follows: First, the sample sheet was pasted on the conductive adhesive of the sample tray; the tray was then vacuumed in a preparation chamber for 12 h (The instrument was filled with nitrogen before operation); finally, the sample in the preparation chamber was moved in situ to the analysis chamber (vacuum state) for scanning. No high temperature treatment during the whole process. Peak fitting was processed with XPSPEAK. The fitting procedure was allowed determining the peak position, height and width.

The temperature programmed reduction of hydrogen (H2-TPR) was conducted on a chemisorption analyzer PCA-1200 (Beijing Builder Electronic Technology Company) with a quartz U-shaped tubular reactor. A 20 mg of each catalyst was packed into the reactor and operated with 30 mL/min of high purity nitrogen at 400 °C for 40 min to purify the sample, and then cooled to 100 °C. The temperature was increased to 900 °C (10 °C/min) with gas flowing of 5 V % H2/Ar (30 mL/min).

On the same chemisorption analyzer, the experiments of oxygen temperature programmed desorption (O2-TPD) detected the oxygen mobility of catalysts. 200 mg samples were degassed with helium (30 mL/min) at 200 °C for 1 h in the quartz microreactor. After cooling down to 50 °C, 20 V % O2/He of 30 mL/min mixture gas was introduced for 1 h. Subsequently, the catalysts were heated at a rate of 10 °C min−1 from 50 to 900 °C in a helium atmosphere.

Results and discussion

Evaluation of catalytic oxidation

The catalytic oxidation performances of the samples in the toluene catalytic combustion are displayed in Fig. 1. As shown in Fig. 1a, all the catalysts exhibited an upward trend as the reaction temperature increased. Pure CeO2 exhibited the T90 (the temperature at which the Ct reached 90%) value at 322 °C. After doping transition metal oxide, the sequence of T90 over the samples was as follows: CeO2–MnOx (261 °C) < CeO2–ZrOx (309 °C) < CeO2–NiOx (316 °C). Hence, the catalytic activity for toluene over CeO2-based catalysts was improved after the addition of transition metal, especially for CeO2–MnOx catalyst. Moreover, the CeO2 catalysts displayed unsatisfactory CO2 selectivity at high temperature in comparison with CeO2–MOx (M: Mn, Zr and Ni), especially for CeO2–MnOx, suggesting that the doping of transition metal oxide could improve the CO2 selectivity at high temperatures. In brief, CeO2–MnOx exhibited the best CO2 selectivity and catalytic performance of toluene, and the value of CO2 yield could reach 100% at 280 °C and above 90% toluene conversion was obtained at 261 °C.

Fig. 1
figure 1

Catalytic performances for toluene destruction over a CeO2, b CeO2–MnOx, c CeO2-ZrOx and d CeO2–NiOx catalysts. Reaction conditions: catalysts volume = 0.1 mL, toluene concentration = 500 ppm, GHSV = 60,000 h−1: (a) conversion of toluene and (b) yield of CO2

Full size image

N2 adsorption–desorption

The N2 adsorption–desorption isotherms of CeO2 and CeO2–MOx (M: Mn, Zr and Ni) catalysts with the BJH pore-size distributions are presented in Fig. 2a, b. As shown in Fig. 2a, all samples displayed the typical type IV isotherms with H3-type hysteresis loops, which indicated the existence of mesoporous structure [5, 21]. Furthermore, the hysteresis loop of CeO2 could be observed at the P/P0 value of 0.8. Compared to CeO2, the hysteresis loop of CeO2–MOx shifted to lower values, which was representative of more mesopores in the samples [18, 22]. It was reported that the higher SBET and pore volume were believed to weaken transfer resistance and optimize the adsorption capacity of VOCs [19]. Thereby, the CeO2–MnOx catalyst showed the excellent catalytic performance compared with CeO2. Fig. 2b shows the pore-size distributions of all samples. Obviously, each sample shows the distribution range from 5 to 50 nm, which was the feature of typical mesoporous material. The textural properties of the prepared catalysts are summarized in Table 1. The SBET of CeO2–MOx was larger than that of CeO2 (49 m2g−1). CeO2- ZrOx exhibited the largest SBET (72 m2g−1), followed by CeO2–MnOx with the SBET of 63 m2g−1, and the CeO2–NiOx catalyst possessed the SBET of 50 m2g−1. Obviously, the structural and surface properties of the prepared catalysts were directly optimized by the type of doped metals, which was benefited to the removal activity of toluene at high temperatures.

Fig. 2
figure 2

(a) N2 adsorption–desorption isotherms and (b) pore-size distributions of the four catalysts (a CeO2, b CeO2–MnOx, c CeO2–ZrOx and d CeO2–NiOx)

Full size image
Table 1 BET surface area, pore volume and Raman results of CeO2 and CeO2–MOx (M: Mn, Zr and Ni) catalysts
Full size table

XRD patterns

Fig. 3 presents the XRD patterns of CeO2 and CeO2–MOx (M: Mn, Zr and Ni) samples. From Fig. 3a, the feature diffraction peaks of cubic CeO2 (PDF# 43-1002) were detected on the samples, and no other diffraction peaks were observed, indicating that the added M species existed the high dispersed state or amorphous metal species [17, 20, 23, 24]. Additionally, the diffraction peaks of CeO2–MOx were slightly shifted to high angle side (shown in Fig. 3b) compared to pure CeO2 due to the incorporation of M ions into the ceria lattice, resulting in the formation of Ce-M solid solution, which contributed to improving the oxygen storage/release capacity [16, 25, 26]. Previous studies reported that the amorphous structure of the metal oxide catalysts or the homogeneous distribution of surface metal oxides might be a reason for superior catalytic activity of VOCs oxidation [27,28,29]. As revealed in Fig. 2, the CeO2–MOx oxides showed superior catalytic activity and CO2 selectivity than CeO2, which was attributed to the formation of Ce-MOx solid solution, high distribution of metal oxides or amorphous structure of the metal oxides.

Fig. 3
figure 3

XRD patterns of a CeO2, b CeO2–MnOx, c CeO2–ZrOx and d CeO2–NiOx catalysts: (a) 2θ:10–80° and (b) 2θ: 27.5–30°

Full size image

Raman studies

Raman spectroscopy was conducted to obtain more structural information for the CeO2 and CeO2–MOx (M: Mn, Zr and Ni) catalysts, and the results are displayed in Fig. 4. The sharp bands centered at about 462 cm−1 were assigned to the symmetric stretching vibrations (F2g) of cubic fluorite ceria [30, 31]. Another weak peak at about 600 cm−1 in CeO2–MOx samples was attributed to the presence of oxygen vacancies, which was proposed to promote the redox reaction of the catalysts [30]. In addition, the characteristic vibrational modes of MnO2 were detected in CeO2–MnOx catalyst at 640 cm−1 [32]. For CeO2-ZrOx and CeO2–NiOx catalysts, no Raman peaks attributed to Zr and Ni metal oxides could be observed in these two samples, respectively, which could be attributed to a relatively homogeneous dispersion of Zr and Ni metal oxides on CeO2 or the incorporation into the lattice of CeO2 [30]. It was worth noting that the deformation and shift of position for ceria F2g bands were produced in the CeO2–MOx catalysts, confirming the incorporation of M (M: Mn, Zr and Ni) metal ions into the CeO2 lattice structure [33]. Especially in the case of the CeO2–MnOx sample, the width of ceria F2g bands was widened in a strong way, and its position shifted toward 453 cm−1, which contributed to the introduction of Mn ions into CeO2 [32, 34]. It was reported that the doping of other cations into ceria lattice contributed to the generation of structural defects, which could correlate with the creation of oxygen vacancies and favor oxygen mobility [26, 35]. Thus, the relative concentration of oxygen vacancies (Cov) was calculated using the ratios of the peak areas of the Raman spectra and the results are summarized in Table 1. Obviously, CeO2–MOx showed higher Cov than that of pure CeO2. The order of the Cov was: CeMnOx (15.5%) > CeNiOx (12.1%) > CeZrOx (11.6%) > CeO2. (7.4%), which was basically consistent with toluene catalytic oxidation activity results, confirming that the oxygen vacancies were crucial sites in the reaction. The CeO2–MOx catalysts exhibited a distinctly higher Cov than pure CeO2, indicating that the transition metal oxides were successfully introduce and thereby the Cov of the Ce-based catalysts increased. Hence, the CeO2–MnOx exhibited superior performance of catalytic removal toluene than pure CeO2.

Fig. 4
figure 4

Raman spectra of a CeO2, b CeO2–MnOx, c CeO2–ZrOx and d CeO2–NiOx materials

Full size image

XPS studies

In order to investigate the surface oxidation states of CeO2 and CeO2–MOx (M: Mn, Zr and Ni) catalysts, Fig. 5 shows the XPS spectra of Ce 3d and O 1 s. According to the previous publications, the six peaks labeled as U, U″, U‴, V, V″ and V‴ in Fig. 5a are connected to Ce4+ ions, while the other two peaks U′ and V′ were attributed to the Ce3+ ions [34,35,36]. It was remarkable that the BE of Ce 3d over CeO2–MnOx catalyst presented a tendency of decline compared with pure CeO2. It might be due to the interaction between manganese and cerium oxides, which affected the electron state of cerium, resulting in an improvement of catalytic oxidation ability and the superior catalytic performance. The surface relative concentrations of Ce3+ ions (Ce3+/(Ce3++Ce4+)) for the samples were quantitatively analyzed and the results are summarized in Table 2. The Ce3+/(Ce3++Ce4+) ratio decreased in the order of CeO2–MnOx (16.2%)>CeO2-ZrOx (15.3%)>CeO2–NiOx (14.1%)>CeO2 (13.1%). It was obvious that the relative concentration of Ce3+ ions of CeO2–MnOx catalyst was the largest among the samples. It was reported [21, 35] that a higher Ce3+ content promoted the formation of the charge imbalance, oxygen vacancies and unsaturated chemical bonds on the surface of catalysts, which could improve the catalytic oxidation property of toluene. Besides, the mobility of the active oxygen species was improved by more amounts of Ce3+, which could enhance the catalytic performance for VOCs oxidation. Thereby, the CeO2–MnOx catalyst showed the excellent catalytic activity.

Fig. 5
figure 5

XPS patterns of Ce 3d (a) for a CeO2, b CeO2–MnOx, c CeO2–ZrOx and d CeO2–NiOx catalysts; O 1 s (b) for a CeO2, b CeO2–MnOx, c CeO2–ZrOx, d CeO2–NiOx and pure MnOx (inset in (c)) catalysts; Mn 2p3/2 (c) for a CeO2–MnOx and b pure MnOx; Ni 2p3/2 (d) for CeO2–NiOx catalyst

Full size image
Table 2 XPS results of CeO2, CeO2–MOx (M: Mn, Zr and Ni) and pure MnOx catalysts
Full size table

Fig. 5b shows that the XPS spectra of O 1 s for all catalysts, which was divided into two peaks. One peak at lower BE was assigned to the lattice oxygen (Olat), and the other peak at higher BE corresponded to surface oxygen (Osuf) including oxygen defect or adsorbed oxygen [34, 37]. Obviously, the binding energy of all peaks (Olat and Osuf) in CeO2–MOx shifted to higher value compared with pure CeO2, which indicated that the interaction between CeO2 and MOx (M: Mn, Zr and Ni) existed. In order to investigate the influence of various oxygen species, the ratios of Osuf/(Olat + Osuf) were calculated by integrating O 1 s spectra peak areas and the values are summarized in Table 2. The Osuf concentration of other catalysts followed by the sequence: CeO2–MnOx (30.3%)>CeO2-ZrOx (26.2%)>CeO2–NiOx (20.2%)>CeO2 (19.1%). The similar trend was observed in the relative concentration of Ce3+ ions. In addition, as previously reported in the literature [35, 38], the higher amounts of surface Ce3+ were probably related to abundant surface oxygen defects, which was helpful for an increase in surface adsorbed oxygen, highlighting the importance of the amount of adsorbed oxygen and oxygen defects on the catalytic activity of samples. It was reported that surface adsorbed oxygen species were vital for most of the catalytic oxidation reactions due to the greater mobility than lattice oxygen [35, 38]. Hence, the CeO2–MnOx catalyst with the highest Osuf/(Olat + Osuf) showed the best toluene conversion.

Additionally, Fig. 5c shows the Mn 2p3/2 XPS spectra of pure MnOx and CeO2–MnOx catalysts, which were further resolved into two components. The peaks at low BE were attributed to the Mn3+ and the high BE peaks were assigned to Mn4+, which indicated the coexistence of Mn3+ and Mn4+ on the surface of catalysts [21, 39]. Consistent with recent reports, the BE of Mn 2p3/2 peak in CeO2–MnOx shifted to higher value compared with pure MnOx, which indicated that the interaction between CeO2 and MnOx and it affected the oxidation state of manganese species [40]. In addition, the surface ion (Mn3+/(Mn4++Mn3+)) ratios and Osuf/(Olat +Osuf) calculated by integrating spectra peak areas are summarized in Table 2. Obviously, the CeO2–MnOx catalyst showed the higher Mn3+/(Mn4+ + Mn3+) ratio (46.6%) than that of pure MnOx (43.7%) and the Osuf/(Olat + Osuf) showed the same trend. Mn3+ was easier to lose the oxygen species than that on Mn4+ at low-temperature due to its weak binding capacity to oxygen [41]. Furthermore, it was reported that the presence of more Mn3+ contributed to the formation of oxygen vacancies and structural defects, which could enhance the redox properties of the catalyst during the reaction processes [42].

The Ni 2p XPS analysis of CeO2–NiOx catalyst is shown in Fig. 5d. Obviously, a broad peak appearing at about 855 eV was corresponded to Ni 2p3/2 and it confirmed the presence of Ni species [35]. Previous studies reported that the Ni 2p3/2 peaks from 850.0 to 857.0 eV could be divided into metallic nickel (Ni0) and nickel oxide (Ni2+) at 852.5 and 855.5 eV, while peaks from 857.0 to 865.0 eV were attributed to satellite peaks [43, 44]. The coexisted of two nickel species on the surface of catalyst contributed to electron transfer and promoted the redox capacity of the catalyst. Unfortunately, it was difficult to distinguish the nickel oxide state due to the weak signal resulting from the uniform dispersion of low nickel content in the bulk catalyst.

H2-TPR

In order to investigate the reducibility of each catalyst, H2-TPR analyses were carried out and the profiles are shown in Fig. 6a. What is more, the corresponding hydrogen consumption of each reduction region is summarized in Table 3. As depicted in Fig. 6a, CeO2 exhibited two reduction peaks centered at around 510 and 760  °C, which were assigned to the reduction of surface CeO2 and bulk CeO2, respectively [30, 45]. For CeO2–NiOx sample, it presented two main reduction peaks with the maxima at 320 and 770  °C, respectively. The latter might be related to the reduction of bulk CeO2, and the former was assigned to the reduction of NiO species and partial ceria [46]. Compared to the pure CeO2 catalyst, the reduction of the former peak was initiated at lower temperatures (< 200  °C) and the intensity was stronger, which indicated a significant improvement in reducibility due to the interaction between NiO and CeO2. As to CeO2-ZrOx sample, it exhibited two reduction features centered at about 550 and 750 °C, which could be assigned to the reduction of surface and bulk CeO2 [45, 47, 48]. Interestingly, compared to CeO2, the peak intensity of CeO2-ZrOx was enhanced and the range was broadened, which was attributed to the higher mobility of lattice oxygen due to the formation of vacancies or structural defects, further promoting the redox property and then improving catalytic performance [48, 49]. For CeO2–MnOx, the peak at 220 °C was related to the reduction of Mn4+ to Mn3+. The peak at about 320 °C was assigned to the reduction of Mn3+ to Mn2+ together with the possible reduction of surface ceria. The peak at 730 °C was the reduction peak of bulk CeO2 [15, 50]. Particularly, the locations of the three reduction peaks shifted to lower temperature and the intensity was greatly enhanced compared to CeO2. These phenomena indicated that the interaction between CeO2 and MnOx improved the redox property and contributed to the improvement of the catalytic oxidation of the toluene. In addition, the better low-temperature reducibility of CeO2–MnOx confirmed by the hydrogen consumption is summarized in Table 3. Obviously, CeO2–MOx exhibited larger amounts of low temperature (< 620 °C) and the H2 consumption as follows: CeO2 (190.7) < CeO2–NiOx (293.9) < CeO2-ZrOx (539.2) < CeO2–MnOx (1871.6). Obviously, the low reducibility of catalysts were improved variously. In addition, the total H2 consumption of the catalysts showed the same trend as the hydrogen consumption at low temperature: CeO2 (1009.8) < CeO2–NiOx (1013.4) < CeO2-ZrOx (1088.3) < CeO2–MnOx (2200.6), which implied that the reducibility of catalyst was improved variously due to the introduction of Mn, Zr and Ni element into CeO2.

Fig. 6
figure 6

(a) H2 -TPR profiles and (b) initial H2 consumption rate (rinitial) at low temperature for a CeO2, b CeO2–MnOx, c CeO2–ZrOx and d CeO2–NiOx catalysts

Full size image
Table 3 The results of H2-TPR and O2-TPD
Full size table

In order to study the catalytic performance deeply, the differences of reducibility for all catalysts were estimated by the initial H2 consumption rate (rinitial) and the profiles are shown in Fig. 6b. The order of low temperature reducibility was as follows: CeO2 < CeO2–ZrOx < CeO2–NiOx< CeO2–MnOx. The H2-TPR results indicated that the incorporation of M metal cations into the CeO2 crystallites could promote the reducibility of the catalyst, which might attribute to the appearance of abundant reducible oxygen species.

O2-TPD experiment

The O2-TPD was used to investigate the mobility of oxygen species in all catalysts, and the results are shown in Fig. 7. All the samples displayed desorption peaks at low temperature (< 500  °C), which could be assigned to the desorption of physisorbed oxygen and chemisorbed oxygen from the surface of materials. Besides, the peak at high temperature (> 500  °C) was ascribed to lattice oxygen (Olat) [10, 37]. Compared with pure CeO2 catalyst, the surface adsorbed oxygen of CeO2–ZrOx catalyst shifted to higher temperatures and the intensity was enhanced. The phenomena indicated that the addition of ZrOx into CeO2 could improve the thermal stability, and then increased the contents of surface oxygen significantly [51]. For the CeO2–NiOx catalyst, the intensities of surface adsorbed oxygen and lattice oxygen were much higher than that of CeO2, which was favorable for high catalytic activity for VOCs combustion [52]. As for CeO2–MnOx catalyst, the surface adsorbed oxygen moved to lower temperature. It was proved that the oxygen species got easier to be released from the surface of CeO2–MnOx catalyst. The desorption of lattice oxygen for CeO2–MnOx deviated to 822  °C, which could be attributed to the interaction between CeO2 and MnOx. In addition, the intensities of desorption peaks of CeO2–MnOx catalyst were obviously stronger than that of pure CeO2, especially the surface adsorbed oxygen [37]. In addition, based on the accurate quantitative analysis of the profiles, the amount of desorbed oxygen in each peak and total desorbed oxygen is listed in Table 3. Obviously, CeO2–MOx showed larger desorbed amounts of surface oxygen and total oxygen and the order total desorbed oxygen was as follows: CeO2 (10.7) < CeO2–ZrOx (17.0) < CeO2–NiOx (18.4) < CeO2–MnOx (40.7), which indicated that CeO2–MOx exhibited the higher capacity for adsorbed oxygen. The phenomena indicated that more surface adsorbed oxygen species formed and the mobility was improved tremendously by introduction of MOx.

Fig. 7
figure 7

O2-TPD profiles at the temperature range of 50–900  °C for a CeO2, b CeO2–MnOx, c CeO2–ZrOx and d CeO2–NiOx catalysts

Full size image

Conclusions

The physicochemical properties of the CeO2 and CeO2–MOx (M: Mn, Zr and Ni) materials were characterized by N2 adsorption–desorption, XRD, Raman, XPS, H2-TPR and O2-TPD characterization techniques. The results showed that CeO2–MOx catalysts were mainly mesoporous structures and the transition metal ions were introduced into the fluorite lattice, which contributed to the interaction between CeO2 and MOx. CeO2–MOx catalysts showed the better catalytic combustion ability of toluene and CO2 selectivity than CeO2.The T90 of toluene combustion was as follows: CeO2–MnOx (261 °C) < CeO2–ZrOx (310 °C) < CeO2–NiOx (316 °C) < CeO2 (322 °C). CeO2–MnOx catalyst exhibited the more concentration of Ce3+, surface adsorbed oxygen, structural defects and the optimal redox properties, and CeO2–MnOx catalyst showed the highest catalytic efficiency.