Oxidative Dehydrogenation of Ethane with CO₂ over Mo/LDO Catalyst: The Active Species of Mo Controlled by LDO

Abstract

A series of the layered double oxides supported molybdenum oxide catalysts were synthesized and evaluated in the oxidative dehydrogenation of ethane with CO₂ (CO₂-ODHE). The 22.3 wt% Mo/LDO catalyst delivered a 92.3%selectivity to ethylene and a 7.9% ethane conversion at relatively low temperatures. The molybdenum oxide catalysts were fully characterized by XRD, BET, SEM, TEM, UV–vis, Raman TG, and XPS. Isolated [MoO₄]²⁻ dominated on the surface of the fresh 12.5 wt% Mo/LDO catalyst. With the increase of the Mo content, the Mo species transformed from [MoO₄]²⁻ to [Mo₇O₂₄]⁶⁻ and [Mo₈O₂₆]⁴⁻ on the 22.3 wt% and 30.1 wt% Mo/LDO catalysts, respectively. The redox mechanism was proposed and three Mo species including [MoO₄]²⁻, [Mo₇O₂₄]⁶⁻, and [Mo₈O₂₆]⁴⁻ showed quite different functions in the CO₂-ODHE reaction: [MoO₄]²⁻, with tetrahedral structure, preferred the non-selective pathway; [Mo₇O₂₄]⁶⁻, with an octahedral construction, promoted the selective pathway; and the existence of [Mo₈O₂₆]⁴⁻ reduced the ability to activate ethane. This work provides detailed insights to further understand the relationship between structure–activity and the role of surface Mo species as well as their aggregation state in CO₂-ODHE.

1. Introduction

After the shale gas revolution in North America, the verified huge natural gas reserve stimulated great interest in exploiting effective ways of utilizing it to produce value-added chemicals [1,2]. Ethane, as the second most abundant component (up to 15 vol%) of shale gas, can be used as feedstock to produce ethylene with a high selectivity in steam cracking [3]. However, this process is intensive in energy consumption while suffering from the coking formation and environmentally harmful gases emission due to the thermodynamic constraint [4,5]. Several alternative approaches that enhance the effectiveness and reduce emissions are being investigated. The oxidative dehydrogenation of ethane combined with CO₂, which has an endothermic feature (C₂H₆ + CO₂ → C₂H₄ + CO + H₂O ∆H = + 177.6 KJ·mol⁻¹) and suppresses the formation of coking, can be an attractive route for producing ethylene [6]. However, the thermodynamic energy of C-H bond breaking in ethane (423 kJ/mol) is less than that of the C-C bond (377 kJ/mol); the non-selective catalytic conversion of ethane often takes place during the reaction process (C₂H₆ + 2CO₂→4CO + 3H₂, C₂H₆→2C(s) + 3H₂, C₂H₆ + H₂→2CH₄, H₂ O + CO ↔ CO₂ + H₂, CO₂ + C(s) ↔ 2CO), resulting in the formation of by-products [6].

Molybdenum-based oxide catalysts have been researched extensively because of their excellent activity in many dehydrogenation reactions [7,8]. The performance of Mo-supported catalytic systems strongly relates to the local structure and surface density of molybdenum species in CO₂-ODHE. Christodoulakis et al. derived the relationship of structure–activity, finding that the variation trend of the Mo-O-Al bond number per Mo in surface MoO_x species was consistent with the reaction rate per Mo [9]. Tsilomelekis et al. reported that terminal Mo=O sites participated in the non-selective reaction turnovers and the reaction routes followed a primarily selective pathway at high coverage [10]. The aggregation state of surface Mo species is the critical factor for this reaction. However, the effective active species in molybdenum oxide catalysts have not been determined yet. Thus, it would be more desirable to verify the effective active species to lay a foundation for the subsequent design and preparation of catalysts.

Layered double hydroxides (LDHs) are lamellar two-dimensional anionic clays composed of positively charged brucite-like layers [11]. [M²⁺_1−xM³⁺_x(OH)₂](Aⁿ⁻)_x/n·mH₂O is the general molecular formula of an LDH, in which M²⁺ and M³⁺ correspond to divalent and trivalent metal cations, respectively. An inorganic or organic anion inserts into the intercalation layers as charge compensation, written as Aⁿ⁻ [12]. The degree of aggregation can be regulated by anion exchange with different amounts of target active species. However, LDHs are easily gelatinized, which restricts its possibility for reuse [13]. The calcination of LDHs is usually adopted in the preparation of catalysts to improve its stability, resulting in layered double oxides (LDOs) as carrier [14]. Because of its unique physical and chemical properties, LDOs can be utilized as supported loading Mo to control the surface Mo species aggregation.

In this article, we first synthesized a series of Mo loaded LDHs by the anion-exchange method, then the target catalysts, Mo/LDO, were subsequently obtained by the calcination of precursors. We expect that the catalyst with the appropriate aggregation degree of Mo species can be prepared by the control of the LDO carrier and explore the most effective active species on this basis. The physicochemical properties of the Mo/LDO were characterized by BET, XRD, SEM, TEM, XPS, TG, UV–vis, and Raman spectroscopy, and the relationship between aggregation degree and the catalytic effect was demonstrated by these characterizations.

2. Results

2.1. Morphological Characteristics and Structural Properties

The powder XRD profiles of the as-prepared Mo/LDO catalysts and used Mo/LDO catalysts are demonstrated in Figure 1a,b, respectively. All the samples exhibited strong characteristic diffraction peaks related to the crystal structure of the cubic phase MgO (JCPDS card no. 45-0946), which were located at 43.8° and 63.0° and corresponded to the crystal planes of (200) and (220), respectively [15]. The minor XRD peak at 36° was attributed to (111) of MgO [16,17]. The crystallite size was estimated from the XRD patterns using the Scherrer equation. The crystallite size of the MgO phase followed the order of 30.1 wt% Mo/LDO (3.7 nm) < 22.3 wt% Mo/LDO (4.0 nm) < 12.5 wt% Mo/LDO (4.4 nm). The reflection planes such as (101), (112), (301), and (303) of CaMoO₄ were well matched with the diffraction peaks at 2θ = 18.7°, 28.8°, 53.2°, and 58.2°, manifesting the appearance of CaMoO₄ in all samples [18,19]. However, there were no diffraction peaks of aluminum oxide: manifesting aluminum oxide existed in the amorphous phase, in agreement with previous studies [20,21]. No diffraction peaks of carbon were observed because the coke formed during the reaction was considered amorphous carbon [22]. Moreover, the characteristic diffraction peaks of crystalline MoO₃ located at 23.7°, 25.8°, and 27.4°did not appear, indicating the good dispersion of Mo species over the catalyst surface [23].

The nitrogen physisorption isotherms and corresponding pore size distribution curves of the synthesized catalysts are shown in Figure 2. According to the IUPAC classification, the curves of all samples displayed a type IV isotherm and contained the hysteresis in the high pressure region of P/P₀ > 0.4, suggesting the obtained materials were mesoporous structural materials [24]. This structure was generated not only through the elimination of inter-layer CO₃²⁻ and layer collapse, but also the elimination of pTOS in the synthesis process, which acted as a template and transformed into a porous mixed oxide [20]. As Mo loading increased, the nitrogen physisorption isotherms for 30.1 wt% Mo/LDO were type IV isotherms with an H4 type hysteresis loop, indicating that plate-like particles were agglomerated and resulted in slit-shaped mesopores [25]. Such a result suggests that the aggregation degree of the catalyst will increase significantly at a high Mo load.

As shown in Figure 2b, the pore size of the 12.5 wt% Mo/LDO showed a distribution in 5–15 nm with a maximum of about 8 nm. As the Mo contents increased, the maximum pore size shifted to 5 nm while the pore diameter distribution extended. For the 30.1 wt% Mo/LDO, the number of larger apertures decreased significantly and the maximum pore size shifted to a lower value, about 3 nm. Moreover, there was a small local peak with a maximum value at 10 nm on the 30.1 wt% Mo/LDO. The difference in pore size distribution implied the changes in specific surface area and pore volume.

According to the isotherms, the detailed textural properties were summarized in

Table 1. Textural properties derived from the different Mo/LDO catalysts.

Sample	SBET (m2·g−1)	Average Pore Diameter (nm)	Total Pore Volume (cm3·g−1)
12.5 wt% Mo/LDO	203.8	5.6	0.29
22.3 wt% Mo/LDO	194.0	5.9	0.29
30.1 wt% Mo/LDO	171.0	4.0	0.17

. There was a decrease in S_BET, from 203.8 m²·g⁻¹ for the 12.5 wt% Mo/LDO to 194.0 m²·g⁻¹ for the 22.3 wt% Mo/LDO. Moreover, the specific surface area of the 30.1 wt% Mo/LDO declined significantly to 171.0 m²·g⁻¹. The decrease of S_BET was ascribed to the blocking of mesopores by the polymeric MoO_x species and the collapse of the layer, particularly in the 30.1 wt% Mo/LDO. There was a pronounced decrease in the average pore diameter and total pore volume with the increase of Mo content. According to the results of BET, the mesoporous material was synthesized, and the aggregation degree of the catalyst increased with the elevation of Mo loading.

SEM and the elemental mapping were further employed to understand the nanostructures and the surface elements distribution of the catalysts. As shown in Figure 3, all the catalysts exhibited the layered structure of solid lamellar, similar to a previous finding [26]. However, there were distinct profiles between the series of Mo/LDO at the 1 µm level. The aggregation of catalysts appeared gradually, and a pronounced agglomeration was exhibited in the 30.1 wt% Mo/LDO, manifesting the collapse of the layer structure and the blocking of the pores.

The elements of Mg and Al were uniformly distributed on the surface of the catalysts, while the density of the Mo elements increased gradually with the elevation of the Mo load. According to the results of elemental mapping, the structure of the LDO formed successfully. and it could be used as a carrier to disperse Mo species aptly.

TEM images of the as-prepared catalysts are shown in Figure 4. The 2D sheet-like crystallites nanosheets with 10–30 nm appeared on all tested samples. Such nanostructures were composed of aggregated needle shapes or folded ribbon-like crystallites [27,28]. As the Mo loading increased, the aggregation of catalysts started, especially on the 30.1 wt% Mo/LDO (Figure 4f). The measured interplanar spacing was 0.21 nm, which matched well with the d-value of the (200) plane of MgO, as shown in Figure 4g,h [25]. The average particle size of MgO on the 22.3 wt% Mo/LDO was about 4.0 nm, which was consistent with the result of XRD. Moreover, the d spacing of 0.31 nm corresponded to the (112) plane of CaMoO₄ [29].

The bonding information and molecular coordination environment of inorganic materials can be analyzed by UV–vis DRS. As shown in Figure 5a, there were absorption bands between 200 and 400 nm in the spectra, which originated from the ligand and metal charge transfer (CT) transitions (O²⁻→Mo⁶⁺) [30]. The absorption band of the 12.5 wt% Mo/LDO was mainly located at about 220–260 nm, which was ascribed to the isolated MoO_x species with a tetrahedral structure and the terminal Mo=O bond of tetrahedral molybdate [31,32]. The absorption band at around 280 nm had been assigned to both characters as tetrahedrally coordinated species and octahedrally coordinated species [33,34]. Therefore, we attributed this band to the overlap of monomer and polymerized molybdate species ([MoO₄]²⁻, [Mo₇O₂₄]^6- and [Mo₈O₂₆]⁴⁻). Moreover, the bond located at 320 nm was assigned to the octahedral species with a Mo-O-Mo bridge bond [35]. The new absorption at 300 nm was detected on the 22.3 wt% Mo/LDO, which was characterized as the overlap of the absorption bands at 280 and 320 nm, indicating the formation of the octahedral species ([Mo₇O₂₄]⁶⁻ and [Mo₈O₂₆]⁴⁻). The absorption band at 300 nm was more pronounced on the 30.1 wt% Mo/LDO, suggesting more octahedral molybdate species were formed by polymerizing the tetrahedral molybdate species. It could be deduced that introducing excessive Mo could result in a higher aggregation of Mo species.

The energy gap transformed from UV–vis DRS can be used to analyze the local structures of Mo oxides. There was an empirical linear correlation between the energy gap (E) and the average number of nearest Mo neighbors (N_Mo): N_Mo = 16–3.8 × E [36]. The average aggregation degree of Mo species increased as the energy gap decreased [37,38]. The energy gap of [MoO₄]²⁻ and [Mo₇O₂₄]⁶⁻ was 4.4 and 3.3 eV, respectively [34,39]. In this work, the energy gap decreased from 3.40 on the 12.5 wt% Mo/LDO to 3.33 on the 22.3 wt% Mo/LDO, suggesting that the structure transformed from isolated [MoO₄]²⁻ to a polymerized molybdate species, [Mo₇O₂₄]⁶⁻. As the Mo content further increased to 30.1 wt%, the energy gap decreased to 3.28, indicating the formation of [Mo₈O₂₆]⁴⁻.

The adhibition of Raman spectroscopy relates to the status of transition metal oxides; this technique can be used to study Mo species on the surface of catalysts. Figure 6 displays the Raman spectra of fresh and used Mo/LDO catalysts.

As is shown in Figure 6a, three peaks were observed in the profiles of the fresh Mo/LDO catalysts. The peaks located at 912 and 980 cm⁻¹ were related to the terminal Mo=O stretching vibration of highly dispersed [MoO₄]²⁻ [40,41]. These two peaks dominated on the surface of the 12.5 wt% Mo/LDO. As the Mo loading increased, the peak intensity of 912 and 980 cm⁻¹ gradually decreased, inferring the disappearance of isolated [MoO₄]²⁻. The peak at 877 cm⁻¹ was attributed to Mo-O-Mo stretching in [Mo₇O₂₄]⁶⁻ and [Mo₈O₂₆]⁴⁻ [40,42]. Oppositely, the peak located at around 877 cm⁻¹ was more pronounced as the Mo loading elevated; this may be ascribed to the formation of polymerized molybdate species ([Mo₇O₂₄]⁶⁻ and [Mo₈O₂₆]⁴⁻) on the 22.3 wt% and 30.1 wt% Mo/LDO. Moreover, the decrease of 877 cm⁻¹ on the 30.1 wt% Mo/LDO was more significant, suggesting the higher aggregation degree of Mo species ([Mo₈O₂₆]⁴⁻).

After the reaction (Figure 6b), the peak at 912 cm⁻¹ could still be observed on the used 12.5 wt% Mo/LDO, inferring the existence of [MoO₄]²⁻. Simultaneously, the peak intensity of 877 cm⁻¹ showed the same trend after the reaction, suggesting the existence of [Mo₇O₂₄]⁶⁻ and [Mo₈O₂₆]⁴⁻ on the 22.3 wt% Mo/LDO and 30.1 wt% Mo/LDO, respectively. The peak at 940 cm⁻¹ which was assigned to the terminal Mo=O stretch of [Mo₇O₂₄]⁶⁻ appeared on all used catalysts [43]. The peak at 940 cm⁻¹ became relatively pronounced and the peak at 912 cm⁻¹ declined over the 12.5 wt% Mo/LDO, indicating the transformation of [MoO₄]²⁻ to [Mo₇O₂₄]⁶⁻. The peak located at 1338 cm⁻¹ related to amorphous carbon appeared after the reaction, and the peak intensity decreased significantly with the increase of Mo content [44]. No peaks were observed at 822 and 994 cm⁻¹, suggesting the absence of crystalline MoO₃ [45].

The above results revealed that the Mo species on the surface of catalysts varied with the Mo content. According to UV–vis, the surface Mo species transformed from isolated [MoO₄]²⁻ to [Mo₇O₂₄]⁶⁻ and [Mo₈O₂₆]⁴⁻ with the increase of Mo content. A substance will be reduced easier because of the higher density of states capable of accepting electrons [10]. Therefore, the catalyst with an appropriate aggregation degree of Mo species exhibited the highest selectivity.

The amount of carbon deposit formed during the reaction was verified by TG. As shown in Figure 7, the first weight loss in the temperature range of 303–473 K was attributed to the removal of physical absorbed water and other volatile compounds [46]. The second weight loss between 473 K and 713 K was ascribed to the burning of amorphous carbon [47,48]. The third weight loss at 713–873 K was due to the decomposition of hydrocarbons [48,49]. The deposition of graphitic carbon was usually above 873 K: no peak was observed above 873 K, indicating the absence of graphitic carbon. [50,51]. The amount of carbon deposit formed during the reaction followed the order of 12.5 wt% Mo/LDO (21.7%) > 22.3 wt% Mo/LDO (17.4%) > 30.1 wt% Mo/LDO (11.1%). According to the results of Raman and UV–vis, isolated [MoO₄]²⁻ dominated on the surface of the 12.5 wt% Mo/LDO, which produced the largest amount of carbon deposition during the reaction, indicating that the catalyst with the isolated [MoO₄]²⁻ species may promote the non-selective pathway. As Mo content increased, the Mo species transformed from isolated [MoO₄]²⁻ to [Mo₇O₂₄]⁶⁻ and [Mo₈O₂₆]⁴⁻ on the surface of the 22.3 wt% and 30.1wt% Mo/LDO, resulting in the gradual disappearance of isolated [MoO₄]²⁻. Meanwhile, the formation of coke reduced as the Mo loading elevated, manifesting [Mo₇O₂₄]⁶⁻ and [Mo₈O₂₆]⁴⁻ that promoted the selective pathway in the reaction process.

The valence state change of the surface Mo species was investigated by XPS, and the spectra are shown in Figure 8. As shown in Figure 8a, all spectra of the as-prepared Mo/LDO catalysts displayed a binding energy peak at 232.6 eV with a satellite peak at 235.9 eV. This doublet was associated with Mo (VI), corresponding to Mo 3d_5/2 and 3d_3/2 peaks, respectively, suggesting that all Mo/LDO catalysts were fully oxidized and the existence of Mo VI species on the surface [52]. The peaks of Mo 3d broadened with the elevation of Mo content, manifesting the presence of more than one type of Mo (VI) species with different chemical characteristics [53]. Meanwhile, the intensity of both Mo 3d peaks increased with the increase of the Mo/(Mg + Al) ratio (as shown in

Table 2. Binding energies (eV) and atomic ratios of Mo/(Mg + Al) determined by XPS.

Sample	Mo 3d5/2	Mo 3d3/2	Mo/(Mg + Al) (%)
12.5 wt% Mo/LDO	232.6	235.9	8.0
22.3 wt% Mo/LDO	232.6	235.9	13.8
30.1 wt% Mo/LDO	232.6	235.9	18.3

), indicating the increase of Mo species on the surface of the samples.

2.2. Reaction Mechanism Analysis

After the reaction, part of the Mo (VI) species was reduced into lower states and new peaks showed up (Figure 8b). The Mo 3d_5/2 binding energy peak appearing at 230.2 eV was attributed to Mo (IV) and the peak at 231.2 eV was related to Mo (V) [54,55]. The concentrations of Mo (VI), Mo (V), and Mo (IV) were 34.91%, 26.87%, and 38.22%, respectively. The Mo (VI) species on the surface of fresh catalyst would be partially transformed into low-valance species under the reaction conditions.

Moreover, as shown in Figure 8c, the residual carbon on the as-prepared catalyst might originate from the carbon-containing compounds during sample preparation, resulting in the existence of peaks in the fresh 22.3 wt% Mo/LDO [56]. The peak at 289.2 eV was attributed to the C=O bond in CO₃²⁻, which may formed due to the adsorbed carbon dioxide [57,58]. The peak located at 286.3 eV was ascribed to the function group of the C-O bond in hydrocarbon [59,60]. The higher intensity of the C-O bond after the reaction also indicated the formation of hydrocarbon. The peaks located at 285 and 284.5 eV were related to the sp³ C-C bond in amorphous carbon and the sp² C-C bond in graphitic carbon; both peaks increased significantly after the reaction, demonstrating the formation of carbon deposition after the CO₂-ODHE reaction [61,62].

According to the previous literature, the mechanism of the Mo-based catalysts in the CO₂-ODHE reaction (as shown in Figure 9) was the redox mechanism [6]. Ethane adsorbed by interacting with the lattice oxygen sites of the dispersed MoO_x; subsequently, the adjacent lattice oxygen achieved C-H bond rupture by extracting one of the hydrogens in ethane. After the initial C-H bond cleavage, the fragment -C₂H₅ eliminated the β-H to form ethylene. The leaving hydroxyl groups at the Mo center recombined and desorbed into H₂O, leaving an oxygen site. CO₂ was adsorbed as an oxidant at the adjacent O site to generate carbonate intermediates, and the O=C=O bond was broken, where O was added into the oxygen site of the reduced Mo species. The remaining CO* continued to adsorb on the original O site. Finally, the CO dissociated and the Mo site returned to its original state to complete the cycle.

3. Discussion

Figure 10 and

Table 3. The detail of catalytic performance over the Mo/LDO catalysts.

Sample	Conversion (%)	Selectivity (%)	Yield (%)	C2H4 Formation Rate, (µmol·h−1·g−1cat)
12.5 wt% Mo/LDO	4.4	90.5	4.0	10.6
22.3 wt% Mo/LDO	7.9	92.3	7.3	15.4
30.1 wt% Mo/LDO	5.9	88.9	5.2	10.2

exhibit the catalytic performance of the Mo/LDO catalysts. The reactivity of the LDO was consistent with the blank test, manifesting that the LDO was not the active species for the reaction. As expected, the 22.3 wt% Mo/LDO delivered an excellent selectivity to ethylene (92.3%) at an approximately 7.9% ethane conversion in the stable stage. All the curves of ethane conversion exhibited a consistent trend that decreased first and stabilized subsequently. The labile stage was attributed to the coke deposition on the catalyst which primarily took place in the initial time-on-stream [63].

The initial formation rate of ethylene exhibited the same trend with the increase of Mo loading. The 22.3 wt% Mo/LDO catalyst reached the highest initial rate of ethylene formation, 15.4 µmol·h⁻¹·g⁻¹_cat, while the ethylene formation rate of the 12.5 wt% Mo/LDO was slightly higher than that of the 30.1 wt% Mo/LDO catalyst, 10.6 and 10.6 µmol·h⁻¹·g⁻¹_cat, respectively, indicating [MoO₄]²⁻ possessed a high activity for ethane activation. However, Raman and UV–vis revealed that isolated [MoO₄]²⁻ dominated on the surface of the fresh 12.5 wt% Mo/LDO catalyst, the C₂H₄ selectivity was only 83% in the initial stage, and side reactions were prone to occur. It has been reported that it is easier for the tetrahedral species to extract oxygen from the terminal Mo=O bonds so that ethane is over-oxidized to form carbon monoxide [64]. Therefore, the catalyst with the isolated [MoO₄]²⁻ species is prone to promote the non-selective pathway. With the increase of Mo content, the Mo species on the 22.3 wt% Mo/LDO catalyst transformed from [MoO₄]²⁻ to [Mo₇O₂₄]⁶⁻, while the C₂H₄ selectivity elevated from 83.0% to 92.5% in the initial stage, indicating [Mo₇O₂₄]⁶⁻ promoted the selective pathway in the reaction process. With the formation of the polymerized molybdate species, [Mo₈O₂₆]⁴⁻, the initial formation rate of ethylene decreased significantly. The new species in a highly polymeric state reduced the ability to activate ethane and resulted in the reduction of the ethylene formation rate.

As shown in Figure 10c, the curve of the 12.5 wt% Mo/LDO in ethylene selectivity increased first and then reached a plateau. A significant change in selectivity could be observed during the initial period of the reaction due to the structural transformation, according to the previous literature [6]. The initial polymerization of Mo led to the formation of the polymer Mo oxo species, resulting in the improvement of the ethylene selectivity.

The variety of reaction performances indicated that [MoO₄]²⁻ was the most active species, whereas it was prone to promote the non-selective oxidation of ethane. The transformation in the active species increased the selectivity to ethylene because of the enhancement in the M=O bond. However, the reaction performance decreased significantly with the formation of [Mo₈O₂₆]⁴⁻, indicating it was not conducive to the reaction. It could be deduced that a suitable proportion of [MoO₄]²⁻ and [Mo₇O₂₄]⁶⁻ species over the surface of the catalyst enabled it to maintain a high selectivity even under a high conversion.

4. Materials and Methods

4.1. Catalyst Preparation

All the materials were used without further purification. Mg(NO₃)₂·6H₂O, Al(NO₃)₃·9H₂O, NaOH, Na₂CO₃, and CaMoO₄ were purchased from Macklin Biochemical Co., Ltd. (Shanghai, China). Para-toluene sulfonic acid (pTOS) and HNO₃ were purchased from Aladdin Reagent Co., Ltd. (Shanghai, China).

4.1.1. Preparation of LDH

The hydrotalcite was synthesized by the co-precipitation method, which was reported with some modifications [12]. Briefly, the precursors Mg(NO₃)₂·6H₂O and Al(NO₃)₃·9H₂O were mingled in deionized water with a proper Mg/Al molar ratio. The basic solution consisting of NaOH and Na₂CO₃ was prepared to adjust the pH of the mixture. Subsequently, the solutions mentioned above were added into a flask by dropwise addition. Simultaneously, the pH value of the mixture solution was maintained at around 9.5 by controlling the titration rate of both solutions. The formative precipitates were aged in their mother liquor at 338 K for 12 h. Finally, the precipitates were filtered, rinsed with distilled water until the pH value was 7, and dried at 353 K overnight.

4.1.2. Preparation of Mo/LDO

The Mo/LDO catalysts were prepared by the anion exchange method [15]. First, Mg-Al/LDH was presented to the anion exchange at pH 4.5 (controlling the pH with HNO₃) with pTOS (an organic swelling agent). After 1 h, the solution containing CaMoO₄ with a molar ratio of Mg: Al: Mo: pTOS = 3: 1: x: 0.79 was dropped into the reaction mixture at a speed of 1 mL/min. The pH of the mixture was adjusted to 4.5, and the forming slurry was stirred for 24 h. Subsequently, the pH of the mixture was modulated to 10 by dropping the solution of NaOH. All the samples were aged at 338 K for 18 h, then filtered and washed with distilled water to pH = 7. The precipitation was dried at 353 K for 12 h and calcined in a muffle furnace at 873 K for 2 h to obtain the final catalyst.

4.2. Catalyst Characterizations

XRD analysis of the powdered samples was conducted on a D/220-PC diffractometer (Rigaku Corporation, Tokyo, Japan) with a Cu Kα radiation (λ = 0.15406 nm), which operated at 40 kV and 30 mA with a scan rate of 5°·min⁻¹, in the range of 2θ from 10° to 75°. The Brunauer–Emmett–Teller (S_BET), total pore volume (V_total) and Barrett–Joyner–Halenda pore diameter (D_BJH) were measured on an ASAP-2460 analyzer (Micromeritics, Norcross, GA, USA). Each sample was evacuated at 150 °C for 5 h for outgassing before the surface area measurement. The morphology and structure of the catalysts were determined by scanning electron microscopy (Carl Zeiss SIGMA, Carl Zeiss AG, Oberkochen, Germany) equipped with elemental analysis and a transmission electron microscope (FEI Tecnai G2 F20, FEI Company, Hillsboro, OR, USA). Laser Raman spectra were collected by a LabRAM HR Evolution (HORIBA Scientific, Kyoto, Japan) with a 532 nm laser at ranges of 600−1500 cm⁻¹. Diffuse reflectance UV–vis spectra were measured by a UV-3600 UV/vis spectrophotometer (Shimadzu Corporation, Kyoto, Japan). The spectra of the catalysts were recorded in the wavelength range of 200–600 nm using BaSO₄ as the standard. The thermogravimetric-differential thermal analysis was measured by a HITACHI STA 7300 (Hitachi, Kyoto, Japan) to determine the formation of carbon deposition during the reaction. X-ray photoelectron spectroscopy (Thermo Scientific K-Alpha, Thermo Fisher Scientific, Waltham, MA, USA) was carried out to measure the surface elemental variation of the as-prepared Mo/LDO and used 22.3 wt% Mo/LDO.

4.3. Catalyst Test

The catalyst testing of the Mo/LDO catalysts were carried out in a continuous fixed-bed flow reactor (Yuanbang Electric Furnace Manufacturing Co., Longkou, China) at atmospheric pressure. Typically, 300 mg catalysts were put in the quartz tube for the catalyst test. Under N₂ atmosphere (15 mL/min), the temperature of the reactor elevated to 873 K with a speed of 10 K/min. Then, the feedstock C₂H₆, CO₂, and N₂ (as a diluent gas) was introduced into the reactor with the ratio of 1: 1: 2 and maintained for 5 h. Utilizing a gas chromatograph (GC-2000, Shanghai Ramiin Instrument Co., Shanghai, China) equipped with a flame ionization detector (FID) detector, the feed and products were analyzed.

The conversion of ethane, the selectivity and yield of ethylene, and the initial formation rate of ethylene were calculated according to the following formula:

$${\mathrm{C}}_2{\mathrm{H}}_6\mathrm{Conversion}\left(\%\right)=\frac{{\mathrm{n}}_{{\mathrm{C}}_2{\mathrm{H}}_6,\mathrm{in}}-{\mathrm{n}}_{{\mathrm{C}}_2{\mathrm{H}}_6,\mathrm{out}}}{{\mathrm{n}}_{{\mathrm{C}}_2{\mathrm{H}}_6,\mathrm{in}}} \times 100$$

(1)

$${\mathrm{C}}_2{\mathrm{H}}_4\mathrm{Selectivity}(\%)=\frac{{\mathrm{n}}_{{\mathrm{C}}_2{\mathrm{H}}_4,\mathrm{out}}}{{\mathrm{n}}_{{\mathrm{C}}_2{\mathrm{H}}_6,\mathrm{in}}-{\mathrm{n}}_{{\mathrm{C}}_2{\mathrm{H}}_6,\mathrm{out}}} \times 100$$

(2)

$${\mathrm{C}}_2{\mathrm{H}}_4{\mathrm{Yield}(\%)=\mathrm{C}}_2{\mathrm{H}}_6{\mathrm{Conversion}(\%) \times \mathrm{C}}_2{\mathrm{H}}_4\mathrm{Selectivity}(\%)$$

(3)

$$\begin{gathered} \mathrm{Initial}\mathrm{formation}\mathrm{rate}\mathrm{of}\mathrm{ethylene}(\mathsf{\mu }\mathrm{mol}·{\mathrm{h}}^{-1}·{\mathrm{g}}_{\mathrm{cat}}^{-1})=\frac{{\mathrm{n}}_{{\mathrm{C}}_2{\mathrm{H}}_6(\mathrm{in},\mathrm{In}\mathrm{Initial}\mathrm{Strage})}}{{\mathrm{m}}_{\mathrm{cat}}} \\ \times \frac{{\mathrm{C}}_2{\mathrm{H}}_6\mathrm{Conversion}}{100} \times \frac{{\mathrm{C}}_2{\mathrm{H}}_4\mathrm{Selectivity}}{100} \end{gathered}$$

(4)

$$\mathrm{Carbon}\mathrm{balance}=\frac{{\mathrm{C}}_{\mathrm{in}}-{\mathrm{C}}_{\mathrm{out}}}{{\mathrm{C}}_{\mathrm{in}}}$$

(5)

where (C) denotes the number of moles of carbon. The maximum total carbon balance at the end of the run did not exceed 5%.

5. Conclusions

A series of LDO-supported molybdenum oxide catalysts were synthesized and characterized by XRD, BET, SEM, TEM, Raman, UV–vis, TG, and XPS. The catalytic performance was tested in the CO₂-ODHE reaction. The 12.5 wt% Mo/LDO catalyst with isolated [MoO₄]²⁻ on the surface exhibited only a 4.4% ethane conversion and produced the largest amount of carbon deposition during the reaction. As the Mo content increased, the Mo species transformed from isolated [MoO₄]²⁻ to a polymerized molybdate species, [Mo₇O₂₄]⁶⁻, on the surface of the 22.3 wt% Mo/LDO, delivering an excellent selectivity to ethylene (92.3%) with a 7.9% ethane conversion. Due to the higher aggregation degree of Mo species on the 30.1 wt% Mo/LDO ([Mo₈O₂₆]⁴⁻), a decreased ethane conversion was found. The CO₂-ODHE reaction over Mo/LDO catalysts in this work is supposed to follow the redox mechanism. The surface Mo species and their aggregation state could determine the reaction route. Isolated [MoO₄]²⁻ preferred the non-selective pathway, [Mo₇O₂₄]⁶⁻ promoted the selective pathway, and the existence of [Mo₈O₂₆]⁴⁻ reduced the ability to activate ethane.