Liquid Phase Exfoliation of Few‐Layer Non‐Van der Waals Chromium Sulfide

Abstract

Exfoliation of 2D non‐Van der Waals (non‐vdW) semiconductor nanoplates (NPs) from inorganic analogs presents many challenges ahead for further exploring of their advanced applications on account of the strong bonding energies. In this study, the exfoliation of ultrathin 2D non‐vdW chromium sulfide (2D Cr₂S₃) by means of a combined facile liquid‐phase exfoliation (LPE) method is successfully demonstrated. The morphology and structure of the 2D Cr₂S₃ material are systematically examined. Magnetic studies show an obvious temperature‐dependent uncompensated antiferromagnetic behavior of 2D Cr₂S₃. The material is further loaded on TiO₂ nanorod arrays to form an S‐scheme heterojunction. Experimental measurements and density functional theory (DFT) calculations confirm that the formed TiO₂@Cr₂S₃ S‐scheme heterojunction facilitates the separation and transmission of photo‐induced electron/hole pairs, resulting in a significantly enhanced photocatalytic activity in the visible region.

2D Cr₂S₃ prepared by liquid phase exfoliation forms a heterojunction with titanium dioxide for photoelectrochemical water splitting. 2D Cr₂S₃ is obtained from bulk Cr₂S₃ by liquid phase stripping in NMP. The results show that 2D Cr₂S₃ can form an S‐scheme heterojunction with TiO₂ giving a PEC water decomposition efficiency of the heterojunction of H₂: 124.3 and O₂: 62.2 µmol cm⁻² h⁻¹.

1. Introduction

Since the discovery of graphene in 2004, 2D layered van der Waals (vdW) nanomaterials have been widely utilized in various technological fields owing to their potentially valuable physicochemical properties.^{[ 1 ]} It is well known that most of the 2D nanomaterials consist of stacked layers and can be prepared through the “bottom‐up” or “top‐down” methods.^{[ 2 ]} Liquid‐phase exfoliation (LPE) is the most adopted “top‐down” method to prepare 2D nanoplates with high quality from layered crystals, producing many star 2D nanoplates in a solution‐processable form, the applications of which have been extensively investigated.^{[ 3 ]} With the rapid development of 2D nanoplates, researchers have attempted to use the “top‐down” method to exfoliate 2D nanoplates from non‐vdW inorganic analogs, and lots of exciting results have already been achieved in recent years.^{[ 4 ]} For example, in 2017, Guan and co‐workers reported exfoliation of atomically thin α‐WO₃ NPs through sonication of bulk WO₃ in bovine serum albumin (BSA) solution with a pH of 4.^{[ 5 ]} Not long after that, ultrathin 2D non‐vdW α‐Fe₂O₃ NPs were successfully obtained using the LPE method in DMF solution.^{[ 6 ]} By now, ongoing efforts are focused mostly on exploring 2D non‐vdW nanomaterials through LPE, given their exceptional properties and potential applications.

Chromium sulfide (Cr₂S₃), a representative of non‐vdW transition metal dichalcogenides (TMDs), exhibits interesting optical and air‐stable ferrimagnetic properties, revealing great potential for advanced optoelectronic applications.^{[ 7 ]} Nevertheless, current research on the preparation of 2D Cr₂S₃ mainly relies on the “bottom‐up” method, specifically chemical vapor deposition (CVD),^{[ 8 ]} due to the intra‐ and inter‐layer covalent bonding in Cr₂S₃. This method typically requires strict experimental conditions such as high temperature and pressure.^{[ 7} , ^{9 ]} To the best of our knowledge, there have been no reports on the exfoliation of 2D Cr₂S₃ from non‐van der Waals solids using a “top‐down” approach. Motivated by this circumstance, we first performed density function theory (DFT) calculations to compare the cleavage energies of different nanomaterials with 2D Cr₂S₃, including the vdW and the mostly experimentally realized non‐vdW materials, as shown in Scheme S1 (Supporting Information). Noteworthy, the exfoliation energy of Cr₂S₃ is much higher as compared to graphene and phosphorene, which can be attributed to the presence of only vdW bonds between layers in the latter materials, allowing for easier exfoliation accomplished even by micromechanical cleavage.^{[ 10 ]} However, the exfoliation energy of 2D Cr₂S₃ is much lower compared to recently exfoliated non‐vdW nanomaterials such as α‐Fe₂O₃ and FeS₂,^{[ 4} , ^{6 ]} revealing the possibility of realizing 2D Cr₂S₃ not only by “bottom‐up” but also by “top‐down” methods.

Herein, we realize a successful exfoliation of a few‐layer non‐vdW 2D Cr₂S₃ in N‐Methylpyrrolidone (NMP) solution by means of the LPE method. We examined the morphology and structure of the 2D Cr₂S₃ using atomic force microscopy (AFM) and transmission electron microscopy (TEM). The experimental results revealed that the obtained 2D Cr₂S₃ exhibits typical n‐type conduction behavior with a band gap of 1.45 eV. In addition, we suggest that this 2D Cr₂S₃ material exhibits uncompensated antiferromagnetic properties. Furthermore, we loaded the 2D Cr₂S₃ onto TiO₂ nanorod arrays to form an S‐scheme heterojunction. The experimental and DFT results indicated that the TiO₂@Cr₂S₃ heterojunction significantly enhances the separation of photo‐induced electron/hole pairs, leading to enhanced photocatalytic activity in the visible region. Specifically, the TiO₂@Cr₂S₃(12 h) heterojunction exhibits a remarkable photocurrent density of 6.8 mA cm⁻² at 1.23 V versus reversible hydrogen electrode (RHE) under AM 1.5G illumination along with the highest applied bias photon‐to‐current efficiency (ABPE) of 2.1% and incident photon to current efficiency (IPCE) of 67.1%. Overall, our study not only provides fundamental insight into 2D nanomaterials but also offers guidance for future exploration of novel and useful non‐vdW 2D inorganic analogs.

2. Results and Discussion

2.1. Exfoliation and the Morphology of 2D Cr₂S₃

The 2D Cr₂S₃ was prepared by exfoliating Cr₂S₃ crystals using the LPE method in an organic solvent NMP, as shown in Figure 1a. The detailed preparation process is presented in the Methods section. In the LPE process, the high‐frequency sound generates bubbles whose collapsing introduces mechanical energy that disrupts the bonds holding the layers together in bulk Cr₂S₃.^{[ 4} , ^{11 ]} As Cr₂S₃ possesses structural anisotropy and the number of interstitial bonds is lower compared to in‐layer bonds, the intralayer bonds can be broken more easily leading to exfoliation. As for the fate of the interstitial Cr atoms post‐exfoliation, they likely remain within the separated layers or are hydrolyzed.^{[ 12 ]} The exfoliation process primarily involves mechanical separation of the layers rather than chemical alteration of the atomic composition. Typically, nanosheets produced via exfoliation methods can range from a few nanometers to several micrometers in lateral size.

Figure 1.

Exfoliation process and 2D morphology of 2D Cr₂S₃. a) Scheme illustration of the exfoliation of bulk Cr₂S₃ crystals to 2D Cr₂S₃ in NMP solution. b) Low‐magnification TEM image of the exfoliated 2D Cr₂S₃. c,d) HR‐TEM and the SAED pattern of the exfoliated 2D Cr₂S₃ along the [001] zone axis. e, f and, g) AFM images and the thickness distribution of the exfoliated 2D Cr₂S₃, each bar represents a range rather than an exact value, e.g., 1.8 represents the range of 1.7–1.9 nm.

The TEM image in Figure 1b, taken at low magnification, confirms the successful exfoliation of the 2D Cr₂S₃. Figure 1c shows a typical high‐resolution TEM (HR‐TEM) image of the exfoliated 2D Cr₂S₃ along the [001] zone axis. Noteworthy, the atoms are continuously arranged in a hexagonal order without any obvious defects, indicating a good crystallinity of the exfoliated Cr₂S₃. The lattice spacing, measured to be 2.97 Å, corresponds to the (110) plane, which is consistent with the previously reported value for Cr₂S₃ synthesized by CVD.^{[ 13 ]} In addition, the corresponding selected area electron diffraction (SAED) pattern (Figure 1d) further demonstrates the single‐crystal nature of the exfoliated 2D Cr₂S₃ with a perfect six‐fold in‐plane symmetry. Figure S1 (Supporting Information) shows a high‐angle annular dark field scanning TEM (HAADF‐STEM) and elements mapping of the 2D Cr₂S₃, the results show that Cr and S are evenly distributed and the atomic ratio is ≈2:3. AFM was further carried out to study the thickness of the exfoliated 2D Cr₂S₃ as revealed in Figure 1e,f. An analysis of the AFM thickness distribution for 100 units of 2D Cr₂S₃ is depicted in Figure 1g. Statistical analysis reveals that the average thickness of the exfoliated 2D Cr₂S₃ is ≈3.4 nm, which is equivalent to two unit cells of Cr₂S₃.

2.2. Characterization of 2D Cr₂S₃

The composition, structure, and optical absorption of Cr₂S₃ were investigated before and after the exfoliation process. Figure S2 (Supporting Information) shows digital images of bulk and 2D Cr₂S₃ with equal quality in NMP. Apparently, the bulk Cr₂S₃ shows ash black color, while 2D Cr₂S₃ has a yellowish‐brown color, which further indicates a reduced optical absorption of Cr₂S₃ after exfoliation. Furthermore, the 2D Cr₂S₃ exhibited better dispersibility in NMP after one month, suggesting good stability and dispersity. X‐ray diffraction (XRD) was carried out to identify the crystal structure of Cr₂S₃ before and after exfoliation (Figure 2a). Four characteristic diffraction peaks, namely (003), (110), (113), (116), and (300) were observed, confirming the presence of a rhombohedral phase for Cr₂S₃.^{[ 7} , ¹³ , ^{14 ]} Figure 2b shows the Raman spectra of bulk and 2D Cr₂S₃. The peak at 251.2 cm⁻¹ corresponds to the in‐plane E _g mode, while the peaks at 174.4, 285.8, and 363.5 cm⁻¹ can be assigned to the out‐of‐plane A _g modes of Cr₂S₃.^{[ 9a ]} Interestingly, all four primary Raman peaks exhibit a slight “blue‐shift” after the exfoliation process. This observation is consistent with previous studies on CVD‐prepared Cr₂S₃, where a shift in peak position was observed with decreasing thickness of Cr₂S₃.^{[ 7c ]} X‐ray photoemission spectroscopy (XPS) measurements were made in order to study the chemical state and surface composition of Cr₂S₃. Figure S3 (Supporting Information) shows survey spectra of the binding energies ranging from 0 to 1200 eV. The peaks corresponding to C, N, and O were detected in both bulk and 2D Cr₂S₃, which can be attributed to surface contamination when the samples were exposed to air.^{[ 12} , ^{15 ]} In the detailed fitted results (Figure 2c), the peaks with the binding energies located around 575.0 and 584.0 eV were assigned to Cr 2p_1/2 and Cr 2p_3/2, respectively,^{[ 9b ]} while the peaks around 577.0 and 587.0 eV were attributed to Cr─O bonds.^{[ 16 ]} It is not surprising that the apparent higher peak intensities of oxidized states increase after exfoliation as the 2D Cr₂S₃ is exposed to the atmosphere during the LPE process.^{[ 17 ]} The exfoliated thinner 2D Cr₂S₃ material has a larger specific surface area, allowing for more oxygen‐containing species to be adsorbed on their surface, leading to an increased number of oxidized species. Similar results were observed in the high‐resolution spectrum of S 2p (Figure 2d), where the peaks around 161.0 and 162.0 eV were assigned to S 2p_1/2 and S 2p_3/2, respectively,^{[ 13a ]} and where a slightly higher peak intensity of oxidized species with bonding energies around 168.5 eV (S─O) was detected in 2D Cr₂S₃.^{[ 18 ]}

Figure 2.

Composition, and structure of bulk Cr₂S₃ crystals and 2D Cr₂S₃. a) XRD pattern. b) Raman spectra. c,d) XPS spectra of bulk and 2D Cr₂S₃.

The UV–Vis absorption spectrum was used to analyze the change in optical absorption of Cr₂S₃ after the exfoliation process (Figure S4, Supporting Information). Apparently, the intrinsic optical absorption of the 2D Cr₂S₃ decreased significantly compared to bulk Cr₂S₃. It is well known that the band gap of 2D nanomaterials will increase as the number of layers decreases, leading to a diminished optical absorption in few‐layer 2D nanomaterials.^{[ 19 ]} The bandgap energy (E _g) of bulk and 2D Cr₂S₃ was determined using the Kubelka‐Munk equation: (αhν)ⁿ = A(hν‐E_g),^{[ 20 ]} where α, h, ν, and A refer to the absorbance, Planck constant, light frequency, and constant, respectively. The band gap of bulk Cr₂S₃ was calculated to be 0.98 eV, which is in excellent agreement with our theoretical results calculated using the HSE06 functional (1.01 eV, Scheme S1, Supporting Information), while the band gap of 2D Cr₂S₃ was fitted to 1.45 eV. To better evaluate the conduction band (CB) and valence band (VB) of 2D Cr₂S₃, the VB spectrum acquired from XPS was performed (Figure S5, Supporting Information) giving the result that the VBM position of 2D Cr₂S₃ is 0.55 eV. According to Equation S1 (Supporting Information), the VB is calculated to be 0.31 eV versus Normal Hydrogen Electrode (NHE). Applying the formula E_CB = E_VB – E_g , we determined the CB of 2D Cr₂S₃ to be −1.14 eV versus NHE_. The Mott‐Schottky (MS) plots of 2D Cr₂S₃ were performed in 0.5 m Na₂SO₄ solution at different frequencies. As shown in Figure S6 (Supporting Information), the flat‐band potential (E_FB ) of 2D Cr₂S₃ was found to be −1.04 eV, which was obtained by the interception of the X‐axis that was extrapolated from the linear part of the MS plots. Normally, the E_FB is close to their CB edge, while for n‐type semiconductors the specific CB value is usually positively shifted about 0.1–0.3 eV.^{[ 21 ]} Therefore, the CB value of 2D Cr₂S₃ can be roughly estimated to be in the range from −1.14 to −1.34 eV versus NHE, which is consistent with the CB value obtained from the UV–Vis and the XPS VB spectra.

2.3. Magnetic Properties of 2D Cr₂S₃

The magnetic properties of exfoliated 2D Cr₂S₃ were investigated using a superconducting quantum interference device magnetometer (SQUID‐MPMS). The field cooling (FC) and zero field cooling (ZFC) processes were conducted under an external field of 1000 Oe (Figure 3 ; Figure S7, Supporting Information). In many reports, the magnetic state of Cr₂S₃ is referred to as ferrimagnetic. Ferrimagnetic materials are characterized by the presence of two magnetic sublattices with differing magnetic moments, which are antiparallelly aligned but unequal in magnitude, resulting in a net magnetic moment. Typically, in such materials, the FC (Field‐Cooling) curve exhibits a higher magnetization compared to the ZFC (Zero Field‐Cooling) curve due to the alignment of magnetic moments with the external field during the cooling process. However, upon closer examination, it becomes evident that the magnitude of magnetic moments in Cr₂S₃ is low in absolute values, as depicted in Figure 3a. This contrasts with typical ferrimagnets like Fe₃O₄, where the FC curve starts from high values, indicating the preservation of magnetic moments in a field at low temperatures.^{[ 22 ]} Ferrimagnets often display a broader temperature range over which the FC curve shows higher magnetization compared to the ZFC curve due to the persistence of magnetic alignment with the external field. In the case of 2D Cr₂S₃, the difference between the FC and ZFC curves is more pronounced at lower temperatures, where magnetic moments are more likely to align, as illustrated in Figure 3. Additionally, the transition between the FC and ZFC curves appears sharper, indicative of anomalies possibly due to the presence of uncompensated magnetic moments, which is consistent with the behavior observed in Figure 3. Based on these observations, we conclude that the synthesized 2D Cr₂S₃ behaves more like an uncompensated antiferromagnet.

Figure 3.

Magnetic performance of the exfoliated 2D Cr₂S₃. a) Temperature‐dependent magnetic susceptibility χ of 2D Cr₂S₃ studied on sapphire substrate with a magnetic field of 0.1 T. FC (Red) and ZFC (blue) refer to the field cooling and zero‐field cooling processes. b) Magnetic hysteresis loops of 2D Cr₂S₃ at 10 and 300 K with magnetic field scanning within ± 6 T. c) The enlarged magnetic hysteresis loops of 2D Cr₂S₃ acquired form (b).

2.4. Photocatalytic Activity of TiO₂@Cr₂S₃

Photoelectrochemical (PEC) water splitting has been considered a promising way to convert solar energy into clean hydrogen and oxygen, contributing to sustainable energy supply and carbon neutrality. Titanium dioxide (TiO₂) is a widely studied photoanode in PEC water splitting owing to its favorable physicochemical properties and environmentally friendly fabrication process.^{[ 23 ]} However, the limitations of TiO₂, such as fast carrier recombination, wide band gap, and slow water oxidation kinetics, have hindered its further application for PEC water splitting. Herein, we successfully constructed an S‐scheme heterojunction composed of TiO₂@Cr₂S₃ by facile soaking 1D TiO₂ nanorod arrays (NRs) in a 2D Cr₂S₃ solution, which is further demonstrated to be an effective way to enhance the photoelectrolysis activity of TiO₂.

The details for the fabrication process of TiO₂@Cr₂S₃(6, 12, 18, and 24 h) are presented in the SI. The diffraction peaks in the XRD pattern in Figure S8 (Supporting Information) reveal a rutile type of the fabricated 1D TiO₂ NRs.^{[ 24 ]} The Raman spectrum of TiO₂@Cr₂S₃(12 h), as shown in Figure S9 (Supporting Information), also exhibits peaks characteristic of rutile type TiO₂. Noteworthy, there are no obvious peaks of Cr₂S₃ that can be detected in either the XRD or Raman spectra owing to the small additive amount of 2D Cr₂S₃. Figure S10 (Supporting Information) shows SEM images of the pristine TiO₂ NRs and TiO₂@Cr₂S₃(12 h) heterojunction on FTO glasses from the front and side view. Notably, aggregated 2D Cr₂S₃ can be observed on the surface of the TiO₂ NRs, and the color of the TiO₂ film changes from greyish to dark brown after the soaking process. Figure S11 (Supporting Information) presents an elemental mapping of TiO₂@Cr₂S₃(12 h), further confirming the presence of Cr₂S₃ on the TiO₂ NRs; the molar ratio of Cr₂S₃ to TiO₂ is ≈0.021:1. Figure S12 (Supporting Information) displays the valence states after the formation of TiO₂@Cr₂S₃, wherein the Cr 2p spectrum of TiO₂@Cr₂S₃ remains largely unchanged compared to the previous 2D Cr₂S₃. Similarly, there are no significant changes observed in the Ti 2p^{[ 25 ]} and O1s^{[ 26 ]} spectra, indicating the stability of the constituent elements in the formation of TiO₂@Cr₂S₃; there is no alteration in valence states or formation of chemical bonds involved.

Normally, the photoelectrolytic activity of photoanode photocurrent density is evaluated using photocurrent density. A standard three‐electrodes PEC system was used to measure the photocurrent densities of pristine TiO₂ and TiO₂@Cr₂S₃ in a 0.5 m KOH electrolyte under AM 1.5G illumination. Figure 4a shows that the photocurrent densities of pristine TiO₂ and 2D Cr₂S₃ are 3.6 and 1.0 × 10⁻⁴ mA cm⁻² at 1.23 V versus RHE under front illumination, respectively. However, the formation of TiO₂@Cr₂S₃ significantly enhances the photocurrent density ‐ TiO₂@Cr₂S₃(6, 12, 18 h and 24 h) reached 5.43, 6.8, 6.36 and 6.1 mA cm⁻² at 1.23 V versus RHE, respectively. This value of TiO₂@Cr₂S₃(12 h) is ≈1.89 and 68000 times higher than that of pristine TiO₂ and 2D Cr₂S₃, respectively. The enhanced photocurrent densities can also be observed at lower applied potentials, with the highest applied bias photon‐to‐current efficiency (ABPE) of TiO₂@Cr₂S₃(6, 12, 18 and 24 h) are 1.62%, 2.1%, 1.88% and 1.87% at 0.6 V versus RHE, respectively (Figure 4b). The overall water splitting experiment of TiO₂@Cr₂S₃ was carried out at 1.23 V versus RHE under AM 1.5G illumination for 2 h as shown in Figure 4c. The H₂:O₂ properties of TiO₂@Cr₂S₃(6, 12, 18 and 24 h)are 96.9:48.22, 124.3:62.2, 116:57.92 and 110.9:55.67 µmol cm⁻² h⁻¹. Clearly, TiO₂@Cr₂S₃(12 h) exhibits the best performance, as evidenced by the gas‐phase chromatogram shown in Figure 4d. Excessive loading of 2D Cr₂S₃ may hinder the light absorption of TiO₂, leading to the decrease in performance of TiO₂@Cr₂S₃(18 and 24 h). The total gas production of TiO₂@Cr₂S₃(12 h) was about 2.5‐fold enhanced, with the H₂ production measured at 50.6 µmol cm⁻² h⁻¹, and O₂ production detected at 25.3 µmol cm⁻² h⁻¹ for pristine TiO₂ (Figure S13 and Table S1, Supporting Information), consistent with the stoichiometric ratio of 2:1. Furthermore, the detected gas production curves align well with the calculated gas evolution, demonstrating a Faradaic efficiency of about 97.8%. The Movies S1, S2 (Supporting Information) showcase the gas production process of pristine TiO₂ NRs and TiO₂@Cr₂S₃(12 h) at 1.23 V versus RHE. Notably, there is only about a 10% decrease in photocurrent density after the continuous 2‐hour period of light illumination (Figure 4e).

Figure 4.

Photocatalytic activity. a) Photocurrent density versus applied potential curves of pristine TiO₂, Cr₂S₃, and TiO₂@Cr₂S₃. b) ABPE curves of pristine TiO₂ and TiO₂@Cr₂S₃. c) H₂ and O₂ evolution of different TiO₂@Cr₂S_3. d) The corresponding gas chromatogram of TiO₂@Cr₂S₃(12 h) water splitting products. e) Photocurrent density versus time of TiO₂@Cr₂S₃(12 h). f) IPCE curves of pristine TiO₂ and TiO₂@Cr₂S₃(12 h). g) The UV‐DRS spectra of pristine TiO₂ and TiO₂@Cr₂S₃. h) Calculated photocurrent density curves of TiO₂ NRs and TiO₂@Cr₂S₃(12 h) by integrating IPCE curves with the standard solar spectrum. i) The PL spectra of TiO₂ NRs and TiO₂@Cr₂S₃(12 h).

The incident photon to current efficiency (IPCE) curves were analyzed to investigate the photoelectric conversion ability of TiO₂@Cr₂S₃(12 h) (Figure 4f). The pristine TiO₂ exhibited a peak of ≈29.8% at 330 nm. However, the IPCE value of TiO₂@Cr₂S₃(12 h) was significantly enhanced, reaching a maximum value of 67.1% at 360 nm. This value is about 4.4‐fold higher than the IPCE value (15.8%) of pristine TiO₂ at 360 nm. These results indicate that the loading of 2D Cr₂S₃ on TiO₂ effectively accelerates the separation of carriers.

The PEC properties of pristine TiO₂ and TiO₂@Cr₂S₃ were further studied to investigate the underlying mechanism for the significantly enhanced photoelectrolysis activity. The UV–Vis diffuse reflectance spectrum (UV‐DRS) of pristine TiO₂ and TiO₂@Cr₂S₃ (Figure 4g) were analyzed in order to observe the differences after the introduction of 2D Cr₂S₃ (Figure S14, Supporting Information). Interestingly, the TiO₂@Cr₂S₃ photoanode exhibited an enlarged excited region compared to the pristine TiO₂ photoanode. Previous studies have reported that enhanced light absorption can generate more photo‐induced electron/hole pairs, which contributes to the improvement of the PEC conversion. To better understand the effect of the enlarged light absorption on the PEC performance, the light absorptance versus wavelength was calculated. The obtained photocurrent densities (J _c) of pristine TiO₂ and TiO₂@Cr₂S₃(12 h) amounted to 3.7 and 7.5 mA cm⁻² according to Equation S2 (Supporting Information), respectively, by integrating the standard solar spectrum with the curves in Figure 4 h. It closely matches the detected photocurrent values (Figure 4a). This suggests that the applied AM 1.5G solar light matches well with the standard solar spectrum in this work and further demonstrates that enhanced light absorption is beneficial for enhancing photocurrent densities. Moreover, the calculated charge separation (η _sep) and surface charge transfer (η _trans) efficiencies confirmed the significant improvement in η _sep and η _trans after the loading of 2D Cr₂S₃ on the TiO₂ array (Figure S15 and Table S2, Supporting Information).

Electrochemical impedance spectroscopy (EIS) and Mott–Schottky (MS) curves were taken to gain a deeper understanding of the effect of the formation of TiO₂@Cr₂S₃ heterojunction and the process of charge separation. From high to low frequency, the EIS curves can be divided into two parts, where the R _ct refers to the interfacial charge transfer resistance and R _s is the series resistance (Figure S16, Supporting Information), respectively.^{[ 27 ]} The values of each part for pristine TiO₂ and TiO₂@Cr₂S₃(12 h) photoanodes were fitted and listed in Table S3 (Supporting Information). As can be seen, there are no significant differences in the values of R _s between the photoanodes, indicating that the series resistance has a negligible influence. However, a noticeable decrease in R _ct is observed after the formation of the TiO₂@Cr₂S₃ heterojunction, suggesting that the introduction of 2D Cr₂S₃ can greatly accelerate the charge transfer process, leading to enhanced photocurrent density.

Photoluminescence (PL) spectroscopy was performed to investigate the carrier dynamics of TiO₂ and TiO₂@Cr₂S₃(12 h). The emission density of TiO₂@Cr₂S₃ was found to be significantly quenched compared to pristine TiO₂ (Figure 4i), suggesting that the introduction of 2D Cr₂S₃ is beneficial for suppressing the recombination dynamics of photo‐induced charge carriers.^{[ 28 ]} Furthermore, time‐resolved transient photoluminescence decay (TRPL) spectroscopy was performed to gain insight into the carrier behavior for both TiO₂ and TiO₂@Cr₂S₃(12 h) photoanodes. As revealed in Figure S17 and Table S4 (Supporting Information), the TiO₂@Cr₂S₃(12 h) shows a shorter average lifetime of 0.64 ns, consistent with previous studies on heterojunctions. This confirms that the loading of 2D Cr₂S₃ reduces the average lifetime and enhances the charge transfer process.^{[ 29 ]} Based on the results above, the synergy of the 2D Cr₂S₃ materials significantly improves the photoelectrolysis activity of TiO₂ arrays.

The positive slopes of the MS curves indicate the formation of an n–n type heterojunction between TiO₂ and TiO₂@Cr₂S₃(12 h). The flat band potentials of the samples can be determined by finding the point where the linear part of the curves intersects the x‐axis (Figure S18, Supporting Information) and were found to be −0.28 and −0.61 V for TiO₂ and TiO₂@Cr₂S₃, respectively. The negatively shifted value can be attributed to the reduction of the Fermi surface level pinning effect, as demonstrated by the open circuit potentials (OCP) results (Figure S19, Supporting Information).^{[ 30 ]} Considering that the band gap of TiO₂ is 3.07 eV, it can be inferred from E _g = E_VB − E_CB that the conduction band edge (E_VB ) of the prepared TiO₂ is 2.69 eV versus NHE. Therefore, the band alignment between TiO₂ and Cr₂S₃ confirms a type‐II heterojunction structure. Compared to the traditional type‐II charge transfer mechanism, the S‐scheme transfer offers unique advantages in photocatalysis. In this mechanism, electrons located partially on the conduction band with low reduction potential in both heterojunctions can recombine with holes on the valence band with low oxidation potential, thereby maintaining the high redox activity of the photogenerated charge carriers in both materials.^{[ 31 ]} Hence, it is crucial to investigate the charge transfer mechanism of TiO₂@Cr₂S₃.

Figure 5a,b shows the DMPO (5,5‐Dimethyl −1‐pyrroline N‐oxide) ‐·OH and DMPO‐·O₂ ⁻ signals of the three materials in water and methanol solutions, where the oxidation‐reduction potentials of ·OH and ·O₂ ⁻ are −1.99 and −0.33 V versus NHE, respectively.^{[ 32 ]} Due to the lower E_VB (0.31 V vs NHE) of the 2D Cr₂S₃ than the oxidation potential of ·OH, no DMPO‐·OH signal could be observed. However, both TiO₂ NRs and TiO₂@Cr₂S₃ exhibited significant DMPO‐·OH signals, with TiO₂@Cr₂S₃ showing a stronger signal, indicating the generation of more ·OH. Similarly, only a small amount of ·O₂ ⁻ radicals were generated due to the lower E_VB (−0.38 V vs NHE) of TiO₂, while clear DMPO·O₂ ⁻ signals were observed for both 2D Cr₂S₃ and TiO₂@Cr₂S₃. If TiO₂@Cr₂S₃ followed the traditional type‐II charge transfer. The lower reduction potential of TiO₂ and the lower oxidation potential of Cr₂S₃ would not match the trend of stronger DMPO‐·OH and DMPO‐·O₂ ⁻ signals.^{[ 33 ]} Therefore, we believe that an S‐scheme charge transfer indeed occurs inside TiO₂@Cr₂S₃, where electrons on the TiO₂ CB recombine with holes on the Cr₂S₃ VB, so retaining holes on the TiO₂ VB and electrons on the Cr₂S₃ CB and realizing an effective separation of the photo‐generated carriers.^{[ 31 ]} Hence, TiO₂@Cr₂S₃ gains stronger redox capability. Further investigation of the interface charge transfer between TiO₂ and Cr₂S₃ was conducted using density functional theory (DFT) calculations. Figure 5c,d show the electrostatic potential of the (001) surface of Cr₂S₃ and the (101) surface of TiO₂, respectively. We employed hybrid long‐range corrected HSE06 functional to calculate the work functions more precisely.^{[ 34 ]} The work function of the (001) surface of Cr₂S₃ is calculated to be 6.86 eV, while the work function of the (101) surface of TiO₂ is calculated to be 7.45 eV. The difference in work functions indicates that, upon contact, the charge will flow from Cr₂S₃ (with a lower work function) to TiO₂ (with a higher work function) until the Fermi level reaches equilibrium. This charge transfer results in a negatively charged TiO₂ and a positively charged Cr₂S₃, creating an internal electric field from Cr₂S₃ to TiO₂. This electric field facilitates the formation of the S‐scheme charge transfer mechanism^{[ 33} , ^{35 ]} discussed above. Figure 5e shows the charge density distribution, with yellow representing charge accumulation and blue representing charge depletion. It is evident that Cr₂S₃ provides electrons at the non‐uniform contact interface and accumulates electrons on the surface of TiO₂. A Bader charge analysis confirms that TiO₂ gains 0.63 electrons from Cr₂S₃, supporting the possibility of spontaneous electron transfer from Cr₂S₃ to TiO₂ at the heterojunction interface of TiO₂@Cr₂S₃.

Figure 5.

Density functional theory calculation and charge transfer. a) DMPO‐·OH in aqueous, b) DMPO‐·O₂ ⁻ in methanol dispersion of 2D Cr₂S₃, TiO₂ NRs and TiO₂@Cr₂S₃. c and d) Electrostatic potentials of Cr₂S₃ (001) surface and TiO₂ (101) surface. Insets show the structural models of the materials for the DFT calculation. e) Image of simulated interface charge density difference between TiO₂ and Cr₂S₃. f) Schematic illustration of TiO₂@Cr₂S₃ heterojunction: internal electric field (IEF)‐induced charge transfer, separation, and the formation of an S‐scheme heterojunction.

Figure 5f illustrates the charge transfer mechanism of the S‐scheme heterojunction between TiO₂ and Cr₂S_3. The higher conduction band edge (−1.14 V vs NHE) of 2D Cr₂S₃ makes it an ideal reducing photocatalyst, while TiO₂ has a deeper valence band edge (2.82 V vs NHE), indicating its strong oxidation capability. When they come into contact, the difference in work functions leads to electron flow toward TiO₂ and the generation of an internal electric field. Additionally, the charge loss in Cr₂S₃ causes an upward band bending, while the electron accumulation in TiO₂ leads to a downward band bending, providing a prerequisite for the formation of an S‐scheme.^{[ 36 ]} Under light excitation, both TiO₂ and Cr₂S₃ generate photo‐generated electrons that migrate to their respective conduction bands. Driven by the built‐in electric field, electrons in the TiO₂ conduction band recombine with holes in the Cr₂S₃ valence band. This heterojunction charge transfer mechanism in the S‐scheme not only enhances the separation of photo‐generated carriers compared to type‐II transfer but also exhibits stronger redox capability. This is the reason behind the enhanced PEC water‐splitting ability of TiO₂@Cr₂S₃. In addition, surface electronegative species may form due to suspended bonds on the surface after stripping of non‐van der Waals materials. According to previous reports, when this surface‐charged negative material is in contact with another material a heterojunction is formed where the holes generated by light can be rapidly transferred to the surface of the material, thus generating an additional driving force to promote charge separation and hole transfer between the two interfaces.^{[ 37 ]} In the TiO₂@Cr₂S₃ system, surface electronegative species cause the photogenerated holes of 2D Cr₂S₃ to be attracted to the surface, and then recombine with the TiO₂ photogenerated electrons at the Cr₂S₃/TiO₂ interface. The charge separation efficiency and S‐scheme charge transfer of TiO₂@Cr₂S₃ are so further improved.

3. Conclusion

This study innovatively developed commercial‐phase chromium sulfide (Cr₂S₃) nanoplates prepared through liquid‐phase exfoliation. The research findings revealed that 2D Cr₂S₃ exhibits suppressed antiferromagnetic and enhanced ferromagnetic interactions compared to its bulk counterpart. Furthermore, the newly formed heterojunction between 2D Cr₂S₃ and rutile‐phase titanium dioxide nanorod arrays significantly enhanced the photocatalytic water splitting activity under simulated sunlight, reaching values as high as H₂: 124.3 and O₂: 62.2 µmol cm⁻² h⁻¹. Density functional theory and experimental data confirm that this improvement can be attributed to the formation of an S‐scheme heterojunction and an enhancement of the light absorption activity. These results demonstrate that liquid‐phase exfoliated 2D Cr₂S₃ with appropriate band gap and band edge positions make ground for an effective strategy for efficient utilization of solar energy in titanium dioxide nanorod arrays.

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supporting Information

Supplemental Movie 1

Supplemental Movie 2

Su W., Kuklin A., Jin L. hua, Engelgardt D., Zhang H., Ågren H., Zhang Y., Liquid Phase Exfoliation of Few‐Layer Non‐Van der Waals Chromium Sulfide. Adv. Sci. 2024, 11, 2402875. 10.1002/advs.202402875

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.