Slow Photon-Enhanced Heterojunction Accelerates Photocatalytic Hydrogen Evolution Reaction to Unprecedented Rates

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE11 Apr 2022Slow Photon-Enhanced Heterojunction Accelerates Photocatalytic Hydrogen Evolution Reaction to Unprecedented Rates Jing Liu†, Yin-Hao Guo†, Zhi-Yi Hu†, Heng Zhao, Ze-Chuan Yu, Lihua Chen, Yu Li, Gustaaf Van Tendeloo and Bao-Lian Su Jing Liu† State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, 430070 Wuhan †J. Liu, Y.-H. Guo, and Z.-Y. Hu contributed equally to this work.Google Scholar More articles by this author , Yin-Hao Guo† State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, 430070 Wuhan †J. Liu, Y.-H. Guo, and Z.-Y. Hu contributed equally to this work.Google Scholar More articles by this author , Zhi-Yi Hu† State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, 430070 Wuhan Nanostructure Research Centre (NRC), Wuhan University of Technology, 430070 Wuhan †J. Liu, Y.-H. Guo, and Z.-Y. Hu contributed equally to this work.Google Scholar More articles by this author , Heng Zhao State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, 430070 Wuhan Google Scholar More articles by this author , Ze-Chuan Yu School of Civil Engineering and Architecture, Wuhan University of Technology, 430070 Wuhan Google Scholar More articles by this author , Lihua Chen State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, 430070 Wuhan Google Scholar More articles by this author , Yu Li *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, 430070 Wuhan Google Scholar More articles by this author , Gustaaf Van Tendeloo Nanostructure Research Centre (NRC), Wuhan University of Technology, 430070 Wuhan EMAT, University of Antwerp, B-2020 Antwerp Google Scholar More articles by this author and Bao-Lian Su *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, 430070 Wuhan Laboratory of Inorganic Materials Chemistry (CMI), University of Namur, B-5000 Namur Google Scholar More articles by this author https://doi.org/10.31635/ccschem.022.202101699 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail In photocatalysis, both the photogenerated charge separation and transport and the induced light utilization greatly influence performance. In this work, highly ordered [email protected] core-shell inverse opal ([email protected]) nanocomposites have been successfully designed as a model to couple the heterojunction system with the slow photon effect for photocatalytic H2 production. Theoretical calculations and experimentation provide direct evidence for the slow photon effect in the [email protected] nanocomposites. The type II heterojunction is responsible for promoting the migration and separation of photogenerated charges, and the slow photon effect is in charge of enhancing light harvesting in the [email protected] nanocomposites. This synergy of two functions gives rise to a significantly enhanced photocatalytic H2 production rate under simulated solar light for the [email protected] nanocomposites. The highest H2 production rate reaches 48.7 mmol g−1 h−1 under simulated solar light with the benchmark performance for all reported [email protected] composites. Our work provides proof-of-principle that coupling the heterojunction system with the slow photon effect can greatly enhance the photocatalytic activity of composite photocatalysts. Download figure Download PowerPoint Introduction Solar light utilization has been extended to limit global warming and decrease carbon emissions. Consequently, photocatalytic hydrogen generation enabled by semiconductors under light illumination has attracted increasing attention. In fact, this strategy has been widely explored as an environmentally friendly method for converting solar energy to stored chemical energy.1–3 Generally, the photocatalytic reaction involves the following steps: light harvesting, charge separation, transportation of photogenerated electrons and holes to the surface of the photocatalyst, and the redox reaction.4 Because the balance of the thermodynamics and kinetics of the photocatalytic reaction processes play the key role in the photocatalytic hydrogen generation efficiency, the single-component photocatalyst, owing to its low absorption of solar light and photocatalytic inefficiency, has hindered its commercial application. To overcome this obstacle, heterojunction photocatalytic water-splitting systems (including type I, type II, and type III heterojunctions, Z-scheme heterojunction, and S-scheme heterojunction systems) have been adopted because of their ability to accelerate the migration of photogenerated electrons and holes to various counterparts for reduction and oxidation reactions to prevent their recombination rate, and to prolong their lifetimes.4–16 Such heterojunction photocatalysts exhibit broadband light response and high hydrogen evolution rates.4,17–19 This internal charge-transfer process leads to a spatial separation of photogenerated electrons and holes, resulting in efficient electron transfer for H2 reduction when the conduction bands (CBs) of the composite photocatalyst couple favorably.2,13 Furthermore, the heterojunction system yields photogenerated carriers that benefit photocatalysis in terms of thermodynamic requirements.4,20,21 For the heterojunction system, the structural design is very important to make full use of the induced light. Hierarchically structural porous materials with high surface area, large pore volume, and interconnected hierarchical porosity of different length scales display significant structural advantages in electron and ion transport, and mass diffusion has roused widespread attention and applications.22–26 As a special hierarchically porous structure, three-dimensional ordered marcroporous (3DOM) structures as good light harvesters have given rise to considerable research activity. They allow deep penetration of light and lead to longer light pathways. They significantly improve induced light utilization efficiency. Previous work on single (TiO2, ZnO, BiVO4, etc.), binary (TiO2/CdS, TiO2/Au, TiO2/ BiVO4, etc.), and ternary (TiO2/BiVO4/ZnO, TiO2/Au/CdS, etc.) semiconductor-based composites with 3DOM structure show that the 3DOM structure can greatly improve the absorption and utilization of light and significantly enhance photocatalytic performance.17,22,27–32 Particularly, the inverse opal photonic crystal structures benefitting the slow photon effect have further exhibited highly enhanced photocatalytic activity.18,29,33–35 The slow photon effect generally occurs when the wavelength of blue and/or red edges of the photonic band gap (PBG) match the electronic band gap (EBG) of a semiconductor. When this happens, light propagates with extremely low group velocity, leading to a longer lifetime of the photons and a largely increased optical path length of the light wave in the material to promote the interaction between light and material.36–38 This greatly improves the photon absorption and conversion efficiency, as broadly evidenced and applied in various kinds of single semiconductors for enhanced photocatalytic activity.17,39 Since the heterojunction system plays a significant role in separating photogenerated electron-hole pairs and slow photon effect contributes a lot to enhancing light harvesting, binary/ternary inverse opal composites seldom provide evidence or theoretical calculations that reveal the slow photon effect in the inverse opal composites. Thus, it is worth systematically investigating and revealing the existence of the slow photon effect in inverse opal composites and demonstrating the advantages of coupling the slow photon effect with the heterojunction system in such an efficient solar-powered photocatalysis. Herein, we choose the very common semiconductors (ZnO, CdS) and report the construction of three-dimensional (3D) [email protected] core-shell inverse opal ([email protected]) nanocomposites as a photocatalyst model by coating CdS nanoparticles onto a ZnO inverse opal (ZnO-IO) open framework using the successive ion layer absorption and reaction (SILAR) method.19,29 To tune and monitor the slow photon effect, we tailor the macroporous sizes of the inverse opal nanocomposite to modulate the relative positions of PBG and EBG in the [email protected] composites, aiming to compare photocatalytic enhancement originating from the slow photons between the red edge and the blue edge.33,36 These results show that by coupling a heterojunction system with the slow photon effect, [email protected] films can enhance light absorption and facilitate the separation of photogenerated electrons and holes, resulting in significantly improved photocatalytic H2 production. This is the first evidence that the heterojunction system can synergistically be coupled with the slow photon effect in such a [email protected] photonic crystal heteronanostructure. The [email protected] nanocomposites demonstrate an unprecedented photocatalytic H2 production rate, proven to originate from the type II heterojunction system with ∼2.7× enhancement under simulated solar light. The slow photon effect enhances photocatalytic H2 production with ∼1.4× and 1.7× at the red and blue edges respectively, providing direct experimental evidence that the slow photon effect occurring at the blue edge can bring a higher photocatalytic enhancement than that in red edge.17,39,40 The highest H2 production rate for [email protected] reached 48.7 mmol g−1 h−1 under simulated solar light with the benchmark performance of ∼fivefold enhancement of [email protected] (10.5 mmol g−1 h−1). Experimental Methods Fabrication of ZnO inverse opal film Highly ordered and continuous ZnO-IO films coated on fluorine-doped tin oxide (FTO) substrate with different air sphere sizes were successfully prepared by a metal salt-based sol–gel infiltration according to our previous work.29 Three different-sized colloidal templates ranging from 230, 290, and 440 nm have been used in this work. After removing the templates by calcination, various ZnO-IO films with different pore sizes were obtained. The samples were then designated as ZnO-IO-230, ZnO-IO-290, and ZnO-IO-440, respectively. The synthesis process of ZnO-IO-3M was the same with ZnO-IO film except using the mixture of three colloidal spheres (230, 290, and 440 nm with the volume ratio of 1∶1∶1) as template. Fabrication of 3D [email protected] nanocomposites To obtain 3D [email protected], the CdS shell layer was coated on the ZnO-IO film by the SILAR technique, as in previous reports.41 Typically, the ZnO-IO film is successively immersed in solutions containing 0.5 M Cd(NO3)2 in ethanol for 5 min, to allow Cd2+ to adsorb onto the ZnO-IO film and is then rinsed with ethanol for 5 min to remove the excess Cd2+. The film was rinsed in a solution containing 0.5 M Na2S in methanol for 5 min, to allow S2− to adsorb onto the ZnO-IO film and to react with preadsorbed Cd2+ to form the desired CdS. The above procedure is termed a SILAR cycle, and the loading of CdS can be increased by repeating these assembly cycles. This immersion cycle was repeated eight times in this study. The final samples were designated as [email protected], [email protected], and [email protected], respectively. The [email protected] was synthesized via the same method to coat the same amount of CdS shell on the ZnO-IO-3M film. Characterizations The crystalline phase of the samples was examined by X-ray diffraction (XRD) with a Bruker D8 ADVANCE diffractometer (Karlsruhe, Germany) using Cu Kα radiation (λ = 1.54 Å) in the 2θ range of 20–70°. Field-emission scanning electron microscopy (FESEM) was performed on a Hitachi S-4800 electron microscope (Tokyo, Japan) equipped with an energy-dispersive spectroscopy facility to characterize the sample morphology and composition. The absorption spectra were collected with a UV2550 (SHIMADZU, Chukyo, Japan) UV–vis spectrometer. The reflection spectra were recorded using an Avaspec 2048/2 fiber-optic spectrometer (Beijing, China). Transmission electron microscopy (TEM) was carried out on an FEI Tecnai Osiris microscope (Massachusetts, United States), operated at 200 kV. The photocatalytic H2 production was performed with an online photocatalytic analysis system (Labsolar-6A, Beijing Perfectlight Technology Co., Ltd., Beijing, China). The gas products were analyzed periodically by an Agilent 7890A (California, United States) gas chromatograph (GC) with a thermal conductivity detector (TCD). Photocatalytic hydrogen reduction Typically, the photocatalytic H2 production was performed in a Pyrex reactor with an entry window of optical quartz glass (Labsolar-6A, Beijing Perfectlight Technology Co., Ltd., Beijing, China). One piece of the obtained [email protected] film coated on FTO substrate (geometrical area of each piece was 2 cm × 2 cm) was placed in the reaction cell with 100 mL of an aqueous solution containing 0.1 M Na2S and 0.1 M Na2SO3 as the sacrificial agents. The light source was irradiated under simulated solar light irradiation (250–780 nm, PLS-SXE-300DUV (Beijing, China) with an UV light intensity of 34 mW cm−2 and a visible light intensity of 158 mW cm−2, Beijing Perfectlight Technology Co., Ltd., Beijing, China). When under visible light irradiation, the light passed through a cutoff filter (λ > 420 nm). The amount of H2 produced was analyzed using on-line GC (Agilent 7890A, California, United States) with TCD, using N2 as the carrier gas. The mass of photocatalyst on one piece of FTO was ∼1 mg. Photoelectrochemical measurements The photocurrent measurements in our experiments were performed on an electrochemical workstation (CHI 660D, Chenhua Instrument Co., Shanghai, China) in a standard three-electrode electrochemical cell with a Pt plate counter electrode and an Ag/AgCl reference electrode. And the electrolyte solution was performed using 0.1 M Na2S and Na2SO3, which was the same as with the sacrificial agents. The working electrodes with the exposed area of 1 × 1 cm2 were illuminated from the front side with solar light sources, the same as with the photocatalytic H2 production. All the [email protected] photoanodes were measured at 0.5 V external potential versus hydrogen electrode (RHE). Apparent quantum yields (AQY) are defined by the following equation because of a 2-electron process for H2 production in the [email protected] with the slow photon effect and type II heterojunction system. The energy density of irradiated solar light with 365 and 500 nm are 19 and 16 mW cm−2. AQY ( % ) = [ Number of the reacted electrons ] / [ Number of incident photons ] × 100 = [ Number of evolved hydrogen molecules × 2 ] / [ Number of incident photons ] × 100 Theoretical simulation of reflectance spectra The theoretical simulation of the reflectance spectra of [email protected] and [email protected] inverse opal films was calculated using the finite-difference time-domain (FDTD) method in the MEEP package.42 Computational details The density functional theory (DFT) calculation was performed by the Cambridge Sequential Total Energy Package. Generalized gradient approximation of the Perdew–Burke–Ernzerhof functional was used as the commutative correlation functional. The cutoff energy was set to 450 eV, and a 30 Å vacuum layer was constructed to eliminate the interaction between the periodic structures of the surface model. The Brillouin area was sampled at 2 × 2 × 1 K points. Results and Discussion Figure 1a illustrates the synthesis process for [email protected] nanocomposites. Figures 1b–1d present the FESEM images of the as-synthesized [email protected] films with different macroporous diameters. The hexagonal close-packed structure is well copied from the hexagonally packed colloidal crystal templates ( Supporting Information Figure S1). The obtained [email protected] films are highly ordered with continuous macroporous structure on a large scale ( Supporting Information Figure S2), similar to those of the ZnO-IO films ( Supporting Information Figure S3). The air-spheres of the ZnO-IO films are distributed uniformly, and the average pore diameter is ∼220, 260, and 380 nm, respectively ( Supporting Information Figure S3). After coating CdS nanoparticles on ZnO-IO structures, the average pore diameter for [email protected], [email protected], and [email protected] was ∼200, 240, and 360 nm. The thickness of the CdS shell was estimated at ∼10 nm via the SILAR method (Figures 1b–1d). SEM-electron dispersive X-ray spectroscopy (EDXS) results further show that the molar ratio of ZnO and CdS in all nanocomposites was close to 1∶1. Supporting Information Figure S4 presents the mapping images of [email protected] ( Supporting Information). These results demonstrate that we have successfully prepared well-ordered, large-area, high-structural quality [email protected] nanocomposites. These impressive core-shell inverse opal films would have an interesting influence on the photocatalytic H2 production based on our previous work for ZnO-IO film.29 Figure 1 | Schematic illustration and structural characterizations of [email protected] films. (a) Schematic illustration of the preparation of [email protected] films. The FESEM images of the as-prepared (b) [email protected] csIO-230, (c) [email protected] csIO-290, and (d) [email protected] csIO-440 heteronanostructures. Download figure Download PowerPoint A hexagonal wurtzite structure (JCPDS No. 079-0206) of the [email protected] films with different macroporous diameters and three strong diagnostic diffraction peaks at (10-10), (0002) and (10-11), respectively were observed based on the XRD patterns ( Supporting Information Figure S5). The diffraction peaks located at 37.76° and 51.75° belong to the FTO substrate. Three characteristic diffraction peaks at 26.46°, 43.89°, and 54.47° were ascribed to the (111), (220), and (222) crystal planes of cubic CdS (JCPDS No. 089-440), respectively. The broader diffraction peaks of cubic CdS implied that very small CdS nanoparticles were decorating the surface of ZnO-IO. Furthermore, the XRD results suggest that the CdS coating did not influence the crystalline phase of ZnO-IO film, indicating the successful synthesis of a heterojunction inverse opal structure. Since the only difference for these [email protected] films was the macroporous diameter, [email protected] was selected to demonstrate the morphological quality of the [email protected] structures. Figure 2a presents a typical TEM image of [email protected] structure, showing the stable core-shell inverse opal structure. Figure 2b presents the high-resolution transmission electron microscopy (HRTEM) image from the enlarged squared area in Figure 2a. The fast Fourier transform (FFT) from the whole area, clearly displayed the main zone axis of ZnO along [2-1-10] and verified the quasisingle crystal domains in the ZnO-IO framework, suggesting that the ZnO nanoparticles formed inverse opal structure with the same orientation (Figure 2b). These quasisingle crystal domains in this ZnO photonic crystal skeleton could improve the separation and the transport of photogenerated electrons and holes43 although details of the formation of such quasisingle crystal domains in inverse opal structure are not clear. However, the CdS (111) ring revealed that the CdS crystals coated on the surface of the framework had different orientations. Figure 2 | Compositional and structural characterizations of [email protected] film. (a) TEM images of low magnification. (b) HRTEM image of the area indicated in (a) and FFT image (inset). (c) Enlarged image of the area indicated in (b) and FFT image (inset). (d) HAADF-STEM image of one typical cell of inverse-opal structure and (e–h) the corresponding STEM-EDXS color maps in red box (d). Download figure Download PowerPoint Figure 2b also displays the ∼10 nm thickness of CdS located on the surface of ZnO, consistent with the SEM observations. The enlarged image of a zone on the surface presents the CdS crystal along the [10-1] zone axis (Figure 2c), further revealing the CdS polycrystalline coat on the ZnO-IO framework. Figures 2d–2h present the CdS distribution on the ZnO-IO framework via high-angle annular dark field-scanning transmission electron microscopy (HAADF-STEM) and STEM- STEM-EDXS images on the randomly selected regions of [email protected] The uniform distribution of Zn, Cd, and S throughout the [email protected] was clearly observed, showing the homogeneous distribution of CdS nanocrystallites on the ZnO-IO structure and intimate contact interface between the CdS and ZnO phases. These results confirm that the CdS nanocrystals were successfully coated onto the ZnO-IO framework to form a core-shell inverse opal structure. And the intimate contact interface is also expected to bring efficient electron transfer and high separation effect. It is well documented that reflectance spectra reflect the morphological quality of the inverse opal photonic crystal structure and the judgment criteria for the occurrence of slow photon effects.17,18,36 The theoretical simulation of the reflectance spectra of [email protected] inverse opal films with 230 and 290 nm templates has been performed to feature the experimental spectral observations (Figures 3a–3d). The FDTD method was implemented by using the MEEP package to calculate the optical response of the ZnO-CdS inverse opal system.42,44 Reflectance spectra with wavelengths ranging from 250 nm to 1000 nm were obtained. In the FDTD simulation, and the entire system was periodic and consisted of five parallel layers stacked along the z axis. The five layers included two perfectly matched layers (PMLs, with thickness equal to the maximum wavelength of the spectra) at the bottom and the top, two vacuum layers (with refractive index equal to 1.0 and thickness twice the maximum wavelength) and one material layer (of which the thickness was around 7.53 times the hole diameter, i.e., 220 and 260 nm) in the middle. The material layer resembled the inverse opal system used in the experiment, which was constructed following the face-centered cubic pattern with its <111> surface perpendicular to the z axis. Materials were modeled by directly defining the refractive index; that is, ZnO was modeled by a dielectric medium with refractive index equal to 2.0 and CdS of 2.5. An electromagnetic wave source was placed in the upper vacuum layer, and two flux recorders were placed at the boundary between the vacuum layer and the PML on both sides. Reflectance was calculated by dividing the upper flux over the sum of the two fluxes. For [email protected] and [email protected], the calculated reflectance peaks were 458 and 542 nm. In both cases, the reflectance peak shifted to longer wavelengths as the macropore size was increased, as expected from the photonic crystal structure of the films. Interestingly, the shoulders were located around the electronic band edge of CdS (see extinction coefficient k spectrum, blue dashed line, in Figure 3d), which provided sufficient evidence of the existence of the slow photon effect. It is then expected that the occurrence of the slow photon effect on the red edge of [email protected] and on the blue edge of [email protected] ZnO-csIO-290 will bring high enhancement in light absorption efficiency and photocatalytic H2 generation. Figure 3 | Theoretical simulation of reflectance spectra. (a–c) Structure of the material layer in the FDTD simulation shown in three side views (Cyan regions indicate medium with refractive index higher than 1 and gray regions are vacuum, where the refractive index is equal to 1). (d) Theoretical simulation of reflectance spectra of [email protected] and [email protected] inverse opal films using FDTD method, and the blue dashed line is the extinction coefficient k spectrum of CdS. (e and f) UV–vis absorption spectra (e) and reflectance spectra (f) of [email protected] with different macroporous diameters. The yellow shaded region indicates the electronic absorption band of CdS. Download figure Download PowerPoint UV–vis spectrophotometry and the reflectance spectra of these [email protected] films were performed in order to study the relative positional relationship between the reflection peak and the semiconductor absorption of the obtained [email protected] composites from an experimental point of view (Figures 3e and 3f). UV–vis spectrophotometry was performed to study the electronic structures and the optical absorption properties of these [email protected] films (Figure 3e). The strong absorption at wavelength in the 350–400 nm range matched the intrinsic interband transition absorption of hexagonal ZnO, showing its UV absorption character. After coating the CdS crystal shell on the surface of the ZnO macroporous framework, the absorption edge was extended to ∼517 nm, representing the characteristics of CdS. This indicated that both UV light and visible light were utilized by the [email protected] films. This is very helpful for the effective utilization of solar light for hydrogen production. It also shows that the larger the size of the macropore, the higher the light absorption for these samples. This is reasonable because the larger macropore enhances light scattering and extends the optical path length of the light inside the inverse opal structure, thereby emphasizing its optical absorption strength.23 For a high quality photonic crystal structure, the PBG can be tuned by changing the macropore size and/or modifying the average refractive index. This is directly revealed by the reflectance spectra of the [email protected] and ZnO-IO films. Before the CdS coating, the PBG of ZnO-IO-230 located at 445 nm ( Supporting Information Figure S6). After the CdS shell deposition, the position of reflection peaks of [email protected] located at 484 nm (Figure 3f), showing a redshift of ∼40 nm due to the high average refractive index of the composites (the refractive index of ZnO = 2.0 and CdS = ∼2.4). This led to the PBG red edge of [email protected] overlapping the electronic absorbance band of CdS (∼517 nm). When the macropore size was increased to 290 nm, the corresponding reflection peak redshifted from 492 nm ( Supporting Information Figure S6) to 539 nm. The PBG blue edge side of [email protected] then overlapped with the electronic absorbance band of CdS (Figure 3f). Both the red and blue edges of the PBG facilitated the occurrence of the slow photon effect.17,40 The PBG of [email protected] was further redshifted from 747 nm ( Supporting Information Figure S6) to 830 nm while its second order PBG was at 510 nm, almost overlapping the electronic excitation wavelength of CdS (Figure 3f). Namely, the stop band reflection suppressed the light absorption to decrease the photocatalytic activity of [email protected] because of the nonoccurrence of the slow photon effect.36 Moreover, the measured reflectance peak of [email protected] and [email protected] were in good correspondence with calculated spectra (the measured reflectance peaks were 484 and 539 nm, and the calculated reflectance peaks were 458 and 542 nm, respectively) (Figure 3d). Interestingly, the shoulders that were observed experimentally on the red edge of [email protected] and the blue edge of [email protected] reflectance spectrum were also found in the calculated spectrum. Both experimentation and theoretical calculations provide direct evidence of the existence of the slow photon effects on the red edge of [email protected] and the blue edge of [email protected], which can bring high enhancement for photocatalytic H2 generation. Photocatalytic H2 generation was initially performed to show the slow photo