Metal-free highly efficient photocatalysts for overall water splitting: C₃N₅ multilayers

Abstract

As a promising means of renewable energy storage, the production of molecular hydrogen and oxygen from photocatalytic water splitting has gained increasing interest. The optimal photocatalyst for water splitting should have high solar energy conversion efficiency and strong photocatalytic redox ability to drive the hydrogen evolution reaction (HER) and oxygen evolution reaction (OER). However, few photocatalysts have been reported to fulfil these two contradictive requirements. Here, we demonstrated from first-principles calculations that the recently synthesized two-dimensional carbon nitride (C₃N₅) multilayers can serve as promising candidates to reach this goal. The intrinsic electric field which is more pronounced in the multilayers alters the band alignment of the photocatalysts, making the HER and OER be driven solely by the photogenerated carriers. The thickness-dependent electronic band gap (2.95–2.16 eV) along with the high carrier mobility broadens the energy range of light absorption and promotes carrier separation and transfer, leading to high solar energy conversion efficiency. Our computational results offer not only low-cost, Earth-abundant and environmentally friendly photocatalysts but also a promising strategy for the design of photocatalysts for highly efficient overall water splitting without using sacrificial reagents.

---

Introduction

With the exploration of energy and environment, solar energy has become one of the focuses of current research as a clean and renewable source of energy that ensures sustainable development of human society.^1–3 Photocatalytic water splitting into H₂ and O₂ that directly converts solar energy into hydrogen fuel has been attracting growing interest since the pioneering work of Fujishima and Honda.⁴ To match the potentials of hydrogen reduction H⁺/H₂ (−4.44 eV) and water oxidation H₂O/O₂ (−5.67 eV), the photocatalyst should have an electronic band gap larger than 1.23 eV.^5,6 But a larger band gap would lead to a lower solar energy conversion efficiency.^7,8 In addition, the photogenerated electrons and holes in the photocatalysts should be capable of fast separation and migration to suppress their recombination; otherwise the energy conversion efficiency will be greatly reduced.^9,10 The lowest solar-to-hydrogen (STH) efficiency that makes photocatalysts economically viable is about 10%.^11,12

Another challenging issue that the current photocatalysts are facing is the need for sacrificial reagents.^13,14 Photocatalytic water splitting contains two half-reactions, i.e., the hydrogen evolution reaction (HER) and oxygen evolution reaction (OER). The redox ability of the photogenerated carriers for the HER (or OER) can be determined by the energy difference between the conduction band minimum (or valence band maximum) and the hydrogen reduction potential H⁺/H₂ (or water oxidation potential, O₂/H₂O). To trigger the overall water splitting, the overpotential of the photogenerated electrons (and holes) should be higher than that of the HER (and OER), otherwise, sacrificial agents are needed to keep the whole reaction running.¹⁵ Notably, improving the redox ability of photogenerated carriers and increasing the energy conversion efficiency are two contradictive goals in a sense.^16,17 High photocatalytic redox ability requires a large band gap, whereas high energy conversion efficiency is always correlated to a narrow band gap to broaden the light absorption range. Therefore, a compromise between the energy conversion efficiency and redox ability is crucial for highly efficient photocatalysts for overall water splitting and becomes the focus of recent studies.¹⁸ For example, on the basis of first-principles calculations, Li et al. predicted that the redox ability of the photogenerated electrons and holes in the PdSeO₃ monolayer is high enough to drive both the HER and OER in water without using sacrificial reagents.¹⁹ Unfortunately, the PdSeO₃ monolayer has a relatively large band gap of ∼2.84 eV, which is unfavorable for high energy conversion efficiency.

Very recently, a new mechanism for photocatalytic water splitting has been proposed and has gained increasing interest.²⁰ In this mechanism, the intrinsic electric field (EF) of the photocatalyst alters the band alignments and breaks the band gap limitation (1.23 eV) for overall water splitting. The EF also accelerates the separation of the photoexcited electrons and holes and suppresses their rapid recombination. Guided by this new mechanism, a number of materials have been proposed for overall water splitting, such as Janus chalcogenide layers,²¹ germanium monochalcogenide monolayer²² and Sc₂CO₂.²³ However, the redox ability for the HER and OER of these photocatalysts has rarely been investigated. Despite the experimental progress in photocatalytic overall water splitting, extensive theoretical studies on this new mechanism are timely and desirable.

Two-dimensional (2D) carbon nitrides (C₃N₄) have been demonstrated to be promising photocatalysts for overall water splitting, due to their structural stability, low cost, easy synthesis, environmental friendliness, etc.^24,25 However, the OER overpotential of the C₃N₄ catalysts is too high to be overcome solely by the voltage generated by the photogenerated electrons.²⁶ Although considerable efforts have been made to reduce the OER overpotential of C₃N₄ catalysts, e.g., by doping transition metal atoms,^27,28 replacing the N atom with a S atom,²⁹ and forming heterostructures with other 2D materials,^30,31 improving the photocatalytic performance of the C₃N₄ catalysts remains a challenging task.

In this work, we demonstrated that the conflict between high broad range light absorption (high energy conversion efficiency) and high photocatalytic redox ability for the HER and OER can be resolved by the intrinsic EF effect. Based on first-principles calculations, we propose a promising candidate 2D material, a C₃N₅ multilayer,³² to realize this mechanism. Our calculations showed that the intrinsic EF in the C₃N₅ multilayer leads to spatial separation of the valence band maximum (VBM) and conduction band minimum (CBM), suppressing the recombination of the photogenerated carriers. Meanwhile, high carrier mobility also facilitates the migration and thus the separation of the carriers. The band alignment of the C₃N₅ multilayer can be modulated by the thickness, enabling a narrow band gap and high redox ability of the photogenerated carriers. Both the HER and OER can proceed, driven solely by the potential of the photogenerated carriers. High STH efficiency with the upper limit up to 12.35% due to broad band solar absorption is predicted in these metal-free carbon nitride photocatalysts. Our theoretical results propose not only a low-cost, Earth-abundant and environmentally friendly photocatalyst but also a general strategy for the design of highly efficient photocatalysts for overall water splitting without using sacrificial reagents.

Computational methods

Our first-principles calculations were performed using the Vienna ab initio simulation package (VASP) within the framework of density functional theory (DFT).³³ The electron–ion interaction was described by the projector-augmented wave (PAW) potential.³⁴ The generalized gradient approximation (GGA) in the form of the Perdew–Burke–Ernzerhof (PBE) functional was adopted for the exchange–correlation functional.^35,36 The van der Waals (vdW) interactions were taken into account by using the DFT-D3 functional, which reproduced well the experimental data of the lattice constants.³⁷ The HSE06 hybrid functional was adopted to give more reliable electronic band structures, band alignment and light absorption properties.³⁸ The Kohn–Sham electron-wave functions were expanded using the plane-wave functions with an energy cutoff of 520 eV. The energy convergence was set to 10⁻⁵ eV and the residual force was set to 0.01 eV Å⁻¹. Brillouin zone (BZ) integration was carried out using 7 × 7 × 1 k-point meshes.³⁹ A vacuum space of more than 20 Å was applied in the z-direction to eliminate interactions between neighboring images. For the electrocatalytic reaction calculations, we used a 2 × 2 supercell to avoid the effects of the adjacent intermediates. The formation energy of the C₃N₅ monolayer was calculated by: E_f = E(C₃N₅) − (mμ_C + nμ_N), where E(C₃N₅) is the total energy of the C₃N₅ monolayer, and m and n are the number of carbon and nitrogen atoms in one unit cell, respectively. μ_C and μ_N are the chemical potentials of carbon and nitrogen atoms, which were obtained from graphene and N₂, respectively. The photon spectra were calculated using the VASP code interface by the Phonopy code.⁴⁰Ab initio molecular dynamics (AIMD) simulations were carried out in the canonical ensemble (NVT) with a time step of 1.0 fs.⁴¹

The Gibbs free energy (ΔG) of the HER and OER was evaluated using the standard hydrogen electrode model proposed by Nørskov et al. for the elemental step, as follows:^42,43

ΔG = ΔE + ΔE_zpe − TΔS + ΔG_U + ΔG_pH

where ΔE represents the absorption energy of the intermediates, and ΔE_zpe and TΔS are the zero-point energy difference and the entropy, respectively. The corresponding data employed in the calculations are listed in Table S1 of the ESI.† ΔG_U = −eU is the relevant electrode potential U. ΔG_pH represents the Gibbs free energy which was corrected to H⁺ concentrations which were taken to be zero for pH = 0 in our calculations.

The OER process in the acidic aqueous solution involves four-electron oxidation steps, which can be written as:⁴⁴

H₂O + * → OH* + H⁺ + e⁻

OH* → O* + H⁺ + e⁻

O* + H₂O → OOH* + H⁺ + e⁻

OOH* → * + O₂ (g) + H⁺ + e⁻

where * denotes the adsorption site, and OH*, O* and OOH* denote the adsorbed intermediates. The HER process considered in this work involves two-electron pathways, including a fast proton/electron transfer step and a fast hydrogen release step:

* + H⁺ + e⁻ → H*

H* + H⁺ + e⁻ → * + H₂ (g)

where H* denotes the adsorbed H⁺ ions.

The STH efficiency of the photocatalyst Q_STH can be evaluated approximately by using the following expression:¹³

where ΔG = 1.23 eV is the potential difference for water splitting, P(hω) is the AM1.5 solar energy flux at the photon energy of hω, E_g is the band gap, and ΔΦ is the vacuum level difference on the two respective surfaces of the C₃N₅ photocatalysts.

Results and discussion

For the non-polar photocatalysts without intrinsic EF, the band gap required for overall water splitting should be larger than 1.23 eV, because both the hydrogen reduction potential (V_H = −4.44 eV) and water oxidation potential (V_O = −5.67 eV) should be located inside the band gap, as shown in Fig. 1(a). In this case, the hydrogen reduction potential and water oxidation potential are aligned with respect to the VBM and CBM according to the same vacuum level which can be determined from the average electrostatic potential of the photocatalysts. Then the redox ability of the photogenerated electrons (or holes) can be evaluated by the energy difference between the CBM (or VBM) and the hydrogen reduction potential, which is denoted as U_e (or U_h), as shown in Fig. 1(a). Here, U_e and U_h correspond to the potential of the photogenerated electrons and holes, respectively. To trigger the HER and OER, they should fulfill the following conditions: U_e ≥ η_HER and U_h ≥ η_OER + 1.23, where η_HER and η_OER represent the overpotentials of the HER and OER. The ideal photocatalyst for water splitting has η_HER = 0 and η_OER = 0, corresponding to the lower limit of band gap of 1.23 eV due to the constraint of E_g = U_e + U_h.

Fig. 1 The schematic representation of the water splitting mechanisms for the photocatalysts (a) without and (b) with the intrinsic EF effect.

For the polar 2D photocatalysts with out-of-plane intrinsic EF, the mechanism for photocatalytic overall water splitting differs significantly from the above case. Let us suppose that the intrinsic EF points from the bottom surface to the top surface, as shown in Fig. 1(b). Driven by the EF, the photogenerated electrons and holes migrate to the bottom and top surfaces of the photocatalysts, ensuring that the HER and OER proceed, respectively, on these surfaces.^45–47 Notably, the vacuum levels are no longer equivalent in the top and bottom regions, but exhibit a difference due to the intrinsic EF effect. The photocatalytic redox ability of the top and bottom surfaces should be evaluated with respect to the vacuum levels of the two regions, as indicated in Fig. 1(b). In this case, the constraint of the band gap becomes E_g = U_e + U_h − ΔΦ. Therefore, high photocatalytic redox ability represented by large U_e and U_h can be achieved in materials with a small band gap.

To realize the above mechanism, we considered the candidate materials, C₃N₅ multilayers, which have been synthesized in recent experiments.³⁴ The C₃N₅ monolayer is composed of triazoles and triazines forming a 2D monoclinic lattice, as shown in Fig. 2. The optimized lattice parameters are a = 5.64 Å, b = 6.68 Å, and γ = 113°, consistent with the experimental data.³² There are four types of nitrogen atoms in the framework: sp³-hybridized nitrogen (N1 and N5), pyridine nitrogen (N2, N4, N6 and N7), pyrrole nitrogen (N8) and graphic nitrogen (N3). The presence of sp³-hybridized nitrogen atoms results in the buckling of the lattice with a buckled height of 1.91 Å. The N–N bond length between the two sp³-hybridized nitrogen atoms (N1–N5) is about 1.40 Å, slightly shorter than that (N1–N8) in the triazole (1.42 Å). The lengths of the C–N bonds are in the range from 1.32 Å to 1.45 Å, which are very close to the values of C₃N₄ (1.33 and 1.39 Å)²⁴ and C₂N (1.34 Å).⁴⁸ The formation energy of the C₃N₅ monolayer (−2.39 eV per atom) obtained from our calculations is very close to that of the C₃N₄ monolayer (−2.47 eV per atom), indicating the energetic stability of the C₃N₅ monolayer. The phonon spectrum and AIMD simulations also confirmed the dynamic and thermodynamic stability of the C₃N₅ monolayer, as shown in Fig. S1.† Although the formation energy of the C₃N₄ monolayer is slightly lower than that of the C₃N₅ monolayer, we didn't find any tendency of conversion from the C₃N₅ monolayer to the C₃N₄ monolayer in our AIMD simulations. Notably, the absence of inversion symmetry and electron transfer between carbon and nitrogen atoms induce out-of-plane electric polarization and intrinsic EF, which are expected to affect the band alignment and relevant electronic properties.

Fig. 2 The relaxed atomic structure of the C₃N₅ monolayer (top (a) and side (b) views). Nitrogen and carbon atoms are represented by blue and brown balls, respectively. The circle gives the atoms in a unit cell and the corresponding bond lengths (Å). (c) The projected density of states of the atomic orbitals of the monolayer.

We determined the energetically most preferable structure of the C₃N₅ bilayer from different stacking patterns and found that it favors a slightly shifted AA-like stacking pattern of monolayers, in good agreement with the results of a previous study.³² Such a slightly shifted AA-like pattern guarantees an improved intrinsic EF effect. Following this stacking pattern, we constructed C₃N₅ multilayers containing two, three, four and five monolayers. The exfoliation energy of a C₃N₅ monolayer from the multilayers was evaluated to be about 13.2 meV Å⁻², which is much smaller than the existing single-layer 2D materials, such as graphene (21 meV Å⁻²), h-BN (32 meV Å⁻²)⁴⁹ and C₃N₄ (36 meV Å⁻²),⁵⁰ as shown in Fig. S2.† These results confirm that the interlayer interactions in the C₃N₅ multilayers are dominated by weak vdW interactions, which verifies the peelability of the multilayer via exfoliation techniques in experiments.

The electronic band structures of the C₃N₅ monolayer and multilayers were then calculated using the HSE06 functional. The electronic band structures of these C₃N₅ materials show that they are semiconductors with an indirect band gap ranging from 2.95 (monolayer) to 2.16 eV (five-layers), as shown in Fig. 3(a)–(e) and Table 1. The VBM resides at M (0.5, 0, 0), whereas the CBM is located between K (0.39, 0.31, 0) and M′ (0, 0.5, 0). From the PDOS shown in Fig. 2(c), one can see that the VBM and CBM are mainly derived from the p_z orbitals of C and N atoms. The intrinsic EF pointing from the bottom surface to the top surface induces the electrostatic potential difference of ΔΦ between the vacuum levels on the two sides of the C₃N₅ nanosheets, as shown in Fig. S3.† The alignment of the hydrogen reduction potential (H⁺/H₂) and the water oxidation potential (O₂/H₂O) was modified correspondingly in the bottom and top regions of the nanosheets, as shown in the right column of Fig. 3(a)–(e). More interestingly, the potential difference of the two sides (ΔΦ) increases with the increases of the thickness of the C₃N₅ photocatalysts, whereas the band gap exhibits an inverse trend, as shown in Fig. 3(f). We attributed this phenomenon to the synergistic effect of intrinsic EF and the quantum confinement effect.⁵¹ Such variations in the trends of ΔΦ and electronic band gap offer a promising approach to optimize the photocatalytic redox ability and energy conversion efficiency in a specific range of thickness. It should be mentioned that the intrinsic EF effect increases with the increase of thickness, which will further reduce the band gap and improve the redox ability of photogenerated carriers. However, this trend will not always continue because it will lead to the divergence of the electric potential which is obviously unphysical. With the increase of the thickness, charge redistribution, e.g., electron transfer along the inverse direction of the intrinsic electric field in response to the band structure modification, will converge the electric field effect. For the bulk C₃N₅ crystal that contains an infinite number of monolayers, our calculations showed that it is a semiconductor with a band gap of 2.26 eV, as indicated in Fig. S3(f).†

Fig. 3 The band structures (left column) and band alignments (right column) with surface potential difference of (a) monolayer, (b) bilayer, (c) trilayer, (d) four-layer and (e) five-layer C₃N₅. (f) The band gap and the intrinsic EF energy difference with respect to the number of layers. The red dotted line represents the band gap of bulk C₃N₅. The folding lines in the right column indicate the H⁺/H₂ and O₂/H₂O potentials relative to the bottom and top surfaces.

Table 1 The band gaps, the surface potential difference (EF), the potentials of photogenerated electrons/holes U_e/U_h, the over potentials for the HER (η_HER) and OER (η_OER) and the STH efficiency Q_STH and corrected STH efficiency Q′_STH of the C₃N₅ photocatalysts

Systems	Gap (eV)	EF (V)	U e (V)	η HER (V)	U h (V)	η OER (V)	Q′STH (%)
Monolayer	2.95	0.31	1.14	0.76	2.11	0.71	2.65
Bilayer	2.73	0.43	1.12	0.74	2.03	0.69	4.54
Trilayer	2.51	0.65	1.10	0.75	2.06	0.68	7.26
4 layers	2.32	0.85	1.11	0.76	2.06	0.69	9.94
5 layers	2.16	1.19	1.17	0.79	1.99	0.69	12.35

---

Another interesting effect of the intrinsic EF is the spatial separation of the Kohn–Sham electron wavefunctions of the VBM and CBM, which reside mainly on the top and bottom layers, as shown in Fig. 4. Therefore, the photogenerated electrons and holes will accumulate at the bottom and top layers, contributing to the HER and OER processes, respectively. This is consistent with the mechanism discussed in the beginning of this section.

Fig. 4 The charge distributions of the Kohn–Sham electron wavefunctions of VBM (yellow) and CBM (blue) for (a) bilayer, (b) trilayer, (c) four-layer and (d) five-layer C₃N₅. The isosurface value is set to 0.004 e Å⁻³.

The redox ability of the photogenerated carriers in the C₃N₅ nanosheets was then evaluated from the band alignments. The potentials of the photogenerated electrons (U_e) and holes (U_h) are listed in Table 1. Although the band gap of the C₃N₅ nanosheets decreases drastically from 2.95 to 2.16 eV with the increase in thickness, U_e and U_h change slightly. For the five-layer C₃N₅, U_e and U_h are only 0.12 V lower than those of the monolayer. This implies that the redox ability of photogenerated carriers is well preserved in the C₃N₅ multilayers with the decrease of the electronic band gap, which is quite crucial for achieving high photocatalytic redox ability in a small band gap semiconductor.

Having investigated the electronic structures of C₃N₅ multilayers, we then turned our attention to the two half-reactions (OER and HER) on the C₃N₅ photocatalysts. Our calculations showed that OH* prefers to adsorb on the C5 site of the top surface, consistent with the experimental finding that C5 has the most positive potential in the structure.³² The calculated Gibbs free energy (ΔG) diagrams of the four steps that occur at the C5 site of a C₃N₅ bilayer are shown in Fig. 5. Meanwhile, we also considered a two-electron process for the OER process which generates H₂O₂ from H₂O.⁵² The free energy of the two-electron process ΔG_H2O2 is higher than that of the four-electron process ΔG_O*, as shown in Fig. S4,† suggesting that the C₃N₅ multilayer prefers a four-electron OER rather than a two-electron OER. Therefore, we mainly focus on the four-electron OER process in the following parts.

Fig. 5 Optimized geometries of the intermediates during the (a) OER and (b) HER on the C₃N₅ bilayer. The pink and red balls represent H and O atoms, respectively. Free energy diagrams for the (c) OER and (d) HER on the C₃N₅ bilayer. U = 2.03 V and U = 1.12 V represent the voltage provided by photogenerated holes and electrons.

The 1^st step of the four-electron OER has the highest ΔG value of about 1.94 eV among the four steps, corresponding to an overpotential of 0.69 V. The overpotential is much smaller than that of the C₃N₄ monolayer (1.56 V),^26,29 implying that C₃N₅ nanosheets have a higher OER activity than C₃N₄ materials. In this case, an external voltage higher than 1.92 V is required for the OER to proceed. Fortunately, the potentials of the photogenerated holes are high enough to drive the OER. Under the electric voltage provided by the photogenerated holes (U_h = 1.99–2.11 eV), all of the four steps become downhill, as shown in Fig. 5(c), indicating that the OER process can proceed, driven solely by the photogenerated holes.

The HER prefers the N6 site of the bottom surface, as shown in Fig. 5(b). The ΔG for the HER at U = 0 is about 0.74 eV for a C₃N₅ bilayer, as shown in Fig. 5(d). This value is lower than the potential of the photogenerated electrons, U_e = 1.10–1.17 V, as listed in Table 1. Under the voltage (1.12 V) provided by the photogenerated electrons, the profile of the Gibbs free energies becomes downhill, as shown in Fig. 5(d), suggesting that the HER can proceed without the need for an external electric field. Therefore, both the HER and OER can proceed spontaneously driven solely by the photogenerated electrons and holes.

The thickness-dependent overpotentials of the HER and OER and the potentials of the photogenerated carriers of the C₃N₅ nanosheets are listed in Table 1. Clearly, the potentials of the photogenerated carriers are high enough to drive the HER and OER processes for these C₃N₅ photocatalysts.

High carrier mobility not only facilitates carrier migration to the surface for the subsequent HER and OER process, but also avoids the rapid combination of the photogenerated carriers. We calculated the carrier mobility of the C₃N₅ monolayer by using the acoustic phonon-limited scattering model. The carrier mobility can be expressed as:⁵³

where e is the electron charge, k_B and ħ are the Boltzmann constant and the Planck's constant, respectively, T is the temperature which was set to 300 K, m*_e is the effective mass and m_d is the average effective mass defined by (m*_xm*_y)^1/2. C_2D is the stiffness constant obtained from the strain energy curve. Eⁱ is the deformation potential calculated by Eⁱ = ΔV_i/(Δl/l₀), where ΔV_i is the shift of the band edge and l₀ is the lattice constant in the x or y direction. The calculated carrier mobility and computational details of the C₃N₅ monolayer are presented in Table 2 and Fig. S6,† where strong anisotropy along different directions is obvious. Surprisingly, the electron mobility is up to 3.8 × 10³ cm² V⁻¹ s⁻¹ along the y direction, which can be ascribed to the small deformation potential. Clearly, the electron mobility is higher than that in MoS₂ (340 cm² V⁻¹ s⁻¹) and PtS₂ (1107 cm² V⁻¹ s⁻¹).⁵⁴ The large difference between the electrons and holes is beneficial for the separation of the photogenerated carriers.⁵⁵

Table 2 Carrier mobility and relevant parameters of the C₃N₅ monolayer

Carrier type	m e*/m0	m e*/m0	E x	E y	C x	C y	μ x	μ y
	x-	y-	(eV)		(J m^−2)		(cm^2 V^−1 s^−1)
Electron	1.89	1.82	1.45	0.52	54.46	164.36	161.08	3777.36
Hole	1.25	2.34	4.71	2.98			24.37	98.04

---

Finally, we evaluated the light absorption properties and solar energy conversion efficiencies of the C₃N₅ photocatalysts. Fig. 6(a) shows the optical absorption of the C₃N₅ monolayer and multilayers calculated at the HSE06 level. Multilayers have higher optical absorption ability than the monolayer in the visible light region, due to their reduced electronic band gaps. The first absorption peak of the C₃N₅ photocatalysts exhibits a clear red-shift as the thickness increases. The absorption wavelengths of the photocatalysts cover the visible-light range, which is beneficial for improving the solar energy conversion efficiency. Assuming 100% efficiency of light absorption and carrier utilization,¹³ the STH efficiencies of the C₃N₅ photocatalysts are roughly estimated to be 2.65% (monolayer), 4.54% (bilayer), 7.26% (trilayer), 10.07% (four-layer) and 12.35% (five-layer), respectively. The upper limit of the STH efficiency of these C₃N₅ nanosheets even exceeds the critical value (∼10%) that makes the photocatalysts economically viable. These STH efficiencies are higher than 2.6% of Al₂S₃ and 6.4% of Ga₂S₃,¹³ and comparable to 9.3% of WSe₂/SnSe₂ and 10.5% of MoSe₂/SnSe₂ heterostructures,⁵⁶ but lower than 15.8% of InGaSSe and 21.4% of InGasTe Janus bilayers.⁵⁷ However, the HER in these materials cannot be driven solely by the photogenerated carriers unless vacancy defects are introduced.^56,57 In view of the recent experimental progress in the synthesis of C₃N₅ nanosheets,³⁴ we believe our theoretical results will motivate the following experimental explorations on the photocatalytic overall water splitting.

Fig. 6 (a) The optical absorption of the monolayer and few-layered C₃N₅ calculated using the HSE06 functional. (b) The STH efficiency of monolayer and few-layered C₃N₅.

Conclusion

In summary, we proposed that the intrinsic electric field in photocatalysts is available for achieving the driving forces and high solar energy conversion efficiency for overall water splitting and demonstrated promising candidate materials, C₃N₅ multilayers. The intrinsic out-of-plane electric field and the high carrier mobility accelerate the separation and migration of the photogenerated carriers and alter the band alignment of the photocatalysts. High photocatalytic redox ability is preserved in the C₃N₅ multilayers, enabling the spontaneous HER and OER in water, driven solely by the photogenerated electrons and holes. The tunable electronic band gap of the C₃N₅ multilayers, 2.95–2.16 eV, broadens the energy range of light absorption and leads to high solar energy conversion efficiency. Our computational results offer not only low-cost, Earth-abundant and environmentally friendly photocatalysts but also a promising strategy for the design of photocatalysts for highly efficient overall water splitting without using sacrificial reagents.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This study was supported by the National Natural Science Foundation of China (No. 21833004) and the Basic Research Project of Natural Science Foundation of Shandong Province (ZR2018ZB0751).

References

1. M. G. Walter, E. L. Warren, J. R. McKone, S. W. Boettcher, Q. Mi, E. A. Santori and N. S. Lewis, Chem. Rev., 2010, 110, 6446–6473.
2. M. R. Hoffmann, S. T. Martin, W. Choi and D. W. Bahnemann, Chem. Rev., 1995, 95, 69–96.
3. G. Girishkumar, B. McCloskey, A. C. Luntz, S. Swanson and W. Wilcke, J. Phys. Chem. Lett., 2010, 1, 2193–2203.
4. A. Fujishima and K. Honda, Nature, 1972, 238, 37–38.
5. M. Ni, M. K. H. Leung, D. Y. C. Leung and K. Sumathy, Renewable Sustainable Energy Rev., 2007, 11, 401–425.
6. T. R. Cook, D. K. Dogutan, S. Y. Reece, Y. Surendranath, T. S. Teets and D. G. Nocera, Chem. Rev., 2010, 110, 6474–6502.
7. J. R. Bolton, S. J. Strickler and J. S. Connolly, Nature, 1985, 316, 495–500.
8. C. Xu, P. Ravi Anusuyadevi, C. Aymonier, R. Luque and S. Marre, Chem. Soc. Rev., 2019, 48, 3868–3902.
9. K. Maeda and K. Domen, J. Phys. Chem. Lett., 2010, 1, 2655–2661.
10. J. Yang, D. Wang, H. Han and C. Li, Acc. Chem. Res., 2013, 46, 1900–1909.
11. L. Liao, Q. Zhang, Z. Su, Z. Zhao, Y. Wang, Y. Li, X. Lu, D. Wei, G. Feng, Q. Yu, X. Cai, J. Zhao, Z. Ren, H. Fang, F. Robles-Hernandez, S. Baldelli and J. Bao, Nat. Nanotechnol., 2013, 9, 69–73.
12. Y. Goto, T. Hisatomi, Q. Wang, T. Higashi, K. Ishikiriyama, T. Maeda, Y. Sakata, S. Okunaka, H. Tokudome, M. Katayama, S. Akiyama, H. Nishiyama, Y. Inoue, T. Takewaki, T. Setoyama, T. Minegishi, T. Takata, T. Yamada and K. Domen, Joule, 2018, 2, 509–520.
13. C. F. Fu, J. Sun, Q. Luo, X. Li, W. Hu and J. Yang, Nano Lett., 2018, 18, 6312–6317.
14. 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 and H. N. Dinh, J. Mater. Res., 2011, 25, 3–16.
15. Z. Zou, J. Ye, K. Sayama and H. Arakawa, Nature, 2001, 414, 625–627.
16. Q. Wang, T. Hisatomi, Q. Jia, H. Tokudome, M. Zhong, C. Wang, Z. Pan, T. Takata, M. Nakabayashi, N. Shibata, Y. Li, I. D. Sharp, A. Kudo, T. Yamada and K. Domen, Nat. Mater., 2016, 15, 611–615.
17. J. Liu, Y. Liu, N. Liu, Y. Han, X. Zhang, H. Huang, Y. Lifshitz, S.-T. Lee, J. Zhong and Z. Kang, Science, 2015, 347, 970–974.
18. S. Hu, C. Xiang, S. Haussener, A. D. Berger and N. S. Lewis, Energy Environ. Sci., 2013, 6, 2984–2993.
19. M. Qiao, J. Liu, Y. Wang, Y. Li and Z. Chen, J. Am. Chem. Soc., 2018, 140, 12256–12262.
20. X. Li, Z. Li and J. Yang, Phys. Rev. Lett., 2014, 112, 018301.
21. Y. Ji, M. Yang, H. Dong, T. Hou, L. Wang and Y. Li, Nanoscale, 2017, 9, 8608–8615.
22. M. Qiao, Y. Chen, Y. Wang and Y. Li, J. Mater. Chem. A, 2018, 6, 4119–4125.
23. C.-F. Fu, X. Li, Q. Luo and J. Yang, J. Mater. Chem. A, 2017, 5, 24972–24980.
24. B. Zhu, L. Zhang, B. Cheng and J. Yu, Appl. Catal., B, 2018, 224, 983–999.
25. G. Gao, Y. Jiao, F. Ma, Y. Jiao, E. Waclawik and A. Du, J. Catal., 2015, 332, 149–155.
26. J. Wirth, R. Neumann, M. Antonietti and P. Saalfrank, Phys. Chem. Chem. Phys., 2014, 16, 15917–15926.
27. H. Li, Y. Wu, L. Li, Y. Gong, L. Niu, X. Liu, T. Wang, C. Sun and C. Li, Appl. Surf. Sci., 2018, 457, 735–744.
28. Y. Zheng, Y. Jiao, J. Chen, J. Liu, J. Liang, A. Du, W. Zhang, Z. Zhu, S. C. Smith, M. Jaroniec, G. Q. Lu and S. Z. Qiao, J. Am. Chem. Soc., 2011, 133, 20116–20119.
29. S. Lin, X. Ye, X. Gao and J. Huang, J. Mol. Catal. A: Chem., 2015, 406, 137–144.
30. W. Che, W. Cheng, T. Yao, F. Tang, W. Liu, H. Su, Y. Huang, Q. Liu, J. Liu, F. Hu, Z. Pan, Z. Sun and S. Wei, J. Am. Chem. Soc., 2017, 139, 3021–3026.
31. K. Chen, Z. Chai, C. Li, L. Shi, M. Liu, Q. Xie, Y. Zhang, D. Xu, A. Manivannan and Z. Liu, ACS Nano, 2016, 10, 3665–3673.
32. I. Y. Kim, S. Kim, X. Jin, S. Premkumar, G. Chandra, N. S. Lee, G. P. Mane, S. J. Hwang, S. Umapathy and A. Vinu, Angew. Chem., Int. Ed., 2018, 57, 17135–17140.
33. G. Kresse and J. Furthmüller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169–11186.
34. P. E. Blöchl, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 50, 17953–17979.
35. 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.
36. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868.
37. S. Grimme, J. Comput. Chem., 2006, 27, 1787–1799.
38. J. Heyd, G. E. Scuseria and M. Ernzerhof, J. Chem. Phys., 2003, 118, 8207–8215.
39. H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Solid State, 1976, 13, 5188–5192.
40. A. Togo, F. Oba and I. Tanaka, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 78, 134106.
41. G. J. Martyna, M. L. Klein and M. Tuckerman, J. Chem. Phys., 1992, 97, 2635–2643.
42. J. K. Nørskov, J. Rossmeisl, A. Logadottir, L. Lindqvist, J. R. Kitchin, T. Bligaard and H. Jónsson, J. Phys. Chem. B, 2004, 108, 17886–17892.
43. Á. Valdés, Z. W. Qu, G. J. Kroes, J. Rossmeisl and J. K. Nørskov, J. Phys. Chem. C, 2008, 112, 9872–9879.
44. H. Xu, D. Cheng, D. Cao and X. C. Zeng, Nat. Catal., 2018, 1, 339–348.
45. M. Faraji, M. Yousefi, S. Yousefzadeh, M. Zirak, N. Naseri, T. H. Jeon, W. Choi and A. Z. Moshfegh, Energy Environ. Sci., 2019, 12, 59–95.
46. K. Chang, Z. Mei, T. Wang, Q. Kang, S. Ouyang and J. Ye, ACS Nano, 2014, 8, 7078–7087.
47. Q. Lu, Y. Yu, Q. Ma, B. Chen and H. Zhang, Adv. Mater., 2016, 28, 1917–1933.
48. J. Mahmood, E. K. Lee, M. Jung, D. Shin, I. Y. Jeon, S. M. Jung, H. J. Choi, J. M. Seo, S. Y. Bae, S. D. Sohn, N. Park, J. H. Oh, H. J. Shin and J. B. Baek, Nat. Commun., 2015, 6, 6486–6492.
49. J. H. Jung, C.-H. Park and J. Ihm, Nano Lett., 2018, 18, 2759–2765.
50. X. Ma, Y. Lv, J. Xu, Y. Liu, R. Zhang and Y. Zhu, J. Phys. Chem. C, 2012, 116, 23485–23493.
51. X. Tian, J. Duan, Y. Wei, N. Feng, X. Wang, Z. Gong, Y. Du and B. I. Yakobson, Adv. Funct. Mater., 2019, 29, 1808721.
52. Z. Li, C. Kong and G. Lu, J. Phys. Chem. C, 2015, 120, 56–63.
53. S. Qi, X. Ma, B. Yang, L. Sun, W. Li and M. Zhao, Carbon, 2019, 149, 234–241.
54. W. Zhang, Z. Huang, W. Zhang and Y. Li, Nano Res., 2014, 7, 1731–1737.
55. S. Lu, C. Li, H. H. Li, Y. F. Zhao, Y. Y. Gong, L. Y. Niu, X. J. Liu and T. Wang, Appl. Surf. Sci., 2017, 392, 966–974.
56. Y. Fan, J. Wang and M. Zhao, Nanoscale, 2019, 11, 14836–14843.
57. Y. Fan, X. Ma, J. Wang, X. Song, A. Wang, H. Liu and M. Zhao, Sci. Bull., 2019 DOI:10.1016/j.scib.2019.10.018.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9nr08416a

---