Effects of oxygen impurity concentration on the interfacial properties of Ta3N5/Ta5N6 composite photoelectrode: A DFT calculation

https://doi.org/10.1016/j.apcatb.2020.119296Get rights and content

Highlights

  • The oxygen doped Ta3N5/Ta5N6 interface shows strong interfacial connection.

  • The oxygen doped Ta3N5/Ta5N6 interface presents covalent character.

  • The contact between Ta5N6 and oxygen doped Ta3N5 forms the Schottky junction.

  • Oxygen doping increases difficulties of electron transfer from Ta3N5 to Ta5N6.

Abstract

The metallic Ta5N6, which is generated during the preparation of semiconductor Ta3N5, interacts with Ta3N5 and forms the Ta3N5/Ta5N6 composite photoelectrode. The oxygen impurity is a natural defect in Ta3N5 and its concentration is probably correlated with the photoelectrochemical performance of Ta3N5/Ta5N6 composite photoelectrode. In this study, density functional theory calculations were performed to investigate effects of oxygen impurity concentration on the properties of Ta3N5/Ta5N6 interface. The results showed that the covalent bonding character became more obvious at the Ta3N5/Ta5N6 interface with the increase of oxygen concentration, leading to the strong connection between Ta3N5 and Ta5N6. The contact between Ta5N6 and oxygen doped Ta3N5 formed the Schottky junction. The built-in potential at the Ta3N5/Ta5N6 interface gradually increased with the oxygen concentration, suggesting that doping with high oxygen impurity concentration was harmful to electron transfer from Ta3N5 to Ta5N6.

Introduction

Solar hydrogen production via semiconductor-based photoelectrochemical (PEC) water splitting system is a promising strategy to ease energy shortage and environmental crisis, because it supplies an environment-friendly approach for splitting water into H2 and O2 gases under irradiation of solar light [1]. In the last decades, a great deal of effort has been devoted to develop more efficient semiconductors [[2], [3], [4], [5], [6], [7], [8], [9], [10]]. Among the semiconductors investigated to date, the tantalum nitride (Ta3N5) has recently attracted a great deal of interest because of its excellent PEC performances [[11], [12], [13], [14], [15], [16], [17], [18]]. Due to the suitable band-gap (about 2.1 eV) and band-edge positions, the Ta3N5 semiconductor is capable of producing H2 and/or O2 gases under visible light irradiation. The theoretical solar-to-hydrogen conversion efficiency of Ta3N5 is as high as 15.9 % (AM 1.5 G) [19], suggesting that it may be one potential water splitting semiconductor in the industrial applications.

In the PEC system, the n-type Ta3N5 acts as photoanode and is often prepared by two steps [20] (Fig. 1a): firstly, the metal Ta foil is oxidized into Ta2O5; secondly, high temperature nitridation (−850 °C) is performed to convert Ta2O5 into Ta3N5. To improve the crystallinity of Ta3N5, the nitridation temperature is usually increased to nearly 1000 °C [[21], [22], [23]]. It is found that, due to the atomic diffusion under higher temperature (Fig. 1b), the sub-nitrides such as TaN, Ta2N and Ta5N6 may generate on one side of Ta3N5 [[21], [22], [23]]. Then, the final sample is actually a Ta3N5/sub-nitride composite photoelectrode. Compared with the semiconductor character of Ta3N5, these sub-nitrides possess metallic character [24,25]. Therefore, the contact between Ta3N5 and sub-nitrides forms the classic semiconductor/metal junction. Since the sub-nitrides act as substrate and accept photo-generated electrons from Ta3N5, the Ta3N5/sub-nitride interface properties are very critical for the final PEC performances. However, whether the Ta3N5/sub-nitride interface is beneficial or harmful to the PEC performances of Ta3N5 is still controversial. In some work the Ta3N5/sub-nitride composite photoelectrode exhibited improved photocurrent [23,26], while in other work it showed detrimental effects on the PEC performance [27,28].

To our knowledge, the oxygen impurity in Ta3N5 is one possible cause of the controversial performances of Ta3N5/sub-nitride interface. The oxygen impurity, which substitutes nitrogen atom (ON), is the most important defect in Ta3N5. Note that, the practical Ta3N5 sample naturally consists of considerable amounts of ON impurities, which may come from the residual O of Ta2O5 after nitridation. Despite different experimental conditions, the ON impurities only change their concentration but cannot be completely eliminated from Ta3N5 [29,30], because they are beneficial to the structural stability of Ta3N5 [31]. Semiconductor defects, especially the shallow donors (or acceptors), are strongly correlated with carrier concentration, which is a critical factor affecting the semiconductor/metal interfacial properties. For example, when the semiconductor/metal contact is Schottky type, the change of electron concentration in semiconductor may change the Schottky barrier height [32,33]. Since the ON impurity is a shallow electron donor in Ta3N5 [34] and its concentration strongly depends on experimental conditions [29], the controversial PEC performances of Ta3N5/sub-nitride interface may be ascribed to the concentration change of ON impurities.

In this study, based the on the density functional theory (DFT) calculation, we want to unravel effects of ON impurity concentration on the Ta3N5/sub-nitride interfacial properties. Among different sub-nitrides, only the Ta5N6 is selected to construct Ta3N5/Ta5N6 interface because of two reasons. Firstly, the ON impurities are major defects in Ta3N5 but not in sub-nitrides. Therefore, the sub-nitrides we selected in this study should be proved to have negligible ON impurities. To our knowledge, only Ta5N6 has been experimentally [35] and theoretically [36] proved to have negligible ON impurities. Secondly, among different sub-nitrides, the Ta5N6(100) surface size matches very well with the Ta3N5(100) surface size. This is helpful to build an interface model and ensure the reliability of calculation results. To further improve the accuracy of our calculation results, the hybrid functional is adopted in this study. Based on the hybrid DFT calculations, the cohesive energies, electronic structures and built-in potentials will be calculated to make an in-depth understanding of the Ta3N5/Ta5N6 interface.

Section snippets

Computational details and models

The DFT calculations are performed by VASP(5.2) [37] with the projected-augmented-wave (PAW) [38] method. The generalized gradient approximation (GGA) [39] in the scheme of Perdew-Burke-Ernzerhof (PBE) [40] is selected for the exchange-correlation functional. Geometry relaxations are carried out until the residual forces on each ion converge to be smaller than 0.02 eV Å−1. For N, O and Ta elements, the 2s22p3, 2s22p4and 5p66s15d4 orbital, respectively, are treated as valence states. The cutoff

Surface properties

Before investigating interface properties, we firstly discuss surface properties. Since the Ta5N6(100) surface, as well as other low-index Ta5N6 surfaces, have been extensively studied in our previous work [36], we will pay attentions to the Ta3N5(100) surface in this section. Although the ON doped Ta3N5(100) surface has been also investigated in some theoretical work [[51], [52], [53]], to our knowledge, effects of different number of ON impurities on the thermodynamic and electronic

Conclusions

In summary, based on the DFT calculations, we investigated effects of oxygen impurity concentration on the interfacial properties of Ta3N5/Ta5N6 composite photoelectrode. We firstly calculated interface cohesive energies and found that doping with more ON impurities is helpful to strengthen interfacial contact between Ta3N5 and Ta5N6. Secondly, by calculating charge density difference, Bader charge and layer decomposed DOS, we unraveled the covalent bonding character at the Ta3N5/Ta5N6

CRediT authorship contribution statement

Jiajia Wang: Investigation, Writing - original draft. Yaqing Jiang: Conceptualization. Aibin Ma: Supervision. Jinghua Jiang: Conceptualization. Jianqing Chen: Writing - review & editing. Baosong Li: Formal analysis. Jianyong Feng: Formal analysis. Zhaosheng Li: Project administration. Zhigang Zou: Supervision.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work is supported by Fundamental Research Funds for the Central Universities (Grant No. 2016B14314), National Natural Science Foundation of China (Grant No. 21503068) and Key Research and development Project of Jiangsu Province (No. BE2016187).

References (69)

  • J. Feng et al.

    Nano Today

    (2020)
  • F. Raziq et al.

    Appl. Catal. B

    (2018)
  • C. Cai et al.

    Appl. Catal. B

    (2020)
  • F. Raziq et al.

    Appl. Catal. B

    (2020)
  • F. Raziq et al.

    Appl. Catal. B

    (2020)
  • X. Wu et al.

    Nano Energy

    (2020)
  • C. Zhou et al.

    Appl. Catal. B

    (2018)
  • A.B. Murphy et al.

    Int. J. Hydrogen Energy

    (2006)
  • Y.B. Li et al.

    Adv. Mater.

    (2013)
  • S.J. Henderson et al.

    J. Solid State Chem.

    (2006)
  • M.K. Hudait et al.

    Physica B

    (2001)
  • J.D. Houmes et al.

    J. Solid State Chem.

    (1996)
  • J. Wang et al.

    Comput. Mater. Sci.

    (2018)
  • J.J. Wang et al.

    J. Catal.

    (2014)
  • L. Qiao et al.

    Surf. Sci.

    (2011)
  • Z. Wang et al.

    Chem. Sci.

    (2016)
  • A. Fujishima et al.

    Nature

    (1972)
  • Z.S. Li et al.

    Energy Environ. Sci.

    (2013)
  • W.J. Luo et al.

    Energy Environ. Sci.

    (2011)
  • J. Feng et al.

    Adv. Funct. Mater.

    (2019)
  • M. Harb et al.

    J. Phys. Chem. C

    (2020)
  • X. Wang et al.

    J. Phys. Chem. Lett.

    (2020)
  • J.M. Morbec et al.

    Phys. Rev. B

    (2014)
  • T. Jing et al.

    RSC Adv.

    (2015)
  • L. Cui et al.

    J. Phys. Soc. Jpn.

    (2014)
  • C. Zhen et al.

    J. Mater. Chem. A

    (2016)
  • Q.-L. Liu et al.

    Phys. Chem. Chem. Phys.

    (2018)
  • W.-J. Chun et al.

    J. Phys. Chem. B

    (2003)
  • B.A. Pinaud et al.

    J. Phys. Chem. C

    (2012)
  • B.A. Pinaud et al.

    Chem. Mater.

    (2014)
  • T. Nobuzo

    Jpn. J. Appl. Phys.

    (1971)
  • D. Li et al.

    RSC Adv.

    (2014)
  • L. Wang et al.

    Adv. Mater.

    (2016)
  • M. Zhong et al.

    Angew. Chem., Int. Ed.

    (2017)
  • Cited by (7)

    • Synthesis of a Z-scheme ternary photocatalyst (Ta<inf>3</inf>N<inf>5</inf>/Ag<inf>3</inf>PO<inf>4</inf>/AgBr) for the enhanced photocatalytic degradation of tetracycline under visible light

      2022, Journal of Physics and Chemistry of Solids
      Citation Excerpt :

      Ta3N5, a typical metal nitride, received much attention because its narrow band gap (2.1 eV) allowed it to absorb the visible light in a long wavelength range up to 600 nm [8,9]. However, the low charge separation efficiency of Ta3N5 in the photocatalytic processes deteriorated its photocatalytic activity and hindered its practical application [10,11]. One of the most common and effective methods to improve the photocatalytic performance was to construct heterojunction materials [12].

    • A DFT study of the water-splitting photocatalytic properties of pristine, Nb-doped, and V-doped Ta<inf>3</inf>N<inf>5</inf> monolayer nanosheets

      2021, Surfaces and Interfaces
      Citation Excerpt :

      They found that doping Ta3N5 with Nb and using polypyrrole co-catalyst narrowed the bandgap and enhanced hydrogen production rate [44]. The theoretical studies have been mostly focused on the electronic structure of bulk and nanostructures of Ta3N5 and improvement of their photocatalytic activities via engineering bandgap, VBM, and CBM [45-54]. The first theoretical study of the band structure of Ta3N5 was performed by Fang et al. and the values of 1.1 and 1.2 eV were obtained for the bandgap of Ta3N5 using the different methods including the generalized gradient approximation (GGA) implemented in Vienna ab initio simulation program (VASP) and the full potential linearized augmented plane wave (FPLAPW) method by WIEN97 code, respectively [45], which were less than the experimental value (2.1 eV) [30].

    View all citing articles on Scopus
    View full text