C3N4 heterojunction photoanode

C3N4 heterojunction photoanode

Electrochimica Acta 324 (2019) 134844 Contents lists available at ScienceDirect Electrochimica Acta journal homepage: www.elsevier.com/locate/electa...

3MB Sizes 1 Downloads 211 Views

Electrochimica Acta 324 (2019) 134844

Contents lists available at ScienceDirect

Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta

Enhanced charge separation and interfacial charge transfer of InGaN nanorods/C3N4 heterojunction photoanode Zhenzhu Xu a, b, Shuguang Zhang a, b, c, Fangliang Gao a, b, c, Peng Gao d, Yuefeng Yu a, b, Jing Lin a, b, Jinghan Liang c, Guoqiang Li a, b, c, * a

State Key Laboratory of Luminescent Materials and Devices, South China University of Technology, Guangzhou, 510640, China Engineering Research Center on Solid-State Lighting and its Informationisation of Guangdong Province, South China University of Technology, Guangzhou, 510640, China c Department of Electronic Materials, School of Materials Science and Engineering, South China University of Technology, Guangzhou, 510640, China d Tianjin Institute of Power Sources, Tianjin, 300384, China b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 26 May 2019 Received in revised form 19 August 2019 Accepted 7 September 2019 Available online 11 September 2019

Semiconducting heterostructures designed with rational engineering of energy bands and interfaces can accelerate electron-hole separation to boost photoelectrochemical (PEC) water splitting. Herein, InGaN nanorods (NRs)/C3N4 heterojunction photoanode has been constructed by directly loading C3N4 on the InGaN NRs surface through a simple chemical vapor deposition method. The working principles and interfacial charge kinetics of the heterojunction have been proposed. The unique heterojunction exhibits efficient charge separation through the potential gradient and enhanced interfacial charge transfer due to the surface passivation. Eventually, the photocurrent density of the InGaN NRs/C3N4 heterojunction photoanode with loading weight ratio of 0.38% reaches up to 13.9 mA/cm2 at 1.23 V vs. RHE under an illumination of ~100 mW/cm2, which is 2 times higher than that of the pristine InGaN NRs. The applied bias photon-to-current efficiency of the designed heterojunction can achieve as high as 2.26% at 0.9 V vs. RHE, 1.65 times higher than the bare InGaN NRs (1.37%). Moreover, the InGaN NRs/C3N4 heterojunction exhibits an obviously improved stability against photocorrosion due to the efficient interfacial charge transfer. This work can open up a novel route for the rational design and construction of heterojunction based photoelectrode to readily enhance the PEC performance. © 2019 Elsevier Ltd. All rights reserved.

Keywords: InGaN nanorods C3N4 Heterojunction Charge separation Interfacial charge transfer

1. Introduction Energy and environment issues have been important topics of research interests and raised worldwide concerns [1]. Solarpowered photoelectrochemical (PEC) water splitting is a renewable approach to directly store abundant solar energy in the form of energy-rich (143 MJ/kg) and carbon-free hydrogen fuel, which is an attractive route for addressing the rising global energy demand and environmental issues [2e4]. The oxidation of water to O2 on the photoanode is a key process in the direct PEC production of fuels, but far from efficient due to the complicated and sluggish fourelectron transfer process [5,6]. The development of novel and efficient photoanodes is a key objective in PEC research. Many

* Corresponding author. State Key Laboratory of Luminescent Materials and Devices, South China University of Technology, Guangzhou, 510640, China. E-mail address: [email protected] (G. Li). https://doi.org/10.1016/j.electacta.2019.134844 0013-4686/© 2019 Elsevier Ltd. All rights reserved.

metal oxides, such as TiO2 [7], WO3 [8], Fe2O3 [9], and BiVO4 [10], have received immense attention as photoanodes for PEC water splitting. However, those metal oxides suffer from limited light absorption or/and severe electron-hole pair recombination [11,12], which limit their overall photoelectron conversion efficiency. As an alternative, one dimensional (1D) InGaN nanorods (NRs) is the ideal material whose bandgap energy can be tuned from 3.4 to 0.7 eV by increasing the In composition [13], thus allowing highly efficient utilization of solar spectrum. Additionally, the valence-band edge of InGaN always lies below the redox potential of O2/H2O, and the conduction-band edge crosses the redox potential of Hþ/H2 with an In content less than 50% [14]. Additionally, 1D InGaN NRs possess large charge carrier mobility and effective surface-to-volume ratio, as well as tunable bandgap and satisfied band-edge positions, are the ideal photoanode material for PEC hydrogen production. Recently, most of researches are concentrated on enhancing the light harvesting of the InGaN NRs based photoanode through bandengineering [15,16], nanostructuring strategies [17], two-electrode


Z. Xu et al. / Electrochimica Acta 324 (2019) 134844

tandem system [18], integrated multi-junction photoelectrodes [19]. However, according to JH2 O ¼ Jmax  habs  hsep  htrans [20], the practical water oxidation photocurrent (JH2 O ) is limited by not only the theoretical maximal photocurrent (Jmax ), the light absorption efficiency (habs ) but also the charge separation (hsep ) and surface charge transfer efficiencies (htrans ) of InGaN NRs based photoanodes. In PEC water splitting, the fast recombination of photo-generated carriers and large interfacial charge-transfer resistance seriously impede the performance of photoanodes [21,22]. To achieve highly efficient PEC water spitting, a crucial issue is the development of novel InGaN NRs based photoanodes with efficient photogenerated charge separation and interfacial charge transfer. It is well known that the heterojunction structure can provide additional driving force for the separation of charge carriers via built-in electric field generated in the junction, such as p-n junction, Schottky junction, etc. [23e27]. However, achieving efficient charge separation via heterojunction for InGaN NRs based photoanodes remains largely unexplored, which may ascribe to the fact that the assembling and coupling of other semiconductors on the 1D NRs with high density is technologically challenging [28]. Therefore, it is necessary to develop a simple and general process for fabricating efficient combined system. Among various semiconductors, the chemically stable and earth-abundant carbon nitride (C3N4) nanosheets are the ideal coupling material for construction of InGaN NRs based heterojunction in the PEC water splitting [29e34]. Its valence band is more positive than water oxidation potential, and both conduction band and valence band are more negative than those of InxGa1-xN NRs with x smaller than 0.44 [35,36]. Therefore these two semiconductors form staggered relative band positions as required to form an effective heterojunction photoanode. In this work, we report for the design and construction of InGaN NRs/C3N4 heterojunction photoanode for PEC water splitting. C3N4 were readily deposited on the surface of InGaN NRs by a simple chemical vapor deposition method, which is a general and facile method for loading C3N4 on arbitrary targeted substrates. The deposition of C3N4 significantly promotes the PEC performance of the heterojunction based photoanode. The working principle of the InGaN NRs/C3N4 heterojunction as well as interfacial carrier kinetics has been investigated. The enhanced PEC performance of InGaN NRs/ C3N4 photoanode is attributed to the effective charge separation between the heterojunction and enhanced charge transfer efficiency in the semiconductor-electrolyte interface.

2. Experimental 2.1. Preparation of InGaN NRs The InGaN NRs were grown on n-type Si(111) substrate with resistivity of 0.01e0.02 U m by radio-frequency plasma-assisted molecular beam epitaxy (MANTIS). Prior to the growth, the Si(111) substrate was chemically treated in diluted HF (10%) to remove native oxide and blown dry with high-purity (7N) nitrogen gas. Subsequently, the substrates were transferred into the highvacuum MBE growth chamber, and annealed at 900  C for 30 min to remove the oxidized surface layer and reconstruct the surface. The substrate temperature was monitored by using a thermocouple that was calibrated by a pyrometer. For the growth of InGaN NRs, the temperatures of Ga and In K-cell are set at 930  C and 810  C, respectively. The impinging metal fluxes were measured in terms of beam equivalent pressure (BEP) by an ion gauge. In the experiment, the InGaN NRs were grown at a substrate temperature of 780  C. The Ga-BEP and In-BEP were determined to be 3.96  108 Torr, 4.92E7 Torr, respectively. The forward power of the radio-

frequency plasma was 400 W, the N2 flux was 2.0 sccm. The total growth time was 3 h. 2.2. Preparation of InGaN NRs/C3N4 C3N4 loaded InGaN NRs were prepared by one-step chemical vapor deposition. The InGaN NRs on Si substrate (1.8 cm  1.0 cm) was put on top of the melamine powders (0.08 g), which was placed in a closed porcelain crucible. The samples were heated at 550  C for 10 min in N2 atmosphere with a heating rate of 10  C/ min. At the high temperature of 550  C, the melamine molecules or/ and the intermediates impinge on the surface of InGaN NRs. Then they attach to the surface of the NRs by the formation of hydrogen bonds due to their abundant amine groups. Thermal condensation of melamine or/and the intermediates led to the formation of C3N4 on the surface of InGaN NRs. In order to obtain different loading content of C3N4 on InGaN NRs surface, different amounts of melamine (0.08, 0.09, 0.10, and 0.12 g) are employed. And the obtained samples are named as InGaN NRs/C3N4. 2.3. Preparation of pure C3N4 The graphitic carbon nitride (g-C3N4) was prepared by polymerization of melamine molecules. In detail, melamine powder (1 g) was put in a crucible with a cover, and heated at 550  C for 10 min under N2 condition with a heating rate of 10  C/min. After cooling down naturally to room temperature, the light yellow C3N4 powder was obtained. 2.4. Structural and compositional characterization The morphology and structure of the samples were characterized by field emission scanning electron microscopy (SEM, NOVA NANOSEM 430) and transmission electron microscopy (TEM, JEOL 3000F). For the TEM analysis, the NRs and NRs/C3N4 were scraped from the Si substrates, dispersed in a small volume of ethanol, followed by dripping the resultant NR-containing solution on a Cu grid with carbon supporting film. Elemental line scans were obtained by an energy dispersive X-ray spectrometry (EDX), which was taken on the JEOL 3000F microscope. The chemical compositions and valence states of the samples were investigated by X-ray photoelectron spectra (XPS) acquired on an ESCALab 250 Xi electron spectrometer. Powder X-ray diffraction (XRD) pattern was measured with a Bruke D8 Advance powder X-ray diffractometer with Cu Ka radiation. In addition, the light absorption characteristics of the samples were studied by an UVeVis spectrophotometer with an integrating sphere (Lambda 950, PerkinElmer). 2.5. Photoelectrochemical measurements The NRs- and NRs/C3N4-based photoelectrodes were prepared by depositing Ti/Au (20/80 nm) layer onto the back of each sample, and annealing rapidly at 400  C in N2 for 30 s to form an ohmic contact. Then, a metal wire was bonded on the contact, and the whole contact area was protected with an insulating epoxy resin. For the preparation of C3N4-based photoelectrode, 5 mg of C3N4 powders was dispersed in 1 mL of 1:1 (v/v) water/ethanol mixed solvent along with 10 mL of Nafion solution. After ultrasonically dispersing for 15 min, 5 mL of the above solution was dropt onto the surface of a glassy carbon disk electrode, which was used as the working electrode. PEC measurements were carried out with an electrochemical workstation (CHI 760C, CH Instruments Inc.) in a 0.5 mol/L H2SO4 aqueous solution. The photocurrents of each sample were

Z. Xu et al. / Electrochimica Acta 324 (2019) 134844

measured in a three-electrode PEC cell with as-grown sample as the photoanode, a graphite electrode as the counter electrode, a Hg/ Hg2Cl2 electrode (saturated KCl) as the reference electrode, and a 300 W Xe lamp (intensity ~100 mW/cm2) as a light source, respectively. All the line sweep voltammetry (LSV) curves were carried out at a scan rate of 50 mV/s. Electrochemical impedance spectroscopy (EIS) measurements were taken at an AC frequency ranging from 0.01 to 105 Hz with an amplitude of 5 mV under illumination. Before the PEC measurements, reference electrode was calibrated with respect to a reversible hydrogen electrode (RHE) by performing cyclic voltammetry scans. The calibration was performed at a scan rate of 2.0 mV/s, using Pt wire, Pt nanosheet, and Hg/Hg2Cl2 electrode as working electrode, counter electrode, and reference electrode, respectively. The average of the two potentials where the current crossed zero was taken as the thermodynamic potential for the hydrogen electrode reactions. In 0.5 mol/L H2SO4 solution, potentials versus the Hg/Hg2Cl2 electrode were converted to RHE scale by the equation ERHE ¼ EHg2Cl2þ0.22 V. 3. Results and discussion 3.1. Construction of InGaN NRs/C3N4 heterojunction The schematic illustration for construction of InGaN NRs/C3N4 heterojunction on Si substrate is shown in Scheme 1. The InGaN NRs were grown on Si(111) substrate by plasma-assisted molecular beam epitaxy (details in the experimental part). Such 1D NRs will enhance the light absorption by the extension of the light propagation path, and reduce the travelling distance for photo-generated carriers to the semiconductor/electrolyte interface [37,38]. Moreover, the significantly enhanced surface area of NRs is beneficial for mass and charge transport in the semiconductor/electrolyte interface. Then, the 2D C3N4 nanosheets were deposited on the upper part of InGaN NRs via a simple chemical vapor deposition method with melamine as precursor at 550  C in N2 atmosphere (details in the experimental part). Briefly, the melamine molecules or/and the intermediates was impinged onto the surface of InGaN NRs as temperature rose, and attached to the surface of the NRs by the formation of hydrogen bonds due to their abundant amine groups [39]. Subsequently, the C3N4 formed a coating layer on the InGaN NRs surface by thermal condensation of melamine or/and the intermediates. Using this simple and general method, C3N4 nanosheets could be easily deposited on arbitrary low dimensional materials including 0D nanoparticles, 1D nanowires/nanorods, and 2D foils. This method paves a way to construct nanostructured


heterojunction. To assert the derived carbon material is indeed C3N4, melamine powders alone were thermally treated at 550  C in N2 atmosphere. The structure of the as-obtained product was examined by XRD technique. As shown in Fig. S1, the as-obtained product displays two diffraction peaks at 13.1 and 27.5 , originating from in-plane repeating tri-s-trizaine ring structure and interlayer stacking of conjugated aromatic units, respectively [40]. These two characteristic peaks are consistent with the previous results, suggesting the successful synthesis of C3N4 [40e42]. According to XPS measurement, the surface C/N molar ratio of as-prepared C3N4 is estimated to be 0.7, which is lower than the theoretical ratio of 0.75 for C3N4. This is ascribed to the existence of dangling NH2 groups in the asprepared C3N4 [42]. To investigate the effect of loaded C3N4 amount on PEC water splitting performance of the InGaN NRs photoanode, InGaN NRs/ C3N4 containing C3N4 in various amounts were prepared by varying the amount of melamine (details in the experimental part). As shown in the SEM images (Fig. S2), the amount of C3N4 loaded on InGaN NRs is obviously increased by increasing the supplied amount of the precursor simply. In order to determine the contents of loaded C3N4, the thermogravimetry (TG) detection was employed from room temperature to 800  C in N2 atmosphere. The TG plots in Fig. S3 indicate that the pure InGaN NRs are stable through the temperature ranging from 28 to 800  C in N2 atmosphere. For pure C3N4 powder, a rapid weight loss is observed at temperature higher than 550  C in N2 condition. A 100% weight loss is obtained at 800  C, indicating complete decomposition of C3N4. Therefore, for the InGaN NRs/C3N4, the flat regions of the weight loss at temperature from 700 to 800  C suggest that the loaded C3N4 is completely decomposed. According to the TG plots, InGaN NRs/C3N4 with different C3N4 wt % of 0.27, 0.38, 0.53, and 2.85 were prepared by increasing the amount of precursor. The morphology and structure of the pristine InGaN NRs and InGaN NRs/C3N4 heterojunction were illustrated by SEM and TEM images. The SEM images (Figs. S4(a) and (b) show that the pristine InGaN NRs are vertically grown on Si substrate with a high cover density of ~213 mm2. The mean length of the NRs is ~246 nm with an average diameter of ~51 nm. The selected area electron diffraction (SAED) pattern of the representative (In)GaN NR suggests that the NR is single crystalline (Fig. S4(c)), and the high-resolution (HR) TEM images display clear and perfect lattice fringes of the NR (Fig. S4(d)). Fig. S5 shows a typical XRD pattern of the aligned InGaN NRs grown on Si(111) substrate. The peak at 34.1 is attributed to (0002) hexagonal plane of InGaN, indicating the epitaxial growth of InGaN NRs along the [0001] direction. All the results reveal a

Scheme 1. Illustration of the synthetic procedures for depositing C3N4 nanosheets on InGaN NRs. The orange, grey balls denote N, C atoms, respectively.


Z. Xu et al. / Electrochimica Acta 324 (2019) 134844

Fig. 1. Top-view (a) and side-view (b) SEM image of InGaN NRs/C3N4 (0.53 wt %) on Si substrate. (c) Typical TEM image of the InGaN NR/C3N4 (0.53 wt %) under low magnification; the inset is the SAED pattern taken from the TEM image. (d) HRTEM images of InGaN NRs/C3N4 (0.53 wt %) obtained from the region of red dashed square in (c). (e) STEM image and EDX elemental mappings of In, Ga, C, and N for the representative InGaN NR/C3N4 (0.53 wt %) heterojunction. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

high crystalline quality of the pristine NRs. After the deposition of C3N4 nanosheets, the SEM images (Fig. 1(b) and (c)) show that C3N4 nanosheets are successfully loaded on the upper part of InGaN NRs. To further reveal the detailed structure of the as-prepared InGaN NRs/C3N4, TEM measurements were carried out as well. It can be clearly seen from Fig. 1(c) and (d) that the sidewall of the InGaN NR is covered by C3N4 nanosheets compared to the pristine InGaN NRs (Fig. S4). The SAED pattern in the inset of Fig. 1(c) proves that the NRs retain single-crystalline nature well even after the deposition of C3N4 at 550  C in N2 atmosphere. HRTEM image in Fig. 1(d) shows intimate contact between the InGaN NR and C3N4 nanosheets, which is beneficial for efficient transport of charge in the heterojunction. The interplanar spacing of 0.28 nm corresponds to the two adjacent (0002) lattice planes of InGaN NRs. The C3N4 nanosheets loaded InGaN NRs were further confirmed by scanning transmission electron microscopy (STEM) image and EDX spectrum. Fig. 1(e) shows the STEM image of InGaN NR/C3N4 heterojunction, illustrating the atomic number contrast between InGaN (brighter) and C3N4 (darker) regions. The elemental mappings in Fig. 1(fei) depict that In, Ga, C, and N were well-defined, which further confirm that C3N4 nanosheets are formed on the surface of InGaN NRs. XPS measurements were further performed to identify the surface chemical composition of the InGaN NRs/C3N4. The full XPS spectra of the pure C3N4 and InGaN NRs/C3N4 reveal the existence

of C3N4 in the heterojunction (Fig. S6). Fig. 2(a) shows the XPS spectra of C 1s for InGaN NRs/C3N4 and pure C3N4. There are two distinct peaks at 287.8 and 284.5 eV, which correspond to SP2-hybridized carbon in the N-containing aromatic structure (NeC]N) and graphitic CeC, respectively [43]. The N 1s spectra (Fig. 2(b)) of pure C3N4 exhibits four peaks centered at 398.3, 399.1, 400.5 and 404.5 eV, representing SP2-hybridized aromatic N (CeN]C), tertiary nitrogen NeC3 groups, amino groups (NeH), and p excitations, respectively [43]. In addition to the NeIn bond (395.1 eV) and NeGa bond (396.8 eV) [44], the N 1s spectra of InGaN NRs/C3N4 also exhibits CeN]C, NeC3, CeNeH peak, at binding energies of 398.5, 399.1, 400.5 eV, respectively. The In3d and Ga3d spectra of InGaN NRs/C3N4 are shown in Fig. 2(c) and (d), respectively. The main peaks centered at 443.7 and 451.3 eV in the In3d spectrum can be attributed to IneN bonds and In3d3 level (Fig. 2(c)) [45,46], and the two peaks centered at 17.2 and 19.3 eV in Ga3d spectrum can be related to In 4d level and GaeN bond (Fig. 2(d)) [46], respectively. The traces of the characteristic peaks of C3N4 in the heterojunction indicate the successful interaction of InGaN surface with the C3N4. 3.2. PEC performance of InGaN NRs/C3N4 heterojunction Before PEC performance, the light absorbance ability of all samples was studied by UVeVis diffuse reflectance spectroscopy, as shown in Fig. S7. The InGaN NRs on Si substrate show a broad light

Z. Xu et al. / Electrochimica Acta 324 (2019) 134844


Fig. 2. High-resolution XPS spectra of C1s (a), N1s (b), In3d (c), and Ga3d (d) for C3N4 and NRs/C3N4 with loading weight ratio of 0.53%.

absorption ranging from 200 to 600 nm, and the pure C3N4 powder has strong absorption under 450 nm. After loading C3N4 with amount smaller than 0.53%, the optical absorption of NRs/C3N4 heterojunction has no obvious change in comparison with pristine NRs. However, the absorption intensity of the NRs/C3N4 heterojunction is decreased obviously, with the loaded C3N4 amount reaching up to 2.85%. This is probably due to the excessive covering of the NRs surface by C3N4 nanosheets. These results demonstrate that the heterojunction can retain the excellent light absorption ability except that the loading amount of C3N4 is excessive. The PEC performance of the different photoanodes was investigated by linear-sweep voltammetry (LSV) and chronoamperometry. More details are provided in the experimental section. Fig. 3(a) shows the LSV curves of different photoanodes in dark and under illumination. Compared to the pristine InGaN NRs, the photocurrent density of the heterojunction is enhanced significantly with increasing C3N4 loading until 0.53%, and then decreased slightly while increases the C3N4 loading up to 2.85%. The photocurrent density of NRs/C3N4 (0.38%) heterojunction at 1.23 V vs. RHE reaches up to 13.9 mA/cm2 under an illumination of ~100 mW/cm2, which is about 2 times higher than that of bare InGaN NRs (6.6 mA/cm2). For a reference, the bare Si substrate and C3N4 based photoanode show a low photocurrent density under illuminated conditions (Fig. S8), which indicates that the intrinsic Si substrate and C3N4 has poor performance for the OER. In order to test whether thermal treatment of the InGaN NRs during the deposition process of C3N4 plays a role in increasing the photocurrent or not, the LSV curve of the NRs after thermal treatment in 550  C was measured. As shown in Fig. S9, no obvious change between the LSV curves of InGaN NRs with and without thermal treatment is observed. These results indicate that the deposition of C3N4 nanosheets on InGaN NRs surface indeed plays an important

role in enhancing the PEC performance. However, the decrease in the photocurrent density of NRs/C3N4 (2.85%) heterojunction may be attributed to the decreased light absorption (Fig. S7) induced by excessive covering of C3N4. Fig. 3(b) exhibits the amperometric I-t curves of the photoanodes in several light on-off cycles. Photocurrent reflects that all the photoanodes are photo-sensitive and have the ability to generate charge carriers under irradiation. The photoconversion efficiency (h) was estimated by applied bias photon-tocurrent efficiency (ABPE) using a two-electrode PEC cell (details in the calculation section of supporting information). Fig. 3(c) shows the variation of h as a function of the applied potential between working electrode and Pt electrode. The maximum value of h for the bare InGaN NRs electrode is about 1.37% at 0.8 V vs. RHE, whereas the maximum efficiency of NRs/C3N4 (0.38%) heterojunction is as high as 2.26% at 0.9 V vs. RHE under the same conditions. Furthermore, the stability of the bare NRs and NRs/C3N4 (0.38%) heterojunction based-photoanode was evaluated by the photocurrent density-time curves at a fixed applied bias of 0.6 V vs. RHE. As shown in Fig. 3(d), the photocurrent of pristine NRs remains 41% of its initial value after working for 2 h at 0.6 V vs. RHE, whereas that of NRs/C3N4 heterojunction remains 67% of its initial value. This result proves the enhanced stability against photocorrosion after the deposition of C3N4 nanosheets. After storing the InGaN NRs and NRs/C3N4 (0.38%) heterojunction based-electrodes under ambient condition for two weeks, the photocurrent densities of them are comparative to those of fresh samples (Fig. S10), which is promising for practical application. All of the above results reveal that the deposition of C3N4 nanosheets can improve the PEC performance of InGaN NRs significantly. To elucidate the mechanism underlying the enhanced PEC performance of the InGaN NRs/C3N4 electrode, energy band structure of the InGaN NRs/C3N4 heterojunction was studied. The bandgap


Z. Xu et al. / Electrochimica Acta 324 (2019) 134844

Fig. 3. (a) Linear sweep voltammetry of different photoanodes in 0.5 mol/L H2SO4 solution, measured under dark and an illumination of 100 mW/cm2. (b) The chronoamperometry of different photoanodes with light on-off cycles while performed at a fixed applied potential of 0.6 V vs. RHE. (c) Photo-conversion efficiency of InGaN NRs and NRs/C3N4 (0.38 wt %) as a function of applied potential vs. Pt. (d) Photoanode stability of InGaN NRs and NRs/C3N4 (0.38 wt %) heterojunction at 0.6 V vs. RHE.

energies of bulk C3N4 and pristine InGaN NRs were measured by Tauc plots of (ahn)n versus hn derived from the absorption spectra. The values of n are determined to be 1/2 and 2 for the InGaN NRs with direct transition nature and C3N4 with indirect transition nature, respectively [47]. As shown in Fig. 4(a), the bandgap energy of as prepared InGaN NRs is estimated to be about 3.01 eV, which is consistent with the In content of 10% for the InGaN NRs determined by SEM-EDS measurement (not shown here). The bandgap of the pure C3N4 is estimated to be about 2.8 eV, which is close to the reported results [35,48]. To determine the positions of the valence band (VB) edges of the InGaN NRs and C3N4, the valence band XPS spectra were measured. As shown in Fig. 4(b), the VB energy levels of the InGaN NRs and C3N4 are estimated to be ~1.67 and ~1.42 eV below the Fermi level. To further reveal the relative band structures of the InGaN NRs and C3N4, the flat-band potentials were estimated from Mott-Schottky (M-S) plots (details in the calculation section of supporting information). As shown in Fig. 4(c), the slope of the linear part of the M-S curve for the InGaN NRs is positive, implying the n-type nature of the NRs [49], which results from the unintentional doping of the NRs by Si. Interestingly, the M-S plot shows a p-type characteristic (negative slope) for the C3N4 bulk. By extrapolating the linear part of the M-S curves to zero, the flat-band potentials of the InGaN NRs and pure C3N4 were determined to be 0.1 and 0.4 V vs. RHE, respectively. Together with the band gaps (Fig. 4(a)), the conduction band (CB) and VB energy levels of InGaN NRs are calculated to be 1.44 and 1.57 V vs. RHE, while the CB and VB energy levels of C3N4 are calculated to be 0.98 and 1.82 V vs. RHE. The relative band structures for the n-type NRs and p-type C3N4 are determined as shown in the left part of Fig. 4(e). After depositing p-type C3N4 nanosheets onto n-type InGaN NRs, a p-n junction is formed in the interface of NRs/C3N4 heterojunction. Within the interface, the Fermi levels of the NRs and C3N4 are aligned to reach a new equilibrium by electron transferring,

resulting in the formation of a built-in potential and therefore the band bending [50]. Considering the very close flat band potentials of the NRs and NRs/C3N4 heterojunction (Fig. 4(c) and (d)), schematic energy band diagram of the NRs/C3N4 heterojunction is determined as shown in the right part of Fig. 4(e). As can be seen, a potential gradient is formed between two semiconductors. The potential gradient drive the injection of electrons from the C3N4 overlayer to InGaN NRs and the transfer of holes out of the InGaN NRs, allowing the effective separation of charges and promoting the ABPE of the heterojunction. The overall PEC performance is determined by several processes, such as light absorption, charge separation, and the charge transfer in semiconductor/electrolyte interface. In addition to the charge separation efficiency, the interfacial charge-transfer efficiency was investigated using electrochemical impedance spectroscopy (EIS). The resulting Nyquist plots for the pristine InGaN NRs and different NRs/C3N4 photoanodes are shown in Fig. 5(a). The inset shows the equivalent circuit diagram, where Rs, CPE, and Rct are the resistance of the electrode or the electrolyte, the capacitance phase element, and the charge transfer resistance across semiconductor/electrolyte interface, respectively [51,52]. It is known that the diameter of semicircle in a Nyquist plot is equal to the interfacial charge transfer resistance of Rct [53]. As can be seen, the NRs/C3N4 photoanodes show a smaller Rct as compared to the pristine NRs, implying a lower charge transfer resistance and higher charge transfer efficiency in the semiconductor/electrolyte interface. To quantify the influence of C3N4 on the enhanced interfacial charge transfer, the hole injection rate was determined by using the methanol as hole scavenger (details in calculation section of the supporting information). As shown in Fig. 5(b), the hole injection yield of the InGaN NRs/C3N4 (0.38%) could reach 84% at 1.23 V vs. RHE, which is higher than that of bare NRs (57%), demonstrating that the formation of InGaN NRs/C3N4 heterojunction could efficiently accelerate the

Z. Xu et al. / Electrochimica Acta 324 (2019) 134844


Fig. 4. (a) Tauc plots of the InGaN NRs and the C3N4 powder derived from absorption spectra. (b) Valence band XPS spectra of InGaN NRs and C3N4 powder. (c) Mott-Schottky plots of the InGaN NRs and C3N4. (d) Mott-Schottky plots for the NRs/C3N4 (0.38%) heterojunction electrode. (e) Energy band diagrams of InGaN NRs, C3N4, respectively, and schematic energy band diagram of the InGaN NRs/C3N4 heterojunction under equilibrium condition.

Fig. 5. (a) Nyquist plots and equivalent circuit of the pristine InGaN NRs and different NRs/C3N4 photoanodes measured at 0.6 V (vs. RHE) in a 0.5 mol/L H2SO4 aqueous solution under illumination. (b) Hole injection efficiency of the InGaN NRs and typical NRs/C3N4 (0.38%) photoanodes.


Z. Xu et al. / Electrochimica Acta 324 (2019) 134844

interfacial charge transfer and injection. It has been reported that the accumulation of photo-generated holes may causes serious photocorrosion to the photoanode [54]. The higher interfacial charge transfer efficiency explains the improved stability of the heterojunction. It has been reported that the trap states on the surface of original semiconductor may accumulate photogenerated carriers and hinder the charge transfer [25]. The photovoltage decay rate is a useful method to detect the presence of trap states in the semiconductors [55]. Since a long relaxation time for the charge transferring from the trap states, a slow delay indicates a high density of trap states. The photovoltage-time spectra for pristine InGaN NRs and NRs/C3N4 (0.38%) photoelectrodes were collected, as shown in Fig. 6. The decay lifetime of photovoltage when the light off can be calculated by fitting the spectra using an exponential function with two time constants [55,56]:

yðtÞ ¼ A0 þ A1 et=t1 þ A2 et=t2

tm ¼ ðt1 t2 Þ=ðt1 þ t2 Þ where tm is the harmonic mean of the lifetime, and the total halflife is log(2tm). Table S1 in the supporting information lists the two time constants (t1, t2) of the two photoanodes in detail. The total half-lifes of pristine InGaN NRs and NRs/C3N4 (0.38%) photoanodes

were calculated to be 0.21 and 0.15 s, respectively. The rapider photovoltage decay of the NRs/C3N4 (0.38%) heterojunction suggests that the deposition of C3N4 nanosheets could possibly reduce surface trap states via surface passivation, and alleviate the trapping of holes, therefore accelerate the holes transfer from electrode surface into electrolyte. On the basis of above results, the working principles for the enhanced PEC performance of InGaN NRs/C3N4 heterojunction photoanode have been proposed, as shown in Scheme 2. Firstly, the InGaN NRs and C3N4 nanosheets are excited under UVeVisible light, and generate holes and electrons in the valence band and conduction band, respectively. The potential gradient in the heterojunction interface drives the injection of electrons from the C3N4 overlayers to the NRs and the transfer of holes out of the NRs. This means that the deposition of C3N4 nanosheets could facilitate the charge separation and extract the photogenerated holes efficiently, allowing most of the holes to participate in the water oxidation reaction and thus improving the efficiency of the photoanode. Moreover, the deposition of C3N4 nanosheets may decrease the number of trapping sites on the surface of InGaN NRs by surface passivation, which can reduce hole transfer resistance and enhance the holes transferring from NRs surface to electrolyte. The efficient interfacial hole transfer contributes to the increased photocurrent as well as the stability of the heterojunction against photocorrosion due to the depression of hole accumulation. To sum up, the synergistic effect of enhanced charge separation and interfacial charge transfer improve the PEC performance of the InGaN NRs/C3N4 heterojunction photoanode significantly. 4. Conclusions

Fig. 6. Open circuit photovoltage decay (OCVD) spectra of the InGaN NRs and the NRs/ C3N4 (0.38%) photoanodes.

In conclusion, we have for the first time successfully designed and fabricated the InGaN NRs/C3N4 heterojunction photoanode by a simple chemical vapor deposition method. The general and facile method paves a way for loading C3N4 nanosheets on arbitrary targeted substrates, especially those with a low dimensional structure. Resultantly, the NRs/C3N4 heterojunction photoanode achieves 2 times higher photocurrent density than the pristine InGaN NRs, and reaches a higher ABPE of 2.26% at 0.9 V vs. RHE. A detailed energy band diagram and EIS analysis reveal that the improved photo-conversion efficiency results from the efficient charge separation forced by the potential gradient in the heterojunction, and the enhanced interfacial charge transfer through the surface passivation. Moreover, the NRs/C3N4 heterojunction based photoanode exhibits an obviously improved photostability due to the suppression of the hole accumulation by efficient interfacial hole transfer. This work will provide insights into the design and

Scheme 2. Schematic illustration of the accelerated charge-separation and interfacial charge transfer processes of the InGaN NRs/C3N4 heterostructure photoanode.

Z. Xu et al. / Electrochimica Acta 324 (2019) 134844

construction of heterojunction based photoelectrode to enhance the PEC water splitting efficiency. Acknowledgements This work was supported by National Science Fund for Excellent Young Scholars of China (No. 51422203), National Natural Science Foundation of China (No. 51572091, 51002052 and 51372001), National Science Fund for Young Scholars of China (No. 61404051), Natural Science Foundation of Guangdong Province, China (2018A030313395), Equipment pre-research fund key projects of China (No. 6140721010102), Joint Fund of Ministry of Education for Equipment Pre-Research (6141A02022435), Key Innovation Project in Industry Chain of Shaanxi Province (2018ZDCXL-GY-01-02-01). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.electacta.2019.134844. References [1] [2] [3] [4] [5] [6] [7] [8] [9]

[10] [11] [12]


[14] [15] [16] [17] [18] [19] [20]

N. Armaroli, V. Balzani, Chem. Eur J. 22 (2016) 32e57. Q. Huang, Z. Ye, X. Xiao, J. Mater. Chem. A 3 (2015) 15824e15837. M. Ye, J. Gong, Y. Lai, C. Lin, Z. Lin, J. Am. Chem. Soc. 134 (2012), 15720-1572. J.X. Feng, J.Q. Wu, Y.X. Tong, G.R. Li, J. Am. Chem. Soc. 140 (2018) 610e617. S. Hu, M.R. Shaner, J.A. Beardslee, M. Lichterman, B.S. Brunschwig, N.S. Lewis, Science 344 (2014) 1005e1009. J.A. Seabold, K. Zhu, N.R. Neale, Phys. Chem. Chem. Phys. 16 (2014) 1121e1131. I.S. Cho, C.H. Lee, Y. Feng, M. Logar, P.M. Rao, L. Cai, D.R. Kim, R. Sinclair, X. Zheng, Nat. Commun. 4 (2013) 1723. S.Q. Yu, Y.H. Ling, J. Zhang, F. Qin, Z.J. Zhang, Int. J. Hydrogen Energy 42 (2017) 20879e20887. K. Zhang, T. Dong, G. Xie, L. Guan, B. Guo, Q. Xiang, Y. Dai, L. Tian, A. Batool, S.U. Jan, R. Boddula, A.A. Thebo, J.R. Gong, ACS Appl. Mater. Interfaces 9 (2017) 42723e42733. Y. Park, K.J. McDonald, K.S. Choi, Chem. Soc. Rev. 42 (2013) 2321e2337. L. Badia-Bou, E. Mas-Marza, P. Rodenas, E.M. Barea, F. Fabregat-Santiago, S. Gimenez, E. Peris, J. Bisquert, J. Phys. Chem. C 117 (2013) 3826e3833. €rtsch, C. F D. Monllor-Satoca, M. Ba abrega, A. Genç, S. Reinhard, T. Andreu, J. Arbiol, M. Niederberger, J.R. Morante, Energy Environ. Sci. 8 (2015) 3242e3254.  _ B. Sebeka, J. Juodkazyte, I. Savickaja, A. Kadys, E. Jelmakas, T. Grinys, S. Juodkazis, K. Juodkazis, T. Malinauskas, Sol. Energy Mater. Sol. Cells 130 (2014) 36e41. P.G. Moses, C.G. Van de Walle, Appl. Phys. Lett. 96 (2010) 21908. B. AlOtaibi, H.P.T. Nguyen, S. Zhao, M.G. Kibria, S. Fan, Z. Mi, Nano Lett. 13 (2013) 4356e4361. M. Gopalakrishnan, S. Gopalakrishnan, G.M. Bhalerao, K. Jeganathan, J. Power Sources 337 (2017) 130e136. Y.J. Hwang, C.H. Wu, C. Hahn, H.E. Jeong, P. Yang, Nano Lett. 12 (2012) 1678e1682. B. AlOtaibi, S. Fan, S. Vanka, M.G. Kibria, Z. Mi, Nano Lett. 15 (2015) 6821e6828. S. Fan, B. AlOtaibi, S.Y. Woo, Y. Wang, G.A. Botton, Z. Mi, Nano Lett. 15 (2015) 2721e2726. H. Dotan, K. Sivula, M. Gr€ atzel, A. Rothschild, S.C. Warren, Energy Environ. Sci.


4 (2011) 958e964. [21] Z. Chen, T.F. Jaramillo, T.G. Deutsch, A. Kleiman-Shwarsctein, A.J. Forman, N. Gaillard, R. Garland, K. Takanabe, C. Heske, M. Sunkara, E.W. McFarland, K. Domen, E.L. Miller, J.A. Turner, H.N. Dinh, J. Mater. Res. 25 (2010) 3e16. [22] W. Li, P. Da, Y. Zhang, Y. Wang, X. Lin, X. Gong, G. Zheng, ACS Nano 8 (2014) 11770e11777. [23] G. Liu, K. Du, S. Haussener, K. Wang, ChemSusChem 9 (2016) 2878e2904. [24] J.S. DuChene, G. Tagliabue, A.J. Welch, W. Cheng, H.A. Atwater, Nano Lett. 18 (2018) 2545e2550. [25] D. Jeon, N. Kim, S. Bae, Y. Han, J. Ryu, ACS Appl. Mater. Interfaces 10 (2018) 8036e8044. [26] Y. Yu, J. Geng, H. Li, R. Bao, H. Chen, W. Wang, J. Xia, W. Wong, Sol. Energy Mater. Sol. Cells 168 (2017) 91e99. [27] M. Karimi-Nazarabada, E.K. GoharshadiSolar, Sol. Energy Mater. Sol. Cells 160 (2017) 484e493. [28] B.A. Pinaud, J.D. Benck, L.C. Seitz, A.J. Forman, Z. Chen, T.G. Deutsch, B.D. James, K.N. Baum, G.N. Baum, S. Ardo, Energy Environ. Sci. 6 (2013) 1983e2002. [29] W. Ong, L. Tan, Y.H. Ng, S. Yong, S. Chai, Chem. Rev. 116 (2016) 7159e7329. [30] H. Ou, L. Lin, Y. Zheng, P. Yang, Y. Fang, X. Wang, Adv. Mat. 29 (2017) 1700008. [31] D.J. Martin, K. Qiu, S.A. Shevlin, A.D. Handoko, X. Chen, Z. Guo, J. Tang, Angew. Chem. Int. Ed. 53 (2014) 9240e9245. [32] Y. Li, Y. Xue, J. Tian, X. Song, X. Zhang, X. Wang, H. Cui, Sol. Energy Mater. Sol. Cells 168 (2017) 100e111. [33] S. Kumar, S. Karthikeyan, A. Lee, Catalyst 8 (2018) 74. [34] M.Q. Wen, T. Xiong, Z.G. Zang, W. Wei, X.S. Tang, F. Dong, Opt. Express 24 (2016) 10205e10212. [35] J. Liu, Y. Liu, N. Liu, Y. Han, X. Zhang, H. Huang, Y. Lifshitz, S.T. Lee, J. Zhong, Z. Kang, ChemInform 46 (2015) 970e974. [36] M.G. Kibria, Z. Mi, J. Mater. Chem. A 4 (2016) 281e282. [37] Y. Hou, X. Yu, Z.A. Syed, S. Shen, J. Bai, T. Wang, Nanotechnology 27 (2016) 455401. [38] D. Wang, A. Pierre, M.G. Kibria, K. Cui, X. Han, K.H. Bevan, H. Guo, S. Paradis, A. Hakima, Z. Mi, Nano Lett. 11 (2011) 2353e2357. [39] M. Shalom, S. Gimenez, F. Schipper, I. Herraiz-Cardona, J. Bisquert, M. Antonietti, Angew. Chem. Int. Ed. 53 (2014) 3654e3658. [40] Y. Yuan, W. Xu, L. Yin, S. Cao, Y. Liao, Y. Tng, C. Xue, J. Hydrogen. Energ. 38 (2013) 13159e13163. [41] P. Niu, L. Zhang, G. Liu, H. Cheng, Adv. Funct. Mater. 22 (2012) 4763e4770. [42] Y. Guo, J. Li, Y. Yuan, L. Li, M. Zhang, C. Zhou, Z. Lin, Angew. Chem. Int. Ed. 55 (2016) 14693e14697. [43] Q. Xu, B. Cheng, J. Yu, G. Liu, Carbon 118 (2017) 241e249. [44] A. De, M. Tangi, S.M. Shivaprasad, J. Appl. Phys. 118 (2015) 25301. [45] M. Tangi, J. Kuyyalil, S.M. Shivaprasad, J. Appl. Phys. 114 (2013) 153501. [46] L. Caccamo, G. Cocco, G. Martín, H. Zhou, S. Fundling, A. Gad, M.S. Mohajerani, , F. Peiro , W. Dziony, H. Bremers, A. Hangleiter, M. Abdelfatah, S. Estrade L. Mayrhofer, G. Lilienkamp, M. Moseler, W. Daum, A. Waag, ACS Appl. Mater. Interfaces 8 (2016) 8232e8238. [47] O. Elbanna, M. Fujitsuka, T. Majima, ACS Appl. Mater. Interfaces 9 (2017) 34844e34854. [48] Y. Chen, W. Huang, D. He, Y. Situ, H. Huang, ACS Appl. Mater. Interfaces 6 (2014) 14405e14414. [49] J. Kamimura, P. Bogdanoff, M. Ramsteiner, P. Corfdir, F. Feix, L. Geelhaar, H. Riechert, Nano Lett. 17 (2017) 1529e1537. [50] Z. Zhang, J.T. Yates, Chem. Rev. 112 (2012) 5520e5551. [51] W. Zhou, Y. Zhou, L. Yang, J. Huang, Y. Ke, K. Zhou, L. Li, S. Chen, J. Mat. Chem. A 3 (2015) 1915e1919. [52] W.J. Jo, J. Jang, K. Kong, H.J. Kang, J.Y. Kim, H. Jun, K.P.S. Parmar, J.S. Lee, Angew. Chem. Int. Ed. 51 (2012) 3147e3151. [53] F. Malara, A. Minguzzi, M. Marelli, S. Morandi, R. Psaro, V. Dal Santo, A. Naldoni, ACS Catal. 5 (2015) 5292e5300. [54] S.H. Kim, M. Ebaid, J. Kang, S. Ryu, Appl. Surf. Sci. 305 (2014) 638e641. [55] B. Mukherjee, W. Wilson, V.R. Subramanian, Nanoscale 5 (2013) 269e274. [56] Y. Pu, G. Wang, K. Chang, Y. Ling, Y. Lin, B.C. Fitzmorris, C. Liu, X. Lu, Y. Tong, J.Z. Zhang, Y. Hsu, Y. Li, Nano Lett. 13 (2013) 3817e3823.