Interface Effects on the Electronic and Optical Properties of Graphitic Carbon Nitride (g-C₃N₄)/SnS₂: First-Principles Studies

Abstract

Heterojunctions have received much interest as a way to improve semiconductors’ electrical and optical properties. The impact of the interface on the electrical and optical properties of g-C₃N₄/SnS₂ was explored using first-principles calculations in this study. The results show that, at the hetero-interface, a conventional type-II band forms, resulting in a lower band gap than that in the g-C₃N₄ and SnS₂ monolayers. When there is no high barrier height, the averaged microscopic and averaged macroscopic potentials can be used to accomplish efficient carrier transformation. Furthermore, the polarization direction affects the absorption spectra. All of these discoveries have significant implications for the development of g-C₃N₄-based optoelectronics.

1. Introduction

Semiconductor photocatalysis is a green technology that provides significant social and economic benefits and has a wide range of applications in antisepsis, water quality improvement, and environmental optimization [1,2,3]. In recent years, it has become increasingly vital to investigate effective photocatalysts. Since Wang demonstrated that g-C₃N₄ can be photocatalytically oxidized and create H₂ under visible light [4], g-C₃N₄ photocatalyst research has received much attention. Because of its high stability and sufficient band gap [5,6], g-C₃N₄ is a better contender among the many photocatalysts available. However, due to the high photogenerated electron–hole recombination rate and poor light absorption caused by the low specific surface area, pure g-C₃N₄ has a low photocatalytic efficiency. As a result, several researchers, including those concerned with g-C₃N₄ functionalization and chemical doping, have concentrated their efforts on improving electron–hole separation and visible optical absorption, which are crucial for improving catalytic efficiency [5,7]. The photocatalytic effectiveness of heterojunctions is shown to be higher than that of previous modification plans. As a result, scientists have worked on a variety of g-C₃N₄/semiconductor heterojunctions, including g-C₃N₄/SnO₂ [8], g-C₃N₄/TiO₂ [9], g-C₃N₄/Zn₂GeO₄ [10], and NiO/g-C₃N₄ [11].

The n-type semiconductor SnS₂ has sparked curiosity because of its low price, nontoxicity, and acceptable band gap (approximately 2.4 eV), all of which make it ideal for photocatalysis. The band edge potentials of SnS₂ and g-C₃N₄ have been found to be extremely comparable [12]. As a result, modeling the heterojunction can be very useful in photocatalytic activity enhancement. G-C₃N₄/SnS₂ has been experimentally synthesized, and its remarkable photocatalytic activity in visible light has been demonstrated [13].

Inspired by this experimental idea, we attempt to study the hybridization between the SnS₂ sheet and g-C₃N₄. To design and increase the photocatalytic effectiveness of the g-C₃N₄/SnS₂ heterostructure, the interfacial effect on the structural, electronic, and optical properties must be studied. Herein, based on density functional theory (DFT) calculations, we try to investigate the interfacial interaction between monolayer g-C₃N₄ and SnS₂ sheet and band structures of g-C₃N₄ and SnS₂. The computation results show a type-II band alignment, which is desirable for visible light water splitting. Moreover, due to the charge transfer between g-C₃N₄ and SnS₂ sheet, it can further enhance the photogenerated electron–hole separation. These results suggest that SnS₂ sheet would be a nice choice to improve the photocatalytic performance of g-C₃N₄ photocatalyst.

2. Computational Methods

The Vienna ab initio Simulation Package (VASP 5.4, Austrian) was used to execute the computations, which were based on the projector augmented wave (PAW) approach and included dispersion correction (DFT-D2) [14,15,16,17,18,19,20]. The exchange and correlation potentials are described using the Perdew−Burke−Ernzerhof (PBE) method of generalized gradient approximation (GGA) [16]. When the energy and force on each ion are decreased below 10⁻⁶ eV and 0.02 eV/Å, structural optimization occurs. The cutoff energy of the plane-wave base group is set to 520 eV, and a Monkhorst–Pack k-grid of 7 × 7 × 1 is utilized [21].

The charge density difference is described as follows to understand how charge transfer occurs [22]:

$$\rho ={\rho }_{g-C_3{N}_4/Sn{S}_2}-{\rho }_{g-C_3{N}_4}-{\rho }_{Sn{S}_2}$$

(1)

where ${\rho }_{g-C_3{N}_4/Sn{S}_2}$, ${\rho }_{g-C_3{N}_4}$, and ${\rho }_{Sn{S}_2}$ are the charge densities of g-C₃N₄/SnS₂, g-C₃N₄, and SnS₂ monolayers in one system, respectively. Positive and negative numbers represent charge accumulation and depletion, respectively.

The interfacial valence electron change is explored to see if the interfacial contact is mostly physisorption. Then, the integrated charge density difference throughout the x–y plane is then integrated along the z direction, which is perpendicular to the interface, and the plane-averaged charge difference $\Delta q$ [23] is defined as follows:

$$\Delta q={\displaystyle {\int }_{-\infty }^{+\infty }{\displaystyle {\int }_{-\infty }^{+\infty }\left({\rho }_{Sn{S}_2/g-C_3{N}_4}-{\rho }_{Sn{S}_2}-{\rho }_{g-C_3{N}_4}\right)}}dxdy$$

(2)

Furthermore, by integrating along the z-axis, the charge displacement curve (CDC) is defined by integrating $\Delta Q$ [23]:

$$\Delta Q={\displaystyle {\int }_{-\infty }^z\Delta qdz}$$

(3)

3. Results and Discussion

3.1. Geometric Structures

In our work, we perform benchmark calculations for the isolated single-layer g-C₃N₄ and SnS₂ sheets to examine the reliability of the computational parameters. After full relaxation, the lattice parameters are a = b = 7.15 Å for g-C₃N₄, and a = b = 3.70 Å for SnS₂. To build a g-C₃N₄/SnS₂ heterojunction, we adopt 1 × 1 and 2 × 2 supercells for g-C₃N₄ (6 C and 15 N atoms) and SnS₂ (including 4 Sn and 8 S atoms), respectively, as shown in Figure 1. For g-C₃N₄ and SnS₂, the mismatch is 3% for the x and y directions, and a vacuum spacing of 15 is used to reduce artificial contact. For this heterogeneous structure, we perform full structure relaxation, and the lattice parameter is a = b = 7.18 Å; meanwhile, the minimum spacing between two monolayers is 3.06 Å. The bond lengths of C1-N1, C2-N2, C3-N3, SnS1, and Sn-S2 are 1.490, 1.353, 1.348, 2.568, and 2.567 Å, respectively.

These findings are comparable with those obtained with g-C₃N₄ and SnS₂ monolayers, showing a modest contact between the two. The interfacial formation energy ($E_F$) [24] is computed using the following equation to determine stability:

$$E_F=E_{g-C_3{N}_4/Sn{S}_2}-E_{g-C_3{N}_4}-E_{Sn{S}_2}$$

(4)

where $E_{\mathrm{g}-{\mathrm{C}}_3{\mathrm{N}}_4/\mathrm{S}\mathrm{n}{\mathrm{S}}_2}$, $E_{\mathrm{g}-{\mathrm{C}}_3{\mathrm{N}}_4}$, and $E_{\mathrm{S}\mathrm{n}{\mathrm{S}}_2}$ are the total energies of the heterojunction and monolayers. The heterojunction’s interfacial formation energy is predicted to be −0.027 eV, demonstrating improved thermodynamic stability. Furthermore, the phonon spectrum of the heterostructure is shown in Supporting Information Figure S1 to elucidate stability.

To investigate the optical properties of g-C₃N₄/SnS₂, the optical absorption spectra are calculated by converting the complex dielectric function to the absorption coefficient α by the following equation [25], where ${\epsilon }_1$ and ${\epsilon }_2$ are the real and imaginary absorption factors, respectively.

$$\alpha ={\displaystyle \frac{4\pi E}{hc}}{\left({\displaystyle \frac{\sqrt{{\epsilon }_1^2+{\epsilon }_2^2}-{\epsilon }_1}{2}}\right)}^{{\displaystyle \frac{1}{2}}}$$

(5)

3.2. Fat Band and Density of States (DOS)

As illustrated in Figure 2, the fat bands and DOS are estimated to indicate high photocatalytic activity. The conduction band minimum (CBM) and valence band maximum (VBM) of g-C₃N₄/SnS₂ are located at the Γ and M points, respectively, in Figure 2a, providing a narrower gap than that of the monolayers (2.197 and 2.864 eV in Figure 2b,c).

The main features of the DOS of Sn, S, N, and C atoms in the g-C₃N₄/SnS₂ heterostructure are similar to the results of the isolated g-C₃N₄ and SnS₂. This is because of the weak interaction and relatively large separation between the g-C₃N₄ and SnS₂. As can be seen in Figure 2a, SnS₂ (g-C₃N₄) contributes to the CBM (VBM) of the heterojunction system. The CBM (VBM) largely disperses over the p states of N and C atoms (p and d states of S and Sn atoms, respectively) in Figure 2b,c. Moreover, the VBM of g-C₃N₄ is 0.22 higher than that of SnS₂, suggesting the interfacial type-II band alignment of the g-C₃N₄/SnS₂ heterostructure, in which the valence band offset (VBO) between g-C₃N₄ and SnS₂ is about 0.22 eV, and the conduction band offset (CBO) is about 1.01 eV. This will be addressed further in the subsequent charge transport investigation. Therefore, it can be concluded that photocatalysis will transport electrons from SnS₂ to g-C₃N₄, resulting in the redox process.

3.3. Charge Transport

The charge redistribution is mainly distributed in the interfacial regions (positive and negative values are expressed as charge accumulation (yellow isosurface) and depletion (blue isosurface) in the adhesion process, respectively), as shown in Figure 3a,b, which can be explained by (i) minor local lattice distortion due to a small lattice mismatch; (ii) the high-strength double-layer honeycomb structure; and (iii) the elimination of lattice faults when the interlayer gap between SnS₂ and g-C₃N₄ is large enough. The averaged microscopic and averaged macroscopic potentials are then determined, as shown in Figure 3c. It shows that there is no high barrier height to influence the carrier transformation. Despite the electron transfer (positive value of the charge displacement curve (CDC, ΔQ) in the interfacial region) during the creation of g-C₃N₄/SnS₂, the x–y plane average charge density difference Δq suggests that no charge accumulates near the g-C₃N₄ layer side, as shown by the difference in the average charge density on the x–y plane q. Therefore, when the lattice mismatch is modest, the structure is strong, and the interfacial defects are minor. Smaller local lattice distortion, lesser interfacial defect, and thin interfacial regions can better hold the physicochemical properties of the component, and the physicochemical parameters of the heterojunction can be easily controlled, limiting the impacts of interfacial trapping states and transition barriers on carrier transport.

The localized PDOS in the normal direction of the g-C₃N₄/SnS₂ interfaces (x–y plane) is depicted in Figure 3d to study charge transport channels at the interfaces. At the heterojunction interfaces, a staggered-gap (type II) offset junction is plainly visible, implying that the charge transfer is fundamentally justified. There is no discernible difference between the g-C₃N₄ and SnS₂ monolayers when the band gap and interface are separated by a greater distance. Close to the interface, however, it drops. When a photon of light is absorbed by g-C₃N₄, it energizes electrons and holes, which are then negatively transported down the CBM and VBM. That is to say, the spontaneous interfacial charge transfer from the g-C₃N₄ to SnS₂ can be imply rationalized in terms of the large difference between two layers, which results in a built-in electric field at the interface, promoting photogenerated electrons from the CB of SnS₂ to the CB of g-C₃N₄, and improving the transformation of photogenerated holes of the VB from g-C₃N₄ to SnS₂. Therefore, the photogenerated electron–hole separation can be enhanced effectively by this built-in interface electric field. The yellow shadow in Figure 3e represents the interface states that emerge at the margin of the g-C₃N₄ conduction band. Due to band edge rearrangement and redistribution, the interface states minimize the heterojunction system’s band gap, promoting efficient charge extraction and providing a viable way to design efficient g-C₃N₄-/SnS₂-based optoelectronic devices.

3.4. Optical Properties

The photocatalytic efficiency of the g-C₃N₄ monolayer and g-C₃N₄/SnS₂ can be calculated using the absorption spectra shown in Figure 4. The peak energy of g-C₃N₄ is around 3.3 eV, as shown in the image. In contrast, the visible absorption spectra of the heterojunction are successfully improved because the absorption edge changes from UV to the near-infrared region, demonstrating polarization-dependent behavior. The optical absorption coefficients in the α direction parallel to the x–y plane are greater than those in the perpendicular direction, as shown in Figure 4.

4. Conclusions

In summary, first-principles DFT simulations were used in this study to investigate the impact of the interface on the electrical and optical properties of the organic–inorganic g-C₃N₄/SnS₂. Because the heterojunction system has a band alignment of type II, it has a smaller band gap than that of g-C₃N₄ and SnS₂ monolayers. The built-in interface electric field within the interface region is desirable for photogenerated carrier separation. The averaged microscopic and averaged macroscopic potentials reveal that carrier transformation is achievable even when the barrier height is low. The electric field and the nice band edge alignment suggest that the g-C₃N₄/SnS₂ heterostructure will have good application prospects in the field of photocatalysis. The absorption spectrum is dependent on the polarization direction and may be successfully regulated. The ramifications of this theoretical study for g-C₃N₄-based heterojunction optoelectronics and photocatalysts are important. In a nutshell, because of the band offset built-in interface polarized electric field, the energy-wasteful electron–hole recombination could be effectively reduced in the proposed g-C₃N₄/SnS₂ heterostructure, and then, the photocatalytic quantum efficiencies would be improved greatly. Therefore, g-C₃N₄/SnS₂ is a promising photocatalyst based on g-C₃N₄.