Nonstoichiometric In–S group yielding efficient carrier transfer pathway in In2S3 photoanode for solar water oxidation

The construction of high‐efficiency photoanodes is essential for developing outstanding photoelectrochemical (PEC) water splitting cells. Furthermore, insufficient carrier transport capabilities and sluggish surface water oxidation kinetics limit its application. Using a solvothermal annealing strategy, we prepared a nonstoichiometric In–S (NS) group on the surface of an In2S3 photoanode in situ and unexpectedly formed a type II transfer path of carrier, thereby reducing the interfacial recombination and promoting the bulk separation. First‐principles calculations and comprehensive characterizations demonstrated NS group as an excellent oxygen evolution cocatalyst (OEC) that effectively facilitated carrier transport, lowered the surface overpotential, increased the surface active site, and accelerated the surface oxygen evolution reaction kinetics by precisely altering the rate‐determining steps of * to *OH and *O to *OOH. These synergistic effects remarkably enhanced the PEC performance, with a high photocurrent density of 5.02 mA cm−2 at 1.23 V versus reversible hydrogen electrode and a negative shift in the onset potential by 310 mV. This work provides a new strategy for the in situ preparation of high‐efficiency OECs and provides ideas for constructing excellent carrier transfer and transport channels.


INTRODUCTION
Photoelectrochemical (PEC) water splitting systems have been intensively studied as the ideal technologies for sustainable energy harvesting and storage, enabling the production of green hydrogen from abundant sunlight and water. 1,2The key component of a PEC cell is a photoanode, which involves an oxygen evolution reaction (OER) with a four-electron process and determines the PEC efficiency. 3,4Metal sulfide semiconductors have narrow band gaps and long average carrier diffusion lengths owing to the constituent S 3p orbital-dependent valence bands, making them promising candidates for photoanodes. 5Furthermore, carrier transport dynamics with poor photogenerated electron-hole separation, sluggish surface water oxidation kinetics, and insufficient carrier transport abilities considerably impact the efficiency of metal sulfide-based photoanodes. 6o address these issues, various dopants (e.g., O, N, P, Zn, and In) [7][8][9][10][11][12] and vacancies (e.g., Zn, S, and S 2 2− ) [13][14][15] are used to improve the bulk transfer behavior and promote electron-hole pair separation in metal sulfides, with some even boosting the OER dynamics because of the surface modification effects.For instance, oxidizing In 2 S 3 and CdS regulates their energy band positions, prolongs their charge carrier lifetimes, and increases the electron densities for efficient photoexcited carrier transport and separation. 7,16Furthermore, introducing S vacancies into CdIn 2 S 4 or ZnIn 2 S 4 photoanodes restrains the surface state distribution, bringing charge accumulations on adjacent atoms and resulting in decreased potential in the rate-determining step of OER. 13,17Doping and vacancy defects in metal sulfides, especially for those introducing S vacancies or O doping, are effective strategies for enhancing carrier transport dynamics.Heteroatom doping or vacancies do not always favor carrier transport and introduce impurities that lead to severe carrier recombination and hinder photogenerated electron-hole pair transfer. 18esides, loading oxygen evolution catalysts (OECs), such as FeOOH, 19 NiOOH, 20 Co-Pi, 21 CoO x , 22 O-S bonds, 23 and V 13 O 16 , 24 affect the surface catalytic behavior of carriers by providing more redox reaction active sites, reducing the OER barrier, and promoting surface hole transfers, making it one of the most direct and attractive strategies to solve the poor dynamics of carriers.Interestingly, some OECs, such as conjugated polycarbazole frameworks, 25 metal-organic frameworks, 26 layered double hydroxides, [27][28][29][30] and clusters, 31 can also simultaneously construct heterojunctions to form band-matched semiconductor/OEC interfaces (type II or Z-scheme) that reduce bulk recombination and improve the bulk transfer behavior of photogenerated carriers.Nevertheless, current OECs are often too thick, affecting the light absorption of the semiconductor material and leading to more carrier recombinations. 32Particularly, OEC loading also causes serious interfacial recombination problems.Multiple strategies, such as heterojunction constructions, 33 surface modifications, 34 nanostructure engineering, 35 interfacial bond designs, 36 and crystal facet regulation, 37 have been used to solve the transfer kinetics of photogenerated carriers and achieve certain results.Cho and coworkers reviewed the latest progress in surface regulation, indicating that constructing the optimal semiconductor surface is crucial for developing efficient photoelectrodes for water splitting. 38However, the question of simultaneously developing efficient bulk and surface transport dynamics still restricts the development of PEC water splitting.Specifically, constructing an efficient carrier transport channel through a simple in situ method that avoids excessive carrier recombination remains a difficult task with great challenges.
Herein, a nonstoichiometric In-S (NS) group was formed in situ on the surface of an In 2 S 3 photoanode using a simple solvothermal annealing strategy to optimize the photogenerated carrier transfer.This in situ synthesis resulted in a type II heterostructure comprising a high-quality interface, promoting the bulk separation of the carriers and reducing the interfacial recombination.Additionally, the amorphous characteristics of NS group accelerated the transfer of holes and exposed more reaction sites.Moreover, the existing NS group as an efficient OEC promoted the surface transfer of the photogenerated holes, reduced the overpotential (η) of the OER, and facilitated the surface catalytic reaction kinetics by precisely affecting the first and third steps (formation of *OH and *OOH) of the four-electron reaction.Finally, the photoanode with optimal carrier transfer dynamics exhibited a considerably higher photocurrent (J) of 5.02 mA cm −2 at 1.23 V versus reversible hydrogen electrode (V RHE ) and the onset potential (V on ) shifted negatively by 310 mV, which is higher than those of current sulfide-based photoanode devices (Table S1).

Synthesize and structural characterization of photoanode
A solvothermal annealing strategy was used to fabricate hybridized photoanode with an NS-rich surface.As shown in Figure 1A, the pyramidal In 2 S 3 was hydrothermally supported on an F-doped SnO 2 (FTO) glass substrate.Next, a methanol (CH 3 OH) solution containing vanadyl acetylacetonate (VO(acac) 2 ) was dropped onto In 2 S 3 surface, which was then placed in a muffle furnace for high-temperature annealing (In 2 S 3 /NA) before being finally immersed in NaOH solution to obtain NS-modified In 2 S 3 (In 2 S 3 /NS).Figures 1B and S1 show scanning electron microscopy (SEM) images of In 2 S 3 with a nanopyramid morphology.The high-temperature annealing resulted in a thicker film of approximately 150 nm covering the In 2 S 3 /NA surface (Figure S2).X-ray diffraction (XRD) (Figure S3) confirmed that the thick film was VO 2 (JCPDS: 25-1003).Soaking In 2 S 3 /NA with NaOH removed the VO 2 to obtain In 2 S 3 /NS, which retained the pyramidal morphology while some flocculent particulates remained on the surface (Figures 1E and S4).Atomic force microscopy and transmission electron microscopy (TEM) demonstrated that the In 2 S 3 /NS surface morphology changed significantly compared with that of In 2 S 3 (Figures S5 and S6) and detected a layer of flocculent particulates that appeared on the In 2 S 3 surface after the heat treatment.In 2 S 3 and In 2 S 3 /NS exhibited similar XRD and Raman peaks (Figure S7), indicating the absence of other materials such as VO 2 or In 2 O 3 on the In 2 S 3 /NS surface.Additionally, high-resolution transmission electron microscopy (HRTEM) (Figure 1C) clearly revealed a lattice spacing of 0.380 nm that corresponded to the (116) plane of In 2 S 3 .In 2 S 3 /NS included the same (116) plane with a lattice spacing of 0.381 nm (Figure 1F), confirming that the bulk phase of In 2 S 3 /NS was still In 2 S 3 .Moreover, an amorphous layer was detected on the In 2 S 3 /NS surface that may expose more reactive sites, which is discussed later.Selected area electron diffraction patterns further confirmed that the crystallinity of In 2 S 3 /NS (Figure 1G) was weaker than that of In 2 S 3 (Figure 1D), 39 which was consistent with the HRTEM results.Figure S8 shows the TEM-energy dispersive spectroscopy (EDS) elemental maps for In 2 S 3 /NS, revealing uniform distributions of In, S, and O on In 2 S 3 /NS.The surface composition of In 2 S 3 /NS was further characterized by TEM-EDS line scan analysis.Figure 1H shows that the In to S ratio in In 2 S 3 /NS was close to 1:1 on the surface but remained normal at 2:3 in the bulk, demonstrating a change in the surface element amounts and the appearance of the NS group on the surface.This conclusion was verified through other EDS line scans (Figure S9).Moreover, measurements of the inductively coupled plasma (ICP) (Table S2) confirmed the presence of In and S in the NaOH solution after In 2 S 3 /NA was soaked in it, which was attributed to the loss of In and S during the solvothermal annealing process.As expected, the content of S was higher than that of In in the solution, confirming an increase in the ratio of In to S on the photoanode, which was consistent with the EDS line scan results.These data provide evidence that solvothermal annealing can be used to successfully form the NS group on the In 2 S 3 surface in situ.

Surface characterization and formation process of NS group
The chemical state of the synthesized surface NS group was further characterized by X-ray photoelectron spectroscopy (XPS).Compared with those of In 2 S 3 , the In 3d, S 2p, and O 1s peaks of In 2 S 3 /NS had lower binding energies (Figure 2A-C), indicating that the formation of the surface NS group increased the electron densities around the In, S, and O atoms. Figure 2C shows the XPS spectrum of O with a peak at 532.2 eV that correlated to the surface-absorbed H 2 O.The higher O content in In 2 S 3 /NS was attributed to the enhanced hydrophilicity of In 2 S 3 /NS, which can be verified from the results shown in Figure S10.In addition, we controlled different soaking times of NaOH (10 s, 5 min, 10 min, and 15 min, named In 2 S 3 /NA-10 s, In 2 S 3 /NA-5 min, In 2 S 3 /NA-10 min, and In 2 S 3 /NA-15 min, respectively) to discuss the effect of NaOH etching on the formation of NS groups.As the soaking time increases, the surface VO 2 is gradually etched away (Figure S11) and the NS groups on the surface gradually become visible.XPS can be used to determine the chemical composition and valence states of elements.The XPS spectrum reveals that as the etching time increases, the content of V gradually decreases and eventually disappears completely after 10 min (Figure S12).This is consistent with the results of SEM.Compared with In 2 S 3 /NA-5 min, the In 3d and S 2p peaks of In 2 S 3 /NA-10 min have higher binding energies (Figure S12), indicating successful etching removal of surface VO 2 .To further characterize the formed NS group, X-ray absorption spectroscopy (XAS) was used to survey the S K absorption edges. 23,40Figure 2D shows the S K-edge X-ray absorption near-edge structure (XANES) curves of In 2 S 3 and In 2 S 3 /NS.The XAS spectra at the S K-edge concern the electronic transitions from the S 1s to hybridized states of In and S 3p orbitals.The XAS experimental spectrum of In 2 S 3 is formed by two well-resolved peaks at 2471.5 and 2475.8 eV, and two rather broad and flat bands at energies higher than 2482 eV.The S K-edge XANES of In 2 S 3 /NS is different compared with In 2 S 3 , indicating that the chemical environment of S is changed.Meanwhile, the NS group was further qualitatively characterized by semi-quantitative XPS analysis.As shown in Figure 2E and Table S3, significant changes in the atomic contents of In and S demonstrated a loss in In and S during the heat treatment.Furthermore, the In to S ratio also changed from 2:3 to 1:1, illustrating the stoichiometric to nonstoichiometric change in the ratio of In to S on the In 2 S 3 surface, which was consistent with the previous results in this study.Interestingly, the XPS spectra displayed no presence of V (Figure 2F).Furthermore, the XPS spectra of a sample with a higher concentration of VO(acac) 2 , named In 2 S 3 /NS-HC, also showed no indication of the presence of V (Figure S13).To verify this conclusion, we performed SEM-EDS and ICP testing on In 2 S 3 /NS (Figure S14 and Table S4), which confirmed that V was absent in In 2 S 3 /NS.The formation of NS groups on the surface can be further explained by etching XPS (Figure S15).Compared to the pristine In 2 S 3 , the In and S peaks of In 2 S 3 /NS exhibit negative shifts, which is due to the formation of surface NS groups increasing the electron cloud density around In and S atoms.Nevertheless, no V signal was observed in depth-etched XPS, which could be attributed to the fact that V overcame lattice constraints and underwent oxidative leaching, forming oxides that were eliminated by NaOH etching.These results verified the formation of the NS groups on the surface with no V in an optimal photoanode.
To explore the role of V in the synthesis, we prepared In 2 S 3 coated with a CH 3 OH solution (CHO) and without the addition of VO(acac) 2 , named In 2 S 3 /CHO.In 2 S 3 /CHO exhibited a surface morphology similar to that of In 2 S 3 (Figure S16), making it clear that the NS group could not be formed without VO(acac) 2 as the V source.Furthermore, we performed the same experiments on FTO to form FTO/NS, which was used to explore the influence of In 2 S 3 on the formation of the NS group.FTO/NS maintained the morphology of FTO (Figure S17), demonstrating that the formation of the NS group was associated with In 2 S 3 and that the resulting phase composition was independent of the solution used.Additionally, we annealed In 2 S 3 at a high temperature in air to remove S and form In 2 O 3 , as demonstrated by XRD (JCPDS: 06-0416; Figure S18), and repeated the In 2 S 3 /NS preparation process (In 2 O 3 /NS).In 2 O 3 /NS also exhibited a surface morphology similar to that of In 2 O 3 (Figure S19), indicating the importance of S in the formation of the NS group.These results can suggest that the surface flocculent particulates originated from In 2 S 3 .Since VO 2 films provide better closed environments for the formation of NS groups, we used the same optimization conditions and performed high-temperature annealing in an N 2 atmosphere to form In 2 S 3 /A.Surface particulates were not formed in this case (Figure S20).Considering that during the preparation process, the photoanode needs to be soaked in NaOH.In order to verify whether it is only the etching effect of NaOH, we conducted a comparative experiment of NaOH direct immersion on the pristine In 2 S 3 (named In 2 S 3 /NaOH).As shown in Figure S21, In 2 S 3 /NaOH reveals the same pyramid morphology as In 2 S 3 and the surface has not changed, indicating that the formation of NS groups is not directly caused by NaOH etching.At the same time, In 2 S 3 /NaOH and In 2 S 3 display the same XRD characteristic peaks and similar PEC properties (Figure S21), indicating that the immersion of NaOH will not affect In 2 S 3 .This manifests that In 2 S 3 is inherently resistant to alkali, eliminating the influence of the presence of NaOH on the device.The above experimental results were combined to produce a schematic diagram of the formation of NS groups, as shown in Figure 2G.For the subsequent high-temperature annealing, we inferred that the V ion entered the In 2 S 3 surface and was impregnated, which reduced the stability of the In 2 S 3 surface during annealing and converted the excess V to VO 2 .Moreover, NaOH etching resulted in the removal of the surface VO 2 with a simultaneous loss of In and S due to the unstable surface In 2 S 3 , which caused the detection of In and S in the NaOH solution by ICP.Briefly, heat treatment produced a thick VO 2 film that completely covered the In 2 S 3 surface and provided a closed environment to avoid converting In 2 S 3 to In 2 O 3 .Simultaneous surface reconstruction occurred following the permeation of V ions and heat treatment, resulting in a loss of In and S and the formation of the NS group on the In 2 S 3 surface in situ.

Bulk separation dynamics of carrier transport
Figure S22 displays the photocurrent-voltage (J-V) curves of In 2 S 3 under different treatment conditions, including the annealing temperature (300 • C-450 • C), annealing time (1-6 h), and V source concentration (0.1-0.5 mol L −1 ).The optimum sample was obtained by annealing at 350 • C for 4 h with a V source concentration of 0.2 mol L −1 .The performance parameters of In 2 S 3 /NS were significantly better than those of pure In 2 S 3 (Figure 3A).J reached a value of 5.02 mA cm −2 at 1.23 V RHE and V on reduced from 0.61 to 0.30 V.The significantly enhanced J and obvious negative shift in the V on indicated that the NS group formed in situ optimized the transport and transfer paths of the carriers and improved their utilization efficiencies.In addition, we conducted a stability test under a hole scavenger (Figure S23A), which showed that the stability improved to a certain extent after solvent thermal annealing.The improvement in stability may be attributed to the modified NS group greatly promoting the surface OER dynamics and reducing the interfacial impedance.Moreover, the XRD patterns of the before/after tested samples were obtained (Figure S23B; the resulting sample was named In 2 S 3 /NSA).The XRD peaks remained basically unchanged, indicating that this process indeed improved the stability of the photoanode to a certain extent.To further explore the role of V, we conducted a series of comparative experiments with different elements (Zn and Al) by varying the composition of the solution (named In 2 S 3 /Zn and In 2 S 3 /Al).Through a similar heat treatment process, In 2 S 3 /Zn and In 2 S 3 /Al were obtained.However, In 2 S 3 /Zn displays obvious XRD peaks of ZnO, indicating that ZnO is not easily etched by NaOH.Owing to the existence of ZnO, the PEC performance of In 2 S 3 /Zn is poor (Figure S24).Slightly different from Zn, Al 2 O 3 can be removed easily by alkaline etching, while there is no surface reconstruction occurs on the surface of In 2 S 3 and generates NS group.Hence, In 2 S 3 /Al still exhibits basically unchanged performance compared with In 2 S 3 (Figure S25).Therefore, the V-induced NS groups are critical for enhanced PEC performance.The In 2 S 3 /NS photoanode exhibited the same light harvesting capability and similar band gap values (E g , 2.18 eV; Figure S26) as those of In 2 S 3 , indicating that these factors do not affect the performance.To examine the improvement of the photoanodic performance, the charge separation (η sep ) and injection (η inj ) efficiencies were also evaluated (Figure S27).The η sep for the as-prepared In 2 S 3 /NS was higher than that of In 2 S 3 (Figure 3B).The competitive carrier migration and recombination behaviors in bulk were assessed by analyzing the band-edge positions relative to the reaction redox potentials.Mott-Schottky (M-S) measurements were performed to determine the flat band potentials of the samples (Figure S28).The flat band potentials of In 2 S 3 and In 2 S 3 /NS were found to be 0.16 and −0.01 V RHE .The conduction band positions of In 2 S 3 and In 2 S 3 /NS were roughly estimated to be −0.04 and −0.21 V RHE , respectively.Combined with the E g value of In 2 S 3 and In 2 S 3 /NS determined from the Tauc plots, the valence band positions were determined to be 2.14 and 1.97 V RHE , respectively.The energy bands of pure In 2 S 3 and In 2 S 3 /NS were type II heterostructures (Figure 3C), demonstrated by the transfers of the electrons and holes to the FTO and electrolyte, respectively.Based on the surface potentials obtained by Kelvin probe force microscopy, the surface potential of pure In 2 S 3 was determined to be approximately 12 mV larger than that of In 2 S 3 /NS (Figure 3D), indicating that the work function of bulk In 2 S 3 was greater than that of In 2 S 3 /NS and the Fermi energy level of In 2 S 3 was lower than that of In 2 S 3 /NS.This was consistent with the earlier M-S results and facilitated the bulk phase transfer of electron-hole pairs.Additionally, atomic insights into the electron-hole transport mechanisms were obtained using density functional theory (DFT) calculations to illustrate the band alignments and free electron transfers of the heterojunctions in these models (Figure 3E,F).The calculated work functions of In 2 S 3 and NS were 5.88 and 5.43 eV (Figure 3G,H), respectively, indicating a more positive Fermi level of NS than that of In 2 S 3 .After reaching a state of Fermi equilibrium, the photogenerated carrier migration followed a type II charge transfer pathway.To verify this phenomenon, the opencircuit potential (OCP) was measured from the voltagetime (V-T) curves (Figure S29).The V-T curve for In 2 S 3 /NS produced a larger OCP value than that of In 2 S 3 , indicating an excellent carrier separation capability.These results showed that the formation of NS groups altered the transfer behavior and facilitated the separation of bulk carriers in the photoanode, thereby enhancing the bulk separation dynamics of photogenerated electron-hole pairs.

Carrier transport dynamics in in situ formed interface
The results suggested that the NS group formed on the In 2 S 3 surface in situ using solvothermal annealing, which was beneficial for the formation of high-quality interfaces, may reduce carrier recombination and effectively suppress the interfacial transfer resistance of carriers.Surface photovoltage spectroscopy (SPV) was used to examine the separation and transport behaviors of the photogenerated electron-hole pairs. 41As shown in Figure 4A, In 2 S 3 /NS displayed a higher SPV response intensity relative to pure In 2 S 3 , indicating improved carrier separation and transport.Room-temperature photoluminescence (PL) spectroscopy (Figure 4B) was conducted to investigate the bulk carrier dynamics.The PL intensity of In 2 S 3 /NS was lower than that of In 2 S 3 , suggesting a faster charge carrier separation.Furthermore, time-resolved PL decay spectra (Figure 4C) were obtained to compare the photogenerated carrier lifetimes.Compared with that of In 2 S 3 (τ = 0.84 ns), In 2 S 3 /NS (τ = 1.71 ns) demonstrated a longer carrier lifetime, which was indicative of suppressed charge recombination (Table S5).These results indicate that introducing an NS group does not introduce excessive interfacial defects while promoting the bulk phase separation of photogenerated electron-hole pairs.Moreover, electrochemical impedance spectroscopy (EIS) under different biases was employed to investigate the interfacial charge transfer and surface catalytic behavior (Figure S30). 42The series resistance (R s ) represented the resistance between the FTO and the sample and the charge transfer resistance included those of the bulk phase (R bulk ) and between the photoanode and electrolyte (R ct ).In 2 S 3 /NS exhibited the smallest R bulk compared with pure In 2 S 3 (Figure 4D), confirming the formation of a type II band structure, where the in situ introduced NS groups were beneficial for the carrier transfer in the bulk and interface and reduced the interface transmission resistance, improving the photoanode performance.

Surface carrier transfer and recombination kinetics
Simultaneously, the R ct values of In 2 S 3 /NS were lower than those of In 2 S 3 over the entire testing range (Figure 5A), indicating that the NS group as an effective OEC reduces the electrochemical impedance and promotes charge carrier transfers at the photoanode/electrolyte interface.The transfer behavior of the photogenerated carriers on the photoanode surface was mainly affected by the charge transfer rate constant (K ct ) and electron-hole recombination rate constant (K rec ).The In 2 S 3 /NS photoanode exhibited an improved K ct (Figure 5B), suggesting that the presence of the NS group promotes the surface transfer of photogenerated holes.The unexpected increase in K rec (Figure S31) could be attributed to the regulation and optimization behaviors of In and S during NS formation, which led to the introduction of some surface recombination sites.The charge transfer efficiency (η trans ) of In 2 S 3 /NS displayed a vast improvement compared with that of pure In 2 S 3 (Figure 5C), demonstrating the characterization of the NS group that promotes carrier transfer.Moreover, the hole lifetimes (τ h ) at the photoanode/electrolyte interface were also used to characterize the interfacial properties.The τ h of In 2 S 3 /NS was significantly lower than that of In 2 S 3 (Figure 5D), confirming that the NS group accelerates the transport of holes to the photoanode surface to participate in water oxidation.These results manifest that the catalytic effect of the NS group is mainly seen in the fast hole transfer and increased OER dynamics.Double-layer capacitance (C dl ) measurements were performed to determine the electrochemical surface areas, which confirmed the increase in active sites induced by the NS group.Cyclic voltammograms were obtained in a non-Faradaic region at different scan rates (50-150 mV s −1 ).The C dl for In 2 S 3 /NS was larger than that for In 2 S 3 (Figure S32).A more positive slope was an indicator of more active reaction sites.In 2 S 3 /NS contained more surface active sites and a larger reaction area than those of In 2 S 3 due to the amorphous properties of the NS group.Meanwhile, the enhanced η inj (Figure S33) in In 2 S 3 /NS confirmed the role of the NS group in boosting surface OER dynamics.The electrochemical OER properties under dark conditions were studied, as shown in Figure S34.The water oxidation current of In 2 S 3 /NS was higher than that of pristine In 2 S 3 under dark reaction conditions.Tafel slopes (Figure S34) were calculated using the dark J-V, the smaller Tafel slope of In 2 S 3 /NS indicated a lower η, further establishing the excellent OER catalytic activity of the NS group.Furthermore, the hole lifetime (τ H ) can be expected from the OCP decay.In 2 S 3 /NS exhibited a carrier lifetime of 0.10 s at the transient when the illumination was switched off (Figure S35), which is less than that of In 2 S 3 (τ H = 0.29 s).This suggests that the formed NS groups promote the transfer of photogenerated holes.These results demonstrate that the NS group is an excellent OEC that promotes photogenerated hole transfer, reduces surface η values, and boosts OER kinetics.To rationalize the improved catalytic performance of the In 2 S 3 /NS photoanode, we obtained the η values for the OERs within the In 2 S 3 and In 2 S 3 /NS systems based on the free energy changes (Table S6) calculated for both structures using the DFT (Figure S36).The computational details are described in the Supporting Information.The value of η decreased from 0.91 V for pristine In 2 S 3 to 0.49 V for In 2 S 3 /NS (Figure 5E,F).The 0.42 V drop in η reliably proves that the formation of the NS groups enhances the catalytic performance of the In 2 S 3 /NS photoanode.Introducing surface NS into the In 2 S 3 system alters the rate-determining step of * to *OH (step 1) in the In 2 S 3 system (Figure 5E) to *O to *OOH (step 3) in the In 2 S 3 /NS (Figure 5F) system, certifying that *OOH formation is preferred for the NS groups and favors O 2 production.These data demonstrate that the surface NS groups act as efficient OECs, have higher active surface areas and lower η, reduce the electrochemical reaction barriers of the photoanode, inhibit surface recombination, and promote OER kinetics.

CONCLUSION
In summary, solvothermal annealing was used to reconstitute In and S and form NS groups on the In 2 S 3 surface in situ, which exhibited an excellent PEC performance with a high J of 5.02 mA cm −2 at 1.23 V RHE and the V on that shifted negatively by 310 mV.The enhancement mechanisms were attributed to multiple factors (Figure 5G).First, the NS groups were formed in situ on the In 2 S 3 surface, resulting in a similar composition and close connection with bulk In 2 S 3 , high-quality photoanode/OEC interface formation, and reduced the interface recombination.Second, the unexpected formation of a type II heterojunction facilitated charge separation and transport and decreased bulk recombination.Finally, the surface acquired the amorphous properties of the NS groups and as an effective OEC exposed more reaction sites, increased the active surface area, lowered the surface η, and boosted the surface OER kinetics by altering the ratedetermining step of * to *OH (step 1) in the In 2 S 3 system to that of *O to *OOH (step 3) in the In 2 S 3 /NS system, thereby promoting the transfer capability of photogenerated holes.Our work provides a new method of surface modifications to promote bulk separation and surface injection by adjusting and optimizing the In to S surface ratio.In addition, it offers new ideas for generating efficient carrier transport and preparing high-efficiency OECs in situ.

EXPERIMENTS
The detailed experimental materials and methods can be found in the Supporting Information section.

C O N F L I C T O F I N T E R E S T S TAT E M E N T
The authors declare no conflict of interest.

F
I G U R E 1 (A) Schematic diagram of the photoanode synthesis.Top-view scanning electron microscopy (SEM) images of (B) In 2 S 3 and (E) In 2 S 3 /NS.High-resolution transmission electron microscopy (HRTEM) images of (C) In 2 S 3 and (F) In 2 S 3 /NS.Selected area electron diffraction (SAED) of the (D) In 2 S 3 and (G) In 2 S 3 /NS surfaces.(H) Transmission electron microscopy-energy dispersive spectroscopy (TEM-EDS) line scans of In 2 S 3 /NS.NS, nonstoichiometric In-S.

F
I G U R E 2 (A) In 3d, (B) S 2p, and (C) O 1s X-ray photoelectron spectroscopy (XPS) for In 2 S 3 and In 2 S 3 /NS.(D) The normalized S K-edge X-ray absorption near-edge structure (XANES) spectra.(E) Changes in the In and S elements based on XPS data.(F) V 2p XPS for In 2 S 3 and In 2 S 3 /NS.(G) Schematic diagram of the formation mechanism of the nonstoichiometric In-S (NS) groups.

F
I G U R E 3 (A) Linear sweep voltammogram curves.(B) η sep curves of In 2 S 3 and In 2 S 3 /NS.(C) Estimated band structures of In 2 S 3 and In 2 S 3 /NS.(D) Atomic force microscopy (AFM) images and the corresponding Kelvin probe force microscope (KPFM) surface potential maps of In 2 S 3 and In 2 S 3 /NS.Simulated optimized structures of (E) In 2 S 3 and (F) In 2 S 3 /NS.Electrostatic potentials of (G) In 2 S 3 and (H) In 2 S 3 /NS.NS, nonstoichiometric In-S.

F
I G U R E 4 (A) Surface photovoltage (SPV), (B) photoluminescence (PL), and (C) time-resolved PL (TRPL) spectra of the In 2 S 3 and In 2 S 3 /NS photoanodes.(D) R bulk of In 2 S 3 and In 2 S 3 /NS obtained from EIS at the different potentials.NS, nonstoichiometric In-S.

F
I G U R E 5 (A) R ct , (B) K ct , (C) η trans , and (D) τ h at the photoanode/electrolyte interface of In 2 S 3 and In 2 S 3 /NS, obtained by electrochemical impedance spectroscopy (EIS) at different potentials.Free energies of the oxygen evolution reaction (OER) steps for (E) In 2 S 3 and (F) In 2 S 3 /NS systems.(G) Schematic illustration of the carrier transport dynamics on the In 2 S 3 and In 2 S 3 /NS photoanodes.NS, nonstoichiometric In-S.
We acknowledge the support from the National Key Research and Development Program of China (No. 2021YFA1500800), the National Natural Science Foundation of China (52025028 and 52202272), and the Priority Academic Program Development of Jiangsu Higher Education Institutions.