Plasmonic Metal/Semiconductor Heterostructure for Visible Light-Enhanced H2 Production.

A plasmonic Ag/Bi2WO6 heterostructure, having Ag NPs deposited on Bi2WO6, is obtained by a hydrothermal and photodeposition method. The synthesized Ag/Bi2WO6 composite exhibits strong visible light absorption with a localized surface plasmon resonance (LSPR) and shows an enhanced photoabsorption property. It is demonstrated that such a Ag/Bi2WO6 heterostructure shows excellent plasmon-enhanced photocatalytic activity in the dehydrogenation of ammonia borane (NH3BH3) solution under visible light irradiation, which is due to the results from the synergetic effect between Ag NPs and emerging W5+ ions. More importantly, the performance of a Ag/Bi2WO6 hybrid is almost eight times higher than that of sole Bi2WO6 nanosheets. The introduction of LSPR of Ag in Bi2WO6 improves the electrical conductivity of the composite and lowers the recombination rate of charge carriers. This study opens up the opportunity of rationally fabricating plasmonic metal/semiconductor heterostructures for highly efficient photocatalysis.

Introduction

Conversion of solar energy to hydrogen fuel from hydrogen storage materials through photocatalysis is accepted to be an excellent technique to harvest energy and to solve worldwide environmental issues.[1,2] Photocatalytic hydrolysis from hydrogen-rich compounds involves a semiconductor material as a photocatalyst, such as TiO2, and a solar light source for hydrogen evolution.[3,4] Alternatively, the localized surface plasmon resonance (LSPR) of metals such as Au, Ag, and Pd along with a semiconductor has proven to be a significant candidate for photocatalysis.[5] Plasmonic nanostructures develop oscillation of electrons when they are incident by the light wave of plasmonic resonance frequency, which produces a bound or localized electromagnetic mode in a confined plasmonic nanostructure. This creates an enhanced electric field that produces energetic electrons and heat, which has been reported to be significant in the conversion of solar light energy to activate chemical reactions.[6] LSPR has proved to be highly effective in heterogeneous photocatalysis, and a single-component photocatalyst has low catalytic efficiency and cannot fulfill the desired requirements. Recently, a number of research studies on noble metal/semiconductor hybrid photocatalysts have been done, and they have proven successful in many reactions such as degradation, hydrogen evolution, hydrogenation, and oxidation.[7,8] The noble metal/semiconductor hybrid has found a strong place in the field of photocatalysis. Loading of a noble metal on a photocatalyst can result in an extended light response and enhances the interfacial charge transfer efficiency.[7−11] This enhanced local electric field leads to the increased interband transition rate, making the energy generated by LSPR higher than the bandgap of a semiconductor and increasing the electron–hole pair separation in the photocatalyst.[12−15] The electrons are directed toward a noble metal, and holes are accumulated on the other edge of a semiconductor. The noble metal acts as an electron trapper and reduces the recombination rate of electron–hole pairs. This process enhances the photoactivity of the photocatalyst. In contrast, too much noble metal reduces the active sites for the reaction and acts as a recombination center. The high concentration of noble metals blocks the active sites of the reaction. Figure S1 in the Supporting Information shows that the noble metal on the semiconductor absorbs the incident light and undergoes surface plasmon oscillation, which excites electrons and holes. However, the plasmonic enhancement of photoconversion is still a great challenge. To attain progress in this method, some basic problems, such as the fabrication size, geometry, and combination (molar ratio) of the noble metal and semiconductor, need to be thoroughly investigated. The energy transfer between plasmonic metals and semiconductors takes place through three mechanisms: light scattering, hot electron injection, and plasmon-induced resonance energy transfer.[16−18] Designing a plasmonic metal/semiconductor photocatalyst is a big challenge. However, some reported work shows that strongly coupled metal/semiconductor nanostructures generated a high intensity of LSPR, but the kind of architecture of the metal/semiconductor heterostructure is still a mystery.[19−21]

There are studies available that have worked on the metal and semiconductor combination. Yu et al., in their work, have shown the LSPR effect of Au-chain@ZnCd1–S and reported 54.6% of H2 production.[22] Simagina and his team have also reported 330 μmol of hydrogen evolution from ammonia borane in 3 h over Ag/TiO2.[23] Zhang et al. reported 3 mL/min H2 production from NH3BH3 in 60 min over PtNi-graphite.[24] Hydrogen production of 1.1 mL/min from NH3BH3 (AB) has also been reported in Pt@SiO2 heterostructures.[25]Table summarizes the hydrogen evolution from ammonia borane over some reported plasmonic photocatalysts. Many noble metals, such as Ru, Rh, Au, and Pd, have resulted in high hydrogen yield from AB solution.[6,26] Silver nanoparticles, being cost-effective, will easily cut off the expenses of the photocatalyst. Herein, we report a green chemistry route to synthesize a plasmonic Ag/Bi2WO6 nanostructure by coupling a Bi2WO6 semiconductor and silver metal nanoparticles (NPs). More importantly, such a Ag/Bi2WO6 hybrid displayed a dramatic plasmon-enhanced photocatalytic activity in the photocatalytic hydrolysis of NH3BH3 solution under visible light irradiation.

Table 1

Hydrogen Evolution from Ammonia Borane over Some Reported Photocatalysts

catalyst	preparation method	time (min)	hydrogen evolution	reference
Bi2WO6	hydrothermal	150	0.050 μmol/min	present work
Ag/Bi2WO6 (1:1)	hydrothermal and photodeposition	150	0.13 μmol/min	present work
Ag/Bi2WO6 (1:2)	hydrothermal and photodeposition	150	6.608 μmol/min	present work
Ag/Bi2WO6 (2:1)	hydrothermal and photodeposition	150	0.57 μmol/min	present work
MoO3	solvothermal	50	63.3 mol %	(27)
WO3	solvothermal	50	10 mol %	(27)
Cu/TiO2	sol–gel	60	90 mol %	(28)
Pt/TiO2-ZnO	sol–gel	300	88 mol %	(29)
CdS-TiO2	electrospinning	60	95 mol %	(30)
TiO2 (nanofiber)	electrospinning	60	34 mol %	(30)
Au chain@ZnxCd1–xS	hydrothermal	60	54.6 mol %	(22)
Ag/TiO2	hydrothermal	180	330 μmol	(23)
PtNi-graphite	impregnation and chemical reduction	60	3 mL/min	(24)
Pt@SiO2	sol–gel and chemical route	60	1.1 mL/min	(25)
TiO2 (carbon nanofiber)	electrospinning	60	55 mol %	(30)

Experimental Section

Synthesis of Bi2WO6

Na2WO4·2H2O (1.23 g) and Bi(NO3)3·5H2O (3.64 g) were added to a Teflon jar containing 150 mL of deionized water under magnetic stirring in a conventional hydrothermal operation. The Teflon tank was sealed in an autoclave and heated for 20 h at 160 °C. The autoclave was allowed to cool naturally to ambient temperature after the reaction period was completed; the sample was centrifuged and washed multiple times with deionized water before being dried in an oven at 80 °C for 10 h. Finally, a yellowish Bi2WO6 nanopowder was synthesized.

Synthesis of Plasmonic Ag/Bi2WO6

The introduction of Ag on the semiconductor was achieved using photoinduced deposition of Ag on Bi2WO6. In a typical synthesis, 0.085 g (0.5 mmol) of AgNO3 was added to a beaker containing 50 mL of deionized water and agitated continuously in the dark for 30 min. The AgNO3 solution was then poured to 0.349 g (0.5 mmol) of hydrothermally generated Bi2WO6 and held in a visible light chamber for 60 min. A gray precipitate was obtained, which was rinsed multiple times with deionized water before being dried for 8 h at 60 °C in an oven. By varying the molar ratios of silver and bismuth tungstate, three distinct Ag/Bi2WO6 molar ratios (1:1, 1:2, and 2:1) were created. The synthesis of Ag/Bi2WO6 is depicted graphically in Figure S2.

Characterization

X-ray diffraction (XRD) patterns were used to verify the phase purity of the produced photocatalysts using a SmartLab diffractometer (Rigaku, Japan). The patterns were captured in the 2θ range of 10–70° using Cu Kα radiation (λ = 1.5416 Å), and the scanning rate was kept at 3° m–1. FTIR and Raman spectroscopy were used to further investigate the detailed structural analyses. An IR-Prestige 21 spectrometer (Shimadzu Corp., Japan) was used to record the FTIR spectra in the frequency range of 400 to 4000 cm–1 using KBr as a diluting agent. A FESEM (Carl Zeiss Microscopy Ltd., Germany) apparatus equipped with an energy dispersive X-ray spectroscope was used to capture morphological images. Using a commercial (Quantachrome Instruments, USA) apparatus, the Brunauer–Emmett–Teller (BET) test was performed to determine the surface area, pore volume, and pore size distribution. Prior to measuring nitrogen adsorption–desorption, the produced sample was degassed at 200 °C for 4 h. Thermogravimetric analysis (TGA) was performed to determine the thermal stability of the materials using a Discovery STD-650 (TA Instruments, USA). A UV–vis spectrometer (PerkinElmer, USA) in the range of 200 to 800 nm and a photoluminescence spectrofluorometer (Shimadzu, Japan) with a 360 nm excitation wavelength were used to measure the optical characteristics. A CH instrument (Novo Control, German) was used to conduct the electrochemical analysis. A PHI 5000 (USA) was used for XPS analysis. TEM images were obtained by a transmission electron microscope (FEI, USA). The dehydrogenation of ammonia borane was studied by a gas chromatograph (Thermo Scientific, USA) with a TCD detector using a molecular sieve and argon as a carrier gas.

Photocatalytic Experiment

Photocatalytic Hydrolysis of Ammonia Borane

Dehydrogenation of NH3BH3 was carried out to assess the catalytic efficacy of bare Bi2WO6 and Ag/Bi2WO6 (1:1, 1:2, and 2:1) catalysts. Typically, a 10 mg Ag/Bi2WO6 sample was placed in a test tube with 5 mL of distilled water and Ar gas was pumped through the apparatus to render it inert. A rubber septum was used to inject 12.8 mg of NH3BH3 dissolved in 10 mL of water into the test tube. With continuous magnetic stirring, the reaction was carried out in the dark and under visible light irradiation. A gas chromatograph with a TCD detector and argon as a carrier gas was used to track the evolution of hydrogen.

Simulation of Surface Plasmon Resonance

The phenomenon of LSPR on Ag/Bi2WO6 can be easily elucidated by finite element method (FEM) simulation using COMSOL Multiphysics. The production of an electric field at the junction of Ag and the Bi2WO6 substrate was demonstrated using COMSOL modeling for a completely spherical silver particle on Bi2WO6 (Model-A) and a half-spherical silver particle on Bi2WO6 (Model-B). The geometry was created using the radius of a 12.5 nm silver nanosphere over a 50 nm × 50 nm Bi2WO6 substrate. The work permittivity of silver was set to −15.243 -i0.40284[31] in our simulation, while the refractive indexes of air and Bi2WO6 were set to 1 and 2.17, respectively.[32,33] For the simulation, a periodic boundary condition was used and the TM-polarized light wave was incident on the silver nanosphere.

Results and Discussion

Structural Study

Figure a shows the diffraction peak of pure Bi2WO6 and Ag/Bi2WO6 (1:1, 1:2, and 2:1) at room temperature. Considering the orthorhombic symmetry of the material, the characteristic diffraction peaks are indexed according to JCPDS no. 73-1126.[34] There are no distinct peaks for metallic silver, which could be due to the metal’s low concentration. A sharp, well-defined diffraction peak corresponds to the crystalline nature of the catalysts.

Figure 1

(a) XRD spectra, (b) FTIR spectra, (c) TGA spectra, and (d) Nyquist plot of Bi2WO6 and Ag/Bi2WO6 (1:1, 1:2, and 2:1).

The chemical composition of the prepared Bi2WO6 and Ag/Bi2WO6 (1:1, 1:2, and 2:1) is studied using the FTIR spectra. Bi2WO6 has an absorption band at 500–1000 cm–1. Figure b shows the stretching modes of Bi–O and W–O at 578.64 and 732.95 cm–1, respectively.[35−37] The peaks at 3464.15 and 1624.06 cm–1 are ascribed to the O–H bending and stretching vibration of adsorbed H2O molecules, respectively.[36,38] Because of the reduced amount of Ag, no separate peaks for Ag were identified, although with Ag loading, the strength of the peak at 1381.03 cm–1 grew, while the peak at 578.64 cm–1 dropped, as can be seen in the spectra. The lower intensity peak at 578.64 cm–1 is due to a reduction in the functional group associated with the Bi–O bonds per unit volume. Because photodeposition of Ag on Bi2WO6 inhibits IR radiation from reaching the molecule, light absorption by the Bi–O bond is reduced. When Ag is incorporated into Bi2WO6 nanoparticles, a new dip appears at 1381 cm–1, which represents the Ag–O bond.[39] The position of the Bi2WO6 peaks did not alter after the silver coating. The peaks have not changed, showing that Ag has not harmed Bi2WO6’s structure. Another intriguing feature of the peak at 732.95 cm–1, which is connected with the stretching mode of W–O, was that it became narrower as the molar ratio of Ag loading increased. This shows that photoinduced Ag deposition on Bi2WO6 has altered Bi2WO6’s reactivity in the IR region.

Figure c shows the TGA curve of Ag/Bi2WO6 (1:1, 1:2, and 2:1) in the temperature range of 25 to 800 °C. The maximum weight loss of 9% is observed in Ag/Bi2WO6 (2:1). The reason for weight loss is the decomposition of the silver metal; silver nanoparticles decompose and lose weight between temperatures of 200 and 450 °C.[40] The weight loss between 25 and 200 °C is due to the evaporation of moisture adsorbed from the atmosphere before performing the test. Ag/Bi2WO6 (1:1) has undergone a weight loss of 7%, and Ag/Bi2WO6 (1:2) has shown a weight loss of 4% within the given temperature range.

Electrochemical impedance spectroscopy (EIS) tests show the charge transfer mechanism in Ag/Bi2WO6 composites with varying molar ratios. The Nyquist plots of Ag/Bi2WO6 (1:1, 1:2, and 2:1) in dark and light conditions are shown in Figure d. The Nyquist plot’s semicircle has a reduced diameter, indicating that the photogenerated electron–hole pair is effectively separated in the materials. The arc radius of Ag/Bi2WO6 (1:2) was substantially lower than that of Ag/Bi2WO6 (1:1 and 2:1) in the current EIS measurement. It shows that the LSPR effect of Ag strengthens charge transportation while weakening the recombination rate, which is consistent with the photoluminescence and UV absorbance results.

FESEM images were used to examine the surface morphologies of produced Ag/Bi2WO6. The temperature of the hydrothermal synthesis was crucial in generating the crystalline and porous nanoflakes of the produced composite. The FESEM images of Ag/Bi2WO6 (1:2) are shown in Figure S3a,b. The micrograph shows several square nanoflakes with a length of about 200 nm. The nanoflakes self-assembled themselves in the form of circular colonies. Due to the relatively small sizes of silver nanoparticles, the Ag content is not visible in the images. The pores and the crystallinity of the nanoflakes were found to enhance the adsorption of the organic compound and the transfer of active species.[41,42] FESEM images of Ag/Bi2WO6 (1:1), Ag/Bi2WO6 (1:2), and Ag/Bi2WO6 (2:1) revealed no discernible differences. The inset spectra in Figure S3a revealed that the compound is made up of Bi, W, O, and Ag components, implying that Ag exists in Bi2WO6. More detailed insights into the morphology of the Ag/Bi2WO6 (1:1, 1:2, and 2:1) composite were investigated by TEM. The images of individual Bi2WO6 and the Ag/Bi2WO6 composite with different molar contents of Ag to Bi2WO6 NPs are shown in Figure a,b. Their selected area electron diffraction (SAED) pattern (Figure S3c) appears as bright concentric circles, which can be indexed to the (131), (200), (202), (331), and (262) planes of the Aurivillius-type layered structure Bi2WO6.[43] The TEM images of Bi2WO6 and Ag/Bi2WO6 with different contents of Ag loaded are similar in size. The darker region in the TEM images represents the area of Ag as it is the area of high electron density. The closer TEM image of Ag/Bi2WO6 (1:2) (Figure b) clearly shows the lattice spacing of 0.31 nm, which corresponds with the (113) lattice plane of Bi2WO6, and the lattice fringes of 0.25 nm match well with the (111) plane of Ag. This result further proves the successful preparation of the Ag/Bi2WO6 composite.[44]

Figure 2

(a) TEM image and (b) lattice spacing of Ag/Bi2WO6 (1:2).

The N2 adsorption–desorption isotherm and pore size distribution of the produced photocatalysts are shown in Figure S4a,b. The Barrett–Joyner–Halenda pore size distribution plot of synthesized Bi2WO6 and Ag/Bi2WO6 (1:1, 1:2, and 2:1) exhibits a limited range of pore size distribution with average pore diameters of 15.18, 8.70, 8.69, and 8.73 nm, respectively, showing the photocatalyst’s mesoporous characteristic. The specific surface area, pore width, and pore volume of Bi2WO6 and Ag/Bi2WO6 with various molar ratios of Ag to Bi2WO6 are shown in Table . There is no appreciable variation observed in the surface area of the prepared composites.

Table 2

Summary of BET Results of Pure Bi2WO6 and Ag/Bi2WO6 (1:1, 1:2, and 2:1)

catalyst	specific surface area (m2/g)	pore diameter (nm)	pore volume (cc/g)
Bi2WO6	26.40	15.18	0.12
Ag/Bi2WO6 (1:1)	26.02	8.70	0.11
Ag/Bi2WO6 (1:2)	20.84	8.69	0.07
Ag/Bi2WO6 (2:1)	20.43	8.73	0.12

X-ray photoelectron spectroscopy (XPS) has been performed on pure Bi2WO6 and Ag/Bi2WO6 (1:1, 1:2, and 2:1). The binding energy in the spectrum is calibrated using that of C 1s (284.62 eV). Figure S5 shows the overall XPS spectrum of the Bi2WO6 and Ag/Bi2WO6 (1:1, 1:2, and 2:1) heterostructures. No peak corresponding to Ag is detected in the overall XPS spectrum of Bi2WO6, whereas composites of Ag/Bi2WO6 show Ag peaks, indicating that the photodeposition method is successful for Ag deposition. The peaks centering in the region of 159.23 and 164.57 eV (Figure a) can be designated to the binding energies of Bi 4f7/2 and Bi 4f5/2 in Bi3+.[45] Also, the peaks centering in the region of 35.44–37.59 eV (Figure b) can be ascribed to W 4f5/2 and W 4f7/2 in the W6+ oxidation state. Ag/Bi2WO6 (1:2) shows the largest positive binding energy shift of 0.3 and 0.4 eV, depicting a higher oxidation state of W in the case of Ag/Bi2WO6 (1:2). This is due to the higher interaction of Bi2WO6 with Ag.[45,46] All the measured values are consistent with the previous reports.[47,48] The peaks centering at 373.74 and 367.72 eV (Figure c) can be ascribed to Ag 3d3/2 and Ag 3d5/2.[49] Considering the binding energies of Ag 3d3/2 and Ag 3d5/2, the valence of Ag in the heterostructure can be identified as +1.[50] The binding energy of O 1s (Figure d) lies at 530.20 eV, and there is a large negative shift observed in the case of the binding energy of O 1s in the Ag/Bi2WO6 (1:2) composite. The strong interaction of the composite with Ag creates an electric field; this weakens the bond, and O atoms in Bi2WO6 get replaced, creating oxygen vacancies.[45,46] The increased surface oxygen vacancy decreased the surface recombination centers and improved the charge separation efficiency, thus enhancing the photocatalytic activity.[51]

Figure 3

XPS spectra of (a) Bi 4f peaks, (b) W 4f peaks, (c) Ag 3d peaks, and (d) O 1s peaks.

Optical Properties

A UV–vis spectrometer is used to examine the optical characteristics of the prepared plasmonic photocatalysts. In Figure a, the absorbance spectra of Bi2WO6 and Ag/Bi2WO6 (1:1, 1:2, and 2:1) are shown. Due to the intrinsic bandgap transition, the pure Bi2WO6 sample exhibits strong photoresponse qualities from UV light to visible light shorter than 430 nm, as demonstrated in the absorption spectra. The LSPR’s synergistic impact with light absorption improves Ag/Bi2WO6 composites’ absorption throughout a larger visible light area.[52,53] The photoabsorption properties of the Ag-loaded Bi2WO6 (1:2) composites are improved in the visible light region. The enhancement in the absorption peak may be attributed to the SPR effect.[54−56] The optical bandgaps of pure Bi2WO6 and Ag/Bi2WO6 (1:1, 1:2, and 2:1) were estimated using the tau plot and found to be 3.06, 2.71, 2.41, and 2.85 eV, respectively (Figure S6). The optical bandgap of the Ag/Bi2WO6 (1:2) nanophotocatalyst is calculated to be 2.41 eV, which is less compared to those of Bi2WO6 and Ag/Bi2WO6 (1:1 and 2:1). This result is in accordance with the XPS result, which shows that higher oxygen vacancy had been created in the Ag/Bi2WO6 (1:2) catalyst.

Figure 4

(a) Absorbance spectra and (b) PL spectra of pure Bi2WO6 and Ag/Bi2WO6 (1:1, 1:2, and 2:1).

The photoluminescence (PL) emission spectra of Ag/Bi2WO6 can explain the electron trapping method by Ag NPs. As the Ag loading increases, the PL emission spectra in Figure b reveal a reduction in PL intensity. Because recombination of electron and hole pairs is a radiative process ascribed to PL emission, suppressing the recombination process diminishes the intensity of the PL spectra.[43,44] This shows that adding Ag NPs to Bi2WO6 can moderately limit the interaction of photogenerated holes and electrons.

SPR Effect

The real phenomenon of SPR on Ag/Bi2WO6 is rather complicated, but it can easily be seen in simulation. Our computation uses two models, a spherical Ag nanoparticle and hemisphere Ag nanoparticle, as shown in Figure a,b. Under normal irradiation of TM-polarized light (λ = 632.8 nm), the simulation result shows the creation of an electromagnetic field at the junction of Ag and Bi2WO6. Figure a indicates that spherical Ag on Bi2WO6 has a higher electromagnetic field enhancement than semisphere Ag on Bi2WO6 (Figure b). In the vicinity of plasmonic Ag, the interface in contact with the noble metal and semiconductor thus plays a significant role in the production of an electric field. As the size of the nanoparticle increases, the distance between the “effective dipole” and its substrate image dipole increases, thereby weakening the nanoparticle and substrate interaction.[57] The contact between Ag nanospheres and the Bi2WO6 substrate appears to be the most important component in determining the local increased electric field according to these simulation results.[58,59]

Figure 5

Electromagnetic field enhancement in spherical (model A) and semisphere (model B) Ag on Bi2WO6.

Dehydrogenation of Ammonia Borane

Hydrogen is a clean energy source with very high energy content (120 MJ/kg). It can serve as energy in vehicular applications in the near future.[60] The low volumetric density of hydrogen makes it difficult to store. To overcome this difficulty, various storage solutions have been developed and a large number of studies have been performed on hydrogen storage materials,[61,62] such as metal hydrides[61] and organic hydrides. Solid hydrogen storage materials have gained significant attention in recent years. Ammonia borane (NH3BH3, AB) has drawn much attention due to its low molecular weight (30.87 g/mol) and high hydrogen capacity (19.6 wt %). AB can release hydrogen by thermal dehydrogenation, but this process requires huge power consumption and high temperature. In contrast, AB is efficient in releasing hydrogen at room temperature via dehydrogenation reaction in the presence of the catalyst under visible light irradiation.[63,64] The hydrogenation reaction of AB proceeds according to the following reaction.

According to eq , 1 mol of AB can produce 3 mol of H2, which means that 12.8 mg (414 μmol) of AB could produce 1243.92 μmol of H2. The process was carried out in a photoreactor in visible light irradiation with continuous stirring.

The hydrogenation activity of pure Bi2WO6 and the Ag/Bi2WO6 composite was investigated in the dark and also under visible light, and enhanced H2 liberation is observed under visible light irradiation (λ ≈ 632 nm). In dark conditions (Figure a), the Ag/Bi2WO6 composite exhibited catalytic efficiency with a steady-going increase in H2 generation. The rate of the reaction exhibited by the prepared composite is very low. Pure Bi2WO6 exhibited a much lower reaction rate of 0.010 μmol/min, and Ag/Bi2WO6 (1:2) displayed 0.66 μmol/min H2 production in 150 min. It is thus displayed that the Ag NPs and the Bi2WO6 composite, specifically the LSPR of Ag and W5+ ions, display a synergistic effect, which enables the Ag/Bi2WO6 hybrid to be more effective in NH3BH3 hydrolysis under visible light irradiation.

Figure 6

Plots of evolved H2 gas as a function of reaction time from an aqueous NH3BH3 solution: (a) μmol of H2 evolved within 150 min in the dark condition, (b) μmol of H2 evolved within 150 min under visible light, (c) H2 yield % from NH3BH3, and (d) wavelength dependence of initial H2 yield rate enhancement upon LED light exposure.

Pure Bi2WO6 produced only 0.050 μmol/min H2 with a yield of 0.61%. The incorporation of Ag on Bi2WO6 facilitates enhanced visible light absorption. The composite Ag/Bi2WO6 (1:2) produced 6.608 μmol/min H2 (Figure b) with a yield of 79.6% in 150 min (Figure c). However, Ag/Bi2WO6 (1:1) and Ag/Bi2WO6 (2:1) produced 0.57 and 0.13 μmol/min H2 in 150 min, with corresponding yields of 6.9 and 1.5%, respectively. Pure Bi2WO6 and the Ag/Bi2WO6 (1:1 and 2:1) catalyst produced lower H2. The high quantity of Ag loading in these composites might have reduced the active sites for reaction and was not appropriate to create an acceptable LSPR effect. Pure Bi2WO6 fails to show any LSPR effect due to the absence of Ag content, and thus, the H2 production, in this case, is very low.

The wavelength dependence for H2 production enhancement was also investigated by using monochromatic LEDs with wavelengths of 470, 530, and 650 nm. It was noted that the largest enhancement in H2 production was observed in the red LED with a wavelength of 650 nm (Figure d). This wavelength dependence clearly elucidates that the photoactivity enhancement is due to the plasmonic effect.[65] Upon excitation by visible light irradiation, a Bi2WO6 support would give rise to hot energetic electrons. Subsequently, these hot electrons will be injected into the adjacent Ag NPs, allowing fast interfacial electron transfer. In this way, the surface Ag NPs are negatively charged and act as an electron trapper.

To understand the mechanism of photocatalytic activity of the prepared plasmonic composite, a scavenger test has been performed. K2S2O8 as an electron scavenger, NaHCO3 as a hole scavenger, and 2-propanol as a hydroxyl radical scavenger are added to the best performing photocatalyst (Ag/Bi2WO6 (1:2)) for the photocatalytic hydrolysis of AB under visible light irradiation in the same condition. Figure a shows that the H2 yield drastically decreased from 6.608 to 0.053 μmol/min in the presence of K2S2O8. When Ag/Bi2WO6 was illuminated by visible light irradiation, it produced an electron–hole pair; K2S2O8, being a negatively charged scavenger, reacts easily with the electrons, thus resulting in activity reduction under visible light irradiation. However, hydrogenation activity decreased marginally after adding NaHCO3 and 2-propanol (3.08 and 2.9 μmol/min, respectively). The scavenger tests signify that a large part of LSPR-induced hot electrons participate in the hydrogenation activity. They get excited because of the LSPR effect under visible light irradiation and generate charge pairs. Here, the negative charge acts as a main active species that directly reacts with AB to dissociate the B–N bond to form NH3, which hydrolyzes to produce H2 and generates an NH4+ ion.

Figure 7

(a) Comparison of H2 production activity from NH3BH3 solution within 150 min with or without NaHCO3, K2S2O8, and 2-propanol scavengers over plasmonic Ag/Bi2WO6 (1:2). (b) Five recycling experiments for NH3BH3 hydrolysis under visible light irradiation within 150 min.

The catalytic stability of the plasmonic Ag/Bi2WO6 (1:2) composite was investigated by recovering the photocatalyst after the reaction. The recovered photocatalyst was tested under the same condition for another five cycles. It showed considerable activity during five repeated cycles (Figure b). Furthermore, the XRD spectra retained the original structural property (Figure S7) even after multiple recycling experiments, which mean the efficient and stable property of plasmonic Ag/Bi2WO6 with high potential application prospect.

Mechanism of Photocatalytic Hydrogen Evolution

To understand the mechanism of photocatalysis with the Ag/Bi2WO6 heterostructure, it is desirable to interpret the synergistic effect between the constituent materials of plasmonic metal/semiconductor nanostructures, which can prove to be highly efficient to design a photocatalytic system for efficient solar energy conversion. To identify the behavior of photogenerated electrons and holes in the hybrids, we have made a band diagram of the Ag/Bi2WO6 hybrid based on the bandgaps and CB edge potentials of Bi2WO6 as well as the Fermi energy (EF) of Ag (−0.5 V vs NHE at pH = 0) (Figure ).[66] The CB edge potential of the Bi2WO6 NPs was estimated to be −1.27 V vs NHE at pH = 0 from their Mott–Schottky plot (Figure S8). Based on the band structure of the Ag/Bi2WO6 heterostructure, together with the results in the photocatalysis experiments, possible transfer routes for the photogenerated charge carriers could be proposed, as shown in Figure .

Figure 8

Band structure and electron hole process in photocatalytic hydrogen evolution for Ag/Bi2WO6 (1:2).

Conclusions

In this paper, the plasmonic nanostructures of pure Bi2WO6 and Ag/Bi2WO6 (1:1, 1:2, and 2:1) were successfully prepared by the hydrothermal and photodeposition technique. The FESEM and TEM studies show the presence of Ag nanoparticles in orthorhombic Bi2WO6 nanosheets. The plasmonic nanostructured Ag/Bi2WO6 (1:2) exhibited a localized surface plasmonic effect and presented very excellent photocatalytic activity toward dehydrogenation of NH3BH3 (AB), producing 6.608 μmol/min H2 with a yield of 79.6%. The combination of the noble metal with the photocatalyst introduced the plasmonic effect, which highly enhanced the photocatalytic activity of the photocatalyst. This study provides a promising strategy in exploring stable and efficient plasmonic semiconductor photocatalysts for solving hydrogen evolution problems as new energy resources and energy carriers.