Effects of intrinsic defects and extrinsic doping on the electronic and photocatalytic properties of Ta₃N₅

Abstract

Ta₃N₅ is a good candidate for an oxygen evolution photocatalyst or a photoanode for a Z-scheme device due to its n-type feature. In the present work, the formation energies and electronic structures of defects contained in Ta₃N₅ are studied by first principles density functional theory in detail. Our results show that the substitution of O for three-coordinated N in Ta₃N₅ possesses a low formation energy and introduces a shallow donor under both N-rich and N-poor conditions, making a major contribution to the n-type conductivity. By investigation of the optical transition levels, we show that the four-coordinated N vacancy in Ta₃N₅ is responsible for the observed 720 nm sub-band gap optical absorption. In addition, for alkali metal doped Ta₃N₅, our results reveal that the interstitial doping can lead to enhanced conductivity and reduced band gap, and the doping of Na and K in Ta₃N₅ are expected to produce higher photocatalytic activity compared to Rb and Cs. These results are useful to understand the recent experimental observations and provide guidance to engineer Ta₃N₅ with improved photocatalytic efficiency.

---

1 Introduction

The direct splitting of water into oxygen and hydrogen by semiconductor photocatalysts has been proposed as a clean and renewable approach to resolve the global energy and environment crisis.^1–3 However, most of the photocatalytic materials possessing the ability to split water allow only UV optical absorption, which greatly restricts their energy conversion efficiency. Aiming at improving the utilization of solar energy, thus, numerous studies have focused on semiconductor materials possessing narrow band gaps.^4–7 It is unanimously accepted that an ideal semiconductor material for the photoelectrode should have a suitable band gap of about 2.0 eV to meet both the dynamic reaction barrier and a wide range of optical absorption. Meanwhile, the conduction band and valence band edges of the semiconductor should straddle the water redox potentials in order to split water without external bias. Therefore, transition metal (oxy)nitrides such as LaTaO₂N^8–10 and Ta₃N₅,^11–14 have attracted massive research interest during the past decades because of the suitable band gap and band edge positions. Despite that great effort has been taken to improve photocatalytic performance, several challenges still remain including stabilizing the surface against photooxidation and improving the quantum yield.

Due to the n-type feature, Ta₃N₅ can be applied as an oxygen evolution photocatalyst or a photoanode for Z-scheme device.¹⁵ Generally, for metal oxides and nitrides such as TiO₂ and GaN, the most common origin of the n-type conductivity is attributed to anion vacancies.¹⁶ But there are a number of exceptions. For example, previous theoretical calculation indicated that O_N anti-site, as a dominant intrinsic defect in TaON, is responsible for its n-type conductivity.¹⁷ On the other hand, some experimental studies reported that O defects as unintentional impurities are included in Ta₃N₅ inevitably during the synthesis even under the anaerobic condition.^18,19 Thus, the oxygen defects may potentially contribute to the n-type conductivity in Ta₃N₅. Previous theoretical calculation has also shown that the formation energy of O substituting for N is smaller than N vacancy.²⁰ However, few theoretical studies focus on the n-type conductive behavior of Ta₃N₅, in particular, the effect of the oxygen defects on this property. It is known that extrinsic doping is an effective approach to improve the n-type conductivity of semiconductor materials, which can enlarge the band bending and facilitate the migration of holes from bulk to surface. Therefore, the quantum yield of Ta₃N₅ can be improved by enhanced n-type conductivity. Indeed, It is previously observed that alkali metal (Na, K, Rb, Cs) doped Ta₃N₅ has enhanced activity for photocatalytic splitting water.^21–23 The main doping effects are a lower band gap and a drastic enhancement in the electrode conduction. Among these materials, Na doped Ta₃N₅ has the highest improvement of photocatalytic activity.²³ Nevertheless, the origin of reduced band gap and enhanced conductivity is still need to further investigate.

In this work, aiming at solving the questions shown above, the formation energies and electronic structures of various intrinsic and extrinsic defect contained Ta₃N₅ are investigated by first principles calculation. Our results show that the substitution of O for three-coordinated N is the most stable among all of the defects in Ta₃N₅ considered here, and is mainly responsible for the n-type conductivity. Despite the different characteristics of electronic structure, both two types of N vacancies have shallow thermodynamic transition level of (0/+1). In addition, the interstitial doping of alkali metals in Ta₃N₅ can produce an enhanced n-type conductivity. Na and K are better donors over Sr and Cs, leading to higher enhancement of electronic conductivity and photocatalytic activity. The reduced band gap of alkali metal doped Ta₃N₅ can be attributed to large lattice distortion induced by interstitial doping.

2 Computational details

The spin-polarized density function theory (DFT) calculations were performed by the Vienna ab initio simulation package (VASP).²⁴ The projector augmented wave (PAW) method was applied for the interactions between the valence electrons and the ionic cores. Generalized gradient approximation (GGA) in the formulation of Perdew–Burke–Ernzerhof (PBE) was employed for exchange-correlation functional.²⁵ The electron wave function was expanded in a basis set of plane waves with a kinetic energy cut off of 500 eV. We simulated the doping effects using a 3 × 1 × 1 repetition of the Ta₃N₅ unit bulk cell with 96-atom, and a 3 × 3 × 3 Monkhorst–Pack k-point set were used to obtain the accurate geometry and electronic structure.²⁶ The lattice parameters and the atomic positions were fully relaxed until that all forces are smaller than 0.01 eV Å⁻¹, and the convergence threshold for self-consistence field iteration was set at 10⁻⁶ eV.

The formation energy of a defect α in charge state q is given by^27,28

H_f(α, q) = E(α, q) − E(bulk, 0) + Σn_iμ_i + qE_f + ΔV

where E(α, q) is the total energy of defects contained Ta₃N₅ supercell in charge state q, E(bulk, 0) is the total energy of a perfect crystal in the same supercell. μ_i (i = Ta, N or O) refers to the atomic chemical potential of the type i atoms, and n_i is the number of atoms of type i added to (n_i < 0) and/or removed from (n_i > 0) the perfect crystal. E_f is the Fermi level referenced to the valence band maximum (VBM). ΔV is the VBM correction term, which can be obtained by aligning the average electrostatic potentials between the defect contained supercell and the defect-free supercell. The electrostatic interactions between the defects resulting from the periodic supercell cannot be neglected.²⁷ Previous theoretical studies have shown that the first-order Makov–Payne correction can be used to removed such spurious electrostatic interaction.^29,30 However, its reliability and generality remain controversial. Here, we correct the defect formation energies by using the extrapolation method, more details are given in Fig. S1 (in ESI†). The thermodynamic transition level of ε_α(q/q⁰) for a given defect, corresponding to the positions of Fermi level at which a given defect changes from charge state q to q⁰, can then be written as

Owing to the all known band gap underestimation of DFT, we carried out a DFT plus on-site repulsion U (DFT+U) scheme with U = 0.06 eV on N 2p states, 3.96 eV on O 2p and 6.47 eV on Ta 5d, since previous theoretical studies demonstrated that it given a reasonable and reliable description for the electron structures of Ta₃N₅.²⁰ Here, our calculated band gap is 2.07 eV, which is in good consistence with the experimental value of 2.1 eV. For alkali metal elements, no + U correction was required.

3 Result and discussion

3.1 Chemical potentials

In generally, the defect formation energies are affected by element chemical potentials during the process of growth. Thus, to avoid the precipitation of element phases, a thermodynamic limit is proposed for the chemical potentials under equilibrium growth conditions.

μ_N ≤ 0, μ_Ta ≤ 0

To keep Ta₃N₅ thermodynamically stable, it is required that μ_Ta3N5 = 3μ_Ta + 5μ_N, the calculated μ_Ta3N5 value is −9.96 eV, which is in consistence with other theoretical values of 9.74 eV (GGA + U)²⁰ and 9.45 eV (hybrid functional method).¹⁸ In order to avoid the formation of secondary phases during the growth of Ta₃N₅, the following conditions should be satisfied

μ_Ta + μ_N < μ_TaN = −1.55 eV,

5μ_Ta + 6μ_N < μ_Ta5N6= −12.57 eV,

As shown in Fig. 1, the calculated accessible ranges of chemical potentials for different elements are plotted in a two-dimensional (μ_Ta and μ_N) plane. The blue line refers to the allowed range of chemical potentials of Ta and N for forming Ta₃N₅. Two specific chemical potential points of A and B, referring to N-rich and N-poor conditions, respectively, are adopted to calculate defect formation energies. For condition A, chemical potentials of μ_Ta and μ_N are −3.32 eV and 0, respectively; while they are −0.44 and −1.73 eV for condition B. Considering that O atoms are included in Ta₃N₅ unavoidably in the process of growth, the secondary phases such as TaON and Ta₂O₅ should also be avoided,

μ_Ta + μ_O + μ_N < μ_TaON = −6.1 eV,

2μ_Ta + 5μ_O < μ_Ta2O5 = −18.18 eV,

Fig. 1 Accessible range of chemical potentials (blue line) under equilibrium growth conditions for Ta₃N₅. A (N-rich condition) and B (N-poor condition) sites are chosen as the representative chemical potentials for the following calculation of defect formation energy.

In Ta₃N_5, there are two types of symmetrically different N, which are denoted as N₃ (the N three-coordinated with Ta) and N₄ (the N four-coordinated with Ta). In this study, the following defects are taken into account: Ta vacancy (V_Ta), N vacancy (V_N3 and V_N4), N interstitial (N_i), Ta interstitial (Ta_i). The cation/anion anti-site defects, such as Ta on the N site and N on the Ta site, are not considered in this study due to their large formation energies. Because of the involvement of O element, the defects of O interstitial (O_i) and substitution of O for N (O_N3 and O_N4) are also contained in our calculations. Additionally, for the alkali metal doping, only cationic substitution and interstitial doping are included in our defect investigation.

3.2 Formation energies and transition levels for intrinsic defects

The formation energies of point defects in Ta₃N₅ as a function of the Fermi level in the two chosen chemical potential environments A and B are given in Fig. 2. Under N rich condition, as shown in Fig. 2a, we can see that O_N3 is the most stable defect in almost the whole range of band gap. N_i is easier to form when the Fermi level is close to the conduction band minimum, and the Fermi level will be pinned at 0.11 eV below the CBM by the charge neutrality between positively and negatively charged defects in the absence of other extrinsic dopants in Ta₃N₅.³¹ In addition, both O_N3 and O_N4 being shallow donors, are stable only in +1 charge state, and O_N4 possesses higher formation energy compared to O_N3 with a difference value of 0.27 eV, which is slightly smaller than previous reported theoretical value of 0.56 eV.¹⁸ Thus, the O_N3 is responsible for the n-type conductivity of Ta₃N₅. The calculated DOS indicates that the defect states induced by O are resonant with the conduction band, as shown in Fig. 3a and b, this is understandable because of the lower energy of O 2p compared to N 2p orbital. Therefore, based on above analysis, the involvement of O_N defects in Ta₃N₅ is extremely advantageous for the improvement of n-type conductivity and can thus lead to the enhanced photocatalytic activity for Ta₃N₅.

Fig. 2 Calculated formation energies of defects in Ta₃N₅ as a function of the Fermi energy under N-rich condition (a), and N-poor condition (b).

Fig. 3 Calculated total and projected density of states for O_N4 (a), O_N3 (b), V_N4 (c) and V_N3 (d). The red dashed lines indicate the Fermi level, which is set to zero.

For the N vacancy defects, as shown in Fig. 2a, it is noted that both V_N3 and V_N4 can be stable in +1, +2 and +3 charge states depending on the position of Fermi level, V_N4 has lower formation energy compared to V_N3 when the Fermi level is high in the band gap, otherwise V_N3 is more stable. In addition, the ionization levels of (0/+1) for both V_N3 and V_N4 are above the CBM, implying shallow donor property. Whereas the (+1/+2) and (+2/+3) levels of V_N3 are located in the band gap with the values of 0.91 and 1.11 eV below the CBM, respectively, while those of 1.36 and 1.48 eV for V_N4. To examine the origin of the shallow (0/+1) ionization levels, the calculated DOS of neutral N vacancy defects is given in Fig. 3c and d. V_N4 defect in Ta₃N₅ creates two localized deep states within the band gap and a delocalized state resonating with the conduction band, thereby this delocalized state will be responsible for the shallow character of (0/+1) ionization level of V_N4 defect.²⁸ To further investigate the defect state distribution, we perform the calculation of decomposed charge density, as plotted in Fig. 4. Indeed we note that upon the removal of N₄ atom in Ta₃N₅, correspondingly three electrons are released, two electrons are trapped by Ta dangling bonds neighboring the N vacancy, while the other one distributes in a large volume, which is characterized by delocalized states. This picture is consistent with the DOS analysis above. On the contrary, the DOS shown in Fig. 3d reveals that V_N3 does not exhibit conventional shallow donor behavior, and three deep localized defect states appear in the band gap. The decomposed charge density shown in Fig. 4b also provides evidence that the three electrons are mainly localized on the Ta dangling bonds around the N₃ vacancy. To investigate how the shallow donor level is produced, the relaxation energy has been calculated since the thermodynamics transition level is not only in relation to the position of electronic states but also to the relaxation energy after electron transition.³² Here, the relaxation energy is defined as the energy difference of V_N defect at +1 charge state between V_N⁺¹ and V⁰_N atomic geometries. As expected, the calculated value for V_N3⁺¹ is 0.37 eV, which is significantly larger than the value of 0.08 eV for V_N4⁺¹. However, this value is still not enough to explain so large difference between the positions of defects states and ionization levels of (0/+1) for V_N3 vacancy, we attribute it to large Coulomb repulsion interaction between these localized electrons. By further examination of the geometry structure of V_N3 defect, it is found that the Ta atoms around the N vacancy move out by about 1.7% after one electron was ionized, it is remarkably larger compared to V_N4, which has only 0.22% relaxation, indicating that a larger geometric relaxation can be realized for V_N3 after one electron is ionized compared to that of V_N4. As mentioned above, the V_N3 and V_N4 defects can contribute to n-type conductivity to some extent due to the shallow nature of (0/+1) level. However, the deep levels of (+1/+2) and (+2/+3) make these defects acting as recombination center in Ta₃N₅, which is harmful for photocatalytic activity.³³

Fig. 4 Calculated band decomposed charge-density for V_N4 (a) and V_N3 (b) with a 0.003 eÅ⁻³ isosurface value.

As shown in Fig. 2a, the interstitial N in Ta₃N₅ is stable only in +1 and −1 charge states with the (−1/+1) transition level at 0.58 eV above the VBM, indicating that it is a negative U defect.³⁴ Therefore, it cannot contribute to the n-type conductivity because it behaves like an acceptor under typical n-type conditions while as a donor under p-type conditions. From the DOS shown in Fig. 5a, we note that N_i introduces two impurity states in the band gap with an occupied one near the VBM and an unoccupied state in middle of band gap. In order to further explore its electronic structure information, the local geometric structure of N_i is given in Fig. 6a. We can see that the most stable N_i structure is the split configuration,³⁵ which refers to the interstitial N atom sharing the same site with a three-coordinated N atom. Based on molecular orbital theory, we propose a model for the local bonding shown in Fig. 7a. Because of the similar energies of the 2p orbitals, the coupling between the two N atoms is so strong that unoccupied σ_u anti-bonding orbital is located above the CBM, and a singly occupied π_g anti-bonding in the band gap. Therefore, these defect states can accept an additional electron from the valance band or donate an electron into the conduction band, explaining both acceptor and donor properties of N_i defect. The interstitial O is stable in neutral charge state across almost whole band gap as shown in Fig. 2a, the donor level (+1/0) is extremely deep, which is at about 2 eV below the CBM. In order to reveal the origin of completely different characteristics between the two interstitial defects, a closer look at the electronic structure is helpful. Fig. 5b shows the absence of defect state in the band gap and the Fermi level just located at VBM. Because of the similar geometry structure as N_i, shown in Fig. 6b, the local bonding structure based on molecular orbital theory are given in Fig. 7b. The lower energy of O 2p compared to N 2p, means that the π_g antibonding orbital of O–N bond is not as strong as N–N bond, resulting in these states within the valance band. A similar split configuration has been reported for interstitial N doped TiO₂ by previous theoretical investigation, however, the transition levels of (−1/+1) are located at 1.54 eV and 1.98 eV above VBM for anatase and rutile, respectively.³⁵ We attribute such difference to that the VBM of Ta₃N₅ is higher in energy compared to TiO₂.

Fig. 5 Calculated local density of states of N–N atom pair in interstitial N doped Ta₃N₅ (a) and O–N atom pair in interstitial O doped Ta₃N₅ (b).

Fig. 6 Local geometric structures of interstitial N (a) and interstitial O (b) doped Ta₃N₅.

Fig. 7 The schematic level diagram of N–N atom pair in interstitial N doped Ta₃N₅ (a) and O–N atom pair in interstitial O doped Ta₃N₅ (b).

The interstitial Ta is stable in +1, +2, +3, +4 and +5 charge states depending on the position of Fermi level, and the transition levels of (+1/+2), (+2/+3), (+3/+4) and (+4/+5) are at 0.05, 0.55, 0.67 and 1.11 eV below the CBM, respectively. Ta_i can donate one electron to the conduction band spontaneously. Nevertheless, the large formation energy implies low concentration of this defect in Ta₃N₅, thus making slight contribution to its n-type conductivity. Similarly, owning to the extremely large formation energy, Ta vacancy is unlikely to form in Ta₃N₅ during growth and we will not discuss it in details.

Under N poor condition, as shown in Fig. 2b, the positions of thermodynamic transition levels are essentially not changed. However, the formation energies of donor defects are significantly lower compared to that in N rich condition, while the formation energies of the acceptors are increased. O_N3 has the lowest formation energy even in whole range of band gap and no acceptor can compensate for it, implying that a stronger n-type conductivity can been obtained during synthesis under this condition. In addition, the formation energies of donor defects such as V_N and Ta_i, whose donor levels are deep in the band gap, decrease with downshift of the N chemical potential, which can increase the sub-band gap light absorption of Ta₃N₅. Indeed, the increase of sub-band gap absorption and charge mobility has been observed from the synthesized Ta₃N₅ samples with rising nitridation temperature from 850 °C to 1000 °C.³⁶

The previous experimental studies have revealed a 720 nm optical absorption by UV-Vis optical absorption spectra measurement, the origin of which was assigned to the reduced Ta⁵⁺ ions.¹⁸ The investigation taken by A. Dabirian has implied that nitrogen vacancy is the most probably candidate for the 720 nm light absorption.³⁷ To provide an assignment for these optical absorption spectra, the optical transition level is calculated since it has been shown to give very good description for experimental optical absorption.^35,38,39 Herein we only consider the N-rich condition since consistent conclusion can be obtained under the two growth conditions. As discussed before, the Fermi level locates at 0.11 eV below the CBM, thus, the stable charge states of both V_N4 and V_N3 are +1, while for N_i and Ta_i they are −1 and +2, respectively. O_N and O_i are unlikely to cause this optical absorption and thus not included in our calculations. Based on the Franck–Condon principle, we can calculate the possible optical transitions associated with these defects. To explain the transitions process qualitatively, for V_N4, the configuration coordinate diagram is shown in Fig. 8a.³⁹ Here V_N4⁺¹ is considered as the ground state while V_N4⁺² as the excited state. As discussed before, the thermodynamic transition levels is 1.36 eV below the CBM. Considering an electron transition from defect state to the conduction band after absorption of a photon, the charge state of V_N4 will be changed from +1 to +2, while the atomic geometry is still kept in that of +1 charge state because the light absorption process is far more quickly than the corresponding lattice relaxation. The optical transition level (+1/+2) is defined as the crossing point of the formation energies between V_N4⁺¹ charge state in V_N4⁺¹ atomic geometry and V_N4⁺² charge state in V_N4⁺¹ atomic geometry, as shown in Fig. 8b. Our calculated optical transition energy is 1.71 eV, which is in excellent agreement with the experimental report of 1.72 eV.^18,37 The defect relaxation energy, defined as the difference between optical transition energy and thermodynamic transition energy, is thus about 0.35 eV. Moreover, the optical transition energies for other defects are also calculated. For V_N3, our calculated value is 1.38 eV. It means that the two types of N vacancy have large difference in sub-band gap optical absorption. For N_i, by considering optical transition: N_i⁻¹ → N⁰_i, the calculated value is 2.07 eV, while the optical transition of Ta_i⁺² → Ta_i⁺³ for Ta_i is 0.89 eV. Therefore, our results consistently show that the optical transition V_N4⁺¹ → V_N4⁺² is responsible for the experimental observed 720 nm optical absorption.

Fig. 8 Configuration coordinate diagram for V_N4 defect (a), E_opt refers to the energy of optical transition. The formation energy as a function of Fermi level for V_N4 defect that can occur in two charge states: +1 and +2 (b).

3.3 The doping effects on photocatalytic properties of Ta₃N₅

Finally, to explain the origin of enhanced conductivity and reduced band gap of Ta₃N₅ through doping of alkali metals, their corresponding formation energies and electronic structure are also investigated. Since the growth temperatures of Ta₃N₅ are commonly high and the alkali metal nitrides are expected to dissociate at such high temperatures, the relevant secondary phases can be excluded. Thus, only the metal phases are adopted for their chemical potential calculation. To examine the origin of the enhanced conductivity, firstly, the formation energies of alkali metals (Na, K, Rb, Cs) doped Ta₃N₅ as a function of Fermi level are given in Fig. S2.† Because the formation energies of cationic substitution doping are found to be remarkable high, only interstitial doping are shown herein. Indeed, we note that the thermodynamic transition levels for all of the alkali metal interstitial defects are very shallow and these defects are stable only in +1 charge state, indicating their excellent donor properties. Thus, these results provide a basis for elucidating the enhanced conductivity of alkali metal doped Ta₃N₅ revealed by experiment studies.^21–23 Moreover, the defect formation energies satisfy this order: Na > K > Rb > Cs. It implies that Na and K have larger solubility in Ta₃N₅, leading to more excellent conductivity compared to others, which is in good agreement with the experimental observation.²³ Due to the larger ions radius of Rb and Cs compared to Na and K, the incorporation of these elements can lead to larger lattice distortion in Ta₃N₅, thereby, the higher formation energies are required.

To further investigate the cause of the reduced band gap, the DOS of alkali metal doped Ta₃N₅ are shown in Fig. S3.† Due to higher energies of the s orbitals of alkali metal compared to Ta 5d orbital, the alkali metal states make almost no contribution to band edges, one excess electron is therefore easily donated into the conduction band of Ta₃N₅, narrating the shallow donor nature of these defects. The calculated band gap of Na, K, Rb and Cs doped Ta₃N₅ are 2.02, 2.01, 1.96 and 1.85 eV, respectively. Compared with pure Ta₃N₅, the incorporation of Na and K ions results in slight reduction of band gap, while Cs doping significantly reduce the band gap. Which is close to the experimentally reported range of 1.7–1.8 eV.²³ Since the ionic radius of Cs atom is the largest among these alkali metal elements, we point out that the lattice distortion induced by doping is expected to affect the electronic structure and reduce the band gap of Ta₃N₅. On the basis of the above analysis, higher concentrations of Na and K doping are considered by incorporating one or two atoms in 2 × 1 × 1 (64 atom) Ta₃N₅ supercell, corresponding to 1.6 and 3.1% atomic concentrations, respectively. Such high doping concentrations are reasonable since Na and K doped Ta₃N₅ have lower defect formation energies. The calculated band gap are 1.94 and 1.97 eV for 1.6% atomic concentration of Na and K, respectively, 1.60 and 1.72 eV for 3.1% atomic concentration, which are in excellent agreement with experimental reports.^22,23

In general, the prepared Ta₃N₅ shows n-type conductivity and is often used for splitting water to produce oxygen. The improvement of conductivity is an effective strategy to enhance the activity of photocatalytic materials. For example, previous experimental studies suggested that the substitution of Mo or W at V sites in BiVO₄ could lead to remarkable improvement of conductivity and promote internal charge separation.^40,41 As shown in this paper, the interstitial doping of alkali metals in Ta₃N₅ could effectively increase the carrier concentration and reduce its band gap to some extent. Moreover, the incorporation of atoms with smaller ionic radius can lead to higher doping concentration and thus larger enhancement of conductivity. Recently, Ba doped Ta₃N₅ has been synthesized, yielding a maximum solar energy conversion efficiency of 1.5%, the enhanced activity was assigned to the synergistic effects of enhanced interlayer conductivity.⁴² Based on our previous results, thus, doping of alkaline-earth metal, such as Mg, Ca and Sr in Ta₃N₅, may also lead to improvement of electronic conductivity to achieve higher photocatalytic efficiency since they possess smaller ionic radius than Ba. Although Mo and W have one more valence electron than Ta, the energies of their d orbitals are lower than that of Ta.⁴³ Therefore, the substitution of Mo or W for Ta in Ta₃N₅ is not good selections for improving the photocatalytic performance.⁴⁴ Our results can serve as a guidance to engineer Ta₃N₅ for improving photocatalytic efficiencies.

4 Conclusion

In summary, we investigated the formation energies and electronic structures of intrinsic and extrinsic defects contained in Ta₃N₅ by first principles calculations, our results demonstrated that:

(1) O_N3 in Ta₃N₅ is the dopant with the lowest formation energy and excellent donor properties among the studied defects, making the most contribution to its n-type conductivity. In addition, higher carrier concentration can be obtained under N-poor condition; despite the different features of defect states, both V_N3 and V_N4 have a shallow (0/+1) level; N and O interstitial defects in Ta₃N₅ can form a split configuration, N_i possesses both donor and acceptor properties, while O_i only act as a deep donor.

(2) The results of optical transition levels show that the sub-band gap light absorption around 720 nm could be assigned to the optical transition V_N4⁺¹ → V_N4⁺², while V_N3, possessing lower absorption energy and higher defect formation energy, is unlikely to contribute to this range of light absorption.

(3) Alkali metals interstitial doping in Ta₃N₅ can lead to enhanced conductivity due to their shallow donor properties, the reduced band gap can be attributed to the lattice distortion induced by interstitial doping. In addition, due to the lower formation energy of Na compared to others, it can lead to larger improvement in conductivity and photocatalytic activity. The present work may provide an effective strategy to improve the photocatalytic activity of Ta₃N₅.

Acknowledgements

This work is supported by the National Basic Research Program of China (973 program, 2013CB632401), National Science foundation of China under Grant 11374190, 21333006, and 11404187 and 111 Project B13029. We also thank the National Supercomputer Center in Jinan for providing high performance computation.

References

1. A. Kudo and Y. Miseki, Chem. Soc. Rev., 2009, 38, 253–278.
2. X. Chen, S. Shen, L. Guo and S. S. Mao, Chem. Rev., 2010, 110, 6503–6570.
3. H. Tong, S. Ouyang, Y. Bi, N. Umezawa, M. Oshikiri and J. Ye, Adv. Mater., 2012, 24, 229–251.
4. Y. Bi, S. Ouyang, N. Umezawa, J. Cao and J. Ye, J. Am. Chem. Soc., 2011, 133, 6490–6492.
5. F. E. Osterloh, Chem. Soc. Rev., 2013, 42, 2294–2320.
6. P. Wang, B. B. Huang, X. Y. Qin, X. Y. Zhang, Y. Dai, J. Y. Wei and M.-H. Whangbo, Angew. Chem., Int. Ed., 2008, 47, 7931–7933.
7. X. Ma, Y. Dai and B. B. Huang, ACS Appl. Mater. Interfaces, 2014, 6, 22815–22822.
8. A. Kasahara, K. Nukumizu, G. Hitoki, T. Takata, J. N. Kondo, M. Hara, H. Kobayashi and K. Domen, J. Phys. Chem. A, 2002, 106, 6750–6753.
9. C. L. M. Leroy, A. E. Maegli, K. Sivula, T. Hisatomi, N. Xanthopoulos, E. H. Otal, S. Yoon, A. Weidenkaff, R. Sanjinesd and M. Gratzel, Chem. Commun., 2012, 48, 820–822.
10. W. Luo, Z. Li, X. Jiang, T. Yu, L. Liu, X. Chen, J. Ye and Z. Zou, Phys. Chem. Chem. Phys., 2008, 10, 6717–6723.
11. W. J. Chun, A. Ishikawa, H. Fujisawa, T. Takata, J. N. Kondo, M. Hara, M. Kawai, Y. Matsumoto and K. Domen, J. Phys. Chem. B, 2003, 107, 1798–1803.
12. P. Zhang, J. Zhang and J. Gong, Chem. Soc. Rev., 2014, 43, 4395–4422.
13. C. Zhen, L. Wang, G. Liu, G. Q. Lu and H.-M. Cheng, Chem. Commun., 2013, 49, 3019–3021.
14. J. Wang, J. Feng, L. Zhang, Z. Li and Z. Zou, Phys. Chem. Chem. Phys., 2014, 16, 15375–15380.
15. S. S. K. Ma, K. Maeda, T. Hisatomi, M. Tabata, A. Kudo and K. Domen, Chem.–Eur. J., 2013, 19, 7480–7486.
16. B. J. Morgan and G. W. Watson, J. Phys. Chem. C, 2010, 114, 2321–2328.
17. S. Y. Chen and L. W. Wang, Appl. Phys. Lett., 2011, 99, 222103.
18. E. Nurlaela, S. Ould-Chikh, M. Harb, S. D. Gobbo, M. Aouine, E. Puzenat, P. Sautet, K. Domen, J.-M. Basset and K. Takanabe, Chem. Mater., 2014, 26, 4812–4825.
19. S. J. Henderson and A. L. Hector, J. Solid State Chem., 2006, 179, 3518–3524.
20. J. Wang, T. Fang, L. Zhang, J. Feng, Z. Li and Z. Zou, J. Catal., 2014, 309, 291–299.
21. S. S. K. Ma, T. Hisatomi, K. Maeda, Y. Moriya and K. Domen, J. Am. Chem. Soc., 2012, 134, 19993–19996.
22. Y. Kado, R. Hahn and C.-Y. Lee, Electrochem. Commun., 2012, 17, 67–70.
23. Y. Kado, C. Y. Lee, K. Lee, J. Muller, M. Moll, E. Spiecker and P. Schmuki, Chem. Commun., 2012, 48, 8685–8687.
24. G. Kresse and J. Furthmüller, Comput. Mater. Sci., 1996, 6, 15–50.
25. J. P. Perdew, J. A. Chevary, S. H. Vosko, K. A. Jackson, M. R. Pederson, D. J. Singh and C. Fiolhais, Phys. Rev. B: Condens. Matter Mater. Phys., 1992, 46, 6671–6687.
26. H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Solid State, 1976, 13, 5188–5192.
27. C. G. Van de Walle and J. Neugebauer, J. Appl. Phys., 2004, 95, 3851–3879.
28. X. Ma, Y. Dai, L. Yu and B. Huang, Sci. Rep., 2014, 4, 3986.
29. W. J. Yin, S. H. Wei, M. M. Al-Jassim, J. Turner and Y. F. Yan, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 83, 11.
30. S. Chen and L.-W. Wang, Appl. Phys. Lett., 2011, 99, 222103.
31. S. Lany and A. Zunger, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 78, 235104.
32. H. Shi and M.-H. Du, Phys. Rev. B: Condens. Matter Mater. Phys., 2014, 90, 174103.
33. P. Reunchan, S. Ouyang, N. Umezawa, H. Xu, Y. Zhang and J. Ye, J. Mater. Chem. A, 2013, 1, 4221–4227.
34. M. E. Grillo, S. D. Elliott and C. Freysoldt, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 83, 085208.
35. J. B. Varley, A. Janotti and C. G. Van de Walle, Adv. Mater., 2011, 23, 2343–2347.
36. B. A. Pinaud, A. Vailionis and T. F. Jaramillo, Chem. Mater., 2014, 26, 1576–1582.
37. A. Dabirian and R. van de Krol, Appl. Phys. Lett., 2013, 102, 033905.
38. J. L. Lyons, A. Janotti and C. G. Van de Walle, Phys. Rev. Lett., 2012, 108, 156403.
39. C. Freysoldt, B. Grabowski, T. Hickel, J. Neugebauer, G. Kresse, A. Janotti and C. G. Van de Walle, Rev. Mod. Phys., 2014, 86, 253–305.
40. W. J. Jo, J. W. Jang, K. J. Kong, H. J. Kang, J. Y. Kim, H. Jun, K. P. Parmar and J. S. Lee, Angew. Chem., Int. Ed., 2012, 51, 3147–3151.
41. D. K. Zhong, S. Choi and D. R. Gamelin, J. Am. Chem. Soc., 2011, 133, 18370–18377.
42. Y. B. Li, L. Zhang, A. Torres-Pardo, J. M. Gonzalez-Calbet, Y. H. Ma, P. Oleynikov, O. Terasaki, S. Asahina, M. Shima, D. Cha, L. Zhao, K. Takanabe, J. Kubota and K. Domen, Nat. Commun., 2013, 4, 2566.
43. J. Osorio-Guillén, S. Lany and A. Zunger, Phys. Rev. Lett., 2008, 100, 036601.
44. J. Feng, D. Cao, Z. Wang, W. Luo, J. Wang, Z. Li and Z. Zou, Chem.–Eur. J., 2014, 20, 16384–16390.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra08563b

---