Binary Ce–Mn oxides confined in carbon nanotubes as efficient catalysts for ethylbenzene dehydrogenation in the presence of carbon dioxide

Abstract

The present work was undertaken to investigate the influence of binary Ce–Mn species confined in carbon nanotube (CNT) channels on the catalytic properties for oxidative dehydrogenation of ethylbenzene (EB) to styrene utilizing CO₂ as a mild oxidant. 7.0 wt% of Ce–Mn oxides were filled in CNTs by an incipient wetness impregnation method. The texture and physicochemical properties of the as-prepared materials were characterized by TEM, XRD, H₂-TPR, Raman and XPS. The binary Ce–Mn oxides confined in CNTs exhibited much higher catalytic activities than those of single Ce or Mn. Among the catalysts tested, the sample CeMn-in-CNTs with Mn/(Ce + Mn) = 0.375 exhibited the highest conversion of EB and selectivity for styrene. The superior catalytic performance could be attributed to the fact that the doped Mn species could accelerate the oxidation of Ce³⁺ towards Ce⁴⁺ by the reduction of the high valence of Mn species, the occurrence of surface oxygen vacancy and activated oxygen resulting from migration of Mn, and the change of equilibrium caused by coupling a reverse water gas shift reaction.

---

1. Introduction

As an important monomer in the modern petrochemical industry, styrene (ST) is used for the production of polystyrene, acrylonitrile–butadiene–styrene resins, styrene–acrylonitrile resins, styrene–butadiene rubber, styrene–butadiene latex, unsaturated polyester resins and so on.¹ ST is industrially produced via catalytic dehydrogenation of ethylbenzene (EB) at high temperatures (873–953 K) over potassium promoted iron oxide catalyst in the presence of a large quantity of overheated steam.² However, this extremely endothermic process suffers from drawbacks such as thermodynamic limitation, high reaction temperature and overheated steam, irreversible deactivation of catalyst and coke deposition,^2,3 according to the following chemical equation (eqn (1)).³

C₆H₅CH₂–CH₃ → C₆H₅CH=CH₂ + H₂ ΔH^θ_{298 K} = +117.6 kJ mol⁻¹ (1)

In order to solve the above problems, the alternative techniques such as dehydrogenation followed by hydrogen oxidation,⁴ membrane technologies,⁵ or the oxidative dehydrogenation (ODH) have been proposed. However for the ODH, which is a exothermic reaction (eqn (2)),⁶ in spite of the absence of detrimental thermodynamic limitations and lower operation temperatures,³ problems remain in terms of the potential explosion of oxygen-containing mixtures and the deep oxidation of alkenes to carbon oxides.

C₆H₅CH₂–CH₃ + 0.5O₂ → C₆H₅CH=CH₂ + H₂O ΔH^θ_{298 K} = −124.3 kJ mol⁻¹ (2)

The ODH of EB using CO₂ as a soft oxidant is an attractive process because it has not only supplied strategic considerations for the control, conversion and utilization of CO₂, which is one of the major greenhouse gases, but has also changed the conventional endothermic dehydrogenation reaction to a reaction (eqn (3)).

C₆H₅CH₂–CH₃ + CO₂ → C₆H₅CH=CH₂ + H₂O + CO ΔH^θ_{298 K} = +159.2 kJ mol⁻¹ (3)

Ceria is often used as the key components of the catalyst for many industrially important reactions^7–13 since ceria can exist under both oxidizing and reducing environment by changing states of Ce³⁺and Ce⁴⁺, and this property allows ceria to readily store and release oxygen from its lattice framework. But the pure CeO₂ is seldom used for its poor thermostability at high temperatures¹⁴ and other transition metal ions are usually introduced into the ceria cubic structure to create a defective fluorite-structured solid solution. Such modifications may enable the new chemicals possess special properties such as better resistance to sintering, higher oxygen storage/release capacities for various reactions.^15–17 Manganese and manganese oxides, which exhibit a multiplicity of oxidation states and can store and release oxygen in the same way as CeO₂ and Ce₂O₃,^18,19 are proved to be good candidates to make part of a ceria-based catalysts, which used for many reactions, as well as the ODH of EB to ST.

The carbonaceous materials have been extensively used in heterogeneous catalysis.^20–25 Carbon nanotubes (CNTs), a member of carbon materials with graphite layers and tubular structure, have been reported as non-metallic catalysts for various reactions^20,26–30 due to its unique features, such as nanometer-sized channel, highly conductive graphitized tube walls, large specific surface areas and high surface energy. CNTs are also used as catalyst support. Its exceptional mechanical, thermal, electronic properties, along with the possibility to chemically modify their surface by doping or grafting, make them an interesting alternative to conventional supports, and its graphitic structure make it possible for their high stability under harsh operation condition.^28,31–33 To date, many metal decorated CNTs have been studied as the catalysts for liquid, gas-phase catalytic reactions,³⁴ and the results have shown that metal nanoparticles filled in CNTs usually exhibit higher catalytic activity than the that loaded outside CNTs due to the confinement effect of CNTs.^32,34–36 However, bimetal oxides filled in CNTs catalysts used for ODH of EB are yet seldom reported so far.

Herein, we report the preparation of binary Ce–Mn oxides filled in CNTs by a simple wet chemical method and the as-prepared catalysts are used for ODH of EB to ST with CO₂ as a soft oxidant. The effect of Mn doping in ceria confined in CNTs on the catalytic performance has been investigated in detail. The proposed promoting mechanism may enlighten the perspective on the design of CNTs filling catalyst and the realization of bimetal oxide components confined CNTs as well as other related catalyst with beneficial performance.

2. Experimental

2.1. Catalyst preparation

A series of catalysts filled with different ratio of Mn/(Ce + Mn) in CNTs were prepared through incipient wetness impregnation methods. Prior to impregnation, the raw CNTs with outer diameter 10–20 nm (Chengdu Organic Chemicals in China) were refluxed with 68% HNO₃ at 393 K for 5 h to remove residual contaminants, amorphous carbon and further to open the caps, and followed by filtering and washing with deionized water until pH reaching around 7, then dried at 333 K for over 12 h.

Cerium nitrate, manganese acetate and acrylamide with different mole ratios were dissolved in certain amount of ethanol. The acid-treated CNTs were impregnated by the above mixed solution. After stirring 30 minutes, another 0.5 ml of N,N-dimethylformamide was added dropwise. After ultrasonic treatment for 20 min, continuously stirring at room temperature for 12 h, the obtained sample was dried at 333 K for 12 h, and then calcined in Ar atmosphere at 773 K with a rate of 5 K min⁻¹ and kept for 3 h. Finally, the obtained sample was labelled as CeMn-in-CNTs-x, Mn-in-CNTs and Ce-in-CNTs according to the filled different components, respectively. Herein x represented the molar ratio of Mn/(Mn + Ce).

The compared samples of loaded outside CNTs were obtained by the same impregnation method using closed CNTs. The raw CNTs were treated by refluxing in dilute nitric acid (37.5%). Then, the CNTs were impregnated with manganese acetate and cerium nitrate mixed solution under the ultrasonic assistance, followed by drying at 333 K for 12 h, and calcining at 773 K under Ar gas for 3 h. The obtained sample was defined as CeMn-out-CNTs-x.

2.2. Catalysts characterization

Catalysts were characterized by transmission electron microscope (TEM, JEM-2100 at 200 kV), energy dispersive spectroscopy (EDS, JEM-2100), inductively coupled plasma-atomic emission spectrometry (ICP-AES, optima 5300DV), X-ray diffraction (XRD, Philips X'pert Pro X-ray diffractometer using Cu K_α radiation of 0.15418 nm, operating at 40 kV and 40 mA over a 2θ range from 10 to 80°) and X-ray photoelectron spectroscopy (XPS, PHI 5000 VersaProbe), in which, all binding energies were calibrated to C 1s at 284.6 eV. The reducibility of samples was studied by temperature programmed reduction (TPR) in a mixture flow of 5 vol% H₂ in Ar, and the temperature had changed from room temperature to 1200 K at a rate of 10 K min⁻¹. In this experiment, each measured sample was controlled strictly at 50.0 mg. Raman spectra were recorded with a Raman spectrometer (Labram aramis, Horiba Jobin Yvon) using a laser excitation line at 532 nm.

2.3. Catalytic reaction tests

The ODH reaction of EB in the presence of CO₂ was carried out in a fixed-bed microreactor under atmospheric pressure. Typically, CO₂ was fed into the reactor at a flow rate of 20 ml min⁻¹ through a mass flow controller and passed through a glass evaporator filled with liquid EB maintained at 308 K (EB vapour pressure of 2.16 kPa). 50.0 mg of catalyst was loaded in a U-type reactor in each run and the reaction temperature was 823 K, GHSV = 18 000 ml g⁻¹ h⁻¹. Prior to the reaction, the catalyst was pretreated at 823 K in the presence of CO₂ for 2 h. The hydrocarbon products were analyzed with a gas chromatograph (GC-950) equipped with flame ionization detector and the gas byproducts were measured by thermal conductivity detector.

3. Results and discussion

3.1. Structure characterization of prepared catalysts

3.1.1 HRTEM. Fig. 1 shows the HRTEM images of the as-prepared catalysts with different components. It can be seen that the majority of metal-oxides nanoparticles are filled inside the channel of CNTs, the statistic result indicates that there are more than 90% of particles inside CNT channel. The CNTs' caps are opened and clean (Fig. S1†) after acid treatment and the solution is sucked into CNTs through the capillary force since the relatively low surface tension of ethanol and the assistance of sonication (Fig. S2†). The corresponding particle sizes are illustrated in the inset of Fig. 1A–C, respectively. It is noticed that the average size (about 2.0 nm) of the binary Ce–Mn nanoparticles is smaller than that (about 3.0 nm) of the single Ce or Mn nanoparticles, suggesting that Mn atoms incorporated into CeO₂ can inhibit the crystal growth of ceria.^37,38 The HRTEM images in Fig. 1A′ and B′ show the clear lattice fringes with the interlayer distance of about 0.31 nm over CeO₂, and 0.34 nm over Mn₂O₃, corresponding to crystal planes (1 1 1) with d-spacing value of d₁₁₁ = 0.31 nm in cubic CeO₂ structure and to planes (1 1 0) with d-spacing value of d₁₁₀ = 0.36 nm in cubic Mn₂O₃ structure, respectively. Fig. 1C′ shows the lattice fringes of binary Ce–Mn oxides with the interlayer distance of about 0.30 nm. Its calculated corresponding lattice parameter is 0.52 nm, less than the value 0.54 nm of the lattice parameter of ceria. The lattice parameters of ceria are decreased with the addition of Mn, suggesting that some Mn³⁺ ions are incorporated into the ceria lattice since the radius of Mn³⁺ ions (0.065 nm) is smaller than that of Ce⁴⁺ ions (0.094 nm).^39–42 Further elemental analysis of CeMn-in-CNTs-0.375 is conducted by EDS analysis (Fig. S3 and Table S1†). It showed that the ratio of Mn/(Ce + Mn) was 0.378, which was close to the stoichiometric value 0.375 during preparation, confirming binary Ce–Mn nanoparticles formation.

Fig. 1 HRTEM images of Ce-in-CNTs (A, A′), Mn-in-CNTs (B, B′) and CeMn-in-CNTs-0.375 (C, C′).

3.1.2 XRD. Fig. 2 shows the powder XRD patterns of the as-prepared samples with different Mn/(Ce + Mn) atomic ratios. The intensive and broad diffractions at 2θ = 26.2° and 44.4° observed in all profiles can be indexed as (0 0 2) and (1 0 1) diffractions of graphite structure in CNTs (JCPDS 75-1621).

Fig. 2 XRD patterns of Ce-in-CNTs (a), CeMn-in-CNTs-0.375 (b), CeMn-in-CNTs-0.5 (c), CeMn-in-CNTs-0.625 (d), Mn-in-CNTs (e).

For the single component Ce sample, the distinct diffraction peaks displayed by line a can be indexed to the CeO₂ (1 1 1), (2 0 0), (2 2 0), (3 1 1) lattice planes, which can be attributed to the cubic structure of CeO₂ (JCPDS 43-1002). When a small amount of Mn is added, e.g. CeMn-in-CNTs-0.375 as shown in line b, the XRD patterns don't show any diffractions of manganese oxides. The intensity of Ce species diffraction peaks decrease significantly, and only weak diffractions ascribed to CeO₂ are observed, which would be of a cubic fluorite structure as Machida et al. reported.³⁸ It is considered that for MnO_x–CeO₂ likely mixed oxides, the diffraction profiles at Mn/(Ce + Mn) ≥ 0.75 showed the crystallization of Mn₂O₃, whereas those at Mn/(Ce + Mn) ≤ 0.5 ascribable only to CeO₂ with a cubic fluorite structure. When the Mn amount is further increased by forming the samples of CeMn-in-CNT-0.5 and CeMn-in-CNT-0.625, as shown in line c and d, no clear diffraction peaks assigned to Mn or Ce species can be clearly identified in XRD patterns due to the small amount loading. For Mn-in-CNTs, as shown in line e, typical diffraction peaks of MnO₂ (JCPDS 81-2261) at 28.6°, 37.3°, 56.6° and Mn₂O₃ (JCPDS 33-0900) at 32.3°, 35.6°, 53.3° are identified. This means that in the CNTs environmental, Mn species not only have +4 oxidation states in MnO₂-type structure, but also +3 oxidation states in Mn₂O₃-type structure.

3.1.3 H₂-TPR. The reduction properties of the catalysts are determined by temperature programmed reduction in hydrogen atmosphere (H₂-TPR) and their main features are summarized in Fig. 3. In general, mainly distinct two reduction peaks at about 773 and 1073 K were associated to the reduction of surface ceria and the bulk CeO₂ (line a in Fig. 3).^43,44 But for Ce species confined in CNTs (line b in Fig. 3), the peaks shift to lower temperatures, and a broad reduction peak with a maximum at about 768 K and with a weak shoulder reduction peak at 556 K are observed. The peak at 768 K can be attributed to the reduction of surface CeO₂, which is lower than the reduction temperature of pure CeO₂ or Ce on Al₂O₃ supports (820 K),⁴⁵ indicating that the reducibility of CeO₂ particles confined in CNTs is enhanced due to the effect of electron deficient in the interior surface of CNTs.^31,36 The peak at 556 K is related to highly dispersed surface cerium species, which also greatly lower than 773 K.

Fig. 3 H₂-TPR profiles of CeO₂ (a), Ce-in-CNTs (b), CeMn-in-CNTs-0.375 (c), CeMn-out-CNTs-0.375 (d), Mn-in-CNTs (e), MnO_x (f).

The TPR results of Mn-in-CNTs are illustrated by line e in Fig. 3. Two distinct peaks at 602 K and 875 K with a slight shoulder about 546 K can be observed. The pure MnO_x showed two overlapped strong reduction peaks at 690 and 780 K with a slight shoulder at about 590 K, (line f in Fig. 3) which were assigned to the reductions of MnO₂ to MnO.^46–48 So the first two peaks (at 546 and 602 K) can be attributed to the reduction of MnO₂ to Mn₂O₃ and Mn₂O₃ to Mn₃O₄, the third peak at 875 K represents the reduction of Mn₃O₄ to MnO. The first two peak temperature of Mn-in-CNTs is lower than that of pure MnO_x, indicating that MnO₂ nanoparticles confined in CNTs can easily be reduced. However, the broad peak with a maximum at 875 K is higher than that of pure MnO_x at 780 K. This different behavior can be interpreted as the oxophilic Mn may remain in a more reduced state inside the channels.³⁵

Line c and d in Fig. 3 illustrate the TPR results of CeMn-in-CNTs-0.375 and CeMn-out-CNTs-0.375. Relative to Ce-in-CNTs (line b) the reduction temperatures of CeO₂ in the both samples shift obviously to lower value, indicating that doping of manganese can improve the reducibility of cerium oxide. However for sample CeMn-in-CNTs-0.375, no significant peak of the reduction of Mn₂O₃ to Mn₃O₄ is observed since Mn species was doped into CeO₂, but the reduction temperature of Mn₃O₄ to MnO is lowered obviously, indicating that cerium can also improve the reducibility of manganese oxide.

3.1.4 Raman. The Raman spectra of samples excited with 532 nm laser line are given in Fig. 4. Two major strong peaks, D band at ∼1350 cm⁻¹ and G band at ∼1583 cm⁻¹, and a shoulder peak named D′ band at ∼1610 cm⁻¹ are obviously observed.

Fig. 4 Raman spectra of CNTs (a), Mn-in-CNTs (b), Ce-in-CNTs (c) and CeMn-in-CNTs-0.375 (d).

Usually the relative intensity of the D and G peaks (I_D/I_G) can be used to estimate the extent of disorder within the samples.^34,49,50 It can be seen that the I_D/I_G of acid treated CNTs is 1.42, subsequently, the changes are not distinct after loading different metal oxides since the load substance is within tubes of CNTs. The lightly increased I_D/I_G ratios with introducing active component into CNTs channel can be ascribed to the increased defects or cavity sites, which are formed on walls of CNTs during the preparation process.

3.2. Catalytic performance

3.2.1 Catalytic activity and selectivity in ODH of EB. The catalytic activities of different samples are estimated by ODH of EB in the presence of CO₂ at 823 K and their catalytic conversions of EB and selectivity for ST are demonstrated in Fig. 5A and B, respectively. In all cases, the selectivities for ST are all over 98%, therefore the byproducts can almost be neglected, and the conversion of EB at 8 h has been used as the main evaluation of catalytic performance. As shown in Fig. 5A, for the monocomponent sample, Mn-in-CNTs shows the lowest catalytic activity with a conversion of 15.1%, and Ce-in-CNTs shows that of 20.2%. Whereas for the bicomponent sample, CeMn-out-CNTs-0.375 exhibits an apparent improvement of catalytic performance, the EB conversion is 30.8%. However for CeMn-in-CNTs-0.375, the catalytic activity sharply rises to 50.9%. The result suggests that the synergistic effect between Mn and Ce species confined in CNTs may promote greatly the ODH of EB in the presence of CO₂. Additionally, it is found that different from other samples, the selectivity of the sample CeMn-in-CNTs-0.375 is not enough high at the beginning of reaction, but it increases obviously with the reaction proceeds. As shown in Fig. 5B, the selectivity rapidly increases from the initial 97.7 to 99.2% at 8 h, and maintains a sustained and stable state. The phenomenon perhaps suggested that the doping of Mn not only increased the catalytic activity but also improved the selectivity of target products due to the confinement effect of CNT.

Fig. 5 EB conversion in function of the time on stream over as-prepared catalysts (A) and corresponding selectivity to ST (B). Mn-in-CNTs (a), Ce-in-CNTs (b), CeMn-out-CNTs-0.375 (c), CeMn-in-CNTs-0.375 (d). Reaction conditions: T = 823 K, m_cat = 50 mg, total flow = 15 ml min⁻¹.

3.2.2 Influence of loading. The catalytic activities of the catalysts confined in CNTs with loading amount 5% and different mole ratios of Mn/(Mn + Ce) are showed in Fig. 6A. It can be seen that the conversion of EB increases gradually with increasing the Mn/(Mn + Ce) ratio. When the Mn/(Mn + Ce) ratio is 0.375, the conversion increases sharply up to 42.4%, if further increasing the Mn/(Mn + Ce) to 0.75, the conversion continuously decreases. Our experimental result demonstrates that the optimal ratio of Mn/(Mn + Ce) is 0.375.

Fig. 6 EB conversions over the catalysts confined in CNTs with loading amount 5% but different Mn/(Mn + Ce) ratio (A) and over the samples CeMn-in-CNTs-0.375 with different loading amount (B). Reaction conditions: T = 823 K, m_cat = 50 mg, total flow = 15 ml min⁻¹, reaction time 3 h.

It is reported that when Mn/(Mn + Ce) ≤ 0.5, MnO_x–CeO₂ mixed oxides would form a MnO_x–CeO₂ solid solutions with a cubic fluorite structure.³⁸ It is interpreted that the steady structure of solid solution and the distorted lattice would play an important role during catalysis. Further the influence of the different total loading amount at fixed Mn/(Mn + Ce) = 0.375 on the conversion of EB is also given in Fig. 6B. It can be seen that the conversion of EB over the pristine CNTs is quite low (only 7.9%), but a little amount of Mn and Ce filled into the CNTs channel can drastically elevate the catalytic activity. The conversion increase with increase of the loading amount, when the total loading amount is up to 7%, the catalyst displays the highest conversion of EB (61.4%). Further increase the loading, the catalytic activity decreases obviously maybe owing to the channels of CNTs being blocked. It was illustrated by the compare of TEM images with CeMn amounts of 3% (Fig. 7A) and 11% (Fig. 7B). A few nanoparticles can be seen in the CNTs channel of the former sample while a large amount of nanoparticles jam in the CNTs channel of the latter. Therefore, the optimum loading mass of the total metal oxides is 7% according to the ratio of Mn/(Mn + Ce) = 0.375.

Fig. 7 TEM images of CeMn-in-CNTs with CeMn amounts of 3% (A) and 11% (B).

3.3. Mild oxidant CO₂

As a mild oxidant, CO₂ can decompose into CO and active oxygen in the ODH reaction,⁵¹ so the gas by-products over the Ce-in-CNTs (Fig. 8A) and CeMn-in-CNTs-0.375 (Fig. 8B) at different reaction temperatures are examined.

Fig. 8 The profile of gas byproducts measured at different temperatures over samples Ce-in-CNTs (A) and CeMn-in-CNTs-0.375 (B). Reaction conditions: m_cat = 50 mg, total flow = 15 ml min⁻¹.

The main by-products CO and H₂ are produced from the decomposition of CO₂ and dehydrogenation of EB. As shown in Fig. 8A, the H₂ and CO are increased as the raise of reaction temperature, indicating that rising temperature can increase the conversion of CO₂ and the yield of ST. But for CeMn-in-CNTs-0.375, it is found that CO content is increased while the H₂ content decrease gradually with the raise of temperature. In addition, it is noticed that the increase of CO content over CeMn-in-CNTs-0.375 is larger than over Ce-in-CNTs at the same temperature interval. This result revealed that the reaction routes for EB dehydrogenation over the two kinds of catalysts may be different due to the doping of manganese. In the case of Ce-in-CNTs, the catalytic reaction obeys the route of ODH reaction, whereas for the CeMn-in-CNTs-0.375, the catalytic reaction may couple a reverse water gas shift reaction: H₂ + CO₂ → H₂O + CO. Thus the content of H₂ decreases with the increasing of CO content.

3.4. Promoting effect of Mn

In order to realize the action of Mn species doping into CeO₂, XPS is applied to investigate the surface chemical states of sample CeMn-in-CNTs-0.375 before and after reaction. Fig. 9A shows XPS survey scan spectrum and distinct C 1s, O 1s, Ce 3d and Mn 2p can be observed. As shown in Fig. 9B, typical three-lobed envelops (around 880–898 eV for Ce 3d_5/2, 899–908 eV for Ce 3d_3/2 and approximately 917 eV for Ce 3d_1/2) are obviously observed and both Ce³⁺ and Ce⁴⁺ oxidation states can be clearly distinguished. After deconvolution, the XPS spectra of Ce 3d could split into ten peaks for evaluating cerium valence states.^52–59 For the fresh sample, the six peaks: v(882.9 eV), v′′(888.6 eV), v′′′(898.4 eV), u(901.2 eV), u′′(907.2 eV) and u′′′(917.0 eV) can be assigned to characteristic spectra of Ce⁴⁺ and other four peaks: v₀(880.7 eV), v′(885.7 eV), u₀(899.3 eV), and u′(904.3 eV) can be assigned to that of Ce³⁺. The relative areas (%) of them, which may be used for semi-quantitative estimation of the relative amount of cerium present in the samples as Ce⁴⁺ or Ce³⁺,⁵² are listed in Table 1. It is found that the fresh sample CeMn-in-CNTs-0.375 has predominant non-stoichiometric with 63.9% of the Ce 3d photoemission being due to Ce⁴⁺ and 36.1% ascribed to surface Ce³⁺, which may exist as Ce₂O₃.^55,60 For the used sample, it can be seen that the Ce³⁺ on the surface of catalyst decreased from 36.1 to 26.4% while Ce⁴⁺ increased from 63.9 to 73.6%, indicating that the valence state of Ce species rises after catalytic reaction.

Fig. 9 XPS survey scan spectrum of CeMn-in-CNTs-0.375 (A), Ce 3d (B), Mn 2p (C) and O 1s (D) spectra of catalysts before and after 3 h reaction at 823 K, fresh (a) and used (b) CeMn-in-CNTs-0.375.

Table 1 The quantitative results estimated by peak-fitting of Mn 2p and Ce 3d spectra for CeMn-in-CNTs-0.375

Sample	Surface composition (at%)				Distribution (%)
	C 1s	O 1s	Mn 2p	Ce 3d	Ce^3+	Ce^4+	Mn^2+	Mn^3+	Mn^4+
Fresh	93.6	5.47	0.37	0.55	36.1	63.9	0	24.7	75.3
Used	95.51	3.67	0.42	0.39	26.4	73.6	4.2	36.0	59.8

---

Fig. 9C represents the envelopes of Mn 2p spectra of sample CeMn-in-CNTs-0.375. The fresh sample has been deconvoluted into two peaks (641.0 and 642.0 eV) assigned to contribution of Mn³⁺ and Mn⁴⁺, respectively.^61–63 The relative areas (%) of Mn⁴⁺ is 75.3% and Mn³⁺ is 24.7% (Table 1), showing that the Mn species on the surface existed mainly in the form of MnO₂. It's worth noting that the Mn²⁺ does not exist in fresh sample, which is consistent with the results of XRD. For the used catalyst, the relative amount of Mn⁴⁺ decreases, whereas the amount of Mn³⁺ increases distinctly, even a small number of Mn²⁺ is observed. Apparently the Mn species has been reduced after catalytic reaction. Combining the change of valence states of Ce and Mn species, it can be considered that the reduction of Mn species promoted the oxidation of Ce species. According to the of Mars and Van Krevelen,⁶⁴ the increase of the percentage of Ce⁴⁺ on the surface of catalyst could enhance the oxygen storage capacity and transfer ability, which is benefit to improve the catalytic activity.

Furthermore, it also can be seen from Table 1 that the Mn species is increased (from 0.37 to 0.42) while the Ce species is decreased (from 0.55 to 0.39) on the surface of catalyst after reaction. It would be interpreted that the part of the doped Mn are moved out the lattice of CeO₂ to the surface of catalyst.^65,66 Therefore the oxygen vacancies also are created.

The XPS spectra of O 1s are shown in Fig. 9D. The oxygen species fall into six categories. For the fresh catalyst, the binding energy of 529.6 and 530.3 eV are indexed to the lattice oxygen of Ce–O and Mn–O, respectively.^67–69 The peaks at 531.1 and 532.2 eV are assigned to surface hydroxyls and adsorbed molecular water, respectively. The peak of 531.6 eV is considered as oxygen which double bonded to carbon from a carbonyl group (C=O), while the 533.5 eV as oxygen which single bonded to carbon (C–O).^8,38,69–71

According to the relative areas of characteristic peaks, we compared the change of different oxygen species and results showed in Table 2. After reaction, the oxygen of Ce–O increase from 6.1 to 8.8% and Mn–O from 12.1 to 14.4%, respectively, indicating that the lattice oxygen on the catalyst surface is enriched due to the compensation of active oxygen from CO₂. The relative area of oxygen from H₂O decrease slightly from 24.9 to 23.8%, indicating that the adsorbed molecular water mostly keep adsorption/desorption equilibrium during the ODH reaction, simultaneously, a little be decomposed at sites of oxygen vacancy to form active oxygen (O*) and active hydrogen (H*).^72,73

Table 2 Assignment and their relative intensities of O 1s XPS data for CeMn-in-CNTs-0.375

Assignment	Fresh		Used
	Peak position (eV)	Contribution to total O content (%)	Peak position (eV)	Contribution to total O content (%)
Ce–O	529.6	6.1	529.3	8.8
Mn–O	530.3	12.1	530.0	14.4
OH	531.1	15.3	530.6	11.2
C=O (COOR)	531.6	19.8	531.3	20.1
H2O	532.2	24.9	532.1	23.8
C–O (C–OH; COOR)	533.5	21.8	533.3	21.7

---

It is found that the oxygen from surface hydroxyls decreases distinctly from 15.3 to 11.2%, suggesting that the hydroxyl groups mould supply the activated oxygen or partial OH groups combine with H* to form water. The carbonyl group and carboxyl group may exhibit few contributions to ODH with tender change before and after reaction according to our previous work.²⁶

3.5. Possible promoting mechanism

On the basis of the above discussion, it is demonstrated that the doping of Mn may alter the ODH of EB to coupling a reverse water–gas shift reaction, promotes the oxidation of Ce³⁺ to Ce⁴⁺ by the reduction of Mn⁴⁺ to Mn³⁺/Mn²⁺, the migration of Mn to surface of catalyst to create oxygen vacancies, and the oxygen from hydroxyl groups participates in the catalytic reaction. Therefore a possible promoting reaction mechanism is proposed as shown in Scheme 1.

Scheme 1 Proposed reaction mechanism for EB dehydrogenation over CeMn-in-CNTs in CO₂.

Being the deviation from planarity, the CNTs have the ability to enrich reactant gases in its cavity due to their hollow structures.³⁵ Therein CO₂ molecules were activated firstly on the Ce–Mn oxide nanoparticles and dissociated into CO and lattice oxygen.^51,74 In the presence of lattice oxygen, along with the cyclic redox between the both redox couples of Ce⁴⁺/Ce³⁺ and Mn⁴⁺/Mn³⁺, and the migration of Mn to the surface of catalyst, the oxygen vacancies, activated oxygen, and activated hydrogen can be generated. According to literatures,^{9,72,73,75–77} the following chain of reactions can be occurred:

2Ce³⁺ + □ + O* → 2Ce⁴⁺ + O²⁻

Mn³⁺ + 3Ce⁴⁺ + O²⁻ → Mn⁴⁺ + 3Ce³⁺ + □ + O*

2Mn⁴⁺ + O²⁻ → 2Mn³⁺ + □ + O*

H₂O + □ → 2H* + O* + □

The generated active oxygen and the oxygen from surface hydroxyl groups, then attacked the EB to make hydroxyl radicals, thus, the hydrogen bonds from ethyl group broken and H₂O and ST formed. Meanwhile, H₂O can also be absorbed and dissociated by oxygen vacancies to generate more O* and H* which combined themselves to generate H₂.⁷² It is worth noticed that in our experiment the H₂ content is decreased along with the conversion of EB. It also may be attributed to a small amount of Mn³⁺ is reduced to Mn²⁺ by H* as shown in Table 1. Therefore, the promoting action of bicomponental catalyst for ODH of EB is actually that the doping of Mn promoted the formation of Ce⁴⁺, oxygen vacancies and activated oxygen, and then improved the catalytic activity and selectivity for ST.

Relative to catalyst outside CNTs, the increased catalytic performance of catalyst inside CNTs are mainly attributed to the confinement effect of CNTs, including the easy reduction of metal oxide in environment of electron deficient, the enrichment of reactants inside CNTs, and faster electroconductibility inside than outside, these factors will promote above reaction chains to occur.

4. Conclusions

In summary, a new CNTs-confined binary Ce–Mn oxide catalyst has been developed for the ODH of EB using CO₂ as mild oxidants. Relative to the single component catalyst and the binary component supported outside CNTs, the superior catalytic performance of the Ce–Mn oxides confined inside CNTs can be attributed to following reasons:

(1) The resultant binary Mn–Ce oxide particles inside CNTs have highly dispersed small particle size since the Mn atoms incorporated into CeO₂ can inhibit the crystal growth of ceria.

(2) In the condition of Mn/(Mn + Ce) = 0.375, the binary metal oxides has stable cubic fluorite structure of CeO₂ doping with Mn₂O₃ and MnO₂.

(3) Compared with the nanoparticles loaded outside of CNTs, the nanoparticles filled inside CNTs exhibit higher reduction ability, higher catalytic selectivity and reactant enrichment due to the confinement effects of CNTs.

(4) Mn doping is effective to improve decomposition of CO₂, coupling a reverse water-gas reaction of CO₂ and enhancing the conversion of EB.

(5) The migration of Mn species promotes the occurrence of surface oxygen vacancy, active oxygen, active hydrogen and the cyclic redox between the redox couples of Mn⁴⁺/Mn³⁺ and Ce⁴⁺/Ce³⁺.

Acknowledgements

This work is financially supported by the Nation Natural Science Foundation of China (no. 21073087) and open Analysis Foundation of Nanjing University.

Notes and references

1. G. R. Meima and P. G. Menon, Appl. Catal., A, 2001, 212, 239–245.
2. D. Su, N. I. Maksimova, G. Mestl, V. L. Kuznetsov, V. Keller, R. Schloegl and N. Keller, Carbon, 2007, 45, 2145–2151.
3. F. Cavani and F. Trifiro, Appl. Catal., A, 1995, 133, 219–239.
4. A. L. Sun, Z. F. Qin and J. G. Wang, Appl. Catal., A, 2002, 234, 179–189.
5. S. Elnashaie, B. K. Abdallah, S. S. Elshishini, S. Alkhowaiter, M. B. Noureldeen and T. Alsoudani, Catal. Today, 2001, 64, 151–162.
6. A. H. de Morais Batista, F. F. de Sousa, S. B. Honorato, A. P. Ayala, J. Mendes Filho, F. W. de Sousa, A. N. Pinheiro, J. C. S. de Araujo, R. F. Nascimento, A. Valentini and A. C. Oliveira, J. Mol. Catal. A: Chem., 2010, 315, 86–98.
7. S. Bernal, G. Blanco, M. A. Cauqui, M. P. Corchado, C. Larese, J. M. Pintado and J. M. Rodriguez-Izquierdo, Catal. Today, 1999, 53, 607–612.
8. A. Trovarelli, G. Dolcetti, C. Deleitenburg, J. Kašpar, P. Finetti and A. Santoni, J. Chem. Soc., Faraday Trans., 1992, 88, 1311–1319.
9. J. Kaspar, P. Fornasiero and M. Graziani, Catal. Today, 1999, 50, 285–298.
10. R. S. Monteiro, L. C. Dieguez and M. Schmal, Catal. Today, 2001, 65, 77–89.
11. D. Andreeva, V. Idakiev, T. Tabakova, L. Ilieva, P. Falaras, A. Bourlinos and A. Travlos, Catal. Today, 2002, 72, 51–57.
12. D. Andreeva, I. Ivanova, L. Ilieva and M. V. Abrashev, Appl. Catal., A, 2006, 302, 127–132.
13. X. D. Wu, J. Fan, R. Ran, J. Yang and D. Weng, J. Alloys Compd., 2005, 395, 135–140.
14. M. Shen, J. Wang, J. Shang, Y. An, J. Wang and W. Wang, J. Phys. Chem. C, 2009, 113, 1543–1551.
15. B. M. Reddy, A. Khan, Y. Yamada, T. Kobayashi, S. Loridant and J. C. Volta, J. Phys. Chem. B, 2003, 107, 5162–5167.
16. K. N. Rao, B. M. Reddy, B. Abhishek, Y.-H. Seo, N. Jiang and S.-E. Park, Appl. Catal., B, 2009, 91, 649–656.
17. J. Xu, B. Xue, Y.-M. Liu, Y.-X. Li, Y. Cao and K.-N. Fan, Appl. Catal., A, 2011, 405, 142–148.
18. L. Jia, M. Shen, J. Hao, T. Rao and J. Wang, J. Alloys Compd., 2008, 454, 321–326.
19. H. W. Nesbitt and D. Banerjee, Am. Mineral., 1998, 83, 305–315.
20. W. Xia, X. Yin, S. Kundu, M. Sánchez, A. Birkner, C. Wöll and M. Muhler, Carbon, 2011, 49, 299–305.
21. S. Maldonado and K. J. Stevenson, J. Phys. Chem. B, 2004, 108, 11375–11383.
22. S. Maldonado and K. J. Stevenson, J. Phys. Chem. B, 2005, 109, 4707–4716.
23. M. F. R. Pereira, J. J. M. Orfao and J. L. Figueiredo, Appl. Catal., A, 1999, 184, 153–160.
24. G. Mestl, N. I. Maksimova, N. Keller, V. V. Roddatis and R. Schlogl, Angew. Chem., Int. Ed., 2001, 40, 2066–2068.
25. P. Serp, Appl. Catal., A, 2003, 253, 337–358.
26. R. Rao, M. Yang, Q. Ling, C. Li, Q. Zhang, H. Yang and A. Zhang, Catal. Sci. Technol., 2014, 4, 665–671.
27. N. V. Qui, P. Scholz, T. Krech, T. F. Keller, K. Pollok and B. Ondruschka, Catal. Commun., 2011, 12, 464–469.
28. J. J. Delgado, X. Chen, J. P. Tessonnier, M. E. Schuster, E. Del Rio, R. Schloegl and D. S. Su, Catal. Today, 2010, 150, 49–54.
29. N. V. Qui, P. Scholz, T. F. Keller, K. Pollok and B. Ondruschka, Chem. Eng. Technol., 2013, 36, 300–306.
30. J. Zhou, H. Song, B. Fu, B. Wu and X. Chen, J. Mater. Chem., 2010, 20, 2794–2800.
31. H. Yang, S. Song, R. Rao, X. Wang, Q. Yu and A. Zhang, J. Mol. Catal. A: Chem., 2010, 323, 33–39.
32. P. Serp and E. Castillejos, ChemCatChem, 2010, 2, 41–47.
33. J.-P. Tessonnier, O. Ersen, G. Weinberg, C. Pham-Huu, D. S. Su and R. Schloegl, ACS Nano, 2009, 3, 2081–2089.
34. D. Wang, G. Yang, Q. Ma, M. Wu, Y. Tan, Y. Yoneyama and N. Tsubaki, ACS Catal., 2012, 2, 1958–1966.
35. X. Pan and X. Bao, Acc. Chem. Res., 2011, 44, 553–562.
36. W. Chen, X. Pan and X. Bao, J. Am. Chem. Soc., 2007, 129, 7421–7426.
37. G. S. Qi and R. T. Yang, J. Phys. Chem. B, 2004, 108, 15738–15747.
38. M. Machida, M. Uto, D. Kurogi and T. Kijima, Chem. Mater., 2000, 12, 3158–3164.
39. D. Zhang, H. Fu, L. Shi, J. Fang and Q. Li, J. Solid State Chem., 2007, 180, 654–660.
40. I. Khalakhan, M. Vorokhta, M. Chundak and V. Matolin, Appl. Surf. Sci., 2013, 267, 150–153.
41. U. Caldino, J. Phys.: Condens. Matter, 2003, 15, 3821–3830.
42. J. A. Perez-Omil, J. J. Delgado, W. Ouahbi, A. B. Hungria, N. Browning, M. A. Cauqui, J. M. Rodriguez-Izquierdo and J. J. Calvino, J. Phys. Chem. C, 2010, 114, 8981–8991.
43. M. Boaro, M. Vicario, C. de Leitenburg, G. Dolcetti and A. Trovarelli, Catal. Today, 2003, 77, 407–417.
44. P. Zimmer, A. Tschope and R. Birringer, J. Catal., 2002, 205, 339–345.
45. M. Wu, X. Y. Wang, Q. G. Dai, Y. X. Gu and D. Li, Catal. Today, 2010, 158, 336–342.
46. X. F. Tang, Y. G. Li, X. M. Huang, Y. D. Xu, H. Q. Zhu, J. G. Wang and W. J. Shen, Appl. Catal., B, 2006, 62, 265–273.
47. J. Trawczynski, B. Bielak and W. Mista, Appl. Catal., B, 2005, 55, 277–285.
48. J. Carno, M. Ferrandon, E. Bjornbom and S. Jaras, Appl. Catal., A, 1997, 155, 265–281.
49. E. Raymundo-Pinero, T. Cacciaguerra, P. Simon and F. Beguin, Chem. Phys. Lett., 2005, 412, 184–189.
50. P. Lespade, R. Aljishi and M. S. Dresselhaus, Carbon, 1982, 20, 427–431.
51. R. Rao, Q. Zhang, H. Liu, H. Yang, Q. Ling, M. Yang, A. Zhang and W. Chen, J. Mol. Catal. A: Chem., 2012, 363–364, 283–290.
52. H. Chen, A. Sayari, A. Adnot and F. ç. Larachi, Appl. Catal., B, 2001, 32, 195–204.
53. D. R. Mullins, S. H. Overbury and D. R. Huntley, Surf. Sci., 1998, 409, 307–319.
54. L. S. Zhong, J. S. Hu, A. M. Cao, Q. Liu, W. G. Song and L. J. Wan, Chem. Mater., 2007, 19, 1648–1655.
55. F. Larachi, J. Pierre, A. Adnot and A. Bernis, Appl. Surf. Sci., 2002, 195, 236–250.
56. A. Laachir, V. Perrichon, A. Badri, J. Lamotte, E. Catherine, J. C. Lavalley, J. Elfallah, L. Hilaire, F. Lenormand, E. Quemere, G. N. Sauvion and O. Touret, J. Chem. Soc., Faraday Trans., 1991, 87, 1601–1609.
57. G. Z. Chen, F. Rosei and D. L. Ma, Adv. Funct. Mater., 2012, 22, 3914–3920.
58. J. P. Holgado, R. Alvarez and G. Munuera, Appl. Surf. Sci., 2000, 161, 301–315.
59. S. Cai, D. Zhang, L. Zhang, L. Huang, H. Li, R. Gao, L. Shi and J. Zhang, Catal. Sci. Technol., 2014, 4, 93–101.
60. M. L. Trudeau, A. Tschope and J. Y. Ying, Surf. Interface Anal., 1995, 23, 219–226.
61. M. C. Alvarez-Galvan, B. Pawelec, V. A. D. O'Shea, J. L. G. Fierro and P. L. Arias, Appl. Catal., B, 2004, 51, 83–91.
62. M. Ferrandon, J. Carno, S. Jaras and E. Bjornbom, Appl. Catal., A, 1999, 180, 141–151.
63. F. Kapteijn, A. D. Vanlangeveld, J. A. Moulijn, A. Andreini, M. A. Vuurman, A. M. Turek, J. M. Jehng and I. E. Wachs, J. Catal., 1994, 150, 94–104.
64. C. Doornkamp and V. Ponec, J. Mol. Catal. A: Chem., 2000, 162, 19–32.
65. G. Li, Q. Wang, B. Zhao and R. Zhou, Fuel, 2012, 92, 360–368.
66. L. Liu, Z. Yao, B. Liu and L. Dong, J. Catal., 2010, 275, 45–60.
67. A. Bensalem, F. Bozonverduraz, M. Delamar and G. Bugli, Appl. Catal., A, 1995, 121, 81–93.
68. C. D. Wagner, D. A. Zatko and R. H. Raymond, Anal. Chem., 1980, 52, 1445–1451.
69. S. Hamoudi, F. ç. Larachi, A. Adnot and A. Sayari, J. Catal., 1999, 185, 333–344.
70. M. Oku and K. Hirokawa, J. Electron Spectrosc. Relat. Phenom., 1976, 8, 475–481.
71. S. B. Deng, H. Liu, W. Zhou, J. Huang and G. Yu, J. Hazard. Mater., 2011, 186, 1360–1366.
72. Y. Sekine, R. Watanabe, M. Matsukata and E. Kikuchi, Catal. Lett., 2008, 125, 215–219.
73. R. Watanabe, M. Ikushima, K. Mukawa, F. Sumomozawa, S. Ogo and Y. Sekine, Front. Chem., 2013, 1, 21.
74. A. L. Sun, Z. F. Qin, S. W. Chen and J. G. Wang, J. Mol. Catal. A: Chem., 2004, 210, 189–195.
75. B. K. Cho, J. Catal., 1991, 131, 74–87.
76. Z. Y. Ding, L. X. Li, D. Wade and E. F. Gloyna, Ind. Eng. Chem. Res., 1998, 37, 1707–1716.
77. L. Gang, R. Yue, Y. Jia, Y. Ni, J. Yang, H. Liu, Z. Wang, X. Wu and Y. Chen, Particuology, 2013, 11, 454–459.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra02048d

---