Xu
Zhang
,
An
Chen
,
Zihe
Zhang
,
Menggai
Jiao
and
Zhen
Zhou
*
School of Materials Science and Engineering, National Institute for Advanced Materials, Institute of New Energy Material Chemistry, Computational Centre for Molecular Science, Nankai University, Tianjin 300350, P. R. China. E-mail: zhouzhen@nankai.edu.cn; Fax: +86 22 23498941; Tel: +86 22 23503623
First published on 11th September 2018
Photocatalytic water splitting is a promising method for the production of clean energy and searching for efficient photocatalysts has received extensive attention. Fabricating type-II heterojunctions is an effective approach to improve the photocatalytic efficiency. Based on the band edge positions and lattice parameters, we found that several kinds of monochalcogenide monolayers can be used to fabricate type-II heterojunctions with C_{2}N monolayers. C_{2}N/GaTe and C_{2}N/InTe van der Waals (vdW) heterojunctions were investigated as potential photocatalysts for water splitting by means of first-principles computations. Both are type-II heterojunctions, and could promote the efficient spatial separation of electron–hole pairs. Their band edges straddle water redox potentials, satisfying the requirements for photocatalytic water splitting. Besides, the high carrier mobility of C_{2}N/GaTe and C_{2}N/InTe heterojunctions implies that the transfer of carriers to reactive sites is easy, and the recombination probability of photo-generated carriers is reduced. The Gibbs free energy calculations indicate that C_{2}N/GaTe and C_{2}N/InTe heterojunctions, especially C_{2}N/InTe, exhibit high catalytic performance towards hydrogen and oxygen evolution reactions. Particularly, C_{2}N/InTe exhibits a direct band gap with strong absorption in both visible and near ultraviolet regions, indicating that it is a very promising candidate for photocatalytic water splitting. This work would provide a new idea for the development of type-II heterojunctions for photocatalytic water splitting.
Nanoscale materials with adjustable dimensions and shapes are one of the most promising photohydrolytic catalysts. Among various kinds of nanoscale materials, two-dimensional (2D) materials possess two main inherent advantages which can be used to improve the photocatalytic efficiency for water splitting.^{11,13–15} First, 2D materials have a high surface to volume ratio, exhibiting a large specific surface area available for photocatalytic reactions and promoting the adsorption of reactants. Besides, the distance that photo-generated electrons and holes have to migrate in 2D materials is short, reducing the possibility of electron–hole recombination, thereby potentially enhancing photocatalytic activity. With the rapid development of high-performance computations, various 2D materials were designed and predicted to be promising photocatalysts for overall water splitting.^{11,12,15,16} The computations have promoted the development of photocatalysts and some 2D materials have been proved to be good photocatalysts for water splitting in experiments, such as g-C_{3}N_{4} (ref. 17) and MPS_{3} (M = Fe or Mn).^{18,19}
Besides, fabricating van der Waals (vdW) heterojunctions is an effective method to further improve the photocatalytic performance.^{20,21} Especially, in type-II heterojunctions, both the conduction band minimum (CBM) and valence band maximum (VBM) of one layer are lower than that of the other layer. Therefore, the lowest energy states of electrons and holes are on different layers, ensuring the effective separation of photo-generated electrons and holes. In our previous reports, we proposed a high-throughput computational material design framework to search for 2D photocatalysts and rationally design type-II heterojunctions for water splitting based on the Materials Project Database.^{15} Various kinds of type-II heterojunctions were successfully designed. However, due to the rapid development of 2D materials, several novel kinds of materials, which have not been included in materials databases, have been successfully prepared in experiments. Therefore, predicting type-II heterojunctions containing novel kinds of 2D materials is necessary to evaluate their photocatalytic activity.
Recently, a novel kind of 2D material, named C_{2}N, has been successfully prepared by simple wet chemical reactions.^{22} By means of first-principles computations, C_{2}N monolayers were predicted to be a promising photocatalyst for water splitting.^{23} In experiments, vdW heterojunctions composed of C_{2}N and aza-conjugated microporous polymers were prepared, which exhibit good photocatalytic performance for overall water splitting.^{24} Enlightened by the computational and experimental results, we screened the 2D monolayer materials in our previous work^{15} to explore the materials which could be used to fabricate type-II heterojunctions with C_{2}N monolayers to further improve their photocatalytic performance. The screening criteria are consistent with our previous report.^{15} First, the CBM and VBM of the layer should be lower or higher than that of C_{2}N, respectively. Then the lattice parameters of the layer should be close to those of C_{2}N (lattice mismatch <10%), which could reduce the strain, improving the stability. This could also be achieved by constructing supercells. Considering the computational cost, small supercells, such as 2 × 2 × 1 supercells, were considered. Our results indicate that transition metal monochalcogenides and dichalcogenides could be used to fabricate type-II vdW heterojunctions with C_{2}N. Recently, C_{2}N/WS_{2} vdW type-II heterojunctions were proposed as promising photocatalysts for water splitting.^{25}
Therefore, in this work, we focused on monochalcogenide/C_{2}N heterojunctions. Based on the calculated band structures,^{26} besides GaTe which was proposed in our previous report,^{15} some other monochalcogenides, including GaSe and InTe, were also investigated. By means of first-principles computations, we designed GaSe/C_{2}N, GaTe/C_{2}N and InTe/C_{2}N vdW heterojunctions and investigated their stability, electronic, optical properties, carrier mobility and catalytic performance towards the hydrogen evolution reaction (HER) and oxygen evolution reaction (OER). The results indicate that GaTe/C_{2}N and InTe/C_{2}N are type-II heterojunctions, implying the efficient separation of holes and electrons, thereby preventing their recombination. The exciton binding energies of GaTe/C_{2}N and InTe/C_{2}N type-II heterojunctions are lower than that of C_{2}N, indicating better separation of the bound holes and electrons to free charge carriers. Both GaTe/C_{2}N and InTe/C_{2}N exhibit high carrier mobility, indicating high photocatalytic activity. Besides, InTe/C_{2}N exhibits an excellent catalytic performance towards the HER and OER, further proving InTe/C_{2}N type-II heterojunctions to be a promising candidate for overall photocatalytic water splitting.
The binding energy of vdW heterojunctions was calculated based on:
E_{binding} = E_{heterojunction} − E_{C2N} − E_{MX} |
For inorganic semiconductors, the phonon scattering dominates the intrinsic mobility which could be described by deformation potential (DP) theory.^{35} DP theory has been successfully used in predicting the carrier mobility of 2D semiconductors, including graphene,^{36} phosphorene,^{37} MoS_{2},^{38} MXenes^{39} and so on.^{16,40} Therefore, the carrier mobility (μ_{2D}) of isolated C_{2}N, GaTe, and InTe monolayers and heterojunctions could be predicted by using the phonon limited scattered mode on the basis of effective mass approximation with DP theory according to:
To investigate the optical absorption, the imaginary part of the dielectric function ε_{2} was calculated based on the following equation:^{41}
The HER catalytic performance of the heterojunctions was evaluated by calculating the Gibbs free energy change of the adsorption of atomic hydrogen (ΔG_{H}):
ΔG_{H} = ΔE_{H} + ΔE_{ZPE} − TΔS_{H} |
Under acidic conditions, the mechanism for the OER process is:
ΔG = ΔE + ΔE_{ZPE} − TΔS + ΔG_{U} + ΔG_{pH} |
In order to evaluate the efficiency of the separation of photogenerated charge carriers, the exciton binding energies (E_{b}) were calculated based on the Mott–Wannier hydrogenic model. In a previous report, the exciton binding energies of C_{2}N containing vdW heterojunctions were calculated based on:^{25}
Then the electronic properties of these monolayers were investigated at the HSE06 level as shown in Fig. S2.† The C_{2}N monolayer exhibits a direct band gap of 2.49 eV at the Γ point of the Brillouin zone, which agrees well with those in previous reports.^{23,46} The MX monolayers (GaSe, GaTe and InTe) exhibit indirect band gaps of 2.97, 2.21 and 2.21 eV, respectively. For GaSe and InTe monolayers, the CBM is at the Γ point while the VBM is located between the Γ and K points. For the GaTe monolayer, the CBM is located at the M point. These computational results are consistent with those in the previous report,^{26} indicating the reliability of our results.
The geometric structures, lattice parameters and binding energies of the most stable structure for the three heterojunctions are summarized in Fig. 1 and Table S3.† The distance between the C_{2}N and MX monolayers in these heterojunctions is about 3.40 Å (Table S3†), implying that the vdW effect plays the primary role in the interlayer interaction. Due to the large lattice mismatch for C_{2}N/GaSe (Table S3†), the binding energy is positive, indicating that it is unstable. For C_{2}N/GaTe and C_{2}N/InTe, their lattice mismatch is lower than 3%, ensuring the experimental feasibility. Besides, the binding energies of C_{2}N/GaTe and C_{2}N/InTe are −10.8 and −9.6 meV Å^{−2}, respectively, which are comparable to those of some typical vdW crystals, including MoS_{2} (−26.0 meV Å^{−2})^{47} and graphene (−12.0 meV Å^{−2}),^{48} and some vdW heterojunctions, such as GaSe/graphene (−18.4 meV Å^{−2})^{49} and C_{3}N_{4}/MoS_{2} (−17.8 meV Å^{−2}).^{50} These results further indicate that C_{2}N/GaTe and C_{2}N/InTe are stable. C_{2}N/GaTe and C_{2}N/InTe heterojunctions were investigated in the following computations.
The band gaps of C_{2}N/GaTe and C_{2}N/InTe heterojunctions are 1.43 and 1.52 eV, respectively, which are lower than those of isolated monolayers. The CBM of C_{2}N/GaTe and C_{2}N/InTe heterojunctions both are located at the Γ point. The VBM of C_{2}N/GaTe is located between the Γ and K points while the VBM of C_{2}N/InTe is at the Γ point. Therefore, C_{2}N/InTe transfers to a direct band gap semiconductor while C_{2}N/GaTe still possesses an indirect band gap. As shown in Fig. 2, the CBM of the heterojunctions is mainly contributed by C_{2}N while the VBM mostly comes from the MX. Therefore, both C_{2}N/GaTe and C_{2}N/InTe possess a typical type-II band alignment, which promotes the efficient spatial separation of electron–hole pairs, overcoming the electron–hole recombination issue. The conduction band offset (CBO) and valence band offset (VBO) were calculated to be 0.65 and 1.00 eV for C_{2}N/GaTe and 0.79 and 0.92 eV for C_{2}N/InTe. The photogenerated electrons in the CBM of GaTe or InTe could transfer to the CBM of C_{2}N by the chemical potential difference of CBO while the photogenerated holes in the VBM of C_{2}N can migrate to the VBM of GaTe or InTe via the chemical potential difference of VBO. The results further ensure the effective spatial separation of electron–hole pairs.
Besides, the partial density of states (PDOS) was also calculated to gain deep insight into the composition of the VBM and CBM of the heterojunctions as shown in Fig. S4.† The CBM mainly originates from the p states of C and N atoms while the VBM is contributed by the p states of Te and Ga (or In) atoms.
A fundamental requirement for photocatalysts for overall water splitting is that the band edges should straddle the water redox potentials. That is to say, the VBM must be lower than the oxidation potential of O_{2}/H_{2}O and the CBM should be higher than the reduction potential of H^{+}/H_{2}. Besides, the water redox potential is influenced by the pH value.^{12,15} Here, the reduction potential of H^{+}/H_{2} is calculated by using: and the oxidation potential of O_{2}/H_{2}O is . The schematic diagram for the band edge positions relative to the vacuum level of the isolated monolayers and heterojunctions is shown in Fig. 3. The band edges of all the isolated monolayers (C_{2}N, GaSe, GaTe and InTe) and C_{2}N/InTe heterojunction straddle the redox potentials of water at pH = 0. However, the VBM of the C_{2}N/GaTe heterojunction is higher than the oxidation potential of O_{2}/H_{2}O at pH = 0. An appropriate pH value of 1.60–4.98 is needed for the C_{2}N/GaTe heterojunction for photocatalytic water splitting. The results suggest that the constructed C_{2}N/GaTe and C_{2}N/InTe heterojunctions are promising candidates for photocatalytic water splitting without an external bias voltage at appropriate pH. Interestingly, the C_{2}N/InTe heterojunction exhibits a direct band gap.
Fig. 3 Band edge positions of isolated monolayers and heterojunctions. The redox potentials of water splitting are represented by orange dashed lines. |
Compared with the isolated monolayers, the band edge positions of the CBM and VBM in the heterojunctions are located closer to the redox potentials of water, suggesting higher photocatalytic efficiency and optical absorption.^{51} In order to investigate the optical absorption, the average absorption coefficients of x, y and z directions were calculated as shown in Fig. 4. Besides, the imaginary parts of the dielectric function are also shown in Fig. S5.† As shown in Fig. 4 and S5,† the C_{2}N monolayer possesses absorption peaks in the visible spectrum while GaTe and InTe monolayers exhibit broad absorption in the near ultraviolet region. Fascinatingly, C_{2}N/GaTe and C_{2}N/InTe heterojunctions exhibit strong absorption in both visible and near ultraviolet regions with an intensity of 10^{5} cm^{−1}. Therefore, both the C_{2}N/GaTe and C_{2}N/InTe heterojunctions are efficient light harvesting photocatalysts.
The effective masses of electrons and holes are computed by fitting parabolic functions to the CBM and VBM along the transport direction. The and along the directions of x and y for the isolated C_{2}N, GaTe and InTe monolayers as well as C_{2}N/GaTe and C_{2}N/InTe heterojunctions are shown in Tables S4† and 1. The of the C_{2}N monolayer is much higher than the of C_{2}N, indicating much lower hole mobility in the C_{2}N monolayer, consistent with previous reports.^{25} For GaTe and InTe, the is also larger than , which agrees well with previously reported data,^{53,54} indicating the credibility of our results.
|m*| | C | |E_{1}| | μ _{2D} | ||
---|---|---|---|---|---|
C_{2}N/GaTe | e_{x} | 0.82 | 222.11 | 2.60 | 685.67 |
h_{x} | 1.06 | 222.11 | 1.55 | 1163.21 | |
e_{y} | 0.38 | 219.03 | 2.55 | 3229.29 | |
h_{y} | 0.86 | 219.03 | 1.82 | 1251.53 | |
C_{2}N/InTe | e_{x} | 0.39 | 207.12 | 3.35 | 1744.56 |
h_{x} | 0.24 | 207.12 | 8.84 | 613.05 | |
e_{y} | 0.42 | 207.04 | 3.54 | 1330.34 | |
h_{y} | 0.31 | 207.04 | 8.88 | 390.98 |
To calculate the in-plane stiffness and DP constant, the relative energy and band edge positions of the CBM and VBM as a function of the uniaxial strain along the transport directions are shown in Fig. 5 and S7,† respectively. Based on the calculated effective masses, in-plane stiffness and DP constant, the carrier mobility of the isolated monolayers and heterojunctions could be obtained and are shown in Tables S4† and 1. In heterojunctions, the electron and hole mobility can become very high, which could be attributed to the improved in-plane stiffness and more dispersed valence band. The carrier mobility of C_{2}N/GaTe and C_{2}N/InTe is higher than that of many 2D semiconductors, such as MoS_{2} (ref. 38) and BN monolayers,^{55} and comparable to that of MXenes^{39} and phosphorene.^{56} The improved carrier mobility indicates that the transfer of carriers to reactive sites is easier during the photocatalytic process. Besides, due to the short distance for the photogenerated electrons and holes to migrate in 2D nanosheets, the recombination of photogenerated carriers is further reduced. Therefore, it can be inferred that the photocatalytic activity of the C_{2}N/MX heterojunctions would be very good.
To further evaluate the separation of the charge carrier, the exciton binding energies (E_{b}) were also calculated as shown in Table S5.† The E_{b} of C_{2}N was calculated to be 1.24 eV, which is consistent with those in previous reports.^{25,57} C_{2}N/GaTe and C_{2}N/InTe exhibit much lower E_{b} values of 0.31 and 0.13 eV, respectively, indicating easier separation of carriers.
Since the CBM of C_{2}N/GaTe and C_{2}N/InTe heterojunctions is mainly contributed by C_{2}N while the VBM mostly comes from the MX, the HER performance of the C_{2}N layer and the OER catalytic activity of the MX layer were calculated. The ΔG_{H} with different hydrogen coverages on the C_{2}N layer is shown in Fig. 6.
Fig. 6 ΔG_{H} with different hydrogen coverages on the C_{2}N layer in (a) C_{2}N/GaTe and (b) C_{2}N/InTe heterojunctions. |
The ΔG_{H} of the first two hydrogen atoms is too negative, indicating that the bonding is too strong. The ΔG_{H} of the third hydrogen could reach 0.08 and 0.03 eV for C_{2}N/GaTe and C_{2}N/InTe, respectively, which are very close to zero, implying that C_{2}N/GaTe and C_{2}N/InTe are potentially excellent catalysts for the HER at an appropriate H coverage. The fourth hydrogen atom adsorbed on the C_{2}N/GaTe and C_{2}N/InTe heterojunctions would be close to the MX layer (Fig. S8†), leading to a positive ΔG_{H}.
Then the OER performance of MX in the heterojunctions was investigated. In order to calculate the limiting step of the OER, the free energy of each elementary step (ΔG_{1} − ΔG_{4}) was computed. The free energy diagrams of OER elementary steps on MX layers are shown in Fig. 7.
For both C_{2}N/GaTe and C_{2}N/InTe heterojunctions, ΔG of the first three steps is positive while the last step is negative. The third step is the rate-limiting step with a ΔG of 2.70 and 2.17 eV for C_{2}N/GaTe and C_{2}N/InTe, respectively. At the standard reduction potential of water oxidation (1.23 eV), the ΔG of the rate-limiting step will decrease to 0.94 eV for C_{2}N/InTe, which is comparable to and even lower than that of many TM@C catalysts or metal oxides.^{66}
The stability of C_{2}N in water has been experimentally confirmed^{24} while the MX has been proven to be stable by calculating their enthalpy of solvation.^{26} Therefore, the C_{2}N/MX heterojunctions would be stable in water. Overall, both C_{2}N/GaTe and C_{2}N/InTe are type-II heterojunctions with strong absorption in both visible and near ultraviolet regions and high carrier mobility. Especially, C_{2}N/InTe is a direct-band-gap heterojunction and exhibits excellent catalytic performance towards both the HER and the OER, indicating that it is a very promising photocatalyst for overall water splitting.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c8na00084k |
This journal is © The Royal Society of Chemistry 2019 |