A Review on Visible Light Active Perovskite-Based Photocatalysts

Perovskite-based photocatalysts are of significant interest in the field of photocatalysis. To date, several perovskite material systems have been developed and their applications in visible light photocatalysis studied. This article provides a review of the visible light (λ > 400 nm) active perovskite-based photocatalyst systems. The materials systems are classified by the B site cations and their crystal structure, optical properties, electronic structure, and photocatalytic performance are reviewed in detail. Titanates, tantalates, niobates, vanadates, and ferrites form important photocatalysts which show promise in visible light-driven photoreactions. Along with simple perovskite (ABO3) structures, development of double/complex perovskites that are active under visible light is also reviewed. Various strategies employed for enhancing the photocatalytic performance have been discussed, emphasizing the specific advantages and challenges offered by perovskite-based photocatalysts. This review provides a broad overview of the perovskite photocatalysts, summarizing the current state of the work and offering useful insights for their future development.


Introduction
Photocatalysis has long been studied for clean energy and environmental applications. Over the past two decades, the number of applications based on photocatalysis has increased sharply, while a wide range of materials systems have been developed [1][2][3][4]. Photocatalysis has been of particular interest in the production of hydrogen from water using solar energy [5]. Further, conversion of CO2 to hydrocarbons (fuels) is also of significant interest, as it is a solution to reduce CO2 emissions across the globe [6,7]. Apart from the clean energy generation, photocatalysis has several promising applications in the environmental field. Some of the applications include degradation of volatile organic compounds (VOC) for water treatment [8], germicide and antimicrobial action [9][10][11], de-coloration of industrial dyes [12][13][14], nitrogen fixation in agriculture [15], and removal of NOx/SOx air pollutants [16][17][18][19]. These applications have driven the development of variety of materials systems which are suitable for specific applications. Although TiO2-based materials are the most studied for photocatalytic applications, ternary and other complex oxide systems have been increasingly explored as photocatalysts. Among the various classes of materials studied, perovskites-based photocatalysts have unique photophysical properties and offer distinct advantages.
Perovskites are the class of compounds presenting the general formula ABO3. Generally, in this crystal structure, the A site is occupied by the larger cation, while the B site is occupied by the smaller cation. Perovskites are one of the most important families of materials exhibiting properties suitable for numerous technological applications [20]. Perovskite compounds such as PbZrO3, BaTiO3, PbTiO3 are most commonly used piezoelectric compounds [21]. BiFeO3 thin films show multiferroic behavior [22], while compounds such as SrTiO3 have shown excellent photocatalytic properties [23,24]. The origin of such properties lies in the crystal structure of perovskites. The perovskite crystal structure has corner connected BO6 octahedra and 12 oxygen coordinated A cations, located in between the eight BO6 octahedra ( Figure 1). The perfect structure of the octahedral connection results in a cubic lattice. However, depending on the ionic radii and electronegativity of the A and B site cations, tilting of the octahedra takes place, which gives rise to lower symmetry structures. As seen from the crystal structure, B site cations are strongly bonded with the oxygen (or any other anion) while, A site cations have relatively weaker interactions with oxygen. Depending on the type of the cations occupying the lattice sites, these interactions could be altered to yield the different perovskite crystal geometries.  For example, different degrees of tilting of the octahedra give rise to different crystal fields, which result in different electronic and optical properties. The degrees of tilting may affect the band structure, electron and hole transport properties, photoluminescence, and dielectric behavior [25,26]. From the point of view of photocatalysis, perovskite structures may offer significant advantages over the corresponding binary oxides for several reasons. Firstly, perovskites could offer favorable band edge potentials which allow various photoinduced reactions. For example, as compared to the binary oxides, several perovskites have sufficiently cathodic conduction band (CB) energies for hydrogen evolution. Secondly, A and B site cations in the lattice give a broader scope to design and alter the band structure as well as other photophysical properties. In the case of double perovskites such as A2B2O6, stoichiometric occupation of two cations at the B site is known to be beneficial for visible light photocatalysis. Thirdly, some studies have shown that it is possible to combine the effects such as ferroelectricity or piezoelectricity with the photocatalytic effect to benefit the photocatalytic activity.
Perovskite photocatalysts have been studied to a great extent because of their promise for being visible light active. A review of the present work on the visible light driven perovskite photocatalyts is essential to provide a broad overview and possible future directions. Shi et al. reported a general review of perovskite photocatalysts active under UV and visible light [27]. The current review article is focused on visible light active perovskite compounds. We emphasize the strategies used to develop or enhance the visible light absorption and subsequent photocatalytic activities. Further, we attempt to shed some light on the underlying principles specific to the perovskite crystal structure which play an important role in the photocatalytic activity, suggesting potential areas in the field where further work is needed. The first section of the article discusses the mechanism and thermodynamics of some of the most important photocatalytic reactions, while the later section reviews the material systems in detail. In the current review, perovskites are broadly divided into simple (ABO3 type) perovskites and complex perovskites (double, layered, etc.).

Overview of Photocatalytic Reactions
Photocatalysis is a process that utilizes the energy input from incident radiation and the catalytic properties of the surface of a material to carry out and/or accelerate certain chemical reactions. To date, numerous chemical reactions have been studied, which are potentially useful in energy generation and environmental cleaning applications. Photocatalysis is known to be able to produce thermodynamically uphill reactions, which otherwise need intense energy inputs in terms of high temperature (or pressure). Understanding the mechanism of photocatalytic reactions is critically important to design and develop new photocatalytic materials. In this section, a brief review of mechanism and thermodynamics of most common photocatalytic reactions is presented. Figure 2 shows the reduction and oxidation levels of some of the common photocatalytic reactions with reference to vacuum and the normal hydrogen electrode (NHE). It is noted that these values provide an insight only on the thermodynamic feasibility of the reaction. It is seen that for the reduction reaction, the energy of the (photoexcited) electron should be higher (on the absolute vacuum scale) than the redox level. Therefore the CB potential of the photocatalyst should be located at a higher energy value than the reduction reaction of interest.

Photocatalytic Water Splitting
One of the most studied reactions is the direct splitting of water into hydrogen and oxygen. Figure 3 shows the schematics of the water splitting reaction according to the 4-photon model [29]. In the water splitting reaction, upon the radiation of photon with suitable wavelength, photoexcited pairs of electrons and holes are generated within a photocatalyst. Typically, electron-hole separation takes place, due to surface charge or co-catalyst loading. Direct oxidation of water molecules, chemisorbed on the surface of the photocatalyst (or co-catalyst) occurs, by the interaction of water molecule and hole in the valence band (VB) of the photocatalyst. This reaction results in liberation of an oxygen molecule and 4-protons. The protons then migrate to the sites of photoexcited electrons to form hydrogen molecules (Equations (1)-(3)).
2H2O → 2H2 + O2 ∆H = + 234 kJ/mol (1) Evolution of hydrogen and oxygen using sunlight is considered as one of the most promising ways to generate hydrogen as a clean and renewable fuel. Like the water molecule, other molecules also undergo decomposition by the process of photocatalysis.

Photooxidation of Organic Molecules
Several organic compounds undergo photooxidation reactions, where a direct oxidation via photogenerated holes occurs or an indirect oxidation via hydroxyl ions takes place [31]. The degradation of organic molecules also takes place by reactive oxygen species ( Figure 4). Organic dyes, aliphatics and aromatic hydrocarbons, and organic acids can be mineralized to CO2 and H2O by photocatalytic processes. Like organic compounds, hydroxyl ions and reactive oxygen species (ROS) are known to inactivate microorganisms by degrading their cell walls [10,32]. The photocatalytic inactivation of microbes is effective in antimicrobial, antifungal and antiviral applications. A later section reviews certain silver-and bismuth-based perovskites which display particularly efficient antimicrobial action under visible light.

Photocatalytic Conversion of CO2 to Fuels
CO2, with a standard enthalpy of formation of −393.5 kJ·mol −1 at 298 K, is one of the most stable molecules. With appropriate adsorption and photocatalytic processes, reduction of CO2 in presence of water could be performed to produce hydrocarbons ( Figure 5). Possible chemical reactions of adsorbed CO2 and protons are presented by the following equations (Equations (4)- (7)). It could be seen that different number of protons in the reactants, results in different hydrocarbons as products. Figure 5. Schematics of CO2 photoreduction reaction over a photocatalyst surface [34].

Photocatalytic Nitrogen Fixation
Like CO2 reduction, atmospheric nitrogen could be reduced to ammonia or nitrates by the photocatalytic processes. The mechanism of nitrogen reduction is similar to that of CO2, where chemically adsorbed nitrogen molecules react catalytically with protons and form compounds of nitrogen and hydrogen (Equations (8)-(10)): H2O (hv/TiO2) → 2H + + 1/2O2 + 2e − N2 + H·→ N2H (10) The photocatalytic reduction of nitrogen is extremely useful in nitrogen photofixation processes for agricultural applications [35][36][37]. Although the process of photocatalytic nitrogen fixation is promising, efforts in this area have been severely limited. It is noted that the mechanism of the photocatalytic processes presented above is a simplified understanding, while the photocatalytic processes are complex in nature.
It is known that a given chemical reaction has a specific photooxidation or photoreduction level (potential) and thus the band potentials of the photocatalyst must satisfy the thermodynamic conditions. Intrinsic properties such as band gap (optical absorption) and band edge potentials determine the thermodynamic feasibility of photoinduced reactions under light irradiation. Apart from the basic conditions, there are several factors which affect the photocatalytic performance of the material system under consideration. Properties such as electron and hole effective mass, exciton lifetime and diffusion length, exciton binding energy affect the electron-hole separation and transport within the lattice. These properties are known to strongly influence the performance (kinetics/efficiency) of the photocatalytic reactions. Defects in the lattice, defect-induced energy states, localization of electrons on specific defect sites could determine the fate of the photoexited electron-hole pair. Finally, the electron transfer across semiconductor-electrolyte interface is significantly affected by surface states, surface band structure (depletion region induced electric field), and band bending. Such electronic properties of materials could be altered to suit specific photocatalytic applications. To date, numerous material systems have been evolved through systematic efforts of understanding and improving the electronic properties of materials. Among these materials perovskites have shown excellent promise for efficient photocatalysis under visible light irradiation, on account of their unique crystal structure and electronic properties. The perovskite crystal structure offers an excellent framework to tune the band gap values to enable visible light absorption and band edge potentials to suit the needs of specific photocatalytic reactions. Further, lattice distortion in perovskite compounds strongly influences the separation of photogenerated charge carriers. The following sections present some groups of materials that have shown visible light activity.

Titanate Perovskites
Titanate perovskites have been studied for photocatalytic applications for a long time. Most of the titanate perovskites have band gap energy (Eg) value more than 3.0 eV, however they show excellent photocatalytic properties under UV radiation [1]. Using these titanates as host materials, doping is widely used to alter the optical properties and induce visible light absorption. TiO2 (anatase) has a band gap of 3.2-3.4 eV and its CB potential is −0.3 to −0.6 eV above the water reduction level [38]. Certain perovskite titanates have CB energies more negative than TiO2, making them more suitable candidates for hydrogen generation. Titanates also offer good photostability and corrosion resistance in aqueous solutions. In this section, a detailed review of MTiO3 (M = Sr, Ba, Ca, Mn, Co, Fe, Pb, Cd, Ni) systems is presented. Figure 6 gives an overview of elements that form perovskite titanates.

SrTiO3
SrTiO3 is a simple cubic (Pm3m, a = 3.9 Å) perovskite with an indirect band gap of 3.25 eV [39]. When loaded with a co-catalyst such as Rh or NiOx, SrTiO3 shows stoichiometric water splitting under UV radiation [40] and has been studied extensively for developing visible light water splitting catalysts. Doping of the Ti site with Mn, Ru, Rh, and Ir was of significant interest in early days [41]. It is found that these dopants induce mid-gap states in the band gap allowing visible light absorption [42]. Mn and Ru doping are found useful for O2 evolution, while dopants like Ru, Rh, and Ir are suitable for H2 evolution [41]. Rh-doped SrTiO3 thin films also shows cathodic photocurrent from overall water splitting under visible light, where 7% Rh doped SrTiO3 showed 0.18% incident photo-to-electron conversion efficiency (IPEC) under 420 nm irradiation [43]. Using Rh-doped SrTiO3 as a H2 evolving photocatalyst, various Z scheme systems have been developed. In a significant demonstration, a novel electron mediator [Co(bpy)3] 3+/2+ was used for Rh-doped SrTiO3 with BiVO4 photocatalyst. Such a system showed a solar energy conversion efficiency of 0.06% under daylight [44]. Efforts in Z scheme photocatalysis have also been targeted towards eliminating the need for electron mediators by preparing composite photocatalysts. In such systems, electrons from an O2 evolving photocatalyst recombine with holes from a H2 evolving photocatalyst at the interface of the composite. The quality of the interface and the band alignment of the two semiconductors are important factors for the successful realization of mediator-free type Z schemes.
Rh-doped SrTiO3 was combined with several O2-evolving photocatalysts such as BiVO4, AgNbO3, Bi2MoO6, WO3, or Cr/Sb-doped TiO2 [45]. In those experiments the authors found that agglomeration of the photocatalyst particles occur under acidic conditions, which results in Z scheme photocatalysis. A combination of Rh-doped SrTiO3 and BiVO4 resulted in the best yield [45]. A schematic of microstructure and mechanism of water splitting of an agglomerated Z scheme photocatalyst is shown in Figure 7. In a recent effort, a composite of 1% Rh-doped SrTiO3 loaded with 0.7% Ru and BiVO4 was successfully prepared. Such a composite showed a stoichiometric water splitting reaction (pH 7) with quantum yield (QY) of 1.6% at 420 nm [23]. These studies successfully establish the feasibility of the "Z scheme" photocatalysis for candidates such as Rh-doped SrTiO3 under visible light. Further, the water splitting efficiency is dependent on the synthesis method used for Rh-doped SrTiO3 [46]. The use of excess Sr in hydrothermal and complex polymerization methods proved useful for improving the apparent yield to 4.2% under 420 nm radiation [46]. Apart from mono-doping, co-doping has been employed in SrTiO3 to pursue visible light driven photocatalysis. Co-doping of Sb (1%) and Rh (0.5%) was found useful for visible light photocatalysis and estimated H2 and O2 evolution rates for 1 m 2 surface area were 26 mL·h −1 and 13 mL·h −1 , respectively [47]. Further, a composite system was prepared from co-doped La-Cr in SrTiO3 and co-doped La-Cr Sr2TiO4 which showed visible light driven H2 evolution (24 µmol·h −1 ·g −1 ) [48]. Composite preparation led to heterojunctions between doped phases and produced a synergistic effect for hydrogen evolution. Further, a solid solution of AgNbO3 and SrTiO3 was discovered to be a visible light active photocatalyst [49]. (AgNbO3)0.75(SrTiO3)0.25 showed promising performance for O2 evolution and isopropanol (IPA) degradation under visible light.
Several efforts have been made to understand and design SrTiO3-based photocatalysts. Particularly DFT-based band structure calculations provide useful insights into the electronic structure and its correlation with photocatalytic activity. It is shown that Rh doping in SrTiO3 produces band-like states above the valence band maximum (VBM) which are responsible for the visible light absorption. The proximity of dopant-induced states to VBM helps efficient replenishment of electrons and suppresses electron trapping from CB [42]. Theoretical calculations indicate that a TiO2-terminated SrTiO3 surface with defects such as O and Sr vacancies would alter its electronic structure and induce visible light absorption peaking around 420 nm [50]. Such strategies could be useful in the development of low dimensional materials. Further, theoretical work on doped SrTiO3 compounds show that certain dopants such as La strongly lower the effective mass of electrons and holes (near the valence band region), increasing the mobility of the photoexited carrier [51]. Along with the ground state band structure calculations, study of electron and hole masses, defect chemistry, photoexcited transport could be carried out to understand this system in detail. Understanding the excited state properties is useful in further development of new photocatalysts.

BaTiO3
Like SrTiO3, Rh doping in BaTiO3 has been carried out and a quantum yield of 0.5% under 420 nm was reported [52]. Being a hydrogen evolving catalyst, this material has also been used as Z scheme with Pt/WO3 for overall water splitting [52].

CaTiO3
Calcium titanate is one of the common perovskite minerals with a band gap of 3.6 eV. Cu doping in CaTiO3 is widely studied and visible light-driven photocatalytic water decomposition has been reported [53]. Cu doping not only induces visible light absorption, but also enhances hydrogen evolution under UV radiation when NiOx co-catalyst is used. Studies of such doped systems, where dopants enhance the photocatalytic activity host materials are important to design efficient photocatalysts. Detrimental effects of doping on the properties such as electron-hole recombination, electron/hole effective mass, and reduced crystallinity should be studied and reported in detail. These studies are useful to gain insights on the photocatalytic activities of the doped systems. Co-doping of Ag and La at CaTiO3 has been done to narrow the band gap and induce visible light absorption [54]. DFT studies also indicate that like SrTiO3, TiO2-terminated CaTiO3 surfaces possess the capability of visible light absorption [55]. Along with alkali titanates, several transition metal titanates show promise for visible light photocatalysis. Figure 8 shows the empirically estimated band diagram of the MTiO3 systems with respect to water oxidation and reduction levels. Alkali metal titanates such as Ba, Ca, and Sr have enough CB potential for hydrogen evolution. However, certain transition metal titanates do not possess the desired CB potential for water reduction, though they have band gap values in the visible region (such as Co, Ni, Fe etc.). These materials could be suitable for degradation of organics or other photooxidation reactions.

CoTiO3
CoTiO3 has a band gap in the visible region (Eg 2.28 eV). Recently this compound has been investigated for photocatalytic O2 evolution reaction without any co-catalyst [57]. A yield of 64 μmol·g −1 ·h −1 was obtained under visible light. CoTiO3 and TiO2 composites have been studied for 2-propanol mineralization, however it is worth noting that the band positions of CoTiO3 are not suitable for transferring photoexcited charges to TiO2, offering no significant advantage [58].

NiTiO3
NiTiO3 has a reported band gap of around 2.16 eV and its light absorption spectra show peaks in visible region corresponding to crystal field splitting [59]. NiTiO3 nanorods have been employed for degradation of nitrobenzene under visible light [59].

FeTiO3
FeTiO3 has a band gap of 2.8 eV and thus it absorbs visible light. Composites of FeTiO3 and TiO2 are studied for the degradation of 2-propenol under visible light. In such composites, TiO2 acts as hole capturing phase, thereby separating the electron-hole recombination [60].

PbTiO3
PbTiO3 has a band gap of 2.75 eV and has been investigated for visible light photocatalysis. Core-shell particles of nano-TiO2 (shell) and micro-PbTiO3 (core) were studied for effective charge separation and photocatalytic performance [62]. It is worth noting that PbTiO3 is ferroelectric in nature while TiO2 is dielectric. It is shown that the ferroelectric behavior helps electron-hole separation at the interface of two particles.

MnTiO3
Further, F doped MnTiO3 shows improved separation of charges and degrades rhodamine B under visible light [63]. Some of the promising titanates photocatalysts have been listed in Table 1, as a part of compilation of perovskite systems.

Tantalate Perovskites
Alkali tantalates are particularly known for efficient overall water splitting reaction under UV irradiation as they possess both VB and CB potentials suitable for water splitting reaction [72][73][74]. To enable visible light photocatalysis, doping of various elements has been studied to achieve visible light activity.

NaTaO3
Our group reported a detailed study on Bi-doped NaTaO3 and showed that the bismuth doping site significantly affects the photocatalytic activity for hydrogen evolution [75,76]. Further, co-doping of La-Co, La-Cr, La-Ir, La-Fe in NaTaO3 have shown successful visible light absorption and subsequent hydrogen evolution [77][78][79][80][81]. Co-doping of La-N in NaTaO3 has been studied for hydrogen evolution by Zhao et al. [82]. These studies have indicated that both anion and cation doping in NaTaO3 is useful for visible light photocatalytic applications. Among the doped NaTaO3 systems, computational studies on the anionic (N, F, P, Cl, S) doping were reported by Han et al. which shows that certain anions like N and P may be useful for visible light absorption [83]. Additionally, doping of magnetic cations such as Mn, Fe, and Co in NaTaO3 has also been studied using DFT-PBE [84]. Recently, our group studied DFT calculations of a number of doped NaTaO3-based photocatalysts by PBE0 hybrid calculations ( Figure 9) [85]. Further, anion doping was also explored in detail using (HSE06) hybrid DFT calculations, where N, P, C, and S doping at O sites were studied. The study also reports the thermodynamics and effect of coupling between N-N, C-S, and P-P on the optical and electronic properties [86]. DFT studies are useful in explaining the properties of existing materials systems and designing new materials. Particularly, use of hybrid functional such as PBE0 or HSE06, is able to accurately define the valence band structure and location of bands or energy states that are crucially important for visible light driven photocatalysis. Hybrid DFT calculations could be useful in predictive modeling, where, band gaps and band edge potential of useful doped photocatalysts are identified. An example of doped tantalate systems is shown in Figure 9.

AgTaO3
AgTaO3 exhibits similar behavior to alkali tantalates, however, it has a smaller band gap value of 3.4 eV. AgTaO3 doped with 30% Nb absorbs visible radiation and shows a stoichiometric overall water splitting reaction under visible light when loaded with NiO co-catalyst [87]. Co-doping of N-H and N-F in AgTaO3 has been studied in detail. The study indicates that co-doping not only balances the charges but also improves the carrier mobility. N-F co-doped AgTaO3 has an effective band gap value of 2.9 eV and shows H2 generation under visible light [88].

KTaO3
KTaO3 (Eg 3.6 eV) photocatalysts have been studied for water splitting under UV radiation. However, work on development of visible light driven KTaO3 based photocatalysts is limited.

KNbO3 and NaNbO3
Both KNbO3 (Eg 3.14 eV) and NaNbO3 (Eg 3.08 eV) have band gap values in the UV-responsive region, however suitable modifications of the band structure have resulted in visible light photocatalysis [89]. N-doped NaNbO3 is a known visible light photocatalyst for the degradation of 2-propanol [90]. Nitrogen doping in KNbO3 has been studied for water splitting as well as organic pollutant degradation [91]. First principles calculations predict that co-doping of La and Bi would induce visible light response in NaNbO3 [92]. Recent work on ferroelectric perovskites of KNbO3-BaNiNbO3 shows that the solid solution of these compounds could absorb six times more light and shows fifty times more photocurrent than others [93]. Although photocatalytic properties are not known, this is an attractive candidate for visible light driven photocatalyst.

AgNbO3
Replacing an A site alkali metal by silver reduces the band gap of the perovskite and AgNbO3 has a band gap of around 2.7 eV. Studies have shown that the photocatalytic activity of AgNbO3 strongly depends on the shape of the particles: polyhedron-shaped particles are favorable for O2 evolution reactions [94]. Further, La doping was found to enhanced the hotocatalytic performance by 12-fold for gaseous 2-propenol degradation [95].

AgVO3
AgVO3 exists in two types of crystal structures, viz. α-AgVO3 (Eg 2.5 eV) and β-AgVO3 (Eg 2.3 eV) [96]. Both phases are photocatalytically active. However, β-AgVO3 shows better photocatalytic performance than the α-phase. The CB potential of AgVO3 is not sufficient for H2 evolution, but it is suitable for the degradation of volatile organic compounds (VOCs) and O2 evolution. β-AgVO3 nanowires show excellent photocatalytic performance in the degradation of Rh B [97]. Composites of AgBr-AgVO3 were reported to display respectable efficiency for Rh B degradation [98], while Ag-loaded AgVO3 has shown good performance for degradation of bisphenol [99].

CuNbO3
CuNbO3 crystallizes in the monoclinic structure and has a band gap of 2.0 eV. It is an intrinsic p-type semiconductor and has shown 5% efficiency for photon to electron conversion when used as a photocathode. Being a stable material under irradiation, more investigations should be carried out on this material [100]. Tantalum, niobium, and vanadium belong to the same group in the periodic table. Perovskite compounds of these elements show decreasing band gap and CB potential values. This trend is attributed to the 3 d, 4 d and 5 d orbital energies in V, Nb and Ta, respectively.

Ferrite Perovskites
Most of the ferrite perovskites have their native band gaps in the visible region. Hematite and other iron oxide compounds have known shortcomings such as short exciton diffusion length, low electron conductivity, and lower conduction band edge potential [101]. However certain ferrite-based perovskites have shown good photocatalytic activities, circumventing the shortcomings seen in binary iron oxides.

LaFeO3
LaFeO3 (Eg 2.1 eV) has been explored for degradation of pollutants as well as hydrogen evolution under visible light. Sol-gel synthesized LaFeO3 loaded with Pt co-catalyst showed high yield of hydrogen evolution (3,315 μmol·h −1 ·g −1 , in the presence of ethanol) under 400 W tungsten light source [68]. Another study on this phase demonstrates high yield of H2 and O2 (1290 μmol and 640 μmol after three hours, respectively), without any co-catalyst loading [102]. Further, Thirumalairajan et al. showed shape dependent photocatalytic activity of LaFeO3 for Rh B dye under visible light (>400 nm) [103]. Doping of Mn in LaFeO3 has also been studied and it shows higher photocatalytic activity [104]. Lanthanum ferrite has demonstrated excellent photocatalytic activity under visible irradiation; however, studies on the fundamental photophysical properties are lacking the literature. Understanding the reasons why LaFeO3 is a better photocatalyst that Fe2O3, in terms of comparative electronic properties such as electron-hole separation, mobility, photoexcited lifetimes etc. is important for further development in ferrite-based photocatalyst.

BiFeO3
BiFeO3 is a known ferroelectric material with a band gap of 2.3 eV. Recent studies have shown that BiFeO3 could be used as a visible light photocatalyst [105]. The ferroelectric properties of BiFeO3 could be utilized to enhance the electron-hole separation and improve the photocatalytic activity. Figure 10 shows the effect of polarization on the band bending of BiFeO3 particles. Band bending affects the separation of electrons and holes in the space charge region and thus affects the photocatalytic activity. Gd-doped BiFeO3 show enhanced photocatalytic degradation for rhodamine B degradation attributed to its ferromagnetic behavior [106]. Ca doping in BiFeO3 leads to improved performance for photocatalytic degradation of Congo Red dye [107].
In another study, cations such as Y, Mg and Al were doped in BiFeO3 and their photocatalytic performance was evaluated by degradation of Rh B under 400 nm radiation. The degradation performance was limited to C/C0 of 0.8 within three hours [108]. The study of the effects of ferroelectric behavior on photocatalytic performance is a relatively new topic and it generally shows impressive activity for the degradation of organic pollutants, however, more experimental evidence and understanding are needed to establish a correlation between ferroelectric behavior and photocatalytic activity. Further studies on the stability and toxicity of bismuth-based materials should be carried out for realizing their practical applications. It is further noted that the tilting of octahedra in the perovskite crystal structure significantly affects its electronic properties. Nevertheless, the effect of tilting of octahedra on the photophysical properties such as electron-hole separation, electron transport, delocalization of charges has not been studied in detail. Such studies will prove useful in establishing the importance of perovskites in the field of photocatalysis. Figure 10. Schematics of changes to band diagram upon polarization of Gd doped BiFeO3 due to ferroelectricity [106].

GaFeO3
GaFeO3 has been reported to show overall water splitting without any co-catalyst loading under visible light (λ > 395 nm) [109]. The authors also reported a yield of 0.10 and 0.04 µmol·h −1 under a 450 nm band pass filter. It is worth noting that the catalytic activity decreased due to deactivation of the catalyst.

YFeO3
Among the other ferrites, YFeO3 has a band gap value of 2.43 eV and showed photocatalytic activity four times that of TiO2-P25 under > 400 nm visible light radiation (Rh B degradation) [110] 3.4.5. PrFeO3 PrFeO3 was evaluated for hydrogen evolution reaction from ethanol-water mixture and showed a yield of 2847 μmol·g −1 ·h −1 under 200W tungsten lamp irradiation [69].

AlFeO3
Composites of AlFeO3 and TiO2 are reported to yield the sunlight driven photocatalytic degradation of methyl orange (MO) and eosin dye [111]. Ferrite based perovskites also offer advantage of magnetic recovery of the particles which is useful in practical applications.

Other Perovskite Systems
Among the other perovskite photocatalysts, compounds of bismuth, cobalt, nickel, and antimony (occupying B sites) have band gap values in the visible region. Pentavalent bismuth perovskites are known to be active photocatalysts under visible light radiation. Perovskites such as LiBiO3 (Eg 1.63 eV), NaBiO3 (Eg 2.53 eV), KBiO3 (Eg 2.04 eV), and AgBiO3 (Eg 2.5 eV) have all been investigated for degradation of organic pollutants [112]. NaBiO3 shows better photocatalytic performance for phenol and methylene blue (MB) degradation as compared to other Bi 5+ containing perovskites as well as P25-TiO2 [112]. NaBiO3 is also reported to have higher photocatalytic activity than N-doped TiO2 photocatalyst [113]. NaBiO3 and BiOCl composites have been studied for enhanced electron-hole separation and subsequent photocatalysis [114]. AgBiO3 was shown to be effective in restricting the growth of Microcystis aeruginosa under simulated solar light [115]. This study shows that AgBiO3 with band gap energy around 2.5 eV could be a potential algaecide under natural light. DFT studies show that NaBiO3 has a strong band dispersion arising from Na 3s and O 2p hybridized orbitals, which contributes to the higher photocatalytic activity [112]. LaCoO3 has a band gap value of 2.7 eV and oxygen deficient LaCoO3 (LaCoO3-δ) has been studied for MO degradation (>400 nm) [116].
Other compounds such as LaNiO3 has been studied for MO degradation using wavelengths greater than 400 nm [117]. Cu doping in LaNiO3 has been studied for improved H2 evolution [118]. Ilmenite type AgSbO3 has an absorption edge onset at 480 nm, and it has been demonstrated for O2 evolution under sacrificial agent as well as degradation of MB, RhB, and 4-chlorophenol [119]. Further, solid solution of LaCrO3 and Na0.5La0.5TiO3 was developed for hydrogen evolution [120]. Recently, Gd2Ti2O7/GdCrO3 composite was reported as photocatalyst p-n junction photocatalyst [71]. The study shows that GdCrO3 has a band gap of 2.5 eV and is responsible for the visible light absorption [72].

Double Perovskites
Compounds with general formula A2B2O6 belong to the double perovskites and they have similar crystal structures to simple perovskites. Double perovskites have the basic framework of corner connected BO6 octahedra and A cations enclosed within, however, the connectivity of the octahedra may differ from structure to structure. Double perovskites could accommodate different cations at the A or/and B sites, taking a general form AA'BB'O6. Accommodation of different cations at the A and B sites alters the photophysical properties of the compound significantly. Among the binary oxides, only a few compounds are known to have band gap values in the visible region (narrow gap), e.g., Fe2O3, WO6, CuO, Bi2O3. However, these materials suffer from insufficient CB potential for hydrogen evolution. Some of the materials also suffer from poor stability and low mobility of photoexcited charges. On the other hand, many of the binary oxides are known to be efficient photocatalysts, however only activated by UV radiation (wide gap). Complex compounds offer a possibility of combining the elements from 'narrow gap' and 'wide gap' binary compounds to exploit the properties of both types of oxides, and thus are potentially useful as visible light photocatalysts. The following section reviews photocatalytic properties of double perovskites which are visible light active.

Sr2FeNbO6
Sr2FeNbO6 has a cubic crystal structure and a band gap of 2.06 eV. In its pristine form, it is a known photocatalyst, while 7% Ti doping at Fe site has shown two time enhancement of the hydrogen generation in methanol solution. A total of 28.5 µmol·h −1 and 650 µmol·h −1 were reported in presence of sacrificial agents and with 0.2% Pt as co-catalyst [121]. W doping in Sr2FeNbO6 has also been studied and demonstrated enhancement to hydrogen evolution under visible radiation [122].

La2FeTiO6
Hu et al. reported higher photocatalytic activity for La2FeTiO6 than for LaFeO3 for degradation of p-chlorophenol under visible light irradiation [123].

Other Double Perovskites
Rare earth and bismuth-based double perovskites were studied for visible light photocatalysis. Compounds with general formula Ba2RBiO6, (R = La, Ce, Pr, Nd, Sm, Eu, Gd, Dy), were prepared and their photocatalytic activity was studied for MB degradation [124]. Authors found rare earth cation dependent photocatalytic performance, where compounds such as Ba2EuBiO6, Ba2SmBiO6, and Ba2CeBiO6 showed high photocatalytic activity. Complex perovskites such as CaCu3Ti4O12 have also been studied for their photocatalytic performance. CaCu3Ti4O12 was found to possess an indirect band gap of 1.27 eV and Pt-loaded CaCu3Ti4O12 shows degradation of MO under radiation greater than 420 nm [125]. Photophysical properties of certain double perovskite compounds have been reported, however, more efforts are needed to investigate photocatalytic properties of these materials. Compounds such as Ba2CoWO6, Ba2NiWO6, Sr2CoWO6 and Sr2NiWO6 are stable perovskites compounds for O2 evolution using sacrificial agents [126].
Certain tantalum-based compounds have been studied for degradation of organic compounds. N-doped K2Ta2O6 is known to absorb visible light up to 600 nm and degrade formaldehyde under visible light [127]. Our group demonstrated that Bi doping in Na2Ta2O6 causes visible light absorption and degradation of MB [128]. Although some work has been done with double perovskites as photocatalysts, understanding of their fundamental properties is limited. More work is needed to discover their advantages as visible light photocatalyst and develop novel materials systems.

Mixed Oxides
Mixture of oxides and nitrides or oxides and sulphides has been developed to engineer the band structure of the oxide photocatalysts suitable for the visible light absorption. Oxynitride and oxysulphide photocatalysts offer distinctive advantages over their doped counterparts. Earlier reports on oxynitrides have revealed that replacing oxygen by nitrogen at lattice sites, in a stoichiometric manner, narrows the band gap of the oxide, by pushing the VBM into the band gap [129]. Such a modification does not produce native point defects, which would otherwise be introduced in the case of N doping. Stoichiometric incorporation of nitrogen also avoids the localized states induced by N doping and reduces the possible electron-hole recombination. Similar composition containing mixtures of sulfur and oxygen i.e., oxysulfides, has been developed for photocatalysts. Mixed oxysulfide perovskites of Sm2Ti2S2O5 (Eg 2.0 eV) is known for water oxidation and reduction reaction for oxygen and hydrogen evolution, respectively in presence of sacrificial agents under low photon energy wavelengths of 650 nm. The band structure of this phase reveals that the presence of sulphur narrows the band gap and enables visible light absorption [130].
Oxynitride compounds such as CaNbO2N, SrNbO2N, BaNbO2N, and LaNbON2 belong to the perovskite type crystal structures [131]. Photocatalytic hydrogen evolution has been reported under visible light from methanol solution. SrNbO2N (Eg 1.8 eV) has been investigated in detail, where the photoelectrode of SrNbO2N on a transparent conducting surface shows water oxidation reaction under no external bias [132]. Tantalum counterparts of these compounds were developed and utilized in the Z scheme photocatalysis. Compounds such as CaTaO2N and BaTaO2N were loaded with Pt co-catalyst and coupled with pt/WO3 for Z scheme water splitting [133]. A solid solution of BaTaO2N and BaZrO3 was formulated for hydrogen and oxygen evolution, which showed improved performance compared with the individual photocatalysts under visible light [134]. LaTiOxNy is another perovskite type compound which shows high photocurrent density under visible light [135].
Apart from the double perovskites belonging to the general formula AA'BB'O6, there are a several other compounds that show crystal structures close to the perovskite type structure, however such compounds are not included in the current review. Theoretically, double perovskites offer a wider scope to design photocatalysts by selecting suitable cations and AA' and BB' sites in the lattice. Work on design and development of double perovskite is currently limited and synthesis and characterization of new materials in this category are needed.

Summary and Outlook
A large number of perovskite-based compounds (over 80) have been studied for visible light driven photocatalytic applications. Perovskite structures offer abundant scope in designing novel compounds based on A and B site occupancy, which gives rise to a wide range of materials systems with unique properties. Among these compounds, photocatalysts with pristine band gaps in the visible region such as LaFeO3, PrFeO3, NaBiO3, and AgBiO3 (Table 2) show promising photocatalytic performance under visible radiation (>400 nm). Although there are reports on the photocatalytic activity of these compounds, detailed studies on these materials are limited. More efforts are needed to understand the structure-property relations in such ferrites and bismuth based compounds and improve their photocatalytic activity. Among the wide band gap semiconductors, SrTiO3-and NaTaO3-based photocatalysts are the most investigated systems. The strategy of doping foreign elements in wide band gap photocatalysts is widely used to induce visible light absorption and subsequently to enable photocatalytic activity. Nonetheless, very limited knowledge is available on the adverse effect of dopants on the photophysical properties of the compounds (e.g., if the benefit derived from dopant-induced visible light activity is overweighed by a loss of UV light activity). This should be properly investigated in the future. Appropriate dopants which retain the beneficial properties of the host materials while inducing visible light responses should be identified. Research work on complex perovskites show that these compounds offer distinct advantages as compared with simple perovskites. Designing complex perovskite compounds with suitable elements at A and B sites to yield desired photocatalytic properties is a challenge. We expect computational design would help shorten the selection process. Recent advances in computational tools such as DFT based band structure calculations are highly effective to design and understand novel materials systems. 2.50 Inhibition of Microcystis [115] As certain perovskite compounds exhibit ferroelectric, ferromagnetic, or piezoelectric effects, there is a need to understand the correlation between these effects and the photocatalytic activity to a greater depth. Such studies would certainly be helpful in the development of efficient visible light photocatalysts. On a final note, there has been significant progress in the development of visible light perovskites in the past years. This development has laid a good foundation for future work in this area. Further understanding of the crystal and electronic structural factors behind photocatalytic activity is needed for the future development of efficient visible light-driven perovskites.