Manipulatable Interface Electric Field and Charge Transfer in a 2D / 2D Heterojunction Photocatalyst via Oxygen Intercalation

: Charge separation is the most important factor in determining the photocatalytic activity of a 2D / 2D heterostructure. Despite the exclusive advantages of 2D / 2D heterostructure semiconductor systems such as large surface / volume ratios, their use in photocatalysis is limited due to the low e ﬃ ciency of charge separation and high recombination rates. As a remedy for the weak interlayer binding and low carrier transport e ﬃ ciency in 2D / 2D heterojunctioned semiconductors, we suggested an impurity intercalation method for the 2D / 2D interface. PtS 2 / C 3 N 4 , as a prototype heterojunction material, was employed to investigate the e ﬀ ect of anion intercalation on the charge separation e ﬃ ciency in a 2D / 2D system using density functional theory. With oxygen intercalation at the PtS 2 / C 3 N 4 interface, a reversed and stronger localized dipole moment and a built-in electric ﬁeld were induced in the vertical direction of the PtS 2 / C 3 N 4 interface. This theoretical work suggests that the anion intercalation method can be a way to control built-in electric ﬁelds and charge separation in designs of 2D / 2D heterostructures that have high photocatalytic activity.


Introduction
With the increasing energy demand, energy conversion and storage are important key factors to meet the demands of humans. At the same time, the energy crisis and environmental pollution problems come one after another because the burning of fossil fuels produces a huge amount of CO 2 . Therefore, renewable energy sources have received intensive attention as an alternative to non-renewable fossil fuels [1][2][3]. Solar energy conversion to hydrogen and hydrocarbon fuels using photocatalysts are considered to be a very promising renewable energy resources [4,5], however, eco-friendly hydrogen production from solar energy is only 5% of the total commercial hydrogen production [6]. To increase hydrogen production by renewable energy sources, the semiconductor-based photocatalyst has received considerable interest in water splitting and pollutant degradation, such as CO 2 reduction [7,8]. However, the most widely used photocatalysts such as metal oxide and precious metals are limited because of the high cost and the band gap within the ultraviolet region [9][10][11]. Therefore, semiconductor photocatalysts require a suitable photocatalytic system for efficient charge separation for electron-hole pairs [12][13][14].
Graphitic carbon nitrides (C 3 N 4 ) have attracted attention for hydrogen and oxygen evolution as a water splitting photocatalyst due to their low cost, easy fabrication, and high chemical and thermal stability [11,15,16]. In particular, the band gap of C 3 N 4 (2.70 eV) is appropriate for a photocatalyst

Results and Discussion
As the impurity intercalation of the heterostructure, O atoms were chosen because previous studies have reported that the photocatalytic performance of monolayer C 3 N 4 is enhanced by O doping [48]. To theoretically examine the stability of the PtS 2 /C 3 N 4 heterostructure after doping, the intercalation energy (E int ) of the PtS 2 /Ox/C 3 N 4 heterostructure was calculated using Equation (1): where E PtS 2 /O x /C 3 N 4 , E PtS 2 /C 3 N 4 , and E O 2 are the total energy of the PtS 2 /O x /C 3 N 4 , the PtS 2 /C 3 N 4 heterostructure, and the O 2 , respectively. The n is the number of O atoms in the heterostructure. The E int of the PtS 2 /O 1.8% /C 3 N 4 heterostructure was −1.78 eV. The optimized configuration of the PtS 2 /O x /C3N4 heterostructure is shown Figure S1. The E int of all considered PtS 2 /Ox/C 3 N 4 heterostructures is shown in Table S1. Based on these results, we can infer that the PtS 2 /O 1.8% /C 3 N 4 heterostructure is energetically stable. This means that the O atoms remain captured between the C 3 N 4 and PtS 2 layers, rather than diffusing between the layers and recombining to yield O 2 or H 2 O gas. The calculated interface binding energy (E b ) of the PtS 2 /C 3 N 4 and the PtS 2 /O 1.8% /C 3 N 4 heterostructures was 12.5 and 10 meV/Å 2 , which was close to the typical vdW-binding energy (13-21 meV/Å 2 ) in previous works [49,50]. Hence, the PtS 2 /C 3 N 4 and PtS 2 /O 1.8% /C 3 N 4 heterostructures belong to a vdW heterostructure.
To explore the charge separation after the O-intercalation in the PtS 2 /C 3 N 4 heterostructure, the charge density difference and planar-averaged charge density difference of the PtS 2 /C 3 N 4 and PtS 2 /O X /C 3 N 4 (X = 1.8, 3.6, 5.4, 7.1 %) heterostructures were calculated. The charge density difference iso-surface can be evaluated with Equation (2): where ρ total , ρ C 3 N 4 , ρ PtS 2 , and ρ X denote the charge densities of the PtS 2 /C 3 N 4 or PtS 2 /O X /C 3 N 4 heterostructure, the monolayer C 3 N 4 , the monolayer PtS 2 , and the O atom, respectively. It is defined as Equation (3): The charge density difference and planar charge density difference of the PtS 2 /C 3 N 4 and the PtS 2 /O 1.8% /C 3 N 4 heterostructures is shown in Figure 1a. The charge density difference iso-surface plots are shown in the inset in Figure 1a. The yellow and sky blue colors indicate the electron accumulation and depletion regions, respectively. The red line of Figure 1a indicates that the planar-averaged charge density difference of the C 3 N 4 /PtS 2 (left) and the PtS 2 /O 1.8% /C 3 N 4 (right) heterostructures along the Z direction exhibited charge redistribution. In both the PtS 2 /C 3 N 4 and the PtS 2 /O 1.8% /C 3 N 4 heterostructures, the charge redistribution mainly occurred around the PtS 2 /C 3 N 4 heterostructure interface and the charge transferred from the C 3 N 4 to the PtS 2 layer in the heterostructure. However, in the case of the PtS 2 /C 3 N 4 heterostructure, little charge transfer showed between the C 3 N 4 and PtS 2 . (Figure 1a) This indicates that the PtS 2 /C 3 N 4 heterostructure shows a low efficiency of separation of photogenerated electron-hole pairs, like other 2D/2D heterostructures [31,51]. Due to the O-intercalation, the PtS 2 /O 1.8% /C 3 N 4 heterostructure effectively boosts the charge separation, better than that of the PtS 2 /C 3 N 4 heterostructure. The charge transferred from the PtS 2 to the C 3 N 4 layer at the interface of the PtS 2 /O 1.8% /C 3 N 4 heterostructure, unlike in the PtS 2 /C 3 N 4 heterostructure. In addition, in the other O concentration, the direction of charge transfer was the same, and the amount of charge transfer increased ( Figure S2). From these results, this study confirmed the impurity intercalation effect on the amount and direction of charge transfer. As the built-in electric field at the interface makes an important contribution to charge separation (drifted charge separation) and other charge separation pathways in semiconductors, it must be conceivable [52,53]. In general, the direction of a built-in electric field is induced by the dipole moment [54,55]. Thus, the dipole moment and built-in electric field of the PtS2/C3N4 and the PtS2/OX/C3N4 heterostructures were calculated, as shown in Figure 1b. Using Equation (3), the dipole moment (μ) is given by Equation (4) where z indicates the z-axis distance. In Figure 1b, the red line denotes the dipole moment. The μ value decreases as O-intercalation increases and the direction of μ is reversed. As the oxygen is inserted, the direction and strength of the μ changes because the electrons on the surface of the PtS2 move to the O, resulting in the surface of the PtS2 having a positive charge. This result indicates that the change in dipole moment is caused by the acceptance of electrons occurring between the PtS2 layer and the O, regardless of the interaction between the O and the C3N4 layer. From these results, it can be predicted that a similar dipole moment change will occur due to electron acceptance between the surface and the anion if the other anion as well as the O are inserted into the 2D/2D heterostructure. Next, a built-in electric field was generated at the PtS2/C3N4 and the PtS2/OX/C3N4 heterostructure. Thus, the built-in electric field was calculated as Equation (5): As the built-in electric field at the interface makes an important contribution to charge separation (drifted charge separation) and other charge separation pathways in semiconductors, it must be conceivable [52,53]. In general, the direction of a built-in electric field is induced by the dipole moment [54,55]. Thus, the dipole moment and built-in electric field of the PtS 2 /C 3 N 4 and the PtS 2 /O X /C 3 N 4 heterostructures were calculated, as shown in Figure 1b. Using Equation (3), the dipole moment (µ) is given by Equation where z indicates the z-axis distance. In Figure 1b, the red line denotes the dipole moment. The µ value decreases as O-intercalation increases and the direction of µ is reversed. As the oxygen is inserted, the direction and strength of the µ changes because the electrons on the surface of the PtS 2 move to the O, resulting in the surface of the PtS 2 having a positive charge. This result indicates that the change in dipole moment is caused by the acceptance of electrons occurring between the PtS 2 layer and the O, regardless of the interaction between the O and the C 3 N 4 layer. From these results, it can be predicted that a similar dipole moment change will occur due to electron acceptance between the surface and the anion if the other anion as well as the O are inserted into the 2D/2D heterostructure. Next, a built-in electric field was generated at the PtS 2 /C 3 N 4 and the PtS 2 /O X /C 3 N 4 heterostructure. Thus, the built-in electric field was calculated as Equation (5): where E and ε 0 are the built-in electric field, the charge density, and the vacuum permittivity, respectively. ∆ρ is the charge density difference obtained in Equation (2). The built-in electric field of the PtS 2 /C 3 N 4 heterostructure (−1.11 × 10 6 V/cm) was smaller than that of the PtS 2 /O 1.8% /C 3 N 4 heterostructure (1.64 × 10 6 V/cm). In addition, the direction of the built-in electric field was reversed after O-intercalation in the C 3 N 4 /PtS 2 heterostructure. As the amount of O-intercalation in the PtS 2 /C 3 N 4 heterostructure increased, the strength of the built-in electric field increased ( Figure 2b). As shown in Figure 2b, as O-intercalation increased, the intensity of the dipole moment and the built-in electric field increased linearly. From these results, it can be inferred that the direction of the built-in electric field can be controlled by adjusting the O-intercalation. This increased built-in electric field indicates that a powerful electrostatic field existed perpendicular to the interface direction. Thus, it can be predicted that the PtS 2 /O X /C 3 N 4 heterostructure will show increased photocatalytic activity because the driving force for charge separation provided by the built-in electric field is consequently enhanced, thus leading to significantly increased photocatalytic activity.
Catalysts 2020, 10, x FOR PEER REVIEW 5 of 12 where E and are the built-in electric field, the charge density, and the vacuum permittivity, respectively. ∆ρ is the charge density difference obtained in Equation (2). The built-in electric field of the PtS2/C3N4 heterostructure ( −1.11 × 10 V/cm) was smaller than that of the PtS2/O1.8%/C3N4 heterostructure (1.64 × 10 V/cm). In addition, the direction of the built-in electric field was reversed after O-intercalation in the C3N4/PtS2 heterostructure. As the amount of O-intercalation in the PtS2/C3N4 heterostructure increased, the strength of the built-in electric field increased ( Figure 2b). As shown in Figure 2b, as O-intercalation increased, the intensity of the dipole moment and the built-in electric field increased linearly. From these results, it can be inferred that the direction of the built-in electric field can be controlled by adjusting the O-intercalation. This increased built-in electric field indicates that a powerful electrostatic field existed perpendicular to the interface direction. Thus, it can be predicted that the PtS2/OX/C3N4 heterostructure will show increased photocatalytic activity because the driving force for charge separation provided by the built-in electric field is consequently enhanced, thus leading to significantly increased photocatalytic activity. The optimized band structure of the PtS2/C3N4 and the PtS2/O1.8%/C3N4 with various O concentration heterostructures was calculated to confirm the photocatalytic activity ( Figure 2). Considering that, in general, the PBE approach underestimates the band gaps (Eg), this study used a band structure calculated using the HSE06 method to increase the reliability in the following analyses. The contribution of each layer in the conduction band minimum (CBM) and valence band maximum (VBM) of the PtS2/C3N4 and the PtS2/O1.8%/C3N4 heterostructure was investigated as shown in Figure 3. Since it is generally known that the trends of hybrid functional calculations and normal DFT calculations are not very different, the contribution of each layer in the CBM and the VBM was verified using the band structure of the PBE approach [56]. The contributions of each layer in the PtS2/O3.6%/C3N4, PtS2/O5.4%/C3N4, and PtS2/O7.1%/C3N4 heterostructures are shown in Figure S3. The C3N4 layer in the PtS2/C3N4 heterostructure largely contributed to the VBM, whereas the CBM was dominated by the PtS2 layer in the C3N4/PtS2 heterostructure (Figure 2a). In addition, this phenomenon was maintained despite the amount the O-intercalation increased in the PtS2/C3N4 heterostructure (Figures 2b and S3). The PtS2/C3N4 heterostructure with and without O-intercalation had staggered gap structures, resulting in Z-scheme or type-II systems. The reason for the difference The optimized band structure of the PtS 2 /C 3 N 4 and the PtS 2 /O 1.8% /C 3 N 4 with various O concentration heterostructures was calculated to confirm the photocatalytic activity ( Figure 2). Considering that, in general, the PBE approach underestimates the band gaps (E g ), this study used a band structure calculated using the HSE06 method to increase the reliability in the following analyses. The contribution of each layer in the conduction band minimum (CBM) and valence band maximum (VBM) of the PtS 2 /C 3 N 4 and the PtS 2 /O 1.8% /C 3 N 4 heterostructure was investigated as shown in Figure 3. Since it is generally known that the trends of hybrid functional calculations and normal DFT calculations are not very different, the contribution of each layer in the CBM and the VBM was verified using the band structure of the PBE approach [56]. The contributions of each layer in the PtS 2 /O 3.6% /C 3 N 4 , PtS 2 /O 5.4% /C 3 N 4 , and PtS 2 /O 7.1% /C 3 N 4 heterostructures are shown in Figure S3. The C 3 N 4 layer in the PtS 2 /C 3 N 4 heterostructure largely contributed to the VBM, whereas the CBM was dominated by the PtS 2 layer in the C 3 N 4 /PtS 2 heterostructure (Figure 2a). In addition, this phenomenon was maintained despite the amount the O-intercalation increased in the PtS 2 /C 3 N 4 heterostructure (Figure 2b and Figure S3). The PtS 2 /C 3 N 4 heterostructure with and without O-intercalation had staggered gap structures, resulting in Z-scheme or type-II systems. The reason for the difference of heterostructure type is due to the difference in the direction and strength of the built-in electric field caused by the difference in charge redistribution occurring by O-intercalation. From these results, it can be inferred that the electrons and holes will move differently in the PtS 2 /C 3 N 4 and the PtS 2 /O X /C 3 N 4 heterostructures when visible-light irradiation is absorbed.
Catalysts 2020, 10, x FOR PEER REVIEW 6 of 12 of heterostructure type is due to the difference in the direction and strength of the built-in electric field caused by the difference in charge redistribution occurring by O-intercalation. From these results, it can be inferred that the electrons and holes will move differently in the PtS2/C3N4 and the PtS2/OX/C3N4 heterostructures when visible-light irradiation is absorbed. The band edge position of the PtS2/C3N4 and the PtS2/Ox/C3N4 heterostructures, which is another important factor that determines the performance of photocatalytic water splitting of a semiconductor, was calculated to confirm the redox capability of the PtS2/C3N4 and the PtS2/Ox/C3N4 heterostructures. The potential level vs. the normal hydrogen electrode (NHE) was converted from the vacuum potential level ( ) using Equation (6) [57]: where is the energy of the free electrons at the hydrogen scale ( = − 4.44 ) and is obtained by the planar-averaged electrostatic potential. To achieve water splitting on the semiconductor, the CBM potential is more negative than the H + /H2 reduction potential (0.0 eV), and the VBM potential is more positive than the H2O/O2 oxidation potential (1.23 eV). Figure 3 shows the charge transfer and band edge position of the PtS2/C3N4 and the PtS2/Ox/C3N4 heterostructures. The band edge position of the CBM (VBM) potential in the C3N4/PtS2 heterostructure was −0.56 (1.66 eV), which was more negative (positive) than the H + /H2 reduction (O2/H2O oxidation) potential. Based on the direction of the built-in electric field, in the case of the PtS2/C3N4 heterostructure, the electrons in the CBM of the PtS2 could transfer to the VBM of the C3N4 to combine with holes. In addition, the electrons in the CBM of the C3N4 were used to reduce H + to H2, and the holes in the VBM of the C3N4 were used to reduce H + to H2. Through this result, it can be inferred that the PtS2/C3N4 heterostructure showed a Z-scheme. In the case of the PtS2/OX/C3N4 heterostructure, the electrons transferred from the CBM of the C3N4 to the CBM of the PtS2 and the VBM holes of the PtS2 transferred into the VBM of the C3N4 structure. Thus, H + /H2 reduction and O2/H2O oxidation occurred in the CBM of the PtS2 and the VBM of the C3N4. In the density of states, there was no defect level caused by O in the energy gap of the PtS2 and the C3N4 ( Figure S4). In the case of the C3N4 layer, the energy gap increased as O was inserted at the interface of the PtS2/C3N4 heterostructure. On the other hand, the energy gap of the PtS2 layer decreased. Due to the decreased The band edge position of the PtS 2 /C 3 N 4 and the PtS 2 /O x /C 3 N 4 heterostructures, which is another important factor that determines the performance of photocatalytic water splitting of a semiconductor, was calculated to confirm the redox capability of the PtS 2 /C 3 N 4 and the PtS 2 /O x /C 3 N 4 heterostructures. The potential level vs. the normal hydrogen electrode (NHE) was converted from the vacuum potential level (E vacuum potential level ) using Equation (6) [57]: E potential level vs NHE = E e − E vacuum potential level (6) where E e is the energy of the free electrons at the hydrogen scale (E e = −4.44 eV) and E vacuum potential level is obtained by the planar-averaged electrostatic potential. To achieve water splitting on the semiconductor, the CBM potential is more negative than the H + /H 2 reduction potential (0.0 eV), and the VBM potential is more positive than the H 2 O/O 2 oxidation potential (1.23 eV). Figure 3 shows  Figure S4). In the case of the C 3 N 4 layer, the energy gap increased as O was inserted at the interface of the PtS 2 /C 3 N 4 heterostructure. On the other hand, the energy gap of the PtS 2 layer decreased. Due to the decreased energy gap of the PtS 2 , the CBM of the PtS 2 was lower than the O 2 /H 2 O oxidation potential. Given the decreased energy gap of the PtS 2 , the CBM of the PtS 2 was lower than the O 2 /H 2 O oxidation potential. As the monolayer C 3 N 4 and monolayer PtS 2 had an indirect band gap, C 3 N 4 and PtS 2 exhibited similar photo absorption efficiency. However, after the intercalation of O at the interface of the heterostructure, the strength of photo absorption in the PtS 2 with a decreased energy gap will increase. From these results, it can be inferred that the strength of photo absorption in the PtS 2 will be stronger than that of the C 3 N 4 when visible-light irradiation is absorbed. The CBM level was located about 0.09 eV under the H + /H 2 reduction potential. Using the thermal energy at room temperature (0.026 eV), the possibility of H + /H 2 reduction occurring in the PtS 2 /O 1.8% /C 3 N 4 heterostructure was approximately 3%. Thus, it was found that the O-intercalation in the PtS 2 /C 3 N 4 heterostructures not only enhanced the separation of photogenerated electron-hole pairs, but also changed the direction of the built-in electric field, resulting in the determination of the heterostructure type.
Considering the generally used synthesis conditions of C 3 N 4 , the possibility of O intercalation in the PtS 2 /C 3 N 4 heterostructure was examined theoretically. The relationship between the equilibrium constants (K p ) for reactions and temperatures was considered based on the thermodynamic energy of intercalation formation. The main reactions occurring in the PtS 2 /O/C 3 N 4 heterostructure can be written as follows: (1) O 2 gas The Gibbs energy for interaction is described in Equation (7): where G PtS 2 /O/C 3 N 4 , G PtS 2 /O/C 3 N 4 , n O , and µ O denote the Gibbs energy of the PtS 2 /O/C 3 N 4 heterostructure, the Gibbs energy of the PtS 2 /C 3 N 4 heterostructure, the coefficient atoms in the chemical equation, and the chemical potential of atoms, respectively. In the case of µ O , it can be written as Equation (8): where µ 0 , k B , T, p 0 , and p i are the standard chemical potential (p 0 = 1 bar), Boltzmann constant, temperature, standard partial pressure, and partial pressure, respectively. The standard chemical potential values were taken from the JANAF-NIST thermochemical table [53]. Using Equation (2) and Equation (6), the Gibbs energy for intercalation can be expressed as Equation (9): where ∆E DFT and K p are the change in total energies of the DFT and the equilibrium constant, respectively. Hence, K p is defined as Equation (10): To consider the concentration of O-intercalation, the Boltzmann distribution is given by Equation (11) x p where x p , x i , G p , and G i denote the molar fraction of pristine and intercalation, and Gibbs free energy of pristine and intercalation in the PtS 2 /C 3 N 4 heterostructure, respectively. Using Equations (10) and (11), K p is given by Equation (12) The calculated K p was plotted for varying temperatures (Figure 4). It is known that C 3 N 4 exhibits good crystallinity under the synthesis conditions between 873 and 973 K [58]. Therefore, to exclude other factors that affect the photocatalytic activity, other than O-intercalation, the temperature range for K P was set from 800 K to 1000 K. When O 2 gas is used as the reactant gas, the partial pressure of the O 2 gas must be 2.83 × 10 21 atm at 800 K and 5.60 × 10 18 atm at 1000 K in order to insert 1 at% O. However, these partial pressures are much higher than the practical processing capabilities of synthesis devices. When H 2 O is the reactant gas for O-intercalation in the PtS 2 /C 3 N 4 heterostructure, the partial pressure of the H 2 should be controlled between 6.3 × 10 −7 atm and 3.9 × 10 −6 atm, which is a feasible range in synthesis devices. Through this result, it is expected that the PtS 2 /O/C 3 N 4 heterostructure can be synthesized using H 2 O gas as the reactant gas.
Catalysts 2020, 10, x FOR PEER REVIEW 8 of 12 Using Equations (10) and (11), is given by Equation (12) = The calculated Kp was plotted for varying temperatures (Figure 4). It is known that C3N4 exhibits good crystallinity under the synthesis conditions between 873 and 973 K [58]. Therefore, to exclude other factors that affect the photocatalytic activity, other than O-intercalation, the temperature range for KP was set from 800 K to 1000 K. When O2 gas is used as the reactant gas, the partial pressure of the O2 gas must be 2.83 × 10 atm at 800 K and 5.60 × 10 atm at 1000 K in order to insert 1 at% O. However, these partial pressures are much higher than the practical processing capabilities of synthesis devices. When H2O is the reactant gas for O-intercalation in the PtS2/C3N4 heterostructure, the partial pressure of the H2 should be controlled between 6.3 × 10 −7 atm and 3.9 × 10 −6 atm, which is a feasible range in synthesis devices. Through this result, it is expected that the PtS2/O/C3N4 heterostructure can be synthesized using H2O gas as the reactant gas.

Calculation Method
All DFT calculations were performed with the Vienna Ab initio package (VASP) [59,60]. The interactions between the valence and core electrons were described using the projector augmented wave (PAW) method [61]. A generalized gradient approximation (GGA) with the Perdew-Burke-Ernzerhof (PBE) functional was employed for the plane-wave expansion with a cutoff energy of 500 eV [62,63]. Brillouin zones were sampled with a Monkhorst-Pack k-point grid of 2 × 2 × 1 in the 4 × 4 PtS2/C3N4 heterostructure with O-intercalation [64]. The energy convergence criteria in the selfconsistent field were set to 10 -5 eV. All geometry structures were fully relaxed until the Hellman-Feynman forces achieved a range of 0.02 eVÅ . The Methfessel-Paxton smearing scheme was applied with a smearing width of 0.1 eV [65]. To avoid the artifact of a neighboring periodic image, the size of the vacuum region in the z direction was set to a value larger than 20 Å. The vdW correction was considered with Grimme's DFT-D3 (BJ) scheme to accurately describe the interaction between the C3N4 and PtS2 layers [66]. The band gap of the PtS2/C3N4 and PtS2/OX/C3N4 (X = 1.8, 3.6, 5.4, 7.1%) heterostructures were determined by the hybrid functional (HSE06) with α = 0.2 [67].

Conclusions
This study demonstrated that anion intercalation in a 2D/2D heterostructure is an effective way to manipulate the direction and strength of the built-in electric fields and dipole moments. The direction of the vertical built-in electric field in the PtS2/C3N4 heterostructure can be reversed with O-

Calculation Method
All DFT calculations were performed with the Vienna Ab initio package (VASP) [59,60]. The interactions between the valence and core electrons were described using the projector augmented wave (PAW) method [61]. A generalized gradient approximation (GGA) with the Perdew-Burke-Ernzerhof (PBE) functional was employed for the plane-wave expansion with a cutoff energy of 500 eV [62,63]. Brillouin zones were sampled with a Monkhorst-Pack k-point grid of 2 × 2 × 1 in the 4 × 4 PtS 2 /C 3 N 4 heterostructure with O-intercalation [64]. The energy convergence criteria in the self-consistent field were set to 10 −5 eV. All geometry structures were fully relaxed until the Hellman-Feynman forces achieved a range of 0.02 eV Å −1 . The Methfessel-Paxton smearing scheme was applied with a smearing width of 0.1 eV [65]. To avoid the artifact of a neighboring periodic image, the size of the vacuum region in the z direction was set to a value larger than 20 Å. The vdW correction was considered with Grimme's DFT-D3 (BJ) scheme to accurately describe the interaction between the C 3 N 4 and PtS 2 layers [66]. The band gap of the PtS 2 /C 3 N 4 and PtS 2 /O X /C 3 N 4 (X = 1.8, 3.6, 5.4, 7.1%) heterostructures were determined by the hybrid functional (HSE06) with α = 0.2 [67].

Conclusions
This study demonstrated that anion intercalation in a 2D/2D heterostructure is an effective way to manipulate the direction and strength of the built-in electric fields and dipole moments. The direction of the vertical built-in electric field in the PtS 2 /C 3 N 4 heterostructure can be reversed with O-intercalation at the interface, as the O-atom bound to the PtS 2 accepts electrons and induces a local dipole moment in the opposite direction. When the O-intercalation density increased, the intensities of the electric field and dipole moment increased nearly linearly. It was also noticeable that the band edge levels in the PtS 2 and the C 3 N 4 significantly shifted with different O-intercalation amounts. Hence, this anion intercalation method can be widely used for 2D/2D heterojunction photocatalysts for effective control of charge separation aspects and redox reactivities.