Citation

Wei Fen, Xue Weichao, Yu Zhiyang, Lu Xue Feng, Wang Sibo, Lin Wei, Wang Xinchen. Dynamic cooperations between lattice oxygen and oxygen vacancies for photocatalytic ethane dehydrogenation by a self-restoring LaVO₄ catalyst[J]. Chin. Chem. Lett., 2024, 35(3): 108313  

Fen Wei¹, Weichao Xue¹, Zhiyang Yu, Xue Feng Lu, Sibo Wang^*, Wei Lin^*, Xinchen Wang^*     

Ethene is a valuable platform chemical in modern chemical industry. The traditional route for ethene production is steam cracking of the naphtha from crude oil [1]. Alternatively, high-temperature cracking of ethane is another ethene production approach, because of the exploitation of shale gas that supplies abundant ethane over the past decade [2–4]. Although being used industrially, operating these two recipes requires harsh conditions, including high reaction temperature (> 900 ℃), coupled with fast outlet gas cooling (< 800 ℃) and short contact time (~0.3 s) to avoid overoxidation of ethene [5–7]. Apart from steam cracking, thermocatalytic oxidative and non-oxidative ethane dehydroge-nation (EDH) to produce ethene have stimulated great interest in recent years [8–10]. The oxidative EDH involving an oxidant (e.g., O₂) is thermodynamically favorable owing to the exothermic feature but suffers easily from overoxidation to form CO₂ [11,12]. Comparatively, the nonoxidative EDH bears an obvious merit in preventing overoxidation, yet this route usually encounters severe catalyst deactivation caused by coke formation and metal sintering at high temperatures [13], which imposes complex and even high-cost processes to regenerate the catalysts [14].

Besides thermal strategies, the technology of light-driven photocatalysis has recently exhibited immense vitality in dehydrogenation of alkanes to olefins [15–24], which enables the break of thermodynamic limitation and the handling of the reactions under mild conditions. For example, Pt⁺-loaded black TiO₂ presents excellent conversion and high stability for cyclohexane dehydrogenation, but a much slower conversion rate for EDH, even at a very negative partial pressure [19]. Another photocatalytic EDH is achieved over Cu/TiO₂ under UV light irradiation, affording moderate conversion (1.7%) and high ethene selectivity (98.41%) using diluted ethane (10%) [20]. Despite these achievements, the study of photocatalytic EDH is just in its infancy, and it is challenging to realize efficient and selective light-initiated EDH for making ethene, especially those mediated by non-noble-metal catalysts in undiluted ethane.

Herein, we demonstrate the facile preparation of an O_v-rich LaVO₄-O_v photocatalyst, which displays high activity and stability for EDH via dynamic collaborations between O_latt. and O_v. The LaVO₄-O_v material is created through a one-pot hydrothermal method and characterized by diverse physiochemical techniques. Under simulated sunlight irradiation, pristine LaVO₄-O_v actively dehydrogenates pure C₂H₆ to form C₂H₄ (15.4 µmol h⁻¹ g⁻¹) and CO (26.8 µmol h⁻¹ g⁻¹) with a conversion of 2.3%. By depositing a small amount of Pt as a cocatalyst, the yield and selectivity of C₂H₄ are promoted to 275 µmol h⁻¹ g⁻¹ and 96.8%. Noteworthily, the photocatalyst appears nearly no coke formation after the reaction. The ¹³C and ¹⁸O labeled isotope reactions validate that C₂H₄ produces from C₂H₆, while CO originates from C₂H₆ and O_latt.. Importantly, the consumed O_latt. can recover readily by contacting the catalyst with O₂ in air, which thus constructs a dynamic cycle of O_latt. and achieves a simplistic catalyst regeneration. The restored LaVO₄-O_v displays good reusability and affords a turnover number of about 305 based on the determined catalytic-active sites. The O_v is revealed to strengthen optical harvesting, boost ethane capture/activation, and facilitate separation and migration of charge carriers, contributing to the augmentation of photocatalytic efficiency. DFT and in-situ DRIFTS identify the key intermediates in the EDH process and support the construction of the possible reaction mechanism.

The LaVO₄-O_v and LaVO₄ samples exhibit similar X-ray diffraction (XRD) patterns with all diffraction peaks indexed to a monoclinic LaVO₄ phase (Fig. 1a) [25]. The Fourier transform infrared spectra (FTIR) indicate that O_v formation affects the chemical feature of the VO₄³⁻ unit (Fig. S1 in Supporting information) [26,27]. The XPS survey spectra show only La, V and O elements, except C, indicating their high purities, and further analysis on La and V elements reveals the forms of La³⁺ and V⁴⁺/V⁵⁺ cations (Fig. S2 in Supporting information) [27,28]. The asymmetric O 1s spectrum comprises three peaks with the dominant one at 529.7 eV (Fig. 1b), attributing to O_latt.. The additional two peaks at 530.5 and 532.0 eV correspond to O_v and surface adsorbed oxygen species (O_ads.), respectively [29–31]. Compared with LaVO₄-O_v, the V 2p and O 1s spectra of LaVO₄ are negatively shifted by about 0.2 eV, meaning that introducing O_v into LaVO₄ endows these elements with enriched electrons [32].

Figure 1

	Download: JPG larger image
Fig. 1. (a) XRD patterns, (b) O 1s spectra, (c) EPR spectra, (d) DRS and colors of LaVO4-Ov and LaVO4. (e) The calculated band structure, (f) DOS, (g) FESEM image, (h) TEM image, (i) HRTEM image, and (j) EDS maps of LaVO4-Ov.

The defective structure of LaVO₄-O_v was probed by electron paramagnetic resonance (EPR). A strong signal (g = 1.977) is detected for LaVO₄-O_v, but not for LaVO₄ (Fig. 1c), which confirms the existence of O_v-related V⁴⁺ species [28]. Mesopores are revealed present in LaVO₄-O_v and LaVO₄ (Fig. S3 in Supporting information), and the former holds a larger specific surface area than the latter, which underlines that formation of O_v endows LaVO₄ with beneficial texture properties for heterogeneous photocatalysis.

The UV–vis diffuse reflectance spectra (DRS) indicate LaVO₄-O_v and LaVO₄ have broad light absorption in UV-visible region (Fig. 1d). Besides, LaVO₄-O_v exhibits a reinforced infrared harvesting, compared with LaVO₄, due to the presence of O_v that elevates the Fermi level and builds defect level in the bandgap [29,33–36]. The Tauc plots determine the bandgap energies (E_g) of LaVO₄-O_v (2.21 eV) and LaVO₄ (2.28 eV) (Fig. S4 in Supporting information). From the Mott-Schottky curves with positive slopes (Fig. S5 in Supporting information), the flat band potentials of LaVO₄-O_v and LaVO₄ are estimated at 0.46 and 0.50 V (vs. NHE), which are considered as their conduction band (CB) potentials. Therefore, the band structures of LaVO₄-O_v and LaVO₄ are defined (Fig. S6 in Supporting information).

The band structures and density of states (DOS) of LaVO₄-O_v and LaVO₄ were theoretically simulated. The modeled LaVO₄-O_v and LaVO₄ feature indirect bandgap with E_g values of about 2.4 and 2.5 eV (Fig. 1e and Fig. S7a in Supporting information), respectively, close to the experimental data. The DOS indicate that the Fermi levels of LaVO₄-O_v and LaVO₄ are located above the valence band (VB) maximum (Fig. 1f and Fig. S7b in Supporting information), due to perfect or lower energy electron filling. Moreover, the VBs are dominated by oxygen p-orbital, inferring that the photo-formed holes will retention on oxygen atoms for oxidation reactions, while the CBs consist of the d-orbitals of V and La with a localized electronic density of states. These results suggest that O_v not only shifts the Fermi level of LaVO₄, but also makes it overlap with the CB. Moreover, the small difference in the E_g values of LaVO₄-O_v and LaVO₄ reflects that the intensified photoabsorption of LaVO₄-O_v should be induced by oxygen defects on the surface, rather than in the bulk.

The field emission scanning electron microscopy (FESEM) image of LaVO₄-O_v presents a morphology of nanoparticles in size around 100 nm (Fig. 1g). The transmission electron microscopy (TEM) image exposes the presence of mesopores (indicated by arrows, Fig. 1h), which might be the result of ammonia gas release during the crystallization process. In the high-resolution TEM (HRTEM) image (Fig. 1i), two lattice spacing values of 0.485 and 0.497 nm, with an inclination angle of 71.5°, are well resolved, correspond to the (011) and (110) facets of monoclinic LaVO₄. Further analysis of the HRTEM image reveals that the particle has high crystallinity and dominantly exposes (110) facets on the surface [37]. Energy dispersive spectrometry (EDS) maps display the uniform distribution of La, V and O elements of the LaVO₄-O_v material (Fig. 1j).

Photocatalytic EDH activities of the samples were evaluated by simulated sunlight in pure ethane with a partial pressure of 1 atm, which are much harsh conditions compared with previous works using diluted ethane at negative pressures [19,20,38]. The LaVO₄ catalyst is active to the EDH reaction with C₂H₄ and CO as the main products (Fig. 2a). In comparison, LaVO₄-O_v presents a enhanced activity, affording C₂H₄ and CO at rates of 15.4 and 26.8 µmol h⁻¹ g⁻¹, respectively, which corresponds to a C₂H₄ selectivity of 47.8% and a C₂H₆ conversion of 2.3%. The moderate C₂H₄ selectivity may result from the strong oxidizing ability of the photoinduced holes that leads to C₂H₄ overoxidation to form CO. However, the EDH efficiency is expected to be controllable by tailoring the active sites and redox potentials of LaVO₄-O_v. Positively, the preliminary results reveal that the yield rate and selectivity of C₂H₄ are improved to 275 µmol h⁻¹ g⁻¹ and 96.8% by depositing Pt as a cocatalyst (Fig. 2b), and detailed studies are ongoing in our group. The high performance of LaVO₄-O_v is mainly attributed to the presence of Ov, serving as active sites for the EDH reaction. Also, O_v enhances light-to-heat conversion to drive the reaction. Specifically, LaVO₄-O_v with a strong infrared absorption delivers a higher photo-to-thermal conversion than LaVO₄ (Fig. S8 in Supporting information), that is, 170 ℃ for LaVO₄-O_v and 130 ℃ for LaVO₄.

Figure 2

	Download: JPG larger image
Fig. 2. (a) Photocatalytic ODE performance of different samples, (b) activities of LaVO4-Ov and Pt/LaVO4-Ov samples, (c) activity cycling tests of LaVO4-Ov after regeneration. MS spectra of (d) C2H4 and (e) CO produced from the 13CH312CH3 isotope experiment, and (f) CO produced from 12CH312CH3 over the used LaVO4-Ov after regenerating with 18O2.

Interestingly, CO produces from EDH without an oxidant, indicating CO should come from C₂H₆ oxidation by O_latt. of the catalysts. The participation of O_latt. is mirrored by the time-dependent activity of LaVO₄-O_V, showing a gradual decay in the 4 h reaction (the first cycle, Fig. 2c), due to consumption of catalytic-active O_latt and formation excessive O_v. To assess the reusability, the used LaVO₄-O_v is employed to catalyze two sets of 4h-reaction, but appears substantial deactivation (Fig. S9 in Supporting information). In another test, when the used LaVO₄-O_v is first exposed to air for regeneration and then re-applied to run the reaction repeatedly, it well preserves the original activity (Fig. 2c). These highlight that O_latt. realizes regeneration by exposure to air, attaining a dynamic cycle of O_latt. and the recovery of its intrinsic activity [21,39,40], together with its color becomes almost the same to the fresh sample (Fig. S10 in Supporting information). In other words, LaVO₄-O_v holds an oxygen storage/release ability, like the well-studied CeO₂ [41,42]. Therefore, CeO₂ is also utilized for photocatalytic EDH, but just exhibits negligible activity (Fig. 2a), which however stresses the superior performance of LaVO₄-O_v.

Control experiments indicate no product yields without LaVO₄-O_v, C₂H₆, or photoirradiation (Fig. S11 in Supporting information), suggesting that excitation of the photocatalyst triggers the EDH reaction. To estimate the contribution of thermal-catalysis, EDH is conducted at 200 ℃ (higher than 170 ℃ in photocatalysis) in dark, just affording a small C₂H₄/CO production (Fig. S11), emphasizing the high EDH efficiency stems primarily from photocatalysis.

XRD, XPS and TEM reveal no crystal, surface, and structural changes for LaVO₄-O_v after activity cycling tests (Fig. S12 in Supporting information), highlighting the outstanding stability of the photocatalyst after regeneration. Moreover, thermogravimetry (TG) and TG-mass spectroscopy (TG-MS) tests disclose nearly no noticeable carbon deposition forms on spent LaVO₄-O_v (Fig. S13 in Supporting information), which underlines the superiority of photocatalysis in preventing coke formation.

¹³C and ¹⁸O labeled isotope experiments were conducted to track the origin of C₂H₄ and CO. First, the fresh LaVO₄-O_v is employed for dehydrogenating ¹³CH₃¹²CH₃. The MS spectrum of C₂H₄ gives dominated fragments of ¹³CH₂¹²CH₂ (m/z = 29) (Fig. 2d and Fig. S14 in Supporting information), while those of CO are indexed to ¹³CO (m/z = 29) and ¹²CO (m/z = 28) (Fig. 2e). Such results confirm that C₂H₄ derives from C₂H₆, while CO comes from C₂H₆ and O_latt.. In the ¹⁸O-labeled experiment, the used LaVO₄-O_v is exposed to ¹⁸O₂ for regeneration, and then applied to catalyze ¹²CH₃¹²CH₃ conversion, which results in the detections of main C¹⁸O (m/z = 30) and C¹⁶O (m/z = 28) (Fig. 2f), proving O_latt. recovery of the catalyst from molecular O₂ [43]. Such isotope experiments solidly validate the sources of C₂H₄ and CO and directly illustrate the dynamic O_latt. loop of LaVO₄-O_v.

The steady-state photoluminescence (PL) and time-resolved PL (TR-PL) spectra reveal that LaVO₄-O_v manifests inhibited recombination and promoted separation of light-induced carriers, compared with LaVO₄ (Fig. S15 in Supporting information). Meanwhile, the electrochemical impedance spectra (EIS) indicate LaVO₄-O_v retains a lowered charge-transfer resistance than LaVO₄ (Fig. S16a in Supporting information), and the transient photocurrent responses reflect electron transfer in LaVO₄-O_v is much faster than that in LaVO₄ (Fig. S16b in Supporting information). These findings suggest O_v formation promotes separation and migration of the charges, thus boosting photocatalytic EDH efficiency.

Temperature-programmed desorption of ethane (C₂H₆-TPD) was carried out to probe the active sites of LaVO₄-O_v and LaVO₄. The C₂H₆-TPD profiles exhibit two broad signals centered at around 250 and 500 ℃ (Fig. 3a), attributing to physical and chemical adsorption of ethane, respectively. The corresponding peak areas of LaVO₄-O_v are greater than that of LaVO₄, indicating O_v of LaVO₄-O_v promotes ethane adsorption. The determined number of active sites on LaVO₄-O_v is about 2.68 µmol/g, and hence, the turnover number of the catalyst in the activity cycling tests reaches 305, confirming the catalytic nature of the photo-driven EDH reaction.

Figure 3

	Download: JPG larger image
Fig. 3. (a) C2H6-TPD, (b) H2-TPR and (c) O2-TPD profiles of LaVO4-Ov and LaVO4. (d) In situ EPR spectra of LaVO4-Ov in C2H6 under varied conditions. (e) In situ DRIFTS spectra of LaVO4-Ov in C2H6 under light irradiation for different durations.

The redox abilities of the catalysts were estimated by temperature-programmed reduction of hydrogen (H₂-TPR) and O₂-TPD. The H₂-TPR profiles of LaVO₄-O_v and LaVO₄ display two peaks at around 485 and 650 ℃ (Fig. 3b), relating respectively to reduction of surface adsorbed oxygen and O_latt. [44]. Besides, the reduction peaks of LaVO₄-O_v appear at the lower temperatures, indicative of its enhanced oxidation ability. Three desorption peaks are observed in the O₂-TPD profile of LaVO₄-O_v (Fig. 3c). The strongest desorption below 300 ℃ links to surface adsorbed oxygen species (O_ads.) [44]. The small peak at around 330 ℃ is ascribed to the desorption of surface lattice oxygen (O_s-latt.), which is reactive and affects catalytic properties [45]. The desorption above 350 ℃ represents the release of bulk lattice oxygen (O_b-latt.) [44,46], reflecting the structural stability. Compared to LaVO₄-O_v, LaVO₄ exhibits a markedly diminished desorption of O_sorb., owing to its absence of O_v that reduces surface O₂ adsorption. Furthermore, no clear desorption of O_s-latt. evolves for LaVO₄ because the lack of O_v makes its O_s-latt. more stable. In addition, the desorption of O_b-latt. of LaVO₄ presents at a lower temperature, revealing the higher structural stability of LaVO₄-O_v. These results confirm LaVO₄-O_v possess high redox ability, rich active sites, and enhanced oxygen storage capability.

To probe the reactive oxygen species and structural changes of the LaVO₄-O_v photocatalyst, in situ EPR measurements were conducted. The EPR signal assigned to O_v-related V⁴⁺ species intensifies regularly according to the prolongation of irradiation time (Fig. 3d), which signifies the enrichment of O_v, due to the release of O_latt. for running the EDH reaction.

In-situ DRIFTS was employed to identify intermediates and elucidate processes of EDH. After adsorbing C₂H₆ in dark, LaVO₄-O_v exhibits five distinct peaks assigned to stretching vibrations of C—H bonds (Fig. S17 in Supporting information) [17,20], suggesting efficient C₂H₆ adsorption by the catalyst. Upon light irradiation, signals of several new species are detected with the intensities heightened by prolonging the photoirradiation (Fig. 3e). The gradually enhanced band at 3680–3058 cm⁻¹ reflects the accumulation of H₂O (Table S1 in Supporting information) [47,48]. The vibrations of C—H (2975 and 2875 cm⁻¹) and CH₂ (2930 cm⁻¹) bonds imply the generation of the key intermediates of CH₃—CH₂· and ·CH₂—CH₂· [17,20,49], which may be generated from dissociation adsorption of C₂H₆ on O_v. The involvements of these key species are also supported by detected vibrations corresponding to CH₃ (1400 and 1380 cm⁻¹) and C-C (1265, 1138, 1118, 1089 and 1031 cm⁻¹) moieties. Notably, observation of featured vibrations of C=O (1763 and 1716 cm⁻¹) and C=C (1650, 1619 and 1545 cm⁻¹) bonds reveals the formation of CO and C₂H₄ [50–52]. In addition, the vibration at 1452 cm⁻¹ points to the evolution of OH species during the reaction [17,48].

Based on the results of photocatalytic EDH, we further performed DFT calculations to explore the thermodynamic favorability of C₂H₆ transformation and O_latt. consumption. The reaction mechanism of alkane dehydrogenation co-steered by O_v and O_latt. has not been thoroughly tracked and deciphered on LaVO₄-O_v. This mysterious mechanism is expected to be decoded on the recyclable LaVO₄ benchmark platform. The LaVO₄-O_v catalyst allows O_v to capture C₂H₆ with a low adsorption energy of −0.66 eV (Fig. 4, from Ⅰ to Ⅱ). The adjoining lattice oxygens are responsible for subsequent bonding with H to generate OH (Fig. 4, from Ⅲ to Ⅳ). The upward trend in free energy until complete desorption of 2OH (Fig. 4, from Ⅴ to Ⅶ), discloses that ethane dehydrogenation to ethene is a challenging process that requires photothermal energy to drive. After that, it is very favorable for the adsorption of O₂ because of a downward trend in energy (Fig. 4, Ⅷ), which can perfectly refill the consumed O_latt. sites after decomposing into two oxygen atoms (Fig. 4, Ⅸ). Such theoretical results are in good agreement with the hao de activity cycling tests. In contrast, LaVO₄ without O_v is inferior in both ethane adsorption and OH desorption (Fig. S18 in Supporting information), demonstrating the effectiveness of more stable intrinsic O_v in EDH, rather than metastable O_v generated by depleting O_latt..

Figure 4

	Download: JPG larger image
Fig. 4. Calculated potential energy diagram and the corresponding optimized geometries of active species over LaVO4-Ov by DFT.

The photocatalytic EDH process is proposed based on the experimental and theoretical results (Fig. 5). Initially, C₂H₆ is preferentially adsorbed on O_v of LaVO₄-O_v (step Ⅰ). Upon light irradiation, the photoinduced holes are trapped by surface O_latt. to produce active O^•− species [21], which enables H abstraction from adsorbed C₂H₆, yielding the C₂H₅/C₂H₄ fractions and adsorbed OH groups (steps Ⅱ and Ⅲ). Concurrently, the electrons reduce V⁵⁺ to form V⁴⁺, which is supported by XPS analyses on used LaVO₄-O_v catalyst (Fig. S19a in Supporting information). Afterward, the C₂H₄ species release from catalyst surface to generate ethene gas (step Ⅳ), while the adsorbed OH groups desorb in form of H₂O (step Ⅴ), rendering the catalyst with decreased O_latt. and increased O_v (Fig. S19b in Supporting information). Due to the strong oxidative ability of photogenerated holes, overoxidation of C₂H₄ also occurs in cooperation of O_latt., resulting in CO release. Lastly, being exposed to O₂, V⁴⁺ cation reduces O₂ to O_latt. positioned in the lattice with itself oxidized to V⁵⁺ (step Ⅵ), achieving a dynamic cycle of O_latt. and catalyst regeneration, which finishes the catalytic loop.

Figure 5

	Download: JPG larger image
Fig. 5. The proposed photocatalytic EDH mechanism over LaVO4-Ov (the black box represents the Ov of fresh LaVO4-Ov, and the red box represents the Ov created from the consumption of Olatt. during the reaction).

In summary, demonstrated herein is the facile synthesis of a LaVO₄-O_v catalyst with O_v and its function in light-driven EDH under mild conditions. The photocatalytic EDH reaction follows a mechanism mediated dynamically by O_latt. and O_v. Bare LaVO₄-O_v grants effective dehydrogenation of undiluted C₂H₆ to produce C₂H₄, together with CO, while Pt-loaded LaVO₄-O_v affords greatly enhanced C₂H₄ production and selectivity. Remarkably, LaVO₄-O_v shows an excellent coking-resistant property and achieves fine reusability via a simplistic regeneration with air exposure to restore the depleted O_latt.; while its O_v promotes optical absorption, ethane adsorption/activation, and charge separation/transport to boost the photocatalytic efficiency. This work may advance the feasibility of photocatalysis technology for EDH with renewable solar resource. Also, the recoverable catalytic-active O_latt. of LaVO₄-O_v is anticipated to extend for driving some oxidant-sensitive organic transformations in a heterogeneous fashion.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

This work is financially supported by the National Key R&D Program of China (Nos. 2021YFA1502100 and 2022YFE0114800) and the National Natural Science Foundation of China (No. 21973014).

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.108313.