ReS₂ Nanosheets with In Situ Formed Sulfur Vacancies for Efficient and Highly Selective Photocatalytic CO₂ Reduction

Abstract

Artificial photosynthesis can provide valuable fuels and positively impact greenhouse effects, via transforming carbon dioxide (CO₂) and water (H₂O) into hydrocarbons using semiconductor‐based photocatalysts. However, the inefficient charge‐carrier dissociation and transportation as well as the lack of surface active sites are two major drawbacks to boosting their activity and selectivity in photocatalytic CO₂ reduction. Recently, ReS₂ has received tremendous attention in the photocatalysis area due to its intriguing physicochemical properties. Nevertheless, the application of ReS₂ in photocatalytic CO₂ reduction is scarcely covered. Herein, a heterojunction formed between ReS₂ nanosheets and CdS nanoparticles is reported, achieving an apparently raised CO production of 7.1 μmol g⁻¹ and high selectivity of 93.4%. The as‐prepared ReS₂/CdS heterojunction exhibits strengthened visible‐light absorption, high‐efficiency electron–hole pair separation/transfer, and increased adsorption/activation/reduction of CO₂ on in situ created sulfur vacancies of ReS₂, thus all favoring CO₂ photoreduction. These are corroborated by advanced characterization techniques, e.g., synchrotron‐based X‐ray absorption near‐edge structure, and density functional theory–based computations. The findings will be of broad interest in practical design and fabrication of surface active sites and semiconductor heterojunctions for applications in catalysis, electronics, and optoelectronics.

Herein, a heterojunction formed between ReS₂ nanosheets and CdS nanoparticles exhibits enhanced photocatalytic CO production together with excellent CO production selectivity. Advanced characterization, e.g., synchrotron‐based X‐ray absorption near‐edge structure, and theoretical computations disclose that the pronounced electronic coupling between ReS₂ and CdS and in situ generated sulfur vacancies contribute to the apparently improved light‐induced CO₂‐to‐CO conversion performance.

1. Introduction

CO₂ emission leads to severe global greenhouse effect. Therefore, various strategies have been developed to relieve this process, including CO₂ fixation and cyclic utilization.^{[ 1} , ² , ³ , ^{4 ]} Meanwhile, the application of clean and renewable energy to fix CO₂ can not only promote the carbon cycle but also relieve the global energy crisis. Semiconductor‐based photocatalytic carbon dioxide (CO₂) conversion represents a carbon‐neutral and sustainable strategy to generate fuels and chemicals using renewable and clean solar energy.^{[ 1} , ² , ³ , ⁴ , ^{5 ]} This process is fundamentally impacted by three steps: 1) light absorption to excite semiconductor‐based photocatalysts, 2) photogenerated electron–hole pair separation and transfer efficiency, and 3) redox reactions on the surface of the photocatalyst.^{[ 6} , ^{7 ]} To improve photocatalytic CO₂ reduction, researchers have generally focused on approaches including introducing defects in the crystal lattice,^{[ 8} , ⁹ , ¹⁰ , ¹¹ , ¹² , ^{13 ]} loading metal cocatalysts,^{[ 14} , ^{15 ]} exposing highly active facets,^{[ 16} , ^{17 ]} and fabricating heterojunctions.^{[ 7} , ^{18 ]} Among these, forming heterojunctions in composites is deemed as an effective strategy due to efficient suppression of charge carrier recombination and highly promoted migration efficiency.^{[ 19} , ^{20 ]} Some cocatalysts broaden the light absorption spectrum to utilize long‐wavelength light.^{[ 21} , ^{22 ]} However, CO₂ molecules are highly thermodynamically stable with a bond energy of 750 kJ mol⁻¹ for C—O.^{[ 6 ]} This implies that the dissociation of CO₂ requires high energy via conventional methods.^{[ 6 ]} The activation of CO₂ molecules on the surface of photocatalysts relies on various active sites, e.g., functional groups, frustrated Lewis pairs, single atoms and vacancies. These possess an affinity for CO₂ and water (H₂O) adsorption and activation. These are however rare on the perfect surfaces of photocatalysts. The deliberate creation of active sites on the surfaces of photocatalysts is therefore a major research approach to practically realize effective photocatalysis.^{[ 6 ]}

Transition metal dichalcogenides (TMDs) have received significant attention in catalysis, rechargeable batteries, and sensing devices.^{[ 23} , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ^{28 ]} For example, Zhou et al. reported that FeS ultrathin nanosheets on a carbon fiber cloth achieved highly efficient hydrogen evolution due to the phase transition triggered by illumination at room temperature.^{[ 29 ]} Similarly, Fu et al. prepared MoReS₃ with a layered structure, a new type of TMDs, and it showed excellent hydrogen evolution in electrocatalysis.^{[ 30 ]} As a new type of TMDs, ReS₂ has been studied in photocatalytic hydrogen evolution (PHE), both experimentally and theoretically.^{[ 31} , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ^{38 ]} Zhang et al.^{[ 34 ]} reported that ReS₂ exhibited significant performance in PHE, a two‐electron catalytic reaction. This is because free electrons are captured by tightly bound excitons to form trions that constrain recombination and thereby promote performance. The corresponding computation indicates that the adsorption energy of the H atoms exhibits a small absolute value (−0.5 eV), revealing the highly efficient PHE performance. With a narrow bandgap of 1.7 eV, ReS₂ possesses strong absorption in the visible‐light range, thus benefiting photocatalytic activity. In addition, ReS₂ exhibits an automatic transition of hydrophobicity–hydrophilicity before and following visible‐light illumination.^{[ 35 ]} This is because defects on its surface change the adsorption configuration of H₂O and oxygen (O₂) to form hydroxyl groups leading to the transmission of hydrophilicity.^{[ 35 ]} This adsorption configuration impacts adsorption of CO₂ and H₂O in CO₂ photoreduction. Zhou et al.^{[ 36 ]} reported that Re sites can absorb hydrogen as the intermediate for CO₂ hydrogenation. Findings suggest ReS₂ nanosheets (NSs) are a potential photocatalyst for CO₂ reduction. A drawback however with most TMDs is that the reactive sites are located at the edge of the NSs, because of the hanging bond created by breakage of the bond between the transmitting metal and sulfur. These limited active sites on the edge and inert surface elsewhere are a significant impediment to applying TMDs as photocatalysts. Therefore, the activation of the inert surface through the introduction of defects on ReS₂ NSs is regarded as an effective strategy to boost their photocatalytic performance.

Herein, we report a simple self‐assembly approach to synthesize a heterojunction composed of ReS₂ NSs and CdS nanoparticles (NPs) for photocatalytic CO₂ reduction. The optimized ReS₂/CdS heterostructure exhibits a boosted photocatalytic CO₂‐to‐CO conversion activity of 7.1 μmol g⁻¹ accompanied by a high selectivity of 93.4%. The enhanced performance is aroused by high‐efficiency interfacial charge transfer between ReS₂ and CdS as well as in situ formed sulfur vacancies on the ReS₂ surface. Results from advanced characterizations, e.g., synchrotron‐based X‐ray absorption near‐edge structure (XANES), together with X‐ray photoelectron spectroscopy (XPS), confirm the strong electronic coupling between ReS₂ and CdS. Density functional theory (DFT)‐based computations highlight the adsorption and activation of CO₂ on the ReS₂ surface with in situ formed sulfur vacancies based on electron transfer and change in C—O bond length and angle. Our work will be of immediate practical interest to a wide range of researchers for the design and synthesis of nanostructured materials in the field of energy conversion and storage.

2. Results and Discussion

ReS₂ NSs were prepared by exfoliating commercial bulk ReS₂ under ultrasonication in deionized water, and CdS NPs were fabricated by a hydrothermal method. Then, the heterojunctions of ReS₂ and CdS were synthesized by physical mixing. The as‐prepared samples were denoted as CdS, CR4, CR8 and CR12, respectively, according to the added volumes of ReS₂ NSs suspension (0, 4, 8 and 12 mL, respectively). The X‐ray diffraction (XRD) patterns of all the as‐prepared samples are shown in Figure S1, Supporting Information. The pattern displayed is ascribed to cubic‐phase CdS (PDF #10‐0454). A weak peak at 14.6° is attributed to the presence of ReS₂ NSs. The intensity of this peak is enhanced with increasing content of ReS₂ in the heterojunction. No apparent alteration of peak positions and intensities is observed for cubic‐phase CdS after its combination with ReS₂, suggesting the weak interaction formed via physical mixing at room temperature does not change the crystal structure.

Furthermore, the morphologies and microstructures of CdS NPs, ReS₂ NSs, and CR12 are characterized by aberration‐corrected high‐angle annular dark field scanning transmission electron microscopy (HAADF‐STEM), transmission electron microscopy (TEM), and HAADF‐STEM elemental mapping. The HAADF‐STEM image of CdS NPs (Figure  1a) exhibits a lattice spacing value of 0.33 nm, ascribed to the (111) facet of cubic‐phase CdS. In addition, the HAADF‐STEM image of a ReS₂ NSs (Figure 1b) shows the typical Re₄ diamond chain (DC) structure (yellow dot line square). The spacing values between the Re₄ diamond in the a [100] and b [010] directions are 0.35 and 0.31 nm, respectively (Figure 1b). This is also displayed in the simulated atomic structure of the ReS₂ monolayer (Figure 1c). It is reported that Re₄ DCs can act as reactive sites for water splitting. This is because these favor water adsorption and activation, and possibly assist proton transfer to participate in CO₂ photoreduction.^{[ 33} , ^{34 ]} The thickness of the ReS₂ NSs was further confirmed to be ≈6.5 nm using atomic force microscopy (AFM; Figure 1d). The ultrathin thickness of the ReS₂ NSs not only endows them with a large surface area to form strong electronic coupling with other materials, but also increases the number of exposed active sites toward catalytic reactions. The combination of a ReS₂ NSs and CdS NPs in CR12 is shown in Figure 1e. Moreover, the HAADF‐STEM image and the corresponding energy dispersive X‐ray (EDX) elemental mapping images further corroborate the hybridization of the ReS₂ NSs and CdS NPs (Figure 1f) in CR12. The colors red, yellow, and green in Figure 1f represent, respectively, the distribution of Cd, S, and Re elements, in agreement with the HAADF‐STEM image.

Figure 1.

Aberration‐corrected HAADF‐STEM images of a) CdS NPs and b) ReS₂ NSs. c) Simulated model for ReS₂. The red and yellow colored spheres denote Re and S, respectively, and the blue‐dotted line denotes the Re₄ diamond chain. d) AFM image of ReS₂ NSs and measured thickness of ReS₂. e) TEM of ReS₂ NSs and CdS NPs. f) EDX mapping of CR12 from red‐dotted rectangle of (e).

Moreover, both surface‐sensitive XPS and synchrotron‐based XANES were conducted to disclose the interactions between CdS and ReS₂ in CR12. The high‐resolution Re 4f XPS spectrum for CR12 (Figure  2a) showed a shift of ≈0.5 eV toward the direction of lower binding energy compared with that for pure ReS₂ (Figure 2b), suggesting electron transfer from CdS to ReS₂ in CR12. In addition, the high‐resolution XPS spectrum of Cd 3 d for CR12 shifts to the direction of higher binding energy (Figure 2c), also indicating the electron migration from CdS to ReS₂ in CR12. Furthermore, synchrotron‐based XANES was performed (Figure 2d). The S K‐edge XANES of CR12 indicates a shift toward high photon energy direction, in comparison to that for CdS. This also supports the electron transfer from CdS to ReS₂ in CR12.

Figure 2.

a) High‐resolution XPS spectrum of Re 4f for CR12. b) High‐resolution XPS spectrum of Re 4f for ReS₂. c) High‐resolution XPS spectrum of Cd 3d for CdS and CR12. d) XANES S K edge of CdS and CR12.

Solid–gas phase photocatalytic CO₂ reduction activities of the as‐prepared samples were examined under visible‐light irradiation (λ ≥ 420 nm). As shown in Figure  3a, CdS shows a limited photocatalytic CO₂ reduction activity with CO and CH₄ production of 2.3 and 1.1 μmol g⁻¹ over 7 h (Figure S2, supporting information). The coupling of ReS₂ and CdS (CR4, CR8, and CR12) leads to apparent enhancement of both activity and selectivity in visible‐light‐driven CO₂‐to‐CO conversion (Figure 3a). In particular, CR12 shows the highest photocatalytic CO₂‐to‐CO conversion activity of 7.1 μmol g⁻¹ and selectivity 93.4%, ≈309%, and 138% times higher than those of CdS alone. This result for the first time demonstrates that coupling with a ReS₂ NS could significantly boost both activity and selectivity in photocatalytic CO₂‐to‐CO conversion. In addition, we conducted three blank experiments under the same conditions but purged with ultra‐high‐purity argon gas instead of CO₂, without visible‐light illumination and without a photocatalyst, respectively. These contrast experiments show no photocatalytic CO₂ conversion performance, suggesting that the products (CO and CH₄) are generated from photoinduced CO₂ conversion. Furthermore, the stability of CR12 was studied via four‐cycle testing with 7 h per cycle (Figure 3b). No apparent deterioration in photocatalytic CO₂ reduction performance of CR12 was found over a 28 h test. The TEM image and EDX spectrum (Figure S3, Supporting Information) of CR12 after photocatalytic CO₂ reduction exhibited no obvious difference from those before reaction. This finding suggests that no apparent change in morphology and chemical composition of CR12 occurs after the 28 h photocatalytic reaction.

Figure 3.

a) Photocatalytic CO₂ reduction activities of CdS, CR4, CR8, and CR12 under visible‐light illumination (λ ≥ 420 nm). b) Photocatalytic CO₂ reduction activity of CR12 recycle test with 7 h per cycle test. c) UV–vis diffuse reflectance spectroscopy and band edge for CdS, CR4, CR8, and CR12. d) TSPL spectra for CdS and CR12.

To further investigate the origin of the enhanced activity and selectivity, as well as the reaction mechanism in CR12 for photocatalytic CO₂ reduction, both experimental characterizations and DFT‐based theoretical calculations were conducted. The light absorption capacity was studied by UV–vis diffuse reflectance spectroscopy. As shown in Figure 3c, increased absorption in the range of 460–800 nm is attributed to the presence of ReS₂ NSs. However, the adsorption edge of CR12 (457 nm) does not display any shift. This finding suggests that the bandgap (2.24 eV) is not changed following combination with the ReS₂ NSs. The bandgap of ReS₂ NSs is 1.7 eV (Figure S4, Supporting Information). This induces the “bulge” at ≈600–800 nm in the UV–vis spectrum of CR12. Such an improved visible‐light absorption probably contributes to the raised activity in photocatalytic CO₂ reduction.

A range of characterization methods, e.g., steady‐state photoluminescence (PL) spectroscopy, transient‐state photoluminescence (TSPL) spectroscopy, electrochemical impedance spectroscopy (EIS), and transient photocurrent (TPC) density measurement, were executed to probe the efficiency of charge‐carrier separation and transportation. The steady‐state PL (Figure S5, Supporting Information) intensity of CR12 is significantly lower than that of CdS. This finding is attributed to the oppressed charge‐carrier recombination after coupling with the ReS₂ NSs in CR12. After fitting of the TSPL curves in Figure 3 d, elongated lifetimes of charge carriers (τ ₁ = 0.33 ns; τ ₂ = 3.44 ns; τ _ave = 1.40 ns) for CR12 were observed, in comparison to those for CdS (τ ₁ = 0.29 ns; τ ₂ = 3.08 ns; τ _ave = 1.04 ns). This also supports the more effective dissociation and migration of photogenerated electrons and holes in CR12. Furthermore, the EIS spectra (Figure S6a, Supporting Information) exhibit a smaller semicircle radius for the Nyquist plot together with a decreased charge‐transfer resistance (R _t = 4672 Ω) for CR12 in contrast with that of CdS (R _t = 4830 Ω). This finding suggests a faster charge‐carrier migration rate in CR12. In addition, CR12 exhibits a greater TPC density than CdS (Figure S6b, Supporting Information). This also confirms the more efficient dissociation of light‐induced excitons, in agreement with the earlier PL and EIS spectra results.

To investigate the CO₂ adsorption, activation, and reduction process, XPS and Raman characterizations were conducted for pristine ReS₂, RS1, RS2, and RS3 (see details in Section 1.5, Supporting Information). The C 1s peaks in Figure  4a–c mainly consist of a peak located at 284.6 eV, ascribed to the contaminated carbon. All other peaks were calibrated using this peak. After deconvolution of the C 1s peaks in Figure 4a–c, four satellite peaks can be obtained and attributed to C—OH, *COOH, b‐CO₂ (chemisorbed and bent CO₂), and l‐CO₂ (physisorbed and linear CO₂), respectively.^{[ 39 ]} Among these adsorption configurations, *COOH is deemed as the precursor for CO, consistent with the major product (CO) of this work.^{[ 40} , ^{41 ]} The contents (mol%) acquired based on the areas of four satellite C 1s peaks are provided in Table S1, Supporting Information). RS1 shows the highest contents of C 1s peaks associated with *COOH and b‐CO₂, in comparison to RS2 and RS3. This is because the water vapor can facilitate the adsorption of CO₂.^{[ 42} , ⁴³ , ^{44 ]} The aforementioned results indicate that the adsorption of CO₂ on the surface of the ReS₂ NSs is obviously enhanced by visible‐light illumination. To further study the key role of illumination on CO₂ adsorption, the high‐resolution XPS spectra of S 2p for pristine ReS₂ NSs, RS1, RS2, and RS3 were collected. The ReS₂ NSs exhibits two peaks located at 162.05 and 163.18 eV, ascribed to the S 2p _3/2 and S 2p _1/2 (Figure S7, Supporting Information). In comparison, after light illumination, satellite peaks located at 162.7 and 163.85 eV appear (Figure 4d–f), attributed to the presence of sulfur vacancy induced by visible‐light illumination. Moreover, the high‐resolution Re 4f XPS spectrum (Figure S7, Supporting Information) of the pristine ReS₂ NSs only exhibits two peaks at 41.65 and 44.06 eV, attributed to the Re 4f _7/2 and Re 4f _5/2 peaks, respectively. In comparison, Re 4f satellite peaks located at lower binding energy positions are observed in the high‐resolution Re 4f XPS spectra of RS1, RS2, and RS3 (Figure S8a–c, Supporting Information), also implying the presence of sulfur vacancies in ReS₂ after light illumination. Furthermore, the aberration‐corrected HAADF‐STEM images of RS1 and pristine ReS₂ are displayed in Figure S9a,b, respectively. In contrast to pristine ReS₂, RS1 exhibits more defects, e.g., vacancies and pores, on its surface after the light illumination. This finding is in agreement with the aforementioned XPS results.

Figure 4.

a,d) High‐resolution XPS spectrum for C 1s and S 2p for RS1. b,e) High‐resolution XPS spectrum for C 1s and S 2p for RS2. c,f) High‐resolution XPS spectrum for C 1s and S 2p for RS3. g) Side‐view (elevation) and h) top‐view (plan) of the electron density distribution of absorbed CO₂ on V_s‐ReS₂. The red, yellow, gray, and brown colored spheres denote O, S, Re, and C atoms, respectively. The isosurface value is set to 0.002e Å⁻³. i) Schematic of photocatalytic CO₂ reduction in CR12 system under visible‐light illumination (λ ≥ 420 nm). The purple, red, yellow, orange, white, and black colored spheres denote Cd, O, S, Re, H, and C atoms, respectively.

These findings are further corroborated by the Raman spectroscopy characterization. Pristine ReS₂ displays a typical E _g‐like peak at ≈305 cm⁻¹ (Figure S10, Supporting Information). In contrast, RS1, RS2, and RS3 all exhibit a blueshift of the E _g‐like peak. This might be caused by the presence of a sulfur vacancy leading to the mass fluctuation at the S position. In addition, RS1, RS2, and RS3 all exhibit peaks at ≈1271 cm⁻¹, ascribed to *COOH (Figure S11, Supporting Information).^{[ 40} , ^{41 ]} In particular, the highest *COOH peak intensity is observed for RS1, compared to those of RS2 and RS3, in coincidence with the aforementioned XPS results (Table S1, Supporting Information). These findings were reinforced in the Fourier transform infrared spectroscopy (FTIR) spectra (Figure S12, Supporting Information). The bands at 1620 cm⁻¹ are attributed to *COOH and the bands at 1404 cm⁻¹ to the symmetric (O—C—O) vibration, indicating the presence of adsorbed CO₂ molecules on the surface of the photocatalyst. These findings are consistent with those from the XPS results.

To verify the possible CO₂ activation on the ReS₂ with the sulfur vacancy (V_s‐ReS₂), DFT‐based theoretical computation was conducted to determine the adsorption of CO₂ and the local electronic structure. As shown in Figure S13, Supporting Information, the atomic structures for V_s‐ReS₂ were constructed and optimized. The adsorption behavior of CO₂ on the surface of V_s‐ReS₂ was studied (Figure 4g,h). The adsorption energy (E _ad) of CO₂ on V_s‐ReS₂ was found to be −0.82 eV. This value demonstrates that the chemisorption of CO₂ on V_s‐ReS₂ is favorable.^{[ 45} , ^{46 ]} The CO₂ molecule loses linearity to change α(OCO) to 117.9°, and the two C—O₁ and C—O₂ bonds are lengthened to 1.209 and 1.398 Å, respectively, thereby denoting activation of the C=O bond on V_s‐ReS₂.^{[ 46} , ^{47 ]}

Bader charge analysis (Figure S14, Supporting Information) was conducted to study the charge distribution on the CO₂‐adsorbed V_s‐ReS₂ and to investigate the electronic impact. It was found that Re atoms and C atoms are electron deficient, whereas S atoms and O atoms are negatively charged (Figure S14a, Supporting Information). In particular, Re near a sulfur vacancy (Re₄, Re_7, and Re₁₃) changed more positively when CO₂ was absorbed onto V_s‐ReS₂ (Figure S14b, Supporting Information).^{[ 38} , ⁴⁶ , ^{47 ]} Charge density difference was used to visualize the electron transfer behavior. The findings showed that Re and carbon atoms donate charges to oxygen and that they are positively charged.^{[ 37} , ⁴⁶ , ^{47 ]} These findings confirm that V_s‐ReS₂ favorably impacts chemisorption and activation of CO₂ and electron transfer between CO₂ and V_s‐ReS₂.

Based on the findings from both the experimental studies and theoretical computations, a possible photocatalytic CO₂ reduction mechanism is proposed. The conduction band (CB) and valence band (VB) edge positions of CdS and ReS₂ are estimated via combining their Mott–Schottky plots (Figure S15, Supporting Information) and UV–vis diffuse reflectance spectrum (Figure S5, Supporting Information). As shown in Figure 4i, CdS and ReS₂ form a type I (straddling‐type) heterojunction. Under visible‐light illumination (λ ≥ 420 nm), the CdS NPs are photoexcited and the electrons transfer from the VB to the CB, whereas the photogenerated holes in the VB of CdS will migrate to the VB of ReS₂, where H₂O molecules are oxidized. Then, photogenerated electrons in the CB of CdS transfer to the CB of ReS₂. The ReS₂ NSs accommodates abundant active sites for the adsorption, activation, and reduction of CO₂ and H₂O to evolve CO and CH₄ (Equation ((1), (2))–(3)). The most likely pathways for the CO₂ photoreduction on the CdS/ReS₂ heterojunction are proposed as

$$*+{\text{CO}}_2+{\mathrm{e}}^{-}+{\mathrm{H}}^{+}\to {\text{COOH}}^{*}$$ (1)

$${\text{COOH}}^{*}+{\mathrm{e}}^{-}+{\mathrm{H}}^{+}\to {\text{CO}}^{*}+{\mathrm{H}}_2\mathrm{O}$$ (2)

$${\text{CO}}^{*}\to *+\text{CO}\uparrow$$ (3)

The asterisks and vertical arrows, respectively, denote reactive sites and the release of gas.

The presence of ReS₂ NSs in CR12 leads to the apparent enhancement in the selectivity of photocatalytic CO₂‐to‐CO conversion. This is probably because the electrons in ReS₂ can be captured by tightly bound excitons to form trions consisting of two electrons and one hole,^{[ 34 ]} which facilitates the two‐electron reduction reaction of CO₂‐to‐CO conversion rather than the eight‐electron reduction reaction of CH₄ production.

3. Conclusions

In summary, we have successfully prepared a heterojunction of ReS₂ NSs and CdS NPs using a facile self‐assembly method via physical mixing at room temperature. This nanocomposite exhibits a significantly boosted visible‐light photocatalytic CO production of 7.1 μmol g⁻¹ together with an increased CO₂‐to‐CO conversion selectivity of 93.4%. Such an improved photocatalytic performance originates from two factors: 1) intimate electronic interaction advancing efficient photogenerated electron–hole separation and migration and 2) in situ generated sulfur vacancies serving as active sites for CO₂ adsorption, activation, and reduction to CO. These are verified by both state‐of‐the‐art characterizations, e.g., synchrotron‐based XANES, and theoretical calculations. Our work demonstrates the promising potency of ReS₂ in light‐driven CO₂ reduction and the intriguing opportunities of applying in situ generated anion vacancies of transitional metal dichalcogenides in catalysis, electronics, and optoelectronics.

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supplementary Material