Electrochemical Water Splitting

Communication Single-Crystalline Nanostructure

www.advmat.de

2D Porous TiO2 Single-Crystalline Nanostructure Demonstrating High Photo-Electrochemical Water Splitting Performance Teera Butburee, Yang Bai, Huanjun Wang, Hongjun Chen, Zhiliang Wang, Gang Liu,* Jin Zou, Pongtanawat Khemthong, Gao Qing Max Lu, and Lianzhou Wang* large accessible surfaces as well as longrange charge diffusion length and structural coherence are highly desired,[10–14] but to meet all criteria in one single nanostructure remains very challenging. The recently developed TiO2 porous single-crystals[15] showed great promise in achieving the desirable properties. However, it is difficult to make these porous single-crystals into thin film electrodes, as the large number of associated grain boundaries significantly reduces charge mobility.[14] An alternative strategy for the synthesis of efficient semiconductor electrodes is to directly grow the porous single-crystalline nanostructures (PSNs) on conductive substrates.[3] Although such configuration is easy to design, the actual synthesis has been a long-standing bottleneck. Previous research has successfully demonstrated the fabrication of nanostructures for photo-electrochemical water splitting[5] and mesoscopic solar cells,[16] such as 1D nanorods[3,17] and nanotubes,[18] 2D nanowalls,[19,20] and 3D host–guest structure,[21] which are vertically aligned on the conductive substrates and show fast charge transport along their longitudinal directions.[17] However, the limited acccessible surface area and/or polycrystalline

Porous single crystals are promising candidates for solar fuel production owing to their long range charge diffusion length, structural coherence, and sufficient reactive sites. Here, a simple template-free method of growing a selectively branched, 2D anatase TiO2 porous single crystalline nanostructure (PSN) on fluorine-doped tin oxide substrate is demonstrated. An innovative ion exchange–induced pore-forming process is designed to successfully create high porosity in the single-crystalline nanostructure with retention of excellent charge mobility and no detriment to crystal structure. PSN TiO2 film delivers a photocurrent of 1.02 mA cm−2 at a very low potential of 0.4 V versus reversible hydrogen electrode (RHE) for photo-electrochemical water splitting, closing to the theoretical value of TiO2 (1.12 mA cm−2). Moreover, the current–potential curve featuring a small potential window from 0.1 to 0.4 V versus RHE under one-sun illumination has a near-ideal shape predicted by the Gartner Model, revealing that the charge separation and surface reaction on the PSN TiO2 photoanode are very efficient. The photo-electrochemical water splitting performance of the films indicates that the ion exchange–assisted synthesis strategy is effective in creating large surface area and single-crystalline porous photoelectrodes for efficient solar energy conversion. Nanostructured semiconductors have been widely explored for application in optoelectronics, photocatalysis, energy storage, and sensing, with extensive studies carried out on their design, fabrication, and characterization.[1–9] For many applications, Dr. T. Butburee, Dr. Y. Bai, H. Wang,[+] Dr. H. Chen, Dr. Z. Wang, Prof. L. Wang Nanomaterials Centre Australian Institute for Bioengineering and Nanotechnology and School of Chemical Engineering The University of Queensland St Lucia, QLD 4072, Australia E-mail: [email protected] Dr. T. Butburee, Dr. P. Khemthong National Nanotechnology Center (NANOTEC) National Science and Technology Development Agency (NSTDA) Klong Luang, Pathumthani 12120, Thailand [+]Present address: Beijing Key Laboratory of Green Reaction Engineering and Technology, Department of Chemical Engineering, Tsinghua University, Beijing 100084, China

DOI: 10.1002/adma.201705666

Adv. Mater. 2018, 1705666

Prof. G. Liu Shenyang National Laboratory for Materials Science Institute of Metal Research Chinese Academy of Sciences Shenyang 110016, China E-mail: [email protected] Prof. G. Liu School of Materials Science and Engineering University of Science and Technology of China 72 Wenhua Road, Shenyang 110016, China Prof. J. Zou Materials Engineering and Centre for Microscopy and Microanalysis The University of Queensland St. Lucia, QLD 4072, Australia Prof. G. Q. Max Lu The University of Surrey Guildford, Surrey GU2 7XH, UK

1705666 (1 of 8)

© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com

www.advmat.de

Figure 1. Morphology characterization of potassium titanate nanostructure. SEM images of various 2D potassium titanate nanostructures: A) 2D, B) E-2D, and C) F-2D, respectively. D) Low magnification top-view SEM image of the typical morphology of E-2D nanostructures. E) Cross-sectional SEM image of the E-2D film on FTO substrate.

nature of such nanostructures restrain their performance in many applications. Introducing porosity into single-crystalline nanostructures is proposed as an effective way to enlarge the accessible surface area without compromising the excellent charge transport properties of single crystals. However, conventional techniques for creating pores, such as the incorporation of sacrificing templates, are not appropriate for the fabrication of PSNs. This is because the brittle nature of many semiconductor nanostructures makes it difficult to introduce any hard templates to their structures, and soft templates such as block copolymers[22] usually lead to poor crystallinity and impurities from the polymer residues. Here, we demonstrate an innovative and versatile synthetic strategy for the direct growth of selectively branched, 2D PSNs of anatase TiO2 on fluorine-doped tin oxide (FTO)-coated glass substrates. The selectively branched, layered potassium titanate (K-titanate) nanostructures were first grown from potassium titanium oxalate (PTO) on FTO substrate via a one-pot hydrothermal synthesis. By soaking the substrate in diluted ethanolic hydrochloric acid (HCl), the 2D K-titanate was easily converted into protonated titanate (H-titanate) via an ion-exchange process in which K+ ions were replaced by H+ species. After annealing, the dehydration and condensation of the layered titanate aroused structural reconstruction, leading to the formation of anatase single-crystalline nanostructures with a large amount of pores. To our knowledge, this is the first demonstration that a simple ion-exchange process can induce well-defined porosity formation in the fabrication of single-crystalline nanostructures. Moreover, additional branches grown out from the plate-like backbone structures can be controlled in a selective manner. Our detailed characterizations reveal that the TiO2 PSN photoanodes not only deliver excellent charge separation and transport efficiency, but also provide highly accessible

Adv. Mater. 2018, 1705666

surface area, leading to high photocurrent of photo-electrochemical (PEC) water splitting with a very low onset potential. The findings herein provide an excellent platform for fundamental studies of the electronic process in porous semiconductors. In addition, the high exposed surface area in these films could open up new possibilities for controlled hybridization and surface reaction. To fabricate 2D porous single-crystalline TiO2 nanostructures, K-titanate was first grown on FTO substrate via a one-pot hydrothermal treatment of PTO in the presence of diethylene glycol (DEG) and water, followed by an ion exchange–induced pore-forming process and calcination (refer to the Experimental Section for more details). Figure 1A shows a typical scanning electron microscopy (SEM) image of the 2D platelike nanostructures (2D hereafter) grown on FTO substrate with minimal water (3 mL, 7.5 vol%). By slightly increasing water content and reaction time, additional branches can be controlled to selectively grow solely on the edges of the backbone plates (E-2D hereafter), as shown in Figure 1B. The surface regions including the sharp surface edges, tips, etc., are the preferred nucleation sites[23] for those additional branches due to the relatively large surface free energy and increased molecular mobility in these regions. The appearance of some cracks in the plates might be caused by the stress generated during the formation of branches.[24–27] Figure S1 (Supporting Information) shows the structural evolution of these 2D nanostructures and their growth mechanism. Further increasing the water content (3.25–5.5 mL, 8.125–13.75 vol%) accelerates the decomposition of oxalate ligand, facilitating the growth of branches that fully cover the entire plate (F-2D hereafter), as indicated in Figure 1C. The low magnification top-view SEM images of E-2D (Figure 1D) shows a uniform and selectively branched nanowall morphology; top-view SEM images of 2D

1705666 (2 of 8)



www.advmat.de

and F-2D can be found in Figure S2 (Supporting Information). In Figure 1E, a crosssectional SEM image of E-2D shows that the vertically grown nanostructures are wellaligned against the substrate. Since the film thickness plays a vital role in determining the optoelectronic property of the semiconductor nanostructures, we demonstrated that the thickness of 2D nanostructured film can be facilely tuned by adjusting the precursor ratio, as shown in Figure S3 (Supporting Information). The detailed parameters are summarized in Figure S4 (Supporting Information). The optimal film thickness is determined by light absorption and charge mobility. The penetration depth (α−1) of UV light on anatase TiO2 has been experimentally studied and well identified as about 2–3 µm.[18,28,29] To ensure >90% absorption of the incident light, the film thickness is normally required to be >2.3 times the value of α−1.[5] Benefiting from the nanostructured rough surface that can significantly enhance light absorption as well as the single crystallinity that enables efficient charge carrier kinetics, the TiO2 PSNs with an optimal thickness of ≈6–7 µm would ensure sufficient light harvesting, and thus high photoelectrochemical performance. In order to increase the accessible surface area, a simple ion exchange–induced poreforming process was designed by soaking the as-synthesized 2D nanostructured films in diluted HCl. The crystallographic structures of the 2D nanostructures before and after ion-exchange treatment were examined by X-ray diffraction (XRD) measurement, and it was found that K-titanate (K2Ti4O9) was dominant in all the as-prepared 2D Figure 2. Crystallographic structure analysis. A) XRD patterns of the samples during ionnanostructures, as confirmed in Figure S5 exchange and calcination processes. (a) as-prepared, (b) ion-exchanged, and (c) calcined (Supporting Information). Figure 2A shows films, respectively. B) TEM image of the E-2D nanostructure after ion exchange and calcination. the XRD change of E-2D crystal structure C,D) SAED patterns of the backbone plate (area 1 in (B)) and branches (area 2 in (B)), respectively. evolution upon ion exchange and calcination. E,F) HRTEM images of the example branch and backbone plate, respectively. Clearly, the characteristic peaks of layered titanate (Figure 2A-a) almost disappeared as a result of the single-crystalline plate. This single-crystalline plate architecture ion-exchange process (Figure 2A-b). Upon heat treatment, a clearly exhibited two orthogonal lattice spacing of ≈0.19 nm, pure anatase phase was obtained, as indexed in the XRD patcorresponding to anatase (200) and (020) planes.[32] These tern (Figure 2A-c). The crystal structures were further studied results indicate the excellent structural coherence of the backby transmission electron microscopy (TEM). Figure 2B shows bone plates, despite a rich amount of pores formed. Figure 2E a TEM image of the sub-microsized plate-like architecture of presents an HRTEM image of an individual nanoscaled branch the E-2D backbone, with nanosized hair-like branches (with (randomly selected from area 2 in Figure 2B), in which the diameters of 5–20 nm and lengths of 20–200 nm) growing single crystalline nature is confirmed. However, it is noted selectively on the edges of the plates. Figure 2C is a selected that the SAED of area 2 showed a ring-like pattern, as shown area electron diffraction (SAED) pattern taken from the plate in Figure 2D. This is because the randomly selected area 2 in part (area 1 in Figure 2B), suggesting that the backbone plate is Figure 2B for SAED pattern analysis contains many branches single crystal[30] and is taken along the [001] zone axis.[31] Highwith disordered orientations and these branches lead to a random scattering of the electrons along the longitude orienresolution TEM (HRTEM) image (Figure 2F) of such backbone tation, resulting in a ring-like electron diffraction pattern and plate reveals that a significant number of nanosized pores thus polycrystalline feature. with an average size of ≈2–6 nm are evenly distributed in the

Adv. Mater. 2018, 1705666

1705666 (3 of 8)



www.advmat.de

Figure 3. Nanostructure evolution mechanism. SEM images of A) K-titanate, B) H-titanate, and C) porous single-crystalline TiO2, respectively. D) Schematic illustration of converting K-titanate to porous single-crystalline TiO2 via ion exchange followed by calcination. E) Atomic crystal model of crystal structure changes during ion-exchange and calcination processes.

To understand the formation mechanism of TiO2 PSNs, we examined the morphological evolution of the nanostructures. Figure 3A–C shows typical SEM images of structural morphologies taken at different stages (as-prepared, ion-exchanged, and calcined). Figure 3A shows that the as-prepared E-2D (K-titanate) has a smooth surface with some cracks on the 2D main plates. The cracks in Figure 3A could be attributed to the stress in the layered crystals, which led to the separation of some weakly contacted layer in this layered K-titanate during the formation of branches.[24–27] By soaking the K-titanate film in diluted HCl (0.02 m), K+ ions in K-titanate can be easily replaced by the smaller H+ ions (protonation),[33] leading to the formation of H-titanate. The elimination of potassium elements by ion exchange is evidenced by the elemental mapping in Figure S6 (Supporting Information) where the amount of K is lower than the detecting limitation. To further confirm the elimination of potassium, we performed X-ray photoelectron spectroscopy (XPS) and energy-dispersive X-ray spectroscopy (EDS) elemental analysis. As indicated in Figure S7 (Supporting Information), the strong characteristic peaks of K2p1/2 (295.5 eV) and K2p3/2 (293.6 eV) shown in the XPS of K-titanate have vanished after ion exchange. The carbon peaks were from carbon tape, which was used as the substrate in XPS measurement. Moreover, the EDS elemental analysis (point analysis) also verified the disappearance of the characteristic peaks of K (≈3.3 and 3.6 keV), as shown in Figure S8 (Supporting Information).

Adv. Mater. 2018, 1705666

Noticeably, these cracks disappeared after ion exchange and simultaneously high porosity was clearly generated in the backbone structure, as shown in Figure 3B. After calcination at 400 °C in air, the protonated titanate was successfully transformed to single-crystalline anatase TiO2 without deterioration of the porous structure, as shown in Figure 3C. Based on the above experimental results, we proposed a hypothesis to explain the structural evolution of sample E-2D. The schematic illustration and corresponding atomic crystal models of different phases involved in the transformation are shown in Figure 3D,E, respectively. The substitution of K+ ions by smaller H+ ions in the protonation process may lead to a structure collapse into many small pieces of H-titanate composed of disordered octahedral TiO6 units, which was verified in the above mentioned XRD pattern of H-titanate (Figure 2A-b). Such ion exchange is a critical step as it can release the stress in the layered crystal structure and thus result in the disappearance of some cracks in the plates shown in Figure 3B.[26,34] Subsequent calcination would lead to the reorganization of these TiO6 units and dehydration, forming crack-free anatase TiO2 PSN with large porosity, as supported by the XRD pattern (Figure 2A-c) and SEM image (Figure 3C). To evaluate their optoelectronic properties, those novel nanostructures grown uniformly on FTO substrates in a large area (Figure S9, Supporting Information) were applied as the photoanodes for PEC water splitting. All the nanostructured

1705666 (4 of 8)



www.advmat.de

Figure 4. Photo-electrochemical performance and optoelectronic property analysis. A–C) Photocurrent density versus applied potential curves, IPCE recorded at the potential of 0.4 V versus Ag/AgCl, and EIS of 2D, E-2D, and F-2D films, respectively. D,E) Active site distribution in the nonbranched 2D single-crystalline structure indicated by the location of photodeposited Au and MnOx particles, respectively. F) Schematic diagram illustrating the working mechanism of E-2D photoelectrode–based PEC water splitting.

photoanode films including 2D, E-2D, and F-2D showed comparable absorption spectra and the bandgap was calculated to be 3.2 eV in the corresponding Kubelka–Munk function plots (Figure S10, Supporting Information). Figure 4A shows photocurrent density versus applied potential (J–V) curves of various nanostructures under simulated solar illumination (AM 1.5G) in 1 m KOH. At 1.23 V versus reversible hydrogen electrode (RHE), the bare 2D sample generated an anodic photocurrent density of 0.44 mA cm−2. By selectively growing branches on the edges of the backbone plates, the photocurrent density achieved from the E-2D sample was drastically increased up to 1.02 mA cm−2, which is close to the theoretical photocurrent density of anatase TiO2 (1.1 mA cm−2).[35–37] Moreover, the steep rise of photocurrent with respect to the potential, promptly reaching to the saturation point at ≈0.4 V versus RHE results in a high fill factor. The current–potential (j–E) curves of most reported photoelectrodes are “S” type, for which there is a slow increase of the current limited by the surface catalytic capability and/or semiconductor charge transfer. But in our case, the j–E curve of E-2D has a sharp increase and is approaching to the ideal curves predicted by the Gartner Model.[38] The near ideal j–E curve suggests that the E-2D is excellent in terms of efficient charge separation, transport, and collection.[39,40] The PEC performance achieved on sample E-2D is among the best in the reported anatase TiO2 nanostructures for PEC water splitting (Table S1, Supporting Information). However, further introducing nanosized branches over the entire plate structure deteriorated the photo current density to 0.52 mA cm−2 at 1.23 V versus RHE for sample F-2D.

Adv. Mater. 2018, 1705666

To evaluate the external quantum efficiency of various nanostructured photoelectrodes, the incident photon-to-current conversion efficiency (IPCE) measurement was performed at 1.23 V versus RHE in the wavelength region from 300 to 450 nm. As seen in Figure 4B, compared to 2D and F-2D, E-2D photo electrode exhibits much higher IPCE over the entire wavelength region with a maximum value of 94.3%, which is in accordance with the observed photocurrents from J–V curves. The IPCE can be rationalized using the following equation[41] IPCE = [ηe−/h + ][ηtran ][ηcoll ] (1) where ηe−/h+ is defined as the fraction of electron–hole (e−/h+) pairs generated per incident photon flux, ηtran represents the efficiency of charge transport to the solid–liquid interface, and ηcoll is the charge collection efficiency at the electrode–electrolyte interface.[41] As the photocarrier generation is determined by the absorption of semiconductor, all the photoelectrodes that exhibit similar light harvesting ability should have comparable ηe−/h+. Therefore, the higher IPCE obtained by the E-2D photoelectrode was believed to be ascribed to the enhanced ηtran and/ or ηcoll. To verify our hypothesis, electrochemical impedance spectra (EIS) analysis was conducted to gain further insight into the charge transfer kinetics of our nanostructured photo electrodes. Figure 4C presents the Nyquist plots of various samples, which were fitted using an equivalent circuit shown as its inset. Each Nyquist plot consists of a high-frequency intercept on the real axis corresponding to the series resistance (RS), a semicircle corresponding to a parallel combination of

1705666 (5 of 8)



www.advmat.de

charge-transfer resistance (R1) and constant phase element 1 (CPE1) at medium frequencies, and finally a linear region at lowfrequency range corresponding to the Warburg diffusion process of the electrolyte.[42,43] The identical Z′ intercept indicates that all samples have comparable RS comprising the intrinsic resistance of the material, ionic resistance of the electrolyte, and contact resistance between the active material and the photocurrent collector.[44] Among all the samples, the 2D and E-2D photoelectrodes exhibited comparable semicircle at the medium frequency range, while the F-2D photoelectrode showed significantly larger semicircle. Smaller semicircle at medium frequency range suggests smaller electron transport resistance and thus more efficient electron transport (ηtran) in the photoelectrode.[43] Despite the significantly enhanced surface area, the F-2D photoelectrode suffered from much larger electron transport resistance which was due to the dramatic increase in grain boundaries and interfaces with excess branches covering the backbone plate. Thus, the interfacial charge recombination was aggravated and the movement of electrons to the FTO was obstructed,[45] leading to increased radius of semicircles and decreased ηtran. This also explains why F-2D electrode needs higher overpotential to reach the saturated photocurrent. Rationally controlling the branch growth only on the edges of the backbone plates (E-2D) will significantly reduce the grain boundaries and interfaces on the backbone plates. It is worth noting that these selectively grown branches can still effectively enhance the surface area and thus create sufficient reactive sites for efficient water oxidation reaction. Moreover, these branches also act as direct pathways to efficiently extract holes from the bulk structure to the semiconductor–electrolyte interface,[11] which results in faster separation of photogenerated electron–hole pairs and facilitated the hole diffusion to the semiconductor/electrolyte interface. This was clearly verified by the steeper slope of the linear region for E-2D in the low frequency range, compared to that of 2D.[46,47] Therefore, the selective growth of branches is critical in achieving efficient charge separation and diffusion in PSNs. Moreover, exposing the nanomesh backbone to the electrolyte may shorten the charge migration to instantly encounter the reactants within a few nanometers, leading to minimized bulk charge recombination loss. As a result, the charge recombination lifetime is expected to be significantly prolonged, which was verified by the open-circuit photovoltage decay (OCPVD) measurement (Figure S11, Supporting Information). To gain further insights on the charge carrier kinetics in PSNs, we fabricated dye-sensitized solar cells (DSSCs) from various films including 2D, E-2D, and F-2D, and the intensity-modulated photocurrent spectroscopy (IMPS) measurement (Figure S12, Supporting Information) was performed on DSSCs. Further evidence from IMPS also showed the efficient charge carrier kinetics of E-2D photoanode in DSSCs which share the same principle with the PEC water splitting. Detailed analysis of IMPS results can be found in the Supporting Information. As discussed above, for the single-crystalline E-2D, the backbone plates featuring high structural coherence and long-range electronic connectivity act as a fast track for electron transfer free from grain boundaries. Furthermore, the selectively grown branches are able to facilitate hole transport to electrode–electrolyte interfaces.[11] To further examine the proposed

Adv. Mater. 2018, 1705666

mechanism, we conducted photodeposition of noble metal (Au) and metal oxide (MnOx),[48,49] respectively, on the nonbranched 2D single-crystalline nanostructure which provides a smooth surface for clearer observation. Reductive area that is electron rich is likely to have Au particles deposited according to the reaction: Au+ + e− → Au, while the hole rich area is likely to have metal oxide anchored according to the reaction Mn2+ + xH2O + (2x–m)h+ → MnOx + 2xH+.[48] As shown in Figure 4D, the observed Au particles at the middle of the plate structure indicate that the photogenerated electrons were extracted rapidly through the backbone plate structure to FTO substrate. On the other hand, MnOx nanoparticles were found to be the predominant particles formed on the nanostructure, mainly distributed on the edges and thin parts of the plate structure (Figure 4E) which manifests that holes are the dominant charge carriers in these area. This evidence explicitly demonstrates a spatial charge separation in the PSNs, and thus the charge recombination is suppressed. In addition to the excellent charge transfer property, the highly accessible surface area also contributed to the higher photocurrent achieved on the E-2D photoelectrode. The surface area was evaluated by measuring the dye-loading capacity, with the results summarized in Table S3 (Supporting Information). Compared to the compact single-crystalline anatase film (Figures S13 and S14, Supporting Information), the porous E-2D film showed 25 times higher surface area (174.7 vs 5.6 nmol cm−2, dye-loading capacity) as well as photocurrent, which was attributed to the large porosity generated via the ion-exchange process. We also conducted Brunauer–Emmett– Teller (BET) measurement to confirm the specific surface area of each sample (Figure S16, Supporting Information). The specific surface areas (BET) of F-2D, E-2D, 2D, and compact TiO2 are 194.2, 183.3, 168.0, and 45.6 m2 g−1, respectively. The trend is consistent with the data derived from dye-loading capacity measurement. Figure 4F presents the schematic diagram illustrating the working mechanism of E-2D photoelectrode–based PEC water splitting. The photogenerated electrons are transported efficiently via the single-crystalline backbone nanostructure as the direct pathways to counter electrode for water reduction. Direct growth of this nanostructure on FTO provides intimate contact between the film and substrate and improves the electron collection efficiency. Moreover, the ultrathin branches not only further increase the accessible surface area, but also effectively promote the hole transfer to the semiconductor–electrolyte interface, leading to more efficient water oxidation kinetics.[11] In conclusion, we have, for the first time, demonstrated an effective synthesis method for direct growth of porous singlecrystalline semiconductor nanostructure on FTO substrates, involving a one-pot hydrothermal reaction and subsequent ion exchange–induced pore-forming process. The vertically aligned, 2D porous single-crystalline anatase TiO2 nanostructures with selectively grown branches not only possess highly accessible surface area but also deliver excellent charge separation and transport efficiency, leading to high performance of PEC water splitting. These findings shed light on the design of synthetic strategy for achieving high-performing nanostructured materials for a variety of applications such as solar cells and supercapacitors.

1705666 (6 of 8)



www.advmat.de

Experimental Section Various 2D-Branched K-titanate Synthesis: Clean FTO-coated glass was seeded by soaking in 0.04 m TiCl4 for 30 min, followed by blowdrying, and calcination at 450 °C for 1 h. In a typical synthesis of 2D-branched K-titanate, 0.438 g of PTO was dissolved in x mL of water (where x = 3, 3.25, and 3.50 for 2D, E-2D, and F-2D, respectively) under mechanical stirring. After stirring for 15 min, 40-x mL of DEG was added with continuous stirring for 30 min. The resultant emulsion was then transferred to a 250 mL Teflon-lined autoclave, which contained the desirable size of seeded FTO glass by putting the conductive side down against the wall of Teflon-lined autoclave. The angle between the FTO glass and the bottom of the Teflon-lined autoclave was around 20°–30°. After fitting in the stainless-steel shell, the reaction vessel was then transferred to the oven set at 180 °C. The reaction was allowed to proceed for 9 h for 2D, and 12 h for E-2D and F-2D. Then, the reaction vessel was cooled down naturally to room temperature before removing the film. The film on FTO was gently rinsed with ethanol to remove excess precursor and possible contaminants, and then dried in the oven set at 70 °C for 12 h. Before carrying out the ion-exchange process, the dried film was calcined at 400 °C (ramp rate = 1 °C min−1, 2 h) in a tube furnace with a flow of compressed air to enhance the contact with FTO substrate and remove possible organic molecules left after synthesis. Ion Exchange to Transform K-titanate to H-titanate: The calcined film was soaked in ethanolic HCl (0.04 m) with gentle agitation for 3 h, and aqueous HCl (0.02 m) for 3 h, respectively (diluted HCl was refreshed every hour). The film was then rinsed with ethanol and dried in the oven set at 70 °C for 12 h. After that, the film was calcined again to transform H-titanate to porous single-crystalline branched 2D anatase. The calcination condition was the same as the first calcination. Synthesis of Compact Single-Crystalline Anatase Film: For comparison with porous single-crystalline 2D anatase films, the compact singlecrystalline anatase film on FTO was synthesized based on the method reported previously.[50] The thickness of the film was controlled to be comparable to that of 2D-branched films (≈6 µm). Photodeposition of Noble Metal and Metal Oxide: The nonbranched 2D single-crystalline films were soaked in a mixture solution (0.025 m) of HAuCl4 and MnSO4, and irradiated under a solar simulator for a desired period of time. Characterization: The morphological characteristics of the obtained nanostructures were observed by field emission SEM (JEOL JSM-7100F) at 15 kV. Their structural characteristics as well as EDS were investigated by TEM (JEOL 2100F operated at 200 kV). Crystalline phases of various films on substrate were analyzed by XRD on a Bruker Advanced X-ray diffractometer (40 kV, 30 mA) with Cu Kα radiation (wavelength = 1.54060). UV–vis light absorption spectra of the films were measured using a Shimadzu 2200 UV–vis spectrophotometer. The XPS measurement was carried out at SUT-NANOTEC-SLRI XAS Beamline (BL 5.3), Synchrotron Light Research Institute, Thailand, with a PHI 5000 Versaprobe-II (ULVAC-PHI, INC) spectrometer. Spectra were corrected to the main line of the carbon 1s spectrum (adventitious carbon) set to 284.8 eV. Spectra were analyzed using MultiPak software (version 9.6). The example large-area films were taken by Nikon D7000. The OCPVD measurement was performed in a similar way as previously reported.[51] The surface area of various films was evaluated by measuring the dyeloading capacity according to the method reported previously.[52] More details can be found in Figure S15 (Supporting Information). Photo-Electrochemical Measurements: PEC, IPCE, and EIS measurements were conducted on a CHI 660 electrochemical workstation (CHI Instrument).[42] The applied bias for IPCE measurement was 0.4 V versus Ag/AgCl. The reaction chamber was the standard 3-electrode configuration, using TiO2-coated FTO, Pt mesh, and Ag/AgCl as working, counter, and reference electrodes, respectively. The electrolyte was 1 m KOH. For linear sweep voltammetry and EIS tests, the light source was simulated sunlight illumination (100 mW cm−2) using a Xenon lamp (150 W, Newport) equipped with AM 1.5G filter with front-side illumination to the working electrode. The

Adv. Mater. 2018, 1705666

reaction chamber was bubbled with N2 for 30 min before the tests. All potentials were measured by Ag/AgCl reference electrode, and converted to RHE reference scale using the Nernst equation E RHE = E Ag/AgCl + 0.0591× pH + 0.1976 V

(2)

The DSSCs based on 2D, E-2D, and F-2D films were assembled following the previous report with slight modification.[53,54] IMPS measurement was performed using Metrohm Autolam PGSTAT302 potentiostat/galvanostat, equipped with Autolab Photokit. The light source was a triple light emitting diode (LED) array (LDC627-Red, λ = 627 nm, 700 mA) driven by the output current of the Autolab LED Driver, which provides both AC and DC components of the illumination in frequency response analyzer mode. The intensity of the incident light was 5.9 mW cm−2. The scan range was 10 kHz–1 Hz. The measurement was carried out with NOVA 2.1 software.

Supporting Information Supporting Information is available from the Wiley Online Library or from the author.

Acknowledgements T.B. and Y.B. contributed equally to this work. The authors acknowledge the financial support from the Australian Research Council through its DP and FF programs. The technical support from the Queensland Node of the Australian National Fabrication Facility is appreciated. The authors are grateful to the Australian Microscopy & Microanalysis Research Facility at the Centre for Microscopy and Microanalysis (CMM), The University of Queensland. The financial support from the National Natural Science Foundation of China (Nos. 51629201, 51422210) and the Major Basic Research Program, Ministry of Science and Technology of China (2014CB239401) is also appreciated. T.B. acknowledges financial support from the NANOTEC Thailand, Royal Thai Government Scholarship. The authors acknowledge Assistant Professor Dr. Montree Sawangphruk, Vidyasirimedhi Institute of Science and Technology, Thailand, for the use of IMPS measurement facility. The authors thank Dr. Pisist Kumnorkaew, NANOTEC, Thailand, for assisting in DSSC fabrication.

Conflict of Interest The authors declare no conflict of interest.

Keywords 2D, ion-exchange, pore-forming, porous single-crystalline TiO2 films, water splitting Received: September 29, 2017 Revised: February 12, 2018 Published online:

[1] P. V. Kamat, J. Phys. Chem. C 2007, 111, 2834. [2] S. Linic, P. Christopher, D. B. Ingram, Nat. Mater. 2011, 10, 911. [3] B. Liu, E. S. Aydil, J. Am. Chem. Soc. 2009, 131, 3985. [4] F. Zaera, Chem. Soc. Rev. 2013, 42, 2746. [5] F. E. Osterloh, Chem. Soc. Rev. 2013, 42, 2294. [6] A. S. Arico, P. Bruce, B. Scrosati, J.-M. Tarascon, W. van Schalkwijk, Nat. Mater. 2005, 4, 366.

1705666 (7 of 8)



www.advmat.de

[7] Y.-G. Guo, J.-S. Hu, L.-J. Wan, Adv. Mater. 2008, 20, 2878. [8] J. Zhang, X. Liu, G. Neri, N. Pinna, Adv. Mater. 2016, 28, 795. [9] G. Konstantatos, E. H. Sargent, Nat. Nanotechnol. 2010, 5, 391. [10] Y. Liu, R. Che, G. Chen, J. Fan, Z. Sun, Z. Wu, M. Wang, B. Li, J. Wei, Y. Wei, G. Wang, G. Guan, A. A. Elzatahry, A. A. Bagabas, A. M. Al-Enizi, Y. Deng, H. Peng, D. Zhao, Sci. Adv. 2015, 1, e1500166. [11] I. S. Cho, Z. Chen, A. J. Forman, D. R. Kim, P. M. Rao, T. F. Jaramillo, X. Zheng, Nano Lett. 2011, 11, 4978. [12] L. Wang, X. Bi, S. Yang, Adv. Mater. 2016, 28, 7672. [13] Y. Peng, Z. Le, M. Wen, D. Zhang, Z. Chen, H. B. Wu, H. Li, Y. Lu, Nano Energy 2017, 35, 44. [14] P. Docampo, S. Guldin, U. Steiner, H. J. Snaith, J. Phys. Chem. Lett. 2013, 4, 698. [15] E. J. Crossland, N. Noel, V. Sivaram, T. Leijtens, J. A. AlexanderWebber, H. J. Snaith, Nature 2013, 495, 215. [16] N. Tetreault, M. Gratzel, Energy Environ. Sci. 2012, 5, 8506. [17] X. Feng, K. Shankar, O. K. Varghese, M. Paulose, T. J. Latempa, C. A. Grimes, Nano Lett. 2008, 8, 3781. [18] Z. Zhang, L. Zhang, M. N. Hedhili, H. Zhang, P. Wang, Nano Lett. 2012, 13, 14. [19] M. Zhou, J. Bao, W. Bi, Y. Zeng, R. Zhu, M. Tao, Y. Xie, ChemSusChem 2012, 5, 1420. [20] M. Zhou, X. W. Lou, Y. Xie, Nano Today 2013, 8, 598. [21] N. Tétreault, É. Arsenault, L.-P. Heiniger, N. Soheilnia, J. Brillet, T. Moehl, S. Zakeeruddin, G. A. Ozin, M. Grätzel, Nano Lett. 2011, 11, 4579. [22] P. Yang, D. Zhao, D. I. Margolese, B. F. Chmelka, G. D. Stucky, Chem. Mater. 1999, 11, 2813. [23] R. Müller, E. D. Zanotto, V. M. Fokin, J. Non-Cryst. Solids 2000, 274, 208. [24] H. L. Bhat, J. N. Sherwood, T. Shripathi, Chem. Eng. Sci. 1987, 42, 609. [25] R. Laghmach, N. Candau, L. Chazeau, E. Munch, T. Biben, J. Chem. Phys. 2015, 142, 244905. [26] L. T. Romano, C. G. Van de Walle, J. W. Ager III, W. Götz, R. S. Kern, J. Appl. Phys. 2000, 87, 7745. [27] Y. Bao, Eng. Fract. Mech. 2005, 72, 505. [28] A. Wahl, J. Augustynski, J. Phys. Chem. B 1998, 102, 7820. [29] J. Zhang, J. Wang, Z. Zhao, T. Yu, J. Feng, Y. Yuan, Z. Tang, Y. Liu, Z. Li, Z. Zou, Phys. Chem. Chem. Phys. 2012, 14, 4763. [30] X. Zou, S. Hovmöller, P. Oleynikov, Electron Crystallography: Electron Microscopy and Electron Diffraction, Vol. 16, Oxford University Press, Oxford, UK 2011. [31] H. G. Yang, C. H. Sun, S. Z. Qiao, J. Zou, G. Liu, S. C. Smith, H. M. Cheng, G. Q. Lu, Nature 2008, 453, 638. [32] C. Z. Wen, J. Z. Zhou, H. B. Jiang, Q. H. Hu, S. Z. Qiao, H. G. Yang, Chem. Commun. 2011, 47, 4400. [33] C.-W. Peng, M. Richard-Plouet, T.-Y. Ke, C.-Y. Lee, H.-T. Chiu, C. Marhic, E. Puzenat, F. Lemoigno, L. Brohan, Chem. Mater. 2008, 20, 7228. [34] P. Strasser, S. Koh, T. Anniyev, J. Greeley, K. More, C. Yu, Z. Liu, S. Kaya, D. Nordlund, H. Ogasawara, M. F. Toney, A. Nilsson, Nat. Chem. 2010, 2, 454. [35] A. B. Murphy, P. R. F. Barnes, L. K. Randeniya, I. C. Plumb, I. E. Grey, M. D. Horne, J. A. Glasscock, Int. J. Hydrogen Energy 2006, 31, 1999.

Adv. Mater. 2018, 1705666

[36] Z.-F. Huang, L. Pan, J.-J. Zou, X. Zhang, L. Wang, Nanoscale 2014, 6, 14044. [37] J. Li, N. Wu, Catal. Sci. Technol. 2015, 5, 1360. [38] D. J. Fermín, E. A. Ponomarev, L. M. Peter, J. Electroanal. Chem. 1999, 473, 192. [39] J. Tang, K. W. Kemp, S. Hoogland, K. S. Jeong, H. Liu, L. Levina, M. Furukawa, X. Wang, R. Debnath, D. Cha, K. W. Chou, A. Fischer, A. Amassian, J. B. Asbury, E. H. Sargent, Nat. Mater. 2011, 10, 765. [40] Y. Li, J. Z. Zhang, Laser Photonics Rev. 2010, 4, 517. [41] Z. Chen, T. F. Jaramillo, T. G. Deutsch, A. Kleiman-Shwarsctein, A. J. Forman, N. Gaillard, R. Garland, K. Takanabe, C. Heske, M. Sunkara, J. Mater. Res. 2010, 25, 3. [42] S. Wang, H. Chen, G. Gao, T. Butburee, M. Lyu, S. Thaweesak, J.-H. Yun, A. Du, G. Liu, L. Wang, Nano Energy 2016, 24, 94. [43] H. Xia, Y. Wang, J. Lin, L. Lu, Nanoscale Res. Lett. 2012, 7, 33. [44] M. Xu, L. Kong, W. Zhou, H. Li, J. Phys. Chem. C 2007, 111, 19141. [45] W.-Q. Wu, H.-L. Feng, H.-S. Rao, Y.-F. Xu, D.-B. Kuang, C.-Y. Su, Nat. Commun. 2014, 5, 3968. [46] S. Devaraj, N. Munichandraiah, J. Electrochem. Soc. 2007, 154, A80. [47] K. Yuan, T. Hu, Y. Xu, R. Graf, G. Brunklaus, M. Forster, Y. Chen, U. Scherf, ChemElectroChem 2016, 3, 822. [48] C. Zhen, C. Y. Jimmy, G. Liu, H.-M. Cheng, Chem. Commun. 2014, 50, 10416. [49] Y. Yang, G. Liu, J. T. Irvine, H. M. Cheng, Adv. Mater. 2016, 28, 5850. [50] S. Feng, J. Yang, H. Zhu, M. Liu, J. Zhang, J. Wu, J. Wan, J. Am. Ceram. Soc. 2011, 94, 310. [51] H. Cui, G. Zhu, Y. Xie, W. Zhao, C. Yang, T. Lin, H. Gu, F. Huang, J. Mater. Chem. A 2015, 3, 11830. [52] M. Lv, D. Zheng, M. Ye, L. Sun, J. Xiao, W. Guo, C. Lin, Nanoscale 2012, 4, 5872. [53] Y. Bai, H. Yu, Z. Li, R. Amal, G. Q. Lu, L. Wang, Adv. Mater. 2012, 24, 5850. [54] H. Yu, J. Pan, Y. Bai, X. Zong, X. Li, L. Wang, Chem. - Eur. J. 2013, 19, 13569. [55] K. Zhu, N. R. Neale, A. Miedaner, A. J. Frank, Nano Lett. 2007, 7, 69. [56] J. van de Lagemaat, N. G. Park, A. J. Frank, J. Phys. Chem. B 2000, 104, 2044. [57] J. Krüger, R. Plass, M. Grätzel, P. J. Cameron, L. M. Peter, J. Phys. Chem. B 2003, 107, 7536. [58] M. Rodríguez-Pérez, I. Rodríguez-Gutiérrez, A. Vega-Poot, R. García-Rodríguez, G. Rodríguez-Gattorno, G. Oskam, Electrochim. Acta 2017, 258, 900. [59] Gurudayal, L. M. Peter, L. H. Wong, F. F. Abdi, ACS Appl. Mater. Interfaces 2017, 9, 41265. [60] Z. Zhang, M. F. Hossain, T. Takahashi, Int. J. Hydrogen Energy 2010, 35, 8528. [61] I. S. Cho, M. Logar, C. H. Lee, L. Cai, F. B. Prinz, X. Zheng, Nano Lett. 2014, 14, 24. [62] Z. Pan, Y. Qiu, J. Yang, M. Liu, L. Zhou, Y. Xu, L. Sheng, X. Zhao, Y. Zhang, J. Mater. Chem. A 2015, 3, 4004. [63] Z. Zhang, P. Wang, Energy Environ. Sci. 2012, 5, 6506. [64] Q. Kang, J. Cao, Y. Zhang, L. Liu, H. Xu, J. Ye, J. Mater. Chem. A 2013, 1, 5766. [65] W. Wang, J. Dong, X. Ye, Y. Li, Y. Ma, L. Qi, Small 2016, 12, 1469. [66] S. Y. Noh, K. Sun, C. Choi, M. Niu, M. Yang, K. Xu, S. Jin, D. Wang, Nano Energy 2013, 2, 351. [67] S. Soedergren, A. Hagfeldt, J. Olsson, S.-E. Lindquist, J. Phys. Chem. 1994, 98, 5552.

1705666 (8 of 8)
