Atomic-scale interface engineering in Bi/Bi₂O₃ heterojunctions for selective CO₂ photoreduction to methanol

Abstract

The strategic engineering of an Ohmic junction at the Bi/Bi₂O₃ (BBO) interface is demonstrated to synergistically enhance photocatalytic CO₂-to-methanol conversion through precisely modulated charge behavior and interfacial energy alignment. This metallic Bi-semiconductor Bi₂O₃ Ohmic junction with local surface plasmon resonance effect induces a robust built-in electric field that promotes the unidirectional electron transfer from Bi₂O₃ to Bi while suppressing charge recombination. Theoretical calculations and experimental evidence reveal that the interfacial Bi sites within the Ohmic junction predominantly facilitate CO₂ adsorption and activation to form *COOH, whereas ensuing protonation steps are favored on metallic Bi sites on BBO Ohmic junction. Furthermore, the Ohmic junction enhances interfacial electron density and strengthens orbital hybridization between Bi 6p and O 2p orbitals, thereby reducing the activation energy of the rate-limiting *CO₂ → *COOH step by 0.6 eV, enabling a CH₃OH production rate of 610 μmol g^-1 under light irradiation. The work deciphers the dual role of Ohmic junctions in simultaneously resolving bulk charge transport limitations and tailoring surface catalytic landscapes, establishing a universal paradigm for metal-semiconductor heterojunction photocatalyst design.

Introduction

The photocatalytic reduction of CO₂ into solar fuels like methanol represents a promising route to tackle both energy scarcity and carbon neutrality goals^1,2,3,4,5. Nevertheless, the practical implementation of this photocatalytic reaction faces three fundamental challenges: rapid recombination of photogenerated charge carriers, high bond dissociation energy (≈750 kJ mol⁻¹) for cleaving stable C=O bonds, and dependency on sacrificial agents for sustaining catalytic activity. Substantial research attention has focused on engineering the bulk phase configuration of photocatalysts, as this structural parameter fundamentally governs charge carrier separation dynamics, including constructing ultrathin architectures, introducing heteroatom dopants, and building heterojunctions^6,7. Among these, Ohmic junctions formed, characterized by minimal contact resistance at material interfaces, offer a distinct advantage by enabling unimpeded charge injection while preserving catalytic sites^8,9,10. However, achieving precise atomic-level modulation of Ohmic interfaces to concurrently enhance bulk charge migration and surface reaction kinetics remains a challenging.

Recent advances in Ohmic junction engineering, including metal-semiconductor systems (e.g., Bi/BiOCl, where the metallic Bi Nanoparticles formed a low-resistance contact with the n-type BiOCl photocatalyst to accelerate electron extraction)^11,12 and semiconductor heterointerfaces (e.g., Mo₂TiC₂ MXene/ZnCdS ohmic junction)¹³, have demonstrated enhanced charge separation and catalytic activity in photocatalytic fields. To further improve the performance of Ohmic junction photocatalysts, leveraging the localized surface plasmon resonance (LSPR) effect presents a compelling approach. The LSPR effect confers unique advantages in photocatalytic CO₂ reduction through multiple synergistic pathways: enhanced broadband light-harvesting capabilities, efficient hot-carrier manipulation, intensified electromagnetic near-fields, and thermally aided catalytic processes^14,15. For example, Zhang et al. reported a plasmonic Fe-nanonecklace-hydrogenated titanium hybrid for advancing solar light-driven NH₃ synthesis via a local heating effect-assisted photothermal pathway, circumventing the equilibrium limit of thermocatalysis¹⁶. He et al. synthesized anisotropic Ti₃C₂T_x MXene with strong LSPR features, which showed enhanced plasmonic charge carrier-induced photocatalytic water splitting activity¹⁷. These developments highlight the great potential of integrating LSPR effects into highly efficient Ohmic junction photocatalysts for CO₂ photoconversion.

Herein, we demonstrate an atomic-level engineering strategy to construct Bi/Bi₂O₃ (BBO) Ohmic junctions with LSPR effect for high-efficiency CO₂-to-methanol conversion. By in-situ growing metallic Bi nanodomains on nanosheets, we constructed an interfacial structure with aligned Fermi levels, thereby facilitating barrier-free electron migration from Bi₂O₃ to Bi. Combining X-ray absorption spectroscopy with density functional theory (DFT) simulations, we elucidate that the Ohmic contact facilitates electron accumulation on Bi and the polarized Bi^δ+ sites stabilize *CO₂^- intermediates via strong 6p-2p orbital hybridization (−ICOHP = 1.85 eV). This interaction reduces the energy barrier for *CO₂ protonation by 0.6 eV and stabilizes *CHO intermediates through orbital hybridization. This synergistic configuration achieves a methanol production rate of 610 μmol g⁻¹, surpassing most reported photocatalytic systems. Our work establishes a paradigm for designing Ohmic interfaces that unify ultrafast charge transport and targeted surface catalysis, overcoming longstanding efficiency bottlenecks in artificial photosynthesis.

Results

CO₂ adsorption and activation mechanism over Bi/Bi₂O₃ Ohmic junction

Figure 1a systematically elucidated the interfacial electronic engineering and catalytic mechanisms of Bi/Bi₂O₃ (BBO) Ohmic junction for enhanced CO₂ reduction. The BBO Ohmic junction revealed higher photocatalytic CO₂ reduction activity to CH₃OH than that of Bi or Bi₂O₃, which was ascribed to the interfacial interaction inducing fast separation of photogenerated carriers and the strong adsorption to *CO for its protonation reaction produces. To further understand the mechanism of CO₂ activation and conversion on BBO, we analyzed detailed molecular orbital interactions. The results of the partial density of states (PDOS) indicated that during CO₂ adsorption and activation on Bi₂O₃, hybridization occurred between the Bi 6p_x orbital and the C 2p_z orbital (Fig. 1b and Supplementary Figs. 1–3). This leaded to partial electron transfer from the Bi 6p_x orbital to the 2π anti-bonding orbital of CO₂, weakening the CO₂ molecule and causing its linear structure to bend. This structural change drove CO₂ adsorption and activation on the Bi₂O₃ surface. In the case of CO₂ adsorption and activation on metal Bi of BBO (Fig. 1c, and Supplementary Figs. 4 and 5), the electronic states of Bi’s 6p_x, 6p_y, and 6p_z orbitals near the Fermi level nearly vanished. Meanwhile, tightly overlapping and continuous hybrid peaks emerged in the electronic states of C 2p_x, 2p_y, and 2p_z orbitals of CO₂. This suggested significant electron transfer from 6p_x, 6p_y, and 6p_z orbitals of Bi to the 2π anti-bonding orbital of CO₂, facilitating CO₂ adsorption and activation on metal Bi of BBO. The transferred electrons filled C 2p_x, 2p_y, and 2p_z orbitals of CO₂, resulting in degeneracy among these filled orbitals. This orbital degeneracy manifested as numerous continuous hybrid peaks in the corresponding PDOS. The projected crystal orbital Hamiltonian population (COHP) was also analyzed to understand the mechanism of *CO protonation. The rise of the antibonding state contributes to the bond strength, which can be seen from the computed integrated COHP (ICOHP) as shown in Fig. 1d, a more negtive ICOHP was observed for BBO Ohmic junction (−1.85 eV) than Bi₂O₃ (0.82 eV), indicating a much stronger bond coupling between BBO Ohmic junction and *CO, favoring the further protonation reaction.

Fig. 1: The mechanism CO₂ adsorption/activation and BBO Ohmic junction formation.

a Reaction mechanisms of CO₂ reduction on Bi, Bi₂O₃, and BBO Ohmic junction. b COHP for the Bi 6p-O 2p of absorbed *CO on the surface of Bi₂O₃, and BBO Ohmic junction. PDOS and IPDOS related to Bi 6p-O 2p interactions before (up) and after (down) CO₂ adsorption on c Bi₂O₃ and d BBO Ohmic junction. e Differential Bader charge of BBO Ohmic junction. (f) Theoretical work function of the surface of Bi (012) and Bi₂O₃ (100). g The electron transfer mechanism of BBO Ohmic junction, thereby enabling the generation of CH₃OH. Source data for Fig. 1b–d, f are provided as a Source Data file.

To clarify the interface charge transfer of BBO Ohmic junction, density functional theory calculation (DFT) calculations with Bader charge were calculated. Supplementary Fig. 6 exhibited the optimized crystal structure of Bi₂O₃ and BBO Ohmic junction. Figure 1e depicted the charge-density difference plots of BBO Ohmic junction, whereas the yellow and cyan areas represented electron density accumulation and deplete, respectively. Theoretical calculations indicated that the coupling of Bi and Bi₂O₃ promoted charge enrichment on Bi₂O₃ (Δq = 1.557 e), which suggested that Bi atoms can adjust the surrounding chemical microenvironment of Bi₂O₃, thus affecting the adsorption of intermediate species. To verity the heterojunction type in theory, the theoretical work functions of Bi and Bi₂O₃ were calculated to be 4.78 and 5.15 eV, respectively (Fig. 1f). The much smaller work function of Bi promoted the charge transfer from Bi to Bi₂O₃ until their consistent Fermi levels after the coupling of Bi and Bi₂O₃, and at the same time, inducing the generation of a strong built-in electric field and band bending (Fig. 1g). This phenomenon can accelerate the transfer and separation of photogenerated carriers for CO₂ photoreduction to CH₃OH (Supplementary Fig. 7).

Structural characterizations

Figure 2a presents the synthetic procedures of BBO Ohmic junction via a hydrothermal and wet-chemical cascade strategy. The Bi₁₉Br₃S₂₇ nanowires have been firstly synthesized by hydrothermal method, then followed by a NaBH₄ reduction method to inducing the dissolution and crystallization of Bi species, and finally the BBO nanosheets were obtained. The X-ray diffraction (XRD) patterns of the samples were shown in Supplementary Fig. 8. The nanowire morphology of Bi₁₉Br₃S₂₇ samples were shown in Fig. 2b and Supplementary Fig. 9. To validate the structure evolution of Bi₁₉Br₃S₂₇ nanowires to form BBO nanosheets, the different reduction time for Bi₁₉Br₃S₂₇ was conducted at the same conditions. Figure 2c–e and Supplementary Figs. 10–12 record the morphologies of BBO-2, BBO-4, and BBO-6. With the increasing of reduction time, the BBO samples exhibited a gradual increased nanosheet structure (Supplementary Figs. 13–15). Moreover, the progressive rise in Bi content of these samples indicated the existence of a rich bismuth environment (Supplementary Tables S2–4). The SEM, TEM and AFM tests of Bi₂O₃ were conducted. The scanning electron microscope (SEM) and transmission electron microscopy (TEM) images indicate that the Bi₂O₃ sample was composed of irregularly nanosheet (Supplementary Figs. S16a, b). The lattice spacings of 0.274 nm can be assigned to the (400) planes of Bi₂O₃ (Supplementary Fig. S16c). The AFM images and the corresponding height profile showed that the thickness of its nanosheet was ~5.06 nm (Supplementary Figs. S16d, e). These data provide a solid foundation for the subsequent comparison with BBO-4 composite materials. The excellent lattice matching and well-defined BBO-4 Ohmic junction were verified by high resolution TEM (HRTEM; Fig. 2f, g), demonstrating the success in fabricating the Ohmic junction. The high-quality coherent heterointerface with a 0.328 nm lattice fringe spacing belonging to the (012) plane of Bi and a 0.274-nm lattice fringe spacing attributing to the (400) plane of Bi₂O₃. Furthermore, the element composition of BBO-4 Ohmic junction was detected by energy dispersive spectroscopy (EDS) element mapping (Fig. 2h–j). Moreover, the Bi content was gradually increased as time extended, and the XRD patterns of BBO samples can be indexed to hexagonal Bi (JCPDS: 05-0519) and Bi₂O₃ (JCPDS: 22-0515) (Fig. 2k, and Supplementary Fig. 17) and no other phase structures were detected (Supplementary Fig. 18)^18,19. To precisely resolve the phase composition, Rietveld refinement of the XRD patterns (Supplementary Fig. 19) was performed to determine the phase composition of the BBO catalysts. The refinement results confirm the exact phase composition. The Bi/Bi₂O₃ mass ratios for BBO-2, BBO-4 and BBO-6 were 53.65/46.35, 91.75/8.25, and 97.74 wt%/2.26 wt%, respectively. The Bi content increased with increasing reduction time. The above results showed there was on other phase structures. The Raman peaks of metallic Bi at 71 and 98 cm⁻¹ corresponded to the asymmetric coupling vibration and symmetric vibration of Bi atoms within the metal (Supplementary Fig. 20). The Raman peaks of Bi₂O₃ at 326 and 465 cm⁻¹ were attributed to the stretching of Bi–O bonds and the symmetric stretching of Bi–O–Bi in Bi–O octahedral units within Bi₂O₃. In the Raman spectrum of BBO-4, characteristic peaks of both metallic Bi and Bi₂O₃ were observed. The above results confirmed the chemical reduction can achieve the structure reconstruction of Bi₁₉Br₃S₂₇ for BBO nanosheet formation.

Fig. 2: Structural characterization of BBO Ohmic junction.

a Schematic illustration of the synthesis of BBO. TEM image of b BBS, c BBO-2, d BBO-4, and e BBO-6. f High-solution TEM image, g lattice fringes, and h–j EDS mapping images of BBO-4. k XRD patterns of Bi, Bi₂O₃, BBO-2, BBO-4, BBO-6. Normalized XANES spectra at the l Bi-L₃ edge and m corresponding Fourier transform R space fitting results. Source data for Fig. 2k–m are provided as a Source Data file.

As shown in Fig. 2l, the absorption edge of BBO-4 in the X-ray absorption fine structure (XAFS) Bi L₃-edge was between the Bi foil and Bi₂O₃, meaning that the valence state was Bi^δ+ (0 < δ < 3), further confirming the existence of metal Bi. In addition, the absorption edge of Bi in the XAFS Bi L₃-edge was between the Bi foil and Bi₂O₃, showing the occurrence of surface oxidation, which was accordance with the X-ray photoelectron spectroscopy (XPS) results. To obtain the quantitative structural parameters around Bi, Fourier transform of the least-squares curve fitting was performed on the EXAFS data of Bi, Bi₂O₃, and BBO-4 (Fig. 2m). BBO-4 exhibited a typical Bi/Bi₂O₃ structure. The presence of both Bi–O and Bi–O–Bi coordination shells confirmed the existence of Bi₂O₃ oxide, while the Bi-Bi metallic coordination indicated the presence of metallic Bi. In contrast, the Bi₂O₃ sample showed only a single Bi–O coordination shell, demonstrating its phase-pure Bi₂O₃ composition. The Bi metal sample, however, revealed both Bi-Bi and Bi-O bonds, suggesting the existence of surface oxidation. To investigate the electronic states and charge transfer of BBO-4, XPS analysis was employed. For Bi₂O₃, the XPS diffraction peaks at 159.00 eV (Bi 4f_7/2) and 164.32 eV (Bi 4f_5/2) were characteristic of Bi³⁺ species (Supplementary Fig. 21a)^20,21. The characteristic Bi 4f peaks of Bi metals could be fitted into two pairs of doublets associated with Bi⁰ and Bi³⁺. The XPS signal of Bi oxidation states in Bi metals was derived from surface oxidation of the catalyst under exposure to air. For BBO-4, the Bi 4f signals deconvoluted into four peaks: two at 156.68 and 161.98 eV (assigned to Bi⁰ in metallic Bi) and two at 158.89 and 164.21 eV (attributed to Bi³⁺ in Bi₂O₃)^22,23,24. Comparative analysis of binding energies across Bi and Bi₂O₃, BBO-4 revealed a 0.11 eV positive shift in the Bi⁰ binding energy for the heterojunction relative to pristine metallic Bi, indicating decreased electron density around Bi⁰. Conversely, the Bi³⁺ binding energy in BBO showed an 0.11 eV downshift compared to Bi₂O₃, signifying increased electron density around Bi³⁺. These XPS results collectively suggested electron transfer from metallic Bi to Bi₂O₃ within the heterojunction, consistent with the electron flow direction in Ohmic junction^25,26,27. The O 1s XPS spectrum of BBO-4 was the same as that of pristine Bi₂O₃, with peaks at 529.52 and 530.85 eV, which were attributed to the chemical bonding of Bi–O and adsorbed oxygen (H–O), respectively (Supplementary Fig. 21b). The results were in accordance with the above work function analysis.

Local surface plasmon resonance effect and photocatalytic CO₂ reduction over BBO

Prior to evaluating photocatalytic performance, we first investigated the material’s light utilization capability for photocatalytic applications. The UV-vis-NIR DRS (Fig. 3a) revealed that BBO-4 demonstrated significantly enhanced broad-spectrum light absorption (280–2100 nm) compared to its constituent materials (Bi and Bi₂O₃). This improved LSPR performance in the BBO Ohmic junction can be attributed to its unique nanosheet architecture. The abundant interfaces between metallic Bi and Bi₂O₃, along with their interfacial interactions, collectively modulated the LSPR intensity. Finite-difference time-domain simulations further corroborated the presence of a strong LSPR effect in BBO (Fig. 3b), demonstrating enhanced local electric field strength surpassing that of individual components^28,29. Notably, the accumulation of LSPR-excited hot electrons at the interface spontaneously elevated the BBO surface temperature to 117.3 °C without external heating (Fig. 3b, and Supplementary Fig. 22). This LSPR-induced phenomenon simultaneously enhanced photon utilization efficiency and enabled localized photothermal energy conversion, thereby synergistically promoting photocatalytic activity. The corresponding bandgaps E_g of Bi₂O₃ was determined to be 2.51 eV, according to Tauc plots: (ahν)^1/2 = A(hν- Eg) (Supplementary Fig. 23). Mott–Schottky analysis indicated flat potentials of −0.20 V (Bi) and −0.57 V (Bi₂O₃) versus NHE (Fig. 3c). Given that semiconductor conduction band potentials typically located at more negative values than their flat potentials, the BBO band alignment satisfied the thermodynamic requirements for CO₂ photoreduction to CH₃OH.

Fig. 3: LSPR effect and photocatalytic CO₂ reduction activity evaluation.

a UV-vis-NIR DRS spectra and b spatial distribution of electric fields enhanced by LSPR effect under the excitation wavelength of 800 nm of Bi, Bi₂O₃, and BBO, and surface temperature of BBO-4 under light irradiation with the increasing time c Mott–Schottky plots of Bi and Bi₂O₃. d Photocatalytic CO₂ activities of the as-prepared samples. e Photocatalytic performance of BBO-4 under different conditions. f Cycling experiments of BBO-4 for photocatalytic CO₂ reduction. The error bars represent the standard deviation derived from three repeated experiments. g Mass spectra for CO₂ reduction of BBO-4 using ¹³CO₂ as the reacting gas. h Photocatalytic CO₂ reduction performances of BBO-4 compared with reported representative photocatalysts (the detailed reaction conditions were listed in Supplementary Table S5). Source data for a, c, d–h are provided as a Source Data file.

The photocatalytic CO₂ reduction performances of synthesized catalysts were evaluated under full spectrum (Fig. 3d, and Supplementary Fig. 24). The primary products of Bi, Bi₂O₃, and Bi₂S₃ resulting from photocatalytic CO₂ reduction were mainly CO. The precursor Bi₁₉Br₃S₂₇ nanowires revealed the generation of CO, CH₄, and CH₃OH. The series of BBO derived from Bi₁₉Br₃S₂₇ nanowires displayed enhanced photocatalytic activity for CH₃OH generation. The optimal BBO-4 samples reached the highest value of 610 µmol g⁻¹, substantially surpassing the rates observed with unmodified Bi, Bi₂O₃, and precursor Bi₁₉Br₃S₂₇ nanowires. The liquild products were further detected by nuclear magnetic spectrum (NMR). As shown in Supplementary Fig. 25, the proton spectrum exhibited a methanol peak at δ 3.28 ppm, and a water peak at δ 4.7 ppm. Based on the (NMR) analysis of the liquid products from photocatalytic CO₂ reduction, it is confirmed that the liquid product generated by the BBO-4 heterojunction photocatalyst was CH₃OH. No impurity peaks or other liquid products were detected.

In contrast, no reduction products were observed under dark conditions, in the absence of a catalyst or Ar atmosphere (Fig. 3e). The above results of controlled experiments demonstrated the generation of products as a result of reduced CO₂ in the photocatalytic system. N₂ adsorption/desorption isotherms and the pore-size distributions of Bi, Bi₂O₃, and BBO-4 were characterized (Supplementary Fig. 26), suggesting surface area was not the influence factor on photocatalytic activity. As depicted in Fig. 3f, the yields of products were basically no significant change after four cycles, and its structure had no obvious agglomeration and transform (Supplementary Fig. 27), demonstrating the excellent stability of BBO-4. Furthermore, precise detection of ¹³CH₄ (m/z = 17), ¹³CO peak (m/z = 29) and ¹³CH₃OH (m/z = 33) utilizing ¹³C as the only carbon source, provided further evidence that CH₃OH, CH₄ and CO were derived from CO₂ photoreduction (Supplementary Fig. 28, and Fig. 3g). To demonstrate the superior activity of BBO-4, we conducted comparative analyzes with over 20 different catalytic systems (Fig. 3h). The results conclusively showed that BBO-4 exhibited favorable photocatalytic performance among the reported photocatalyst.

Identification of CO₂ photoreduction intermediates

To investigate the mechanistic pathway of photocatalytic CO₂ reduction, we conducted in situ diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS) to monitor the dynamic evolution of adsorbed intermediates on pristine Bi, Bi₂O₃, and BBO-4 catalysts under light irradiation (Fig. 4a–f, and Supplementary Figs. 29–31). The characteristic peak at 1546 cm⁻¹ (Fig. 4a) corresponds to combined bending and stretching vibrations of the *COOH intermediate on metallic Bi^30,31,32, a feature consistently observed in both Bi₂O₃ and BBO-4 systems under illumination (Fig. 4b, c). Notably, comparative analysis revealed that BBO-4 exhibited weaker b-CO₃⁻ signal intensity but enhanced m-CO₃²⁻ formation compared to pristine Bi and Bi₂O₃. This distinct behavior suggests that m-CO₃²⁻ species generated through single oxygen-site adsorption at Bi centers serve as key intermediates in the CO₂ reduction pathway on BBO-4^31,33,34. The diagnostic *CO intermediate (2076 cm⁻¹) was detected across all three catalysts, with its presence indicating a critical branching point for subsequent hydrocarbon formation, particularly CH₄ and CH₃OH^35,36,37. Temporal evolution studies (Fig. 4d–f) demonstrated progressive intensification of *COOH signals over 5–60 min, consistent with continuous CO₂ activation and hydrogenation processes. Crucially, BBO-4 exhibited significantly stronger *CO adsorption compared to its constituent materials (Bi and Bi₂O₃), as evidenced by enhanced peak intensity. This pronounced adsorption capacity favored further hydrogenation steps over CO desorption, thereby directing the reaction pathway toward CH₃OH products. Based on comprehensive analysis of these in situ DRIFTS results, we proposed a stepwise reaction mechanism for photocatalytic CO₂ reduction on BBO-4 (Fig. 4g). The process initiates with CO₂ chemisorption at Bi–O sites, followed by sequential protonation steps generating *COOH intermediates. Subsequent cleavage of C–O bonds yields adsorbed *CO species, which underwent progressive hydrogenation through surface-bound reactive hydrogen species to form CH₄ and CH₃OH.

Fig. 4: In-situ FTIR spectra.

Investigation of CH₃OH generation during photocatalytic CO₂ reduction. In situ ATR-FTIR experiments for (a, d) Bi, (b, e) Bi₂O₃, and (c, f) BBO-4 over 1000–2200 cm⁻¹, 1500–1600 cm⁻¹, and 2000–2200 cm^-1 under full spectrum light illumination. g Proposed molecular mechanisms of CO₂ adsorption and CO₂-to-CH₃OH photosynthesis on the surface of BBO-4 nanosheets. Source data for Fig. 4a–f are provided as a Source Data file.

Carrier dynamics

Through in situ Kelvin probe force microscopy (KPFM) coupled with controlled light illumination, we directly visualized the spatial-temporal dynamics of photoinduced charge carriers in BBO-4. Under dark conditions, surface potential line profiles revealed distinct differences among the materials: Bi₂O₃ and BBO-4 exhibited surface potentials of 0.50, 0.97 mV, respectively (Fig. 5a, b). Light irradiation induced dramatic surface potential reduction in BBO-4 (ΔV = 0.62 mV, Fig. 5c), demonstrating efficient photoinduced electron migration to the material surface under built-in electric field guidance. Notably, BBO-4 demonstrated the much more pronounced surface potential heterogeneity, suggesting enhanced charge redistribution capabilities. Thickness-normalized analysis (Fig. 5c, and Supplementary Fig. 14) quantified this effect, showing normalized potential gradients of 0.21 and 0.13 mV nm⁻¹ for BBO-4 with or without light irradiation, respectively. However, the normalized potential difference of BBO-4 decreased to 0.08 mV nm⁻¹ (Supplementary Fig. 32), indicating that the electric field can be optimized by appropriate Bi content¹. Ohmic junction generated favorable localized electric fields that facilitate electron accumulation. This light-responsive behavior confirmed that BBO-4 Ohmic junction enhanced charge separation through two synergistic mechanisms: (1) creation of strong interfacial electric fields and (2) directional charge transport mediation, which were also supported by the photocurrent and electrical impedance spectra (Supplementary Fig. 33).

Fig. 5: Surface potential and carrier dynamics of BBO.

KPFM images and surface potential analysis of a Bi₂O₃ in the dark and b, c BBO-4 in the dark and under light irradiation, respectively (ΔA stands for differential absorbance). TA spectra and corresponding 3D contour plots of d, e Bi₂O₃ and g, h BBO-4. Normalized transient absorption kinetics for f Bi₂O₃ and i BBO-4. Source data for Fig. 5a–i are provided as a Source Data file.

To elucidate the interfacial charge transfer mechanisms in Ohmic junction systems, we conducted comparative femtosecond transient absorption (fs-TA) spectroscopy studies on Bi₂O₃ and BBO-4 heterostructures. Bi₂O₃ (Fig. 5d, e) and BBO-4 (Fig. 5g, h) demonstrated extended photoinduced absorption spanning 450–570 nm. The absorption intensity of BBO-4 was stronger than that of Bi₂O₃, indicating enhanced light-matter interactions in BBO-4 Ohmic junction system through interfacial electronic state hybridization. Figure 5f–i reveals the fs-TA kinetic profiles at 550 nm of Bi₂O₃ and BBO-4, and their signal can be well-fitted with triexponential results. The fs-TA spectra of Bi₂O₃ and BBO-4 were fitted by a tri-exponential function and the fitting parameters are τ₁ = 0.191, τ₂ = 6.53, τ₃ = 1490 ps for Bi₂O₃, and τ₁ = 0.366, τ₂ = 44.1, τ₃ = 994 ps for BBO-4. Photogenerated electron relaxation in the CB of Bi₂O₃ and BBO-4 occurred through three pathways: exciton-mediated states (τ₁), defect trap states (τ₂), and recombination of e^- and h⁺(τ₃). Notably, the lifetime components of BBO-4 (τ₁ and τ₂) were much longer than that of Bi₂O₃ (τ₁ = 2.95 ps), which was attributed to the ineffective utilization of photoelectrons within a single photocatalyst, hindering efficient photocatalytic activity^38,39. The shorter lifetime of τ₃ (994 ps) for BBO-4 compared with Bi₂O₃ (τ₃ = 1490 ps) was favorable to the interfacial charge transfer of BBO-4 Ohmic junction (Supplementary Fig. 34).

Reaction mechanism of CO₂ photoreduction

The interfacial engineering strategy significantly enhances gas-solid interactions in the BBO-4 Ohmic junction, as evidenced by complementary gas adsorption analyzes. CO₂ adsorption isotherms (Fig. 6a) demonstrated the adsorption capacities of Bi, Bi₂O₃ and BBO-4 were 0.3402, 0.4826, and 0.5632 cm³g⁻¹. This indicated that BBO-4 may provide more active sites for adsorbing and converting CO₂ molecules during the photocatalytic CO₂ reduction process, thus facilitating the occurrence of the reaction. The enhanced CO₂ adsorption effect favored the adsorption and activation of CO₂. The CO₂ and CO TPD profiles (Fig. 6b, c) revealed intensified desorption signals spanning 100–450 °C, which was indicative of optimized physisorption/chemisorption site distribution. Moreover, the hydrophilicity of Bi₂O₃ and BBO-4 was also observed by contact angle test. The introduction of Bi reduced hydrophilicity of BBO-4, which was beneficial to inhibit the competing reaction of H₂ generation (Fig. 6d, and Supplementary Figs. 35 and 36). The interfacial engineering of BBO-4 revealed improved CO₂ activation capabilities through synergistic electronic modulation, as evidenced by comprehensive adsorption analyzes (Fig. 6e). Comparative structural studies revealed that CO₂ chemisorption induced characteristic molecular distortion across all models: Bi₂O₃, BBO-4 surface, and BBO-4 interface. BBO-4 interface exhibited larger C=O bond elongation and smaller angular deformation compared with the surface of Bi₂O₃ and BBO-4, confirming effective CO₂ polarization on BBO-4 Ohmic junction. The Bader charge analysis displayed that 0.381 e of BBO-4 interface transferred to CO₂ during CO₂ adsorption, which were much more than those of Bi₂O₃ (0.042 e) and BBO-4 surface (0.224 e). The CO₂ adsorption energy of BBO-4 interface (−0.52 eV) was larger than that of Bi₂O₃ (−0.47 eV) and BBO-4 surface (−0.49 eV). These results confirmed that the interface effect was more conducive to CO₂ activation in CO₂ photoreduction process.

Fig. 6: Photocatalytic CO₂ reaction mechanism.

a CO₂ adsorption isotherm curves, b CO₂-TPD, c CO-TPD signals, and d contact angle measurements of Bi₂O₃ and BBO-4. e Calculated models, charge difference and adsorption energy of CO₂ for Bi, Bi₂O₃, BBO (metallic Bi site), and BBO with interfacial site. Free energy diagram for CO₂ photoreduction to CH₃OH over the surface of f Bi, g Bi₂O₃ and h BBO with metallic Bi sites, and i the interfacial sites of BBO. Source data for a–c, f–i are provided as a Source Data file.

Figure 6f–i, and Supplementary Figs. 37–40 display the free energy profiles and the corresponding intermediate configurations. For the models of Bi, Bi₂O₃, BBO with metallic Bi site, and BBO with interfacial site, from a thermodynamic standpoint, COOH* formation represented a rate-limiting step in the CO₂ reduction process with free-energy barriers of 2.07, 3.45, 1.76, and 1.51 eV, respectively. It is obvious that the Ohmic interface reduced the free-energy barrier of rate-determining step by 1.94 eV. The formation of Ohmic junction promoted charge redistribution at the interface, inducing the generation of built-in electric field, further accelerating the adsorption and activation of CO₂. The *CO → CO (g) and *CO → *CHO were critical step which determine the selectivity of CO₂ reaction. Due to the considerably lower free energy of *CO(g) compared to *CHO on both Bi and the metallic Bi sites in BBO, the formation of CO became the thermodynamically favored and primary product. For Bi₂O₃ and interface BBO, the *CHO formation energy was much lower than that of CO (g), suggesting the formation of *CHO was easier than that of CO (g). Moreover, the fabrication of the Ohmic junction interface greatly reduced the energy barrier of the rate-determining step. Experimentally and theoretically, BBO exhibited a better photocatalytic CO₂ reduction performance into CO or CH₃OH than Bi₂O₃.

Discussion

In summary, this work successfully constructed a BBO Ohmic junction photocatalyst via a sequential approach involving hydrothermal synthesis and in situ chemical reduction. The heterostructure exhibits enhanced broad-spectrum solar utilization, CO₂ adsorption capacity, and photocatalytic performance for methanol synthesis. Mechanistic investigations indicate that interfacial Bi sites within the Ohmic junction play a dominant role in promoting CO₂ adsorption and activation, while subsequent protonation reactions preferentially occurred on metallic Bi domains. The optimized catalyst achieved a remarkable methanol yield of 610 μmol g⁻¹ under simulated solar irradiation, surpassing the performances of representative semiconductor photocatalysts and retaining stable activity across four consecutive reaction cycles. Theoretical calculations demonstrate that the BBO interface significantly lowers the energy barrier for *CO intermediate conversion, especially facilitating the pivotal *CO → *CHO transition during methanol formation. The enhancement in photocatalytic performance is attributed to the synergy between optimized charge transport mediated by the Ohmic junction and LSPR effects. This research not only presents a versatile strategy for designing high-performance heterojunction photocatalysts but also underscores the promise of plasmon-enhanced bismuth-based materials in advancing sustainable solar fuel generation via CO₂ valorization.

Methods

Materials

Bismuth(III) nitrate pentahydrate (Bi(NO₃)₃·5H₂O, 99%), sodium hydroxide (NaOH, ≥96%), sodium borohydride (NaBH₄, ≥98%), sodium carbonate (Na₂CO₃, 99.95–100.05%), nitric acid (HNO₃, 65%), bismuth(Iii) Bromide (BiBr₃, 99.99%), Thiocarbamide (CH₄N₂S, 99%), Ethylenediaminetetraacetic acid (EDTA, 99.5%), ethanol (CH₃CH₂OH, 99.7%) were all purchased from Beijing InnoChem Science & Technology Co., Ltd. All chemical reagents were used without further purification, and ultrapure water was used throughout the experiment.

Preparation of Bi₂S₃

In a typical synthetic procedure, First, 0.1455 g of Bi(NO₃)₃·5H₂O was dissolved in 20 mL of ethylene glycol under vigorous stirring for 15 min. Subsequently, 0.114 g of CH₄N₂S was added to the solution and stirred vigorously for another 15 min. The mixture was transferred to a 50 mL Teflon-lined stainless-steel autoclave, sealed, and heated in an oven at 180 °C for 24 h. After natural cooling to ambient temperature, the resulting precipitate was washed separately with deionized water and ethanol, then dried at 70 °C for 8 h to obtain Bi₂S₃.

Preparation of Bi nanosheets

Firstly, 2.4254 g of Bi(NO₃)₃·5H₂O was weighed and dissolved in 50 mL of deionized water under vigorous stirring for 15 min. Subsequently, 4 g of NaOH was added to the solution, and vigorous stirring was continued for another 15 min. The mixture was transferred to a 50 mL Teflon-lined stainless-steel autoclave, sealed, and heated in an oven at 120 °C for 10 h. After natural cooling to ambient temperature, the resulting precipitate was washed separately with deionized water and ethanol, then dried at 60 °C for 8 h. Next, 0.2 g of the dried sample was dissolved in a 50 mL mixed solution of deionized water and ethanol (volume ratio of ethanol to deionized water = 1:1) under vigorous stirring for 30 min. Then, 2 g of NaBH₄ was added to the mixture, and stirring was continued vigorously for an additional 2 h. The final precipitate was washed separately with deionized water and ethanol, followed by drying at 60 °C for 6 h to obtain metallic Bi.

Preparation of Bi₂O₃ nanosheets

Firstly, 5.21 g of Bi(NO₃)₃·5H₂O was dissolved in 27 mL of 1 M nitric acid to form solution A. Separately, 6.78 g of Na₂CO₃ was dissolved in 107 mL of deionized water to prepare solution B. Solution B was then added dropwise into vigorously stirred solution A, followed by continuous vigorous stirring for 15 min. The mixture was allowed to stand at 60 °C for 12 h. The resulting precipitate was washed with deionized water and ethanol, respectively, and dried at 60 °C for 6 h. The dried sample was calcined at 350 °C with a heating rate of 10 °C min⁻¹ for 30 min to obtain Bi₂O₃.

Preparation of Bi₁₉Br₃S₂₇ nanorods

2.2436 g of BiBr₃, 0.3502 g of CH₄N₂S (thiourea), and 0.2 g of EDTA were weighed and dissolved together in 60 mL of ethanol under vigorous stirring for 1 h. The mixture was transferred to a 100 mL Teflon-lined stainless-steel autoclave, sealed, and heated in an oven at 180 °C for 72 h. After natural cooling to ambient temperature, the resulting precipitate was washed separately with deionized water and ethanol, then dried at 80 °C for 4 h to obtain Bi₁₉Br₃S₂₇.

Preparation of Bi/Bi₂O₃

Bi/Bi₂O₃ (BBO) was synthesized via an in-situ reduction method. Specifically, 0.2 g of Bi₁₉Br₃S₂₇ was dissolved in 50 mL of a mixed solution of deionized water and anhydrous ethanol (volume ratio of ethanol to deionized water = 1:1) under vigorous stirring for 30 min. Subsequently, 2 g of NaBH₄ was dissolved in 10 mL of deionized water, and the NaBH₄ solution was slowly added dropwise into the Bi₁₉Br₃S₂₇ suspension while maintaining vigorous stirring. After complete addition, stirring was continued for 2, 4, or 6 h. The resulting precipitate was washed five times with deionized water and dried at 60 °C for 6 h to obtain Bi/Bi₂O₃ (BBO-2, BBO-4, and BBO-6).

Photocatalytic CO₂ reduction activity measurement

In a typical reaction, 20 mg of the catalyst was dispersed in a centrifuge tube containing 2 mL of deionized water and sonicated for 5 min. The dispersed catalyst suspension was then drop-cast onto a quartz plate using a pipette, and the quartz plate was dried in a vacuum oven at 60 °C until completely dry. The catalyst-loaded quartz plate was placed on a quartz column inside a photoreactor (400 mL), followed by the addition of 5 mL of deionized water and sealing of the reactor. High-purity CO₂ (99.999%) was used as the carbon source. A multi-channel atmosphere controller evacuated and purged the sealed photoreactor with high-purity CO₂ 10 times to ensure a CO₂-saturated environment. The sealed photoreactor was then irradiated under a 300 W xenon lamp for 4 h. Gas samples (1 mL) were collected hourly using a microsyringe and analyzed by gas chromatography (GC) equipped with thermal conductivity (TCD) and flame ionization (FID) detectors. The CH₃OH product quantitative analysis was employed on an on-line gas chromatograph (GC-Agilent Technologies 8860, USA) equipped with thermal conductivity detector (TCD) and a flame ionized detector (FID). Furthermore, FID detector was connected with a HP-PLOT Q capillary column (Agilent, length 30 m, ID 0.53 mm, film 40 μm) for separation of C1-C6 paraffin and olefin hydrocarbons, alcohols and oxygenated compounds. The quantification of the products was based on the external standard and the use of a calibration curve.

Electrochemical and photoelectrochemical measurements

Transient photocurrent response, electrochemical impedance spectroscopy, and Mott–Schottky measurements were performed using a CHI760E electrochemical workstation (CHI, China) in a three-electrode quartz cell. A 300 W xenon lamp (MC-PF300C, Beijing MerryChange Technology Co., Ltd, China) served as the light source. The three-electrode system consisted of an Ag/AgCl reference electrode, a Pt wire counter electrode, and a working electrode made from the photocatalyst-coated fluorine-doped tin oxide (FTO, 1.5 × 2.2 cm). The electrolyte was 0.5 M Na₂SO₄ aqueous solution. The working electrode was fabricated as follows: 10 mg of the as-prepared photocatalyst was dispersed in 1 mL of ethanol, followed by the addition of 40 μL of Nafion solution. The mixture was ultrasonicated for 10 min. Then, 400 μL of the resultant suspension was drop-cast onto the FTO substrate and dried under near-infrared (NIR) irradiation.

Computational method

The first-principle calculations were carried out within DFT, using the Vienna Ab Initio Simulation Package code⁴⁰. The interaction between atomic cores and valence electrons was described by the frozen-core projector augmented-wave method⁴¹, while the electron-electron exchange and correlation interactions were treated with the generalized gradient approximation in the form of the Perdew–Burke–Ernzerhof functional⁴². A kinetic energy cutoff of 500 eV was adopted to ensure full convergence of the calculations. Supplementary Data 1–4 for details on the models of Bi, Bi₂O₃, and Bi/Bi₂O₃ heterojunction with different sites have been built based on a collection of characterization results. Owing to the large cell of Bi (012)/Bi₂O₃(100) heterostructure, the Brillouin zone was sampled with the k-grid of 2 × 2 × 1. Structural relaxation was performed using the conjugate gradient (CG) method, by simultaneously minimizing the total energy and interatomic forces. The convergence threshold for the total energy was set to 10⁻⁵eV. To properly describe van der Waals interactions, the dispersion-corrected DFT-D3 scheme was employed. In addition, a vacuum layer of 1.5 nm was introduced along the c-axis to avoid spurious interactions⁴³. Gibbs free energies for each gaseous and adsorbed species were calculated at 298.15 K, according to the expression:

$${{{\rm{G}}}}={{{{\rm{E}}}}}_{{{{\rm{DFT}}}}}+{{{{\rm{E}}}}}_{{{{\rm{ZPE}}}}}-{{{\rm{TS}}}}$$

$${{{{\rm{E}}}}}_{{{{\rm{ZPE}}}}}={\sum }_{i}\frac{1}{2}{{\hslash }}v$$

$${\Theta }_{{{{\rm{i}}}}}={{\hslash }}{vi}/{{{\rm{k}}}}$$

$${{{\rm{S}}}}={\sum }_{i}{{{\rm{R}}}}[{{\mathrm{ln}}}{\left(1-{{{{\rm{e}}}}}^{-\Theta {{{\rm{i}}}}/{{{\rm{T}}}}}\right)}^{-1}+\Theta {{{\rm{iT}}}}{\left({{{\rm{e}}}}\Theta {{{{\rm{i}}}}}^{/{{{\rm{T}}}}}-1\right)}^{-1}]$$

where E_DFT is the electronic energy calculated for specified geometrical structures, E_ZPE the zero-point energy, S is the entropy, $\hslash$ is the Planck constant, $v$ is the computed vibrational frequencies, Θ_i is the characteristic temperature of vibration, k is the Boltzmann constant, and R is the molar gas constant. For adsorbates, all 3N degrees of freedom were treated as frustrated harmonic vibrations.

For each step, the ΔG is the difference of product or intermediate and reactant.

ΔG[CH₃OH*] = G[CH₃OH *] + G[H₂O (g)]−(G[*] + G[CO₂] + 6*G[H⁺+e⁻])

$$\Delta {{{\rm{G}}}}[{{{{\rm{CH}}}}}_{3}{{{\rm{O}}}}\ast ]= {{{\rm{G}}}}[{{{{\rm{CH}}}}}_{3}{{{\rm{O}}}}\ast ]+{{{\rm{G}}}}[{{{{\rm{H}}}}}^{+}+{{{{\rm{e}}}}}^{-}]+{{{\rm{G}}}}[{{{{\rm{H}}}}}_{2}{{{\rm{O}}}}({{{\rm{g}}}})]\\ - ({{{\rm{G}}}}[\ast ]+{{{\rm{G}}}}[{{{{\rm{CO}}}}}_{2}]+6\ast {{{\rm{G}}}}[{{{{\rm{H}}}}}^{+}+{{{{\rm{e}}}}}^{-}])$$

$$\Delta {{{\rm{G}}}}[{{{{\rm{CH}}}}}_{2}{{{\rm{O}}}}\ast ]= {{{\rm{G}}}}[{{{{\rm{CH}}}}}_{2}{{{\rm{O}}}}\ast ]+2\ast {{{\rm{G}}}}[{{{{\rm{H}}}}}^{+}+{{{{\rm{e}}}}}^{-}]+{{{\rm{G}}}}[{{{{\rm{H}}}}}_{2}{{{\rm{O}}}}({{{\rm{g}}}})] \\ - ({{{\rm{G}}}}[\ast ]+{{{\rm{G}}}}[{{{{\rm{CO}}}}}_{2}]+6\ast {{{\rm{G}}}}[{{{{\rm{H}}}}}^{+}+{{{{\rm{e}}}}}^{-}])$$

$$\Delta {{{\rm{G}}}}[{{{\rm{CHO}}}}\ast ]= {{{\rm{G}}}}[{{{\rm{CHO}}}}\ast ]+3\ast {{{\rm{G}}}}[{{{{\rm{H}}}}}^{+}+{{{{\rm{e}}}}}^{-}]+{{{\rm{G}}}}[{{{{\rm{H}}}}}_{2}{{{\rm{O}}}}({{{\rm{g}}}})]\\ - ({{{\rm{G}}}}[\ast ]+{{{\rm{G}}}}[{{{{\rm{CO}}}}}_{2}]+6\ast {{{\rm{G}}}}[{{{{\rm{H}}}}}^{+}+{{{{\rm{e}}}}}^{-}])$$

$$\Delta {{{\rm{G}}}}[{{{\rm{CO}}}}\ast ]= {{{\rm{G}}}}[{{{\rm{CO}}}}\ast ]+4\ast {{{\rm{G}}}}[{{{{\rm{H}}}}}^{+}+{{{{\rm{e}}}}}^{-}]+{{{\rm{G}}}}[{{{{\rm{H}}}}}_{2}{{{\rm{O}}}}({{{\rm{g}}}})] \\ - ({{{\rm{G}}}}[\ast ]+{{{\rm{G}}}}[{{{{\rm{CO}}}}}_{2}]+6\ast {{{\rm{G}}}}[{{{{\rm{H}}}}}^{+}+{{{{\rm{e}}}}}^{-}])$$

$$\Delta {{{\rm{G}}}}[{{{\rm{COOH}}}}\ast ]= {{{\rm{G}}}}[{{{\rm{COOH}}}}\ast ]+5\ast {{{\rm{G}}}}[{{{{\rm{H}}}}}^{+}+{{{{\rm{e}}}}}^{-}] \\ - ({{{\rm{G}}}}[\ast ]+{{{\rm{G}}}}[{{{{\rm{CO}}}}}_{2}]+6\ast {{{\rm{G}}}}[{{{{\rm{H}}}}}^{+}+{{{{\rm{e}}}}}^{-}])$$

$${{{\rm{G}}}}[{{{{\rm{H}}}}}^{+}+{{{{\rm{e}}}}}^{-}]=1/2{{{\rm{G}}}}[{{{{\rm{H}}}}}_{2}]-{{{\rm{eU}}}}$$

where * is the substrate, U is the applied overpotential and e is elementary charge. In this study, U = 0 V versus reversible hydrogen electrode.