Band gap and Morphology Engineering of Hematite Nanoflakes from an Ex Situ Sn Doping for Enhanced Photoelectrochemical Water Splitting.

In this article, we report a simple ex situ Sn-doping method on hematite nanoflakes (coded as MSnO2-H) that can protect the nanoflake (NF) morphology against the 800 °C high-temperature annealing process and activate the photoresponse of hematite until 800 nm wavelength excitation. MSnO2-H has been fabricated by dropping SnCl4 ethanol solution on hematite nanoflakes homogeneously grown over the conductive FTO glass substrate and annealed at 500 °C to synthesize the SnO2 nanoparticles on hematite NFs. The Sn-treated samples were then placed in a furnace again, and the sintering process was conducted at 800 °C for 15 min. During this step, structure deformation of hematite occurs normally due to the grain boundary motion and oriented attachment. However, in the case of MSnO2-H, the outer SnO2 nanoparticles efficiently prevented a shape deformation and maintained the nanoflake shape owing to the encapsulation of hematite NFs. Furthermore, the interface of hematite/SnO2 nanoparticles became the spots for a heavy Sn ion doping. We demonstrated the generation of the newly localized states, resulting in an extension of the photoresponse of hematite until 800 nm wavelength light irradiation. Furthermore, we demonstrated that SnO2 nanoparticles can effectively act as a passivation layer, which can reduce the onset potential of hematite for water splitting redox reactions. The optimized MSnO2-H nanostructures showed a 2.84 times higher photocurrent density and 300 mV reduced onset potential compared with a pristine hematite nanoflake photoanode.

Introduction

Hematite iron oxide (α-Fe2O3) is one of the promising candidates for the production of hydrogen via photoelectrochemical water splitting (PEC-WS) owing to its suitable band gap (2.1 eV) providing a high theoretical solar-to-hydrogen efficiency (15.3%), high chemical and photocorrosion stability, as well as earth abundance.[1−3] It is an n-type semiconductor and hence is utilized as the photoanode material in the PEC-WS cells. However, the performance of the hematite photoanodes is still limited due to several drawbacks such as a high electrical resistivity, short hole-diffusion length (Lh ≈ 2–4 nm), high charge recombination rate, and poor oxygen evolution reaction kinetics taking place on its surface.[4−6] A number of approaches have been introduced to address these problems including, particularly, various nanostructuring and morphology controls (e.g., thin film,[7] nanoflake,[8] nanorod,[9] and nanoporous structure[10]), doping by foreign elements (Sn4+,[11] Ti4+,[12] Pt4+,[13] B3+,[14] etc.), deposition of cocatalysts on the hematite surface (e.g., Zn-Co LDH,[15] Co-Pi,[16] and FeOOH[17]), and/or deposition of isocrystalline ultrathin passivation layers.[18,19]

Morphologically controlled hematite nanostructures are highly desired in terms of increasing light absorption because of their high specific area and addressing problems associated with a short diffusion length of photoholes, which both lead to a significantly enhanced PEC-WS activity.[20−22] Elemental doping is the efficient way to increase a donor density in hematite and thus the electrical conductivity.[23] Tin (Sn) is the most typical example of metal impurity used for the hematite doping. The substitution of iron by Sn4+ ions in the hematite lattice has led to an increase in the donor density ND from 7.86 × 1019 to 1.95 × 1020 cm–3.[24] Consequently, the dynamics of photocharges is enhanced because of better electrical conductivity. Both in situ and ex situ methods have been described to incorporate tin ions into hematite structures. In the case of in situ doping, Sn dopants were added to the precursor solution of hematite during hydrothermal growth.[25]In situ doping is a plain process that provides flexibility in the selection of dopants and uses relatively low sintering temperatures without additional doping steps.[26] However, deformation of morphology and changes in crystallinity are observed from in situ-doped hematite.[23] The most frequent strategy for ex situ doping is a thermal diffusion of tin ions coming from the FTO glass substrate (the glass substrate is coated by a conductive film of fluorine-doped SnO2) at high temperatures (∼800 °C).[15] Correspondingly, one can use a tin-containing precursor solution deposited on the hematite or akaganeites (β-FeOOH) surface, which is then thermally treated under similar temperatures.[27] The key advantage of the ex situ processes is that various hematite and/or akaganeite nanostructures (nanorods,[28] nanotubes,[29] nanoflakes,[30] etc.) can be utilized as the starting substrates. Furthermore, Sn doping also leads to a slightly reduced optical band gap and enhanced optical adsorption coefficient by structural distortion of the hematite lattice.[31] In addition, from computational approaches a band gap reduction of hematite is expected by newly localized energy levels, resulting from orbital hybridization between Fe ions (3d orbital) and Sn ions (5s and 5p orbitals).[32] However, there are only a few studies dealing with the relationship between PEC performance and increased optical property of hematite.[31,33] Especially, the photoresponse at long wavelength (λ ≥ 650 nm) of the incident light without plasmon resonances by novel metals and/or light-trapping structures, which is an important issue for improving the PEC performance of hematite, has been rarely reported.[34,35]

The passivation overlayers, for example, TiO2,[18] Al2O3,[19] or Ga2O3[36] conformally and homogeneously covering the hematite surface generally provide two beneficial effects; namely, they improve stability of photoanodes by preventing a direct contact between electrolyte and hematite, as well as they passivate the detrimental surface states that work as recombination centeres. As a consequence of the surface state inactivation and thus of reduced recombination of the charge carriers, the photocurrent density increases, and simultaneously, the onset potential is shifted to a more cathodic region, which results in significantly higher PEC-WS activity.[37]

In this work, we report morphology-controlled hematite nanoflakes that were treated with a SnO2 coating, which acts as a passivation layer (MSnO2-H) and provides photoresponses at long wavelength (∼800 nm) using ex situ surface treatments and a high-temperature annealing process. The photocurrent density of the MSnO2-H obtained is 0.847 mA/cm2 at 1.23 VRHE, which is about 16-fold enhanced PEC activity over pristine hematite nanoflakes (P-H). We attribute the significant enhancement not only to a reduced recombination by surface passivation effects that can be defined using an intensity-modulated photocurrent spectroscopy (IMPS) analysis but also to an enhanced conductivity and extended photoresponses (650 ≤ λ ≤ 800 nm) by newly localized energy levels resulting from a high concentration of Sn4+ doping.

Results and Discussion

The fabrication of hematite nanoflake-based photoanodes follows several steps, as depicted in Figure . First, a 600 nm thick film of pure iron was deposited onto a FTO substrate by using magnetron sputtering. It is worth noting that it is a novel strategy for preparing fully transparent hematite photoanodes with a nanoflake coverage, that is, we use a total thermal oxidation of the iron thin film to obtain this morphology.[38] In the second step, the FTO/iron film substrates were annealed at 400 °C for 2 h to grow pristine hematite nanoflakes (hereafter, the samples denoted as P-H). In the third step, a precursor solution of SnCl4 diluted in ethanol was drop-casted over the P-H samples and then annealed at 800 °C for 15 min (hereafter, the samples are denoted as SnO2-H). Alternatively, prior to the high-temperature treatment, the samples were annealed first at 500 °C for 1 h in order to preserve the nanoflake morphology (hereafter, the samples are denoted as MSnO2-H). In other words, during the 500 °C annealing process, the thin and small SnO2 NPs are uniformly formed on the surface of hematite nanoflakes, which can prevent a morphology deformation resulting from the restricted grain boundary motion and oriented attachment.[20] It is a strategy similar to the SiO2 encapsulation described by Sivula et al.[39]

Figure 1

Schematic illustration of the fabrication procedures for (a) Fe film, (b) P-H (Fe2O3 NF), (c) SnO2-H, and (d) MSnO2-H, respectively.

The top-view scanning electron microscopy (SEM) images of the hematite photoanodes treated under different conditions are displayed in Figure . The as-grown hematite samples (Figures a, S1b) show a dense array of well-developed nanoflakes, having a length varying between 0.3 and 2 μm and thickness between ∼150 and ∼20 nm at the bottom and their terminating tips, respectively. During the single annealing process at 800 °C of bare hematite nanoflakes (i.e., those without applied SnCl4 solution), an undesirable sintering process of the nanoflakes clearly occurred. It means that the nanoflake morphology transformed to a much sharper nanowire-like shape due to grain boundary diffusion, as demonstrated by SEM and transmission electron microscopy (TEM) images (Figures b and S2). In this procedure, the tin doping in the hematite lattice took place via the mechanism of the thermal diffusion of Sn4+ ions realized from the bottom FTO substrate (the sample is hereafter coded as Sn-H).[15,25] A semi-preserved nanoflake structure was achieved despite the same high-temperature annealing in the case of SnCl4-treated hematite nanoflakes (Figure c). Finally, it can be observed in Figure d that the samples treated with the SnCl4 solution and then annealed in two sequential steps, first at 500 °C for 1 h and then at 800 °C for 15 min (see Figure ), well-maintained the desired nanoflake morphology (MSnO2-H, Figure S3). During the first annealing step, the dried drops of the SnCl4 precursor oxidized to SnO2 NPs, which uniformly covered the hematite nanoflake surface. These nanoparticles then provide two functions during the following second annealing step; namely, they form an encapsulation scaffold to prevent morphology change (a simpler approach than previously reported silica encapsulation) and are a simultaneous source of the ex situ Sn4+ doping of hematite.

Figure 2

SEM images of (a) P-H, (b) Sn-H, (c) SnO2-H, and (d) MSnO2-H, respectively.

The TEM images of P-H, SnO2-H, and MSnO2-H are well consistent with corresponding SEM images, as shown in Figure . The energy-dispersive spectroscopy (EDS) mapping (Figure e–g) visualizes the distribution of Sn, Fe, and O ions in MSnO2-H. In particular, the quantitative analysis of a single SnO2 nanoparticle on Fe2O3 NF, which is calculated from EDS mapping (Figure S4), demonstrates that Sn ions from SnO2 and Fe ions from Fe2O3 NF can be interchangeably diffused into each other during high-temperature annealing, leading to the formation of Fe-doped SnO2 (F/SnO2) and Sn-doped Fe2O3 in MSnO2-H. As shown in Figure h, the particle size distribution of Fe/SnO2 is from 1.0 to 8.1 nm with a lattice spacing of 0.33 nm, which corresponds to the (110) plane of tetragonal SnO2 (inset of Figure h).[40]

Figure 3

TEM images of (a) P-H, (b) SnO2-H, and (c,d, and h) MSnO2-H. EDS mapping images of (e) Sn-K, (f) Fe-K, and (g) O-K signals from a MSnO2-H. (h) TEM image of Fe/SnO2 NPs. Inset in (h): lattice constant of Fe/SnO2 NPs. (i) XRD data of P-H, Sn-H, SnO2-H, and MSnO2-H, respectively, (j) Fe 2p scans for the P-H and MSnO2-H, and (k) band gap calculation from an (Iphhv)1/2vs photon energy (hv) plot of P-H and MSnO2-H.

The crystalline structure of the photoanodes was determined by X-ray diffraction. All the samples tested showed a well-crystalline hematite structure with the preferential orientation of the crystallites along the (110) crystalline plane, which provides the highest electrical conductivity. One of the important features is that only the P-H sample contained a contribution of the magnetite iron oxide (Fe3O4) phase, which commonly accompanies iron oxide nanoflakes prepared by thermal oxidation of iron substrates. The presence of magnetite, however, decreases the overall efficiency of the hematite photoanodes. Thanks to the second annealing step at 800 °C, the magnetite phase was fully transformed into the pure hematite phase.

To study the chemical compositions of the P-H and MSnO2-H samples, X-ray photoelectron spectroscopy (XPS) was performed (Figures i and S4). Figure i presents the Fe 2p scans showing the Fe 2P3/2 (711 eV) and the Fe 2P1/2 (724.5 eV) peaks that are well consistent with those reported for hematite. Furthermore, the coexistence of Fe 2P peaks and an additional peak at 715.8 eV, which is indicative of the Sn 3P3/2 in MSnO2-H, demonstrates the decoration of SnO2 particles on hematite nanoflakes. In addition, the Fe ion diffusion from hematite NFs into SnO2 (Fe/SnO2) because of the ion diffusion due to the high-temperature sintering process can be observed by EDX mapping (Figure S4d).[15,41] The XPS analysis of the MSnO2-H shows two major peaks at 487.3 eV (Sn 3d5/2) and 495.7 eV (Sn 3d3/2), which are matched with the reported XPS data for Sn-treated hematite, indicating substitutional doping of Sn4+ ions in the hematite lattice (Figure S5).[23]Figure S6 shows the valence band maximum (VBM) of P-H (1.20 eV) and MSnO2-H (1.55 eV) assessed by linearly fitting the leading edge of the valence band retrieved from the XPS measurements. The analysis indicates that the Fermi level of MSnO2-H is shifted toward the conduction band (CB) compared with pristine P-H. This Fermi-level shifting demonstrates that hematite nanoflakes are doped by Sn4+ ions. A concentration of 12.08% of Sn on MSnO2-H was estimated from the XPS analysis. This value includes the Sn-ion doping in Fe/SnO2 NPs. To calculate the Sn concentration of MSnO2-H more precisely, we can assume that the spots of the highest Sn concentration in EDS (red spots in Figure S7) correspond to a Sn concentration of pristine SnO2 (33.73 at. %[42]). This leads to a Sn concentration at the interface of Fe/SnO2/hematite NFs of approximately 10–12.5% and the average Sn concentration of hematite NFs in MSnO2-H of 7.5%, respectively. The band gap for the samples can be determined from the plot (iphhν)1/2versus photo energy (hν), where iph is photocurrent density, as shown in Figure k. In particular, the band gap of P-H and MSnO2-H can be read as approximately 2.0 and 1.87 eV, respectively, which are the values well corresponding to those reported for hematite iron oxide. Interestingly, from the plot corresponding to the MSnO2-H sample, potentially, one more value of a band gap of 1.53 eV can be observed caused by newly localized energy levels resulting from the high concentration of Sn-ion doping (ca. 10–12.5%). This result is well consistent with the computational simulation, which shows the reduced band gap (1.57 eV) of hematite with 12.5 atom % of Sn.[32]

The linear sweep voltammetry of P-H, Sn-H, SnO2-H, and MSnO2-H in a 1 M KOH solution with the Ag/AgCl reference electrode under an AM 1.5 solar simulator is shown in Figure a. The PEC performance of the bilayer structure of hematite nanoflakes and thin film using an optimized process for its preparation has been already reported by our group elsewhere.[38] The charge compensations between the photoexcited electrons from hematite nanoflakes and photoexcited holes from the bottom hematite film layer can effectively reduce the probability of a recombination rate of photoanode. The P-H had a photocurrent density of 0.38 mA/cm2 at 1.50 VRHE and an onset potential of 1.21 VRHE, and the photocurrent density of Sn-H increased by 0.63 mA/cm2 at 1.50 VRHE with an onset potential of 1.13 VRHE. The increased PEC performance of the Sn-H compared with P-H is caused by not only the reduced magnetite layer located between hematite nanoflakes and FTO glass that hinders the electron transfer, as shown in X-ray diffraction (XRD) data (see Figure i), but also the enhanced conductivity by Sn-ion doping from bottom FTO glass due to the 800 °C annealing process. When the SnCl4 was treated on Fe2O3 NF, as shown in Figure , the photocurrent density of SnO2-H reached 0.92 mA/cm2 at 1.50 VRHE with the onset potential of 0.93VRHE due to surface passivation and Sn4+ ion doping. In addition, the photocurrent density of MSnO2-H is further enhanced to 1.08 mA/cm2 at 1.50 VRHE with an onset potential of 0.93 VRHE. This can be ascribed to the larger surface area of MSnO2-H than that of SnO2-H obtained by a simple morphology engineering, which helps to preserve the shape of nanoflakes similar to P-H compared with SnO2-H (nanowire-like shape, Figure b). As a consequence, more light, including long wavelength, can be absorbed by the nanoflakes, and thus a higher photocurrent density can be reached. The IPCE data in Figure b shows a higher photoresponse of SnCl4-treated samples compared to that of P-H and Sn-H according to the wavelength at 1.50 VRHE. Although the IPCE values of P-H from 500 to 650 nm wavelength could not be detected, the IPCE values of SnO2-H and MSnO2-H can clearly be observed from 500 to 650 nm due to decreased recombination and the electron–hole pairs (EHPs), which are generated by newly localized states due to a high concentration of Sn doping. To compare the photoresponse of samples at 800 nm wavelength, the chronoamperometry was measured in a 1 M of KOH electrolyte and at an applied bias of 1.50 VRHE under chopped light illumination. The acquired data are presented in Figure c. The photocurrent of the MSnO2-H is found to be 1.5 μA, whereas that of Sn-H and P-H is approximately zero. The higher photocurrent of MSnO2-H than SnO2-H may be attributed to the higher amount of absorbers resulting from a wider hematite nanoflake because of the morphology-controlled process, as shown in Figures and 3.

Figure 4

(a) Linear sweep voltammetry for P-H, Sn-H, SnO2-H, and MSnO2-H with chopped light under 1 sun condition. (b) IPCE of P-H, Sn-H, SnO2-H, and MSnO2-H measured at 1.50 VRHE. (c) Chronoamperometry measurements of P-H, Sn-H, SnO2-H, and MSnO2-H at 1.50 VRHE under illumination (λ = 800 nm). (d) Charge transfer efficiency of P-H and MSnO2-H calculated from the IMPS measurement.

To further clarify the photoresponses of P-H and MSnO2-H, the photocurrent was measured at the applied potential bias of 1.50 VRHE and particularly under different wavelengths of light illuminating the samples. The results are shown in Figure S8a,b. Although the photoresponse of P-H was negligible for the incident wavelengths higher than 620 nm, which is consistent with its band gap energy (2.0 eV) of hematite, the photoresponse of MSnO2-H extended to the wavelength of 800 nm. As reported elsewhere, the localized Fe2+ 3d derived states are generated above the top of the valence band (VB) of hematite when the extra electron from Sn4+ converts a neighboring Fe3+ to a localized Fe2+.[32,43] The Fe2+-localized gap states in Sn-doped hematite have been observed by using a combination of XPS VB spectra and DFT + U calculation.[44] With increasing Sn contents in Fe2O3, the number of localized Fe2+ 3d gap states increases—this is in line with the previous literature.[44] The extended absorption of MSnO2-H can be ascribed to the newly created localized Fe2+ 3d states caused by Sn doping between the VB and CB of hematite, which can absorb photons of smaller energy to generate photoexcited EHPs, which can participate in the PEC reaction.[32,45]

To elucidate the role of the Fe/SnO2 NPs as a passivation layer, the charge transfer (Kt) and surface recombination (Kr) rate constants of P-H and MSnO2-H are calculated by using IMPS responses, as shown in Figure S9. The calculation for P-H (Figure S9b) shows that Kr slightly decreases until 1.4 VRHE and significantly decreases at 1.5 VRHE, while the Kt remains constant over the entire potential. Although the Kt of MSnO2-H shows a similar trend with Fe2O3 NF after 1.0 VRHE, the Kr of the MSnO2-H shows a steep negative slope from 0.9 VRHE, which indicates reduced recombination, as shown in Figure S9c. Figure d shows the charge transfer efficiency [ηtr, defined as Kt/(Kt + Kr)], which is involved in the charge transfer reaction as a function of potential for P-H and MSnO2-H.[6,46] The ηtr of P-H is lower than 0.2 until bias potential of 1.3 VRHE, which means that most of the holes are trapped by surface states of hematite, and it is barely increased to 0.6 at a bias potential of 1.5 VRHE. The ηtr of SnO2-H (Figure S10) and MSnO2-H is higher than that of the P-H after 1.0 VRHE, which demonstrates that Fe-doped SnO2 NPs on the surface of hematite help facile charge transfer from the space charge region to the electrolyte by effectively decreasing the recombination resulting from passivated surface states of the hematite. This indicates that tiny Fe-doped SnO2 NPs (1.0–8.1 nm) can be a superior passivation layer on hematite.

Last but not the least, to elucidate the charge transfer mechanism between the surface of the hematite and electrolyte, electrochemical impedance spectroscopy (EIS) was carried out in the frequency range of 100 kHz to 0.1 Hz at 1.50 VRHE of applied bias and with 369 nm light source.[15] Nyquist plots for the P-H, Sn-H, SnO2-H, and MSnO2-H were fitted by using the equivalent circuit model shown in the inset image of Figure . The equivalent circuit consisted of the series resistance of the electrochemical cell, namely Rs, the trapping/detrapping resistance of electrons, R1, the space charge capacitance of the bulk hematite, C1, the charge transfer resistance between the electrolyte and the surface of the hematite, R2, and the space charge capacitance at the interface between the electrolyte and the surface of the hematite, C2. The Rs values for every sample show similar values, indicating the same experimental condition in the electrochemical measurement.[47] Since the Sn4+ ions by diffusion from bottom FTO glass can effectively increase the conductivity of hematite, the R1 and R2 value of Sn-H (3,550 and 9,449 Ω) is lower than that of the P-H (4,608 and 12,380 Ω), as shown in Figure and Table S2.[48] After the decoration of the SnO2 NPs on Fe2O3 NF by an ex situ SnCl4 solution treatment, the R1 and R2 value of MSnO2-H is decreased to 649.3 and 3,510 Ω, respectively. This indicates that the SnO2 NPs on hematite NF act as a passivation layer, which strongly suppresses the surface recombination owing to the reduced surface defect states, causing the high charge recombination and inefficient water oxidation by photogenerated holes.

Figure 5

Nyquist plots of P-H, Sn-H, SnO2-H, and MSnO2-H under a 369 nm LED source at 1.50 VRHE. The inset image is an equivalent circuit model for fitting the experimental data.

Taking into consideration XPS, band gap calculation, photoresponse, and IMPS measurement, the band diagram of P-H and MSnO2-H can be described, as illustrated in Figure a,b, respectively. When P-H absorbs the light, the electron–hole pairs (EHPs) are generated. Although the EHPs, which are generated in the depletion region, can be flowed to the electrolyte to participate in the water splitting reaction, the photoexcited holes in P-H are captured by surface states, resulting in recombination at the surface of P-H, as aforementioned. Meanwhile, the EHPs in the MSnO2-H can be easily extracted from the electrolyte without recombination due to the passivation layer consisting of Fe/SnO2 NPs. Furthermore, a high concentration of Sn doping can generate a localized Fe2+ 3d state above the VB, which can reduce the band gap, as described in Figures k and 6b. The Sn doping level at the interface of Fe/SnO2 and hematite is higher than that of the plain surface of MSnO2-H, as shown in EDS analysis (Figure S4b,c). Since the energy level of the localized Fe2+ 3d state is variable by the Sn doping amount, the interface between Fe/SnO2 and hematite could be the site for absorbing the long wavelength (800 nm).[44] This demonstrates that ex situ doping by a high concentration of Sn4+ ions on hematite not only creates the passivation layer consisting of Fe/SnO2 NPs but also reduces the band gap, which can generate additional EHPs at a long wavelength.

Figure 6

Schematic illustration of the energy band diagram and charge transfer process in (a) P-H and (b) MSnO2-H.

In conclusion, we explored shape-preserved hematite nanoflakes with Fe-doped SnO2 NPs fabricated by a simple ex situ decoration method to produce an enhanced PEC performance. The SnO2 NPs on hematite nanoflakes not only sustain the original shape of pristine hematite nanoflakes but also act as a passivation layer. Furthermore, MSnO2-H showed greatly reduced recombination and presented a photocurrent current density of 1.08 mA/cm2 at 1.50 VRHE, with a photoresponse even at a long wavelength (800 nm). Our strategy demonstrates the simple way to reduce the EHP recombination of hematite NF and extend the capability of light absorption of hematite for the PEC reaction to the near-infrared region.

Experimental Section

Preparation of HNF and Sn-HB Samples

The iron films with a thickness of 600 nm were deposited on FTO glass by using the magnetron sputtering technique, as reported in our previous paper.[38] To grow the hematite NFs on the iron film, the prepared iron film/FTO samples were heated at 400 °C for 2 h in air using a furnace. The SnCl4/ethanol solution was prepared by mixing ethanol/SnCl4 = 200:1 (v/v). Then, 20 μL of the SnCl4 ethanol solution was dropped on the hematite NFs/FTO and dried in air for 20 min. The dried SnCl4 treated sample was annealed at 500 °C in the furnace for 1 h and then cooled down to room temperature. The sample was annealed once again at 800 °C for 15 min to prepare MSnO2-H. For SnO2-H (without the morphology control), the annealing procedure at 500 °C was skipped. In the case of Sn-H, the hematite NF was placed in the furnace for 15 min at 800 °C without SnCl4 treatment.

Characterization of the Samples

The morphology of the samples was investigated using SEM (Hitachi, S-4800). The EDS analysis was carried on by using a TEM machine (HRTEM, Titan G2). XRD was performed with an X’pert Philips MPD (equipped with a Panalytical X’celerator detector) with graphite monochromatic Cu Kα radiation (λ = 1.54056 Å). The chemical composition and the VBM were characterized by XPS (PHI 5600), and the peak positions were calibrated according to the C 1s peak at 284.8 eV.

Photoelectrochemical Measurements

The three-electrode system, which consists of a Pt mesh and an Ag/AgCl (3 M KCl) electrode as the counter electrode and Ag/AgCl reference electrode, respectively, was used to measure the photoelectrochemical performances of the samples under AM 1.5 illumination in 1 M KOH electrolyte. Photocurrent density was measured based on the linear sweep voltammetry method by using a potential range from −0.5 to 0.7 V at a scan rate of 2 mV s–1. All potentials were converted to the reversible hydrogen electrode (RHE) using the following equation: ERHE = 0.210 + EAg/AgCl + 0.059 pH, where EAg/AgCl is the experimentally measured potential. The incident photocurrent conversion efficiency measurements were carried out in the wavelength range from 300 to 700 nm at the fixed potential of 1.23 VRHE and in the 1 M KOH electrolyte. The EIS data were obtained at 1.23 VRHE in the frequency range from 100 kHz to 0.1 Hz with a 369 nm L.E.D light source. The IMPS analysis was conducted in the range from 1.0 to 1.7 VRHE under 452 nm light illumination at an intensity of 10 mW/cm2.