Download files
Abstract
Solar energy is regarded as a promising alternative renewable energy to replace traditional energy sources, such as fossil fuels. However, the problem with solar energy is finding a way to store it. It can be stored as hydrogen produced from water using sunlight energy through some methods, such as photocatalytic and photoeletrochemical water splitting. For this, photoactive semiconductors that can efficiently absorb sunlight and use the solar energy to split water are needed. TiO2 is one semiconductor material that can do this, but it has low efficiency for solar-to-hydrogen conversion because of its large band gap, which prevents light absorption at visible wavelengths. Monoclinic BiVO4 is a potential alternative photoactive material and it is active for water splitting under visible light. However, the band gap of pure monoclinic BiVO4 (~2.4 eV) is still larger than the ideal value (2.0 eV) and the poor electron transport ability, moderate water oxidation kinetics and low carrier mobility also limit the application of BiVO4. Doping is a potential strategy to further decrease the band gap of BiVO4.
In this work, several metal dopants are co-doped with nitrogen and fluorine to investigate their effects on the structural and electronic properties of BiVO4 by Density Functional Theory (DFT) calculations with the CRYSTAL code. Three different metal dopants (W, Mo and Sn) are tested at both the Bi site and V site, co-doped with N/F, such that the two dopant atoms are either bonded together or separated from each other.
It is found that co-doping can, in some cases, lead to narrower band gaps compared with undoped BiVO4. The smallest band gap (1.92 eV) is obtained when replacing one V atom with W and one O with N such that the two dopants are separated from each other (WV-Nseparated). In general, clear band gaps (i.e. no mid-gap states) are obtained when one dopant atom has fewer valence electrons than the atom it replaces and the other has more valence electrons, such that one dopant compensates the other. In addition, even though it may intuitively be expected that separation of the co-dopants should give weaker charge transfer effects than when the co-dopants are bonded together, the opposite is sometimes observed. For example, in the case of co-doping BiVO4 with W (replacing V) and N (replacing O), it is found that the separated condition results in more charge transfer related to greater interaction between the dopants and their neighbouring atoms than in the condition when the dopants are bonded.
It is found that the N, W and Mo dopants often have significant contributions to the conduction band (CB) or the valence band (VB), while the F and Sn dopants generally have very little or even no contribution to the either the CB or the VB.
Doping causes changes to the atomic charges of the Bi, V and O atoms. Relationships are identified between the atomic charges and the interatomic bond lengths, and thus it can be concluded that the structural distortion caused by the dopants is an important factor in determining the charge transfer that occurs between the dopants and their surrounding atoms and hence on the electronic properties of doped BiVO4. Moreover, V-O bond lengths have stronger effects than Bi-O bond lengths.
Calculated formation energies show that formation of most single- and co-doped BiVO4 is thermodynamically unfavourable, indicating that energy is required to insert the dopant atoms into pure BiVO4. However, the binding energies for most co-doped BiVO4 indicate that the co-doped structures are more stable than the corresponding single-doped cases. The exception is the WV-N co-doped case, which is more unstable compared with the two corresponding single-doped cases.
Overall, it is concluded that co-doping can lead to narrower band gaps for BiVO4, which indicates potential for enhanced photocatalytic performance under visible light.