Effects of defects in g-C3N4 on excited-state charge distribution and transfer: Potential for improved photocatalysis

Journal Pre-proof Effects of defects in g-C3N4 on excited-state charge distribution and transfer: Potential for improved photocatalysis Yitong Zhang, Cong Shen, Xuemei Lu, Xijiao Mu, Peng Song PII:

S1386-1425(19)31077-7

DOI:

https://doi.org/10.1016/j.saa.2019.117687

Reference:

SAA 117687

To appear in:

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy

Received Date: 4 July 2019 Revised Date:

18 October 2019

Accepted Date: 20 October 2019

Please cite this article as: Y. Zhang, C. Shen, X. Lu, X. Mu, P. Song, Effects of defects in g-C3N4 on excited-state charge distribution and transfer: Potential for improved photocatalysis, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy (2019), doi: https://doi.org/10.1016/ j.saa.2019.117687. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier B.V.

Effects of Defects in g-C3N4 on Excited-state Charge Distribution and Transfer: Potential for Improved Photocatalysis

Yitong Zhang,a Cong Shen,a Xuemei Lu,*a Xijiao Mu,*b and Peng Song*a

Effects of Defects in g-C3N4 on Excited-state Charge Distribution and Transfer: Potential for Improved Photocatalysis

Yitong Zhang,a Cong Shen,a Xuemei Lu,*a Xijiao Mu,*b and Peng Song*a

a

Department of Physics, Liaoning University, Shenyang, Liaoning 110036, PR China.

b

School of Mathematics and Physics, Center for Green Innovation, Beijing Key Laboratory

for Magneto-Photoelectrical Composite and Interface Science, University of Science and Technology Beijing, Beijing, 100083, China.

Corresponding Authors: [email protected] (P. Song); [email protected] (X.J. Mu); [email protected] (X.M. Lu); Tel: +86-24-62202365.

Abstract:Graphite phase carbon nitride (g-C3N4) with triazine ring structures is a polymeric metal-free semiconductor with a medium bandgap and two-dimensional layered structure. g-C3N4 has attracted attention because of its photocatalytic applications, such as the photodegradation of pollution and hydrogen production via water splitting. Defective elements and sites are two essential factors in rationally designing highly-efficient photocatalysts based on g-C3N4 at the nanoscale. When the molecule absorbs energy and enters an excited state, electrons migrate and the charge distribution changes accordingly. The properties of the excited states of g-C3N4 are related to its defect elements and sites. Therefore, it is necessary to understand the effects of defects on excited states in the design of g-C3N4 catalysts. In this paper, the excited-state characteristics of intrinsic g-C3N4 and g-C3N4 with C- and N-atom defects are analyzed by density functional theory. We apply quantum chemistry and wave function analysis to determine the hole-electron distributions and charge transfer directions. To measure and discuss the characteristics of electron excitation using quantitative numerical methods, the D, Sr, H, and t indices are calculated. Our results promote a deeper understanding of the roles of defective elements and sites in photocatalysis by g-C3N4.

Keywords: g-C3N4; Defect; Excited-state; Density functional theory; Charge transfer; Charge distribution.

1. Introduction In recent years, semiconductor photocatalysis technology has attracted extensive attention and research due to its important applications in mitigating environmental pollution and energy shortages. Graphite-phase carbon nitride (g-C3N4) is composed of the most abundant elements on the earth C and N has stable chemical properties. Due to its unique band structure, suitable band edge position, high chemical and thermal stability, and predictable response to visible light, it has become one of the most widely studied and promising photocatalysts [1, 2]. In 2009, Wang et al.[1] reported for the first time that g-C3N4 can produce hydrogen by photolysis of water under visible light irradiation, causing g-C3N4 to rapidly become a hot topic in the field of photocatalytic materials. At present, g-C3N4 is used mainly in the photocatalytic decomposition of pollutants, the photolysis of water to produce hydrogen and oxygen, photocatalytic organic synthesis, and photocatalytic oxygen reduction reactions [3-9]. Although g-C3N4 has a wide range of potential applications, it has not been used with high efficient in many actual applications due to its characteristics of rapid electron-hole recombination and low specific surface area. Therefore, the modification of g-C3N4 is one of the most important research topics in the field of catalysis. For this reason, researchers have developed a variety of methods to improve the photocatalytic performance of gC3N4, such as physical composite modification,[10-15] chemical doping modification,[16-18] heterostructured nanocomposite fabrication,[19] and micro-structural adjustment [20, 21]. Compared with bulk g-C3N4, ultrathin or monolayer g-C3N4 nanosheets confer superior charge carrier separation and more abundant surface active sites [22, 23]. The exploitation of defects is an important way to improve the photocatalytic efficiency of g-C3N4 and has attracted increasing attention in experimental studies [24]. In this paper, g-C3N4 (with a triazine ring as the basic structural unit) with three kinds of defects (Figure 1) was selected as the research object, using time-dependent density functional theory (TD-DFT) with a set of extensive multidimensional visualization software to study the excited

state characteristics of intrinsic g-C3N4 and its C and N atomic defects. The effects of defect types and defect sites on the excited states of g-C3N4 were studied based on the hole-electron distribution and charge transfer matrix (CTM) of the excited states. Our results provide a theoretical basis and empirical reference for the design of g-C3N4 catalytic materials. 2. Theoretical method and computational detail The wave functions generated in quantum chemical calculations contain much information of great significance to the study of electronic structure and transition behavior. The electron excitation process can be expressed as “an electron leaves hole and goes to electron”, where hole and electron are real space functions [25, 26]. The excited state wave function can be written as occ vir

ϕ exe = ∑ wia Φ ia ≡ ∑∑ wia Φ ia i→a

i

a

where i and a respectively run over all occupied and all virtual molecular orbitals(MOs), similarly hereafter. Φ ia is the configuration state wave function corresponding to moving an electron from originally occupied MO i to virtual MO a. Where w and w' correspond to configuration coefficient of excitation and de-excitation. Hole and electron expressions can be written in the following form: hole hole (r ) + ρ cross (r ) ρ hole (r ) = ρ loc

= ∑ ( wia ) 2 ϕ i (r )ϕ i (r ) + ∑ i →a

∑ w w ϕ (r )ϕ

i → a j ≠i → a

a i

a j

i

j

(r )

ele ele (r ) + ρ cross (r ) ρ ele (r ) = ρ loc

= ∑ ( wia ) 2 ϕ a (r )ϕ a (r ) + ∑ i →a

∑w w ϕ

i → a i →b ≠ a

a i

Note that the notions used here occ vir

∑ ≡ ∑∑

i→a

i

a

occ occ vir

∑ ∑ ≡ ∑∑∑

i→a j ≠i →a

i

j ≠i

a

b i

a

(r )ϕ b (r )

hole Where ρ loc (r r ) = ∑ ( wia ) ϕiϕi − ∑ ( w'ia ) 2ϕiϕi 2

i→a

hole ρ cross (r ) = ∑

i ←a

∑ w w ϕ ϕ − ∑ ∑ w' w'

i →a j ≠i →a

a i

a j

i

j

i ←a j ≠i←a

a i

a j

ϕiϕ j

ele ρ loc (r r ) = ∑ ( wia ) ϕ aϕ a − ∑ ( w'ia ) 2ϕ aϕ a 2

i→a

ele ρ cross (r ) = ∑

i ←a

∑ w w ϕ ϕ − ∑ ∑ w' w'

i → a i →b ≠ a

a i

b i

a

b

i ← a i ←b ≠ a

a i

b i

ϕ aϕb

Where ϕ denotes MO wavefunction, r is the coordinate vector, i or j are occupied orbital labels, a or b are empty orbital labels, so ∑ represents each i →a

excitation configuration of the cycle, and

∑

represents each de-excitation

i ←a

configuration of the cycle. "loc" and "cross" stand for the contribution of local term and cross term to the hole and electron distribution, respectively. "local" means "oneself", the object is a transition orbital, that is, a transition orbital is purely its own contribution. The hole distribution (ρhole) and the electron distribution (ρele) are divided into two parts: a local term and a cross term. Local terms are generally dominant, reflecting the contribution of the configuration function itself, and the cross terms are not negligible, otherwise the quantitative is not accurate, which reflects the influence of the coupling between the configuration functions on the hole and electron distribution. Due to the orthonormality of MOs and the fact that the sum of square of all configuration coefficients is 1.0, it is clear that

∫ρ

hole

(r )dr = 1

∫ρ

ele

(r )dr = 1

This is an important property that any reasonable definition of hole and electron distribution should satisfy, it indicates that one electron is excited. In order to conveniently measure and discuss the characteristics of electron excitation by some quantitative numerical methods, we have calculated some indices which can reflect the characteristics of excited states, these indices can describe the characteristics of excited states more comprehensively. The overlap function between hole and electron distribution can be defined as [27]

S r ( r ) = ρ hole (r ) ρ ele (r )

To characterize overlapping extent of hole and electron, Sr index is defined as follows

Sr index = ∫ S r (r )dr ≡ ∫ ρ hole (r ) ρ ele (r )dr Centroid can be calculated to reveal most representative position of hole and electron distribution. For example, X coordinate of centroid of electron is written as X ele = ∫ xρ ele (r )dr

where x is X component of position vector r. The charge transfer (CT) length in X/Y/Z can be measured by distance between centroid of hole and electron in corresponding directions: [27, 28]

Dx = X ele − X hole

Dy = Yele − Yhole

Dz = Z ele − Z hole

The total magnitude of CT length is referred to as D index: D index = D ≡ ( Dx ) 2 + ( Dy ) 2 + ( Dz ) 2

The RMSD of hole and electron can be used to characterize their extent of spatial distribution. For example, X component of RMSD of hole is expressed as

σ hole , x =

∫ (x − X

hole

)2 ρ

hole

( r )dr

H measures average degree of spatial extension of hole and electron distribution in X/Y/Z direction, HCT is that in CT direction, and H index is an overall measure H λ = (σ ele,λ + σ hole ,λ ) 2

λ={x, y, z}

H CT = H ⋅ uCT H index = ( σ ele + σ hole ) 2

Where UCT is unit vector in CT direction and can be straightforwardly derived using centroid of hole and electron. The |σhole| and |σele| are referred to as σhole and σele indices, they measure overall RMSD of hole and electron, respectively. t index is designed to measure separation degree of hole and electron in CT direction: t index = D index − H CT

If t index<0, it implies that hole and electron is not substantially separated due to charge transfer. Clear separation of hole and electron distributions must correspond to evidently positive t index. All quantum chemical calculations were carried out using the Gaussian 09 software package [29]. The ground-state geometries of intrinsic g-C3N4 and defective g-C3N4 were optimized using the Gaussian 09 package with density functional theory (DFT), Becke's three-parameter hybrid exchange functional with Lee-Yang-Parr gradient-corrected correlation (B3LYP functional) and the 6-31G (d) basis set [30, 31]. Excited states were computed with time-dependent density functional theory (TD-DFT), the CAM-B3LYP functional and the 631G (d) basis set [32-36]. The long-range-corrected functional (CAM-B3LYP) was employed for the non-Coulomb part of the exchange functional. Based on the excited-state data calculated by Gaussian 09, the hole-electron distribution and charge transfer were investigated in detail using the Multiwfn 3.6 program [37]. The electron-hole distribution and Chole and Cele isosurface were visualized in VMD software [38].

Figure 1. The optimized structures of intrinsic g-C3N4 (a), and Type A (b), Type B (c), and Type C (d) gC3N4.

3. Results and discussion 3.1 Structural analysis

Geometrical optimization of intrinsic g-C3N4, and g-C3N4 with an interstitial N-atom defect (Type A), C-atom defect (Type B) and N-atom defect (Type C) in the ground state were performed using Gaussian 09 with DFT, the B3LYP functional, and the 6-31G (d) basis set. Their geometrical structures are shown in Figure 1, in which the key atoms and partial important geometric parameters of each structure are labeled. For convenience, all data are accurate to three decimal places. Figure 1 (a) shows the structure of intrinsic g-C3N4, which has a planar structure consisting of alternately arranged triazine rings and has the same size vacancies. With an N-atom defect on the interstitial of g-C3N4, a vacancy is created on the interstitial, as shown in Figure 1 (b). It can be seen from Figure 1 (a) that the structure of g-C3N4 has good symmetry, with two symmetrical bond angles, C2-N3-C4 and C14-N15-C16, of 119.720° and 119.719°, respectively. When the interstitial N-atom defect exists, the two angles are 119.556° and 119.554°, respectively. Hence, a very small reduction has taken place, so we can assume that the size of the angle has not changed. In other words, the

symmetry of these two angles does not change with the existence of interstitial N-atom defects. However, it can be seen that the distance between C8 and C10 is significantly reduced, and the distances between C8 and C10 in intrinsic and Type A g-C3N4 are 2.516 Å and 1.930 Å, respectively. This is due to the absence of interstitial N atoms causing the two triazine rings to be close to each other. According to the overall structure and angle change of Type A, the interstitial N-atom defect does not have a large effect on the structure of gC3N4, and the symmetry of the g-C3N4 structure does not change.

Figure 2. Localized orbital isosurface of Type B between N12 and N13.

Compared with intrinsic g-C3N4, the structure of Type B (Figure 1 (c)) is quite different. Firstly, because of the absence of a C atom in the triazine ring, the defective triazine ring is no longer a closed six-membered ring. Compared with the distance between N12 and N13 in g-C3N4, the corresponding distance in Type B is significantly reduced, from 2.519 Å to 1.390 Å, so that GaussView judges N12 and N13 to be connected by a single bond. To determine whether the bonds between N12 and N13 are really bonded, we performed a Mayer bond order analysis. The Mayer bond order between N12 and N13 was determined to be 0.967, which is very close to 1.0, so we can judge that there is a single bond between N12 and N13. To further determine the existence of a single bond between N12 and N13, orbital localization analysis was performed for Type B. Figure 2 shows an isosurface map of an occupied localized orbital between N12 and N13, which indicates that the two atoms are bonded. The decrease in distance between the N12 and N13 atoms is accompanied by changes in other bond angles. The bond angle between N5-C4-N12 increases from 123.040° to 131.351°, which is caused by the upward movement of N12. The reduction in distance between N12 and N13 is caused not only by the upward movement of the N12 atom but also by the downward movement of the

triazine ring where the N13 atom is located, as can be seen from the apparent tilt of the triazine ring above. The increase in the N1-C2-N3 bond angle from 117.419° to 124.873° indicates that the degree of tilt of the triazine ring is relatively large. With the N-atom defect (Figure 1 (d)), the size of the triangular vacancy between the triazine rings becomes larger. The bond angle C2-N3-C4 is reduced from 119.720° to 111.284° due to the inward contraction of C4. The angle between C8-N9-C10 decreases from 119.911° to 114.745°, which is due to the increase in angle C6-N7-C8. This indicates that due to the N-atom defect, the remaining atoms in the triazine ring diffuse outward. Interestingly, the increase in bond angle C6-N7-C8 is almost equal to the decrease in bond angle C8-N9-C10. Through analysis, we know that different defects have different effects on the structure of g-C3N4. The interstitial N-atom defect have little effect on the structure of g-C3N4, but the C-atom defect on the triazine ring (Type B) and Natom defect on the triazine ring (Type C) has a great influence and even changes the structure of g-C3N4.

Figure 3. Normalized absorption spectra of intrinsic and defective g-C3N4.

3.2 Absorption spectra

In order to explore the absorption properties of this material, electron transition data were calculated at the CAM-B3LYP/6-31G (d) level. Simulated absorption curves based on the calculated transitions are depicted in Figures 3 and 4. Figure 3 shows the normalized absorption spectra of intrinsic g-C3N4 and its defects. From the picture, we can see that intrinsic g-C3N4 and its defects have similar absorption profiles. There are two absorption peaks for each

absorption spectrum. One is sharp, with the most intense absorptions occurring at 301.7/241.8/243.7/273.1 nm for intrinsic g-C3N4/Type A/Type B/Type C, respectively. The other intense absorptions are at 254.9/307.6/343.3/327.1 nm, respectively. Compared with intrinsic g-C3N4, the most intense absorption peak gradually moves to the short wavelength direction due to its defects, while the second strong absorption peak moves to the longer wavelength direction. The simulated optical absorption spectra of intrinsic and defective g-C3N4 are shown in Figure 4. By comparing these, we found that the absorption intensity of defective g-C3N4 was higher than that of intrinsic g-C3N4; particularly that of Type B, which was greatly improved.

(a)

(b)

(c)

(d)

Figure 4. Optical absorption spectra of g-C3N4: intrinsic (a), Type A (b), Type B (c), and Type C (d)

3.3 Analysis of transition characteristics

When a molecule absorbs a certain amount of energy, the electrons are excited to a higher level. There could be some significant changes in the electron distribution of a molecule in an excited state; that is, the electrons will transition. With electron migration, the original positions of the electrons are

left empty, producing “holes”. Compared with intrinsic g-C3N4, the holeelectron distribution of the excited states of the defective g-C3N4 molecules changes significantly. Some excited states with special hole-electron distributions will provide corresponding reactive sites for photocatalytic reactions. To better illustrate the hole-electron distribution and charge transfer on the surface of intrinsic and defective g-C3N4, the structure was divided into 12 fragments. The divisions and sequence numbers of the fragments are shown in Figure 5, where different colors represent the atoms contained in different fragments.

Figure 5. Division and labeling of fragment.

Figure 6 shows the hole-electron distribution isosurface and charge transfer matrix (CTM) map of S13 (In this work,we use the form of capital S and digital subscript to express the corresponding excited state serial number, S13 represents the thirteenth excited state) and S32 for intrinsic g-C3N4, respectively. It can be seen from the isosurface diagrams that the distribution of electrons and holes are relatively large and cover almost the entire surface of the g-C3N4 molecule. However, the overlap of holes and electrons is very high, which indicates that their degree of separation is very small. From a quantitative point of view, the D index of S13 and S32 are very small, at 0.410 Å and 0.453 Å, respectively, indicating that the centroids of the holes and electrons are relatively close. Therefore, although the H index is very large, S13 and S32 belong to local excited states.

Figure 6. Hole-electron distribution maps and charge transition matrix maps of S13 and S32 for intrinsic gC3N4.

The heat map in Figure 6 represents the CTMs of S13 and S32, where the xaxis and y-axis represent the fragment numbers. The colored blocks on the diagonal represent the intra-fragment charge transfer of each fragment, and the non-diagonal blocks represent the inter-fragment charge transfers. We can see that the charge transfer of S13 is mainly concentrated between fragment 5 and fragment 6, and the red blocks 5 and 6 on the diagonal line represent strong intra-fragmental charge transfer in fragments 5 and 6, respectively. Of course, there are some charge transfers between other fragments but they are mutual and have little effect on the separation of electrons and holes. The CTM of S32 also shows that charge transfer occurs mainly in fragment 2. So, from the point of view of charge transfer, S13 and S32 still belong to local excited state. For Type A, the hole-electron distribution map (Figure 7) shows that the holeelectron distribution of S14 is very similar to that of S13 in intrinsic g-C3N4. For S45, the D index is very small (D index = 0.239 Å, Table 1), while the Sr value is relatively large (Sr index = 0.706). This indicates that the overlap between holes and electrons is very high, which means that there is little electron

migration for S45. It can also be seen from the CTM that charge transfer comes mainly from the intra-fragmental charge transfer of fragment 10. By analyzing the excited states of Type A, it can be concluded that a defect in interstitial N-atoms does not change the characteristics of the excited states. The D index is very small, the Sr index is very large, and the t index < 0, so the type of electronic excitation in the excited states has not changed and is also the local excited states.

Figure 7. Hole-electron distribution maps and charge transition matrix maps of S14 and S45 for Type A.

From the hole-electron distribution map of intrinsic and Type A g-C3N4, it can be seen that the distributions of holes and electrons are complex. The isosurface maps have many nodes and the distributions alternate with each other, which brings complexity to the investigation. To visualize the holes and electrons more easily, it is necessary to investigate the distance of charge transition and the hole-electron spatial distributions in another way. Chole and Cele functions are defined as follows [25-27]. The function behavior of Chole and Cele is similar to Gaussian function, they are highly smooth functions, the value asymptotically approaches zero from centroid of hole/electron.

 ( x − X ele ) 2 ( x − Yele ) 2 ( x − Z ele ) 2  − − Cele (r ) = Aele exp  −  2 2 2 2σele 2σele 2σele   ,x ,y ,z  ( x − X hole ) 2 ( x − Yhole ) 2 ( x − Z hole ) 2  Chole (r ) = Ahole exp − − −  2 2 2σhole 2σhole 2σ2hole, z  ,x ,y 

σ = Where i ,l

∫ (x − X ) i

2

ρ i (r )d (r )

,

i represents electron or hole, l={x, y, z}, x, y, z are three Cartesian components of coordinate vector r, σ index reflecting the breadth of the distribution of electrons or hole, Xi, Yi, Zi, represent the x, y, and z coordinates of electrons or holes, respectively. The factor A is introduced so that Chole and Cele are normalized. The central position of Cele and Chole isosurface is exactly the centroid position of electrons and holes.

Figure 8. The Chole and Cele isosurfaces of intrinsic and Type A g-C3N4.

Figure 8 shows a Cele and Chole isosurface diagram of the corresponding excited states of intrinsic and Type A g-C3N4. Blue and green isosurfaces represent holes and electrons, respectively. From these graphs, we can more intuitively see the overall distributions of electrons and holes. Although the distribution ranges of holes and electrons in these excited states are very large, they overlap greatly, which corresponds to the values of Sr and H index in

Table 1. The hole and electron distributions of S45 in these excited states almost completely overlap in Figure 8, which corresponds to the maximum Sr value of S45 in Table 1. Table 1. Indices that reflect the excited-state characteristics of intrinsic and defective g-C3N4.

Intrinsic g-C3N4

Type A Type B Type C

Transition state

Wavelength (nm)

Oscillator strength

D(Å)

Sr

H(Å)

t(Å)

S13

300.58

0.0084

0.410

0.508

5.270

-2.006

S32 S14 S45 S2 S49 S15 S35 S43

251.05 304.26 241.37 346.57 240.98 327.21 274.48 260.84

0.0053 0.0097 0.1800 0.2416 0.2337 0.2668 0.3394 0.0668

0.453 0.416 0.239 2.705 1.307 0.272 1.109 2.616

0.528 0.525 0.706 0.292 0.680 0.666 0.634 0.486

5.177 5.167 5.524 2.874 5.633 3.264 3.672 3.854

-2.732 -1.741 -2.728 0.568 -2.667 -1.975 -1.493 0.007

The hole-electron distributions and CTMs of g-C3N4 with C-atom defects are shown in Figure 9, which correspond to S2 and S49, respectively. Compared with intrinsic and Type A g-C3N4, the excited-state charge distributions and modes of charge transfer have changed significantly in Type B. From the charge distribution of S2, we can see that the holes and electrons are no longer distributed over the surface of g-C3N4 as before, but are mainly distributed on fragments 2 and 5. The holes and electrons are clearly separated and most of the holes are distributed on the triazine ring (fragment 5) where the defect is located, while the electrons are mainly distributed on fragment 2, which is adjacent to fragment 5. The t index is > 0 and the smaller value of Sr (see Table 1) also indicate that the holes and electrons are separated greatly, so we can judge that S2 of Type B belong to charge transfer excited state. From Table 1, we can see that the value of the D index of S2 is 2.705 Å, which indicates that the distance between the centroid of the hole and electron distributions are relatively large. This can also be seen intuitively from the Chole and Cele isosurface (Figure 10). From the CTM of S2, we can clearly see that charge transfer occurs almost exclusively between fragments 2 and 5, where electrons transfer from fragment 5 to fragment 2. Compared with S2, the ranges of the hole-electron distributions of S49 are larger than those of S2.

Figure 9. Hole-electron distribution maps and charge transition matrix diagrams of S2 and S49 for Type B.

Although the holes and electrons are partially separated, they still have an overall large overlap with a Sr value of 0.680. This feature can be seen intuitively in the Chole and Cele isosurface (Figure 10), so S49 belongs to local excited states. The same judgment can be made according to the indices of S49 in Table 1. From the CTM map of S49 in Figure 9, we can clearly see that the electron excitation of S49 is mainly multi-center local excited states. Similar to S2, the triazine ring in which the C atom is located has an outward transfer of electrons but no inward electron transfer.

Figure 10. The Chole and Cele isosurfaces of S2 (left) and S49 (right) for Type B g-C3N4.

According to the hole-electron distribution maps of S15 and S35 for Type C (Figure 11), the holes and electrons are almost entirely distributed in the triazine ring where the N-atom defect is located. It can be seen that even though the holes and electrons are distributed on the same triazine ring, they are distributed in different positions and do not overlap. It is obvious that the electrons are distributed on two C atoms which are connected to the defective N atom. The charge transfer of S15 is mainly within the fragment, in which electrons are transferred from one end of the atom (far away from the defective N) to two C atoms that are connected with the defective N atom. In S35, in addition to the charge transfer within fragment 5, electrons are mainly transferred from fragment 4 to fragment 5 where the defect is located. The charge distribution of S43 (Figure 11) with an oscillator strength of 0.0668 is different from those of S13 and S35. The holes and electrons are distributed on different fragments with a D index = 2.616 and t index > 0, which indicates that S43 is a charge transfer excited state, especially as electrons are concentrated on the triazine ring where the defect is located. By analyzing the charge distribution and charge transfer characteristics of intrinsic and defective g-C3N4 structures, we found that the excited state characteristics do not change for Type A. However, the C-atom or N-atom defects on the triazine ring have great influence on the excited state of intrinsic g-C3N4. The excited state corresponding to the absorption peak not only has a local excited state but also has a charge transfer excited state. Compared with intrinsic and Type A g-C3N4, electrons and holes are separated both in local excited states and charge transfer excited states in Type B and Type C, which can be clearly seen in the hole-electron distribution diagram. The charge transfer is no longer confined to intra-fragmental charge transfer and multicenter charge transfer between fragments, and the charge transfer between fragments is centered on the triazine ring where the defect is located. On the one hand, from the point of view of the charge distribution ranges, those of Types B and C are no longer mostly distributed over the whole surface of intrinsic g-C3N4 but are localized around the triazine ring where the defect is located. On the other hand, due to the different types of defects on the triazine

ring, the charge distribution on the triazine ring where the defect located is also different. For Type B, the holes are mainly distributed on the triazine ring where the defect is located. For Type C, a large number of electrons are distributed on the triazine ring where the defect is located, and the charge mainly accumulates on the C atom connected to the defective N atom.

Figure 11. Hole-electron distribution maps and charge transition matrix maps of S15, S35, and S43 for Type C g-C3N4.

4. Conclusions

In this work, based on DFT and TD-DFT methods with a B3LYP/6-31G (d)//CAM-B3LYP/6-31G (d) calculation level, the optimal structures and excited states of intrinsic and defective g-C3N4 were obtained. Compared with intrinsic g-C3N4, the absorption spectrum of defective g-C3N4 was red-shifted and the light absorption intensity was greatly improved, especially for C-atomdefect g-C3N4. In addition, the effects of defects on the hole-electron distributions and charge transfer directions in the excited state of g-C3N4 were also studied in detail. Due to the different defects, the charge distributions

showed different characteristics; as the defects appeared, the charge migration distance also increased. The existence of defects effectively promotes the separation and migration of photogenerated charges. The characteristics of charge distribution may provide active sites for photocatalytic reactions. Our results have potential significance in the design of g-C3N4 2D photocatalytic materials.

Acknowledgements This work was supported by the Innovative Talent Support Program of Liaoning Province (Grant No. LR2017062), the Liaoning Provincial Department of Education Project (Grant No. LFW201710), the Shenyang High-level Innovative Talents Program (Grant No. RC180227), the LiaoNing Revitalization Talents Program (Grant No. XLYC1807162)

References [1] X. Wang, K. Maeda, A. Thomas, K. et al., A metal-free polymeric photocatalyst for hydrogen production from water under visible light, Nat. Mater 8 (2009) 76. [2] K. Maeda, X. Wang, Y. Nishihara, et al., Photocatalytic activities of graphitic carbon nitride powder for water reduction and oxidation under visible light, J. Phys. Chem. C 113 (2009) 4940-4947. [3] G. Xin, Y. Meng, Pyrolysis, et al., Synthesized g-C3N4 for photocatalytic degradation of methylene blue, J. Chem-NY 2013 (2012). [4] H. Ji, F. Chang, X. Hu, et al., Photocatalytic degradation of 2, 4, 6-trichlorophenol over g-C3N4 under visible light irradiation, Chem. Eng. J 218 (2013) 183-190. [5] X.-H. Li, J. Zhang, X. Chen, et al., Condensed graphitic carbon nitride nanorods by nanoconfinement: promotion of crystallinity on photocatalytic conversion, Chem. Mater 23 (2011) 4344-4348. [6] A.B. Jorge, D.J. Martin, M.T. Dhanoa, et al., H2 and O2 evolution from water halfsplitting reactions by graphitic carbon nitride materials, J. Phys. Chem. C 117 (2013) 7178-7185. [7] F. Wu, Y. Liu, G. Yu, et al., Visible-light-absorption in graphitic C3N4 bilayer: enhanced by interlayer coupling, J. Phys. Chem. Lett 3 (2012) 3330-3334. [8] G. Ding, W. Wang, T. Jiang, et al., Highly selective synthesis of phenol from benzene over a vanadium-doped graphitic carbon nitride catalyst, ChemCatChem 5 (2013) 192-200. [9] H.-S. Zhai, L. Cao, X.-H. Xia, Synthesis of graphitic carbon nitride through pyrolysis of melamine and its electrocatalysis for oxygen reduction reaction, Chinese. Chem. Lett 24 (2013) 103-106. [10]S.-W. Cao, Y.-P. Yuan, J. Fang, et al., In-situ growth of CdS quantum dots on g-C3N4 nanosheets for highly efficient photocatalytic hydrogen generation under visible light irradiation, Int. J. Hydrogen. Energ 38 (2013) 1258-1266. [11]X. Zhou, B. Jin, R. Chen, et al., Synthesis of porous Fe3O4/g-C3N4 nanospheres as highly efficient and recyclable photocatalysts, Mater. Res. Bull 48 (2013) 1447-1452. [12]W. Liu, M. Wang, C. Xu, et al., Significantly enhanced visible-light photocatalytic activity of g-C3N4 via ZnO modification and the mechanism study, J. Mol. Catal. AChem 368 (2013) 9-15. [13]S. Zhao, S. Chen, H. Yu, et al., g-C3N4/TiO2 hybrid photocatalyst with wide absorption wavelength range and effective photogenerated charge separation, Sep. Purif. Technol 99 (2012) 50-54. [14]H. Xu, J. Yan, Y. Xu, et al., Novel visible-light-driven AgX/graphite-like C3N4 (X=Br, I) hybrid materials with synergistic photocatalytic activity, Appl. Catal. B-Environ 129 (2013) 182-193. [15]J. Yu, S. Wang, J. Low, et al., Enhanced photocatalytic performance of direct Zscheme g-C3N4-TiO2 photocatalysts for the decomposition of formaldehyde in air, Phys. Chem. Chem. Phys 15 (2013) 16883-16890. [16]S. Yan, Z. Li, Z. Zou, Photodegradation of rhodamine B and methyl orange over boron-doped g-C3N4 under visible light irradiation, Langmuir 26 (2010) 3894-3901. [17]Y. Zhou, L. Zhang, J. Liu, et al., Brand new P-doped g-C3N4: enhanced photocatalytic activity for H2 evolution and Rhodamine B degradation under visible light, J. Mater. Chem. A 3 (2015) 3862-3867. [18]Y. Bu, Z. Chen, Effect of oxygen-doped C3N4 on the separation capability of the photoinduced electron-hole pairs generated by O-C3N4@ TiO2 with quasi-shell-core nanostructure, Electrochim. Acta 144 (2014) 42-49. [19]Q. Li, N. Zhang, Y. Yang, et al., High efficiency photocatalysis for pollutant degradation with MoS2/C3N4 heterostructures, Langmuir 30 (2014) 8965-8972.

[20]X.L. Wang, W.Q. Fang, Y. Yao, et al., Switching the photocatalytic activity of g-C3N4 by homogenous surface chemical modification with nitrogen residues and vacancies, RSC. Adv 5 (2015) 21430-21433. [21]Y. Meng, J. Shen, D. Chen, et al., Photodegradation performance of methylene blue aqueous solution on Ag/g-C3N4 catalyst, Rare. Metals 30 (2011) 276-279. [22]Q. Han, B. Wang, J. Gao, et al., Atomically thin mesoporous nanomesh of graphitic C3N4 for high-efficiency photocatalytic hydrogen evolution, ACS. Nano 10 (2016) 2745-2751. [23]F. Liang, Y. Zhu, Enhancement of mineralization ability for phenol via synergetic effect of photoelectrocatalysis of g-C3N4 film, Appl. Catal. B-Environ 180 (2016) 324329. [24]L. Kong, X. Mu, X. Fan, et al., Site-selected N vacancy of g-C3N4 for photocatalysis and physical mechanism, Appl. Mater. Today 13 (2018) 329-338. [25]Lu, T. Multiwfn Manual, version 3.6, Section 4.13.6. http://sobereva.com/multiwfn.2019. [26]M.J. Peach, C.R. Le Sueur, K. Ruud, et al., TDDFT diagnostic testing and functional assessment for triazene chromophores, Phys. Chem. Chem. Phys 11 (2009) 44654470. [27]T. Le Bahers, C. Adamo, I. Ciofini, A qualitative index of spatial extent in chargetransfer excitations, J. Chem. Theory. Comput 7 (2011) 2498-2506. [28]C.A. Guido, P. Cortona, B. Mennucci, et al., On the metric of charge transfer molecular excitations: a simple chemical descriptor, J. Chem. Theory. Comput 9 (2013) 3118-3126. [29]M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, and D. J. Fox, Gaussian, Inc., Wallingford CT, 2013. [30]W. Kohn, L.J. Sham, Self-consistent equations including exchange and correlation effects, Phys. Rev 140 (1965) A1133-A1138. [31]V.A. Rassolov, J.A. Pople, M.A. Ratner, et al., 6-31G* basis set for atoms K through Zn, J. Chem. Phys 109 (1998) 1223-1229. [32]F. Furche, R. Ahlrichs, Adiabatic time-dependent density functional methods for excited state properties, J. Chem. Phys 117 (2002) 7433-7447. [33]J.L. Whitten, Coulombic potential energy integrals and approximations, J. Chem. Phys 58 (1973) 4496-4501. [34]E. Tapavicza, G.D. Bellchambers, J.C. Vincent, et al., Ab initio non-adiabatic molecular dynamics, Phys. Chem. Chem. Phys 15 (2013) 18336-18348. [35]T. Yanai, D.P. Tew, N.C. Handy, A new hybrid exchange-correlation functional using the Coulomb-attenuating method (CAM-B3LYP), Chem. Phys. Lett 393 (2004) 51-57. [36]J.W. Ochterski, G.A. Petersson, J.A. Montgomery Jr, A complete basis set model chemistry. V. Extensions to six or more heavy atoms, J. Chem. Phys 104 (1996) 25982619.

[37]T. Lu, F. Chen, Multiwfn: a multifunctional wavefunction analyzer, J. Comput. Chem 33 (2012) 580-592. [38]W. Humphrey, A. Dalke, K. Schulten, VMD: visual molecular dynamics, J. Mol. Graph. Model 14 (1996) 33-38.

Highlight: 1. the excited-state characteristics of intrinsic g-C3N4 and g-C3N4 with C- and N-atom defects are analyzed. 2. the hole-electron distributions and charge transfer directions are determined. 3. the characteristics of electron excitation is discussed.

Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Peng Song, Ph.D Liaoning University Shenyang 110036, China Phone and fax: 86-24-62202258 E-mail: [email protected]