Engineering oxygen vacancies on ceria via vanadium oxide dispersion for selective photocatalytic cleavage of lignin C–C bonds

Abstract

Lignin, the most abundant natural aromatic polymer on Earth, offers a promising feedstock for the production of value-added aromatic products through selective C–C bond cleavage. However, achieving both high product yield and selectivity under mild conditions, such as under visible-light irradiation, remains a significant challenge. In this study, we address this issue by employing ceria (CeO₂) as a support to precisely engineer the dispersion of vanadium oxide through two distinct catalyst pretreatment methods: impregnation–calcination (Im-V/CeO₂) and physical mixing–calcination (Pm-V/CeO₂). The results demonstrate that while CeO₂ alone yields 52% and Pm-V/CeO₂ yields 45%, the Im-V/CeO₂ catalyst exhibits superior performance, converting 1,2-diphenylethanol to benzaldehyde with an impressive 89% yield and 98% selectivity. This enhanced performance of Im-V/CeO₂ is attributed to the introduction of low-polymerized vanadium oxide (VO_x) species. Unlike highly polymeric V₂O₅ crystals, these species facilitate the separation of photogenerated electron–hole pairs and reduce charge transfer resistance, thereby improving photocatalytic efficiency. Additionally, the polar nature of VO_x species enhances the activation of lattice oxygen, creating more sites for oxygen adsorption and activation. In-depth mechanistic studies reveal that superoxide radicals play a dominant role in the C–C bond cleavage process. This innovative strategy for achieving high-efficiency photocatalytic C–C bond cleavage through the controlled dispersion of vanadium oxide on a cerium oxide support offers valuable insights for the sustainable valorization of lignin.

---

Green foundation

1. Our work develops cerium oxide nanorod photocatalysts modified with low-polymerized VO_x species via a simple impregnation–calcination method, enabling visible-light-driven valorization of lignin β-1 model compounds and analogues under mild conditions, which proceeds without the need for hazardous solvents, external oxidants, chemical additives, or UV-light irradiation.

2. The anchored low-polymerized VO_x not only enhances the intrinsic oxidative activity of CeO₂ but also extends vanadium-catalyzed reactions from homogeneous to heterogeneous systems, eliminating the need for ligands and reducing cost.

3. The catalyst operates efficiently with a vanadium loading of only 2.6 wt%, achieving 90% conversion of the β-1 model substrate with 98% selectivity toward benzaldehyde. Moreover, it demonstrates broad applicability to vicinal diols, affording high yields of aromatic monomers. These advantages establish the promise of such cerium–vanadium catalysts for biomass valorization and the selective synthesis of aromatic fine chemicals.

Introduction

Modern industry heavily relies on non-renewable fossil fuels (e.g., coal, petroleum, and natural gas) as essential chemical feedstocks and energy sources.¹ However, excessive consumption of these resources has led to a dramatic increase in CO₂ emissions, projected to reach 200 million tons by 2050, and poses a significant environmental burden.² Therefore, replacing fossil fuels with renewable green energy and restructuring the energy structure is critical to achieving the ambitious target of carbon neutrality.³ Biomass, as a sustainable carbon feedstock, offers a viable alternative for producing green chemicals and biofuels.⁴ It is estimated that global biomass production reaches 170 billion tons annually, yet only 3–4% is currently utilized.⁵ Hence, enhancing biomass utilization efficiency is pivotal for advancing green and low carbon development.

Lignin, which accounts for approximately 30% of the weight of lignocellulosic biomass, serves as the largest renewable source of aromatic compounds.⁶ As a highly irregular polymer, lignin mainly consists of three phenylpropanoid subunits—p hydroxyphenyl (H), guaiacyl (G), and syringyl (S)—interconnected by C–O and C–C linkages.^7–9 The efficient valorization of lignin into valuable aromatic chemicals necessitates the cleavage of these bonds while preserving the aromatic rings.^10,11 Notably, over 60% of these interunit linkages are C–O bonds, which are more readily cleaved due to their lower bond dissociation energies (217.6–334.7 kJ mol⁻¹) compared to C–C bonds(288.7–514.6 kJ mol⁻¹).¹² While this reactivity disparity has historically focused research on C–O cleavage, conventional delignification processes often promote condensation reactions that form additional recalcitrant C–C bonds.^13–16 Consequently, the cleavage of these stable C–C linkages becomes an indispensable, yet challenging, prerequisite for efficient lignin utilization. Catalytic lignin oxidation has thus emerged as a promising strategy to address this challenge by targeting the selective cleavage of interunit linkages, particularly the thermodynamically demanding C–C bonds.¹⁷ In current research, the β-O-4 linkages have attracted widespread attention due to their high abundance (50–60%) in lignin and the coexistence of both C–C and C–O bonds within the structure. In addition to β-O-4 linkages, lignin also contains other C–C bond types, such as β-1, β-β, and 5-5 linkages.^18–21 However, owing to their relatively low natural abundance and high bond dissociation energies, strategies for cleaving these C–C bonds remain considerably less explored.

The β-1 linkage, a typical C–C bond connecting phenylpropane units in lignin, has garnered interest despite its low natural abundance (1–15%) due to its potential as a cleavage target for valorization.²² Current thermal catalytic strategies primarily employ homogeneous vanadium (V) or copper (Cu) complexes, which facilitate oxidative cleavage at ca. 100 °C via substrate coordination.^23–25 However, these systems necessitate expensive oxidants, harsh additives, and exhibit high sensitivity to ligand design and reaction conditions, resulting in high operational costs and poor industrial viability.^11,22 Consequently, photocatalysis has emerged as a promising platform for C–C bond cleavage under mild conditions. The mechanism typically involves sequential hydroxyl oxidation, dehydration, and C–C bond cleavage. A critical challenge herein is the overoxidation of the hydroxyl group to a ketone, which significantly increases the C_α–C_β bond strength and severely impedes selective cleavage.^26,27 Although homogeneous vanadium photocatalysts, as reported by Liu et al., achieve a high benzaldehyde yield of 86% by utilizing the strong oxidizing capability of high-valent vanadium and coordination with the hydroxyl group of the substrate to form a photoactive complex, they still suffer from metal leaching and low quantum efficiency.^8,11 Despite the recognized efficacy of homogeneous vanadium complexes in C–C bond cleavage, this knowledge has not been translated into heterogeneous systems; the development of efficient photocatalytic protocols using common vanadium oxide species (e.g., VO_x or V₂O₅) remains largely unexplored.^28,29

In contrast to homogeneous systems, heterogeneous catalysts such as titanium oxide (TiO₂) and graphitic carbon nitride (g-C₃N₄) offer superior structural stability.^30–34 However, they often lack the requisite selectivity to suppress undesired overoxidation. For instance, Hou et al. reported that engineering TiO₂ nanotubes with CuO_x and CeO₂ clusters could achieve near-complete conversion with 98% selectivity toward benzaldehyde through precise control over the distribution of CuO_x species.⁶ Nevertheless, this approach requires precise control over support pretreatment and metal loading ratios, which imparts substantial process complexity and cost, thereby hindering scalability. Therefore, a compelling need persists for the development of simple, low-cost, and scalable synthetic strategies to fabricate heterogeneous photocatalysts that simultaneously deliver high selectivity and efficiency for β-1 C–C bond cleavage.

Herein, we report a CeO₂-based photocatalyst (Im-V/CeO₂) decorated with low-polymerized vanadium oxide (VO_x) species and enriched with oxygen vacancies (O_v), synthesized via a simple impregnation–calcination method. This catalyst enables highly selective photocatalytic oxidative cleavage of the C–C bond in 1,2-diphenylethanol to benzaldehyde at room temperature. To elucidate the structure-performance relationship, systematic characterization reveals that the pretreatment method engineers the dispersion state of vanadium oxide species from low-polymerized to crystalline forms, and enables detailed structural analysis. The significantly enhanced catalytic activity is primarily due to an increased oxygen vacancy concentration, resulting from the stronger interaction between low-polymerized VO_x (with highly polar V=O bonds) and the CeO₂ support compared to crystalline V₂O₅. This electronic modulation further improves charge separation and reduces charge transfer resistance, as supported by photoelectrochemical measurements. Moreover, mechanistic studies using electron paramagnetic resonance (EPR), radical trapping, and substrate expansion have successfully identified the superoxide radical (˙O₂⁻) as the key active species responsible for C_α–C_β bond cleavage. Based on the reaction product distribution and previous mechanistic reports, a coherent reaction mechanism is proposed. We believe that the present work can provide insights for the design of simpler and more rational photocatalytic systems, particularly through surface modification engineering, since current strategies in this field predominantly rely on the construction of heterojunctions.

Experimental

Chemicals

1,2-Diphenylethanol (Leyan 98%), 1,2-diphenylethanone (Aladdin 98%), benzaldehyde (Macklin 98.5%), benzoic acid (Aladdin 99.9%), 1,2-diphenylethane (Macklin 99%), 1-phenyl-1,2-ethanediol (Macklin 98%), 1,2-diphenyl-1,2-etanediol (Adamas 98%), benzoin (Macklin LR), 2-phenoxy-1-phenylethanol (Energy 98%), 1,4-benzoquinone (p-BQ Macklin 99%), ammonium oxalate ((NH₄)₂C₂O₄ Aladdin 98%), potassium persulfate (KPS Macklin 99.5%), N-dodecane (N-C₁₂H₂₆ Macklin ≥99%), NaOH (Hushi AR), Ce(NO₃)₃·6H₂O (Macklin 99.9%), NH₄VO₃ (Macklin 99%), La₂O₃ (Macklin 99.99%), La(NO₃)₃·6H₂O (Macklin 99%), (NH₄)₆H₂W₁₂O₄₀·H₂O (Macklin 99.5%), Fe(NO₃)₃·9H₂O (Macklin 98.5%), Al(NO₃)₃·3H₂O (Aladdin AR), Acetonitrile (CH₃CN Hushi 99.9%), γ-Al₂O₃ (Macklin 99.99%), Fe₂O₃ (Heowns 98%), WO₃ (Macklin 99.9%), Na₃PO₄ (Macklin 98%). All chemicals were used as received without further purification. Deionized water was used throughout the experiments.

Synthesis of CeO₂ with different morphologies

Rod-shaped (CeO₂-R) and cubic (CeO₂-C) CeO₂ nanostructures were synthesized via a hydrothermal method.³⁵ Briefly, 3.472 g of Ce(NO₃)₃·6H₂O and 38.4 g of NaOH were dissolved in 20 mL and 140 mL of deionized water, respectively, under vigorous stirring. The Ce(NO₃)₃·6H₂O solution was then added dropwise to the NaOH solution under continuous stirring at room temperature for 30 min. The resulting mixture was transferred to a Teflon-lined stainless-steel autoclave and subjected to hydrothermal treatment at 100 °C (for CeO₂-R) or 180 °C (for CeO₂-C) for 24 h. The solid product was collected by centrifugation, thoroughly washed with deionized water, and dried overnight at 80 °C. The final CeO₂ powders were obtained by calcination in air at 600 °C for 2 h.

To prepare the octahedral (CeO₂-O) CeO₂ nanostructure, 1.716 g of Ce(NO₃)₃·6H₂O and 0.015 g of Na₃PO₄ were dissolved in 20 mL and 140 mL of deionized water, respectively, under vigorous stirring. The Ce(NO₃)₃·6H₂O solution was added dropwise to the Na₃PO₄ solution and stirred at room temperature for 30 min. The resulting suspension was then transferred to a Teflon lined autoclave and heated at 170 °C for 12 h. The product was recovered by centrifugation, washed with deionized water, dried at 80 °C, and subsequently calcined in air at 600 °C for 2 h, following the same protocol used for CeO₂-R and CeO₂-C.

Synthesis of Im-Vx/CeO₂

Im-Vx/CeO₂ catalysts were prepared using ammonium metavanadate (NH₄VO₃) as the vanadium precursor via a conventional impregnation–calcination (IMC) method.³⁶ In a typical procedure, 50 mg of pre-synthesized rod-like CeO₂ and 5.8 mg of NH₄VO₃ were dispersed in 25 mL of deionized water under ultrasonication for 1 h. The resulting suspension was then stirred overnight at room temperature on a heating/stirring plate. On the following day, the mixture was further heated and stirred at 60 °C for 6 h. After impregnation, excess water was removed by vacuum filtration, and the collected solid was dried overnight at 70 °C. The dried sample was then calcined in air at 500 °C for 5 h in a muffle furnace and allowed to cool naturally to room temperature. The final catalyst was denoted as Im-V/CeO₂, where x represents the actual vanadium content determined by ICP-OES. The theoretical vanadium loading from 5.8 mg NH₄VO₃ corresponds to 5%, while the measured vanadium content was 2.58%.

Synthesis of Pm-V/CeO₂

Pm-V/CeO₂ catalysts were synthesized using a physical mixing–calcination (PMC) approach, with ammonium metavanadate (NH₄VO₃) as the vanadium precursor.³⁷ Based on the optimal vanadium loading determined from the IMC method, 300 mg of rod-like CeO₂ and 18 mg of NH₄VO₃ were thoroughly ground together in an agate mortar for 10 min to ensure homogeneous mixing. The resulting mixture was then calcined in a muffle furnace under air at 500 °C for 5 h and allowed to cool naturally to room temperature. The resulting catalyst was denoted as Pm V/CeO₂. The theoretical vanadium loading was 2.58%, while the actual vanadium content determined by ICP-OES was 2.49%, which is comparable to that of the Im V/CeO₂ sample. For comparison, V₂O₅ is also prepared by direct calcination of ammonium metavanadate under static air at the same temperature.

Synthesis of Im-M/CeO₂

A series of metal-modified CeO₂ catalysts, denoted as Im-M/CeO₂ (M = W, Al, Fe, La), were synthesized via a conventional impregnation–calcination (IMC) method. Ammonium metatungstate, aluminum nitrate nonahydrate, ferric nitrate nonahydrate, and lanthanum nitrate hexahydrate were used as the corresponding metal precursors. Taking Im-W/CeO₂ as an example: 50 mg of rod-like CeO₂ (previously synthesized) was dispersed in 25 mL deionized water along with 40.2 mg of (NH₄)₆H₂W₁₂O₄₀·H₂O, corresponding to a theoretical W loading of 5%. The mixture was subjected to ultrasonication for 1 h, followed by magnetic stirring overnight. On the following day, the suspension was stirred at 60 °C for an additional 6 h. After impregnation, excess water was removed by vacuum filtration, and the obtained solid was dried at 70 °C overnight. The dried sample was then calcined in air at 500 °C for 5 h and allowed to cool naturally to room temperature, affording the final catalyst Im-W/CeO₂. Using the same procedure, Im-Al/CeO₂, Im-Fe/CeO₂ and Im-La/CeO₂ were synthesized by replacing (NH₄)₆H₂W₁₂O₄₀·H₂O with 34.8 mg of Al(NO₃)₃·9H₂O, 18.1 mg of Fe(NO₃)₃·6H₂O, and 8 mg of La(NO₃)₃·6H₂O, respectively.

Characterization

Transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) images were obtained using a JEOL JEM-2100EX microscope operated at an accelerating voltage of 200 kV. Powder X-ray diffraction (XRD) patterns were recorded by Ultima IV X-ray diffractometer using Cu Kα radiation (λ = 1.5418 Å), with a scanning range of 10–80° and a scan rate of 10° min⁻¹. The Raman spectra were acquired on a LabRAM HR Evolution system under ambient temperature using the 532 nm excitation line of an Ar + laser (Spectra Physics). The Fourier-transform infrared (FT-IR) spectra were measured using a Bruker VERTEX 80 Infrared spectrometer (Germany) in the range of 4000–500 cm⁻¹. X-ray photoelectron spectroscopy (XPS) and valence band XPS (VB-XPS) were carried out on a Thermo Scientific ESCALAB 250Xi spectrometer using an Al Kα source (hν = 1486.6 eV, work function Φ = 4.2 eV). In situ XPS measurements were performed in the same instrument. The procedure involved first collecting a spectrum after the sample was kept under dark conditions with continuous O₂ flow for 15 min, and then acquiring a second spectral point after a subsequent 15-minute irradiation with 455 nm blue light. All binding energies were calibrated using the C 1s peak at 284.8 eV. The XPS spectra were deconvoluted using the Avantage software (version 5.99), and the peaks corresponding to V, O and Ce were fitted with a Gaussian–Lorentzian species mixed function. The vanadium content in the catalysts was quantified by inductively coupled plasma atomic emission spectroscopy (ICP-OES 720, Varian). The Brunauer–Emmett–Teller (BET) surface areas and porosity were determined using N₂ adsorption–desorption isotherms at −196 °C on a Micromeritics TriStar 2460 analyzer. Oxygen temperature-programmed desorption (O₂-TPD) was performed on a JW-HX100 instrument, where 100 mg of the catalyst was first heated at 300 °C for 1 h in a 3 vol% O₂/Ar flow (30 mL min⁻¹), cooled to 50 °C, and then purged under Ar for 1 h. The pretreated sample was subsequently heated from 50 to 700 °C at a ramp rate of 10 °C min⁻¹ in an Ar stream, with the O₂ concentration in the effluent monitored by a thermal conductivity detector (TCD). The UV-vis diffuse reflectance spectra (UV-vis DRS) were recorded on a PerkinElmer Lambda 950 spectrophotometer in the wavelength range of 200–800 nm. The photoluminescence (PL) and time-resolved photoluminescence (TRPL) spectra were measured on a Hitachi F-4700 FL spectrofluorometer and an FLS 980 instrument (Japan), respectively, with an excitation wavelength of 330 nm. Electron paramagnetic resonance (EPR) measurements were conducted on a Bruker A300 spectrometer. Electrochemical impedance spectroscopy (EIS) and transient photocurrent measurements were performed in a three-electrode system using a CHI-760E electrochemical workstation (China). The working electrode was prepared by drop-casting 15 μL of a suspension (5 mg catalyst dispersed in 1 mL anhydrous ethanol and 50 μL 5 wt% Nafion solution) onto pretreated fluorine-doped tin oxide (FTO) glass, followed by drying at 60 °C in air. Ag/AgCl (saturated KCl) and a Pt foil (1 × 1 cm²) were used as the reference and counter electrode, respectively. The electrolyte was a 0.2 mol L⁻¹ Na₂SO₄ electrolyte solution (pH ≈ 6.8). The EIS measurements were conducted in the frequency range of 10⁻¹–10⁵ Hz. Transient photocurrent responses were recorded using a 300 W Xe lamp (PLS-SXE300+, Perfectlight, Beijing) as the light source.

Photocatalytic measurements

Photocatalytic aerobic oxidation reactions were carried out in a 10 mL side-arm quartz Schlenk tube. Typically, 0.05 mmol of 1,2-diphenylethanol (substrate 1) and 10 mg of photocatalyst were added to 3 mL of acetonitrile in the quartz reactor. The system was purged with O₂ for 1 min and sealed with a rubber septum and Parafilm. After purging, the mixture was ultrasonicated to ensure uniform dispersion of the substrate and catalyst in the solvent. The reaction was then initiated under irradiation with a 30 W blue LED lamp (λ = 455 nm). After each cycle, the photocatalyst was recovered by centrifugation, washed thoroughly with acetonitrile and anhydrous ethanol, and reused in subsequent runs. Control experiments were performed by adding 0.05 mmol of specific scavengers (e.g., p-BQ, (NH₄)₂C₂O₄, and KPS) to the reaction mixture to investigate the role of reactive species in the photocatalytic process. Safety note! Proper protective eyewear must be worn during the reaction to prevent eye injury from 455 nm blue light exposure.

Product analysis

The conversion of 1,2-diphenylethanol (substrate 1) and its analogues, as well as the yields of the corresponding products, were quantified by gas chromatography (GC) and qualitatively confirmed by gas chromatography-mass spectrometry (GC-MS). Specifically, after the reaction, a fixed amount of N-dodecane (internal standard) was added to the reaction mixture. The solution was ultrasonicated to ensure complete dissolution and then filtered through a 0.22 μm organic membrane filter prior to analysis.

GC quantification analysis

Analyses were performed on a Shimadzu GC-2014 plus AFSC system using nitrogen (N₂) as the carrier gas. The GC was equipped with an SH-Rxi-5SilMS capillary column. The oven temperature was initially held at 50 °C, then ramped to 280 °C at a rate of 20 °C min⁻¹ and held for 3 min.

GC-MS qualitative analysis

Analyses were performed on a Shimadzu GCMS-QP2020NX W/O RP, with helium (He) as the carrier gas and an Rtx-5MS column. The temperature program was identical to that of GC: initial temperature 50 °C, ramped to 280 °C at 20 °C min⁻¹, and held for 3 min. The retention times of the analytes were as follows (in order of elution): benzaldehyde (1a), benzoic acid (1c), N-dodecane (internal standard), 1,2-diphenylethane (1d), 1,2-diphenylethanol (1), and 1,2-diphenylethanone (1b).

The conversion (X) of substrate 1 and the yields (Y) of products were calculated based on the following equations:

X₁ (%) = (n_{1 reacted} × 100%)/n_{1 added}

Y_1a(1c) (%) = (n_{1a(1c) produced} × 100%)/(n_{1 added} × 2)

Y_1b(1d) (%) = (n_{1b(1d) produced} × 100%)/n_{1 added}

where n is the amount of substance (mol).

Results and discussion

Catalyst characterizations

Rod-shaped CeO₂ catalysts with varying vanadium loadings were synthesized through a simple impregnation–calcination method and denoted as Im-Vx/CeO₂, where x represents the actual vanadium content determined by inductively coupled plasma optical emission spectrometry (ICP-OES). The sample with an optimal loading of 2.58% is referred to as Im-V/CeO₂. To investigate the influence of preparation methods on vanadium distribution, a control catalyst Pm-V/CeO₂ with similar V loading (2.49%) was prepared by physical mixing NH₄VO₃ with rod-shaped CeO₂, followed by grinding and calcination (Fig. 1a and Table S1). X-ray photoelectron spectroscopy (XPS) analysis revealed that the surface vanadium content was significantly higher on Im-V/CeO₂ (5.96%) compared to Pm-V/CeO₂ (4.33%) (Table S2). Given their comparable bulk V loadings, this pronounced surface enrichment in the impregnated catalyst demonstrates that the wet impregnation method promotes selective vanadium deposition on the ceria surfaces.²

Fig. 1 (a) Schematic diagram of the preparation of Im-Vx/CeO₂ and Pm-Vx/CeO₂. Representative (b and c) HRTEM and (d) HAADF-STEM images, together with the corresponding elemental mappings (i.e., V, Ce, and O) of Im-V/CeO₂. Representative (e and f) HRTEM and (g) HAADF-STEM images, together with the corresponding elemental mappings (i.e., V, Ce, and O) of Pm-V/CeO₂.

To visually characterize the dispersion state of vanadium on the CeO₂ surface, high-resolution transmission electron microscopy (HRTEM) was conducted. As shown in Fig. 1b, c and Fig. S1, the representative HRTEM images of Im-V/CeO₂ confirmed that the rod-like morphology was retained after impregnation and calcination. Meanwhile, the observed lattice spacing of 3.12 Å corresponds to the CeO₂ (111) plane. Notably, the absence of V₂O₅ lattice fringes—which aligns with the XRD evidence indicating the absence of crystalline vanadia phases—indicates that vanadium species exist in a highly dispersed or amorphous state.³⁸ Furthermore, the high-angle annular dark-field scanning TEM (HAADF-STEM) coupled with the corresponding energy dispersive X-ray (EDX) element mappings (Fig. 1d) further verified the uniform distribution of V, Ce, and O throughout Im-V/CeO₂. For Pm-V/CeO₂, HRTEM images (Fig. 1e and f) also showed a preserved rod-like structure. Lattice spacings of 3.12 Å and 1.56 Å correspond to the CeO₂ (111) and (222) planes, respectively, while a 2.76 Å lattice fringe—matching the V₂O₅ (011) plane—confirms the presence of crystalline V₂O₅.³⁷ Meanwhile, HAADF-STEM and EDX mappings (Fig. 1g) demonstrate that V, Ce, and O were homogeneously distributed in Pm-V/CeO₂ as well.

X-ray diffraction (XRD) analysis was conducted to confirm the findings from TEM observations. As shown in Fig. 2a, all CeO₂ based catalysts exhibited characteristic fluorite CeO₂ diffraction peaks at 28.6°, 33.1°, 47.5°, 56.3°, 59.1°, 69.4°, 76.7°, and 79.1° (PDF #34-0394).⁶ Notably, the Im-V/CeO₂ catalyst does not show any diffraction peaks corresponding to crystalline V₂O₅. This absence, supported by TEM and EDX mappings (Fig. S2a), demonstrates that vanadium species are highly dispersed on the CeO₂ surface without forming crystalline V₂O₅. In contrast, the Pm-V/CeO₂ sample displayed additional peaks at 15.3°, 20.3°, 26.1°, 31.0°, 32.4°, and 34.3° (Fig. S3a), corresponding to orthorhombic V₂O₅ (PDF #41-1426).³⁹ The peak at 32.4° matches the (011) plane of V₂O₅ observed in TEM, confirming the formation of crystalline V₂O₅. These results indicate that the nature of vanadium species varies significantly depending on the precursor treatment: physical mixing and calcination yield detectable crystalline V₂O₅, whereas impregnation-calcination results in vanadium being in an amorphous or super dispersed form.

Fig. 2 (a) XRD patterns, (b) Raman spectra, (c) FT-IR spectra, (d) Ce 3d XPS spectra, (e) V 2p XPS spectra, (f) O 1s XPS spectra of CeO₂, Im-V/CeO₂, and Pm-V/CeO₂.

To further elucidate the nature of vanadium species introduced by impregnation, Raman spectroscopy was performed (Fig. 2b). All CeO₂-based catalysts exhibit the characteristic F 2g vibration mode of fluorite CeO₂ at 460 cm⁻¹. In the case of Im-V/CeO₂, this band remains at 460 cm⁻¹, but its intensity is significantly reduced compared to the pristine CeO₂, indicating lattice distortion or defect formation upon vanadium incorporation (Fig. S2b).^40,41 For Pm-V/CeO₂, the Raman spectrum shows additional modes corresponding to crystalline V₂O₅, in addition to the CeO₂ F2 g vibration mode, confirming the formation of V₂O₅via physical mixing and calcination.^42,43 The distinct Raman signatures between Im-V/CeO₂ and Pm-V/CeO₂ further confirm the non-crystalline nature of vanadium species in Im-V/CeO₂. These findings align with the XRD results, ruling out the presence of crystalline V₂O₅ in Im-V/CeO₂.

To further clarify the structural origin of the amorphous vanadium species in Im-V/CeO₂, as summarized in Table S3 and illustrated in Fig. S4, N₂ adsorption–desorption isotherms, along with corresponding Brunner–Emmett–Teller (BET) surface areas and pore size distributions, were evaluated for all catalysts. All samples exhibit typical type IV isotherms with an H3 hysteresis loop, indicative of mesoporous structure.⁴⁴ Compared to the pristine CeO₂ support, Im-V/CeO₂ shows a 30% decrease in BET surface area, suggesting that the impregnation process facilitates more effective ionic deposition of vanadium species within the CeO₂ mesopores, partially blocking pore channels and reducing accessible surface area. In contrast, Pm-V/CeO₂ demonstrates negligible changes in surface area and pore size, indicating minimal structural alteration by physical mixing. Therefore, the impregnation process, unlike physical mixing, causes a significant reconstruction of the CeO₂ structure, which directly dictates the resulting dispersion morphology of the vanadium species. The surface vanadium atomic density was calculated to be 5.5 V nm⁻² (surface density = number of surface V atoms/BET surface area), which aligns with previously reported vanadium surface densities (4.2–7.7 V nm⁻²) for VO_x species loaded on the surface of TiO₂ and Al₂O₃.^44–46

To directly distinguish between the VO_x and V₂O₅ crystalline phases, Fourier-transform infrared (FT-IR) spectroscopy was performed (Fig. 2c and S2c). The broad band at 3428 cm⁻¹ corresponds to surface hydroxyl (–OH) vibrations,⁴⁵ while the peak at 1623 cm⁻¹ corresponds to Ce–O stretching of the fluorite CeO₂ lattice.⁴⁷ According to literature reports, the bands at 1044 cm⁻¹ and 977 cm⁻¹ are assigned to V=O terminal stretching vibrations (characteristic of isolated or low-polymerized VO_x) and V–O–V bridging vibrations (typical of crystalline V₂O₅), respectively.^48,49 The Im-V/CeO₂ sample exhibits a single, intense V=O band at 1044 cm⁻¹ without observable V–O–V peaks, indicating the predominant formation of highly dispersed VO_x species. In contrast, Pm-V/CeO₂ displays both the 1044 cm⁻¹ (V=O) and 977 cm⁻¹ (V–O–V) bands, whose intensity ratio and peak positions mirror the spectral features of pristine crystalline V₂O₅ (Fig. S3b). These observations align with Raman results, confirming the absence of crystalline V₂O₅ peaks in Im-V/CeO₂ and demonstrating that physical mixing leads to crystalline V₂O₅ aggregates, while impregnation promotes the formation of isolated or low-polymerized VO_x species.

To further clarify how the distinct vanadium species (dispersed VO_xvs. crystalline V₂O₅) interact with the CeO₂ support and potentially govern their catalytic behavior, X-ray photoelectron spectroscopy (XPS) and oxygen temperature-programmed desorption (O₂-TPD), supplemented by in situ XPS, were performed. The Ce 3d spectrum of Im-V/CeO₂ (Fig. 2d) reveals ten peaks, which could be assigned to Ce⁴⁺ (882.4, 889.0, 898.4, 901.1, 907.7, and 916.7 eV) and Ce³⁺ (880.5, 885.2, 898.8, and 903.8 eV), respectively.⁵⁰ Compared to pristine CeO₂, the Im-V/CeO₂ catalyst exhibits a positive shift in Ce 3d binding energy, indicating electron depletion from the CeO₂ lattice. Conversely, its V 2p spectrum (Fig. 2e and S5) shows a negative shift relative to pristine V₂O₅, suggesting electron transfer from CeO₂ to vanadium species.^51–53 While a similar trend was observed for Pm-V/CeO₂, the interaction strength between the vanadium species and the CeO₂ support differs significantly. In the case of Im-V/CeO₂, the strong interaction between VO_x and CeO₂, facilitated by the strongly polarized V=O bond, results in a marked decrease in surface Ce³⁺ content. This is accompanied by an increased V⁴⁺ fraction and the largest positive shift in lattice oxygen binding energy, indicative of promoted lattice oxygen activation and oxygen vacancy generation (Fig. 2d and f).⁵⁴ In contrast, Pm-V/CeO₂ shows negligible changes in Ce³⁺ content, V⁴⁺ fraction, and lattice oxygen binding energy, confirming a weaker interaction between the physically mixed V₂O₅ and CeO₂. These differences in interaction strength directly influence catalytic behavior: the strong VO_x–CeO₂ coupling enhances lattice oxygen activation and facilitates the generation of more oxygen vacancies, while the weak V₂O₅–CeO₂ interaction leads to inferior catalytic performance. This conclusion is corroborated by our subsequent O₂-TPD results.

Building on the O₂-TPD results, we perform in situ XPS under an O₂ atmosphere to explicitly probe the dynamic evolution involving low-polymerized VO_x species, lattice oxygen (O_L), and oxygen vacancies (O_v). As shown in the in situ XPS spectra (Fig. 3a and c), the Ce 3d peak shifts downward by 0.6 eV, while the lattice oxygen (O_L) peak shifts upward by 0.4 eV compared to normal conditions under an O₂ atmosphere in the dark. These opposing shifts indicate the electron redistribution from O_L to cerium, which leads to an increase in Ce³⁺ and O_v concentration by the O 1s spectrum.⁵⁵ Meanwhile, the V 2p XPS spectrum (Fig. 3b and S7) shows a decrease in V⁴⁺ content from 63% to 58%, suggesting an increase in the average vanadium oxidation state of vanadium in the surface VO_x species. Compared with the pristine support (Fig. S6 and Table S4), the introduction of VO_x species results in a marked weakening of the adsorbed oxygen desorption peak at around 130 °C, with the amount of desorbed O₂ drastically reduced from 0.321 mmol g⁻¹ to 0.063 mmol g⁻¹, accompanied by a shift of the peak toward higher temperature. species.^54,56 This indicates effective activation and conversion of adsorbed oxygen species, while the VO_x sites are oxidized by the adsorbed oxygen, resulting in an overall higher oxidation state. Concurrently, the lattice oxygen desorption peak at high temperatures (ca. 650 °C) intensifies, with the amount of desorbed O₂ increasing from 0.009 mmol g⁻¹ for the pristine support to 0.076 mmol g⁻¹, accompanied by a shift to a lower temperature. This confirms the facilitated activation and migration of lattice oxygen, which is consistent with the increased oxygen desorption amount and provides a channel for oxygen vacancy formation.⁵⁷ Furthermore, light irradiation leads to a further decrease in V⁴⁺ content with a concomitant increase in the proportion of O_v.

Fig. 3 In situ XPS spectra of (a) Ce 3d, (b) V 2p, and (c) O 1s for the Im-V/CeO₂ catalyst under dark and light irradiation conditions in an O₂ atmosphere.

In summary, O₂-TPD and in situ XPS collectively reveal a dynamic cycle: low-polymerized VO_x species adsorb and activate O₂ at the interface, leading to their oxidation (decreased V⁴⁺), which in turn triggers lattice oxygen activation, electron transfer to Ce (increased Ce³⁺), and oxygen vacancy formation. Light irradiation markedly enhances this entire process via photogenerated charge carriers, thereby stabilizing additional oxygen vacancies. This mechanism accounts for the superior oxidative activity of the catalyst under simultaneous O₂ and light conditions.

Photoelectrochemical performance

To analyze the optical properties and charge separation efficiency of the catalysts, a series of measurements was conducted, including UV-Vis diffuse reflectance spectroscopy (UV-vis DRS), valence band X-ray photoelectron spectroscopy (VB-XPS), and time-resolved photoluminescence (TR-PL) analysis. UV-vis DRS was performed on CeO₂, Im-V/CeO₂, and Pm-V/CeO₂ (Fig. 4a). Im-V/CeO₂ exhibits the strongest absorption at 455 nm compared to CeO₂ and Pm-V/CeO₂, indicating its superior ability to absorb visible light compared to the other catalysts.

Fig. 4 (a) UV-Vis DRS, (b) Tauc plots, (c) VB-XPS, (d) bandgap diagrams, (e) transient photocurrent responses, and (f) PL spectra of CeO₂, Im-V/CeO₂, and Pm-V/CeO₂.

The bandgaps (E_g) were determined using Tauc plots (Fig. 4b), revealing values of 2.93 eV for CeO₂, 2.47 eV for Im-V/CeO₂, and 2.60 eV for Pm-V/CeO₂. These findings suggest that Im-V/CeO₂ has a narrower band gap, enhancing its potential for visible-light-driven catalytic applications. VB-XPS was employed to determine the valence band positions (E_VB-XPS) of the catalysts, which were 1.91, 2.27, and 2.36 eV for CeO₂, Im-V/CeO₂, and Pm-V/CeO₂, respectively (Fig. 4c). These values were converted to the Normal Hydrogen Electrode (NHE) scale using E_VB(NHE) = E_VB-XPS + Φ − 4.44 (the instrument work function (Φ) is calibrated to 4.2 eV).⁵⁸ The conduction band potentials (E_CB) were further calculated according to E_VB = E_CB + E_g. Therefore, the energy band structures of the catalysts were shown in Fig. 4d. Both Im-V/CeO₂ and Pm-V/CeO₂ exhibited positive shifts in their conduction band potentials relative to CeO₂, indicating their capability to activate O₂ for catalytic reactions. Additionally, the positive shifts in their valence band potentials enhanced their oxidation ability, favoring substrate conversion.^6,11

To evaluate the efficiency of the separation of photogenerated electrons and holes, transient photocurrent responses, electrochemical impedance spectroscopy (EIS), steady-state photoluminescence (PL), and time-resolved photoluminescence (TR-PL) were performed. As shown in Fig. 4e, Im-V/CeO₂ exhibited the highest photocurrent density under visible light irradiation, demonstrating superior electron–hole separation efficiency. The EIS Nyquist plots revealed the smallest semicircle for Im-V/CeO₂, signifying the lowest interfacial charge-transfer resistance, whereas Pm-V/CeO₂ presented the largest resistance (Fig. S8a). As shown in the steady-state PL spectra, Im-V/CeO₂ displayed the lowest emission intensity compared to CeO₂ and Pm-V/CeO₂ (Fig. 4f), suggesting markedly enhanced photogenerated carrier transfer efficiency and suppressed electron–hole recombination in Im-V/CeO₂ photocatalyst, likely due to the presence of VO_x species.⁵⁹ Further evidenced by TR-PL analysis (Fig. S8b), Im-V/CeO₂ (1.80 ns) displayed the shortest average lifetime decay compared to CeO₂ (1.98 ns) and Pm-V/CeO₂ (3.67 ns), further confirming the highest non-radiative decay transition rate and fastest charge transfer in Im-V/CeO₂. Collectively, these results demonstrate that the strongly polarized V=O bonds in VO_x species enhance the preferential electron transfer from CeO₂ to VO_x, thereby improving the separation efficiency of photogenerated electrons and holes. These findings align with the XPS results, which showed more significant shifts in the Ce 3d and V 2p binding energies for Im-V/CeO₂, indicating a more robust interaction between VO_x and CeO₂ compared to the physical mixing approach.

Photocatalytic performance

To evaluate the photocatalytic performance of the catalysts, experiments were conducted in a standard reactor using gas chromatography (GC) with an internal standard for quantification (Fig. S9). Initial tests focused on comparing different CeO₂ morphologies (e.g., rod-shaped, cubic, and octahedral) for the oxidation of 1,2-diphenylethanol (1) to benzaldehyde (1a) (Fig. S10a). Results showed that rod-shaped CeO₂ achieved 57% conversion of 1 and 52% yield of 1a, significantly outperforming cubic (32% conversion) and octahedral (11% conversion) counterparts, while all materials maintained >90% selectivity for 1a. Therefore, rod-shaped CeO₂ was chosen for further optimization.

Under identical conditions, Im-V/CeO₂ exhibited superior performance, achieving an 89% yield of 1a. In comparison, Pm-V/CeO₂ achieved only 45%, and bare CeO₂ delivered 52% (Fig. 5a and S11), demonstrating that wet impregnation greatly increases active-site density. Moreover, the performance varied with V loading and catalyst amount. Yield increased up to a V loading of 2.58% but plateaued beyond that (Fig. 5b). The optimal yield was achieved at 5 hours of reaction time, after which slight over-oxidation to benzoic acid occurred (Fig. 5c). Similarly, conversion improved with increasing catalyst loading up to 10 mg (Fig. 5d), though yield plateaued likely due to light attenuation. The reaction atmosphere played a critical role, with O₂ providing the highest conversion and yield, followed by air, and N₂ showing a marked performance decline (Fig. 5e and S12). This observation underscores the critical role of O₂ in facilitating C–C bond cleavage and further suggests highlights the essential role of O₂ in facilitating C–C bond cleavage and suggests distinct mechanistic pathways under O₂versus N₂ atmospheres.⁸ Further investigations with other metal precursors (e.g., W, Al, Fe, and La) under optimal conditions showed negligible activity (Fig. S10b and c), reinforcing VO_x as the key active species driving the catalytic performance. Furthermore, the remarkable enhancement over previously reported catalysts in both benzaldehyde selectivity and formation rate highlighted the efficacy of our catalyst design strategy (Table S5). Beyond activity and selectivity, we further assess the green credentials of the Im-V/CeO₂ system by evaluating its E-factor and atom economy.⁶⁰ Specifically, we further quantify the mass of waste generated during the reaction and determined that the Im-V/CeO₂ system exhibits an E-factor of only 0.012 in the oxidative C–C bond cleavage of 1,2-diphenylethanol. As summarized in Fig. S14 and Table S6, this value is significantly lower than those reported for conventional photocatalytic systems, highlighting the superior environmental compatibility of our catalyst. To better assess the green and sustainable nature of this system, we also analyse its atom economy. As illustrated in Fig. S15, suppressing the over-oxidation of benzaldehyde to benzoic acid increases the atom economy from 30% to 47.7%. Furthermore, by inhibiting the initial oxidation step that leads to 1,2-diphenylethanone, the atom economy rises substantially to 92.2%, which aligns well with our initial catalyst optimization strategy. Finally, the stability of Im-V/CeO₂ was demonstrated through five consecutive stability tests, during which the catalyst maintained near-constant conversion and yield. Post-reaction XRD and FT-IR characterizations confirmed no structural changes (Fig. 5f and S13), thereby showcasing the excellent stability of the catalyst.

Fig. 5 Optimization of (a) pretreatment, (b) V loading, (c) reaction time, (d) catalyst dosage, and (e) reaction atmosphere for the Im-V/CeO₂ catalyst. Reaction conditions: 10 mg of substrate 1 (0.05 mmol), 3 ml of acetonitrile, 10 mg of catalyst, O₂ atmosphere, 455 nm Blue LED, 5 h. (f) Stability test of Im-V/CeO₂. Reaction conditions: 10 mg of substrate 1 (0.05 mmol), 3 ml of acetonitrile, 10 mg of Im-V/CeO₂, O₂ atmosphere, 455 nm Blue LED, 1 h. The yields were determined by GC with N-dodecane as the internal standard. X and Y represent the conversion and yield, respectively.

Mechanism investigation

To elucidate the mechanism of photocatalytic C–C bond cleavage in lignin β-1 model compounds to produce benzaldehyde, radical scavenging and substrate extension experiments were performed (Fig. 6). To first establish that the reaction is indeed photocatalytic, we performed a set of control experiments (Fig. S16). The results confirm that the oxidative C–C bond cleavage proceeds only in the simultaneous presence of the Im-V/CeO₂ catalyst, light irradiation, and the 1,2-diphenylethanol. Having confirmed the photocatalytic nature, we seek to identify the key active species involved. Given that previous studies have shown that photogenerated electrons reduce O₂ to form superoxide radicals (˙O₂⁻), which are believed to play a critical role in the cleavage process.⁶¹

Fig. 6 (a) Scavenger screening for the photocatalytic oxidation of 1,2-diphenylethanol, using p-BQ (˙O₂⁻), (NH₄)₂C₂O₄ (h⁺), and KPS (e⁻) as trapping agents. Reaction conditions: 10 mg of substrate 1 (0.05 mmol), additive (0.05 mmol), 3 ml of acetonitrile, 10 mg of Im-V/CeO₂, O₂ atmosphere, 455 nm Blue LED, 5 h. (b) Quantification of oxygen vacancies by EPR in CeO₂, Im-V/CeO₂, and Pm-V/CeO₂. (c) In situ EPR detection of light-induced ˙O₂⁻ radicals (DMPO-trapped) on the same catalysts under O₂ atmosphere upon 10 min of irradiation. (d) In situ EPR detection of DMPO-˙C_β adducts in Im-V/CeO₂ under dark and light-irradiated conditions in O₂ atmosphere. (e) Reactions of several analogues of 1 under visible light. Reaction conditions: substrate 1b–5 (0.05 mmol), 3 ml of acetonitrile, 10 mg of Im-V/CeO₂, O₂ atmosphere, 455 nm Blue LED, 5 h. The yields were determined by GC with N-dodecane as the internal standard.

When 1 equivalent of p-benzoquinone (p-BQ), a superoxide radical scavenger, was added, the yield of benzaldehyde (1a) decreased to 7%, while the formation of 1,2-diphenylethanone (1b) increased to 55%. This observation confirms that ˙O₂⁻ is the key species responsible for C–C bond cleavage and highlights the competitive interplay between hydroxyl oxidation and C–C bond scission.^6,11 Furthermore, when 1 equiv. of ammonium oxalate ((NH₄)₂C₂O₄) or potassium persulfate (KPS) was used to trap holes or electrons, respectively, the yield of benzaldehyde decreased to 9% and 26%. This demonstrates that both photogenerated holes and electrons are essential for the reaction, underscoring the importance of charge carriers in the photocatalytic process (Fig. 6a).

To investigate the role of oxygen vacancies in photocatalytic processes, electron paramagnetic resonance (EPR) studies were conducted to quantify oxygen vacancy concentrations.² As shown in Fig. 5b, Im-V/CeO₂ exhibits a significantly stronger signal at g = 2.003 compared to CeO₂ and Pm-V/CeO₂, which correlates directly with its highest activity. This observation, when compared to V₂O₅, provides evidence for strong electronic coupling between VO_x species and the CeO₂ lattice.

The polar V=O bonds interact with lattice oxygen, promoting oxygen vacancy formation and enhancing the catalyst's performance in activating O₂. Moreover, in situ EPR in a 5,5-dimethyl-1-pyrroline N-oxide (DMPO)/methanol solution under illumination (Fig. 6c) demonstrated that Im-V/CeO₂ generated the most intense ˙O₂⁻ signal after 10 min, further confirming the role of oxygen vacancies in enhancing O₂ activation. Importantly, this signal intensity was quantitatively correlated with the catalytic activity, showing a clear linear relationship (Fig. S17), which further confirms the critical role of ˙O₂⁻ radicals in the reaction. To gain a more comprehensive understanding of radical intermediates, in situ DMPO-EPR experiments were performed (Fig. 6d). These revealed no detectable signals in the dark, but progressively intensified six-line peaks under light irradiation, matching β-carbon-centered radicals. These findings corroborated by isotopic evidence, confirm the involvement of a carbon-centered radical pathway.^62–64 Substrate scope studies (Fig. 6e) using analogues of substrate 1 under identical conditions provided further insights into the reaction mechanism. Ketone (1b) and 1,2-diphenylethane (1d) remained completely unreacted, ruling out 1b as an intermediate and reinforcing the essential role of the benzylic hydroxyl group in photocatalytic C–C bond scission. Additionally, substrate 2 exhibited negligible conversion, indicating high selectivity for β-1 models. Substrates 3–5, which bear benzylic hydroxyl groups, were fully converted, with 3 and 4 showing superior performance compared to 5. This suggests that side-chain functionality also plays a role in determining selectivity. Collectively, these findings demonstrate that oxygen vacancies enhance the concentration of reactive sites and improve charge separation on the catalyst surface. This facilitates the synergistic action of ˙O₂⁻ radicals, electrons, and holes, enabling efficient C–C bond scission and high yields of benzaldehyde.

Photocatalytic conversion of methanol-extracted poplar and birch lignin catalyzed by Im-V/CeO₂ was conducted within 24 h.⁶⁵ After the reaction, the initially crimson lignin solution in acetonitrile faded markedly (Fig. 7a), suggesting catalyst-induced oxidative cleavage. To verify that this change corresponded to actual depolymerization, we analyzed the liquid products by GC-MS and identified various low-molecular-weight aromatic compounds (Fig. 7b).¹¹ Further evidence was provided by gel permeation chromatography (GPC), which showed a clear decrease in the number-average molecular weight (M_n) of birch lignin from 1832 to 1208 g mol⁻¹ after the reaction, along with peak broadening and a shift to higher elution volumes (Fig. 7c)—collectively indicating polymer chain scission.⁶⁶ These results demonstrate the efficacy of Im-V/CeO₂ in cleaving C–C bonds in native lignin. It should be noted, however, that the theoretical maximum yield of cleavage products remains limited to approximately 1%, due to the low natural abundance of β-1 linkages.⁶⁷

Fig. 7 (a) Schematic illustration of the photocatalytic depolymerization process of poplar and birch lignin. (b) GC-MS chromatograms of the liquid products from the photocatalytic depolymerization of birch and polar lignin. (c) GPC curves of birch lignin before and after the photocatalytic reaction. Reaction conditions: 50 mg of methanol-extracted lignin, 3 ml of acetonitrile, 10 mg of catalyst, O₂, 455 nm, 24 h.

Based on the experimental results outlined above and previous studies, we propose three possible photocatalytic conversion pathways for the conversion of the β-1 model compound using Im-V/CeO₂ as the catalyst (Fig. 8). According to the main product types, the pathways can be categorized into two C–C bond cleavage pathways (paths I and II) and one preliminary oxidation pathway (path III). The low selectivity for ketone byproducts (below 2%) suggests that preliminary oxidation is not the prevailing mechanism. This aligns with existing studies and is further confirmed by the ˙O₂⁻ trapping experiments discussed earlier, which indicate that preliminary oxidation and C–C bond cleavage are competitive processes.¹¹ Path I, representing the primary reaction pathway, occurs under light irradiation. Oxygen vacancies on the catalyst surface adsorb and activate O₂, which is subsequently reduced to ˙O₂⁻ by photogenerated electrons. Simultaneously, photogenerated holes attack the C_β–H bond to generate a C_β radical intermediate. These radical captures ˙O₂⁻ and protons, forming a six-membered ring transition state. This is followed by intramolecular electron transfer, leading to C–C bond cleavage and the formation of two benzaldehyde molecules. The synergistic participation of photogenerated electrons, holes, oxygen vacancies, and molecular oxygen in this pathway explains the significant decrease in benzaldehyde yield during trapping experiments, thereby confirming path I as the dominant mechanism.

Fig. 8 Proposed reaction pathways for the photocatalytic conversion of 1,2-diphenylethanol catalyzed by Im-V/CeO₂.

Path II resembles the homogeneous vanadium-catalyzed C–C cleavage mechanism reported by Liu et al., which operates in the absence of oxygen.^8,11 In this pathway, the benzylic hydroxyl group of the substrate is activated by holes to form a radical, leading to direct cleavage and the production of benzaldehyde and a benzyl radical. The benzyl radical, unable to undergo further oxidation, couples to form 1,2-diphenylethane. In contrast, under a nitrogen atmosphere, only trace amounts of benzaldehyde and 1,2-diphenylethane are observed, further supporting the conclusion that path II is not a dominant mechanism (Fig. S12). Path III corresponds to a preliminary oxidation process, where the hydroxyl group (–OH) is directly oxidized by holes to yield ketone products. However, the low ketone yield suggests that this pathway is not the primary mode of action. It becomes significantly more active only when ˙O₂⁻ is inhibited, indicating that it competes with C–C cleavage rather than dominating it.

Conclusions

In summary, we have developed a rod-shaped CeO₂ photocatalyst with surface-decorated VO_x species that enables highly selective C–C bond cleavage of lignin model compounds under mild visible-light irradiation, leading to sustainable benzaldehyde production. Under optimized conditions, the lignin model achieves a 90% conversion with an impressive 98% selectivity for benzaldehyde. Compared to bare CeO₂, the VO_x-decorated catalyst demonstrates significantly enhanced C–C bond scission activity and effectively suppresses side reactions. Comprehensive characterizations indicate that the incorporation of surface VO_x onto CeO₂ induces a positive shift in both the valence and conduction bands, thereby enhancing oxidative capability while maintaining O₂ activation for reactive oxygen species generation. EPR analysis reveals that VO_x on the Im-V/CeO₂ system exhibits stronger electronic coupling with CeO₂ compared to V₂O₅, resulting in a substantial increase in oxygen vacancy concentration and additional sites for O₂ adsorption and activation. The generated ˙O₂⁻ radicals are identified as the primary active species in the photocatalytic cleavage of lignin C–C bonds. Through comparative studies of impregnation and physical-mixing pretreatments, we achieved precise control over the surface VO_x species and modulated the oxygen vacancy density on CeO₂, which significantly influenced the catalytic performance. This work not only provides a novel strategy for the efficient and environmentally friendly valorization of lignin but also offers valuable insights for the design and application of heterogeneous cerium-based catalysts.

Author contributions

Yongqi Zhuang was responsible for conducting all the experiments, handling software-related tasks, performing data analysis, and drafting the original manuscript. Yuguo Dong contributed to the methodology and revised the manuscript. Yanmei Zheng, Wenjun Zhang, Lin Dong and Zupeng Chen revised the manuscript and played a role in obtaining funding. Additionally, all authors participated in data analysis, discussion, and the revision of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data that support the findings of this study are available in the supplementary information (SI) of this article. Supplementary information is available. Representative TEM image of CeO₂-R; XRD, Raman and FT-IR spectra of catalysts with varying vanadium loadings prepared by impregnation-calcination method; Magnified views of the characteristic peaks in the XRD and FT-IR spectra for Pm-V/CeO₂ and Im-V/CeO₂; The N₂ adsorption-desorption isotherms of CeO₂, Im-V/CeO₂ and Pm-V/CeO₂; V 2p XPS spectrum of pristine V₂O₅; O₂-TPD spectra of CeO₂, Im-V/CeO₂, and Pm-V/CeO₂; Quantitative comparison of V⁴⁺, Ce³⁺, and Ov fractions for the Im-V/CeO₂ catalyst under normal, O₂-dark, and O₂-light conditions from in situ XPS; EIS spectra and TRPL spectra of CeO₂, Im-V/CeO₂ and Pm-V/CeO₂; The standard photocatalytic reaction Setup; Screening of CeO₂ morphologies, evaluation of supported metal oxides and activity screening of metals impregnated at optimal loadings; Post-reaction GC chromatograms of Im-V/CeO₂ and Pm-V/CeO₂ under identical reaction conditions; Post-reaction GC-MS chromatogram of Im-V/CeO₂ under N₂ reaction conditions; XRD patterns and FT-IR spectra of Im-V/CeO₂ before and after stability test; Summary of E-factor analysis from literature for the photocatalytic conversion of 1,2-diphenylethanol; Atom economy (AE) analysis of different reaction pathways in the photocatalytic conversion of 1,2-diphenylethanol; Control experiments in the absence of photocatalyst, light, or Dpol; Quantitative relationship between the EPR signal intensity of superoxide radicals (˙O₂⁻) and the catalytic activity for CeO₂, Im-V/CeO₂, and Pm-V/CeO₂; Vanadium loading determined by ICP-OES analysis; The surface content of vanadium measured by XPS; BET surface area, pore volume, and average pore diameter determined by N₂ physical adsorption; Quantification of desorbed oxygen species corresponding to different desorption peaks in the O₂-TPD profiles; Comparison of performances and E-factor of Im-V/CeO₂ with typical photocatalytic systems for oxidative cleavage of C–C bond in 1,2-diphenylethanol. See DOI: https://doi.org/10.1039/d5gc05517b.

Acknowledgements

This work was supported by National Key Research and Development Program of China (2023YFD2200505), National Natural Science Foundation of China (22205113, 22202105, 22403034), the Natural Science Foundation of Jiangsu Higher Education Institutions of China (21KJA150003), and the Innovation and Entrepreneurship Team Program of Jiangsu Province (JSSCTD202345).

References

1. C. Chatterjee, F. Pong and A. Sen, Green Chem., 2015, 17, 40–71.
2. H. Fan, J. Su, E. Zhao, Y. Zheng and Z. Chen, J. Colloid Interface Sci., 2025, 684, 140–147.
3. J. Yu, J. González-Cobos, F. Dappozze, F. J. López-Tenllado, J. Hidalgo-Carrillo, A. Marinas, P. Vernoux, A. Caravaca and C. Guillard, Appl. Catal., B, 2022, 318, 121843.
4. G. W. Huber, S. Iborra and A. Corma, Chem. Rev., 2006, 106, 4044–4098.
5. W. Deng, Y. Feng, J. Fu, H. Guo, Y. Guo, B. Han, Z. Jiang, L. Kong, C. Li, H. Liu, P. T. T. Nguyen, P. Ren, F. Wang, S. Wang, Y. Wang, Y. Wang, S. S. Wong, K. Yan, N. Yan, X. Yang, Y. Zhang, Z. Zhang, X. Zeng and H. Zhou, Green Energy Environ., 2023, 8, 10–114.
6. T. Hou, N. Luo, H. Li, M. Heggen, J. Lu, Y. Wang and F. Wang, ACS Catal., 2017, 7, 3850–3859.
7. D. V. Evtuguin, C. P. Neto, A. M. S. Silva, P. M. Domingues, F. M. L. Amado, D. Robert and O. Faix, J. Agric. Food Chem., 2001, 49, 4252–4261.
8. H. Liu, H. Li, N. Luo and F. Wang, ACS Catal., 2019, 10, 632–643.
9. S. K. Hanson and R. T. Baker, Acc. Chem. Res., 2015, 48, 2037–2048.
10. M. Wang and F. Wang, Adv. Mater., 2019, 31, 1901866.
11. X. Wu, J. Lin, H. Zhang, S. Xie, Q. Zhang, B. F. Sels and Y. Wang, Green Chem., 2021, 23, 10071–10078.
12. C. Li, X. Zhao, A. Wang, G. W. Huber and T. Zhang, Chem. Rev., 2015, 115, 11559–11624.
13. L. Shuai, J. Sitison, S. Sadula, J. Ding, M. C. Thies and B. Saha, ACS Catal., 2018, 8, 6507–6512.
14. S. Guadix-Montero and M. Sankar, Top. Catal., 2018, 61, 183–198.
15. L. Dong, L. Lin, X. Han, X. Si, X. Liu, Y. Guo, F. Lu, S. Rudić, S. F. Parker, S. Yang and Y. Wang, Chem, 2019, 5, 1521–1536.
16. N. Li, K. Yan, T. Rukkijakan, J. Liang, Y. Liu, Z. Wang, H. Nie, S. Muangmeesri, G. Castiella-Ona, X. Pan, Q. Zhou, G. Jiang, G. Zhou, J. Ralph, J. S. M. Samec and F. Wang, Nature, 2024, 630, 381–386.
17. M. Wang, J. Ma, H. Liu, N. Luo, Z. Zhao and F. Wang, ACS Catal., 2018, 8, 2129–2165.
18. C. T. Palumbo, E. T. Ouellette, J. Zhu, Y. Román-Leshkov, S. S. Stahl and G. T. Beckham, Nat. Rev. Chem., 2024, 8, 799–816.
19. H. Chen, K. Wan, F. Zheng, Z. Zhang, Y. Zhang and D. Long, Renewable Sustainable Energy Rev., 2021, 147, 111217.
20. J. Xu, P. Zhou, C. Zhang, L. Yuan, X. Xiao, L. Dai and K. Huo, Green Chem., 2022, 24, 5351–5378.
21. Z. Luo, C. Liu, A. Radu, D. F. de Waard, Y. Wang, J. T. Behaghel de Bueren, P. D. Kouris, M. D. Boot, J. Xiao, H. Zhang, R. Xiao, J. S. Luterbacher and E. J. M. Hensen, Nat. Chem. Eng., 2024, 1, 61–72.
22. E. Amadio, R. Di Lorenzo, C. Zonta and G. Licini, Coord. Chem. Rev., 2015, 301–302, 147–162.
23. B. Sedai, C. Díaz-Urrutia, R. T. Baker, R. Wu, L. A. P. Silks and S. K. Hanson, ACS Catal., 2013, 3, 3111–3122.
24. M. Wang, L. H. Li, J. M. Lu, H. J. Li, X. C. Zhang, H. F. Liu, N. C. Luo and F. Wang, Green Chem., 2017, 19, 702–706.
25. Z.-z. Zhou, M. Liu and C.-J. Li, ACS Catal., 2017, 7, 3344–3348.
26. S. Kim, S. C. Chmely, M. R. Nimlos, Y. J. Bomble, T. D. Foust, R. S. Paton and G. T. Beckham, J. Phys. Chem. Lett., 2011, 2, 2846–2852.
27. Y. Meng, J. Li, H. Liu, T. Liu, J. Hu and H. Li, Small Methods, 2025, 9, 2401510.
28. H. P. Toan, D.-V. Nguyen, P. D. M. Phan, N. H. Anh, P. P. Ly, M.-T. Pham, S. H. Hur, T. D. T. Ung, D. D. Bich, M. C. Nguyen, N. L. Nguyen, D. T. Huyen, W. J. Yu and H.-T. Vuong, ACS Appl. Mater. Interfaces, 2024, 16, 29421–29438.
29. P. Hu, P. Hu, T. D. Vu, M. Li, S. Wang, Y. Ke, X. Zeng, L. Mai and Y. Long, Chem. Rev., 2023, 123, 4353–4415.
30. J. Xu, Z. Zhou, M. Liu, J. Wang and L. Zhang, Int. J. Biol. Macromol., 2023, 245, 125476.
31. G. Xiao, Z. Wang, Y. Jin and F. Wang, Nanotechnology, 2024, 35, 5704.
32. L. Wei, K. Qu, J. Huang, Z. Li, X. Ning, X. Yang, M. Li, H. Zhang and J. Yang, Int. J. Biol. Macromol., 2025, 311, 143849.
33. J. Xu, T. Gong, X. Zhang, M. Liu and J. Wang, Surf. Interfaces, 2023, 41, 103210.
34. J. Xu, X. Zhang, J. Long and C. Ge, J. Environ. Chem. Eng., 2023, 11, 110854.
35. Y. Zou, L. Dong, S. Yan, J. Liu, L. Mu, L. Li, Y. Hu, H. Qi, S. Mao and Z. Chen, J. Catal., 2024, 429, 115241.
36. V. Bratan, P. Chesler, C. Hornoiu, M. Scurtu, G. Postole, P. Pietrzyk, A. Gervasini, A. Auroux and N. I. Ionescu, J. Mater. Sci., 2020, 55, 10466–10481.
37. M. A. Zeleke and D. H. Kuo, Chemosphere, 2019, 235, 935–944.
38. X. Ji, Y. Chen, P. Zhang, S. Song, J. Liu and W. Song, Chin. Chem. Lett., 2025, 36, 110719.
39. A. Kumar and P. P. Sahay, J. Sol-Gel Sci. Technol., 2020, 95, 34–51.
40. T. Taniguchi, Y. Sonoda, M. Echikawa, Y. Watanabe, K. Hatakeyama, S. Ida, M. Koinuma and Y. Matsumoto, ACS Appl. Mater. Interfaces, 2012, 4, 1010–1015.
41. J. Li, W. Su, F. Chen, L. Fu, S. Ding, K. Song, X. Huang and L. Zhang, J. Phys. Chem. C, 2019, 123, 3900–3907.
42. A. Badreldin, M. D. Imam, Y. Wubulikasimu, K. Elsaid, A. E. Abusrafa, P. B. Balbuena and A. Abdel-Wahab, J. Alloys Compd., 2021, 871, 159615.
43. B. Liu, D. Yin, F. Zhao, K. K. Khaing, T. Chen, C. Wu, L. Deng, L. Li, K. Huang and Y. Zhang, J. Phys. Chem. C, 2019, 123, 4193–4203.
44. J. Shi, J. Chen, J. Wang, R. Yin, B. Wang, S. Xiong, Z. Wang, H. Liu and J. Li, Fuel, 2023, 332, 126189.
45. T. Wu, G. Liu, L. Zeng, G. Sun, S. Chen, R. Mu, S. A. Gbonfoun, Z.-J. Zhao and J. Gong, AIChE J., 2017, 63, 4911–4919.
46. L. Schumacher, J. Shen, K. Hofmann and C. Hess, Catal. Today, 2024, 426, 114387.
47. A. Xie, X. Zhou, X. Huang, L. Ji, W. Zhou, S. Luo and C. Yao, J. Ind. Eng. Chem., 2017, 49, 230–241.
48. Y. Dimitriev, V. Dimitrov, M. Arnaudov and D. Topalov, J. Non-Cryst. Solids, 1983, 57, 147–156.
49. Y. Jiang, X. Wen, Y. Li, Y. Li, Y. Peng, W. Feng, X. Li, J. Yang, L. Song, L. Huang, H. Sun and J. Shi, J. Mater. Chem. A, 2025, 13, 645–653.
50. C.-J. Yoo, D.-W. Lee, M.-S. Kim, D. J. Moon and K.-Y. Lee, J. Mol. Catal. A: Chem., 2013, 378, 255–262.
51. C. Yan and L. Liu, Chem. Eng. J., 2020, 392, 123738.
52. H. Wang, D. Huang, Q. Zhuge and Y. Wu, Diamond Relat. Mater., 2023, 139, 110413.
53. G. Silversmit, D. Depla, H. Poelman, G. B. Marin and R. De Gryse, J. Electron Spectrosc. Relat. Phenom., 2004, 135, 167–175.
54. J. Jiang, X. Li, X. Liu, L. Shi, F. Su and T. Zhu, Sep. Purif. Technol., 2025, 361, 131494.
55. X. Liang, Y. Tang, Y. Cui, Q. Tian, Q. Dong, X. Li, F. Yu, J. Zhao and C. Wang, Chem Catal., 2025, 5, 101490.
56. B. Cui, L. Zhou, K. Li, Y.-Q. Liu, D. Wang, Y. Ye and S. Li, Appl. Catal., B, 2020, 267, 118670.
57. L. Zhang, G. Spezzati, V. Muravev, M. A. Verheijen, B. Zijlstra, I. A. W. Filot, Y.-Q. Su, M.-W. Chang and E. J. M. Hensen, ACS Catal., 2021, 11, 5614–5627.
58. X. Li, B. Kang, F. Dong, Z. Zhang, X. Luo, L. Han, J. Huang, Z. Feng, Z. Chen, J. Xu, B. Peng and Z. L. Wang, Nano Energy, 2021, 81, 105671.
59. S.-m. Chang and W.-s. Liu, Appl. Catal., B, 2011, 101, 333–342.
60. R. A. Sheldon, M. L. Bode and S. G. Akakios, Curr. Opin. Green Sustainable Chem., 2022, 33, 100569.
61. Y. Meng, J. Huang, J. Li, Y. Jian, S. Yang and H. Li, Green Chem., 2023, 25, 4453–4462.
62. Y. Ai, Y. Zhao, X. Huang, X. Liu, S. Kuang, H. Ding, Y. Zeng, H. Liu and G. Liu, Green Chem., 2025, 27, 8936–8949.
63. Z. Chen, J. Deng, W. Zhang, L. Dong and Z. Chen, Chin. J. Catal., 2024, 61, 135–143.
64. H. Jiang, W. Yang, X. Lian, M. Liu, M. Zhu, H. Li and F. Zhang, J. Mater. Chem. A, 2024, 12, 28744–28752.
65. L. Li, L. Dong, D. Li, Y. Guo, X. Liu and Y. Wang, ACS Catal., 2020, 10, 15197–15206.
66. J. Deng, C. Zhou, Y. Yang, B. Nan, L. Dong, L. Cai, L. Li, Z.-J. Wang, X. Yang and Z. Chen, Chem. Eng. J., 2023, 462, 142282.
67. R. Rinaldi, R. Jastrzebski, M. T. Clough, J. Ralph, M. Kennema, P. C. A. Bruijnincx and B. M. Weckhuysen, Angew. Chem., Int. Ed., 2016, 55, 8164–8215.

---

Footnote

† These authors contributed equally to this work.

---