First-Principles Calculation of the Bulk Photovoltaic Effect in KNbO$_{3}$ and (K,Ba)(Ni,Nb)O$_{3-delta}$

The connection between noncentrosymmetric materials structure, electronic structure, and bulk photovoltaic performance remains not well understood. In particular, it is still unclear which photovoltaic (PV) mechanism(s) are relevant for the recently demonstrated visible-light ferroelectric photovoltaic (K,Ba)(Ni,Nb)O$_{3-delta}$. In this paper, we study the bulk photovoltaic effect (BPVE) of (K,Ba)(Ni,Nb)O$_{3-delta}$ and KNbO$_{3}$ by calculating the shift current from first principles. The effects of structural phase, lattice distortion, oxygen vacancies, cation arrangement, composition, and strain on BPVE are systematically studied. We find that (K,Ba)(Ni,Nb)O$_{3-delta}$ has a comparable BPVE with that of the broadly explored BiFeO$_{3}$, but for a much lower photon energy. In particular, the Glass coefficient of (K,Ba)(Ni,Nb)O$_{5}$ in a simply layered structure can be as large as 12 times that of BiFeO$_{3}$. Furthermore, the nature of the wavefunctions dictates the eventual shift current yield, which can be significantly affected and engineered by changing the O vacancy location, cation arrangement, and strain. This is not only helpful for understanding other PV mechanisms that relate to the motion of the photocurrent carriers, but also provides guidelines for the design and optimization of PV converters.

First principles calculations of the Shift Current Bulk Photovoltaic Effect in Ferroelectrics

We calculate the bulk photovoltaic response of the ferroelectrics BaTiO$_3$ and PbTiO$_3$ from first principles by applying shift current theory to the electronic structure from density functional theory. The first principles results for BaTiO$_3$ reproduce eperimental photocurrent direction and magnitude as a function of light frequency, as well as the dependence of current on light polarization, demonstrating that shift current is the dominant mechanism of the bulk photovoltaic effect in BaTiO$_3$. Additionally, we analyze the relationship between response and material properties in detail. The photocurrent does not depend simply or strongly on the magnitude of material polarization, as has been previously assumed; instead, electronic states with delocalized, covalent bonding that is highly asymmetric along the current direction are required for strong shift current enhancements. The complexity of the response dependence on both external and material parameters suggests applications not only in solar energy conversion, but to photocatalysis and sensor and switch type devices as well.