In Situ Growth of the Bi2S3 Nanowire Array on the Bi2MoO6 Film for an Improved Photoelectrochemical Performance.

A single-crystalline Bi2S3 nanowire array (Bi2S3NWA) is synthesized by an in situ hydrothermal reaction on the surface of a Bi2MoO6 film. As no additional source of Bi3+ is provided during the process, the Bi2MoO6 layer acts as the Bi3+ source for the synthesis of Bi2S3 nanowires. The fabricated Bi2MoO6/Bi2S3NWA electrode exhibited an increased photoelectrochemical (PEC) sulfite oxidation activity, which is attributed mainly to the effective interface obtained by the in situ hydrothermal growth, compared to other Bi2S3 electrodes. The generated electron from the Bi2S3 conduction band rapidly transfers to that of Bi2MoO6, yielding an enhanced electron separation of Bi2S3. Furthermore, the single-crystalline Bi2S3 nanowire can provide a fast electron pathway to Bi2MoO6 through its single domain, which also contributes to the improved PEC activity.

Introduction

Utilization of sunlight has been considered a promising solution to overcome the exhaustion of fossil fuels.[1] The development of semiconductor photoelectrodes has attracted considerable attention for the efficient conversion of solar energy.[2−6] The examination of the semiconductors in photoelectrochemical (PEC) cells is a fast and simple method to characterize their electrochemical behaviors.[3] Although many semiconductors have been examined as photoelectrodes,[7−18] most of them have modest PEC efficiencies owing to the narrow visible light absorptions and unavoidable electron–hole recombinations. To decrease the bulk recombination, nanostructures, such as nanoparticles and nanorods, have been studied as photoelectrodes,[19] which can enhance the kinetic parameters of the PEC reactions because of the reduced hole diffusion length. The shape, size, and connection of the nanostructures can affect the charge transport properties and thus, play a critical role in the efficiency of PEC cells. However, these nanostructures exhibit several disadvantages, such as a reduced space charge region and high surface recombination.[6] Therefore, further studies are required to optimize the sizes and shapes of nanostructures for improved PEC performances.

Bismuth sulfide (Bi2S3) has attracted large interest as a sensitizer for photovoltaic and PEC cells owing to its narrow band gap of approximately 1.3 eV and large absorption coefficient.[19−25] However, the reported photocurrent is considerably lower than its theoretical maximum (∼30 mA/cm2). The recombination at the surface and in the bulk state of Bi2S3, leads to a lower PEC conversion efficiency. One-dimensional Bi2S3 structures such as nanorods and nanowires have been studied as photoelectrodes, which can provide the advantages of a reduced hole diffusion length and enhanced charge transport properties.[26−30] On the other hand, the Bi2S3 nanowires must be grown on a conductive substrate to enable efficient interactions for improved PEC properties. As most Bi2S3 nanowires are prepared as powdered samples, their deposition on substrates is another challenging process.[31] To decrease the interface resistance between the Bi2S3 nanowire and the conductive substrate, an in situ synthetic method is required. Thus, a simple synthesis on the substrate is required to reduce the fabrication complexity of the PEC devices.

In this paper, we report a facile synthesis of a Bi2S3 nanowire array (Bi2S3NWA) on a Bi2MoO6 layer by an in situ hydrothermal method. The Bi2S3NWA was uniformly fabricated and well connected on the Bi2MoO6 layer. Structural characterizations were carried out by transmission electron microscopy (TEM), Raman spectroscopy, and X-ray diffraction (XRD). Compared to other Bi2S3 electrodes, the fabricated Bi2MoO6/Bi2S3NWA composite exhibited an increased PEC sulfite oxidation activity, which is attributed mainly to the effective interface obtained by the in situ growth. This unique composite also has dual electrochemical functions, as a light absorber (Bi2S3NWA) and an electron acceptor (Bi2MoO6 layer). The generated electron from the Bi2S3 conduction band rapidly transfers to that of Bi2MoO6, yielding an enhanced charge separation of Bi2S3. In addition, the single-crystalline Bi2S3 nanowire can provide a fast electron pathway through its single domain, which also contributes to the increased PEC activity.

Results and Discussion

Fabrication of the Bi2MoO6/Bi2S3NWA Composite Electrode

Bi2S3NWA was synthesized in situ on the surface of a Bi2MoO6 film (Scheme ). The Bi2MoO6 film was fabricated by a drop-casting method of a precursor solution onto a fluorine-doped tin oxide (FTO) substrate and subsequent annealing at 500 °C for 3 h in air (Figure a). Bi2S3NWA was then directly grown on the Bi2MoO6 film by a hydrothermal reaction (Bi2MoO6/Bi2S3NWA). Figure b shows a scanning electron microscopy (SEM) image of Bi2S3NWA fabricated on Bi2MoO6. The nanowires have an average width of 87 ± 14 nm and a length of 1.2 ± 0.2 μm (Figure S1).

Scheme 1

Schematic of the Fabrication of the Bi2MoO6/Bi2S3NWA Composite on the FTO Substrate

Photograph courtesy of J. H. Kim and K. M. Nam. Copyright 2019.

Figure 1

SEM images and XRD patterns of (a,c) Bi2MoO6 and (b,d) Bi2MoO6/Bi2S3NWA composite on the FTO substrate, respectively.

Figure c,d shows XRD patterns of the Bi2MoO6 and Bi2MoO6/Bi2S3NWA composite films, respectively. The diffraction peaks of Bi2MoO6 are indexed to the orthorhombic structure (Joint Committee on Powder Diffraction Standards (JCPDS) no. 21-0102), indicating its high crystallinity (Figure c). Figure d shows the XRD pattern of the Bi2MoO6/Bi2S3NWA composite. New peaks are detected in addition to the Bi2MoO6 peaks, which correspond to the Bi2S3 phase (orthorhombic, JCPDS no. 17-0320). Additional phases such as Bi2O3 and MoO3 were not discovered in the XRD patterns (Figure d).

TEM and high-resolution TEM (HRTEM) analyses were carried out to identify the structural composition of the Bi2MoO6/Bi2S3NWA composite (Figures and S2). The HRTEM images of an edge of a single Bi2S3 nanowire show the lattice spacings of 0.25 and 0.20 nm corresponding to the (420) and (002) planes (Figure S2a,b).[30,32] The fast Fourier transform patterns indicate the single-crystalline Bi2S3 nanowire, reflecting the fast crystalline growth toward the [001] direction (Figure S2b). Energy-dispersive X-ray spectrometry of the nanowires showed an average atomic ratio of 39:60 (Bi/S), indicative of the 2:3 atomic composition. A spatial elemental mapping was performed on a Bi2S3 nanowire to evaluate the distribution of each element (Figure c–f). The nanowire was composed of Bi and S in the ratio of 2:3 (Figure d,e), while the substrate was composed of Bi and Mo in the ratio of 2:1 (Figure S2). The elemental mapping confirmed that the Bi2S3 nanowires were directly grown on the Bi2MoO6 layer (Figure S2).

Figure 2

(a) TEM and (b) HRTEM images of the Bi2S3 nanowire. (c) Spatial elemental map of the Bi2S3 nanowire and corresponding elemental distribution maps of (d) Bi (red), (e) S (blue), and (f) Mo (green).

Figure shows Raman spectra of the Bi2MoO6 and Bi2MoO6/Bi2S3NWA composite films in the range of 50–900 cm–1. The Raman peaks of the Bi2MoO6/Bi2S3NWA composite at 98, 188, 236, and 255 cm–1 are in agreement with those of the orthorhombic Bi2S3 (red line in Figure ).[33] Bi2MoO6 peaks are not observed in the spectrum of Bi2MoO6/Bi2S3NWA, indicating that the Bi2S3 layer is uniformly grown on the surface of the Bi2MoO6 layer.[34,35] The surface states of both Bi2MoO6 and Bi2MoO6/Bi2S3NWA films are further characterized by X-ray photoelectron spectroscopy (XPS) (Figure ).[36] The two high peaks at 164.1 and 158.8 eV are assigned to Bi 4f (Figure a,b). The S 2s peak at 224.8 eV (S2– in the metal sulfide) is detected only for the Bi2MoO6/Bi2S3NWA composite film, but not for the Bi2MoO6 film (Figure c,d). Therefore, the XRD, SEM, TEM, Raman spectroscopy, and XPS results verify the presence of Bi2S3NWA on the Bi2MoO6 surface.

Figure 3

Raman spectra of Bi2MoO6 and Bi2MoO6/Bi2S3NWA electrodes.

Figure 4

XPS spectra of (a) Bi 4f, (c) Mo 3d for Bi2MoO6, (b) Bi 4f, and (d) Mo 3d and S 2s for the Bi2MoO6/Bi2S3NWA electrode.

Sulfide anions (S2–) were produced by the decomposition of thiourea during the hydrothermal reaction.[24] Small amounts of Bi3+ and Mo6+ ions were also generated by the dissolution of Bi2MoO6 during the reaction. The solubility of Bi2S3 in an aqueous solution is considerably lower than that of molybdenum sulfide, which have Ksp values of 1.0 × 10–97 and 2.2 × 10–56,[37−39] respectively. The Bi3+ ions reacted more rapidly than the Mo6+ ions with S2– ions forming the Bi2S3. As no additional source of Bi3+ was provided during the process, the Bi2MoO6 layer acted as the Bi3+ source for the growth of Bi2S3 nanowires. As the Bi2S3 has an extremely low solubility (Ksp of 1.0 × 10–97), the hydrothermal sulfidation of the Bi2MoO6 may spontaneously follow the etching and regrowth mechanisms proposed by Chen et al.[38] The Bi2S3 nanowires grow rapidly along the direction vertical to the substrate in a large amount of S2–.[38,40]

To obtain a reliable growth, the reaction solution with the Bi2MoO6 substrate was heated at a relatively low temperature (60 °C for 4 h) in the presence of a large amount of S2– (0.2 M of Na2S). The surface of Bi2MoO6 was initially white, but after the reaction for 4 h, it turned to black brown, indicating the formation of Bi2S3 (Figure S3). The Bi2S3 layer grew larger and then reorganized forming amorphous Bi2S3 particles, but not nanowires. However, several Bi2S3 nanowires were grown on the surface of the Bi2MoO6 electrode at 100 °C for 4 h. Therefore, both large amounts of S2– and a high reaction temperature (above 100 °C) were needed to follow the etching and regrowth mechanisms for the growth of Bi2S3 nanowires on the Bi2MoO6 surface. The unreacted Mo6+ ions were analyzed by inductively coupled plasma atomic emission spectroscopy. The Mo6+ ions remained in the solution throughout the reaction without the formation of a molybdenum sulfide film.

PEC Characterization of the Bi2MoO6/Bi2S3NW Composite Electrode

Various Bi2MoO6/Bi2S3 heterostructures were prepared by well-known synthesis methods to compare the PEC performances of the Bi2MoO6/Bi2S3NWA electrodes (Scheme ). The structures are denoted as Bi2MoO6/Bi2S3(Drop) and Bi2MoO6/Bi2S3(Dip). Furthermore, pure Bi2S3 nanowires (Bi2S3NW) were fabricated by the hydrothermal method for comparison.[30]Figure S4 shows SEM images of the Bi2MoO6/Bi2S3(Dip), Bi2MoO6/Bi2S3(Drop), and Bi2S3NW electrodes. The Bi2MoO6/Bi2S3NWA and Bi2MoO6/Bi2S3(Drop) samples were adjusted with Bi2S3 layers of the same thickness using their UV–visible absorption spectra (Figure S5).

Scheme 2

Schematic Image of the Preparation of Bi2MoO6/Bi2S3 Hetero-Structures

Photograph courtesy of J. H. Kim and K. M. Nam. Copyright 2019.

A sacrificial reagent, sodium sulfite (Na2SO3), was used to examine the PEC performance. Sulfite anions (SO32–) are an efficient hydroxyl radical (•OH) scavenger and react in most diffusion-controlled rates.[41] Therefore, the SO32– oxidation is a suitable model reaction for the measurement of the degree of bulk recombination of the semiconductor. It is considered to provide a surface transfer efficiency of almost 100% during the PEC measurement owing to the fast kinetics at the semiconductor–solution interface.[42]

The PEC performances of the electrodes were investigated using linear sweep voltammetry (LSV) for sulfite oxidation (0.1 M Na2S + 0.1 M Na2SO3). The LSV was carried out in the range of −0.9 to −0.1 V versus Ag/AgCl at a scan rate of 20 mV/s with a chopped UV–visible irradiation (Figures and S6). All electrodes successfully generated anodic photocurrents, which confirmed their n-type characteristics. The photocurrent of the Bi2MoO6/Bi2S3NWA electrode was 7.6 mA/cm2 at −0.2 V (vs Ag/AgCl), while that of Bi2MoO6/Bi2S3(Drop), fabricated by drop-casting the Bi2S3 nanowires onto the Bi2MoO6 electrode, was 0.8 mA/cm2 at the same potential (black bars in Figure b). The fabricated Bi2MoO6/Bi2S3NWA electrode exhibited highly increased PEC activity compared to the Bi2MoO6/Bi2S3(Drop), which is attributed to the effective composite interface obtained by the in situ growth. The photocurrent of the Bi2MoO6/Bi2S3(Dip) and Bi2S3NW were 2.2, and 0.7 mA/cm2, respectively, at the same condition (black bars in Figure b), which indicates the importance of single-crystalline Bi2S3 nanowires on the Bi2MoO6 substrate. The photocurrent of the Bi2MoO6/Bi2S3NWA electrode was also several times higher than those of the Bi2MoO6/Bi2S3(Drop), Bi2MoO6/Bi2S3(Dip), and Bi2S3NW electrodes under the visible light irradiation (red bars in Figure b).

Figure 5

(a) LSVs of the Bi2MoO6/Bi2S3NWA composite under a UV–visible illumination (back side) in a solution of 0.1 M Na2SO3 and 0.1 M Na2S recorded at a scan rate of 20 mV/s and a light intensity of 100 mW/cm2. (b) Photocurrent densities of the electrodes having Bi2MoO6/Bi2S3 heterostructures under UV–visible (black) and visible (red) illuminations at an applied potential of −0.20 V vs Ag/AgCl. (c) Action spectrum and (d) current–time response curve of the Bi2MoO6/Bi2S3NWA composite at an applied potential of −0.50 V vs Ag/AgCl.

Figure c presents the action spectrum of the Bi2MoO6/Bi2S3NWA electrode, which shows the photocurrent depending on the wavelength, acquired with an interval of 10 nm. The portion of the visible light region (>425 nm) is about 85%, which matches well with Figure a. Furthermore, the band gap was determined using the wavelength of the onset photocurrent. The action spectrum indicates a band gap of 1.35 eV. It is well matched with the absorption spectrum (Figure S5). The Bi2MoO6/Bi2S3NW electrode exhibits an identical onset wavelength as that of Bi2S3 (about 1.3 eV),[25] which indicates that Bi2S3NWA is the main absorber for the PEC reaction. The band gaps of the electrodes were also calculated by the Tauc equation.[43] The direct band gaps of the heterostructures (Bi2MoO6/Bi2S3(Drop) and Bi2MoO6/Bi2S3NWA) are almost identical (Figure S5), and thus, the band gap difference is not the main factor for the PEC activity in the case of the Bi2MoO6/Bi2S3 heterostructures.

To evaluate the stability of Bi2MoO6/Bi2S3NWA, chronoamperometry was carried out under a UV–visible illumination (Figure d). The current transient upon the light irradiation was usually attributed to the dynamic balance of photogenerated electrons and their consumption at the semiconductor–solution interface.[44,45] This is a characteristic indication of surface recombination even in the sacrificial reagent.[46] Although the Bi2MoO6/Bi2S3NW electrode exhibited initial decreases in the photocurrent, the photocurrents stabilized at steady-state values for 2 h in the presence of the sacrificial reagent (Figure d).

In order to understand the mechanism of the increased PEC activity of the Bi2MoO6/Bi2S3NWA electrode, the flat-band potentials of Bi2MoO6 and Bi2S3 were measured. The Mott–Schottky (MS) plots were recorded using a Na2SO4 solution (Figure a,b).[30] The MS plot indicated that the flat-band potential of Bi2S3 is approximately 0.15 V [vs normal hydrogen electrode (NHE)] with an n-type behavior. The flat band potential of Bi2MoO6, estimated using the MS plot was approximately 0.25 V (vs NHE). The conduction band edge (ECB) is considered to be more negative than the flat band potentials by approximately 0.1 V.[23,30] Therefore, the typical ECB and valence band edges (EVB) of Bi2S3 are approximately 0.05 and 1.40 eV, respectively, (vs NHE), while those of Bi2MoO6 are approximately 0.15 and 2.95 eV (vs NHE), respectively, which are matched well with the literatures.[30]Figure c shows possible photogenerated electron–hole pathways based on the MS plots: (1) electron transfer from the ECB of Bi2S3 to that of Bi2MoO6, (2) hole transfer from the EVB of Bi2MoO6 to that of Bi2S3, and (3) electron–hole recombination at the Bi2S3–Bi2MoO6 interface.

Figure 6

Mott–Schottky plots of (a) Bi2MoO6 and (b) Bi2S3 obtained by ac impedance–capacitance measurements in a 0.1 M Na2SO4 solution at 500 Hz. (c) Photogenerated electron–hole pathways between the two semiconductors (Bi2MoO6 and Bi2S3).

ECB and EVB of Bi2MoO6 are more positive than those of Bi2S3, favorable for the charge separation of Bi2S3. The origin of the increased PEC activity of the heterostructure compared to that of the bare Bi2S3 electrode is attributed to the enhanced charge separation of Bi2S3 on the Bi2MoO6 electrode (pathways 1 and 2 in Figure c) under the solar light irradiation. Furthermore, the Bi2MoO6/Bi2S3NWA electrode fabricated in situ has an effective interface with a lower resistance (reduced pathway 3 in Figure c) than that of the Bi2MoO6/Bi2S3(Drop) electrode. In addition, the Bi2S3 nanowires can provide fast electrical pathways to Bi2MoO6 through their single domains, compared to that of the Bi2MoO6/Bi2S3(Dip) electrode.

Conclusions

The Bi2MoO6/Bi2S3NWA electrode was prepared by the in situ growth reaction. The Bi2S3 nanowires were homogeneously grown on the Bi2MoO6 layer. As no additional source of Bi3+ was provided during the process, the Bi2MoO6 layer acted as the substrate and the Bi3+ source for the synthesis of Bi2S3 nanowires. The spontaneous growth of Bi2S3 nanowires from the Bi2MoO6 by the etching and regrowth mechanisms provided a suitable interaction with the Bi2MoO6 substrate. The fabricated Bi2MoO6/Bi2S3NWA electrode exhibited an increased PEC activity compared to other electrodes. The increase in activity was attributed mainly to three factors: (1) increased electron separation of Bi2S3 on the Bi2MoO6 layer as ECB and EVB of Bi2MoO6 were more positive than those of Bi2S3, encouraging charge separation of Bi2S3, (2) a suitable interface obtained by the in situ synthetic method, and (3) single-crystalline Bi2S3 nanowires which provided a fast electron pathway through the nanowire.

Methods

Materials

An FTO- coated glass (TEC 15, WY-GMS) was used as the substrate for the thin-film electrode. Bi(NO3)3·5H2O (99.999%, Sigma-Aldrich) and (NH4)6Mo7O24·4H2O (99.98%, Sigma-Aldrich) were used as metal precursor salts. Thiourea (≥99%, Sigma-Aldrich), hydrochloric acid (36.5%, Junsei), ethylene glycol (≥99%, Sigma-Aldrich), isopropyl alcohol (99.5%, Junsei), sodium sulfite (≥98%, Sigma-Aldrich), sodium sulfate (99.0%, Daejung Chemicals), and sodium sulfide nonahydrate (96.0%, Junsei) were used as received. Deionized water was used as the solvent in all electrochemical experiments.

Fabrication of the Bi2MoO6 Film on FTO

FTO substrates were cleaned in deionized water and ethanol, and then sonicated in ethanol for at least 1 h. Drop-casting was carried out to fabricate the thin-film electrodes. A solution of 15 mM Bi2MoO6 precursor (an atomic ratio of Bi/Mo of 2:1) in ethylene glycol was prepared and then applied onto the FTO substrate (15 mM, 200 μL). The prepared film was annealed at 500 °C for 3 h (with a ramp time of 3 h) in air to form the Bi2MoO6 thin film.

Fabrication of Bi2S3NWA on Bi2MoO6/FTO

Bi2S3 nanowires were directly grown on the surface of the Bi2MoO6 layer using an in situ hydrothermal reaction. The Bi2MoO6 layer on the FTO substrate was horizontally placed into a 50 mL Teflon-lined stainless-steel autoclave. A reaction solution was prepared by adding thiourea (7.9 mmol) into deionized water (30 mL). The reaction solution was transferred into an autoclave, and then heated in an electric oven at 140 °C for 4 h. The resulting Bi2MoO6/Bi2S3NWA composite electrode was dried at 50 °C in air.

When the reaction solution was maintained at 140 °C for 48 h, Bi2S3 nanowires were grown directly on the surface of the FTO substrate. The nanowires fully covered the FTO substrate, and the majority were slightly tilted on the substrate.