Red phosphorous (RP) is one of the most abundant, inexpensive, less toxic, and easily available materials^[1⇓-3]. It possesses strong visible light absorption, and has great potential for photocatalytic applications^[4]. However, the photogenerated electron and hole are easily recombined in photocatalytic efficiency, which greatly limits its application because of RP large particle size and agglomeration^[5]. RP is combined with other semiconductor material to construct heterojunction, which can effectively overcome the above disadvantages and improve the photocatalytic activity^[6], such as SnO₂/RP^[7], Bi₂O₃/RP^[8], Bi₂O₂CO₃/RP^[9], g-C₃N₄/RP^[10] and Bi₂Fe₄O₉/RP^[11]. Therefore, it is necessary to find suitable semiconductor matching RP to build heterojunction.

Tungsten trioxide (WO₃), as a typical n-type semiconductor with narrow bandgap energy (2.4-2.8 eV)^[12], was widely used in the management of environmental remediation^[13]. Although remarkable advances have been made, the narrow band gap of WO₃ also causes fast recombination rate of photo-induced charges, leading to poor photocatalytic activity^[14-15]. Compared with pure WO₃ and RP, WO₃/RP composite showed better hydrogen evolution, which could reduce the recombination of charge carriers^[16]. However, WO₃/RP composites via a low temperature calcination method led to aggregation and decrease of the active sites of RP.

In this study, the WO₃/HRP S-scheme heterojunction was constructed through hydrothermal treatment. The photocatalytic performance was estimated by photocatalytic organic degradation and hydrogen evolution. Finally, the S-scheme heterojunction photocatalytic mechanism of WO₃/HRP composites was directly evidenced by in situ X-ray photoelectron spectroscopy (XPS) and electron paramagnetic resonance (EPR) results.

Preparation of WO₃: Na₂WO₄·2H₂O (60.63 mmol·L^-1) was dissolved in 50 mL of ultrapure water, then 0.9 mL of lactic acid was dropped into the solution and stirred for 15 min. Subsequently, pH of solution was adjusted to 1.0 using 6.0 mol·L^-1 HCl. The solution was stirred for 30 min, and heated at 120 ℃ for 12 h, centrifuged. The obtained samples were washed with deionized water, dried at 60 ℃ for 5 h, and then calcined at 500 ℃ for 2 h.

Preparation of WO₃/HRP composites: HRP was treated according to Ref.[17]. The composite with HRP was obtained based on WO₃ mass fraction (3%, 5% and 7%), and the solution was hydrothermally treated at 150 ℃ for 4 h, centrifuged and washed for several times, and freeze-dried for 18 h to obtain WO₃/HRP composite and marked as 3%WO₃/HRP, 5%WO₃/HRP, and 7%WO₃/HRP, respectively.

A 300 W Xe lamp (UV light was cutoff-filtered, 140 mW/cm²) was used, and RhB was used as target pollutant. 5 mg photocatalyst was added into 20 mL RhB solution (10 mg·L^-1), and then stirred in dark for 30 min to ensure absorption-desorption equilibrium. During the irradiation process, the suspension was removed every 1 min for centrifugation and the absorbance of the residual RhB solution was determined by UV-Vis spectrophotometer at the maximum absorption wavelength (λ_max=554 nm).

Photocatalytic hydrogen production was carried out by using a 300 W Xe lamp (UV light was cutoff-filtered, 140 mW/cm²). The generated hydrogen was in situ detected periodically using an online gas chromatography (GC-7900) with a thermal conductivity detector. 50 mg photocatalyst was dispersed in 80 mL solution containing Na₂SO₃ (0.35 mol·L^-1) and Na₂S (0.15 mol·L^-1) as the sacrifice agent, and the suspension was magnetically stirred under visible light irradiation.

The crystal structures of the HRP, WO₃ and 5%WO₃/HRP composite were analyzed by XRD patterns (Fig. 1). For pure HRP, the characteristic diffraction peak at 2θ=14.9° corresponded to (102) crystal plane, consistent with the reported results in the literature^[17]. The characteristic peaks of WO₃ indexed at 2θ=22.8°, 24.1°, and 33.9° correspond to the planes (002), (020) and (200) facets (JCPDS 83-0951), according to the lattice parameters of monoclinic tungsten oxide^[18⇓-20]. For 5%WO₃/HRP composite, the diffraction peaks of the HRP and WO₃ were obviously present.

The surface chemical compositions of all samples were characterized by FT-IR spectroscopy (Fig. S1). For pure HRP, the vibration absorption of the P-P-O, P-O and P=O bond located at 1008, 1178 and 1641 cm^-1, which was consistent with the characteristic absorption peaks of RP^[21]. For pure WO₃, the absorption peaks at 956, 823 and 753 cm^-1 were referred to W=O, O-W-O and W-O-W bonds^[22]. The characteristic peaks of WO₃ and HRP functional groups were presented in the FT-IR spectrum of 5%WO₃/HRP, indicating the coexistence of WO₃ and HRP in the composite.

The morphologies of as-prepared samples were examined by SEM. HRP removed impurities from the surface of commercial RP, and its surface existed micropore (Fig. S2(a, d)). The typical morphology of WO₃ was rectangular nanoplates with a certain thickness (Fig. S2(b, e)). 5%WO₃/HRP composite revealed a large number of loose particles, which facilitates the improved light adsorption, and also provides abundant spaces for the anchoring of other particles (Fig. S2(c, f)). Furthermore, the basic elements of W, O, and P were present in the WO₃/HRP composite (Fig. S2(g, h). Thus, HRP and WO₃ coexisted in the WO₃/HRP composites.

The crystal structure and lattice orientation of each sample were further investigated through TEM. HRP surface was amorphous, and tiny folded, which was favorable to its surface adsorption properties (Fig. 2(a, b)). The WO₃ nanostructure exhibited regular sheet-like morphology structure, which was consistent with the SEM results (Fig. S2(c)). The lattice size was 0.37 nm that matched the distance (020) planes of monoclinic WO₃ structure, according to the lattice parameters of monoclinic WO₃ (JCPDS 83-0951)^[18]. TEM images of 5%WO₃/HRP showed that there was an additional layer of homogeneous WO₃ on the surface of HRP and closely connected to each other, avoiding aggregation of HRP and forming a heterogeneous structure that facilitates interfacial charge transfer (Fig. 2(e, f)).

XPS survey spectra were conducted to investigate the chemical compositions and bonding configurations of the samples^[23]. C, O, P and W were present in the composite (Fig. S3(a)). The high-resolution P2p spectrum of HRP (Fig. 3(a)) was deconvoluted into two peaks at 129.82 and 130.61 eV, corresponding to P2p_3/2 and P2p_1/2, respectively, and the binding energy peak at 134.19 eV corresponded to P-O bond, indicating the presence of phosphorus oxides on the surface of HRP. While in 5%WO₃/HRP composite, the P-O bond disappeared and the binding energies of the P2p_3/2 and P2p_1/2 were shifted towards higher binding energy, 130.12 and 130.97 eV, respectively, indicating that combination can prevent further oxidation of HRP. The W4f region of WO₃ (Fig. 3(b)) was divided into two peaks at 35.52 and 36.33 eV, assigned to W4f_5/2 and W4f_7/2 orbitals of W⁵⁺. Another two peaks at 37.66 and 38.32 eV were corresponded to the W4f_5/2 and W4f_7/2 orbitals of W⁶⁺. W⁵⁺ was derived from the unsaturated W-O bonds on the surface of WO₃. The presence of W⁵⁺ indicated the exfoliation of bulk WO₃ into ultrathin WO₃ nanosheets. The binding energy at 41.63 eV was associated with W5f_3/2^[24]. While, the binding energies of W4f_5/2 and W4f_7/2 in the 5%WO₃/HRP composite shifted to low binding energy, and the ratio of W⁶⁺/W⁵⁺ increased, which facilitated its energy conversion. Thus, when HRP was closely contacted with WO₃, the electrons could flow into WO₃ from HRP though the interface. Therefore, the heterojunction was formed between WO₃ and HRP, leading to strong electronic interaction.

The photodegradation performance of these catalysts was tested under visible light irradiation with RhB as the target pollutant (Fig. 4(a)). The equilibrium state of adsorption-desorption between the photocatalyst and the target pollutant was achieved after 30 min of reaction in the dark. Obviously, WO₃/HRP composites showed better adsorption capacity for RhB. Under visible light condition, the photodegradation performance of WO₃/HRP composites were higher than those of HRP and WO₃, and 5%WO₃/HRP composite was demonstrated the highest photocatalytic activity, and its removal rate reached 97.6% after 4 min. However, with further increasing of the WO₃/HRP ratio, the photocatalytic performance decreased since the excess WO₃ hinder the transfer of photogenerated carriers and reduce the exposure of active sites.

The photodegradation reactions of RhB in the synthesized samples conformed to the pseudo-first-order reaction model (Fig. 4(b)):

where k is the rate consistent (min⁻¹), C₀′ is the concentration of pollutants in the equilibrium of adsorption- desorption, and C_t is the concentration of contaminants for the remaining time after irradiation. The k of HRP, WO₃, 3%WO₃/HRP, 5%WO₃/HRP and 7%WO₃/HRP were 0.17, 0.01, 0.68, 0.75 and 0.69 min^-1, respectively. Among them, the k of 5%WO₃/HRP composite was highest, which was 4.5 and 75 times of those of HRP and WO₃. Therefore, the formation of heterojunction could enhance the photocatalytic activity. Furthermore, after five cycles of photocatalytic degradation of RhB, 5%WO₃/HRP composite still had high photocatalytic activity (90.4%), showing good stability with practical application potential.

The photoreduction ability of HRP, WO₃ and WO₃/HRP composites were examined by photocatalytic water splitting into hydrogen. HRP possessed a relatively low photocatalytic activity with hydrogen production rate of 240.5 μmol·h^-1·g^-1, and the pure WO₃ was almost absent, whereas a sharp increase in the rate of hydrogen production was observed in the WO₃/HRP composite (Fig. 4(c)). Fascinatingly, the hydrogen evolution rate of 5%WO₃/HRP composite was 870.69 μmol·h^-1·g^-1, which was 3.62 times of that of pure HRP. The apparent quantum efficiency (AQE) was calculated according to the following formula^[26]:

Where N_A is the Avogadro’s constant (6.02×10²³ mol^-1), M is the average H₂ generation rate (mol·s^-1), E is power of lamp source, h is the Plank’s constant (6.626×10^-34 J·s), λ is the excitation wavelength, and c is the speed of light (3.0×10⁸ m·s^-1).

AQE of 5%WO₃/HRP composite was calculated to be 11.61%. Further increasing the amount of WO₃ beyond 5% lead to a decrease in photocatalytic hydrogen evolution rate, which could originate from shielding of the light absorption by excess amount of WO₃. Therefore, appropriate WO₃ content had a remarkable effect on the activity enhancement of WO₃/HRP composite.

The above RhB photodegradation and photocatalytic hydrogen evolution results exhibited that the 5%WO₃/ HRP composite possessed superior photocatalytic activity. The introduction of appropriate WO₃ content inhibits the overgrowth and agglomeration of HRP, and the heterojunction between WO₃ and HRP promotes the separation of electrons and holes and accelerates the carrier migration.

The optical absorption properties of prepared samples were investigated by UV-Vis DRS (Fig. 5(a)). HRP had significant absorption of visible light, meaning it was typical visible light photocatalytic material. For pure WO₃, the absorption edge was located at around 446 nm. Compared with pure WO₃ and HRP, the 5%WO₃/HRP composite had a red-shifted absorption edge, which illustrated enhancement of absorption in both UV and IR regions. The UV-visible light absorption performance of 5%WO₃/HRP composite combined the advantages of two materials, which successfully formed a heterojunction, and enhanced light-trapping ability. The band gap (E_g) values of the pure HRP and WO₃ were obtained through transformation with Kubella-Munk function, and were 1.9 and 2.78 eV, respectively (Fig. 5(b)).

The generation, migration, and recombination processes of photocarriers were investigated by PL spectra, transient photocurrent (I-t) curves and electrochemical impedance spectroscopy (EIS) measurements^[29]. The 5%WO₃/HRP composite had lower PL signal, indicating lower recombination of photoinduced electrons and holes (Fig. S3(c)). In addition, the photocurrent signal intensity of the 5%WO₃/HRP composite was 3.3 μA/cm², which was 4~5 times of those of HRP (0.8 μA/cm²) and WO₃ (0.7 μA/cm²), respectively (Fig. 5(c)). These results implied that the lifetime of the photoexcited electron-hole pairs was significantly prolonged due to the more efficient separation, after combining HRP with WO₃.

The arc radius of 5%WO₃/HRP composite was smaller than those of HRP and WO₃, indicating the minimum charge impendence and the fastest reaction speed (Fig. 6(a)). This finding was consistent with the higher photocatalytic activity of the 5%WO₃/HRP composite due to the tight interface between WO₃ and HRP, which promoted carrier separation migration. Similarly, the equivalent circuit was analyzed to provide more intuitive understanding of the internal charge and surface charge- transfer mechanism of the catalyst during the reaction (insert in Fig. 6(a)). Here, R₁ and R₂ were denoted as the electrolyte solution and charge transfer resistances, which included the resistances of the photocatalyst, ITO substrate, electrolyte solution, and wire connections throughout the circuit. CPE₁ was the constant phase element that represents the bilayer capacitance of the charge transfer, and W₁ was the resistance with interfacial diffusion. R₁ of HRP, WO₃ and 5%WO₃/HRP were 10.1, 11.71 and 12.53 Ω, respectively. Under the same concentration, R₂ of HRP, WO₃, HRP and 5%WO₃/HRP were 672, 167 and 93.4 Ω, respectively. Obviously, the interfacial charge resistances of heterojunction composites were lower, indicating easier charge transfer. Therefore, the superior photocatalytic activity of 5%WO₃/HRP composite contributed to the highest photoelectric conversion efficiency, smaller interfacial transfer impendence and enhanced visible light absorption ability. The carrier lifetime (τ_e) of samples are based on the EIS bode plots (Fig. S4(a, b)), proving the carrier transfer process, according to the equation^[30]:

where f_max is the maximum frequency peak position, the calculated τ_e of the HRP and 5%WO₃/HRP composite were 1.68 and 2.7 ms, respectively. The carrier lifetime of 5%WO₃/HRP heterostructure composite was much longer than those of pure HRP, suggesting that the formation of the heterostructure between HRP and WO₃ could greatly prolong the lifetime of the photo-generated electrons, and enhanced the photocatalytic activity.

The energy band structures of samples were revealed by Mott-Schott (M-S) measurements (Fig. 6(b)). The flat potentials (E_fb) of HRP and WO₃ were -0.8 and -0.66 V (vs. Ag/AgCl), respectively, and then converted to the hydrogen standard electrode potential based on the formula:

Where E_AgCl is 0.197 V, and pH of the electrolyte was 6.8. Here, E_fb (vs. NHE) of HRP and WO₃ were -0.2 and -0.06 V, respectively. The slopes of WO₃ and HRP curves were positive, both samples were n-type semiconductors, and E_fb is 0.1-0.3 V higher than its conduction band potential (E_CB). Therefore, the E_CB of HRP and WO₃ were -0.4 and -0.26 V (vs. NHE). In accordance with the formula: E_g = E_VB-E_CB, the E_VB were 1.5 and 2.52 V, respectively.

The hydroxyl radicals (·OH) was important species in the photocatalytic reactions, EPR spectra were presented in Fig. 6(c). HRP had relatively weak signal because of the weak oxidation potential of photogenerated holes, WO₃ displayed moderately strong signal due to its more positive E_VB, and 5%WO₃/HRP composite had strong DMPO-·OH signal, so the photogenerated holes remained in WO₃ and did not transfer to the VB of HRP. Because the potential of OH^-/·OH pair was +2.38 V (vs. NHE)^[31], which was positive than that of HRP (+1.5 V (vs. NHE)) and negative than that of the WO₃ (+2.52 V (vs. NHE), inferring that the ·OH radical was produced by WO₃. These results indicated that the photogenerated electrons and holes in 5%WO₃/HRP composite were present in the CB of HRP and VB of WO₃, respectively, and the charge transfer belonged to S-scheme heterojunction.

Based on above discussions, the S-scheme mechanism was proposed in Fig. 7. HRP is a reducing photocatalyst with smaller work function (5.61 eV)^[17] and higher Fermi level. WO₃ is an oxidizing photocatalyst with large work function (6.23 eV)^[32] and lower Fermi level (Fig. 7(a)). When the WO₃ photocatalysts was in close contact with HRP, electrons spontaneously transferred from HRP to WO₃ until their E_f reached the same level. During the migration of electrons, the interface region of WO₃ possesses a positive charge due to the loss of electrons, which leads to the formation of electron depletion layer and the upward bending of the energy band. While interface region near the HRP is negatively charged due to the gain of electrons, which leads to the formation of an electron accumulation layer and the downward bending of the band edge. As a result, an internal electric field is formed at the interface of the WO₃/HRP heterojunction, impeding the continuous flow of electrons from HRP to WO₃ (Fig. 7(b)). Under visible light irradiation, the electrons were excited from VB to CB of WO₃ and HRP. Driven by the internal electric field, band bending and Coulomb interaction, the photogenerated electrons in the CB of WO₃ spontaneously slid toward HRP, and recombined with the holes on the VB of HRP. However, the useful electrons and holes of strong redox ability could be retained (Fig. 7(c)). Therefore, the improved photocatalytic performance of the WO₃/HRP composite was mainly ascribed to the formation of S-scheme heterojunctions, which contribute to the strong redox capacity for the degradation of organic water pollutants and hydrogen generation.

In this study, the S-scheme WO₃/HRP heterojunction photocatalysts were prepared by the hydrothermal method. The heterojunctions displayed significantly enhanced photocatalytic RhB degradation and hydrogen evolution. The 5%WO₃/HRP heterojunction photocatalysts degraded 97.6% RhB within 4 min. Meanwhile, the photocatalytic hydrogen evolution was almost 3.62 times of pure HRP. The enhanced photocatalytic performance was attributed to the S-scheme heterojunction between WO₃ and HRP, which effectively transferred the photo-generated carriers, and suppressed the recombination of electron-hole pairs. This work could provide new prospect for design and construction of novel heterojunction photocatalyst.