1 Introduction

Electroreduction of carbon dioxide (CO2RR) into high-value fuels and feedstocks offers a compelling pathway not only to meet the increasing energy demand, but also to alleviate the environmental crisis caused by CO2 emissions [1,2,3]. According to the gross-margin model, formate is considered to be one of the most economically feasible products in the CO2RR, which can be widely used as an important raw material in the chemical and pharmaceutical industries, as well as a potential hydrogen carrier and the liquid fuel for proton-exchange membrane fuel cell [4,5,6]. Up to now, various metal-based electrocatalysts, such as Pd, In, Hg, Pb, Cd, and Sn, have been exploited to achieve the CO2 electroreduction to formate [7,8,9,10,11,12,13]. Among these electrocatalysts, Sn-based materials have attracted considerable attention due to their advantages of earth abundance, non-toxicity, and low cost. Unfortunately, the catalytic performance of most Sn-based materials is still limited by the high energy barrier for CO2 activation, which is usually attributed to the poor stabilization of CO2*− intermediates [14,15,16,17,18]. To this end, it is of great significance to develop an efficient and durable Sn-based catalysts for the CO2 electroreduction to formate.

Given that the CO2 molecule activation is closely related to the number and inherent activity of active sites, many effective strategies have been employed to tailor the active sites of electrocatalysts for enhancing the efficiency of the CO2 electroreduction to formate [19,20,21]. The surface chemistry modification, as a powerful strategy, has attracted great interest in adjusting electronic properties of active sites to target intermediate adsorption energy as well as harvest high selectivity [22,23,24,25]. For example, Xie et al. developed a general amino acid modification approach on Cu electrodes for the selective electroreduction of CO2 toward hydrocarbons [26]. Previous theoretical calculations have confirmed that the *OCHO binding energy is closely associated with the oxophilicity of the catalyst surface, which can be achieved by modifying the surface of the electrocatalyst with oxygen atom [27]. For instance, Gao et al. reported a phenomenon that partially oxidized atomic cobalt layers effectively adjusted the electronic structure, promoted the activation of CO2, and stabilized the relevant key intermediates, thereby enhancing the efficiency of the CO2 electroreduction to liquid fuel [28]. As another example, Won et al. prepared hierarchical Sn dendrites and found that the natural oxygen content is closely related to the stability of CO2*− intermediates and the selectivity of formate [29]. To improve the catalytic performance of Sn-based materials, oxygen modification is a promising strategy to regulate the surface oxophilicity of the catalysts and further manipulate their electronic structure. In fact, most of the catalysts with surface chemical modification have undergone structural evolution of the active phase under operation conditions, leading to deviations in the understanding the nature of the active site. Therefore, monitoring the structural evolution of Sn-based catalysts with surface oxygen modification under realistic working conditions is crucial for understanding the nature of the active phase and the rational design of targeted CO2RR catalysts.

Herein, the SnS2 nanosheets arrays on the carbon paper with surface oxygen modification were rationally designed under the guidance of density function theory (DFT) to effectively electroreduce CO2 into formate and syngas (CO and H2). The introduction of oxygen into the surface of SnS2 nanosheets achieved the exposure of Sn active sites and optimal Sn electronic states, thereby enhancing the adsorption and activation of CO2. Specifically, the SnS2 nanosheets with surface oxygen modification exhibit a remarkable Faradaic efficiency of 91.6% for carbonaceous products at −0.9 V vs RHE, including 83.2% for formate production and 16.5% for syngas with the CO/H2 ratio of 1:1. Operando X-ray absorption spectroscopy unravels that the in situ surface oxygen doping into the matrix under working conditions effectively changes the local electronic state of Sn, thereby providing an optimized electronic structure to improve CO2RR performance. In addition, operando synchrotron radiation infrared spectroscopy and DFT calculations further confirm that the local electronic state of Sn is manipulated through surface oxygen modification, thereby promoting the CO2 activation and enhancing the affinity for HCOO* species.

2 Experimental Section

The experimental details are provided in Supporting Information (SI). This section briefly summarizes the synthesis measurements.

In a typical synthesis of SnS2-xOx/CC, 5 mmol of SnCl4·5H2O and 15 mmol of thioacetamide were dissolved in 40 mL of deionized water. The mixture and carbon paper (2 × 2) were then transferred into a Teflon-lined stainless-steel autoclave, followed by being heated to 190 °C for 8 h. After the mixture was cooled down naturally to room temperature, the SnS2/CC was washed by water three times and ethanol twice to remove any possible ions, followed by being dried under vacuum at 60 °C for 12 h. The SnS2-xOx/CC was prepared by placing the SnS2/CC in the muffle furnace that had been heated at 300 °C for several minutes.

3 Results and Discussion

3.1 Preparation and Characterization of SnS2-xOx/CC

At first, to gain insight into the effect of surface oxygen-injection engineering on electronic properties of SnS2 nanosheets, we conducted DFT calculations by using the SnS2 slab with/without oxygen injection as the models (Fig. 1a, b). Compared with the pristine SnS2, the surface oxygen injection leads to a new additional state near the Fermi level (Fig. 1c, d), which is beneficial to manipulate the local electronic structure of Sn and expose the active site of Sn at the edges. Notably, O 2p states also contributed the unoccupied part of these levels, making them serve as the highly catalytically active sites. Furthermore, the electronic localization functions (ELF) exhibit that the charge density is mainly derived from the S atoms for both the SnS2 with/without oxygen injection (Fig. S1). Owing to the introduction of oxygen atoms, the electron density of the whole system has undergone distinctly change, further indicating surface oxygen injection effectively tailors the local electronic structure of Sn.

Fig. 1
figure 1

Top view and side view of the model for a pristine SnS2 and b SnS2-xOx. Calculated DOS of c SnS2 and d SnS2-xOx slabs

Full size image

Then, the SnS2 nanosheets arrays with partially oxidized surface on the carbon paper (denoted as SnS2-xOx/CC) were prepared, as schematically illustrated in Fig. 2a. Specifically, the pristine SnS2 nanosheets arrays were directly grown on the carbon paper by a simple hydrothermal method. Afterward, the SnS2-xOx/CC was further synthesized by the low-temperature calcination of the as-prepared SnS2/CC under the air atmosphere. The morphology of the SnS2-xOx/CC was characterized by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). As shown in Fig. 2b, the final products present a hierarchical nanosheets arrays composed of SnS2-xOx nanosheets and flexible carbon paper. The TEM images of the SnS2 and SnS2-xOx took the nanosheet morphology (Figs. 2c and S2a), whereas the SnS2 nanosheets were completely oxidized into SnO2 nanoplatelets (Fig. S3). The high-resolution transmission electron microscopy (HRTEM) image in Fig. 2d shows that the SnS2-xOx lattice fringes with an interplanar distance of 0.32 nm indexed to the (002) facets of SnS2, confirming the as-obtained SnS2-xOx nanosheets retain its pristine crystal structure (Fig. S2b, c) [30]. Besides, there is an obvious circle of amorphous layer at the edge of the SnS2-xOx nanosheet, which is attributed to the partial oxidation on the surface of the SnS2 nanosheet. In addition, the homologous fast Fourier transform (FFT) pattern indicates the SnS2 phase recorded from [002] orientation (inset in Fig. 2e). The element of O was uniformly distributed on the whole SnS2-xOx nanosheet which can be further confirmed by the high-angle annular dark-field energy-dispersive X-ray spectroscopy (HAADF-EDS) elemental mapping and EDS spectrum (Figs. 2f and S4).

Fig. 2
figure 2

a Schematic illustration of the manufacture of SnS2-xOx nanosheets. b SEM image of SnS2-xOx/CC. c TEM image of SnS2-xOx nanosheets. d, e HRTEM image and the homologous FFT pattern of SnS2-xOx nanosheets (inset in e). f HAADF-STEM and STEM-EDX elemental mapping images of an individual SnS2-xOx nanosheet. g Fourier transform spectra of Sn K-edge for Sn foil, SnS2, SnS2-xOx, and SnO2 from EXAFS

Full size image

To further investigate the phase composition and electronic structure of the SnS2-xOx/CC, we performed X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS) measurement. As evidenced by the XRD patterns in Fig. S5, the SnS2/CC and SnS2-xOx/CC exhibited the diffraction peaks at 30.74°, 32.09°, and 44.98°, which were indexed to the (200), (101), and (211) planes of hexagonal SnS2 (JCPDS No. 23–0677) [31]. Remarkably, no additional peaks corresponding to the phases of SnO2 could be found, indicating that the surface oxygen injection did not change the crystalline phase of SnS2. We further carried out the XPS measurements to clarify the form of O existing in SnS2-xOx/CC. As disclosed by XPS survey spectra, a weak signal of O was recorded in SnS2-xOx/CC, further confirming successful introduction of O (Figs. S6 and S7). In addition, the peaks at 495.3 and 486.8 eV were attributed to Sn 3d3/2 and Sn 3d5/2 of SnS2-xOx/CC, respectively (Fig. S8) [32]. Compared with the SnS2/CC, the Sn 3d3/2 and 3d5/2 peaks for SnS2-xOx/CC shifted to higher binding energies, due to the larger electronegativity of O than that of S.

The surface-sensitive synchrotron radiation soft X-ray absorption structure (XAS) was further employed to investigate the changes in the local electronic structure of SnS2 caused by surface oxygen-injection engineering. As shown in Fig. S9, O K-edge XAS spectra for SnS2-xOx/CC and pure SnO2/CC displayed similar shapes, implying SnO2 species were formed on the surface of SnS2-xOx/CC, further confirming the surface oxygen injection successfully replaced the S atoms. In addition, S L-edge XAS spectra exhibited the two characteristic peaks located at 163.3 eV (S-Sn π* peak) and 166.6 eV (S-Sn σ* peak) observed in pure SnS2/CC and the SnS2-xOx/CC (Fig. S10). Compared with pristine SnS2/CC, the relative strength of S-Sn for SnS2-xOx/CC was slightly reduced, which was attributed to the substitution of partial S atoms on the surface of SnS2-xOx/CC by O atoms. Meanwhile, the X-ray absorption near-edge structure (XANES) measurement was employed to further investigate the effect of surface oxygen injection. Compared with pristine SnS2/CC, the white line peak of the SnS2-xOx/CC shifted to the high-E region, due to the electronegativity of O being greater than that of S, in consistent with the results of Sn 3d XPS spectrum (Fig. S11). Given that the white line peak of Sn K-edge intensity originating from the transition of the 1 s to 5p orbital, the increase in the white line peak intensity after surface oxygen injection indicates the increases in the possibility of electron transition from the 1 s-5p orbital. The above results reveal the surface oxygen injection effectively manipulates the local electronic structure of Sn.

Furthermore, the Fourier transform (FT) k2-weighted extended XAFS (EXAFS) spectrum of the Sn K-edge was employed to further reveal the effect of surface oxygen injection on the local electronic structure of Sn at the atomic level. Considering the surface oxygen injection into the SnS2 nanosheets, we performed out least-squares EXAFS curve fitting analysis for Sn by considering two backscattering paths, including Sn–S and Sn–O. Compared with the SnS2/CC, the Sn K-edge FT-EXAFS curve for SnS2-xOx/CC presented a new peak at 1.49 Å, which is ascribed to the Sn–O coordination (Fig. 2a) [33]. By quantitative EXAFS curve fitting analysis, the coordination number of Sn-S for SnS2-xOx/CC is confirmed to be 4.3, smaller than that of pristine SnS2/CC (6.0), and the coordination number of Sn–O is verified to be 2.1, further confirming surface oxygen injection successfully replaced the S atoms (Table S1 and Fig. S12). Moreover, the wavelet transform (WT) of Sn K-edge EXAFS oscillations exhibited the intensity maxima at 4.3 Å−1 and around 8.2 Å−1 of SnS2-xOx/CC, which associate with Sn–O and Sn–S contributions, respectively (Fig. S13). Taken together, the successful injection of surface oxygen effectively manipulated the local electronic structure of SnS2.

3.2 Electrocatalytic CO2RR Performances of the SnS2-xOx/CC

Surface oxygen-injection engineering provides a potential prospect for enhancing the CO2 electroreduction. The electrocatalytic CO2 reduction activities of the three Sn-based catalysts were evaluated using a three-electrode H-cell in CO2-saturated 0.5 M KHCO3. The linear sweep voltammetric (LSV) curves in Fig. S14 revealed that the SnS2-xOx nanosheets exhibited higher current density than that of pristine SnS2 nanosheets, confirming the injection of oxygen effectively enhanced the electrocatalytic activity of SnS2/CC. Particularly, the geometrical current density of SnS2-xOx/CC achieved 19.68 mA cm−2, which was 2.7 times higher than that of pristine SnS2/CC at overpotential of −0.8 V vs RHE (Fig. 2a). For these three Sn-based catalysts, H2, CO, and formate were the main catalytic products, which are quantified by online gas chromatography and 1H NMR analysis (Fig. S15). Figure 2b exhibits partial current density for carbonaceous products (CO and formate), respectively. At all applied potentials, the SnS2-xOx/CC presented the largest current density among the three electrocatalysts, demonstrating the high activity for CO2 electroreduction. As shown in Fig. 2c, the SnS2-xOx/CC displayed the highest Faradaic efficiency (FE) for carbonaceous products among the three electrocatalysts, while the pristine SnS2/CC exhibited the lowest FE value. At −0.9 V vs RHE, the SnS2-xOx/CC exhibited the FE of 91.8% for carbonaceous products, including the FE of 83.5% for formate production and the FE of 16.5% for syngas with the H2/CO ratio of 1:1. It is worth noting that such syngas ratio is optimal for multiple chemical synthesis (e.g., Fischer–Tropsch synthesis, fermentation and alcohol synthesis, and hydroformylation processes). Furthermore, the as-prepared SnS2-xOx/CC displayed an excellent durability for 10-h potentiostatic test with the less than 3% decay in current density, together with the FE for formate and CO keeping steady at −0.9 V vs RHE (Fig. 2d). The above results demonstrate that the SnS2-xOx/CC represents a promising catalyst for persistently producing formate and syngas toward CO2RR.

Inspired by surface oxygen injection to improve the CO2 electroreduction performance of SnS2/CC, we studied the microscopic reaction kinetics of the pristine SnS2/CC and SnS2-xOx/CC. Based on cyclic voltammogram measurements at different scan rates, the double-layer capacitance (Cdl) value increased from 3.27 mF cm−2 of the pristine SnS2/CC to 3.75 mF cm−2 of the SnS2-xOx/CC, indicating that surface oxygen injection effectively increases the electrochemical active surface area (ECSA) of the electrocatalysts (Figs. S16 and S17). Given that the ECSA of the electrocatalysts is positively correlated with the active sites, we have reason to believe that the surface oxygen modification effectively exposes the active site of Sn. The Tafel plots were further employed to verify the rate-limiting step of the Sn-based catalysts in the CO2RR process. The Tafel slopes of the Sn-based catalysts were all close to 118 mV dec−1, demonstrating that the activation of CO2 served as the rate-limiting step (Fig. S18) [34,35]. In addition, the Nyquist plots were used to confirm the facilitated electron transfer process [36]. The SnS2-xOx/CC displayed the charge transfer resistance (RCT) of 12.1 Ω, which was smaller than that (15.8 Ω) of SnS2/CC (Fig. S19). Therefore, surface oxygen injection effectively accelerates the charge transfer process of SnS2/CC during the CO2RR (Fig. 3).

Fig. 3
figure 3

a Geometrical current densities over SnS2/CC, SnS2-xOx/CC, and SnO2/CC. b Current densities for carbonaceous product (C-product) over SnS2/CC, SnS2-xOx/CC, and SnO2/CC. c Faradaic efficiencies for formate, H2, and CO production over SnS2/CC, SnS2-xOx/CC, and SnO2/CC. d Plot of geometrical current density (j) and Faradaic efficiencies for C-product versus time over the SnS2-xOx/CC at a constant potential of −0.9 V vs RHE

Full size image

3.3 Operando X-ray Absorption Spectroscopy Study

Given that the bulk phase stability of transition metal chalcogenides with heat treatment is destroyed, the bulk phase is in a relatively unstable state [37]. Based on the equilibrium theory of crystalline chemistry, the catalyst in the electrolyte driven by both energetical and kinetical force will tend to freely optimize the structure of the entire bulk, so that the bulk tends to a relatively stable state [38,39,40,41]. Therefore, we employed operando XAFS measurements to monitor the structural evolutions of the SnS2-xOx/CC under realistic working conditions. Figure 4a shows the operando Sn K-edge XANES spectra at different applied potentials, along with the data for Sn foil, SnS2, and SnO2 as references. When cathodic potentials were applied, the absorption edge of Sn K-edge XANES spectra shifted toward low-E side compared to the case of the open-circuit condition, indicating the decrease in the Sn valence state during CO2RR process. Furthermore, when a cathodic potential of −0.9 V versus RHE was applied, the white line peak intensity was significantly increased in relation to the case (−0.4 V versus RHE), indicating more 5p electrons participate in the reaction. After the reaction, the white line peak approximately returned to the state (−0.4 V versus RHE), further confirming that the SnS2-xOx/CC catalyst undergone in situ reconstruction during the reaction and tended to form a relatively stable state.

Fig. 4
figure 4

a Operando XANES spectra at the Sn K-edge. b Operando k2-weighted FT spectra at the Sn K-edge. c Operando SR-FTIR spectra of the SnS2-xOx/CC under the working condition. d Schematic of the whole CO2RR mechanism on SnS2-xOx nanosheets. e Gibbs free energy diagrams for CO2 reduction to HCOOH on SnS2 and SnS2-xOx slabs. f Projected density of states (PDOS) of HCOO* adsorbed on the SnS2 and SnS2-xOx slabs. The * represents an adsorption site

Full size image

Furthermore, the EXAFS was further employed to reveal the atomic reconstruction of the SnS2-xOx/CC catalyst under working conditions (Figs. 4b and S20). At first sight, the Fourier transform curves of SnS2-xOx/CC displayed a significantly dampening in the Sn-S coordination peak and a heightening in the Sn-nonmetallic coordination peak under working condition. Specifically, at the applied potential of −0.4 V versus RHE before the occurrence of CO2RR, the EXAFS fitting results showed that the Sn–O coordination number increased from 2.5 to 3.6, which may be ascribed to the further doping of surface oxygen into the SnS2 lattice during the reaction. To further verify the above conjecture, we performed XRD on the SnS2-xOx/CC after reaction. As expected, the intensity of the diffraction peaks of SnS2 was significantly reduced after the reaction, and the characteristic peak of SnO2 appeared in the SnS2-xOx/CC after reaction (Fig. S21). Moreover, at the potential of -0.9 V versus RHE during CO2RR, Sn–O coordination number arose from 3.6 to 4.2 and the Sn-S coordination number remained unchanged, which may be attributed to the adsorption of the reaction intermediate species. When the cathode potential was removed, the Sn–O coordination number recovered to the state at −0.4 V, while the coordination number of Sn–S remained unchanged (Table S2). The above results indicated that the SnS2-xOx/CC had undergone dynamic surface reconstruction and surface oxygen doping plays a critical role under reaction conditions.

3.4 Operando Synchrotron Radiation Fourier Transform Infrared Spectroscopy Study

Operando synchrotron radiation Fourier transform infrared spectroscopy (SR-FT-IR) was further employed to investigate the catalytic mechanism for the well-designed SnS2-xOx/CC during the CO2RR. All SR-FTIR spectra were recorded with the electrocatalysts at CO2RR catalytic process (open circuit, − 0.6, − 0.7, − 0.8, and − 0.9 V) to reveal the production and transformation of key intermediates. As displayed in Fig. 4c, the monodentate carbonate groups (m-CO32−) appeared at the peaks of ~ 1520 cm−1, demonstrating that more CO2 was adsorbed on the surface of electrocatalyst with the decrease in applied voltage. Meanwhile, a new characteristic peak appeared in the SR-FTIR spectra of ~ 1694 cm−1 (CO2·− radicals) and the intensity of peak continually increased as the applied potentials decreased, indicating that the CO2 molecules adsorbed on the catalyst surface were activated to CO2·− radicals during the reaction [[42]]. Meanwhile, as the cathode potential decreases, the peak intensity at ~ 1541 cm−1 (HCOO−) increased further confirming the excellent proton trapping ability of CO2·− radicals [33]. The peaks at ~1354 cm−1 and ~1660 cm−1 is ascribed to the symmetry vibration of the HCOO* intermediates, which corresponds to the key intermediates or the products for CO2 electroreduction [43]. Based on the above-mentioned operando SR-FTIR analysis, the pathway of electroreduction from CO2-to-HCOOH conversion by the SnS2-xOx/CC could be proposed as the following reactions (Fig. 4d):

$${\text{Step 1}}:{\text{ CO}}_{2} \left( {\text{g}} \right){\mkern 1mu} + {\text{ e}}^{ - } + ^{*} \to {\text{CO}}_{2}^{*} - \left( {{\text{Activation process}}} \right)$$
(1)
$${\text{Step 2}}:{\text{ CO}}_{2} ^{* - } + {\text{ HCO}}_{3}^{ - } \left( {{\text{aq}}} \right) \to {\text{HCOO}}^{*} + {\text{ CO}}_{3}^{{2 - }} \left( {{\text{Surface reaction}}} \right)$$
(2)
$${\text{Step 3}}:{\text{ HCOO}}^{*}{\mkern 1mu} + {\text{ HCO}}_{3}^{ - } \left( {{\text{aq}}} \right)^{{}} + {\text{ e}}_{{}}^{ - } \to {\text{HCOOH }}\left( {\text{l}} \right){\mkern 1mu} + {\text{ CO}}_{3}^{{2 - }} ({\text{Desorption process}})$$
(3)

3.5 Density Functional Theory (DFT) Calculations

DFT calculations were employed to elucidate the catalytic contribution from partial oxidation at SnS2-xOx for CO2RR. The models for pristine SnS2 and SnS2-xOx were chosen for the simulation. Figure S22 shows optimized adsorption configurations of HCOO* intermediates on the armchair edges of the pristine SnS2 slab and SnS2Ox slab with distinguishable Sn-O distances. from the material [44]. Specifically, the Sn–O bond length (dSn-O) is 2.28 Å for SnS2, while the bond length dSn-O for SnS2-xOx is reduced to 2.24 Å, implying that the surface oxygen injection effectively enhances the binding to the HCOO* intermediate. Besides, DFT calculations were further conducted on the Gibbs free energy (ΔG) with multiple elementary reaction steps over SnS2 with/without oxygen injection. As exhibited in Fig. 4e, for both SnS2 and SnS2-xOx, the formation of HCOO* is further confirmed to be the rate-limiting step for formate, which is consistent with the results of Tafel slopes. For the SnS2-xOx slab, the ΔG for HCOO* formation (ΔGHCOO*) was calculated to be 0.97 eV, which is much lower than that for pristine SnS2 slab (1.31 eV), indicating surface oxygen injection enhanced the activation of CO2 and correspondingly facilitated the formation of HCOO*. To gain an in-depth insight into the nature of surface oxygen doping enhancing the intrinsic activity of SnS2, we calculated the projected density of state of HCOO* absorbed SnS2 and SnS2-xOx (Fig. 4f). In the HCOO* PDOS, the dominant features are that HCOO* exhibits strong interaction with the valence band region of SnS2 and SnS2-xOx, which leads to strong chemical adsorption. Notably, the state density of HCOO* overlaps more with the orbital of Sn (5p) in SnS2-xOx with regard to that in SnS2, and the higher occupied state of HCOO* is near the Fermi level, indicating that HCOO* has a stronger interaction with SnS2-xOx, which is consistent with the calculation result of ΔG HCOO*. Furthermore, the charge density differences calculations also show that more electrons gather around the adsorption site in SnS2-xOx, indicating that the surface oxygen injection makes the SnS2 edges exhibit a stronger affinity for HCOO* species (Fig. S23). The above results confirm that surface oxygen injection alters the local electronic structure of Sn atom with optimal ΔGHCOO* to effectively facilitate the production of formate over CO2RR. Particularly, the above theoretical calculation results are consistent with the previous experimental results.

4 Conclusions

In conclusion, we developed SnS2 nanosheets with surface oxygen modification for CO2 electroreduction to formate and syngas (CO and H2). The surface oxygen-injection engineering achieved exposure of Sn active site and optimal Sn electronic states, thereby enhancing the adsorption and activation of CO2. Surface oxygen injection on SnS2 nanosheets significantly improved electrocatalytic activity and selectivity of CO2 reduction to formate and syngas (CO and H2). Specifically, at −0.9 V vs RHE, the SnS2-xOx/CC exhibits the highest FE of 91.6% for carbonaceous products, including the FE of 83.2% for formate production and the FE of 16.5% for syngas with the H2/CO ratio of 1:1. Moreover, the as-prepared SnS2-xOx/CC displays an excellent durability for 10-h potentiostatic test with less than 4% decay in current density. Operando XAS unravels that the in situ surface oxygen doping into the matrix under working conditions effectively modulates the Sn local electronic state. Operando SR-FTIR and DFT calculations reveal that the surface oxygen doping enhanced the affinity for HCOO* species by manipulating the Sn electronic states and accelerated the CO2 activation. This work opens a span-new door for the design of advanced catalysts for CO2 electroreduction.